Handbook of Electoral System Choice

Handbook of Electoral System Choice Edited by Josep M. Colomer Foreword by Bernard Grofman

Handbook of Electoral System Choice

Also by Josep M. Colomer GAME THEORY AND THE TRANSITION TO DEMOCRACY: The Spanish Model POLITICAL INSTITUTIONS: Democracy and Social Choice POLITICAL INSTITUTIONS IN EUROPE (editor) STRATEGIC TRANSITIONS: Game Theory and Democratization

Handbook of Electoral System Choice Edited by

Josep M. Colomer Research Professor in Political Science, Higher Council of Scientific Research, Barcelona, Spain

Foreword by

Bernard Grofman Professor of Political Science and Mathematical Behavioral Science, University of California, Irvine, USA

Chapter 6 © Alberto Díaz-Cayeros and Beatriz Magaloni 2004 Chapter 16 © Patrick Dunleavy and Helen Margetts 2004 Remaining chapters © Palgrave Macmillan Ltd 2004 All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No paragraph of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, 90 Tottenham Court Road, London W1T 4LP. Any person who does any unauthorised act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The author has asserted his right to be identified as the author of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2004 by PALGRAVE MACMILLAN Houndmills, Basingstoke, Hampshire RG21 6XS and 175 Fifth Avenue, New York, N.Y. 10010 Companies and representatives throughout the world PALGRAVE MACMILLAN is the global academic imprint of the Palgrave Macmillan division of St. Martin’s Press, LLC and of Palgrave Macmillan Ltd. Macmillan® is a registered trademark in the United States, United Kingdom and other countries. Palgrave is a registered trademark in the European Union and other countries. ISBN 1–4039–0454–5 This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data Handbook of electoral system choice / edited by Josep M. Colomer; foreword by Bernard Grofman. p. cm. Includes bibliographical references and index. ISBN 1–4039–0454–5 (cloth) 1. Elections. 2. Comparative government. I. Colomer, Josep Maria. JF1001.H26 2004 324.6′3—dc22 10 13

9 12

8 11

7 10

2003062675 6 09

5 08

4 07

3 06

2 05

1 04

Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham and Eastbourne

Contents

viii

List of Tables

x

Acknowledgements

xi

Foreword by Bernard Grofman

xxi

List of Contributors

Part 1 Introduction 1 The Strategy and History of Electoral System Choice Josep M. Colomer

3

Part 2 The Americas 2 The Americas: General Overview Josep M. Colomer

81

3 Argentina: Compromising on a Qualified Plurality System Gabriel L. Negretto

110

4 Brazil: Democratizing with Majority Runoff Jairo M. Nicolau

121

5 Costa Rica: Modifying Majoritarianism with 40 per cent Threshold Fabrice Lehoucq 6 Mexico: Designing Electoral Rules by a Dominant Party Alberto Díaz-Cayeros and Beatriz Magaloni

133 145

7 The United States: the Past – Moving from Diversity to Uniform Single-Member Districts Erik J. Engstrom

155

8 The United States: the Future – Reconsidering Single-Member Districts and the Electoral College Richard L. Engstrom

164

Part 3 Western Europe 9 Western Europe: General Overview Josep M. Colomer

179

v

vi

Handbook of Electoral System Choice

10 France: Reform-mongering Between Majority Runoff and Proportionality Gerard Alexander

209

11 Germany: Partisan Engineering of Personalized Proportional Representation Marcus Kreuzer

222

12 Italy: Lofty Ambitions and Unintended Consequences Diego Gambetta and Steven Warner

237

13 Spain: from Civil War to Proportional Representation Josep M. Colomer

253

14 Sweden: Introducing Proportional Representation from Above Leif Lewin

265

15 Switzerland: Introducing Proportional Representation from Below Georg Lutz

279

16 The United Kingdom: Reforming the Westminster Model Patrick Dunleavy and Helen Margetts

294

Part 4 Eastern Europe 17 Eastern Europe: General Overview Carlos Flores Juberías

309

18 The Baltics: Independence with Divergent Electoral Systems Evald Mikkel and Vello Pettai

332

19 The Czech Republic: Entrenching Proportional Representation Petr Kopecký

347

20 Hungary: Compromising Midway on a Mixed System John W. Schiemann

359

21 Poland: Learning to Manipulate Electoral Rules Marek M. Kaminski and Monika A. Nalepa

369

22 Russia: Compromising a Long-Lasting Transitional Formula Olga V. Shvetsova

382

Part 5 Africa 23 Africa: Dictatorial and Democratic Electoral Systems since 1946 Matt Golder and Leonard Wantchekon 24 Africa: Electoral Systems in Emerging Democracies Shaheen Mozaffar

401 419

Contents vii

25 South Africa: Proportional Representation in the Puzzle to Stabilize Democracy Andrew Reynolds

440

Part 6 Asia and the Pacific 26 Asia and the Pacific: General Overview Allen Hicken 27 Australia: Replacing Plurality Rule with Majority-Preferential Voting Marian Sawer 28 India: Majoritarian Democracy from Above Subrata K. Mitra

453

475 487

29 Indonesia: Transition and Change but Electoral System Continuity Andrew Ellis

497

30 Japan: Manipulating Multi-Member Districts – from SNTV to a Mixed System Junichiro Wada

512

31 New Zealand: Reform by (Nearly) Immaculate Design Jack H. Nagel

530

Appendix: Notes and Sources for Summary Tables

544

Glossary and Index

547

List of Tables

Summary tables 1A 1B 2A 2B 3A 3B 4A 4B 5A 5B 6A 6B

Assembly electoral systems Presidential electoral systems The Americas: Assembly (lower or single chamber) The Americas: Presidency Western Europe: Assembly (lower or single chamber) Western Europe: Presidency Eastern Europe: Assembly (lower or single chamber) Eastern Europe: Presidency Africa: Assembly (lower or single chamber) Africa: Presidency Asia and the Pacific: Assembly (lower or single chamber) Asia and the Pacific: Presidency

74 77 95 105 193 208 325 331 436 439 467 474

Chapter tables 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 1.10 1.11 1.12 1.13 1.14 3.1 3.2 4.1 4.2 5.1 6.1

Majority rules and procedures Proportional representation formulas Major changes of assembly electoral system Changes of assembly electoral system Major changes of assembly electoral system in present democratic periods Changes of assembly electoral system in present democratic periods Changes of presidential electoral system Changes of presidential electoral system in present democratic periods Number of basic assembly electoral systems over time Number of assembly electoral systems over time Number of presidential electoral systems over time Effective number of parties and electoral system change Duration of assembly electoral systems Duration of presidential electoral systems Argentina: Electoral College presidential elections, 1854–1989 Argentina: presidential elections, 1983–99 Brazil: example of seat allocation Brazil: Chamber of Deputies electoral results, 1982–2002 Costa Rica: presidential selection, 1882–2002 Mexico: Chamber of Deputies, 1961–2000 viii

32 43 55 57 57 59 59 60 60 61 61 64 66 67 112 117 126 128 134 148

List of Tables ix

6.2 Mexico: presidency electoral results, 1964–2000 7.1 US presidential electors, from diversity to uniformity, 1788–1836 10.1 France: National Assembly electoral results, 1945–2002 11.1 Germany: bargaining rounds over the 1949 electoral law 11.2 Germany: Bundestag electoral results, 1949–2002 12.1 Italy: Chamber of Deputies electoral results, 1992–2002 13.1 Spain: electoral results, 1931–36 13.2 Spain: electoral results, 1977–2000 14.1 Sweden: transitional electoral results, 1905–14 15.1 Transitional elections in Switzerland, 1911–19 16.1 United Kingdom: key features of electoral systems enacted or proposed, 1997–98 16.2 United Kingdom: House of Commons electoral results, 1945–2001 18.1 Electoral performance in the Baltic states, 1992–2003 18.2 Estonia: State Council (Riigikogu) electoral results, 1992–2003 18.3 Latvia: Parliament (Saeima) electoral results, 1993–2002 18.4 Lithuania: Diet (Seimas) electoral results, 1992–2000 19.1 Czech Republic: Chamber of Representatives electoral results, 1992–2002 20.1 Hungary: National Assembly electoral results, 1990–2002 21.1 Changes in the Polish Sejm electoral laws, 1993–2001 21.2 Poland: simulated and actual seat shares and electoral reform 21.3 Poland: Sejm (Lower Chamber) electoral results, 1991–2001 22.1 Russian Assembly elections, 1993–99 23.1 Africa: dictatorial and democratic elections, 1946–2000 23.2 Africa: colonial legacy and democratic electoral systems 23.3 Africa: effective number of parties 24.1 Africa: patterns of electoral system choice in third-wave democracies 24.2 Africa: political consequences of electoral systems 25.1 South Africa: seat allocations under hypothetical PR thresholds 25.2 South Africa: ethnic diversity of the National Assembly members 25.3 South Africa: National Assembly electoral results, 1994–99 27.1 Australia: House of Representatives, 1990–2001 27.2 Australia: transitional electoral results, 1914–25 28.1 India: Lok Sabha electoral results, 1952–71, 1977–98 29.1 Indonesia: elections, 1955 29.2 Indonesia: electoral results and simulations, 1999 30.1 Japan: lower house and cabinets, 1987–2001 30.2 Japan: upper house and cabinets, 1987–2001 30.3 Japan: winners in single-member districts of mixed system 31.1 New Zealand: electoral systems and political consequences 31.2 New Zealand: electoral results, 1984–2002

151 157 219 229 234 250 258 259 276 285 296 304 342 343 344 345 353 367 374 376 379 392 402 409 412 422 429 446 447 449 484 485 489 509 510 524 526 528 538 541

Acknowledgements

The idea for this book was conceived several years ago. Preparatory work included the organization of more than a dozen panels at the American Political Science Association, the Public Choice Society annual meetings and the First General Conference of the European Consortium for Political Research. Among those who helped to organize the panels, published a few outlines, found the appropriate authors or facilitated data and sources, the following, at least, must be mentioned: Gerard Alexander, Klaus Armingeon, Carles Boix, Gary Cox, Alberto Díaz-Cayeros, Keith Dowding, Patrick Dunleavy, Carlos Flores, Bernard Grofman, Blanca Heredia, Colin Hughes, Marek Kaminski, Jan-Erik Lane, Michael Laver, Fabrice Lehoucq, Arend Lijphart, Georg Lutz, Beatriz Magaloni, Helen Margetts, Nicholas R. Miller, Shaheen Mozaffar, Jack H. Nagel, Thomas F. Remington, Carlos M. Rodríguez-Arechavaleta, Kenneth Shepsle, Rein Taagepera and Jakub Zielinski. Travels, computing tools and materials were financed by the Spanish Ministry of Science and Technology, SEC2000–1186. For the present edition, Palgrave Macmillan editors Alison Howson and Steven Kennedy were enthusiastic supporters and three anonymous reviewers and Mathew S. Shugart made very useful suggestions to improve the project and the final product. All the co-authors have also collaborated at collecting data and building tables, although responsibility for errors remains that of the book editor. The author and publisher have made every attempt to contact copyright-holders. If any have inadvertently been overlooked, the publisher will be pleased to make the necessary arrangement at the first opportunity. J.M.C.

x

Foreword Bernard Grofman

During the latter part of the nineteenth century and the early part of the twentieth century, there was a dramatic change in electoral systems throughout Western Europe and the industrialized world, a shift from majoritarian methods to proportional representation (PR) that left only the English-speaking nations unaffected. Once PR systems were in place, for most of the twentieth century changes occurred only at the margins.1 Now, however, is a great time to be studying electoral laws, thanks to the many new democracies we can study, and the fact that a number of long-term democracies (e.g. Great Britain, Italy, Japan, New Zealand) have made major changes in their electoral laws within the past decade after a long period during which their electoral systems were largely frozen in place. There has been a remarkable resurgence of interest in recent decades in the effects of electoral laws on representation and other aspects of politics. Moreover, the ‘state of the art’ has improved dramatically. When Arend Lijphart and I co-edited Choosing an Electoral System (1984) and Electoral Laws and Their Political Consequences (1986) the electoral systems literature was not that large, and a good part of it was polemic. Moreover, with a handful of exceptions such as Katz (1980), the theory had not advanced much beyond Rae’s seminal work on The Political Consequences of Electoral Laws (1967, 1971). When we turn to the present we find a number of books making important original theoretical contributions, such as Lijphart (1984, 1994, 1999), Taagepera and Shugart (1989), Cox (1997) and Di Cortona et al. (1999), an edited series on the world’s major electoral systems which has already had four of its five planned volumes see the light of day (Grofman et al., 1999a; Bowler and Grofman, 2000a; Shugart and Wattenberg, 2001; Grofman and Lijphart, 2002a), work which studies electoral law developments in new democracies (e.g. Reynolds, 1999b, to name but one of many studies), and numerous overview volumes that integrate a large amount of information on comparative electoral systems (e.g. Reynolds and Reilly, 1997). When we turn from books to articles there is an even greater cornucopia of riches. In particular, there is something new to be learned from every issue of Electoral Studies (a journal which did not exist when I first began writing on electoral systems issues more than a quarter century ago), and articles on the effects of electoral laws and theories of electoral engineering have become a mainstay in top political science journals, especially those that deal with comparative politics. There are a number of key differences between the electoral systems literature of today and that of earlier times. First, there is a greater variety of theoretical perspectives to draw upon. Increasingly important is the game-theoretic approach, with an emphasis on incentives for strategic behaviour and a search for equilibria, exemplified in the work of Gary Cox (e.g. Cox, 1990, 1997). Closely related is the xi

xii

Bernard Grofman

work in social choice theory with its emphasis on the axiomatic underpinning of voting methods (see, for example, Balinski and Young, 1982). Another important research paradigm derives from the work of Rein Taagepera inspired by the physical sciences (e.g. Taagepera, 1986; Taagepera and Shugart, 1989, 1993; Taagepera, 2001), with its emphasis on boundary conditions and statistical ‘law of large number’ effects, a limited set of key variables generating law-like relations, and exponential rather than linear models. Finally, more traditional methods, e.g. the use of regression methods to test empirical generalizations, have been honed and applied to an ever wider set of data by Arend Lijphart (1994) and numerous other scholars. Second, the links between the study of electoral systems and the study of party systems (see, for example, Grofman and Lijphart, 2002a), and between electoral systems research and the study of constitutional design, have become key topics for investigation. Lijphart (1984, 1999) makes the point that choice of electoral systems appears closely linked to other aspects of constitutional design (see especially his discussion of the congruence among features of the Westminster model versus the consensus model). Shugart and Carey (1992), and others, have begun to look at how the presence of a presidential system, the rules for electing the president and the timing of parliamentary and presidential elections interact with the choice of electoral system for parliamentary elections to produce political consequences. My own recent work has stressed that electoral systems and their consequences are embedded in a larger political and institutional framework and has proposed an ‘embedded systems’ research design to conduct longitudinal research on electoral system effects (Grofman et al., 1999b). Thus, on the one hand, it is clear that electoral institutions have ramifications that extend beyond the immediate electoral arena, while, on the other hand, there is an increasing recognition that election systems cannot be understood as operating in a vacuum. The effects of electoral rules are mediated by other aspects of political institutions and political culture, as well as past history and the shape of party constellations. Seemingly identical electoral rules may give rise to very different types of outcomes in different political settings (Bowler and Grofman, 2000b). Third, there is increasing interest in going below the national level to look at local electoral systems and at the impact of national electoral systems on regional and local politics. In particular, political geographers have emphasized how the geographic distribution of partisan support is a key intermediating factor that shapes the extent to which electoral institutions (or changes in them) affect outcomes, especially electoral fairness in the translation of votes into seats (Gudgin and Taylor, 1979; Taylor, Gudgin and Johnston, 1986; Johnston et al. 2001). Fourth, the range of questions considered has broadened considerably. Much of the earlier work on electoral systems dealt with one of three questions: the proportionality of seats–votes relationships, the effect of the electoral system on party proliferation and the effect of the electoral system on cabinet durability. Now topics include racial and gender representation (Karnig and Welch, 1982; Grofman, Migalski and Noviello, 1986; Davidson and Grofman, 1994; Grofman and Davidson, 1994; Rule, 1987; Reynolds, 1999a), the structure of ideological representation (Cox,

Foreword

xiii

l990; Greenberg and Weber, 1985; Myerson and Weber 1993; Robertson, 1976; Sugden, 1984), incentives to cultivate a personal vote through particularistic appeals (Cain, Ferejohn and Fiorina, 1987; Carey and Shugart, 1995; Myerson, 1993a, 1993b; McCubbins and Rosenbluth, 1995), incentives for strategic voting (Cox, 1997) and effects on turnout (Blais and Carty, 1990) – to name but several. Fifth, there is greater recognition that electoral rules that appear identical may significantly differ in their consequences when we look below the surface to consider differences such as in the average number of representatives elected per district (Sartori, 1968; Taagepera and Shugart, 1989), or in national vote thresholds (Taagepera, 2002), or in candidate nomination procedures,2 or in even more finegrain features such as ballot format, or rules affecting what counts as a legal ballot, or rules restricting campaigning, or rules that affect how easy it is for parties to get on the ballot, or for independent candidates to run. Sixth, it is now more widely recognized that the full effects of changes in electoral systems may not occur immediately, since it may take time for key actors to realize the nature of the behaviours that constitute optimizing strategies in the new system (Reed, 1990). Moreover, changes in election systems may give rise to equilibrating forces that moderate the consequences of the changes as voters, candidates and parties adapt their behaviour to the new institutional environment in ways that compensate for the changes so as to partially restore significant elements of the status quo (Christensen and Johnson, 1995; Taagepera and Shugart, 1989). The present volume makes an important addition to the growing body of literature on electoral systems whose key features I have outlined above. First, a principal focus of this volume is the study of a topic that has been relatively neglected until quite recently, namely the origins of electoral systems and the reasons for (and not just effects of) changes in electoral rules. The chapters in this volume provide a set of what my late colleague, Harry Eckstein (1975, 1992), referred to as theoretically driven case studies on this topic, covering a remarkably wide range of countries and time periods.3 They offer a nuanced portrait of real-world changes, illustrating, for example, how change can occur at different levels of a political system, how electoral reform efforts can fail again and again, how exogenous events, both short-term and long-term (e.g. demographic and social shifts), may trigger changes in the calculations of relative party advantage associated with different electoral mechanisms, how the timing and scope of actual changes will be tied to the bargaining/threat power of the parties, how institutional rules for electoral system change (e.g. the ability of groups to go outside legislative channels to institute change through popular initiative) can affect the pace and nature of change, and how, in Professor Colomer’s words, ‘intellectual creativity’ can lead to the invention of ‘new rules and procedures’ that can ‘reshape actors’ institutional preferences and political strategies’. Thus this volume goes a long way in providing the necessary detailed historical evidence about the roots of electoral system change for a large and diverse set of democratic nations throughout recent history.

xiv

Bernard Grofman

Second, this volume makes a major addition to the new institutionalist paradigm in rational choice to which it is squarely anchored. As I have characterized that paradigm, it consists of the belief that ‘institutions exist as both objects of choice and as constraints on choice’, or, in John Ferejohn’s apt phrasing, ‘Preference for outcomes conditions preference for institutions’ (quoted in Grofman, 1989). A central question motivating this volume is why do parties (and other actors) have the preferences among electoral system rules that they do, and why do they make the choices among electoral rules that we observe. A key element in the approach of Professor Colomer and the other authors is that of rational calculation, and a body of evidence is gathered in support of the notion that changes in electoral system reflect parties’ calculations of relative advantage as mediated by the strategic context in which parties find themselves. Colomer summarizes this approach by observing: ‘The performance of the existing electoral rules will likely be evaluated by political actors for the types of winners and losers they tend to produce, that is for the opportunities they create for the survival of different political parties, the attainment of seats and offices within the institutional structure, the possibility to implement their preferred policies, and the likelihood to be re-elected.’4 Colomer also notes that observed (or expected) ‘bad’ outcomes can lead political actors to ‘experiment with alternative formulas’.5 Third, while there have been recent important theoretical contributions about the causes of electoral system changes that go beyond the classic work of Stein Rokkan (see especially Boix, 1999),6 and there is an ongoing debate about the extent to which changes in electoral rules shape party systems as opposed to party systems being primarily a construct of underlying social cleavages in society,7 Professor Colomer offers a new and highly original theory of electoral change which lays the groundwork for a radical revision of what has become the common wisdom about the effects of electoral laws. Colomer’s work challenges Duverger’s famous proposition that electoral laws have a direct causal effect on the number of parties via a combination of a mechanical and a psychological effect (Duverger, 1959).8 More specifically, it offers a new theory about the conditions under which changes in electoral law will be in the direction of greater proportionality. In this theory, nations with a low effective number of parties will be more likely to change electoral rules in ways that reduce proportionality, while systems with a high effective number of parties are more likely to change electoral rules in ways that increase proportionality. In Professor Colomer’s own words, changes in electoral rules act ‘not so much to “produce” or even “permit” or “restrict” the number of parties, but mostly to crystallize, consolidate, or reinforce the previously existing party system.’ Fourth, in addition to the substantial contributions of the country-specific chapters, this volume contains important overview chapters providing descriptive summaries and quantitative analyses of a large cross-national and longitudinal data set that includes information about both legislative and presidential election rules.9 These chapters include the most comprehensive test to date of the well known proposition that the general trend in electoral systems change has been in the direction of greater proportionality,10 and they also include a test of (and

Foreword xv

validation of) Professor Colomer’s new hypothesis that exceptions to this trend can often be understood in terms of the consolidation of existing party systems under rules that reinforce previous winners. These analyses make good use of what another colleague of mine, A. Wuffle (cited in Grofman, 1999; see also the introductions to Davidson and Grofman, 1994, and Grofman and Lijphart, 2002), has called the ‘TNT principle’, i.e. comparisons across time, across nations and across types of institutions – the principle that is at the heart of comparative political analysis. They illustrate the power of theory by laying bare patterns that have not previously been noticed, and could not be noticed until one understood what to look for.11 In sum, the chapters in this volume provide us with a wealth of new information and new theoretical insights that help us account for both stability and change in electoral systems. They show the extent to which changes in electoral rules be traced to strategic calculations and power balances among the relevant actors, especially to the instantiation of rules that mesh more closely with (and act to solidify) existing party constellations. Bernard Grofman Professor of Political Science and Institute for Mathematical Behavioral Science University of California, Irvine

Notes 1. Shugart (1992). 2. See, for example, Coleman (1972), Aranson and Ordeshook (1972) and Owen and Grofman (1995) on the effects of party primaries on the ideological structure of two-party competition. 3. Almost all of the recent work on electoral origins has been on changes in the post-Second World War period, with much of it related to the adoption of electoral rules in former communist countries (see, for example, Bawn, 1993; Shugart, 1992; Ishiyama, 1996; Grofman, Mikkel and Taagepera, 1999), or in Africa (see, for example, Reynolds, 1999b; Mozaffar, 2003; Mozaffar and Vengroff, 2002). One notable exception is Grofman and Lijphart (2002a), which deals with electoral changes in the five Nordic nations over a one hundred year plus span. 4. While we normally associate such a perspective with contemporary scholars in political science who draw on game-theoretic ideas, it is useful to be reminded that ‘rational choice’ ideas are found in the work of such major earlier scholars as V. O. Key and Stein Rokkan. For example, Rokkan (1970), in seminal work, argues that the primary force behind the introduction of PR in the Nordic countries and in some other parts of Western Europe was the desire of conservative parties (then dominant) to avoid complete elimination in light of the expected socialist gains when the working class was enfranchised, coupled with the view of challengers that PR would guarantee them equitable representation. (For some important emendations to Rokkan’s thesis see various essays (and the editors’ introduction) in Grofman and Lijphart, 2002a.) Rokkan (1970) also offers a ‘rational choice’ argument about why Sweden, Denmark and Norway in the 1950s changed from the d’Hondt rule to the modified Sainte-Laguë form of list PR (with an initial divisor of 1.4). In his view, this was done as a means of improving the chances of ‘middle-sized’ parties to achieve equitable representation. (For some important emendations to this Rokkan thesis see Elklit (1999) and the Elklit, Aardal and Särlvik chapters in Grofman

xvi

5. 6.

7.

8.

9.

10.

Bernard Grofman and Lijphart (2002a). In Finland, the attempts by the Swedish People’s Party to replace d’Hondt with modified Sainte-Laguë were rejected several times on grounds that the change might increase party fragmentation (Sundberg, 2002), while in Iceland, d’Hondt was eventually (1987) replaced with greatest remainder as a way of increasing proportionality in small constituencies (Hardarson, 2002).) In this context, it is useful to remember that not all calculations of self-interest need be accurate ones. Other important approaches to electoral system origins include the historical and largely class-based analysis of Rueschmeyer, Stephens and Stephens (1992), and work that views the choice of electoral system as conditioned by ‘familiarity’ or the ‘spirit of the times’. The adoption of first-past-the-post (plurality) single-member district elections by most former British colonies after independence is often attributed largely to habit (Mozaffar, 2003). Today, the attractiveness of mixed member systems may be based at least in part on a widespread perception that they are the ‘wave of the future’ (Shugart and Wattenberg, 2001). Relatedly, there may be important contagion effects, as occurred with the adoption of the modified form of the Sainte Laguë formula for list PR in the early 1950s in the Nordic nations as a replacement for the d’Hondt formula (Elklit, 1999; Aardal, 2002). Another potential source of electoral law choices is the view of key protagonists about principles of good government. For example, Reynolds (2001) argues that, in planning for the 1994 elections in South Africa, ANC leaders opted for proportional representation even though that was not the electoral system most beneficial to the ANC, in large part due to a belief that proportionality of representation was a desirable outcome for the multi-racial democracy they hoped South Africa would become (see also Reynolds and Grofman, 1995). My own views about forces affecting electoral system origins and changes are synopsized in Grofman and Lijphart (2002b) and are very similar to those of Professor Colomer. See Taagepera and Grofman (1985) for an introduction to this debate (cf. Grofman et al., 1999b; Bowler and Grofman, 2000b). Taagepera (see Taagepera and Grofman, 1985; Taagepera and Shugart, 1989), following up on some empirical work in Lijphart (1984), has found both theoretical and empirical support for the generalization that N = I + 1, i.e. that the effective number of parties contesting elections, N, is a simple function of I, the number of major cleavage dimensions in a society. He and his co-authors suggest that it may make just as much sense to think of party constellations as determined by issue cleavages as it does to write the equation I = N − 1, and assume that the number of parties competing sets limits on the number of issue dimensions that can be expressed in the political arena – in the way that two points give us a line (one-dimensional competition), three points define a plane (two-dimensional competition), etc. Duverger’s insights have been refined by scholars such as Steven Reed (1990), Gary Cox (1997) and Rein Taagepera. In particular, Taagepera’s early work (e.g. Taagepera and Shugart, 1989, 1993) provides a statistical basis (making use of known boundary conditions and appeals to uncertainty) for the claim that ne (the number of parties/candidates receiving seats) will be approximately M1/2 in any given district, while, nationwide (at least if districts are of roughly similar magnitude), the expected number of parties receiving seats (including independents) will be given by (MS)1/4. Here we might also note that, in his approach to comparative analysis, Professor Colomer’s methodology has important similarities to the work of Rein Taagepera. Paralleling Taagepera’s recent work, Professor Colomer stresses three aspects of electoral rules that can be manipulated (1) S, (2) M and (3) the degree of proportionality in the electoral rule – a concept he operationalizes in terms of what he refers to as an electoral rule’s quota. (This parameter usually may be expressible as some function of M.) For example, Jackson and McRobie (1998: 2) observe that ‘changes in the electoral systems in the twentieth century have trended from majoritarian to proportional.’ However, they also observe that ‘recently . . . there has been some movement in the opposite direction.’

Foreword

xvii

11. The comparative analyses also demonstrate the importance for the development of a theory of electoral system change of studying not just ‘big’ electoral changes (e.g. from plurality to PR), but smaller electoral changes as well.

References Aardal, Bernt (2002) ‘Electoral System in Norway’, in Bernard Grofman and Arend Lijphart (eds), The Evolution of Electoral and Party Systems in the Nordic Countries. New York: Agathon Press, pp. 167–224. Aranson, Peter and Peter C. Ordeshook (1972) ‘Spatial Strategy for Sequential Elections’, in R. Niemi and H. Weisberg (eds), Probability Models of Collective Decision Making. Columbus, OH: Merrill. Balinski, Michel L. and H. Peyton Young (1982) Fair Representation: Meeting the Ideal of One Man, One Vote. New Haven, CT: Yale University Press. Bawn, K. (1993) ‘The Logic of Institutional Preferences: German Electoral Law as a Social Choice Outcome’, American Journal of Political Science, 37, 4: 965–89. Blais, André and R. K. Carty (1990) ‘Does Proportional Representation Foster Voter Turnout?’, European Journal of Political Research, 18, 2: 167–81. Boix, Carles (1999) ‘Setting the Rules of the Game: The Choice of Electoral Systems in Advanced Democracies’, American Political Science Review, 93, 3: 609–24. Bowler, Shaun and Bernard Grofman (eds) (2000a) Elections in Australia, Ireland and Malta under the Single Transferable Vote. Ann Arbor: University of Michigan Press. Bowler, Shaun and Bernard Grofman (2000b) ‘Introduction: STV as an Embedded Institution’, in Shaun Bowler and Bernard Grofman (eds), Elections in Australia, Ireland and Malta under the Single Transferable Vote. Ann Arbor: University of Michigan Press, pp. 1–14. Cain, Bruce, John A. Ferejohn and Morris Fiorina (1987) The Personal Vote: Constituency Service and Electoral Independence. Cambridge, MA: Harvard University Press. Carey, John M. and Matthew Shugart (1995) ‘Incentives to Cultivate a Personal Vote: A Rank Ordering of Electoral Formulas’, Electoral Studies, 14, 4: 417–40. Christensen, Raymond and Paul Johnson (1995) ‘Toward a Context-Rich Analysis of Electoral Systems: The Japanese Example’, American Journal of Political Science, 39: 575–98. Coleman, James S. (1972) ‘The Positions of Political Parties in Elections’, in R. G. Niemi and F. Weisberg (eds), Probability Models of Collective Decision Making. Columbus, OH: Charles E. Merrill, pp. 332–57. Cox, Gary W. (1990) ‘Centripetal and Centrifugal Incentives in Electoral Systems’, American Journal of Political Science, 31: 82–108. Cox, Gary W. (1997) Making Votes Count: Strategic Coordination in the World’s Electoral Systems. Cambridge and New York: Cambridge University Press. Davidson, Chandler and Bernard Grofman (eds) (1994) Quiet Revolution in the South: The Effects of the Voting Rights Act, 1965–1990. Princeton, NJ: Princeton University Press. Di Cortona, Pietro Grilli et al. (1999) Evaluation and Optimization of Electoral Systems. Philadelphia, PA: Society for Industrial and Applied Mathematics. Duverger, Maurice (1959) Political Parties: Their Organization and Activity in the Modern State. London: Methuen/New York: Wiley. Eckstein, Harry (1975) ‘Case Study and Theory in Political Science’, in Fred Greenstein and Nelson W. Polsby (eds), Political Science: Scope and Theory, Handbook of Political Science, Vol. I. Reading, MA: Addison-Wesley, pp. 79–137 (reprinted in Eckstein, 1992: 117–78). Eckstein, Harry (1992) Regarding Politics: Essays on Political Theory, Stability and Change. Berkeley: University of California Press. Elklit, Jørgen (1999) ‘What Was the Problem if a First Divisor of 1.4 was the Solution?’, in Erik Beukel, Kurt Klaudi Klausen and Poul Erik Mouritzen (eds), Elites, Parties and Democracy. Festschrift for Professor Mogens N. Pedersen. Odense: Odense University Press, pp. 75–101.

xviii

Bernard Grofman

Greenberg, Joseph and Shlomo Weber (1985) ‘Multiparty Equilibria under Proportional Representation’, American Political Science Review, 79, 3: 693–703. Grofman, Bernard (1989) ‘The Federalist Papers and the New Institutionalism: An Overview’, in Bernard Grofman and Donald Wittman, The Federalist Papers and the New Institutionalism. New York: Agathon Press. Grofman, Bernard (1999) ‘Preface: Methodological Steps toward the Study of Embedded Institutions’, in Bernard Grofman, Sung-Chull Lee, Edwin A. Winckler and Brian Woodall (eds), Elections in Japan, Korea, and Taiwan under the Single Non-Transferable Vote. Ann Arbor: University of Michigan Press, pp. i–xvii. Grofman, Bernard and Chandler Davidson (1994) ‘The Effect of Municipal Election Structure on Black Representation in Eight Southern States’, in Chandler Davidson and Bernard Grofman, Quiet Revolution in the South: The Effects of the Voting Rights Act, 1965–1990. Princeton, NJ: Princeton University Press, pp. 301–34. Grofman, Bernard and Arend Lijphart (eds) (2002a) The Evolution of Electoral and Party Systems in the Nordic Countries. New York: Agathon Press. Grofman, Bernard and Arend Lijphart (2002b) ‘Introduction’, in Bernard Grofman and Arend Lijphart (eds), The Evolution of Electoral and Party Systems in the Nordic Countries. New York: Agathon Press, pp. 1–13. Grofman, Bernard, Michael Migalski and Nicholas Noviello (1986) ‘Effects of Multimember Districts on Black Representation in State Legislatures’, Review of Black Political Economy, 14, 4: 65–78. Grofman, Bernard, Evald Mikkel and Rein Taagepera (1999) ‘Electoral Systems Change in Estonia, 1989–1993’, Journal of Baltic Studies, 30, 3: 227–49. Grofman, Bernard, Sung-Chull Lee, Edwin A. Winckler and Brian Woodall (eds) (1999) Elections in Japan, Korea, and Taiwan under the Single Non-Transferable Vote. Ann Arbor: University of Michigan Press. Gudgin, Graham and Peter J. Taylor (1979) Seats, Votes and the Spatial Organization of Elections. London: Pion. Hardarson, Ólafur Th. (2002) ‘The Icelandic Electoral System 1844–1999’, in Bernard Grofman and Arend Lijphart (eds), The Evolution of Electoral and Party Systems in the Nordic Countries. New York: Agathon Press, pp. 101–66. Ishiyama, T. John (1996) ‘Electoral Systems Experimentation in the New Eastern Europe: The Single Transferable Vote and the Additional Member System in Estonia and Hungary’, East European Quarterly, 29, 4: 487–507. Jackson, Keith and Alan McRobie (1998) New Zealand Adopts Proportional Representation. Christchurch, New Zealand: Hazard Press. Johnston, Ron et al. (2001) From Votes to Seats: The Operation of the UK Electoral System Since 1945. Manchester: University of Manchester Press. Karnig, Albert and Susan Welch (1982) ‘Electoral Structure and Black Representation on City Councils’, Social Science Quarterly, 63: 99–114. Katz, Richard S. (1980) A Theory of Parties and Electoral Systems. Baltimore, MD: Johns Hopkins University Press. Lijphart, Arend (1994) Electoral Systems and Party Systems. A Study of Twenty-Seven Democracies 1945–1990. New York and Oxford: Oxford University Press. Lijphart, Arend (1999) Patterns of Democracy: Government Forms and Performance in Thirty-Six Countries. New Haven, CT: Yale University Press. Lijphart, Arend and Bernard Grofman (eds) (1984) Choosing an Electoral System. New York: Praeger. McCubbins, Matthew D. and Frances M. Rosenbluth (1995) ‘Party Provision for Personal Politics: Dividing the Votes in Japan’, in Peter Cowhey and Matthew D. McCubbins (eds), Structure and Policy in Japan and the United States. New York: Cambridge University Press, pp. 35–55. Mozaffar, Shaheen and Richard Vengroff (2002) ‘A “Whole System” Approach to the Choice of Electoral Rules in Democratizing Countries: Senegal in Comparative Perspective’, Electoral Studies, 21, 4: 601–16.

Foreword

xix

Myerson, Roger B. (1993a) ‘Incentives to Cultivate Favored Minorities under Alternative Electoral Systems’, American Political Science Review, 87, 4: 856–69. Myerson, Roger B. (1993b) ‘Effectiveness of Electoral Systems for Reducing Government Corruption: A Game Theoretic Analysis’, Games and Economic Behavior, 5: 118–32. Myerson, Roger B. and R. J. Weber (1993) ‘A Theory of Voting Equilibria’, American Political Science Review, 87, 1: 102–14. Owen, Guillermo and Bernard Grofman (1995) Expressive Voting and Equilibrium in Two-Stage Electoral Competition Involving Both Primaries and a General Election. Paper prepared for delivery at the Annual Meeting of the Public Choice Society, Long Beach, California, 24–26 March 1995. (A revised version presented at the Conference on Strategy and Politics, Center for the Study of Collective Choice, University of Maryland, College Park, MD, 12 April 1996.) Rae, Douglas ([1967] 1971) The Political Consequences of Electoral Laws, 2nd edn. New Haven, CT: Yale University Press. Reed, Stephen R. (1990) ‘Structure and Behavior: Extending Duverger’s Law to the Japanese Case’, British Journal of Political Science, 20: 335–56. Reynolds, Andrew (1999a) ‘Women in the Legislatures and Executives of the World: Knocking at the Highest Glass Ceiling’, World Politics, 51, 4: 547–72. Reynolds, Andrew (1999b) Electoral Systems and Democratization in Southern Africa. New York and Oxford: Oxford University Press. Reynolds, Andrew (2001) The Architecture of Democracy: Constitutional Design, Conflict Management, and Democracy. New York and Oxford: Oxford University Press. Reynolds, Andrew and Bernard Grofman (1992) The Main Proposals for Electoral Reform in South Africa. Paper presented at the UCI Focused Research Program in Democratization, Conference on Constitutional Design, University of California Irvine. Reynolds, Andrew and Ben Reilly (1997) The International IDEA Handbook of Electoral System Design. Stockholm: International IDEA. Robertson, David (1976) A Theory of Party Competition. New York: Wiley. Rokkan, Stein (1970) Citizens, Elections, Parties. Approaches to the Comparative Study of the Processes of Development. Oslo: Universitetsforlaget. Rueschmeyer, Dietrich, Evelyne Huber Stephens and John D. Stephens (1992) Capitalist Development and Democracy. Cambridge: Polity Press. Rule, Wilma (1987) ‘Electoral Systems, Contextual Factors and Women’s Opportunity for Election to Parliament in Twenty-three Democracies’, Western Political Quarterly, 40, 3: 477–98. Sartori, Giovanni (1968) ‘Political Development and Political Engineering’, in J. D. Montgomery and A. O. Hirschman (eds), Public Policy, 17: 261–98. Shugart, Matthew S. (1992) ‘Electoral Reform in Systems of Proportional Representation’, European Journal of Political Research, 21: 207–24. Shugart, Matthew S. and John Carey (1992) Presidents and Assemblies: Constitutional Design and Electoral Dynamics. New York and London: Cambridge University Press. Shugart, Matthew S. and Martin Wattenberg (2001) Mixed Member Electoral Systems: The Best of Both Worlds? Oxford: Oxford University Press. Sugden, Robert (1984) ‘Free Association and the Theory of Proportional Representation’, American Political Science Review, 78, 1: 311–43. Sundberg, Jan (2002) ‘The Electoral System of Finland: Old, and Working Well’, in Bernard Grofman and Arend Lijphart (eds), The Evolution of Electoral and Party Systems in the Nordic Countries. New York: Agathon Press, pp. 67–100. Taagepera, Rein (1986) ‘Reformulating the Cube Law of Elections for Proportional Representation Elections’, American Political Science Review, 80: 489–504. Taagepera, Rein (2001) ‘Party Size Baselines Imposed by Institutional Constraints: Theory for Simple Electoral Systems’, Journal of Theoretical Politics, 13: 331–54. Taagepera, Rein (2002) ‘Nationwide Effective Threshold of Representation’, Electoral Studies, 21: 383–401.

xx

Bernard Grofman

Taagepera, Rein and Bernard Grofman (1985) ‘Rethinking Duverger’s Law: Predicting the Effective Number of Parties in Plurality and PR Systems. Parties Minus Issues Equals One’, European Journal of Political Research, 13: 341–52. Taagepera, Rein and Matthew Soberg Shugart (1989) Seats and Votes: The Effects and Determinants of Electoral Systems. New Haven, CT: Yale University Press. Taagepera, Rein and Matthew Soberg Shugart (1993) ‘Predicting the Number of Parties: A Quantitative Model of Duverger’s Mechanical Effect’, American Political Science Review, 87: 455–64. Taylor, Peter J., Graham Gudgin and R. J. Johnston (1986) ‘The Geography of Representation: A Review of Recent Findings’, in Bernard Grofman and Arend Lijphart (eds), Electoral Laws and Their Political Consequences. New York: Agathon Press.

List of Contributors General Editor Josep M. Colomer, Higher Council of Scientific Research, Barcelona, Spain

Foreword Bernard Grofman, University of California, Irvine, USA

Region and country studies Gerard Alexander, University of Virginia, USA Alberto Díaz-Cayeros, Stanford University, California, USA, and ITAM, Mexico Patrick Dunleavy, London School of Economics and Political Science, UK Andrew Ellis, International IDEA, Stockholm, Sweden, and National Democratic Institute, Jakarta, Indonesia Erik J. Engstrom, University of North Carolina, Chapel Hill, USA Richard L. Engstrom, University of New Orleans, Louisiana, USA Carlos Flores Juberías, University of Valencia, Spain Diego Gambetta, Oxford University, UK Matt Golder, New York University, USA Allen Hicken, University of Michigan, USA Marek M. Kaminski, University of California, Irvine, USA Petr Kopecký, University of Sheffield, UK, and University of Leiden, the Netherlands Marcus Kreuzer, Villanova University, Pennsylvania, USA Fabrice Lehoucq, CIDE, Mexico Leif Lewin, University of Uppsala, Sweden Georg Lutz, University of Berne, Switzerland Beatriz Magaloni, Stanford University, California, USA, and ITAM, Mexico Helen Margetts, University College, London, UK Evald Mikkel, University of Tartu, Estonia Subrata K. Mitra, Ruprecht Karls University, Heidelberg, Germany Shaheen Mozaffar, Bridgewater State College, Massachusetts, USA Jack H. Nagel, University of Pennsylvania, USA Monika A. Nalepa, Columbia University, New York, USA Gabriel Negretto, CIDE, Mexico Jairo M. Nicolau, IUPERJ, Rio de Janeiro, Brazil Vello Pettai, University of Tartu, Estonia Andrew Reynolds, University of North Carolina, Chapel Hill, USA Marian Sawer, Australian National University, Canberra, Australia John W. Schiemann, Fairleigh Dickinson University, New Jersey, USA xxi

xxii Handbook of Electoral System Choice

Olga A. Shvetsova, Binghampton University and California Institute of Technology, USA Junichiro Wada, Yokohama City University, Japan Leonard Wantchekon, New York University, USA Steven Warner, Oxford University, UK

Research assistants Luis-Eduardo Escatel, Andira Hernández, Luis Venegas, CIDE, Mexico

Part 1 Introduction

This page intentionally left blank

1 The Strategy and History of Electoral System Choice Josep M. Colomer

It is on account of the importance of the consequences of the electoral system, especially on the number of political parties and the political composition of assemblies and governments, that the choice of system matters. However, electoral systems – as are other prominent political institutions – are also a consequence of already existing political parties in assemblies and governments, each of which tends to prefer those institutional formulas and procedures that can consolidate, reinforce or increase their relative strength. As will be argued and discussed in the following pages, political configurations in which there is a single dominant party or two rather balanced parties tend to produce choices in favour of rather restrictive or exclusionary electoral systems, such as those based on the majority principle, while pluralistic settings with multiple parties tend to support choices in favour of more inclusive electoral formulas, such as those using rules of proportional representation. More generally, the choice of electoral system seems to follow what could be called the ‘Micro-mega rule’, by which the large prefer the small and the small prefer the large: a few large parties tend to prefer small assemblies, small district magnitudes and rules based on small quotas of votes for allocating seats, while multiple small parties tend to prefer large assemblies, large district magnitudes, and large quotas. In a nutshell, large parties prefer small institutions in order to exclude others from competition, while small parties prefer large institutions able to include them within. This may be seen as Duverger’s laws upside down. As the French political scientist brilliantly sketched in the mid-twentieth century and several generations of scholars have further developed and re-elaborated, electoral rules and procedures have indeed paramount influence on relevant aspects of public life, including the number of candidacies or political parties that are created by would-be leaders and the number of them that prove durable, the number of individuals and different groups in the electorate that attain representation in councils and assemblies, their relations with elected representatives, the political party composition of the legislative and the executive, and the degree of social acceptance or rejection of the polity and its constitutional structure. But we should take into consideration that electoral systems – and formal political institutions and rules more generally – are not absolutely independent variables, but also subjects of political choices by 3

4

Josep M. Colomer

already existing political actors, which tend to make decisions in their own interest. Thus it can be expected that electoral systems will usually crystallize, consolidate or reinforce previously existing political party configurations – and the other political features that can be associated with them – rather than generate new party systems or political outcomes by themselves. It is exogenous changes in the parties’ relative strengths or expectations, whether by the emergence of new parties or, in general, by coordination failures of the existing parties – in spite of or against the incentives provided by the existing electoral system – that can induce further electoral system changes. These statements do not deny, but, on the contrary, assume Duverger’s laws as granted, that is that electoral systems have clear consequences on the number of parties, but it inserts this relationship within a more encompassing analytical framework including the choice of electoral system by already existing political parties.1 It seems reasonable to assume that, under restrictive formulas such as majority rule, political actors facing the effects of their own failure at coordinating themselves into a small number of candidacies and the emergence of new issues and new contenders for seats and offices may shift their preference to electoral institutions able to reduce the risks of competing by giving all participants greater opportunities to obtain or share power. When there are only a few parties, they can be satisfied with majoritarian electoral systems, but when the number and the size of new parties increase, the incumbent parties may begin to fear the risk of becoming absolute losers and try to shift to more inclusive electoral formulas. As will be shown in the following pages, electoral system changes indeed tend to move overwhelmingly in favour of increasingly inclusive, less risky formulas: from indirect to direct elections, from unanimity to majority rule, and from the latter to mixed systems and to proportional representation (in other words, towards large assemblies, districts and quota rules). Three basic variables emerge for consideration at the time of choosing a specific electoral system: the performance of the existing rules and the changing actors’ expectations of electoral results under them; the availability or invention of alternative electoral rules and formulas with different expected effects; the actors’ relative capabilities for institutional decision-making. The performance of the existing electoral rules will likely be evaluated by political actors for the type of winners and losers they tend to produce, that is for the opportunities they create for the survival of different political parties, the attainment of seats and offices within the institutional structure, the possibility of implementing their preferred policies, the likelihood of being re-elected, and the fairness of the overall distribution of power positions. It can be postulated that the more restrictive and exclusionary the existing electoral system – such as those based on majority rule – and the higher the number of parties with popular support or reasonable expectations of getting it, the stronger the pressures for electoral system change in favour of more permissive and inclusive formulas will be. In the analysis which follows, we will use the ‘effective number’ of parties in votes (that is, the number of parties weighed by their size in votes) as a proxy for the potential parties suffering from under-representation under the existing rules and able to expect benefits from such a change.

Strategy and History of Electoral System Choice

5

Two caveats can be taken into consideration. First, of course, the type of change will also depend on the knowledge of alternative formulas and, in absolute terms, on the stock of mathematical and legal inventions available at the moment. But it is ‘bad’ – that is, unfavourable or unpalatable – performances that can be more likely to induce political actors to experiment with alternative formulas or suggest that scholars and experts should invent new rules and procedures able to produce different outcomes than those already enforced. Intellectual creativity can then feed new expectations and reshape actors’ institutional preferences and political strategies accordingly. Second, it could be expected that electoral institutions – as with any other enforceable rules of the game in social interactions and exchanges – may have some capability of imposing costs on those not complying, fostering learning processes by which even the losers in the game can obtain relative net benefits from playing (in comparison with the costs of undertaking changes and playing under a new electoral system), and attracting actors’ adaptation to the constraints and incentives provided. Thus the longer the duration of an existing system, the higher could be the probability that adapted actors will give the system endogenous support while institutionally-induced losers will be eventually absorbed or excluded and decreasingly able to challenge the rules. In other words, the relative capability of institutional decision-making may decrease over time for those actors which are victims of the existing institutional rules. In a first, very simple approach, it can thus be hypothesized that: 1. electoral system changes will move mostly in favour of more inclusive formulas permitting representation of a relatively high number of parties; 2. they should be more probable the higher the effective number of existing parties; and 3. the shorter the duration of the previous electoral system had been. From this perspective, this introductory chapter presents three different approaches – logical, historical and empirical – to the problem of electoral system choice. First, it presents a ‘strategic’ model of how electoral systems are chosen by decisive political actors. Second, it displays a very long-term historical panorama of the invention and adoption of different electoral rules and formulas which is organized around four basic principles: unanimity, lottery, majority and proportionality. Finally, several worldwide tests of the choice of electoral systems on 154 occasions in 94 countries since the nineteenth century strongly support the hypotheses just sketched above.

A model of electoral system choice Let us, first, present a ‘strategic’ model of how electoral systems are chosen by relevant political actors, whether in situations of general change of institutional regime or within more stable frameworks. A few assumptions can be stated in a preliminary way in order to develop a deductive reasoning which will be further tested and discussed.

6

Josep M. Colomer

First of all, let us assume that voters and leaders are motivated to participate in elections and voting decisions in order to win, that is to see their preferred candidates elected or their values or policies adopted as collective, enforceable decisions, or at least to achieve a situation in which the winning alternatives are as close as possible to their most preferred ones. Self-interested individuals thus will prefer and, if given the opportunity, choose or contribute to choosing the electoral rules and voting procedures from which they can expect a higher likelihood of winning or achieving satisfactory outcomes. Other criteria may, doubtless, appear in the public debate and the corresponding exchanges and negotiations, if any. But external references such as electoral systems that had been used in some past experience of the community, imitation of neighbours or other communities, institutional import, as well as inventing and creative engineering, are likely to be dependent on internal, existing criteria and choices. Let us now consider several hypothetical situations faced by those self-interested political actors just identified. If the distribution of power in a community is such that one single group is institutionally dominant and expects to be the secure winner with the existing electoral rules and voting procedures, these will not likely be changed – institutional stability can be expected. More specifically, if the existing winners are sufficiently powerful to impose or maintain their preferred rules, and at the same time they are optimistic about their future electoral chances, they will likely choose in favour of rather restrictive rules based on the majority principle and on single, categorical ballots giving each voter a single choice. Since this type of electoral rules tends to produce a single absolute winner, they can provide the dominant group with more opportunities to remain the winner and retain control. There can be, in contrast, other situations more prone to institutional change, resulting either from uncertainty or threat. One of these situations may exist when no single group of voters and leaders, including in particular the incumbent rulers in a well-organized political community, is sufficiently sure about its support and the corresponding electoral prospects in future contests – in other words, when there is high uncertainty regarding the different groups’ relative strengths. Uncertainty can be the result of partial or confusing information regarding the different groups’ relative support and their subsequent inability to refine their electoral prospects. Also, uncertainty may result from the enforcement of rules and procedures producing unpredictable, rather arbitrary winners and collective decisions, thus making every actor, even those with more reliable voter support, insecure about the likelihood of winning. Another situation prone to make electoral system change enter the agenda of relevant political actors may appear when, with reasonable certainty, the incumbent winners feel that new groups are emerging, gaining increasing support among voters, challenging their domination and threatening to replace them with an alternative set of candidates, values or policies as a result of elections with the existing rules and procedures. Political parties’ failure at coordinating their candidacies, especially their inability to concentrate support under the constrictions of majority rules, may transform winners into losers and vice versa, even if social support of the interests and opinions represented by the different candidacies has not changed a great deal.

Strategy and History of Electoral System Choice

7

Let us adopt now the standard assumption that most actors are risk-averse when they make their choices, including, for our problem, choices with regard to electoral rules and procedures. A logical implication not difficult to make from the assumptions just presented is that, under conditions of uncertainty or threat, self-interested, risk-averse actors will prefer and tend to choose electoral rules and voting procedures creating less opportunities for them to become absolute losers. More specifically, if the existing winners cannot impose their preferred rules because increasing demands and pressures from opposition groups force them to cede or negotiate or, even if retaining sufficient institutional power to impose their preferences, they are pessimistic about their electoral prospects, they will try to assure some partially winning positions for themselves through appropriate, that is more inclusive, electoral rules involving multiple ballots or a system approaching proportional representation. Certainly, existing electoral systems, even the most restrictive ones, may have powerful self-reinforcing mechanisms. On the one hand – as suggested above – not only usual winners but also permanent losers may adapt to play within the existing rules of the game as a consequence of the high expected costs and uncertain benefits of a hypothetical institutional replacement. An electoral system can be considered to be relatively stable if, given the opportunities it provides for actors to gain or share institutional and political power, none of them would find it worthwhile to undertake new initiatives to change the rules of the game. However, for anticipated losers or threatened winners, institutional change can be a rational strategy if the expected advantages of alternative formulas surpass those of playing by the existing rules. Change can thus be facilitated by several processes, such as: accumulative learning from the consequences that can be reasonably expected from different electoral formulas, thanks to both political experience and scholarly analysis; imitation of or contagion from changes in other communities where alternative formulas have produced desirable effects; and reduction of risks by parallel agreements and guarantees among the relevant actors regarding some basic statutes or rights. Under different conditions of information, two types of electoral system choice can be distinguished, which can be called, respectively, propositive and reactive. The former type should better correspond to situations in which actors’ preferences over different rules and procedures are clearly defined. These may lead to straightforward decisions when certain actors have a well-profiled role of political initiative and sufficient decision power to introduce institutional reforms. Cases fitting this type, and to be analysed in this book, include, for instance: Sweden in 1907 with the sudden, simultaneous introduction of universal suffrage and proportional representation by a Conservative government under threat from the rise of the Socialists; the replacement of simple plurality rule with a majoritypreferential system in Australia in 1918 as a result of pressure from farmers to open up space for a third, agrarian party; or the establishment of proportional representation in Switzerland in 1918 or in South Africa in 1994, among many other countries, as a result of a previously existing, solid multi-partism (as explained, respectively, by Leif Lewin, Marian Sawer, Georg Lutz and Andrew Reynolds in the corresponding chapters).

8

Josep M. Colomer

But when none of the actors is sufficiently strong to impose its own alternative, well-informed and well-defined preferences may also lead to formal negotiations between opposite political sides producing intermediate compromises. These can be, for instance, a mixed majority-proportional system as in Hungary in 1989, or a qualified plurality rule midway between simple plurality and absolute majority as in Argentina in 1994 (as explained, respectively, by John W. Schiemann and Gabriel Negretto in the corresponding chapters). ‘Propositive’ choices can, thus, reflect quite directly pressures from excluded actors, negotiations between the government and the opposition, or anticipation of risks by incumbent rulers, but they produce, in general, changes in favour of less risky, more inclusive electoral systems, such as those involving multiple ballots, the combination of different principles of representation and higher degrees of proportionality. In contrast, the reactive type of electoral system choice may occur in situations in which actors face more deficient information regarding possible institutional alternatives, more uncertain expectations regarding each actor’s future voting support, stronger constitutional restrictions on more drastic reforms, or confusing and entangled decision processes. Political actors can then ‘react’ against the negative features of the actual electoral system, for instance because it may produce hazardous or inconsistent results (such as giving more seats to the least voted party or candidate, as in parliamentary or in presidential college elections by plurality rule), or for being too rigid to transform new citizens’ demands or political initiatives into institutional representation. Under conditions where there is significant ignorance, actors typically fall back on ‘learning’ from their own experience, perhaps embellished as (local) ‘historical lessons’ (rather than learning from others’ instructive experiences or from scholarly work producing accumulative knowledge). ‘Reactive’ choices tend to adopt formulas ‘opposite’ to those in existence the consequences of which have been shown to be undesirable, thus replacing, for instance, single-member districts with multi-member districts, majority rule with proportional representation or closed lists of candidates with open lists, and vice versa, etc. Political parties with different interests and prospects may coincide in favouring changes of this type if they share rejection of the existing rules, but they may not necessarily forecast accurately the likely effects of the new combinations of institutional elements on which they can converge. Thus ‘reactive’ choices may produce patchwork electoral systems with unanticipated new undesirable consequences. Cases include, for instance, the electoral system adopted in Spain by the new Republic established in 1931 in reaction to the previous ‘monarchical’ system, which induced high polarization and conflict; the system enforced in Colombia from 1991, which destroyed the internal consistency of its political parties; or the Italian mixed system introduced in 1993, which did not reduce the number of parties but replaced centripetal competition around moderate positions with high polarization (as explained by the author Diego Gambetta in the corresponding chapter). Certainly, political leaders and organizations can also learn from their own experience of tentative, failed or mistaken decisions and progressively refine their later choices – as clearly

Strategy and History of Electoral System Choice

9

happened, for instance, in Poland in the post-Communist period (as explained by Marek Kaminski in his chapter). Reactive choices, however, can also forge fortunate new inventions able to produce widely satisfactory results which become reinforced through use or which may even be imported from elsewhere. This has happened, for instance, with qualifiedplurality rules for presidential elections, as introduced in Costa Rica in 1936 and further imitated by other Latin American countries (according to Fabrice Lehoucq), or with the personalized proportional representation system negotiated in Germany after the Second World War, which, thanks to its ability to reunite both proportionality and closeness between citizens and representatives, was also adopted, among others, by apparently somewhat myopic political leaders in New Zealand in 1993 (as explained by Marcus Kreuzer and Jack Nagel in their respective chapters). ‘Reactive’ choices may thus imply changes in the direction of either more inclusive or more exclusive formulas, mistakes in calculations and a higher than usual number of unintended consequences. However, since majority rules and exclusionary formulas are more likely to create dissatisfaction among actors than the corresponding alternatives, changes in this direction are also more likely to be cancelled and revised. The ‘strategic’ approach proposed here can thus enlighten not only the formation of different actors’ preferences regarding electoral systems, but also their preference changes as a consequence of miscalculations under rapidly evolving relations between forces, mistaken expectations, myopic reactions to undesired features of the previous electoral system or unanticipated electoral results which are observed in retrospect. If, for instance, initially optimistic rulers maintaining some restrictive electoral system based on majority rules turn out to be defeated in the corresponding election, they may feel induced to shift their preferences in favour of more inclusive formulas. Conversely, unexpected winners may be tempted to project their present strength into the future and bet for risky electoral rules producing a single absolute winner. Usually there may be asymmetric information among different groups’ leaders regarding the likely effects of electoral rules. Certain institutional reformers may even prefer difficult-to-understand electoral systems constraining the reactions or strategic behaviour of rival leaders and voters. General ignorance of electoral system experience and literature or objective difficulties in identifying factors of advantage or disadvantage for certain parties or groups can also lead to myopic or blind bets for certain institutions and strategies. But, in general, it should be expected that, in the long term, more inclusive electoral rules will develop stronger endogenous support than restrictive systems and thus will tend to endure more and create more solid institutional equilibria. In other words, the very leaders, groups and political parties whose political and institutional power has been made possible by the existing electoral system will support it and will resist the introduction of adverse changes. The actors’ electoral participation and their play within the existing rules of the game tend to reinforce the rules themselves. Thus electoral rules which favour multiple winners and, for this effect, are able to promote actors’ strategies complying with them can become relatively more stable, self-enforcing sets of rules.

10

Josep M. Colomer

This reasoning can be applied to both foundational institutional choices, such as those made at the time of creating a new state or political community, at some crucial moment in a process of democratization or within a general process of political regime change, as well as to partial or ‘incremental’ institutional reforms that can be promoted and implemented within more stable institutional frameworks. Generally speaking, we should expect that constitutional choices in favour of relatively inclusive and less hazardous electoral rules and procedures, and in particular in favour of proportional representation, would likely be confirmed in further periods. This tendency to institutional self-reproduction can be derived from actors’ self-interest since, once electoral systems producing multiple winners exist, it would be highly risky for the latter to bet on alternative rules favouring the production of a single absolute winner and many losers. Thus the same reasoning can be applied to founding or extremely innovative institutional decisions and to further stages. Restrictive rules can survive under conditions of low uncertainty among the electoral winners regarding their future capability of obtaining relatively favourable electoral results. In contrast, under alternative conditions of higher uncertainty about their future electoral strength in successive elections, they may prefer to introduce additional institutional changes likely to produce wider distribution of spoils. The reality of this model does not always require the open revelation of all actors’ preferences or even of explicit exchanges and agreements. Certain electoral reforms can be promoted by threatened incumbent rulers for self-interested motives even if they only anticipate their future troubles and are not forced to negotiate with a challenging opposition. But even if institutional choices are made ‘from above’ by rulers in control of the decision process, the new rules can increase the institutional strength of the opposition and favour the emergence of new challenging groups, which can lead the risk-averse incumbents to accept further innovative formulas producing wider distributions of power. In contrast, attempts to make the existing electoral system more restrictive are likely to find resistance from those groups whose formation, activity and institutional power respond to the incentives provided by the existing rules. From single-winner to multiple-winner rules Roughly speaking, and according to the classification and analysis that will be extensively developed below, electoral systems based on the majority principle, which tend to produce a single, absolute winner and subsequent absolute losers, must be considered a more risky choice than those based, say, on proportional representation, a principle forged to create multiple partial winners and much fewer losers than majority rule. Within each of these two principles, different formulas and voting procedures may also make a difference with regard to risk. For elections under the majority principle, voting procedures giving each voter the opportunity to vote for only a single candidacy or multiple candidates in bloc2 are a more risky choice for self-interested political actors than those permitting multiple votes, whether categorical, cumulative or ordinal, for different candidates or alternatives. Likewise, for elections based on the principle of proportional representation, certain formulas may be considered more risky because they tend to favour relatively

Strategy and History of Electoral System Choice

11

large groups to a greater extent than others, while voting procedures requiring from voters a single vote for a party’s closed list of candidates may also increase the probability of creating absolute losers than others permitting voters to choose both among parties and among individual candidates. A testable implication of this reasoning is that certain electoral systems producing a single absolute winner may remain stable for relatively long periods of time to the extent that a single dominant group or party can expect to repeat victories in the future, as well as in situations in which two rather balanced groups or parties expect to alternate as winners with high and somewhat regular frequency. Another implication is, however, that choosing, establishing or maintaining this type of electoral system can be a mistaken or highly risky decision for all those groups not expecting to be secure winners in the future, whether because they are minority groups in opposition or because they are threatened rulers with uncertain prospects. In the short term, thus, we should expect electoral system stability in many cases in which political parties are few, well organized and solidly entrenched among voters’ preferences. But in the mid or long term, electoral system change is likely to emerge. The corresponding results to be found in a long-term historical perspective should thus be increasing numbers and proportions of electoral system choices in favour of those formulas and procedures producing multiple winners, as well as a relative reduction of existing electoral systems producing a single absolute winner. Naturally, specific decisions will be constrained by the set of available electoral formulas at the time – although, as mentioned, new inventions are also fostered by the pressures of choice. Thus, within the period in which the unanimity principle was considered to be superior – to be examined and illustrated below – different procedures were gradually invented and adopted to make unanimous decisions feasible. These included, in particular, the acceptance of qualified majorities, such as those formed by wiser, older or more zealous people or by two-thirds or three-quarters of voters, as sufficiently compelling to make reluctant or small minorities withdraw their objections and acquiesce to a unanimous decision. Eventually, these artifacts were simplified into the rule that a simple numerical preponderance would in itself suffice to declare a winner or make an enforceable decision. Likewise, under the further dominion of the majority principle – also to be examined in the following pages – voting procedures producing absolute victories for a single group, such as the bloc ballot in multi-member districts by plurality rule, were progressively replaced with others permitting certain minorities to share power with the winners, such as the limited ballot in multi-member districts or single-member districts. Later on, some of these procedures became encouraging platforms for designing new, even less risky formulas able to produce proportional representation for different groups. Also under the principle of proportional representation, certain formulas and procedures – which will be analysed in detail below – were increasingly preferred and chosen over others for their higher ability to satisfy large majorities of voters and leaders and thus reduce the risks of losing. As mentioned, intermediate, mixed settings for less straightforward institutional choices can also exist. If, for instance, the existing winners have to concede to

12

Josep M. Colomer

disturbing opposition pressures but are still optimistic about their electoral prospects in a more open competition, they will try to persuade the opposition leaders to participate in the election with minimal institutional changes, and in any case with appropriate rules still giving the incumbents significant opportunities to win. To the extent that they can be rather pessimistic regarding their own electoral strength, they can propose or negotiate more open and innovative electoral rules and procedures, but in such a way that they can find themselves among the winners created by these more complex arrangements. Institutional combinations can include, for instance, ‘mixed’ electoral systems in which the likely winner shapes a high number of single-member districts by plurality rule while giving the opposition chance to compete for a portion of seats by proportional representation. Actually, many electoral system choices and reforms are embedded in larger sets of institutional choices, especially regarding the enlargement of voting rights to new population groups, the apportionment of seats with regard to population, the gerrymandering of electoral districts, the establishment of electoral authorities and tribunals oriented to reduce electoral fraud and other elements usually included in electoral laws, such as rules for candidate eligibility, restrictions on campaigning, party finance regulations, opposition access to public media and procedures for ensuring fair voting counts. In particular, the introduction of universal suffrage and processes of democratization have been, even since the late nineteenth century and in further waves, paramount occasions for incumbent rulers and challenging opposition groups to decrease the global costs of changing political institutions and, specifically, deal with innovative electoral rules and formulas. For this reason, the variables mentioned above as favourable to electoral system change – the restrictions of the existing system and a high number of effective parties – do not have a deterministic effect. Under favourable conditions, electoral system change will be more likely if a general political opportunity for institutional change exists, as when suffrage rights are broadened, democratization is fostered or a constituent assembly is called. In these situations, risk-averse actors may also try to enlarge the opportunities to compete for positions of power by creating multiple institutional levels to be submitted to elections, such as the separation of the presidency from the assembly, the embodiment of regional governments or decentralization into local units. This strategy may enlarge the scope of electoral system choices, but it can also reduce innovativeness at the national level. For instance, the degree of multipartism in the assembly can be constrained by the introduction of simultaneous direct presidential elections, which are always submitted to some majority rule and thus foster polarization – as happened, for instance, in France after the constitutional changes introduced during the period 1958–62. As another example, federalism or territorial representation in large countries with diverse populations may work as a substitute for proportional representation by giving different ethnic, regionally-based groups opportunities to enter institutions thus preventing a major electoral reform – which may explain the persistence of plurality rule elections in multi-party systems such as in Canada and India since independence or in Lebanon and Papua New Guinea in more recent times.

Strategy and History of Electoral System Choice

13

Specific analyses of electoral system choices, such as those that are presented in the region and country chapters of this book, need thus to place the question in the context of the global relationship of forces among the relevant political actors in parallel settings for multiple institutional choices. The general test presented at the end will also select all those opportunities for enlargement of voting rights or general regime change that can be considered favourable for a new electoral system choice. Further reading • Bawn, Kathleen (1993) ‘The Logic of Institutional Preferences: German Electoral Law as a Social Choice Outcome’, American Journal of Political Science, 37, 4: 965–89. • Boix, Carles (1999) ‘Setting the Rules of the Game: the Choice of Electoral Systems in Advanced Democracies’, American Political Science Review, 93: 609–24. • Colomer, Josep M. (ed.) (2001) ‘The Strategy of Institutional Change’, Journal of Theoretical Politics, Special Issue, 13, 3. • Dummett, Michael A. E. (1997) Principles of Electoral Reform. Oxford and New York: Oxford University Press. • Duverger, Maurice (1950) L’influence des systèmes électoraux sur la vie politique, Cahiers de la Fondation Nationale des Sciences Politiques, 16. Paris: Armand Colin. • Duverger, Maurice (1951) Les parties politiques. Paris: Seuil (English trans. Political Parties: Their Organization and Activity in the Modern State. New York: Wiley, 1954). • Elster, Jon, Claus Offe and Ulrich K. Preuss (eds) (1998) Institutional Design in Post-Communist Societies: Re-building the Ship at Sea. New York and Cambridge: Cambridge University Press. • Grumm, John G. (1958) ‘Theories of Electoral Systems’, Midwest Journal of Political Science, 2, 4: 357–76. • Lijphart, Arend and Bernard Grofman (eds) (1984) Choosing an Electoral System. Issues and Alternatives. New York: Praeger. • Lijphart, Arend and Carlos H. Waisman (eds) Institutional Design in New Democracies. Berkeley: University of California Press. • Miller, Nicholas R. (1983) ‘Pluralism and Social Choice’, American Political Science Review, 21. • Noiret, Serge (ed.) (1990) Stratégies politiques et réformes électorales: Aux origins des modes de scrutin en Europe aux XIX et XX siècles. Baden-Baden: Nomos (English trans. Political Strategies and Electoral Reform. Origins of Voting Systems in Europe in the 19th and 20th Centuries. Baden-Baden: Nomos). • North, Douglass C. (1990) Institutions, Institutional Change, and Economic Performance. New York and Cambridge: Cambridge University Press. • Przeworski, Adam (1991) Democracy and the Market. New York and Cambridge: Cambridge University Press. • Riker, William H. (1983) ‘Political Theory and the Art of Heresthetics’, in Ada Finifter (ed.), Political Science: The State of the Discipline. Washington, DC: American Political Science Association. • Riker, William H. (1996) The Strategy of Rhetoric: Campaigning for the American Constitution. New Haven, CT: Yale University Press. • Sartori, Giovanni (1994) Comparative Constitutional Engineering: An Inquiry into Structures, Incentives, and Outcomes. New York and London: New York University Press/Macmillan. • Shepsle, Kenneth A. (1986) ‘Institutional Equilibrium and Equilibrium Institutions’, in Herbert Weisberg (ed.), Political Science: The Science of Politics. New York: Agathon.

The invention of electoral rules and procedures Human history has witnessed continuous inventions and reinventions of electoral formulas. A somewhat evolutionary view tracing the historical panorama of the

14

Josep M. Colomer

invention and choice of electoral rule and procedure in disparate parts of the world in the course of several centuries will be presented below. The presentation will be ordered around four basic principles: unanimity, lottery, majority and proportionality – although in reality, of course, there also exist many intermediate formulas or ‘mixed’ systems combining two or more of these principles. Each of them can be associated with different virtues and vices, has been prevalent in different periods and has proved more adapted to some specific type of collective decision. But, in fact, the adoption of each of the principles, as well as the invention of the huge variety of rules and procedures with which they can be applied in the practice of voting and electing, have been broadly encouraged by actual opportunities and challenges of choice. For example, the disadvantages of unanimity rule in making effective decisions or electing a ruler – say, a pope in the high Middle Ages, for instance – moved those embroiled in such kinds of decisions and elections, like priests, monks and theologians, to explore, study, imitate and experiment with different variants of majority rule. Likewise, the drawbacks of majority rule in new mass elections during the nineteenth century – say, in parliamentary elections in countries with highly divisive social or ethnic issues – led politicians, mathematicians, lawyers and other scholars to rediscover or invent new procedures of voting, as well as formulas able to allocate seats and offices to different groups and parties instead of admitting only one absolute single winner. As will be mentioned in the following historical presentation, many voting procedures and formulas have been reinvented, even after several centuries of previous discoveries, under the pressure of finding new, better solutions to practical electoral problems. Intellectual creativity, which is usually fostered by necessity and convenience, has been, on the matter of electoral systems, typically the work of actual or prospective losers with strong motivation to win. Unanimity Approval of proposals and election of delegates by unanimity are almost instinctive procedures in relatively simple, rather homogeneous and not highly populated gatherings and assemblies. Families, groups of friends, urban gangs, neighbourhood meetings, corporation partners and club members tend to make collective decisions under conditions of general agreement. A variety of historical evidence also suggests that at a more general level Ancient Mesopotamian, Assyrian and Sumerian assemblies, Athenian and Spartan agoras, early Christian communities, Germanic tribes and communes, pre-Columbus Amerindian peoples, English and American small town-hall assemblies and other comparable units usually made collective enforceable decisions and reached agreements by virtual unanimity. Also, consuls were elected by traders, bishops by priests and believers, magistrates by citizens, etc., on the basis of large consensus. Decision procedures in these disparate communities included silent acquiescence, clashing spears against shields, shouts of commendation or acclamation, murmurs in favour or cries against the proposer, rising to one’s feet and other ‘viva voce’ expressions rather than formal voting sessions.

Strategy and History of Electoral System Choice

15

The invention of the unanimity principle More formally, regulation of decisions by unanimity may be found in the Justinian code of Rome, which established the principle that ‘What concerns similarly all ought to be approved by all’. Also the Christian Church adopted in the fifth century the principle that ‘He who governs all should be elected by all’. These principles, however, entailed some confusion between participation and decision, since general suffrage was associated with unanimous consent. As they would correspond, in modern times, to families, business companies or other corporations of private law, it was assumed that the members of those old communities had a clear common interest – to the fulfilment of which each of them contributed not on an egalitarian basis but according to their resources – and that a single decision in favour of the common good should thus be easily identified. Of course, organized factions or parties were not even conceivable. Individuals’ disagreements would lead to their exit, since they would reveal that the presumed common interest no longer existed. Lack of unanimity could thus produce secession, split, schism or divorce and the formation of new units with sufficiently small size and homogeneous composition to be able to reach unanimous agreements. In practice, however, several procedures were implemented to make decisions which could be accepted by all participants even in communities that were not highly homogeneous. These include the following: explicit acquiescence of the dissidents to the collective decision; preliminary voting followed by formal, public expression of the decision by all the community members; acceptance of decisions made by a qualified part of the voters to whom the others submitted (in accordance to members’ unequal contributions or different rank); and others. This eventually led to the replacement of the requirement for unanimity with qualified- or simple-majority rules. Regarding the need for acquiescence, minorities may consider the expression of their own position too costly if it requires, for instance, resisting the roar of acclamation, standing up in the middle of the assembly or leaving the room. Actually, many moral, legal or coercive devices can be implemented to force the minority to obey and make the collective decision binding. In certain ancient Middle East assemblies, those having revealed their dissident opinion during deliberation were required to kneel down in front of the assembly as a form of assent. In the medieval German communes, after the masters of households in the commune voted, usually an oath of membership and obedience was taken. In Nordic law, the dissident minority was threatened with punishment by means of a fine, as in Denmark, or exile, as in Iceland. In Russian law, even physical constraints were implemented. On a more sombre note, monarchist philosopher Gottfried W. F. Leibniz, as far back as the early eighteenth century, accepted citizens’ political participation in public affairs only as a way to make ‘the people themselves agree to what is good for them’, while preventing the arbitrariness that was found in assemblies ‘when cabals and animosities prevail over reason’. He praised practices, such as those in the United Provinces (Dutch) parliament, by which unanimity was reached through the force of persuasion and ‘friendly coming to terms’ (or

16

Josep M. Colomer

‘composition amiable’), but he also supported other ‘practical and quite efficacious’ means to deal with ‘stubborn and malevolent people’ resisting the common will (Leibniz, 1701). As in modern private communities, these different ways of forcing explicit acquiescence of the dissidents were means to assure that they would obey the collective decision, respect the elected or contribute with their corresponding effort or duties, in spite of their previous disagreements. Church, empire and assemblies Unanimous decisions were particularly hard to reach within the Christian Church, especially because the principle of unanimity was invested with a mystical and theological notion of the Church’s unity. Whereas voters’ unanimity was believed to be the way to discover God’s will, any failure to obtain such a total consensus was seen as the instigation of the devil. Dissent and disagreement were dominated by passionate condemnations under the assumption that ‘vox populi, vox Dei’ (taken from the Bible: ‘the voice of the people . . . is the voice of the Lord’, Isaiah 66: 6; also I Samuel 8: 7). In the first centuries of the Christian era it had been customary for bishops and other clerical bosses to appoint their successors, although a few choices by unanimous agreement under the direct inspiration of the Holy Spirit are recorded during the second and third centuries – including the descent of a dove to land upon the head of the selected candidate. Since the fifth century bishops, abbots and abbesses, and priests began to be elected, respectively, by the bishops and the clergy of the province, the monks and nuns of the monastery and the convent and the faithful of the corresponding diocese, parish or other jurisdiction – ‘the Christian people’ – typically by acclamation. Even the pope – actually the bishop of Rome – had to be elected by unanimity of the 16 bishops of the province for a long time (Eusebius Pamphilus 324, book 3, ch. 11; book 6, chs 29, 39). The requirement of unanimity, however, produced frequent violent conflicts, schisms and simultaneous elections of two (or more) different popes and anti-popes, even by the mid-third century. During schismatic periods, some monasteries had two abbots and two priors, some parishes two priests, and so on. An early attempt to replace such an ineffective electoral rule was introduced by Pope Simaccus in 500 who decreed that an elected pope should have the unanimous support of the clergy or, in case of division, the support of the majority, but this provision was hardly used. In fact, most of the popes from the fourth to the twelth centuries were appointed or confirmed by the Emperor, who had to act as arbiter in many electoral disputes. More successful was the qualification of decisions though the dignity of the voters who had supported them. From the eleventh century the college of cardinals of the Roman church was given the decisive role in papal elections. In order to attain a unanimous decision, priority was given to cardinal-bishops who were entrusted with gaining the assent of the cardinal-priests and the cardinal-deacons, as well as the approval of the other members of the clergy and the people. The principle of the sounder and greater part (‘sanior et maior pars’) – vaguely inspired by Saint Benedict’s rule for electing abbots – more solemnly gave qualified weight

Strategy and History of Electoral System Choice

17

to the preferences of those voters higher up the hierarchy with stronger zeal and greater seniority (including age and length of time in official capacity). Most canonists gave the interpretation that elections were valid if the ‘greater part’ included the ‘sounder part’. But the two parts often did not coincide, an occurrence which gradually moved decision-makers to lean to the side of qualifying voting results not by the dignity of voters but by numbers. It was Pope Alexander III who formally replaced the unanimity principle with the rule of two-thirds qualified majority for electing popes and bishops after 1179 (the election of bishops by the clergy was not officially abandoned until 1917). Monastic orders, including Benedicts, adopted the same rule. Alexander III probably took inspiration from previous experiences with less-than-unanimity rules for elections of abbots, as well as from Venice’s sophisticated voting procedure to elect the city’s Duke (Baldwin, 1968). For papal elections, the requirement of two-thirds majority was initially aimed at obtaining the support of at least two of the three orders of cardinals but it soon became just a numerical criterion. As formulated by Pope Gregory X, ‘Not zeal to zeal, nor merit to merit, but solely numbers to numbers are to be compared’ (VI Decretalium, book I, tit. VI, chap. 9, in Alberigo, 1991). Note that once a candidate is elected by two-thirds qualified majority, the losers would have to persuade a majority of the winner’s original supporters to change their mind. Faced with this requirement, the losing coalition could hardly have been expected to continue the fight. Divine inspiration of the corresponding choices was still claimed, as in Pope Pius II’s words about his own election in the mid-fifteenth century: ‘What is done by two thirds of the sacred college [of cardinals], that is surely of the Holy Ghost, which may not be resisted’ (Gragg and Gabel, 1959). But further canonists expeditiously identified the ‘sounder part’ with just ‘the greater part’ (Colomer and McLean, 1998). Similar problems emerged from the eleventh century on the occasion of elections of Frankish, Carolingian as well as Bohemian, Hungarian and Polish kings. These princes were elected by their peers, usually gathered at colleges of electors formed by dukes, marquises and counts, archbishops, bishops and abbots, at which unanimity was expected. Higher elections of the Holy Roman-German emperor followed the same model. But they ended with a split among the electors on three occasions between the end of the twelfth century and early fourteenth century, thus producing pairs of kings and anti-kings that undertook violent expeditions to affirm their respective rights. On the three occasions, the corresponding popes were called as arbiters – mirroring the role of the Emperor acting as arbiter in several divisive papal elections, as mentioned above. Two of the popes leaned to the side of the candidate they considered more qualified for the job having been supported by the ‘sounder’, although minority, part of electors. Thus here quality prevailed over quantity, a principle that would be defended theoretically by Marsilius of Padua who identified the ‘valentior pars’ of the princes’ college as representative of all (Defensor pacis, 1324). The defeated candidate for emperor in 1257, Alphonse X the Wise, king of Castile and Leon, who had received a majority of votes, remarked that the emperor

18

Josep M. Colomer

would have authority only ‘after being chosen by all those with choice power or by the major part’ (Alfonso X, Siete partidas, Especulo II, 1, 2). But the winner by both majority and soundness in 1314, Ludwig of Bavaria, insisted on the unanimity thesis that the elected by majority should be considered elected ‘in concord, not in discord’. According to successive regulations, the grand electors of the German emperor had to commit themselves to accept a majority decision as a unanimous one. But it was the Golden Bull introduced by Charles X of Luxembourg in 1356 which minutely regulated the election procedure, including the seclusion of the college electors in a locked chamber along the lines of previous experiences by the Dominican monastic order, several Italian cities and the papal conclave. The Golden Bull explicitly required the grand electors’ unanimity, the election being valid only if the minority conformed with the majority. The formal, public and sophisticated ceremony of investiture of the elected was actually the moment at which unanimity was constructed when even those having voted for losing candidates participated in a newly created common will (Ruffini, 1925). Comparable evolutions can be identified in late medieval towns, assemblies and parliaments called by kings and emperors, especially in Central and Southern Europe. From the eleventh century on, many civil assembly regulations established that decisions should be made by ‘consensus and acclamation’, ‘approval and consent’, ‘unanimity’, with ‘no discrepancy’, with ‘no contradiction’, by ‘free veto’, and so on. In assemblies of municipalities, as well as in their leagues, such as the German Hanse and the Swiss confederation, since a valid decision could not usually be reached if the minority persisted in its opposition, the subsequent rhetorical device was that ‘the minority must follow the majority’. The meaning was that the minority should withdraw its objection and agree with the majority in order that a unanimous common will could be formed. As mentioned, North Italian towns also made collective decisions by very broad agreements. In Florence, Genoa, Pavia, Pisa, Siena and Venice, the general assembly of all citizens approved the appointment of the Consulate, usually by acclamation. But the requirement of unanimous consent did not always produce quick, agreeable decisions in increasingly complex, urban settings, in contrast to easier general coincidences in simpler, more homogeneous, rural environments. Several mechanisms were thus invented to produce acquiescence where a single unanimous preference did not exist. Certain Italian communes had already adopted less-than-unanimity rules, such as those requiring two-thirds, four-sevenths or other qualified majorities, together with indirect elections in several stages and other devices. But ‘a major source of political frustration’ – as noted by a reputed historian – resulted from ‘a belief that unity was essential, and of an etiquette that frowned upon the exposure of irreconcilable viewpoints’ (Brucker, 1963). The Aragonese and the Catalan parliaments or ‘Cortes’ since the thirteenth century, as well the Polish Diet since the sixteenth century, were outstanding examples of enduring representative assemblies working by the unanimity rule. However, unanimous decisions also turned out to be difficult to make, especially when differentiation of social interest increased among the different corporative bodies

Strategy and History of Electoral System Choice

19

represented. Analogous procedures to those previously mentioned for other institutions developed. For instance, public and ordered voting rounds, starting with the higher or ‘sounder’ part, could be organized in the expectation of attracting voters from the other branches to jump on their bandwagon and forget their differences. The Parliament of Catalonia formally asked the king to replace the unanimity rule with that of the ‘major e de la pus sana part’. Several qualified majority rules were also introduced, such as those requiring the agreement of two of the estates, for instance the nobles and the citizens, to make decisions valid even in the face of disagreement or abstention by the clerical representatives. In general, however, majority decisions were considered a means of arriving at the necessary unanimity by imposing the duty of assent upon the minority (Coroleu and Pella, 1876; Konopcynski, 1930; González Antón, 1978; Ferro, 1987). Even the English parliament maintained the fiction of unanimity until very late on. Regarding the election of representatives, by the mid-sixteenth century a judge sentenced that ‘the election might be made by voices or by hands, or such other way, wherein it is easy to tell who has the majority, and yet very difficult to know the certain number of them’, so discarding the count of numbers and inducing the acceptance of the elected by a potential but not acknowledged opposition. Within the House of Commons, even as late as the early nineteenth century, it was commonly assumed that majority voting meant unanimity, most decisions being made by acclamation, that is by loudly shouting ‘yea’ or ‘nay’ (as was still observed, for instance, by Jeremy Bentham in his parliamentary tracts: Bentham, 1834). Only the formation of modern, well-organized political parties led to the adoption of more formal voting procedures requiring numerical precision. In modern times, the aim of achieving unanimous agreement has been transferred to the more ethereal world of international relations, where each of the sovereign parties has veto rights. Some of the other defining features of the earlier settings analysed above can also be found in international organizations: corporate (now state) suffrage; the search for consensus; burdens distributed in proportion to contributions or resources; offices held by turns or lots; decision-making limited to those issues in which a general common interest can be presumed (such as peace, monetary stability, environmental protection). Increasing complexity and the salience of more divisive issues, however, have also led to the adoption of les-than-unanimity decision rules.

Further reading • Benson, Robert L. (1968) The Bishop-Elect. A Study in Medieval Ecclesiastical Office. Princeton, NJ: Princeton University Press. • Buchanan, James and Gordon Tullock (1962) The Calculus of Consent. Logical Foundations of Constitutional Democracy. Ann Arbor: University of Michigan Press. • Colomer, Josep M. (2000) ‘On the Geometry of Unanimity Rule’, Journal of Theoretical Politics, 11, 4: 543–53. • Colomer, Josep M. and Iain McLean (1998) ‘Electing Popes: Approval Balloting and QualifiedMajority Rule’, Journal of Interdisciplinary History, 29, 1: 1–22.

20

Josep M. Colomer

• Esmein, Adhémer (1907) ‘L’unanimité et la majorité dans les élections canoniques’, Mélanges Fitting, vol. I. Montpellier: Imprimerie Générale du Midi, pp. 357–82. • Gaudemet, Jean (1960) ‘Unanimité et majorité (Observations sur quelques études récentes)’, in Études historiques à la mémoire de Noël Didier. Paris: Montchrestien, pp. 141–62. • Konopczynski, Ladislas (1930) Le liberum veto. Étude sur le dévelopement du principe majoritaire. Paris: Honoré Champion/Warsaw: Gebethner et Wolff. • Marongiu, Antonio (1968) Medieval Parliaments. A Comparative Study. London: Eyre & Spottiswoode. • Moulin, Léo (1953) ‘Les origines réligieuses des techniques électorales et déliberatives modernes’, Revue Internationale d’Histoire Politique et Constitutionelle, III: 106–48. • Moulin, Léo (1959) ‘Sanior et maior pars. Studio sull’evoluzione delle tecniche elettorali negli ordini religiosi dal VI al XIII secolo’, Studi Politici, VI, II, 1–2: 48–75. • Petrani, Alexis (1957) ‘Genèse de la majorité qualifiée’, Apollinaris, XXX, 1–2: 430–6. • Rae, Douglas W. (1975) ‘The Limits of Consensual Decision’, American Political Science Review, 69: 1270–94. • Ruffini Avondo, Edoardo (1925) ‘Il principio maggioritario nelle elezioni dei Re e Imperatori romano-germanici’, Atti della Reale Accademia delle Scienze di Torino 1924–1925, vol. 60, pp. 392–414, 441–92, 557–74. • Ruffini Avondo, Edoardo (1925) ‘Il principio maggioritario nella storia del Diritto Canonico’, Archivio Giuridico Filippo Serafini, IX. • Ruffini, Eoardo ([1927] 1976) Il principio maggioritario. Profilo storico. Milan: Adelphi.

Lottery The choosing of public officers by drawing lots is an old tradition that can be found in ancient and medieval local democracies, modern private settings and some international organizations – as has just been mentioned in the previous section. In all these cases, lotteries are a device to preserve the central role of the assembly of members to make decisions on the most relevant issues, typically by acclamation, assent or other forms of unanimity, on the assumption that the identification of a common interest should not be too difficult a task. The most relevant historical experience of selection of delegates, representatives or public officers by lots was developed in Athens during the democratic period from the mid-fifth century to the end of the fourth century BC. On the basis of this experience, the philosopher Aristotle built his concept of democracy, which included the possibility of ‘ruling and being ruled by turns’. Aristotle introduced a sharp distinction by which ‘the appointment of magistrates by lot is thought to be democratic, and the election of them oligarchic’ (Aristotle, 325 BC, book 4, passim). By ‘democratic’ he meant self-governed by the people, while ‘aristocratic’ pointed to the idea of government by the best – which could also lead to a perverse form of oligarchy. This classical criterion was retaken, many centuries later, by the French provincial Charles-Louis de Secondat, baron de la Brède et de Montesquieu, for whom ‘the suffrage by lot is natural to democracy’. According to Montesquieu, the advantages of making choices by lot are, first, that it ‘is unfair to nobody, and [second, that] it leaves each citizen a reasonable hope of serving his country’ (Montesquieu, 1748, book 2, ch. 2). Likewise, Genevan philosopher Jean-Jacques Rousseau, also in the eighteenth century, highlighted the role of lots in an ideal democracy, in which – according to his Athens-inspired, assembly-based model – public offices should be considered ‘a burdensome

Strategy and History of Electoral System Choice

21

charge’ and administrative acts be reduced as much as possible (Rousseau, 1762, book 4, ch. 3). The choosing of public officers by lots may have, thus, two types of advantage. First, by holding frequent choosings and establishing short terms of office, it can produce a high rotation of members in administrative or arbitral posts, thus preventing the formation and self-reproduction of a permanent, closed, rather impermeable elite – whether in the form of an economic oligarchy or just a class of professional politicians. By replacing public officials very frequently and opening public jobs to very wide layers of society, no one can be blamed for making or implementing unpleasant decisions – but no one can be praised either. Only the members’ assembly remains ultimately responsible for the consequences of collective decisions. Certainly, the principle of filling public offices by lots is contrary to the ambition to be ruled, in an ideal world, by ‘the best’. But in comparison to other actual processes of choosing rulers, the average quality of those chosen by lots may not be so different in fact from that obtained with less unpredictable procedures. Officers chosen by lotteries, especially when there is a large number of candidates, are, as a whole, ‘typical’ representatives of the citizens or potential candidates, in terms of their characteristics, capabilities and values. Mediocrity, thus, rather than excellence should be expected to prevail. But this is precisely the outcome which is aimed at with such a procedure. Only those offices requiring particularly high levels of professional or technical skill or experience have always been excluded from selection by lots. From this perspective, lots can be considered a clearly better and less risky procedure for selecting public officials than hereditary succession, cooptation by incumbent office-holders or the buying (or auctioning) of offices, by which the probability that outliers, eccentric or incapable candidates may be appointed is dramatically high. Also, likely winners by lottery may not be so different – in personal characteristics and professional skills – from those selected by certain modern political processes involving convoluted internal interactions of party political organizations, hazardous electoral campaigns and disparate voting procedures. This similarity, however, is usually veiled in the public opinion only by the retrospective attribution of talent or merit to those elected – if for no other reason than the fact that they have been able to get elected. The second type of advantage to choosing by lots and the subsequent rotation in public offices is that they produce wide dispersal of political and administrative knowledge of public affairs among the citizens. The experience of learning and becoming familiar with the problems of satisfying collective common interests can be a good platform for further occasions of participation in voting and elections, such as the assembly’s decision-making and the choice of some other public officials. Thus a lottery can be an appropriate procedure to select public officers where there is a clear identification of the common interest of the members of the community, there are relatively low technical requirements to fill the public jobs submitted to the lottery and there are strong instruments by which the community can make other, very important decisions, including control of those appointed by lots.

22

Josep M. Colomer

Athens, communes and international organizations Public choice by lots has never been an exclusive procedure, as mentioned. The Athenian democracy was based on the central role of the assembly or ‘agora’, open to all citizens, both to elect certain officers and to vote on war and peace and other important decisions. The frequency of assembly meetings, which were called by the blast of a trumpet and the lighting of a bonfire and were attended by a few thousand citizens, increased over the years from about ten to about forty a year. The scope of public decisions was also broadened until the assembly became the major legislator and decision-maker. Typically, most issues required only yes–no decisions, thus permitting the formation of a clear majority. Thus in different periods voting was possible mostly by acclamation but also by show of hands or pebbles (in Greek ‘psephos’ – from which is coined the typically British term ‘psephologist’ for students of and experts in elections). In contrast, elections to fill some highly specialized offices for which several candidates could run required more sophisticated procedures, including the casting of ballots in the form of pieces of tile or potsherd or by written waxed tablets in vessels inside a wooden enclosure. In order to preserve the central role of the assembly to make decisions by unanimity, the Council in charge of setting the agenda for the assembly, formed of 500 individuals, was appointed, first, by an electoral college elected by the assembly, later by annual elections, and then by turns in order determined by lot. The Council was formed of 50 members selected by lot by each of the ten tribes which had evolved from military into basic administrative units. The permanent committee of the Council and its president were also selected by lots. No citizen could hold any office more than once, so it is estimated that about half of the citizens aged 30 or more became members of the Council at some point in their lives. About 600 of the other 700 public officers were also selected by lots from among candidates previously presented. These included: the ten members of the archonship, approximately equivalent to the post of the modern attorney-general, as well as the body in charge of organizing religious ceremonies who were appointed by lots from a pool of candidates previously selected by each tribe, also by lots; the tribunal members, chosen by lots from a pool of all adult citizens, who were in charge of passing judgement on the legality of the conduct of public officials; and a number of administrative jobs, encompassing treasurers, those in charge of settling public contracts and collecting public revenues, and those supervising streets or inspecting markets. Only a few offices requiring special technical skills and qualifications, including high military commands and financial administrators, were excluded from lotteries and submitted to direct election. In contrast to those appointed by lots, the elected for these latter posts were, of course, not ‘typical’ representatives of the citizenry, but they were also subjected to stronger scrutiny and accountability at successive elections. The procedure of selecting candidates for public offices by lots was based, initially, on candidates drawing white and black beans from a container with an open top, the total number of beans being equal to the number of candidates and

Strategy and History of Electoral System Choice

23

the number of white beans equal to the number of places to be filled. In a further development, Athenians also used allotment machines, usually a tube in which balls could be inserted at random and released at the other end (Headlam-Morley, 1981). The choice by lots implies the acceptance of average mediocrity as a collective criterion for choosing public officers, as well as some uncertainty regarding specific outcomes and decisions. On some occasions, it was, however, covered by the belief that it would involve some divine sanction, in other words that choosing by lots was the way to make right decisions as directed by the gods. The Greek philosopher Plato, most notably, praised lots as a way to let the gods make their choices, especially for officers of the temples (Plato, 360 BC, book 6). The Old Testament also included a number of instances in which decisions by lot – as did other events that seemed accidental – were meant to reveal God’s will (Proverbs 16: 33). The Lord also instructed Moses to divide the land by lot among the different tribes (Numbers 16: 52–56, 33: 54). The first apostles of Jesus also drew lots to select the substitute for the traitor Judas (Acts of the Apostles I: 23–26). Apparently, they were imitated by some early non-orthodox Gnostic Christians, who drew lots at each of their meetings to elect priests, bishops and other officers – only for them to be replaced at the following meeting. This device could also be aimed at preventing a sacerdotal oligarchy from developing (Goodwin, 1992: 44). But the Christian Church condemned such a practice as blasphemy and solemnly forbade the choice of priests, bishops or other prelates by lot (more formally after the thirteenth century). Lots have also been used for choosing public officers in other settings based on an open assembly making decisions by virtual unanimity, typically by acclamation, with similar precautions to those previously referred to in order to prevent the formation of oligarchies and voter manipulation, and always in combination with other election procedures. The Roman Republic, for instance, had used lots to establish the order of voting by tribes or other political units, but most decisions were made by voting procedures involving majority rules – as will be explained in the next section. A number of Italian communes – which were identified in the previous section as examples of where decisions were made by assembly acclamation – also used lots after the thirteenth century for choosing magistrates and allocating officers in charge of implementing assembly decisions. This was, in particular, the case in Venice for the indirect election of the Duke (Doge) from the thirteenth century – following direct election by the people’s assembly from the end of the seventh century. The popularly elected Great Council adopted an increasingly complicated procedure to choose the Duke with up to nine stages of approval ballots and lots, which was conceived with the aim of making manipulative manoeuvres impossible (Lines, 1986). Likewise, the Florentine republics during the fourteenth and fifteenth centuries, and again in the late fifteenth and early sixteenth centuries, elected its main governmental body, the Lordship (Signoria), chaired by the standard-bearer of justice (Gonfaloniere), by means of a complex system of approval ballots and very frequent lots. Again, the aim was to prevent fraud, manipulation and the commune’s domination by a few powerful families.

24

Josep M. Colomer

Also, in Barcelona, the popularly elected Council of 100 chose the Consuls of Sea, that is the judicial body for commercial and maritime affairs, by an indirect procedure involving lots. Certainly in the fifteenth century, and probably earlier, the procedure required a child less than seven years old to distribute beans from a jar full of water and covered by cloth among the candidate-members from each of the estates – the ‘honest’, the merchants, the artists and the artisans – the number of white beans being equal to the number of electors to be selected from each estate. Afterwards, the electors so chosen, together with the city councillors, elected the consuls by casting ballots in an urn (Carrère, 1967). Finally, lots were still being used, in combination with several stages of indirect elections, in Spain and the Spanish colonies in the Americas in the early nineteenth century. First, elections called by the central Junta formed to organize the resistance against Napoleon’s troops were held in 1809, with municipalities electing candidates for deputies which were finally selected by lot. New elections in 1810 to form an extraordinary assembly (‘Cortes’), which gathered in Cádiz and produced a new constitution, were also held by a combination of indirect elections in three stages and a final selection of one deputy in each district by lots among the two or three candidates previously chosen (Fernández, 1992; Rueda, 1998). After the approval of the so-called constitution of Cádiz in 1812, this type of procedure was not used again in Spain, but it was followed in some further elections in Spanish America. Specifically, in Buenos Aires, indirect elections of colleges (usually called ‘juntas’) led to the final selection of members of the provincial assembly by lots, according to regulations enforced in 1811, while a mixed procedure of voting, lots and final popular vote by plurality was used for the election of governors in 1815. In Chile a combination of lots and plurality voting was still being used in 1822 (Chiaramonte, 1995). In Mexico, local elections in 1812 still involved some stage of selection of candidates by lots (Emmerich, 1985). Lots remained the usual practice in indigenous communities that were not politically integrated into the new independent state’s political institutions. Ironically, they became part of the supposedly traditional ‘usages and customs’ of the indigenous people to be preserved in the twentyfirst century – but were actually the most visible legacy of Spanish colonial rule (Favre, 1986). In modern times, lots are only used as a method to distribute goods and responsibilities in some private corporations, as well as for allocating temporary jobs, vacation periods or household tasks in other private settings. For public affairs, they are used in certain countries for such tasks as selecting jurors for jury trials, for appointing election administrators, for breaking election ties or for selecting candidates for military service. At the more overarching political level, certain institutional settings formed by units retaining their sovereign rights, that is confederal or international organizations making decisions by near-unanimity rules, use different procedures of rotation by turns of high public offices which produce the same effect as lotteries without replacement. A paramount example is the Helvetic Confederation of Switzerland, which is still mainly an instrument for preserving popular self-government of the cantons. Nowadays, it is governed by surplus multi-party coalitions in the Federal

Strategy and History of Electoral System Choice

25

Assembly and the Federal Council, usually encompassing more than 80 per cent of votes. While they make very broad consensual decisions under the threat of revocation by referenda in the cantons, the presidency of the Federal Council – comparable to the republic’s president – is filled in rotation among the Council members by turns. Likewise, in the European Union, the European Council, which is formed by the chief executives of the member states, and the Council of Ministers (officially called Council of the European Union), formed by representatives of the member states at ministerial level, have very important agenda- and decisionmaking powers. While they usually make decisions by unanimity or highly qualified majorities, the Presidency is held by six-month turns among the member states. Finally, the United Nations Organization (UNO) also distributes some high offices by informal rotation and turns among its member states. The Presidency of the General Assembly is filled by ‘symmetric rotation’ among countries of the five regional groups (Africa, Asia, Latin America and the Caribbean, Western Europe and other developed countries, and Eastern Europe). The Security Council, which works by near-unanimity decisions of the five permanent members (France, Russia, China, the United Kingdom and the United States), is also formed by a number of temporary members rotating in post for periods of two years. They are formally elected by the Assembly, but they must also be distributed fairly from among the different world regions. (In the initial period from 1946, there were two for Latin America, and one for Western Europe, the British Commonwealth, Eastern Europe and the Middle East. After general decolonization and the subsequent enlargement of the UNO membership, ten temporary members have been distributed, since 1963, in the proportions of two for Latin America and the Caribbean, two for Western Europe, one for Eastern Europe and five for Africa and Asia.) Similar proportions are used for filling, by informal turns, the posts in the Economic and Social Council and other committees (Marín-Bosch, 1994). Further reading • Burnheim, John (1981) ‘Statistical Representation’, Radical Philosophy, 27, Spring. • Elster, Jon (1989) Solomonic Judgments. New York and Cambridge: Cambridge University Press. • Gataker, Thomas (1619) On the Nature and Use of Lots: A Treatise Historical and Theological. London: E. Griffin. • Gibbard, Allan (1977) ‘Manipulation of Schemes that Mix Voting with Chance’, Econometrica, 45: 665–81. • Goodwin, Barbara (1992) Justice by Lottery. London: Harvester-Wheatsheaf. • Headlam-Morley, James W. (1981) Election by Lot at Athens. Cambridge and New York: Cambridge University Press. • Manin, Bernard (1995) Principes du gouvernement représentatif. Paris: Calmann-Lévy (English trans. The Principles of Representative Government. Cambridge and New York: Cambridge University Press, 1997). • Mulgan, Richard G. (1984) ‘Lots as a Democratic Device of Selection’, Review of Politics, 46, 4: 539–60.

Majority Majority rule had already been used, in combination with other rules and procedures, for collective decisions and the election of public officers in some

26

Josep M. Colomer

ancient and medieval institutions. But the majority principle was rediscovered and more explicitly invested with the power to create authority in more modern times. During the Roman Republic, that is for about five centuries immediately before our era, many collective decisions were made and most public officials were chosen by procedures involving majority rules. As in other historical processes previously discussed, majority decisions emerged from the truncated traditions of unanimous decisions, typically made by acclamation. Specifically, the ‘centurial comitia’, organized into 193 units of 100 men each, were transformed from military units in the army into units of vote in a political assembly open to all citizens. The traditional show of allegiance made by the soldiers to their commanders eventually became formal voting sessions to declare war, legislate, elect consuls and censors (in the executive), praetors (in charge of provincial governments) and temporary dictators, as well as to dictate capital judicial sentences. Later on, the ‘tribal comitia’ by which the people elected its tribunes by acclamation also evolved into ‘councils of the plebs’, organized in up to 35 large tribes, from which patricians were excluded. They held frequent voting sessions to elect ‘ediles’, ‘quaestors’, lower offices and special commissioners promoting specific demands, to approve general legislation and ‘plebiscites’, and to impose sentences on crimes against the state and on all those punishable by fines. These early experiences of voting and elections with majority procedures should be analysed as forms of indirect elections, having developed the same kind of biases and arbitrary outcomes as when elections in two or more stages were developed again in more modern times. In the Roman experience, first, each ‘centuria’, tribe or voting unit made an internal decision. Oral voting by acclamation was eventually replaced, especially in the plebeian tribes, with individual oral voting before a questioner and, later, with written ballots cast into a urn made of wicker or stone. When several officers had to be elected at the same time, each unit worked as a multi-member district with individual multiple ballots in which decisions were made by simple plurality or relative majority. Once the units had reached internal decisions, then voters were summoned to the Campus of Mars, Capitol Hill or the Forum and each unit’s vote was publicly announced in turn. An alternative or candidate could succeed by winning by a majority or plurality in a majority of units, that is by obtaining, for instance, as little as one-fourth of the total citizens’ votes. Usually, proceedings terminated as soon as an alternative obtained sufficient support or a number of candidates equal to the number of places to be filled had been proclaimed, thus ignoring the size of the winner’s support and the number of votes cast in favour of other alternatives or candidates. The decisive unit’s vote was openly celebrated, usually by acclamation, as a moment of revelation. The winner became the focal point around which to gather general support. For this reason, although in the previous aristocratic ‘curial comitia’ wealthier groups had enjoyed precedence in the voting and thus in making winners, in centurial and tribal assemblies the voting unit which was called upon to vote first was selected by lot, so ensuring some rotation and fair distribution of decisionmaking power.

Strategy and History of Electoral System Choice

27

As can be seen, the most relevant features of these procedures were much later replicated in modern formulas for indirect elections, as in the US electoral college to elect the president, as well as in the further processes of primary elections and party conventions to select presidential candidates, always organized under majority rules. Common developments under this kind of framework included victory by majority at later stages with a minority of popular votes, random selection of decisive voters and the building of instant unanimity around the proclaimed winner. But direct connections between ancient and modern practices of voting by majority rules are difficult to establish, since the fall of the Roman Republic and Empire apparently broke the continuity with and even any memory of those earlier institutional procedures. The reinvention of majority rule In medieval times, decisions and elections by qualified and simple majorities were again gradually accepted in Germanic law and the Church’s canon law as an expedient procedure when unanimous decisions turned out to be impossible, as discussed in the first section of this chapter. Initially, the minority was summoned to acquiesce to the majority will as an expression of unanimous support, as illustrated above in several ways. But the formation of reluctant, repeated, stable, defeated minorities eventually tended to make majority decisions acceptable and presumably enforceable even if the minority did not explicitly assent to share a unanimous common will. Thus majority rule typically emerged as an acceptable decision rule from the experience of two-sided contests – respectively becoming the winners and the losers – rather than the other way around (that is, as if two-party systems were ‘created’ by majority rules). When the losers were divided among several sides, they would seek the adoption of more inclusive electoral rules, as will be discussed in the following section. Thus the majority principle seems to be intrinsically associated with the previous formation of relatively stable factions or parties within the community, logically requiring the respect of minority will. However, in a similar way to that in which certain medieval theologians had praised unanimity rule for its presumed capability to reveal God’s will, early modern naturalist philosophers also associated majority rule with the discovery of the people’s will, that is the new presumed truth and good. Hugo Grotius, for instance, explicitly stated in the early seventeenth century that ‘the majority would naturally have the right and authority of the whole’ (Grotius, 1624: book 2, ch. 5, 17). The Englishman John Locke, by the end of the same century, formally established that, in a civil society formed on the basis of every individual’s rights and consent, collective decisions should be made ‘only by the will and determination of the majority’. In particular, he postulated that, in representative assemblies, ‘the act of the majority passes for the act of the whole and determines, as having by the Law of Nature and Reason, the power of the whole’ (Locke, 1689: II, 96). Likewise, the Genevan Jean-Jacques Rousseau believed, in the mid-eighteenth century, that the general will, which was supposed to be ‘always right’, could be discovered by voting by majority rule, so that ‘the vote of the majority always binds the rest’ (Rousseau, 1762: book 4, ch. 2). Still, the French revolutionary abbot Emmanuel-Joseph

28

Josep M. Colomer

Sieyès praised majority as an ‘incontestable maxim’, while daringly identifying the majority, that is ‘the third estate’, with ‘all’ (Sieyès, 1789: ch. 6). In parallel to these honours, recourse to majority rule for simple reasons of expediency was also argued by other theoreticians, such as the Englishman Thomas Hobbes in the mid-seventeenth century. ‘If the representative’, said Hobbes, ‘consists of many men, the voice of the greater number must be considered as the voice of them all. For if the lesser number pronounce, for example, in the affirmative, and the greater in the negative, there will be negatives more than enough to destroy the affirmatives, and thereby the excess of negatives, standing un-contradicted, are the only voice the representative has’ (Hobbes, 1651: ch. 16). Thus the weight of force, rather than the search for truth, was, at least in the English and Scottish utilitarian tradition, another very good reason to adopt majority rule. Actually, in late medieval and early modern times collective decisions were expeditiously made by majority in local communes, commercial organizations and other assemblies. At some point in time, they would come to be made not by an absolute majority rule, that is by requiring more than half of the voters to coincide on a single alternative, but by adopting the alternative supported by only a relative majority or ‘plurality’, that is by a higher number of voters than any other alternative but not requiring any particular number, proportion or threshold of votes. In practice, this made it possible for generally binding elections presumably decided by ‘majority’ to actually be won by only a minority of voters. Most notably, a very long-lived statute adopted in England by the mid-fifteenth century for the election in the shires of members of Parliament had established that the sheriff in charge of electoral assemblies, after listening to the shouts of supporters of different candidates or counting by heads, should declare elected those with ‘the greatest number’ of supporters. This was probably intended as a means not to have to count votes, especially those supporting defeated candidates, as mentioned before. But, while the fiction of unanimous decisions was maintained, especially within the parliament, this regulation permitted in practice the popular election to be won by a relative, not absolute, majority. Other than deciding by simple plurality, there were different proposals to make majority rule enforceable, although many of them were not widely adopted in the practice of voting and elections. Old procedures include the very early proposal of exhaustive pair-wise voting presented in the thirteenth century by the Catalan philosopher Ramon Llull. By this procedure, an election was to be made by holding multiple rounds of voting between pairs of candidates and declaring the winner the one having won the greatest number of pair-wise comparisons. In different texts, Llull presented this procedure and some variants in a scholarly, general manner, and also with narrative illustrations, possibly taking inspiration from the elections of abbots and abbesses in monasteries and convents in which such a procedure had been used (Llull, c.1274, 1283, 1299). A similar, normatively more ambitious, but much less applicable procedure was devised five hundred years later in France, under the pressures of Enlightenment and Revolution, by Jean-Antoine-Nicolas de Caritat, Marquis de Condorcet (1785).

Strategy and History of Electoral System Choice

29

The Condorcet procedure requires a winner, by majority principle to be decided, to be preferred in all pair-wise comparisons, that is by majority against every other candidate. When the candidates are perceived by the voters as ordered along a single linear dimension (such as the left–right axis in modern times), the winner by Condorcet’s exhaustive pair-wise voting is always the one preferred by the median voter. This is so because, by definition, the median voter is the one having less than half of the voters on one side (such as the left on the ideological axis just mentioned) and less than half on the other side (the right), in a way to be always included in any consistent majority of voters. Thus the Condorcet-winner, when it exists, can be considered to be highly efficient in terms of social utility, since the median voter’s position is the one for which the sum of distances from all the voters’ preferences is minimum. However, in non-linear ideological or policy spaces, that is those involving multiple dimensions, the Condorcet procedure may fail to produce a winner since no candidate can be capable of winning against every other in pair-wise comparisons. The frequency of failures with the Condorcet procedure depends on the number of issue-dimensions in the policy space and the number of candidates. Assuming that all possible distributions of voters’ preferences are equally likely (that is, a high multi-dimensional space), computing calculations have shown that the Condorcet procedure of voting would fail to give a winner in about 9 per cent of times for three candidates, but increasing more than proportionally, reaching, for instance, 32 per cent of times for six candidates (Fishburn, 1973). Although Condorcet himself had also sketched a simplified, more efficacious version of his procedure close to Lull’s (Condorcet, 1787), the procedure usually associated with his name has never been used in mass political elections, especially on account of its ineffectiveness in producing a decision or a winning candidate. However, its hypothetical results can be used as a reference for evaluating results produced by other, more expedient rules and procedures – as will be suggested below. Specifically, we will refer to the likelihood that other simpler and more effective procedures may produce the Condorcet-winner, that is the hypothetical winner in pair-wise comparisons against every other alternative, as well as the Condorcet-loser, that is the alternative which would be defeated by every other alternative in pair-wise contests and can thus be considered the least socially efficient. A comparable voting procedure, currently known as rank-order count, was also designed in the late Middle Ages by another Catholic thinker, Nicholas of Cusa (Nicolaus Cusanus). He had been motivated by failures in the elections not only of Christian popes but also of Holy Roman-German emperors, as discussed above. According to this author, his procedure would require the voter to mark ‘after the name [of each candidate] a clear digit – 1, 2, 3, and so on . . . until he comes to the candidate who is in his judgment the best’. The winner is the candidate having collected the highest sum of points (Nicolaus Cusanus, 1433: book 3, ch. 37). About 350 years later, the same voting rule was devised by a member of the French Academy of Sciences Jean-Charles de Borda. He also presented his invention with an alternative procedure of voting similar to the one introduced by Condorcet

30

Josep M. Colomer

(who, as permanent secretary of the Academy, was in charge of publishing it). An exhaustive pair-wise comparison of candidates would be developed, but the scores for each candidate against the others would be aggregated – the result being the same as with the rank-order count (Borda, 1781). As in the other procedures just mentioned, the invention of Nicolaus and Borda thus also required from every voter a complete ordering of preference of all the candidates. But while, as with Llull’s procedure and in contrast to Condorcet’s, the Nicolaus/Borda procedure of rank-order count always delivers a clear winner, it is highly vulnerable to insincere or ‘strategic’ voting – in contrast to both the Llull and Condorcet versions of pair-wise comparisons. Those with information regarding other voters’ preferences can give lower points than would actually correspond to their sincere preference to rival alternatives with high probabilities of winning thus truncating the expression of voters’ preferences. This weakness induced the two inventors of rank-order count to warn potential users with similar notes of caution. While Nicolaus stated that ‘no method of election can be conceived which is more holy, just, honest, or free . . . if the electors act according to conscience’, Borda remarked that his procedure had been conceived ‘only for honest men’. More practical proposals emerged during the French revolutionary Convention in 1793, which although they were not finally incorporated into the subsequent constitution, were somehow reinvented or adopted later for mass political elections in different countries. Condorcet himself proposed two innovative procedures, both conceived as more feasible substitutes in terms of time and the enlightenment required of voters, for his ‘rigorous method of majority judgments between the candidates taken two by two . . . [which] is often impracticable’. The first was a form of limited ballot by which the election of a number of representatives in a multimember district would be made by giving each voter a lower number of votes than seats to be filled – specifically, in primary elections for indirect formation of the assembly, each voter would have two votes to elect between 9 and 18 representatives in the district in Condorcet’s constitutional project. By declaring elected those with higher numbers of votes, this formula would allow representation from more than one party in each district, thus introducing some degree of pluralism. The second procedure was an apparently slight variant of the first by which the two votes per voter would be distinguished between first and second preferences. Only if no candidate obtained a majority of voters’ first preferences would the second votes be counted – that is the procedure currently known as supplementary vote (Condorcet, 1793a, 1793b). In the same setting, and as a reply to the Girondin constitutional project mostly written by Condorcet, the Jacobin constitutional project presented by AntoineLouis-Leon de Saint-Just proposed voting in multi-member districts with single individual votes. Specifically, he proposed to elect the National Assembly of France in a single national district with 341 seats by giving each voter one vote for only one candidate, and declaring elected all those 341 candidates having obtained the greatest numbers of votes (Saint-Just, 1793: ch. V). This formula, in modern times known as the single non-transferable vote, is usually considered capable of approaching

Strategy and History of Electoral System Choice

31

proportionality of representation very closely. But, in the Saint-Just version, it would obviously have produced the highest level of ‘multi-partism’: likely to be only individual representation, with no collective organization of representatives at all. Multi-member districts In fact, many of the theoretical inventions just mentioned were not directly or immediately relevant for actual choices of voting systems based on the majority principle in real committees, councils, assemblies and parliaments, which in general tended to adopt the simplest possible devices. One of the most basic, almost spontaneous, oldest and widespread procedures of voting to choose delegates or representatives from a community can be described with the following features. All people with a common interest in the election gather in assembly at their common place; they agree to choose a committee or council to be formed by a few individuals; each voter selects one or more candidates of his or her preference (the number of votes per voter being equal to, lower or higher than the number of posts to be filled); the candidates with the highest numbers of votes are elected. This is, basically, the type of procedure that is very frequently found in elections at schools, universities, housing condominiums, neighbourhood associations, trade unions, professional organizations, corporations and some local elections in small towns. At the broader level, it is also the type of procedure which has been used since the thirteenth century for elections of officers and deputies in German and Swiss communes and cantons (Lloyd, 1907), in French municipal assemblies, provincial Estates and the Estates-General (Babeau, 1882, 1894; Cadart, 1952), in English counties, cities and boroughs to form the Parliament (Bishop, 1893), and in Virginia and other North American colonies in other periods (Sydnor, 1962; Pole, 1971). This voting procedure was also used for the indirect election of assemblies, parliaments and presidential colleges in most states of Europe, the United States and Latin America during most of the nineteenth century. In some cases, only those candidates receiving an absolute majority of votes were initially elected and new rounds of votes were taken for each of the remaining seats; the latter could be filled by requiring again a majority and, if no candidate attained such support, by holding a second-round runoff between the two most voted candidates. Voting in some of these historical institutions was by show of hands, while in others voters expressed orally their choice one by one, but secret vote in written ballots can, of course, also be compatible with this type of rule. Political science has created a precise jargon which would define this basic procedure as consisting of multi-member districts, open ballot and simple plurality or absolute majority rules. It can also include different forms of approval ballot, whether permitting a vote for as many candidates as wished or only for as many seats as are to be filled. We know from both logical analysis and empirical evidence that this type of procedure can be highly inclusive and permit somewhat varied representation, since different preference groups encompassing in total broad proportions of voters can see their chosen candidates elected. Spontaneity in defining the rules – which look almost ‘natural’ to the unsophisticated eye – and inclusiveness of actual results can explain the duration over several centuries of this

32

Josep M. Colomer

type of procedure in very different institutions and organizations such as those mentioned, as well as its frequent adoption in non-political, small-scale associations in all periods. The results of these and the following procedures by majority rules are illustrated in stylized manner in Table 1.1. As can be seen in the very simple example of a three-member district with one hundred voters, representation by open ballot, by which each voter can vote for as many candidates as he or she wants up to the number of seats to be filled (in this case from one to three), can fall into the hands of three candidates holding different preferences (in the example in the table the seat allocation is preference x: 1 seat, preference y: 1 seat, and preference z: 1 seat).

Table 1.1

Majority rules and procedures

Candidacies w1 w2 w3 x1 x2 x3 y1 y2 y3 z1 z2 z3 W X Y Z

– Individuals: – Parties: Total voters:

40

30

20

20 20 0

10 10 10

0 0 20

10/100

Per area:

A B C

Rules

Votes

M = 3, Open ballot, Plurality rule

e.g. w1: 40, x1: 30, y1: 20, z1: 10

1

1

1

0

M = 3, Block ballot, Plurality rule

W: 40, X: 30, Y: 20, Z: 10

3

0

0

0

M = 3, Block ballot,

1st round: W: 40, X: 30, Y: 20, Z: 10 2nd round, e.g.: W: 40, X: 60

0

3

0

0

M = 3, Limited ballot (V = 2), Plurality rule

w1: 40, w2: 40, x1: 30, x2: 30, y1: 20, y2: 20, z1: 10, z2: 10

2

1

0

0

M = 3, Limited ballot (V = 1 single non-transferable), Plurality rule

w1: 40, x 1: 30, y1: 20, z1: 10

1

1

1

0

M = 3, Cumulative ballot, Plurality rule

w1: 40–120, x 1: 30–90, y1: 20–60, z1: 10–30

1–3

0–1

0–1

0

M=1, Single ballot, Plurality rule

A – w1: 20, x1: 10, y1: 0, z1: 5 B – w2: 20, x2: 10, y2: 0,z2: 5 C – w3: 0, x3: 10, y3: 20, z2: 0

2

0

1

0

Majority-runoff rule

5 5 0

Three-seat allocations

Strategy and History of Electoral System Choice

33

The consequences of this simple type of election procedure can change dramatically if well-organized factions or parties are formed. With this we are just referring to groups of candidates and voters with a common purpose that coordinate their behaviour in order to win as many possible seats in the election. In the type of multi-member election previously sketched, the emergence of factional candidacies able to attract some voters’ support may produce a set of winners corresponding to only one of the preference or interest groups in the community at the expense of all the rest. To attain this result, it is not necessary that most or even many voters follow the advice of faction leaders to vote for all the members of a list of candidates and only for them; it is sufficient that a few people do so, since, even if they are very few in number, they can make a difference in comparison to the previous scenario without organized factions and make all the candidates in the list winners. This is, of course, due to the fact that plurality rule does not require any specific amount of votes to make a winner only one more vote than the other candidates. In face of an organized faction able to accumulate all the seats in contest, the best corresponding strategy by other candidates and voters is, certainly, the organization of comparable factions able at least to neutralize the advantage obtained by the first faction. This will eventually transform the competition from one among individuals to one among factions. Voters then vote in bloc for a list of candidates rather than choose their preferred individuals. In more precise terms, the procedure of voting moves from open ballot to bloc ballot or closed list. The new distribution of seats in a multi-member district by plurality rule when collective candidacies are organized is also shown, by way of simple illustration, in Table 1.1. In contrast to previous results with open ballot, which was able to produce, as in the example, an allocation of 1–1–1 seats to three different preference candidates, now plurality bloc or closed list ballot produces an allocation in which all the seats are given to the candidates of the largest faction: 3–0–0, even if it has obtained the votes of only a minority of voters (40 out of 100 in the example). The introduction of absolute majority rule with a second round between the most voted candidates at the previous round may produce similar results if all major candidacies induce voting in bloc for a list of candidates. Majority rule with a second- or third-round runoff was used early on in indirect elections to multimember districts in France for the election of the Estates General in 1789, and in direct elections to the French Assembly until 1871, as well as in at least six other European countries during the second quarter of the nineteenth century. As illustrated in Table 1.1, these rules and procedures also concentrate all the seats on a single factional candidacy – the winner at the second round, which can be different from the winner by plurality at the first round (in the example in Table 1.1, the corresponding allocation of seats is 0–3–0). This perspective may shed light on why ‘faction’ is usually associated in classical political literature with bad intentions to the disadvantage of general or at least broadly collective interests. In contrast, ‘party’ has enjoyed somewhat better consideration in modern times – at least in academic circles. But actually early political

34

Josep M. Colomer

parties were formed just as electoral factions and they are still so, especially if the electoral system induces or forces voters to vote in bloc for a closed list of candidates, despite having developed different kinds of complex organization. In some early modern times, when elections in relevant large-scale political institutions were of the type presented above, the distinction between ‘faction’ and ‘party’ was difficult, as witnessed, for example, by the Scottish and British David Hume who – reflecting a general mood of the time – initially considered that ‘sects and factions [should] be detested and hated .. . [because they] subvert government, render law impotent, and beget the fiercest animosities’ (1741). Later on, however, he pondered that ‘to abolish all distinctions of party may not be practicable, perhaps not desirable, in a free government’. Very interestingly, when in a later period Hume had to face the undesirable but perhaps unavoidable existence of political parties, he turned to wish that government were in the hands not of a single party but of multiple-party coalitions in order ‘to prevent all unreasonable insult and triumph of the one party over the other, to encourage moderate opinions, to find the proper medium in all disputes, to persuade each that its antagonist may possibly be sometimes in the right, and to keep a balance in the praise and blame, which we bestow on either side’ (1758). A similar evolution can be detected in the Virginian and American James Madison – who was actually influenced by reading Hume – who initially also condemned factions but later saw them as unavoidable and even a necessary evil. Madison perceptively noted that, with organized parties, in small communities there would likely be a single dominant party embedded in local prejudices and schemes of injustice – and ‘the smaller the number of individuals composing a majority, and the smaller the compass within which they are placed, the more easily will they concert and execute their plan of oppression’. In contrast, greater political units would provide ‘the greater security afforded by a greater variety of parties, against the event of any one party being able to outnumber and oppress the rest . . . In an equal degree does the increased variety of parties comprised within the [large] Union increase this security’ (1788, No. 10; see also the discussion in Dahl, 2002: 29–37). Perhaps it was Jean-Jacques Rousseau who best realized and anticipated the dramatic influence of factions in elections held in multi-member districts by plurality rule – as used in his native Geneva. ‘When factions arise’, he observed in a rather well known but diversely interpreted passage, ‘and partial associations are formed at the expense of the great association, the will of each of these associations becomes general in relation to its members, while it remains particular in relation to the State: it may then be said that there are no longer as many votes as there are men, but only as many as there are associations . . . Lastly, when one of these associations is so great as to prevail over all the rest, the result is no longer a sum of small differences, but a single difference; in this case there is no longer a general will, and the opinion which prevails is purely particular.’ Most remarkably, Rousseau also pondered that if factions or parties were unavoidable, that is ‘if there are partial societies, it is best to have as many as possible and to prevent them from being unequal’ (1762).

Strategy and History of Electoral System Choice

35

All the authors mentioned – who were, of course, among the most influential on constitution-making and criticism in England, the emerging United States, revolutionary France and many other countries at the time – coincided, thus, on the following observations. First, they realized that the formation of electoral factions or parties produced worse results than previous practices based on individual votes among individual candidates. Second, they eventually acknowledged that in large, complex societies in which the homogeneity of interests and values which had prevailed in small, simple communities during the Middle Ages had visibly decreased, the formation of political factions was unavoidable and even necessary to make possible the political representation of a diversity of groups. Finally, they considered that, once parties had begun to intervene in the electoral process, the existence of multiple parties was better – in the sense of re-establishing to some extent the previous capacity of elections to give representation to different preference groups – than the existence of just a few able to provoke the concentration of power in the hands of a single party. In other words, if parties were unavoidable, multi-partism was better than bi-partism. Basically two different types of procedure were then devised to make organized faction- or political party-competition somehow compatible with pluralistic, inclusive representation, while majority was still the basic principle of decision. First, several ballot procedures were invented in order to produce more-than-one-party representation in multi-member districts by plurality rule. Second, the alternative of single-member districts, whether by plurality or by majority second-round rules, was diffused, also in the expectation of reducing the high concentration of power in a single party produced by bloc ballot in multi-member districts. Of course, small, disadvantaged but rising groups favoured the adoption of these relatively more inclusive electoral procedures, while large, dominant and wellsettled incumbents resisted such innovations in favour of maintaining better opportunities for concentrating all seats. Limited, cumulative ballots The two most diffused innovative ballot procedures in multi-member districts have been limited ballot and cumulative ballot. Limited ballot – which had been initially proposed by Condorcet in 1793, as mentioned above – was designed precisely to permit at least the second minority in a district to obtain a minority of seats. By this procedure, each voter is allowed to vote for a number of candidates which is less than the total number of seats to be filled in the district, while the candidates with highest numbers of votes are elected. With high proportions of limited votes (for instance, two-thirds of the seats to be filled), this procedure will give as many seats as votes per voter to candidates of the largest party while the remaining minority of seats will be allocated to candidates of the second party in votes. Only if the largest party obtains a very high proportion of votes and is able to coordinate its followers’ votes to be distributed efficiently among its different candidates can it expect to receive all the seats in the district. Different forms of limited ballot with several votes per voter were used tentatively in a minority of districts in Britain from 1867, as well as in Italy, Portugal and Spain

36

Josep M. Colomer

a few years later. More daringly, limited ballot was adopted as the general formula to fill all the seats in the Assembly in Brazil after 1882, as well as in Argentina, Bolivia, Peru, Portugal and Spain at the end of the nineteenth century. With this procedure, as illustrated in Table 1.1, while the most voted party list obtains a majority of seats, a second party can also receive some seats (in the example in the table, the corresponding allocation of seats is 2–1–0). A particular form of limited ballot is the one giving only one vote per voter – as was proposed by Saint-Just in 1793, as previously mentioned. This formula was innovatively used in Spain from 1865 to 1876. In districts with more than two seats, this form of limited single ballot can distribute the seats among more than two parties, actually producing a really pluralistic and approximately proportional representation, as illustrated in Table 1.1. In modern times, this procedure has been called single non-transferable vote and has been used, most notably, in Japan after 1900 and again after the Second World War. The procedure of cumulative ballot – which had been proposed in Britain by member of Parliament James Garth Marshall (1853) – gives each voter as many votes as there are seats to be filled in the district, but permits the accumulation of more than one vote on a single candidate. The results of this procedure depend on voters’ information regarding the expected support of the different candidates and their ability to coordinate their accumulations. In general, it can be expected that seats will be allocated to more than one party. But only if the larger party fails to coordinate its supporters’ accumulated votes while the smaller parties succeed in doing so can the allocation of seats be as pluralistic as with limited ballot. Cumulative ballot was used in three-member districts for the election of the Illinois state legislature from 1869 to 1980, as well as in Chile, Brazil and Peru during late nineteenth and early twentieth centuries. Table 1.1 also illustrates some hypothetical allocations of seats under this procedure in the simple example of a three-member district. (The incentives for strategic voting and the likely consequences of both limited and cumulative ballots were insightfully analysed by Charles Dodgson, better known as Lewis Carroll, in 1884; see also Black, 1996.) Single-member districts As an alternative to these procedures oriented to permit more-than-one-party representation in multi-member districts, the division of large multi-member districts into smaller single-member districts can also produce some pluralistic representation. In order to show how this happens, Table 1.1 illustrates the simple example of a three-member district which becomes three single-member districts. As shown in the table, a single party able to receive all seats in a multi-member district may fail to become the largest party in some of the smaller single-member districts, thus permitting some other party to obtain representation. (In the example in the table, the new allocation of seats, instead of favouring a single party, 3–0–0, as produced in a three-member district, is now 2–0–1; observe that the party with no seats has wider total support than the party with one seat, but the latter is the strongest in one district.)

Strategy and History of Electoral System Choice

37

In fact, single-member districts can be combined with different rules, mainly plurality, majority with a second round by plurality, and majority with a secondround runoff by majority between the two most voted candidates in the first round. If, according to the first of the rules just mentioned, the seats of an assembly are elected in single-member districts by plurality rule, certain global minorities with relatively strong support in small local areas can see their candidates elected, which is a more inclusive result than the one produced by multi-member districts. Single-member districts, however, can work against large minorities with widespread support across the territory if they are not the largest group in any significant area, up to not giving representation to a relatively large preference group. The adoption of single-member districts in elections by plurality rule apparently occurred, first, in the British colonies in North America and were introduced in England after 1707. They developed in the late eighteenth and early nineteenth centuries for election to House of Commons in Ireland, Scotland and Wales (in contrast to most of England, where two- and three-member districts prevailed), as well as in certain other states and the United States for election to the House of Representatives. They, however, did not become the general formula in Britain until 1885 – at the Conservatives’ demand as a protective device in exchange for a new enlargement of the franchise which they felt to be threatening. In the United States single-member districts expanded gradually, but it was not until 1970 that they finally replaced at large and other multi-member districts with bloc voting in a few states. In contrast Denmark for one had been forming its parliament using only this formula since 1848. As was noted by the English philosopher and politician John Stuart Mill on the occasion of discussing electoral reforms in Britain, a parliamentary election in singlemember districts by plurality rule – as an indirect election with factions, as previously discussed – can give victory to ‘a majority [of voters] of the majority [of districts], who may be, and often are, but a minority of the whole’ (Mill, 1861). Actually, Mill was still short in his insight since, by this procedure as implemented in Britain, a parliamentary majority can be given to a relative majority, so to a minority of voters of the majority of districts, which can indeed be a very small minority of the whole number of voters, and even smaller than some other groups. The election of a single winner by plurality rule was also implemented for the direct election of presidents in regimes with separation of powers between the Presidency and the Assembly. The earliest cases were Colombia in 1853 (replacing the US-inspired electoral college) and Brazil after 1892 (replacing the Portuguese monarchy). Later on, this rule was adopted for direct presidential elections in many countries of Latin America, as well as in the Philippines, South Korea and some African countries under former British control. As can also happen in any parliamentary single-member district, a plurality-winning president may be an extreme candidate receiving a strong rejection from a majority of voters that may have divided their votes between several relatively closer, defeated candidates. In other words, the winner by plurality can be the Condorcet-loser, that is a candidate that would be defeated by any other candidate in pair-wise contests. As has happened, a number of political crises and breakdowns of democratic regimes were produced

38

Josep M. Colomer

in Latin American, Asian and African countries after the election of minority extreme presidents by plurality rule, especially by the mid-twentieth century. Absolute majority rules Comparable but relatively less bad results can be obtained in parliamentary single-member districts, as well as in presidential elections, if majority rule with a second round is used. A first variant, requiring an absolute majority at the first round and only a plurality at the second round, was adopted for the election of the National Assembly of France in 1875 and has been used there most of the time up to the present. This system can only make a difference with simple plurality rule if the candidates coordinate their withdrawals at the second round in favour of those who are relatively closer in their preferences (or more willing to enter into favour exchanges) and have a higher likelihood of winning. In conditions of suboptimal coordination, however, the winner may be as extreme and in the minority in voters’ first preferences and provoke a similar rejection as the winner by simple plurality rule in a single round. Another variant requires an absolute majority at the first round and reduces the choice to the two most voted candidates at the second round. This system was used for parliamentary elections in Germany after 1871. For presidential elections, a precedent of second-round runoff rules can be found in the procedure establishing a first direct election round by majority and a second-round runoff in Congress between the two or three most voted candidates. This formula was adopted in France in 1851, as well as in Bolivia in 1871 where it is still used now and in at least six other Latin American countries in different periods. Direct presidential elections by majority with a second-round runoff between the two most voted candidates has been used in France since 1965, as well as in African countries under former French control, in most countries of Latin America in the most recent democratic period (starting with Ecuador in 1978), and in post-Communist regimes in Central and Eastern Europe since the 1990s. A further variant of absolute majority rule requiring a single round of voting was initially devised by Harvard college professor William R. Ware in 1871 by simplifying some of Thomas Hare’s proposals (to be mentioned below) and imported by mathematician E. J. Nanson in 1882 for elections in Australia, where for election to the House of Representatives it was adopted after 1918. Usually known as majority-preferential or alternative vote, this procedure requires voters to order their preferences among candidates. Any candidate obtaining a majority of first preferences is elected; otherwise, the candidate with the lowest number of votes is eliminated and his or her supporters’ votes are transferred to their next or further preferences for successive rounds of counting until a majority candidate results. The procedure, thus, enables voters to indicate for which candidate they would have voted at a second round if their favourite candidate were eliminated at the first round – for this reason it is usually known as instant runoff in the United States. Majority rule, whether with a second-round runoff or with alternative vote, does not guarantee that the winner is the candidate with the relative widest acceptance and least rejection among voters, since he or she can be eliminated at the first

Strategy and History of Electoral System Choice

39

round of voting or at some preliminary round of counting. Even if the candidates are perceived by most voters as ordered along a single linear dimension, such as the left–right axis, the median voter’s candidate, who is always able to form a majority around him or her against any other candidate, may obtain fewer votes than two other, more extreme candidates and be eliminated. In other words, the winner by majority with any of these procedures may not be the Condorcet-winner, that is the candidate able to win against any other candidate in pair-wise contests, but one that could be defeated by absolute majority by some of the eliminated candidates. However, at least two caveats can be introduced when comparing the expected outcomes by simple plurality rule with those by absolute majority runoff rule. Assuming a single dimension of voters’ preferences in which the median voter can be clearly identified and the same number of candidates running, if no candidate obtains an absolute majority of votes at the first round, majority runoff creates double the probability of plurality rule that the final winner will be the candidate with the median voter’s support. This is due to the fact that if the median voter’s candidate is present at the second round, he or she must be expected to win by majority against any other rival. Thus the probability for the median voter’s candidate to win is equal to the probability of being one of the two most voted candidates in the first round, which, under the above assumptions, a priori and with whatever number of candidates, is double the probability of being the single most voted candidate in the first round. The second caveat is linked to the first. While simple plurality rule can give victory to the Condorcet-loser, that is the candidate that would be defeated by every other candidate in pair-wise contests as mentioned above, the winner by absolute majority runoff rule or its variants cannot be the Condorcet-loser since he or she will have been able to win at least one of the pair-wise contests or comparisons, that is the one between the two survivors at the second round of voting or the final round of counting. (This point was made early on by E. J. Nanson in 1882.) This means that in parliamentary single-member districts, as well as in presidential or other elections producing a single absolute winner, it should be expected that the winner by majority rule with a second-round runoff or alternative vote will be, on average, relatively more accepted and less rejected by the voters than the winner by simple plurality rule. Of course, specific elections can produce winners with characteristics opposite to those just mentioned, but we are just presenting here a general statement regarding the results that can be expected a priori from every electoral rule.

The choice of majority rules This classification of majoritarian electoral rules and procedures for including a variety of political groups and producing relatively acceptable winners permits the establishment of some basic strategic criteria for institutional choice. The choice of procedures with relatively higher inclusiveness, such as limited or cumulative ballot or single-member districts, or of rules with higher requirements of votes for winning absolutely, such as the majority runoff rule, implies less risk, while the

40

Josep M. Colomer

bet on exclusionary procedures, such as multi-member districts with ‘bloc’ closed lists, or for rules with no specific vote requirements for winning absolutely, such as simple plurality rule, should be expected to correspond to more risk-prone actors. In the perspective supplied by the above analysis, it should be expected that a single dominant party with optimistic expectations to remain in such a position will tend to prefer electoral formulas like multi-member districts with ‘bloc’ closed lists and plurality rule to other formulas permitting minority representation. In contrast, in situations in which the incumbent rulers are seriously challenged by a rising opposition or by new emerging parties, as well as in situations of relative ignorance or high uncertainty regarding the likely support of different candidacies – for instance, when a new regime with broad voting rights is established – the relatively more inclusive electoral formulas – that is limited or cumulative ballot, single-member rather than multimember districts, and majority runoff rule rather than simple plurality rule – can be less risky choices for self-interested actors and are capable at the same time of producing more socially acceptable results. A further cross-country, long-term analysis will test these implications on electoral system choice. Further reading • Arrow, Kenneth ([1951] 1963) Social Choice and Individual Values. New York: Wiley. • Brams, Steven J. and Peter C. Fishburn (1983) Approval Voting. Boston: Birkhäuser. • Favre, Pierre (1976) La décision de majorité. Paris: Presses de la Fondation Nationale des Sciences Politiques. • Grofman, Bernard, Sung-Chull Lee, Edwin A. Winckler and Brian Woodall (eds) (1999) Elections in Japan, Korea, and Taiwan under the Single Non-Transferable Vote. The Comparative Study of an Embedded Institution. Ann Arbor: University of Michigan Press. • Hägele, Günther and Friedrich Pukelsheim (2001) ‘Llull’s Writings on Electoral Systems’, Studia Lulliana, 41: 3–38. • Heinberg, John G. (1926) ‘History of the Majority Principle’, American Political Science Review, 20, 1. • Heinberg, John G. (1932) ‘Theories of Majority Rule’, American Political Science Review, 26, 3. • McGann, Anthony J., William Koetzle and Bernard Grofman (2002) ‘How an Ideologically Concentrated Minority Can Trump a Dispersed Majority: Nonmedian Voter Results for Plurality, Run-off, and Sequential Elimination Elections’, American Journal of Political Science, 46, 1. • McLean, Iain and Arnold B. Urken (eds) (1995) Classics of Social Choice. Ann Arbor: University of Michigan Press. • May, Kenneth O. (1952) ‘A Set of Independent, Necessary and Sufficient Conditions for Simple Majority Decision’, Econometrica, 20. • Merrill III, Samuel (1988) Making Multicandidate Elections More Democratic. Princeton, NJ: Princeton University Press. • Miller, Nicholas R., Bernard Grofman and Scott L. Feld (1989) ‘The Geometry of Majority Rule’, Journal of Theoretical Politics, 1. • Nurmi, Hannu (1987) Comparing Voting Systems. Dordrecht: Kluwer. • Pukelsheim, Friedrich (2002) ‘Auf den Schultern von Riesen: Llull, Cusanus, Borda, Condorcet et al.’, Litterae Cusanae, 2: 3–15. • Reilly, Benjamin (2001) Democracy in Divided Societies. Electoral Engineering for Conflict Management. New York: Cambridge University Press. • Riker, William H. (1982) Liberalism Against Populism. A Confrontation Between the Theory of Democracy and the Theory of Social Choice. San Francisco: Freeman. • Saari, Donald (1994) Geometry of Voting. New York: Springer. • Saari, Donald (1995) Basic Geometry of Voting. New York: Springer.

Strategy and History of Electoral System Choice

41

• Staveley, E. S. (1972) Greek and Roman Voting and Elections. Ithaca, NY: Cornell University Press. • Taylor, Lily Ross (1966) Roman Voting Assemblies. From the Hannibalic War to the Dictatorship of Caesar. Ann Arbor: University of Michigan Press.

Proportionality The principle of proportional distribution of shares, burdens and voting rights is usually adopted in private communities and companies on the basis of the number of inhabitants, resources, property or contribution to the provision or finance of collective goods. Remote precedents of proportional representation for distributing seats in more political settings can be found in ancient leagues of cities and tribes. In early modern times, confederal unions and a number of peace plans in Europe also addressed the question of fair representation. More formally, and for domestic political elections, the principle of proportional representation was first formulated in late eighteenth-century France during the holding of the traditional, corporative elections which would trigger the Revolution. The basic mathematical formulas that would make the principle operable were also invented about the same time in the process of setting up the new representative institutions of the United States of America. However, these formulas were not used in the United States for the allocation of institutional seats to political candidacies but only for apportioning seats in the House of Representatives among the differently populated states. Proportional representation formulas were reinvented several times in Europe in the second half of the nineteenth century before they began to be used for the allocation of seats in modern assemblies and parliaments.

The invention of the proportionality principle In terms of ancient confederal leagues, the most remote precedent of proportional representation can probably be found in the Lycian League in Asia Minor from 168 BC to c.50 AD. It reunited 23 cities (including, for instance, Olympos and Xathros) that were represented in the governing Council with one, two or three votes each, according to their size. They also bore a proportional share of financial responsibility. But the most considerable of the confederacies of antiquity was the Amphictyonic League, which was organized around Delphos for religious and mutual defence affairs, with 12 Hellenic tribes, including Athens and Sparta. The representatives of all units met twice a year to vote on decisions which were executed by the Senate. Although working with equal numbers of votes per unit during most of its existence, during a short period from 31 BC to 14 AD the Amphictyonic League was organized with a more proportional distribution of votes: three cities had six votes, three cities had two votes and six cities or groups of small tribes had one vote each (which was cast by one of its components by turns). Although not precisely distributed according to a mathematical formula, the different numbers of votes roughly corresponded to each city’s or group’s population (Freeman, 1863). In their approach to similar problems, a number of confederal plans and attempts to unite different political units for common purposes of defence can be identified

42

Josep M. Colomer

during the late Middle and early Modern Ages. One of the most remarkable was developed in the late seventeenth century by English colonizer William Penn, whose peace plan for Europe included the formation of a Diet (also called Estates or Parliament) at which each sovereign state would be represented in proportion to ‘its value’ (to be calculated not only in terms of population but also of wealth and military strength). Up to 15 units were devised (from the German Empire, France, Russia, Spain and Turkey to the smallest Swiss cantons and German ducats) with a total of 90 votes distributed in a scale with eight values from 12 to 1 (Penn, 1693). Similar problems were, of course, faced by modern confederations and federal states forming representative assemblies – from the United States in the late eighteenth century to the European Union in the late twentieth century – as well as by all kinds of parliaments and assemblies whose members were to be elected in territorial units of different sizes and populations. For the purpose of organizing popular elections, an early formulation of the principle of proportional representation of the different groups of voters in political assemblies came from politician and electoral candidate Honoré Gabriel Riqueti, Count De Mirabeau, in the very first weeks of 1789. In his address to the provincial estate of Provence on the occasion to elect the delegates to the Estates-General of France, Mirabeau famously stated that ‘The Estates-General are [ought to be] to the nation what a chart is to its physical configuration; in all its parts, and as a whole, the copy should at all times have the same proportions as the original’ (Mirabeau, 1789, 1: 7). In a lesser known passage of the same speech, Mirabeau remarked that he did not aim only at obtaining a faithful or fair representation of the different groups of society, but was also concerned with inducing socially efficient outcomes from the corresponding political institutions. In this orientation, representation of all parts should prevent the two dominant estates, the aristocracy and the clergy, from prevailing over the whole nation. He noted: ‘The nation is not there if those who call themselves the representatives have not been chosen in free and individual elections, if the representatives of groups of equal importance are not equal numerically and in voting power . . . In order to know the will of a nation, the votes must be collected in such a way so as to prevent the mistake of taking the will of an estate for one other, or the particular will of certain individuals for the general will’ (Mirabeau, 1789, 1: 7–8). Proportional quotas About the same time as Mirabeau formulated his principles, some of the leading politicians of the newly created United States of America were inventing for the first time the basic mathematical formulas that would make proportional representation feasible. As suggested, they did not invent them with the intention of promoting fair, pluralistic and socially efficient representation of different groups of voters in the corresponding institutions, but only for the sake of fair territorial representation in the new federation, as mandated by the US Constitution (Article 1, section 2). For the first US rulers and electoral law-makers, the basic problem was to define a quota of state inhabitants worth a seat in the House of Representatives – which is, formally, the same problem as to define a ‘quota’ of party votes worth a seat in

Strategy and History of Electoral System Choice

43

any council, assembly or parliament, as would be addressed later on. Three types of quota were then and in further periods defined, which can be called, respectively: exact, sufficient and fixed quotas. Table 1.2 shows an example of how different quota formulas can be applied. The exact or simple quota was first defined by Alexander Hamilton, one of the main authors and propagandists of the new United States federal constitution, for the first House elections in 1791. The exact proportional quota is the divisor between the total number of inhabitants (or votes) and the total number of seats, that is 1/M (where M is the magnitude or number of seats in the district). This formula is quite intuitive from the notion of proportionality, but has the big drawback of not being decisive or complete since the distribution of inhabitants among states, or that of votes among parties, hardly comes in exact multiples of such a quota. Therefore, the remaining seats not able to be allocated on the basis of the exact quota have to be allocated by using some other supplementary formula which may introduce significant distortion. As this formula was initially proposed both by Hamilton and, in the context of parliamentary elections, by English lawyer and politician Thomas Hare about seventy years later, the remaining seats were to be allocated to those states or parties having the ‘largest remainders’ of inhabitants or votes, respectively (Hare, 1859). In actual electoral systems, however, the ‘exact’ or Hamilton-Hare quota has been used together with disparate supplementary formulas for allocating the remaining seats, such as giving preference to the larger parties, or to the smaller parties, or by random, or by applying some other formula of proportional representation to the remaining votes, or – as one of the most distinguished critics of the Hamilton proposal voiced – ‘according to any other crotchet which ingenuity may invent,

Table 1.2

Proportional representation formulas Parties: Votes:

Rules Hamilton-Hare Exact quota: 100/6 = 16.6

Jefferson-d’Hondt Sufficient quota: [40/3 = ] 13 Webster-Sainte-Laguë Modified quota: 100/6 + 1 = 14

Gergonne-Gilpin Fixed quota, e.g. 15 votes

W 40

X 30

Y 20

Z 10

Total/100

2 6.6

1 3.3

4

1

0 10] 1 1

2 6

6 seat allocations

By quotas [Vote remainders By largest remainders Total seats

2

1 13.3 1 2

By quotas

3

2

1

0

6

By quotas By half-quota (7) Total seats

2

2

1

2

2

1

0 1 1

5 1 6

By quotas

2

2

1

0

5

44

Josep M. Colomer

and the combinations of the day give strength to carry’ ( Jefferson, 1792). In order to reduce the number of seats to be allocated by a supplementary criterion, London barrister Henry R. Droop proposed a somewhat lower quota, 1/M + 1 (rounded up); with a lower quota value, a higher number of seats can be allocated by quotas, but this formula would still need a supplementary one to allocate at least a few remaining seats (Droop, 1868, 1869). Logically, the combination of two different formulas for the allocation of the same slot of seats may produce some unexpected and undesirable consequences. Even the standard Hamilton-Hare formula allocating some seats by the exact quota and the rest to the largest vote remainders has been found capable of producing some paradoxical allocations, such as giving fewer seats to a party (or state) which has increased its proportion of votes (or inhabitants) – usually known as the ‘Alabama paradox’. The second formula, which can be called sufficient quota, was first conceived by Virginian politician and later US president Thomas Jefferson in 1792 as a response to Hamilton’s quota – the latter being finally vetoed by president George Washington at Jefferson’s persuasion. Seeking support in the instructions of the US Constitution and in the appeal of simplicity, Jefferson proposed to define a quota able to allocate all the seats by a single criterion. Literally, he stated: ‘We must take, as the ratio of distribution, the common divisor which, applied to every state, gives to them such numbers as, added together, come nearest to 120 [the total number of seats in the House of Representatives at the time]’. The idea of a common divisor for all allocations of seats also motivated Belgian civil law professor Victor d’Hondt to reinvent this quota about a hundred years later (d’Hondt, 1878, 1882). Of course, to fulfil its aim of allocating all the seats, the Jefferson-d’Hondt quota must be smaller than the Hamilton-Hare quota. Although it has been common to confuse the Jefferson-d’Hondt ‘sufficient’ quota formula with some of the calculation methods to establish its value, the basic meaning of the formula should not be forgotten. In fact, there are several methods to find such a quota and produce the corresponding allocation of seats, including trial and error, as proposed initially by Jefferson; using a series of divisors until obtaining the M-th highest quotient or ‘highest average’ (M being the number of seats to be allocated), as promulgated by d’Hondt and known as the method of ‘divisors series’; or starting tentatively with the quota 1/M + 1 (rounded down) and decreasing it down until fitting, as suggested by Swiss physics professor and local politician Eduard Hagenbach-Bischoff (1888). Several variants of this type of quota have also been invented. For instance, US Senator Daniel Webster began, in 1832, to modify the quota by increasing its value to the nearest whole number and then, if necessary, giving extra seats to those states or parties having more than half the quota. This formula, which is known as ‘major fractions’, would be used for seat apportionment in the US in the early twentieth century. It was later reinvented, together with another calculation method based on the divisors series of odd numbers, by French mathematician André Sainte-Laguë (1910) and has been used for parliamentary elections, in modified versions, especially in Scandinavian countries.

Strategy and History of Electoral System Choice

45

Finally, what can be called the fixed quota is just an absolute number of votes established a priori as worthy of a seat. This was actually the basic criterion that had been used in a few of the ancient and early modern leagues and confederations previously mentioned, in which the number of each unit’s representatives was established on the basis of some fixed amount of inhabitants. For political elections, this formula was first proposed by French mathematician Joseph-Diaz Gergonne in 1820. In his own words: ‘At the elections the voters should group themselves freely according to their opinions, their interests, or their desires, and any citizen would become a Deputy from the department in the elective chamber who bore a mandate from two hundred voters [his arbitrarily proposed quota]’ (Gergonne, 1820). Similarly, Philadelphia manufacturer and activist Thomas Gilpin proposed a fixed quota in a pamphlet published in 1844. The context of Gilpin’s proposal, against which he tried to react, was the system previously discussed of multi-member districts by plurality rule by which each voter had to vote in bloc for a closed list, thus producing a single dominant party. His intention was to prevent further successive self-reproductions of rulers in power and make ‘the elected body to be an exact representation of the public interests’. ‘Otherwise’, he warned, ‘is the virtual source of oppression and injury the cause of great discontent, and when not subversive of the administration, it is subversive of the harmony and confidence which ought to be afforded to it’. Gilpin’s proposal was a fixed quota and open lists in which voters could express their ordinal preferences regarding individual candidates (Gilpin, 1844). A similar, but more practically usable formula, actually inspired by Gergonne, was further proposed in the United States by J. Francis Fisher (1863). However innovative and well-intentioned these proposals might be, it is remarkable that they were apparently formulated in complete ignorance of the proportional quota formulas that had been and were at the time being used in the same political system for the apportionment of seats. ‘Fixed’ or Gergonne-Gilpin quotas may have some advantages, such as encouraging turnout and producing uniform distributions of seats among parties independently of district numbers, magnitudes and apportionment. However, they are not able to establish a previously known number of seats in any district or in the total of the assembly. With varying assembly sizes, this type of quota – which was being used in the German Land of Baden-Baden – was adopted by the so-called ‘Weimar Republic’ of Germany in 1919 with a rather demanding quota of 60,000 votes per seat. Of course, different quotas may produce different allocations of seats in favour of either larger or smaller parties and can be chosen by electoral law-makers according to their expectations. But, as Webster established: ‘The divisor may be anything which produces accurate and uniform division’. 3 There are thus two crucial elements in proportional representation electoral systems which may produce different results to the advantage of different parties: the number of seats and the quota formula. With ‘number of seats’ we refer to the total number of seats or ‘assembly size’, as well as to the ‘magnitude’ or number of seats per district. The number of seats can be small if the assembly size is small or, even for a large assembly, if a high number of districts makes their magnitudes small.

46

Josep M. Colomer

The ‘quota’ as the basic ground for distributing seats among parties can be relatively large, like the Hamilton-Hare ‘exact’ quota, relatively small, like the Jeffersond’Hondt ‘sufficient’ quota, intermediate, like the Webster-Sainte-Laguë ‘major fraction’ quota, or have any value, like the Gergonne-Gilpin ‘fixed’ quota. Further modifications can be introduced by the requirement of legal thresholds to enter the process of allocation, whether in the form of a given percentage of votes or as the requirement to obtain one ‘exact’ quota or a portion of it, actually decreasing the effective quota for the parties above the threshold. Mal-apportionment of seats with regard to the population and, in particular, the allocation of a minimum number of seats to every district, even the smallest ones, can also distort representation based on quotas of votes. Note, however, that the smallest quota is the one which is applied with plurality rule: literally, the quota can be any value higher than zero, only depending on the number of parties or candidates. Second-round formulas also imply very small quotas since passing into the second round is a contest by plurality rule, at which no specific threshold is required, and, in some of its variants, the final winner may also be just the most voted candidate at the second round. The consequences of the different electoral formulas just revised can be summarized as follows. First, the smaller the number of seats (both in the total assembly and in every district), the higher the number of votes which can be sufficient for a party to receive a seat – this is the value usually called the threshold of exclusion. The smaller the number of seats and the higher the corresponding exclusion threshold, the more disadvantaged are, of course, the small parties and the more advantaged are the larger ones. Second, the smaller the quota, the more advantaged are the larger parties which can obtain higher numbers of quotas rewarded with seats. In particular, plurality rule, whose quota is, as mentioned, whatever value higher than zero, however small, maximizes the representation of the largest party. Among the basic formulas of proportional representation previously discussed, the Jefferson-d’Hondt ‘sufficient’ quota, which is relatively small, maximizes the representation of those parties larger than the average. The choice of proportional representation This analysis may allow us to identify some basic strategic criteria of electoral system choice. It can be expected, in short, that the large will prefer the small and the small will prefer the large, specifically that dominant, optimistic or risk-prone and would-be large parties will prefer: (1) small assemblies; (2) small district magnitudes; and (3) small quotas. In other words, large parties should be expected to prefer assemblies with a relatively low number of seats, apportionments creating small numbers of seats per district, as well as plurality rule rather than proportional representation formulas and, among the latter, the Jefferson-d’Hondt formula which implies relatively small quotas rather than the Hamilton-Hare formula or other variants. Of course, high numbers of districts and high legal thresholds belong to the same set of preferences. Conversely, seriously challenged, pessimistic or risk-averse and would-be small parties should be expected to prefer: (1)

Strategy and History of Electoral System Choice

47

large assemblies; (2) large districts; and (3) electoral formulas implying large quotas, in particular those of proportional representation rather than majoritarian ones. These electoral system preferences and choices have been justified using different arguments. First, fairness and social efficiency of the subsequent political decisions can be called in support of proportional representation. The Conservative Thomas Hare, for instance, was said to have devised a system which, in contrast to the concentration of power created by plurality rule, would secure representation of all classes, including minorities. However, he remarked that his project did not aim at simply securing the ‘representation of minorities’ but to ‘end the evils of corruption, violent discontent and restricted power of selection or voter choice’ (Hare, 1859). Hare’s main propagandist, the liberal John Stuart Mill, also remarked that proportional representation would give all interests or classes ‘protection against the class legislation of others without claiming the power to exercise it in their own’ (Mill, 1861). French socialist leader Jean Jaurès, on his own, made this didactic presentation of the choice of electoral systems: ‘This will kill that. So is the formula of voting in [a single-member] district. These will kill those. So is the formula of voting by list without proportional representation. These and those will have their fair share. So is the formula of voting by list with proportional representation’. However, in addition to fairness arguments, group self-interest has also sometimes been explicit. For instance, the leader of the German socialist party, Wilhelm Liebknecht, addressed the annual conference of his party in 1890 with these words: ‘Our party would derive great advantages from the introduction of this electoral system [proportional representation] . . . Under the present electoral system the greater part of our votes is lost – whereas under proportional representation our strength in parliament would be doubled or tripled’ (in Rustow, 1950: 121). In fact, majoritarian systems have been supported by successful actors in single dominant or frequently alternating two-party systems – typically formed by conservatives and liberals or, in more recent times, by conservatives and socialists. In contrast, proportional representation is usually favoured in emerging multiparty systems by different types of actors: on the one side by incumbent conservatives or liberals under threat of losing their privileged position, on the other side by minority but growing opposition parties, traditionally including socialists, christians and ethnic parties. In a summary survey it can be observed that, among the earliest countries to adopt proportional representation in the late nineteenth and early twentieth centuries, ‘Belgium had essentially a three-party system before it adopted proportional representation in 1900, Denmark had a four-party system during most of the period in which it employed a British-type of electoral system, Norway had at least three major parties on the eve of the adoption of proportional representation in 1921, Switzerland had four parties in the last elections under the plurality system, and Germany had anywhere between a six- to twelve-party system depending on the criterion used to define it’ (Grumm, 1958: 374). A more systematic test is presented in the following section of this chapter, showing that, indeed, multi-party systems already existed in most countries when electoral systems of proportional representation were chosen in modern times.

48

Josep M. Colomer

Mixed systems Less straightforward calculations than here generically assumed, as well as successive refinements in the invention and design of electoral formulas and procedures, can account not only for specific combinations of proportional representation formulas and district magnitudes, but also for the choice of intermediate or mixed electoral systems. Three basic categories of so-called ‘mixed’ systems can be distinguished, which have been called coexistence (involving different rules for different sets of voters and seats), multiple tier allocation (involving one vote per voter and two or more rules) and parallel systems (involving two votes and two rules) – although loose vocabulary is widespread. Coexistence means that different rules and procedures are used in different districts to fill different portions of the assembly seats. This has been the case, for instance, in Costa Rica since 1913, Greece in 1956 or Panama since 1983, where some districts are decided by plurality rule and others by proportional representation. Multiple tier allocation involves a single ballot but two or more rounds of counting the same votes by different rules to allocate different portions of seats in the assembly. A typical combination includes one round of counting by plurality rule and another round by some formula of proportional representation – as in South Korea in 1960 and again since 1987, Mexico since 1986 or Taiwan since 1991. This type of system allows electoral law-makers to make well-tuned calculations and adjustments, since while it establishes complex and sophisticated voting counts, it reduces voters’ choice to a single decision not permitting much strategic behaviour. Finally, parallel systems give each voter two votes to select candidates for two different portions of seats which are allocated by two different rules. This type of system may result from intermediate compromises, at either explicit negotiations or underlying calculations, between less-than-secure incumbent rulers which would have preferred to be able to maintain or establish plurality or majority rules, and increasingly emerging but not overwhelming opposition groups preferring proportional representation systems. In contrast to the previous type of multiple-tier allocation, parallel mixed systems allow voters to split their ballot between candidates or lists belonging to different parties, thus permitting a rather subtle revelation of preferences. However, the resulting composition of the assembly can be distorted by the use of different allocation principles and rules. Of course, the advantage of the larger parties and the degree of disproportionality in representation will be higher the higher the portion of seats allocated by plurality or majority rule. An early case combining plurality rule and proportional representation in ‘parallel’ mixed systems was Brazil in 1933. A number of ‘parallel’ systems have also been established in recent processes of democratization in which institutional formulas have been chosen under rather balanced relationships of forces between the incumbent rulers and the opposition, as in Russia and several countries in Central and Eastern Europe in the 1990s. Likewise, parallel mixed systems have been compromised in better-established democratic systems in which a long-term dominant party has entered into clear decline but has not been completely

Strategy and History of Electoral System Choice

49

replaced in voters’ support by the opposition, as in Italy under the tension between the governmental christian-democrats whose popularity had worn away, and the communists who were always in contention but never won, and in Japan after some wearing away of the long-term governing liberal-democrats under pressure from the socialists and other minority and divided parties in opposition, in both cases since the mid-1990s. In contrast to these types of mixed systems, other combinations of rules and procedures are much more easily to be classified as either majoritarian or proportional. The first category may include electoral systems in which an absolute majority rule is used to allocate most seats, and only in the case where no party obtains such a majority is a formula of proportional representation applied. This type of system involves one vote per voter and one rule, with another rule just in reserve (or, as has been suggested, as ‘supplementary’ or ‘conditional’). A few would-be dominant parties betted on this type of system in France in 1919 and 1951, and in Italy in 1923 and 1953 – in most cases without obtaining the majority representation they were expecting. On the other side, proportional representation systems can also be organized by using different formulas at different tiers, that is at districts including different distributions of voters such as national, regional and local. Cases include Belgium since 1919, Denmark and Germany since 1920, Greece in a number of experiments with three tiers since 1926, the Netherlands since 1933, Italy since 1946, or Ecuador since 1978. Finally, personalized proportional representation (also called mixedmember proportional systems or additional member systems, among other names) distinguishes each vote as valid both for an individual candidate competing in single-member districts by plurality rule and for a party list in a large district by proportional representation. In contrast to mixed systems, however, the allocation of seats is ‘compensatory’, which means that those seats obtained by a party at single-member district level are discounted from those obtained from the total pool of the assembly seats by proportional representation. This type of system was pioneered in Denmark in 1915. As the system was established in the Federal Republic of Germany in 1953, each voter has two votes. In fact each party receives a number of seats based on its proportion of party list votes; then about half of the total seats in parliament are occupied by those individual candidates with higher numbers of individual votes in single-member districts, while the rest are selected from the party’s closed lists. With this combination, a high proportionality between a party’s votes and seats and closeness and accountability of some individual candidates to voters can be achieved – a combination that most properly deserves the epithet of ‘the best of both worlds’. This type of system was also adopted in Bolivia, New Zealand and, temporarily, in Venezuela in the 1990s, during processes to open or broaden the political system to more pluralistic party configurations. Open and closed ballots The complexity of formulas such as those just reviewed suggests that electoral systems can be chosen not only for the inter-party allocations of seats which are

50

Josep M. Colomer

the result, but also for their consequences on intra-party competition between individual candidates. Roughly speaking, while strongly independent individual candidates should be expected to prefer systems permitting voters the opportunity to chose among individuals, compact, disciplined parties will prefer to establish nomination systems of candidates by which the choice of voters will be determined by party labels rather than by their individual characters. Four types of formula are revised in the following paragraphs along these lines: (1) primary elections; (2) open ballot or panachage, as well as ordinal single transferable vote; (3) preferential vote for individual candidates, including limited ballot in majority systems, and double vote and open lists in proportional representation; and (4) categorical vote, including single ballot in majority systems and closed lists in proportional representation. The order in which these formulas are discussed presumably corresponds to a gradation from higher to lower opportunities for individual candidates to campaign and attract votes. Regarding the choice of electoral rules and formulas, it can be expected that the more open the selection of individual candidates, the lower the pressure that can be expected for ‘opening’ the electoral system, as will be discussed below. First, primary elections, which began to be used to nominate presidential candidates in the United States in 1912, are presently held in the US, as well as in a number of other countries, to select single-winner individual candidates for single-member offices such as president, governor and congressmen in singlemember districts, or top list candidates. Primary elections can include a relatively large number of candidates located at various positions along the policy–ideology spectrum. This is in contrast to the usual restrictions in competitions for singlemember offices, which tend to focus on a couple of major candidates. In this sense, a primary election can be compared to the first round in a majority electoral system with two rounds for their role in selecting the very few candidates that will actually compete for office. A high number of candidates can enter the race and find the subsequent results acceptable for the opportunities that have been given to compete. However, since, in contrast to the first round in an open election, each primary election is reserved to select only one political party’s or coalition’s candidate, voters’ participation tends to be much lower in the latter than in the former. In the United States, for example, participation in presidential primary elections for the two major parties since they were generalized in 1972 has been, on average, about one-third of the participation in the corresponding real presidential election. All students of these experiences have noted that primary voters are non-representative of their parties’ voters, and even less of the whole electorate. In general, primary voters tend to favour more extreme or outsider candidates than the electors, so distorting electoral competition in a remarkable way. The probability that none of the candidates in the election corresponds to the median voter’s preference, as revealed regarding the candidates running in the primary election, is, thus, much higher than the probability that no median voter’s candidate survives at the second round when the first round is open to competition among all parties.

Strategy and History of Electoral System Choice

51

Second, the single transferable ballot can also promote high intra-party individual competition among candidates. Actually, this was the first type of ballot invented for proportional representation electoral systems. Early proposals oriented to making voters choose among individual candidates were gradually refined: by English schoolteacher Thomas W. Hill in 1821 calling voters to gather in groups with the same preferences, as was the subject of an experiment in the Southern Australian colony of Adelaide in 1839; by Danish conservative prime minister Carl-Christopher-Georg Andrae for the election of a minority portion of seats for the joint assembly of Denmark and Schleswig in 1855 (not the Danish parliament); and by English lawyer Thomas Hare, first in two books published in 1857 and 1859 with the impractical proposal of establishing a single national district which would require enormous computing effort, and only precisely establishing the procedures to transfer ballots from voters’ first to second and further preferences in 1865. In its more elaborated form, proportional representation with single-transferable ballot requires each voter to rank individual candidates. Seats are allocated to candidates who have obtained a quota of voters’ first preferences, while in the remaining ballots votes are transferred to the following candidates in voters’ ordinal preferences. Thus for some voters only their first preference is accounted for, while for others favouring less popular candidates and selected at random several preferences are taken into account. Intra-party competition among individual candidates is, with this ballot, at a maximum, and should correspond to weakly organized parties. This is the form of proportional representation that has prevailed in former British colonies, including Ireland since 1922, the Australian Senate since 1949, and a number of cities in the United States, especially from the 1920s to the 1940s. Third, ballots based on party lists of candidates were proposed by socialist Víctor Considerant and began to be used in some Swiss cantons after 1861, as well as for nationwide parliamentary elections in Belgium and Serbia after 1899 and in most of the other countries adopting proportional representation in the years following. Open lists, preferential ballot as well as personalized proportional representation and voting for ‘lemas’ (brands) and ‘sublemas’ allow the voter to select a party candidacy and one or more individual candidates. Variants include compulsory or optional preferences, the latter permitting a vote for a party list as given. The basic procedure requires all preferential votes for either individual candidates or party lists to be counted as votes for the corresponding party; after each party has been allocated the number of seats corresponding to those votes, they are filled partly with those individual candidates who have received higher numbers of preferences and partly by other candidates provided in the list. Of course, the higher the number of individual preferences that each voter can express, the higher the individual candidates’ competition to fill the corresponding seats can be expected to be. Open lists permitting the voter to select a few candidates, usually from one to three, have been used in most countries of Western Europe, presently including Austria, Belgium, Denmark, Finland, Greece and Sweden, as

52

Josep M. Colomer

well as in Latin American countries like Brazil and Peru, while personalized proportional representation is used, with double vote, in Germany, Bolivia and New Zealand, as mentioned, and ‘lemas’ in Uruguay. In this perspective, the single-transferable ballot previously discussed, as well as the open ballot or ‘panachage’ used in Switzerland, could be considered extreme cases, beyond the opportunities supplied by open lists or double vote, since they permit candidates to be selected from different parties and thus promote very high intra-party individual competition. At the other extreme, the system of closed lists actually restricts the voters’ choice to a party’s label, since it does not permit any modification in the order of candidates as given by the party. From the point of view of individual candidates’ selection, closed lists, which are presently used, for instance, in Argentina, Portugal, Spain and Venezuela, can be compared to the single ballot in single-member districts, as used in Britain and the United States. None of these categorical formulas permits anything other than total acceptance or total rejection of the candidacy as presented by one of the parties in contest. Regarding the choice of electoral rules, candidate selection procedures and ballot formulas with relative openness, such as primary elections, single-transferable ballot, open lists and double vote, can have a counterweighting effect on demands for replacing the existing electoral system. The more open the selection of candidates, the lower the pressure that can be expected for ‘opening’ the electoral system, since the latter’s restrictions may be somehow reduced by the former feature. This relation may help to explain, for instance, the stability of such restrictive electoral rules as single-member congressional districts, single-winner presidential elections, plurality rule and bipartism in the United States, due, at least in part, to the openness of primary elections. It may also be significant that primary elections were generalized to almost all states in the US in response to increasing pressures from new groups and political proposals since the 1960s. The experience of primary elections in Latin America, especially in Venezuela since the late 1960s and in Argentina, Chile and Uruguay since the late 1990s also suggests that this device can be adopted in well-established presidential regimes when they receive new demands for greater inclusion from various political groups against the restrictions inherent in any presidential electoral system which tends to centre competition on a very few potentially winning candidates. Conversely, and by the same kind of argument, we can expect that the more open the electoral system regarding the selection of individual candidates, the lower can be expected to be the pressure for primary elections. Further reading • Blais, André (1988) ‘The Classification of Electoral Systems’, European Journal of Political Research, 16: 99–110. • Bogdanor, Vernon and David E. Butler (1983) Democracy and Elections. Electoral Systems and Their Political Consequences. Cambridge and New York: Cambridge University Press. • Bowler, Shaun and Bernard Grofman (eds) (2000) Elections in Australia, Ireland, and Malta under the Single Transferable Vote. Reflections on an Embedded Institution. Ann Arbor: University of Michigan Press.

Strategy and History of Electoral System Choice

53

• Carey, John and Matthew S. Shugart (1995) ‘Incentives to Cultivate a Personal Vote: A Rank Ordering of Electoral Formulas’, Electoral Studies, 14, 4: 417–39. • Colomer, Josep M. (2001) Political Institutions. Democracy and Social Choice. New York and Oxford: Oxford University Press. • Cotteret, Jean-Marie and Claude Emeri (1973) Les systèmes électoraux. Paris: Presses Universitaires de France. • Cox, Gary W. (1997) Making Votes Count. Strategic Coordination in the World’s Electoral Systems. New York: Cambridge University Press. • Duverger, Maurice (1950) L’influence des systèmes électoraux sur la vie politique, Cahiers de la Fondation Nationale des Sciences Politiques, Vol. 16. Paris: Armand Colin. • Duverger, Maurice (1951) Les parties politiques. Paris: Seuil (English trans. Political Parties: Their Organization and Activity in the Modern State. New York: Wiley, 1954). • Farrell, David M. (2001) Electoral Systems. A Comparative Introduction. London and New York: Palgrave – now Palgrave Macmillan. • Grofman, Bernard and Arend Lijphart (eds) (1986) Electoral Laws and Their Political Consequences. New York: Agathon. • Hoag, Clarence and George Hallett (1926) Proportional Representation. New York: Macmillan. • Katz, Richard S. (1980) A Theory of Parties and Electoral Systems. Baltimore, MD: Johns Hopkins University Press. • Katz, Richard S. (1997) Democracy and Elections. New York and Oxford: Oxford University Press. • Lakeman, Enid (1974) How Democracies Vote: A Study of Electoral Systems. London: Faber & Faber. • LeDuc, L., Richard Niemi and Pippa Norris (eds) (1996) Comparing Democracies: Elections and Voting in Global Perspective. London: Sage. • Lijphart, Arend (1994) Electoral Systems and Party Systems. A Study of Twenty-Seven Democracies, 1945–1990. New York and Oxford: Oxford University Press. • Martin, Pierre (1994) Les systèmes électoraux et les modes de scrutin. Paris: Montchrestien. • Massicotte, Louis and André Blais (1999) ‘Mixed Electoral Systems: A Conceptual and Empirical Survey’, Electoral Studies, 18: 341–66. • Penadés de la Cruz, Alberto (2000) Los sistemas elementales de representación. Madrid: Instituto Juan March. • Perrineau, Pascal and Dominique Reynie (2001) Dictionnaire du vote. Paris: Presses Universitaires de France. • Rae, Douglas W. (1967) The Political Consequences of Electoral Laws. New Haven, CT: Yale University Press. • Reynolds, Andrew, and Ben Reilly (eds) (1997) The International IDEA Handbook of Electoral System Design. Stockholm: International Institute for Democracy and Electoral Assistance. • Shugart, Matthew S. and Martin P. Wattenberg (eds) (2001) Mixed-Member Electoral Systems. The Best of Both Worlds? New York and Oxford: Oxford University Press. • Taagepera, Rein (2001) ‘Party Size Baselines Imposed by Institutional Constraints: Theory for Simple Electoral Systems’, Journal of Theoretical Politics, 13, 4: 331–54. • Taagepera, Rein and Mathhew S. Shugart (1989) Seats and Votes. The Effects and Determinants of Electoral Systems. New Haven, CT: Yale University Press.

The electoral system evolution According to the discussion presented in the previous sections, several general tests of the basic propositions regarding actors’ motives and electoral system changes and their observable implications will be here developed. Data are collected in five pairs of summary tables at the end of each of the introductory chapters to the five world regions covered in this book – the Americas, Western Europe, Eastern Europe, Africa and Asia and the Pacific – in each case for the

54

Josep M. Colomer

Assembly (lower or single chamber) and for the Presidency. The relevant information is summarized in Summary Tables 1A and 1B (at the end of this introductory part). The tables include 289 electoral systems, all those having been used in 2,145 elections (1,601 for assemblies an 544 for presidencies) with minimally acceptable levels of political competition and available data in 94 countries since the early nineteenth century. Countries are selected for having more than one million inhabitants and some democratic experience. The latter criterion is operationalized as having appeared as ‘free’, or with scores of 3 or lower out of 7, in 30 Freedom House annual reports (1972–2002) for a period encompassing at least two successive elections. For the countries selected, all electoral systems used in minimally competitive elections in any period are included (even in not fully democratic elections with restrictions on voting rights or candidacies). We exclude only dictatorial periods and fake elections with a single candidacy or won by the incumbent with open fraud (which are alternative ways of choosing an ‘electoral system’). Roughly speaking, available data are from the early nineteenth century onwards for most countries in Western Europe and the Americas, a few in Central Europe in the 1920s, since independence after the Second World War in a number of countries in Africa, Asia and the Pacific area, and from more recent democratization periods in other countries in these latter regions and in Eastern Europe. Data allow us to count 154 electoral system changes in 56 countries. In order to proceed to checks with complete and more homogeneous data, tests are also conducted for changes only produced during democratic periods and having lasted to the present (end of 2002), including 36 electoral system changes in 29 countries. Electoral systems are classified separately for assembly and for presidential elections. Consistent with the previous discussion, four basic categories of electoral systems for the assembly are distinguished, with a few subcategories to distinguish between several formulas and procedures. From less to more inclusive, they are: • indirect elections; • majority rule, including three subcategories: – majority 1 – multi-member districts with bloc ballot; – majority 2 – multi-member districts with limited (including single nontransferable) or cumulative ballots; – majority 3 – single-member districts; • mixed systems, including ‘coexistence’, ‘parallel’ systems and ‘multiple tier’ allocations combining plurality or majority rules with proportional representation; • proportional representation, including three subcategories: – proportional1 – average district magnitude lower than nine, any formula and procedure; – proportional2 – average district magnitude higher than eight, with closed lists; – proportional3 – average district magnitude higher than eight, with open lists, open ballot (‘panachage’) or personalized PR.

Strategy and History of Electoral System Choice

55

Likewise, presidential electoral systems are classified as follows: • college or indirect election; • simple plurality in direct elections (as in the following categories); • qualified plurality (33 per cent, or 40 per cent, or 45 per cent, etc.) with a second round (either by plurality, by majority runoff, or by the Assembly); • majority with a second round (either by plurality, by majority runoff, or by the Assembly); • majority with single transferable or supplementary vote. The most relevant propositions and observable implications previously presented are submitted to empirical scrutiny in the following paragraphs. Changes to electoral systems should be expected to be produced in favour of increasingly inclusive formulas rather than against. So we should find a significantly higher number of changes from indirect to direct elections, from unanimity to majority rules, from majority rules to mixed systems, and from mixed systems to proportional representation, or other changes between categories in this direction than the other way around. Summary Tables 1A and 1B summarize data of 154 electoral system changes. We proceed to separate analysis of the 126 assembly and 28 presidential electoral system changes. For assembly elections, we analyse, first, all cases collected since the nineteenth century and, second, only changes during the present democratic periods. For total data available since the nineteenth century, we have counted 82 major electoral system changes for assemblies between the four basic categories indirect, majority, mixed and proportional in 41 countries, as shown in Table 1.3. The countries with the highest numbers of changes are, of course, among those with the longest periods registered, but, within this set, they are, interestingly, concentrated in Mediterranean Europe: Greece with 9 major changes, France with 7, Italy and Portugal with 6 changes each, and Spain with 4 (in the rest of the world, only Brazil with 6 major changes and Denmark with 4 are above three changes). The countries mentioned have experienced high political regime instability, of which electoral system changes

Table 1.3

Major changes of assembly electoral system New system

Previous system Indirect Majority Mixed Proportional

Indirect

Majority

Mixed

Proportional

* 2 0 0

12 * 1 7

1 13 * 6

5 27 8 *

Note: Changes in the upper-right part are toward more inclusive systems (66 changes), while those in the lower-left part are toward more exclusive ones (16 changes).

56

Josep M. Colomer

are a component, but also some specific inclination to electoral system manipulation since in other countries with comparable degrees of political regime instability, particularly in Latin America, there have been fewer major changes of electoral system and those usually limited to fewer formulas. In total, 66 changes have been in the expected direction toward more inclusive formulas, while only 16 have implied reverse moves – a proportion of more than 4 to 1. More specifically, 18 electoral systems based on indirect elections have been replaced with direct elections, while on only 2 occasions have indirect elections been re-established (in both cases also to be replaced with direct elections later on); 40 majority rule systems have been replaced with mixed or proportional rule systems, while on only 2 occasions with indirect elections (as just mentioned); 8 mixed systems have been replaced with proportional representation systems, while only 6 reverse moves can be identified. In total there have been 40 moves to proportional representation (5 from indirect elections, 27 from majority systems and 8 from mixed systems), while reverse moves from proportional systems number only 13 (7 to majority and 6 to mixed systems). Actually the trend may be under-evaluated for lack of information regarding some countries in which indirect elections or plurality rule systems were replaced in remote periods. It is interesting to note that most of the new mixed systems have been the result of changes from indirect or majority rule systems, while only a few have been established from previous proportional representation systems. Specifically, while 1 mixed system was established from indirect elections and 13 mixed systems have been the result of reforms from majority rule systems, only 6 have resulted from previous proportional systems. These findings are confirmed for all world regions. The balance between ‘inclusive’ and ‘exclusive’ moves is positive in the Americas: 19 to 2; Western Europe: 31 to 12; Eastern Europe: 8 to 1; and Asia-Pacific: 8 to 0. In Africa there have been a significant number of changes in the expected direction since the independence movements in the 1960s, but we have only registered 1 reverse change within the present democratic periods. Taking also into account the eight subcategories of electoral system above distinguished, we have counted 126 changes in 49 countries, as shown in Table 1.4. Again, the countries with higher numbers of changes are France with 12 changes, Spain with 10, Greece and Portugal with 9 each, and Italy with 8. In total, 96 changes have been in the expected direction toward more inclusive formulas, while only 30 have been in the opposite direction – a proportion of 3.2 to 1. The general tendency is confirmed for changes between all pairs of systems. Within majority rule systems, multi-member districts with bloc ballot have been replaced with forms of limited or cumulative ballot on 2 cases and with single-member districts on 12, a total of 14, while opposite direction moves have numbered only 7. Regarding proportional representation systems, in 12 cases systems based on small districts (magnitudes lower than nine) producing low proportionality have been replaced with larger districts, while the opposite move can only be observed on 3 occasions; finally, within proportional systems with large district magnitudes, we have registered 2 changes from closed lists to open lists and 2 in the opposite

Strategy and History of Electoral System Choice

Table 1.4

57

Changes of assembly electoral system New system

Previous system Indirect Majority1 Majority2 Majority3 Mixed Proportional1 Proportional2 Proportional3

Indirect

Maj1

Maj2

Maj3

Mixed

Prop1

Prop2

Prop3

* 2 0 0 0 0 0 0

7 * 0 7 0 2 0 0

1 2 * 2 1 0 0 0

4 12 2 * 0 4 1 0

1 3 2 8 * 3 0 3

3 6 5 5 3 * 0 3

1 1 1 4 2 1 * 2

1 0 0 5 3 11 2 *

Note: Indirect: indirect elections; Maj1: majority rules in multi-member districts; Maj 2: majority rules with limited or cumulative ballot; Maj3: single-member districts; Mixed: mixed systems; Prop1: proportional representation, average district magnitude < 9; Prop2: proportional representation, average district magnitude > 8, closed lists; Prop3: proportional representation, average district magnitude > 8, open list or double vote. Changes in the upper-right part are toward more inclusive systems (96 changes in total), while those in the lower left part are toward more exclusive ones (30 changes in total).

direction. The general trend is also confirmed for all world regions: 27 more inclusive to 4 more exclusive changes in 13 countries of the Americas, 48 to 20 in 15 countries in Western Europe, 10 to 4 in 12 countries in Eastern Europe, and 11 to 2 in 10 countries in Asia-Pacific. When the set of electoral system changes is limited to those produced during present democratic periods, major changes between the four basic categories – indirect, majority, mixed and proportional – are 19, as shown in Table 1.5. Of these, 14 have been in the expected direction toward more inclusive formulas, while only 5 have been in the opposite direction – a proportion close to 3 to 1. The relatively less overwhelming proportion of inclusive changes within present democratic regimes in comparison to all previous periods is, of course, produced by the fact that indirect assembly elections have not been in use longer in present democracies and the proportion of

Table 1.5 Major changes of assembly electoral system in present democratic periods New system Previous system Majority Mixed Proportional

Majority

Mixed

Proportional

* 0 2

8 * 3

3 3 *

Note: Changes in the upper-right part are toward more inclusive systems (14 changes in total), while those in the lower-left part are toward more exclusive ones (5 changes in total).

58

Josep M. Colomer

majority rule systems has been reduced by previous changes toward more inclusive formulas. In spite of this, however, the number of changes from majority rule systems toward mixed and proportional representation systems is still very significant: 8 and 3, respectively, so 11 in total, as opposed to only 2 in the opposite direction. More interesting is the tie that can be observed in changes between mixed and proportional systems: 3 in each direction. But these refer to fewer countries: 2 having moved toward proportional systems (Bulgaria and Croatia) and 2 toward mixed systems (Madagascar and Italy, but the latter with a total of 3 changes, back and forth). The total balance does not make mixed systems a focal point of attraction from any other system: again, most of the new mixed systems have been the result of changes either from dictatorships or from majority rule electoral systems, while only a few have been established from previous proportional representation systems. Specifically, while 7 mixed systems have been established in recent processes of democratization from authoritarian regimes (not counted in Tables 1.3 and 1.4), 8 have been the result of reforms from majority rule systems, most of them also in new electoral democracies or semi-democracies moving fast towards more inclusive formulas in the 1990s (Albania, Armenia, Lithuania, Philippines, Russia, Thailand and Ukraine, while Japan is the only case having experienced such a move among older democracies). In contrast, only 2 countries have adopted mixed systems from proportional representation (Italy twice and Madagascar). This suggests that mixed systems are, in most cases, an ersatz for proportional representation in countries recently emerged from authoritarian regimes or restrictive majority rule systems, as well as an intermediate step towards proportional representation systems when democracy endures. In total, reverse changes from proportional representation to mixed or majority rule systems within present democratic regimes are very few, only 5 moves in 4 countries, while the number of new proportional systems established is still higher: 6. The balance between ‘inclusive’ and ‘exclusive’ moves is also positive in Eastern Europe: 6 to 0, and in Asia: 6 to 0. In Western Europe it looks negative, 2 to 4, but again this is due to several changes in just two countries: France moved twice from proportional representation to a majority rule system (in 1958 and 1988) with one inverse move in between (1986), and Italy followed the same trajectory from proportional to mixed twice (1953 and 1993) with proportionality re-established in between (1956). Taking also into account the eight subcategories of electoral system distinguished above, we have counted 30 changes within presently existing democratic regimes, as presented in Table 1.6. Of these, 22 have been in the expected direction toward more inclusive formulas, while only 8 have been in the opposite direction – a proportion of 2.75 to 1. For presidential electoral system changes, we have in total 28 cases in 14 countries, as shown in Table 1.7. Of these cases, 19 have been in the expected direction, while only 9 in the opposite – a proportion of 2.1 to 1. Actually 7 countries have experienced more than one change between electoral systems (while three other countries not included in the present count have suppressed direct presidential or prime

Strategy and History of Electoral System Choice

Table 1.6

59

Changes of assembly electoral system in present democratic periods New system

Previous system Majority1 Majority2 Majority3 Mixed Proportional1 Proportional2 Proportional3

Maj1

Maj2

Maj3

Mixed

Prop1

Prop2

Prop3

* 0 1 0 0 0 0

0 * 0 0 0 0 0

2 0 * 0 2 0 0

1 1 6 * 1 0 2

0 0 1 0 * 0 1

0 0 0 2 0 * 1

0 0 2 1 5 1 *

Note: Maj1: majority rules multi-member districts; Maj2: majority rules, limited or cumulative vote; Maj3: single-member districts; Mixed: mixed systems; Prop1: proportional representation, average district magnitude < 9; Prop2: proportional representation, average district magnitude > 8, closed lists; Prop3: proportional representation, average district magnitude > 8, open list or double vote. Changes in the upper-right part are toward more inclusive systems (22 changes in total), while those in the lower-left part are toward more exclusive ones (8 changes in total).

Table 1.7

Changes of presidential electoral system New system

Previous system Electoral college Simple-plurality rule Qualified-plurality rules Majority second-round Single transferable vote

College

Plurality

Q.-plurality

Majority

STV

* 2 0 1 0

6 * 0 4 0

1 2 * 2 0

4 5 1 * 0

0 0 0 0 *

Note: Changes in the upper-right part are toward more inclusive systems (19 changes in total), while those in the lower-left part are toward more exclusive ones (9 changes in total).

ministerial elections in recent periods: Germany, Estonia and Israel). All changes but one are in Latin America (the other is the replacement of college with direct elections in Finland). In summary, 11 electoral colleges have been replaced with direct elections, while 2 reverse moves in Argentina and Colombia have also been replaced with direct elections later on. On 7 occasions, simple-plurality rule has been replaced with more inclusive formulas (2 qualified-plurality rules and 5 absolute-majority rules with second rounds). Of the 4 reverse moves registered towards plurality rule, 3 have also been replaced with majority second-round rules in further periods. Looking at the whole story, 13 countries have experienced changes in different periods and produced a present electoral system different from the one in the most remote period registered. Of these, 10 countries have, in the end, moved in the expected direction toward more inclusive electoral systems, while only 3 have moved in the opposite direction. Again, the trend can be under-evaluated for not

60

Josep M. Colomer

including college or plurality elections in certain countries in some remote periods. With the available information, at least one country, Peru, has passed through all the stages in the expected order of increasing inclusiveness: electoral college in 1828, plurality rule in 1896, qualified-plurality (33 per cent) rule in 1931, and majority rule with a second-round runoff since 1978. The trend is even stronger during the present democratic period, in which 7 changes are registered, 6 in the expected direction and 1 in the opposite as presented in Table 1.8. Specifically, 2 countries have abandoned electoral colleges for direct elections with a second round (by qualified-plurality in Argentina and by absolute-majority in Finland), and 4 have replaced plurality rule with secondround procedures (by qualified-majorities in Nicaragua and by absolute-majority in Colombia, Dominican Republic and Uruguay). Only Ecuador has moved within second-round rules from absolute-majority to qualified-plurality. 4 The number and proportion of indirect elections and majority electoral systems should decrease over time, while countries adopting proportional representation should not abandon it in significant numbers. The evolution of electoral systems in the countries selected is, in fact, the result of the processes of choice analysed in the previous paragraphs. Results are summarized in Tables 1.9, 1.10 and 1.11. The dates chosen – 1874, 1922, 1960 and 2002 – correspond to peaks in successive waves of democratization, so that the number of cases in each date is higher than in immediately previous and following years. The number of cases increase over time

Table 1.8 Changes of presidential electoral system in present democratic periods New system Previous system Electoral college Simple-plurality rule Qualified-plurality rules Majority second-round Single transferable vote

College

Plurality

Q.-plurality

Majority

STV

* 0 0 0 0

0 * 0 0 0

1 1 * 1 0

1 3 0 * 0

0 0 0 0 *

Note: Changes in the upper-right part are toward more inclusive systems (6 changes in total), while those in the lower-left part are toward more exclusive ones (1 change in total).

Table 1.9

Number of basic assembly electoral systems over time

Electoral system

1874

1922

1960

2002

Indirect elections Majority rules Mixed systems Proportional representation

6 14 – –

2 11 2 18

– 16 – 23

– 20 18 51

Total countries:

20

33

39

89

Strategy and History of Electoral System Choice

Table 1.10

61

Number of assembly electoral systems over time

Electoral system

1874

1922

1960

2002

Indirect elections Majority rules1 Majority rules2 Majority rules3 Mixed systems Proportional rep1 Proportional rep2 Proportional rep3

6 6 3 5 – – – –

2 2 2 7 2 10 2 6

– 1 1 14 – 12 3 8

– 2 – 18 18 19 12 20

20

33

39

89

Total countries:

Note: Maj1: majority rules multi-member districts; Maj2: majority rules, limited or cumulative vote; Maj3: single-member districts; Prop1: proportional representation, average district magnitude < 9; Prop2: proportional representation, average district magnitude > 8, closed lists; Prop3: proportional representation, average district magnitude > 8, open list or double vote.

Table 1.11

Number of presidential electoral systems over time

Electoral system

1874

1922

1960

2002

Electoral college Simple-plurality rule Qualified-plurality rules Majority second-round Single transferable vote

6 – – – –

4 4 – 6 –

3 9 1 3 1

1 10 5 33 2

Total countries:

6

14

17

51

until covering all 89 countries with more than one million inhabitants that can be considered democratic in 2002. (Four of the countries included in the analysis of past periods have no present democratic regimes – Belarus, Cuba, Nigeria, Pakistan – and one no longer exists – Czechoslovakia.) The data collected strongly confirm the hypothesis that exclusive electoral systems are more easily replaced. Indirect assembly elections decreased and virtually disappeared in the early twentieth century. Majority rule, which was the basic formula for broadening suffrage rights, opening political competition and democratizing in the late nineteenth and early twentieth centuries, was replaced in its appeal by proportional representation, especially after the First World War; this trend has intensified in recent processes of democratization. Mixed systems have also spread widely in the most recent periods, although – as mentioned – this has been more a result of changes from majority rule systems than from proportional representation ones. In total, for assembly elections, much more than half of the present democracies in countries with more than one million inhabitants use proportional representation (57 per cent), while less than one-fourth use majority rule systems (23 per cent) and one-fifth use mixed systems (20 per cent).

62

Josep M. Colomer

The evolution can also be observed within subcategories. Within majority rule systems, multi-member districts, which were still 43 per cent of cases in the late nineteenth century, have been reduced to 10 per cent; within proportional representation systems, those with small districts (magnitudes lower than nine) were a majority still in the 1920s, but they are now barely a third; within those proportional systems with large district magnitudes, almost two-thirds use open lists and similar procedures rather than closed lists. For presidential electoral systems, the evolution is also very strongly in the expected direction. The number of cases also increase over time up to 51 countries. Electoral college procedures, which were once popular in the Americas, have been reduced to the single case of the United States. Simple plurality rule, which was a typical formula for new direct presidential elections in the past, is currently used by only one-fifth of the countries. Absolute-majority rule with a second round had already been popular in the early twentieth century, although at the time in all but one of the six cases registered the second round was transferred to Congress. Nowadays, relatively inclusive majority rule and qualified-plurality rules (the latter a recent innovation in a few countries), both with a direct second-round runoff between the two most voted candidates, encompass together three-fourths of the countries with presidential elections. The single transferable and supplementary votes remain the experience of only two countries in the world. The change of an electoral system should be more likely the higher the number of effective parties in the initial institutional setting. The focus of our analysis is the immediately previous political situation from which a decision to change the electoral system may be made, including constituent assemblies, electoral support of the incumbent government, and negotiation or threat power of opposition parties. The formula for the effective number of parties captures the number of parties weighted by their size (N = 1/Σ p i2, where pi is the proportion of votes for each party i). It can be presumed that the higher the effective number of parties, the weaker the expectation for any single party to become the sure winner, and thus the more likely its preference for an inclusive electoral system permitting multi-partism to develop. Although the party’s support in votes may not be transferred to a corresponding degree of negotiation power within the institutions as a consequence of being distorted by the existing restrictive electoral system, it can be considered a rough proxy for their strength and expectation of support, which is the basis of the corresponding threat or negotiation power. Expectations, however, may also form from recent trends marked by the emergence of new, still small, but already rising parties or pessimistic forecasts regarding its own support. Data on the effective numbers of parties in votes have been calculated for elections immediately prior to 49 major electoral system changes in favour of more inclusive electoral formulas, that is from indirect or majority rule systems to mixed or proportional representation systems for the assembly and from college or plurality rule systems to second-round formulas for the presidency, as well as

Strategy and History of Electoral System Choice

63

for some pre-eminent occasions for such changes, such as constituent assemblies, that have not produced a major change to already existing inclusive electoral systems. The only cases excluded are those in which a new electoral system has been preceded by a significantly long authoritarian period without elections, thus making the former party system irrelevant for the new electoral system choice, or those during the nineteenth century with unavailable data. The average number of effective parties immediately before 49 electoral system changes in favour of more inclusive formulas in 36 countries is 3.8, thus making it clear that multiparty systems are a fact prior and not only subsequent to the adoption of assembly rules with proportional representation or presidential rules with a second round. The observations collected here suggest that when the number of effective parties is not much higher than three, one or two larger parties can expect to become absolute winners under the existing majoritarian rules and, consequently, they will tend to maintain them, while when the effective number of parties increases, in contrast, any party can risk becoming an absolute loser under majority rules, so they may prefer to move to mixed or proportional representation systems securing them a fair portion of seats.5 A sample of changes in specific regions, time periods and type of electoral systems is presented in Table 1.12. First, data for nine early, innovative changes of electoral system in Western European countries from majority rule systems to proportional representation in the early twentieth century show that the average number of effective parties in the previous assembly elections was 4.0. This means that, just before the introduction of proportional representation, multi-party systems already existed, certainly not as a consequence of the existing majority rule electoral system but in spite of it and as a factor for its change. As a contrast, the same index has been calculated for a few failed reforms of the electoral system. The introduction of proportional representation or a mixed system in the United Kingdom failed in both 1918 and 1998 in the environment of a low degree of multi-partism, which can be estimated at 2.4 and 3.1 effective parties in votes, respectively (and, of course, much more reduced in seats). In particular, on the latest occasion, the Labour Party programme for introducing proportional rules could have been encouraged by the relatively high degrees of multi-partism in votes in the two previous elections, in 1987 and 1993, as a consequence of the higher dispersion of votes between Conservatives, Labour and the Liberal/Social Democratic alliance. The Labour victory in 1997, however, somehow dismissed this trend and re-established a single-party domination. Reverse changes from proportional representation to majority rule or mixed systems have also been attempted on a few interesting occasions in countries of Western Europe. In Germany it was formally promoted in 1967 by the would-be dominant Christian-Democrats, who could find encouragement in the relatively reduced and decreasing numbers of effective parties in the two previous elections –3.5 in 1961 and 3.1 in 1965. But the degree of pluralism was still sufficiently high to provoke the rejection of a move towards a more exclusive electoral system not only from the smaller parties but also from the second-in-size but

64

Josep M. Colomer

Table 1.12

Effective number of parties and electoral system change

Country

Year

Previous election Year

Assembly: Changes from majority rule to proportional representation Belgium 1899 1898 Sweden 1909 1908 Denmark 1915 1913 Germany 1918 1912 Netherlands 1918 1913 Switzerland 1918 1917 Italy 1919 1913 Norway 1919 1918 Greece 1932 1928

Type

As. As. As. As. As. As. As. As. As.

Average Costa Rica Chile Colombia Cuba Brazil Bolivia Argentina Ecuador Peru Guatemala

1913 1925 1931 1940 1945 1952 1963 1978 1979 1985

1913 1921 1930 1940 1945 1951 1962 1978 1978 1984

1918 1998

1910 1997

3.7 2.6 3.9 5.9 6.6 3.4 2.7 3.6 3.4 4.0

Pdt. As. Pdt. Ct.As. As. Pdt. As. Pdt. Ct.As. Ct.As.

Average Failed changes: UK

ENP

2.9 5.6 2.0 7.5 3.7 3.3 6.7 4.8 4.8 7.7 4.9

As. As.

2.4 3.1

Assembly: Changes from proportional representation to majority rule or mixed systems France 1958 1956 As. 6.1 1988 1986 As. 4.7 Italy 1953 1948 As. 2.9 1993 1992 As. 6.6 Failed changes: Germany Netherlands

1967 1977

1965 1972

As. As.

Presidential elections: Changes from college or plurality rule to qualified-plurality or absolute-majority with second round Costa Rica 1936 1932 Pdt. Argentina 1972 1965 As. 1994 1994 Ct.As. Ecuador 1978 1978 Pdt. Peru 1979 1978 Ct.As. Guatemala 1985 1984 Ct.As. Brazil 1986 1982 As. Colombia 1991 1990 Ct.As.

3.1 6.9

2.8 4.9 3.0 4.8 4.8 7.7 2.7 2.2

Strategy and History of Electoral System Choice

Dominican R. Uruguay

1995 1996

1994 1994

As. As.

Average Failed changes: United States Honduras Venezuela

1956 1969 1981 1999

1954 1968 1980 1998 1998

65

2.8 3.3 3.9

As. As. Ct.As. Ct.As. Pdt.

2.0 2.0 2.1 6.3 2.1

Note: ENP: Effective number of parties = 1/∑ pi2, where pi is the proportion of votes for party i. Previous elections: As. – Assembly; Pdt. – Presidency; Ct.As. – Constituent Assembly. Internal factionalization of parties has not been taken into account, although it was high, for instance, in Brazil and Colombia.

wary Social-Democrats. In the Netherlands the Social-Democrats promoted a similar change in 1977, but, in the face of a very high level of multi-partism – measurable by the 6.9 effective parties in the previous election – they blatantly failed in their purpose. In two other Western European countries moves from proportional representation to majority or mixed systems have been more successful, as previously mentioned. In France, as is well known, the moves in 1958 and 1988 were completed in spite of having systems with 6.1 and 4.7 effective parties in the respective previous elections. Major institutional regime changes, initially implemented through a coup d’état, helped success on the first occasion, while on the second it implied the re-establishment of the previously existing electoral system after a single election with the new one. In Italy, as also mentioned, the first change from a proportional to a mixed system was attempted in 1953, when the governing Christian-Democrats were encouraged by a dramatic reduction in the degree of multi-partism in the previous two elections, from 5.6 effective parties in 1946 to 2.9 in 1948. They, however, failed to obtain the 50 per cent of popular votes that they themselves had targeted as the condition for the majority rule component of the new system to be applied and re-established proportional representation. The second reform to a new mixed system was introduced in 1993 in the environment of a high degree of multi-partism, measurable at 6.6 effective parties, but under the illusion among some left circles that the ongoing dissolution of the ChristianDemocratic party would open a new period of hegemony of the left. During the first three subsequent elections in which the new system has been used the number of parties has decreased slightly but not dramatically, so making the system still vulnerable to further changes. Consistent with the general hypothesis here discussed, for ten innovative changes of assembly electoral system from majority rules to proportional representation in countries of Latin America during the twentieth century we also find a high average number of effective parties in the constituent assembly or the immediately previous election, at 4.9, as shown in Table 1.12. Similarly, ten changes of presidential

66

Josep M. Colomer

electoral systems from electoral college or plurality rule formulas to qualifiedplurality or absolute-majority formulas with a second round have been introduced, most recently, in nine countries in which the average number of effective parties was 3.9, as also shown in Table 1.12. Likewise, a few formal occasions of changing presidential electoral rules toward more inclusive formulas have failed in the context of pure two-party systems: in the United States in 1956 and 1969 where the corresponding constitutional amendments to replace the electoral college with direct elections passed the House but died in the Senate, and in Honduras in 1981 where the constituent assembly retained plurality rule. In contrast, plurality rule for the election of president was also maintained in the new constitution of Venezuela in 1999, in spite of having been approved by a constituent assembly elected with 6.3 effective parties, together with other restrictive institutional reforms concentrating powers in the presidency. This decision was encouraged by the results of the previous presidential election, won at a single round by a large coalition majority with 56 per cent of votes and only 2.1 effective parties in the system. However, in spite of the very high concentration of power in the presidency, the high degree of legislative multi-partism sustained further pressure for opening up the political system in Venezuela. Change of an electoral system could be more likely the shorter its duration. If this hypothesis were correct, the duration of the electoral systems that have been changed should be, as a whole, lower than the presently existing systems. Testing this hypothesis, however, is difficult because we do not know whether and when the present electoral systems will be changed in the future; their duration, as measured until the end of 2002, will thus be undervalued. The duration of the four basic types of assembly electoral system in terms of the number of elections at which they have been used is summarized in Table 1.13. Indirect elections disappeared, as previously observed. For each of all the presently existing basic electoral system types – that is those based on majority rules, mixed systems and proportional representation – the data collected seems to support the hypothesis presented above: electoral systems that were replaced in the past had been, on average, less durable than those presently existing. Also, restrictive electoral systems were replaced even though they had endured a long Table 1.13 Duration of assembly electoral systems Electoral system Indirect

Majority

Mixed

Proportional

Systems replaced in the past No. of electoral systems counted Average no. of elections/system

17 9.4

40 10.2

9 2.9

12 3.9

Current systems (end of 2002) No. of electoral systems counted Average no. of elections/system

0 0

22 15.7

18 3.5

50 11.2

Strategy and History of Electoral System Choice

67

time, while more inclusive systems could only be replaced if they had not had the time to obtain endogenous support. Specifically, systems based on indirect elections or majority rules lasted for about ten elections each on average before they were replaced, while newer and more inclusive systems, that is mixed and proportional systems, were changed very soon: in most cases after just one election (and after less than three or four elections on average, respectively) – in other words, before they were entrenched in the organization and action of already existing political parties and electoral candidates. With regard to the presently existing electoral systems, both majority rule and proportional representation systems have endured some considerable time on average, which would imply that most of them may be difficult to change in the future. The duration of present majority systems, however, relies significantly on the extreme case of the House of Representatives of the United States, which accounts for almost a third of the total number of elections held with presently existing majority systems; if this is dropped from the sample, the average duration of presently existing majority systems is reduced to 11.3, only slightly higher than those that have been replaced in the past. In contrast, most of the presently used mixed systems are still recent, having been used in most cases for only two or three elections, which may suggest that a number of them (that is, the actors playing within the existing systems) may be able to present softer resistance to change than most of the majority and proportional current systems. The available data on the duration of presidential electoral systems, as shown in Table 1.14, are not able to give support to the hypothesis previously presented regarding the longer duration of existing electoral systems. This may be partly due to the inherent exclusiveness of all presidential electoral systems, which necessarily produce a single absolute winner and may provoke subsequent rejection by all the losers, in contrast to many assembly electoral systems permitting multiple parties to exist and share representation. It can be noted that, in the past, the average duration of presidential electoral systems was even shorter than the average duration of assembly electoral systems that were also based upon plurality or majority rules (shown in Table 1.13). Low numbers of elections for current presidential electoral systems also reflect the fact that most of them exist in recent democracies in Latin America and are thus still necessarily short-living. All the current formulas Table 1.14

Duration of presidential electoral systems Electoral system College

Plurality

Q. plurality

Majority

STV

Systems replaced in the past No. of electoral systems counted Average no. of elections/system

11 8.2

9 7.3

1 6.0

7 7.7

0 0

Current systems (end of 2002) No. of electoral systems counted Average no. of elections/system

1 54

11 6.9

5 5.4

34 4.1

2 6.5

68

Josep M. Colomer

of direct presidential elections have been used, so far and on average, during shorter periods than those that were changed in the past – in most cases, for only two or three elections – which may suggest that they may still be significantly vulnerable to further changes.

Notes 1. To be fair, Maurice Duverger also subtly noted that ‘the first effect of proportionality is to maintain an already existing multiplicity . . . On the whole, proportional representation maintains virtually without change the party system existing at the time of its adoption’, but he did not elaborate (Duverger, 1950: 346). For an early, although also brief, statement more explicitly putting the hypothetical line of causality upside down, see John G. Grumm: ‘the generally-held conclusions regarding the causal relationships between electoral systems and party systems might well be revised. . . . it may be more accurate to conclude that P.R. is a result rather than a cause of the party system in a given country’ (Grumm, 1958: 375). More extensively, Leslie Lipson developed some historical analysis from the premise that ‘chronologically, as well as logically, the party system is prior to the electoral system’ (Lipson, 1964: 343); similarly, Bo Särlvick (1982) developed a strategic analysis of the choice of electoral systems in Scandinavian countries; Vernon Bogdanov echoed the hypothesis that electoral systems could be the dependent variable, but he denied any possibility of generalizing (Bogdanor, 1982: 254–61). Other authors could be mentioned that pointed out a relation between social structures rather than political parties and electoral systems, a subject deserving to be dealt with elsewhere. For sources and discussion, see also Gary Cox (1997: 14–19). In more formal literature, Anthony Downs assumed endogeneity of the electoral system (Downs, 1957), while David P. Quintal sketched a formal model of electoral system choice based on benefit-cost analysis (Quintal, 1970). 2. Terms in bold italics throughout the text may be found in the Glossary. 3. Other quotas proposed and used in the United States for apportioning a high number of seats among a relatively smaller number of states, such as those invented by former president John Quincy Adams and by mathematicians Edward V. Huntington and Joseph A. Hill, cannot be used for elections with more candidates than seats; see Balinski and Young (1982), Marshall, Olkin and Pukelsheim (2002). 4. The findings presented here, although based on differently formulated hypotheses and different categories of electoral systems, are consistent with other empirical tests. Specifically, Arend Lijphart analysed electoral system changes in 16 democratic countries from 1945 to1999, finding that more than two-thirds had been in favour of greater proportionality (Lijphart, 1994: 52–6). André Blais and Louis Massicotte analysed electoral systems existing in 166 democratic and non-democratic countries around 1995, finding, among other correlations, that the more democratic a country the more likely it is to adopt proportional representation and the less likely it is to have a majority system (Blais and Massicotte, 1997). 5. In addition to the 49 cases reported here of electoral system changes in favour of more inclusive formulas or constituent assemblies confirming proportional representation or presidential runoff rules, we also have 7 cases in 3 countries of changes in the opposite direction. In order to increase the number of cases with changes towards more exclusive formulas and to correct some selection bias, a logit regression has been run with all those cases mentioned plus the present (end of 2002) assembly electoral systems in all 89 democratic countries with more than one million inhabitants. The latter addition is based on the assumption, central in the present work, that the number of parties is not only produced or permitted by the electoral system but it also supports its continuity. The result of the logit regression between effective number of parties and electoral systems adopted in the past or supported in the present for 142 cases is in the expected sign and significant, thus supporting our hypothesis.

Strategy and History of Electoral System Choice

69

Logit regression: Dichotomous dependent variable: log (X/1 − X), with X = 1 for changes in favour of more inclusive systems or present proportional representation, mixed or second-round presidential systems, and X = 0 for changes in favour of more exclusive systems or present indirect elections, assembly majority rules, or college or plurality presidential systems. Independent variable: the effective number of parties in the election immediately prior to electoral system change or constituent assembly or the latest election with the present systems before 2003. Number of cases: 142. Coefficient: 0.259. Standard error: 0.154. Significance: 0.092. Cases included: from indirect to majority to mixed or to proportional systems – Albania 1992, Argentina 1963, Argentina 1972, Argentina 1994 (presidential (P)), Armenia 1995, Belgium 1899, Bolivia 1952, Brazil 1945, Brazil 1986 (P), Bulgaria 1991, Chile 1925, Colombia 1931, Colombia 1991 (P), Costa Rica 1913, Costa Rica 1936 (P), Croatia 2000, Cuba 1940, Denmark 1915, Dominican Republic 1995 (P), Ecuador 1978, Ecuador 1997, El Salvador 1984, France 1946, France 1986, Germany 1918, Greece 1932, Greece 1936, Greece 1955, Greece 1958, Guatemala 1985, Indonesia 2002, Italy 1919, Italy 1946, Italy 1956, Japan 1994, Moldova 1993, Netherlands 1918, New Zealand 1993, Norway 1919, Peru 1933, Peru 1979, Philippines 1995, Russia 1993, South Africa 1994, Sweden 1909, Switzerland 1918, Thailand 1997, Ukraine 1993, Uruguay 1996 (P); in the opposite direction – France 1958, France 1988, Greece 1928, Greece 1933, Greece 1952, Italy 1953, Italy 1993. Present electoral systems: all 94 countries in Summary Table 1A except Belarus, Czechoslovakia, Cuba, Nigeria and Pakistan.

References Alberigo, Giuseppe et al. (eds) (1991) Conciliorum Oecumenicorum Decreta. Bologna: Dehoniane (English trans. from a previous edition by Norman Tanner (ed.), Decrees of the Ecumenical Councils. Washington, DC: Georgetown University Press, 1990). Alfonso X (1990) Siete partidas (Spanish Law Code). Madison, WI: Hispanic Seminary of Medieval Studies. Andrae, Poul (1905) Andrae og Hans Opfindelse Forholdstals Valginaaden. Copenhagen (fragments trans. Andrae and His Invention. The Proportional Representation Method, in Clarence Hoag and George Hallett, Proportional Representation. New York: Macmillan, 1926). Aristotle ([325 BC] 1997) The Politics. Chapel Hill and London: University of North Carolina Press. Babeau, Albert (1882) Le village sous l’Ancien Régime. Paris: Didier. Babeau, Albert ([1894] 1972) La province sous l’Ancien Régime. New York: AMS Press. Baldwin, Marshall W. (1968) Alexander III and the Twelfth Century. New York: Newman. Balinski, Michel L. and H. Peyton Young ([1982] 2001) Fair Representation. Meeting the Ideal of One Man, One Vote. Washington, DC: Brookings Institution Press. Bentham, Jeremy ([1834] 1999) Political Tactics. New York and Oxford: Oxford University Press. Bishop, Cortlandt F. (1893) History of Elections in the American Colonies. New York: Columbia College. Black, Duncan (1996) A Mathematical Approach to Proportional Representation: Duncan Black on Lewis Carroll, eds Iain McLean, Alistair McMillan and Burt L. Monroe. Boston: Kluwer. Bogdanor, Vernon (1982) ‘Conclusion: Electoral Systems and Party Systems’, in Vernon Bogdanor and David Butler (eds), Democracy and Elections. Electoral Systems and Their Consequences. Cambridge and New York: Cambridge University Press, pp. 247–62. Borda, Jean-Charles de (1781) ‘Mémoire sur les elections au scrutin’, in Mémoires de l’Académie royale des sciences. Paris (English trans. ‘On Elections by Ballot’, in Iain McLean and Arnold B. Urken (eds), Classics of Social Choice. Ann Arbor: University of Michigan Press, 1995). Brucker, Gene (1963) Florentine Politics and Society, 1343–1378. Princeton, NJ: Princeton University Press.

70

Josep M. Colomer

Cadart, Jacques (1952) Le régime electoral des États Généraux de 1789 et ses origines (1302– 1614). Paris: Librairie du Recueil Sirey. Carrère, Claude (1967) Barcelone, centre économique à l’époque des difficultés 1380–1462. ParisThe Hague: Mouton. Chiaramonte, José Carlos (1995) ‘Vieja y nueva representación: los procesos electorales en Buenos Aires, 1810–1820’, in Antonio Annino (ed.), Historia de las elecciones en Iberoamérica, siglo XIX. Mexico: Fondo de Cultura económica, pp. 19–64. Colomer, Josep M. and Iain McLean (1998) ‘Electing Popes: Approval Balloting and Qualified-Majority Rule’, Journal of Interdisciplinary History, 29, 1. Condorcet, Jean-Antoine-Nicolas de Caritat, Marquis De (1785) Essai sur l’application de l’analyse à la probabilité des décisions rendues à la pluralité des vois. Paris: Imprimerie royale (reprinted New York: Chelsea, 1972. English trans. ‘An Essay on the Application of Analysis to the Probability of Decisions Rendered by a Plurality of Vote’, in Iain McLean and Arnold B. Urken (eds), Classics of Social Choice. Ann Arbor: University of Michigan Press, 1995). Condorcet, Jean-Antoine-Nicolas de Caritat, Marquis De (1787) ‘New Haven Letters’ (English trans. in Condorcet. Foundations of Social Choice and Political Theory, eds Iain McLean and Fiona Hewitt. Aldershot: Edward Elgar, 1994). Condorcet, Jean-Antoine-Nicolas de Caritat, Marquis De (1793a) Plan de constitution présenté à la Convention nationale. Paris: Imprimerie nationale (English trans. ‘Outline for the French Constitution’, in Condorcet. Foundations of Social Choice and Political Theory, eds Iain McLean and Fiona Hewitt. Aldershot: Edward Elgar, 1994). Condorcet, Jean-Antoine-Nicolas de Caritat, Marquis De (1793b) ‘On Elections’ (English trans. in Condorcet. Foundations of Social Choice and Political Theory, eds Iain McLean and Fiona Hewitt. Aldershot: Edward Elgar, 1994). Coroleu Inglada, Josep and Josep Pella Forgas ([1876] 1993) Las Cortes catalanas: estudio jurídico y comparativo. Valencia: Librería Paris. Cox, Gary W. (1997) Making Votes Count. Strategic Coordination in the World’s Electoral Systems. New York: Cambridge University Press. Dahl, Robert A. (2002) How Democratic is the American Constitution? New Haven, CT: Yale University Press. D’Hondt, Victor (1878) La répresentation proportionelle des parties. Ghent. D’Hondt, Victor (1882) Système pratique et raisonné de répresentation proportionelle. Brussels: C. Muquardt. Dodgson, Charles L. (Lewis Carroll) (1884) Principles of Parliamentary Representation. London: Harrison (in Iain McLean and Arnold B. Urken (eds), Classics of Social Choice. Ann Arbor: University of Michigan Press, 1995). Downs, Anthony (1957) An Economic Theory of Democracy. New York: Harper. Droop, Henry R. (1868) Methods of Electing Representatives. London: Macmillan. Droop, Henry R. (1869) ‘On the Political and Social Effects of Different Methods of Electing Representatives’, Papers Read Before the Juridical Society, vol. 3, pp. 469–507. Duverger, Maurice (1950) L’influence des systèmes électoraux sur la vie politique, Cahiers de la Fondation Nationale des Sciences Politiques, 16. Paris: Armand Colin. Duverger, Maurice (1951) Les partis politiques. Paris: Seuil (English trans. Political Parties: Their Organization and Activity in the Modern State. New York: Wiley, 1954). Emmerich, Gustavo-Ernesto (1985) ‘Las elecciones en México, 1808–1911: ¿sufragio efectivo? ¿no reelección?’, in Pablo González Casanova (ed.), Las elecciones en México, evolución y perspectivas. Mexico: Siglo XXI, pp. 41–68. Eusebius Pamphilus ([324] 1953) Ecclesiastical History. Washington, DC: Catholic University of America Press. Favre, Henri (1986) L’indigénisme. Paris: Presses Universitaires de France. Fernández Domínguez, Arturo (1992) Leyes electorales españolas de diputados a Cortes en el siglo XIX. Estudio histórico y jurídico-politico. Madrid: Civitas.

Strategy and History of Electoral System Choice

71

Ferro, Víctor (1987) El dret públic català. Les institucions a Catalunya fins al Decret de Nova Planta. Vic: Eumo. Fishburn, Peter C. (1973) The Theory of Social Choice. Princeton, NJ: Princeton University Press. Fisher, J. Francis (1863) The Degradation of Our Representative System and Its Reform. Philadelphia. Freeman, Edward A. (1863) History of Federal Government. From the Foundation of the Achaian League to the Disruption of the United States. London and Cambridge: Macmillan. Gergonne, Joseph-Díaz (1820) ‘Arithmétique politique. Sur les élections et le système représentatif’, Annales de Mathématiques. Gilpin, Thomas (1844) On the Representation of Minorities of Electors to Act with the Majority in Elected Assemblies. Philadelphia: John C. Clark (reprints in 1872, in Proceedings of the American Philosophical Society, vol. 4; and in 1896, in E. J. James (ed.), An Early Essay on Proportional Representation (by Thomas Gilpin). Philadelphia: American Association for Political and Social Science). González Antón, Luis (1978) Las Cortes de Aragón. Zaragoza: Librería General. Goodwin, Barbara (1992) Justice by Lottery. London: Harvester-Wheatsheaf. Gragg, Florence A. and Leona C. Gabel (eds) (1959) Memoirs of a Renaissance Pope. The Commentaries of Pius II. New York: Putnam’s Sons. Grotius ([1624] 1948) De Juri Belli ac Pacis. The Hague: Nijhoff. Grumm, John G. (1958) ‘Theories of Electoral Systems’, Midwest Journal of Political Science, 2, 4, 357–76. Hagenbach-Bischoff, Eduard (1888) Die Frage der Einführung einer Proportionalvertretung. Basel. Hare, Thomas (1857) The Machinery of Representation. London: Maxwell. Hare, Thomas (1859 [and further editions in 1861, 1865, 1873]) On the Election of Representatives. Parliamentary and Municipal: A Treatise. London: Longmans Green. Headlam-Morley, James W. (1981) Election by Lot at Athens. Cambridge and New York: Cambridge University Press. Hobbes, Thomas ([1651] 1996) Leviathan. Cambridge and New York: Cambridge University Press. Hume, David ([1741] 1987) ‘Of Parties in General’, in Essays Moral, Political, and Literary. Indianapolis, IN: Liberty. Hume, David ([1758] 1987) ‘Of the Coalition of Parties’, in Essays Moral, Political, and Literary. Indianapolis, IN: Liberty. Jefferson, Thomas ([1792] 1904) ‘Opinion on Apportionment Bill’, in The Writings of Thomas Jefferson. Washington, DC, vol. 3. Konopcynski, Ladislas (1930) Le liberum veto. Étude sur les developments du principe majoritaire. Paris: Librairie Ancienne Honoré Champion/Warsaw: Librairie Gebethner et Wolff. Leibniz, Gottfried W. F. ([1701] 1988) ‘Letter to Thomas Burnett, II’, in Political Writings, ed. Patrick Riley. Cambridge and New York: Cambridge University Press. Lines, Marji (1986) ‘Approval Voting and Strategy Analysis: A Venetian Example’, Theory and Decision, 20: 155–72. Lipson, Leslie (1964) The Democratic Civilization. New York: Oxford University Press. Lloyd, Henry D. (1907) A Sovereign People. A Study in Swiss Democracy. New York: Doubleday. Llull, Ramon [c. 1274] Artificius electionis personarum (English trans. in Günter Hägele and Friedrich Pukelsheim, ‘Llull’s Writings on Electoral Systems’, Studia Lulliana, 41: 3–38). Llull, Ramon ([1283] 1982) Blanquerna. Barcelona: Edicions 62 (fragment trans. in Iain McLean and Arnold B. Urken (eds), Classics of Social Choice. Ann Arbor: University of Michigan Press, 1995, as well as in Günter Hägele and Friedrich Pukelsheim, ‘Llull’s Writings on Electoral Systems’, Studia Lulliana, 41: 3–38). Llull, Ramon ([1299] 1937) De arte electionis, in H. Finke (ed.), Gesammelte Aufsätze zur Kulturgeschichte Spaniens, I, 6. Münster: Aschendorffschen (English trans. in Iain McLean and

72

Josep M. Colomer

Arnold B. Urken (eds), Classics of Social Choice. Ann Arbor: University of Michigan Press, 1995, as well as in Günter Hägele and Friedrich Pukelsheim, ‘Llull’s Writings on Electoral Systems’, Studia Lulliana, 41: 3–38). Locke, John ([1689] 1963) Two Treatises of Government. New York: Cambridge University Press. Madison, James ([1788] 1961) ‘No. 10’, in Alexander Hamilton, James Madison and John Jay, The Federalist Papers. New York: New American Library. Marín-Bosch, Miguel (1994) Votos y vetos en la Asamblea General de las Naciones Unidas. Mexico: Fondo de Cultura Económica. Marshall, Albert W., Ingram Olkin and Friedrich Pukelsheim (2002) ‘A Majorization Comparison of Apportionment Methods in Proportional Representation’, Social Choice and Welfare, 19: 885–900. Marshall, James G. (1853) On Minorities and Majorities. London. Marsilius of Padua ([1324] 2001) Defensor pacis (English trans.). New York: Columbia University Press. Mirabeau [Honoré-Gabriel Riqueti, Count De Mirabeau] ([1789] 1834) ‘Discours prononcé à la Tribune Nationale, États de Provence. Aix, 30 janvier 1789, in Oeuvres. Paris: Lecointe et Pougin, Vol. I. Mill, John Stuart ([1861] 1963) Considerations on Representative Government, in Collected Works. Toronto: University of Toronto Press, vol. 18. Montesquieu [Charles-Louis de Secondat, baron de la Brède et de Montesquieu] ([1748] 1964) L’esprit des lois, in Oeuvres complètes. Paris: Seuil (English trans. The Spirit of Laws. Cambridge and New York: Cambridge University Press, 1989). Nanson, E. J. ([1882] 1995) ‘Methods of Election, 1882’, in Iain McLean and Arnold B. Urken (eds), Classics of Social Choice. Ann Arbor: University of Michigan Press. Nicholas of Cusa (Nicolaus Cusanus) ([1433] 1995) De concordantia catholica, in Iain McLean and Arnold B. Urken (eds), Classics of Social Choice. Ann Arbor: University of Michigan Press. Penn, William (1693) Toward the Present and Future Peace of Europe by the Establishment of a European Diet, Parliament or Estates. London. Plato ([360 BC] 1988) The Laws. Harmondsworth and New York: Penguin. Pole, Jack R. (1971) Political Representation in England and the Origins of the American Republic. Berkeley: University of California Press. Quintal, David P. (1970) ‘The Theory of Electoral Systems’, Western Political Quarterly, 23, 4, 752–61. Rousseau, Jean-Jacques ([1762] 1997) The Social Contract. Cambridge and New York: Cambridge University Press. Rueda, José-Carlos (ed.) (1998) Legislación electoral española (1808–1977). Barcelona: Ariel. Ruffini Avondo, Edoardo (1925) ‘Il principio maggioritario nelle elezioni dei Re e Imperatori romano-germanici’, Atti della Reale Accademia delle Scienze di Torino 1924–1925, vol. 60, pp. 392–414, 441–92, 557–74. Rustow, Dankwart A. (1950) ‘Some Observations on Proportional Representation’, Journal of Politics, 12, 1. Saint-Just, Antoine-Louis-Leon De ([1793] 1908) ‘Essai de Constitution’, in Oeuvres complètes, ed. Charles Vellay. Paris: Librairie Charpentier et Fasquelle, vol. 1, pp. 433–45. Sainte-Laguë, André (1910) ‘La répresentation proportionelle et la méthode des moindres carrés’, Comptes Rendus Hébdomadaires des Séances de l’Académie des Sciences, vol. 151, pp. 377–8 (English trans. ‘Proportional Representation and the Method of Least Squares’, in Arend Lijphart and Robert W. Gibberd (1977) ‘Thresholds and Payoffs in List Systems of Proportional Representation’, European Journal of Political Research, 5, Appendix 2: 241–2). Särlvik, Bo (1982) ‘Scandinavia’, in Vernon Bogdanor and David Butler (eds), Democracy and Elections. Electoral S7ystems and Their Consequences. Cambridge and New York: Cambridge University Press, pp. 123–48.

Strategy and History of Electoral System Choice

73

Sieyès, Emmanuel-Joseph ([1789] 1989) Que’est-ce que le Tiers-état?, in Oeuvres. Paris: Edhis (English trans. What Is the Third Estate? New York: Praeger, 1964). Sydnor, Charles S. (1962) American Revolutionaries in the Making. Political Practices in Washington’s Virginia. New York: Collier.

74

Josep M. Colomer

Summary Table 1A • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

• • • •

• • • • •

Assembly electoral systems

Albania: Maj3 – 1991 – l; Mix – 1992 – 3 Argentina: Maj1 – 1826 – 1 // 1853 – 9; Maj2 – 1912 – 10 // Maj3 – 1946 – 2 // 1957 – 4 // PR2 – 1963 – 1 // 1973 – 1 // 1983 – 10 Armenia: Maj3 – 1990 – 1; Mix – 1995 – 1; 1999 – 1 Australia: Maj3 – 1902 – 6; 1918 – 33 Austria: PR 1 – 1918 – 4 // 1945 – 8; PR3 – 1970 – 9 Bangladesh: Maj3 – 1971 – 1 // 1979 – 1 // 1991 – 3 Belarus: Maj3 – 1996 – 2 Belgium: Maj1 – 1831 – 17; 1877 – 11; PR1 – 1899 – 8; PR 3 – 1919 – 7 // 1946 – 18 Benin: PR1 – 1991 – 3 Bolivia: Maj2 – 1924 – ?; PR2 – 1952 – 5 // 1966 – 1 // 1980 – 4; PR3 – 1994 – 2 Bosnia-Hercegovina: PR1 – 1996 – 4 Botswana: Maj3 – 1965 – 8 Brazil: Ind – 1824 – 4 // 1860 – 4; 1875 – 1; Maj3 – 1881 – 2; Maj1 – 1890 – 1; Maj2 – 1892 – 4; 1904 – 5 // 1927 – 2; Mix – 1932 – 1 // PR3 – 1945 – 3; 1950 – 3 // 1982 – 6 Bulgaria: Mix – 1990 – 1; PR2 – 1991 – 4 Canada: Maj3 – 1867 – 2; 1874 – 25; 1972 – 9 Cape Verde: PR 1 – 1992 – 2 Central African Rep.: Maj3 – 1993 – 2 Chile: Maj1 – 1833 – ?; Maj2 – 1879 – ?; PR 1 – 1925 – 1 // 1932 – 11 // 1980 – 3 Colombia: Maj1 – 1857 – ?; Maj3 – 1886 – ?; Maj2 – 1910 – 5; PR1 – 1931 – 10; Mix – 1951 – 2 // PR 1 – 1958 – 6; 1970 – 10 Costa Rica: Ind – 1893 – 10; Mix – 1913 – 15 // PR 1 – 1953 – 13 Croatia: Mix – 1995 – 1; PR 2 – 2000 – 1 Cuba: Ind – 1901 – 10 // PR2 – 1940 – 6 // Czechoslovakia: PR1 – 1920 – 4 // PR 3 – 1990 –1 Czech Rep.: PR3 – 1992 – 3; 2002 – 1 Denmark: Ind – 1834 – ?; Maj3 – 1848 – 22; 1901 – 7; PR3 – 1915 – 1; PR1 – 1920 – 11; PR3 – 1953 – 18 Dominican Rep.: PR1 – 1966 – 9; 2002 – 1 Ecuador: Maj1 – 1895 – 6 // 1928 – 1 // PR2 – 1946 – 7 // 1978 – 8 El Salvador: Maj1 – 1886 – 1 // PR 1 – 1963 – 8 // 1984 – 6 Estonia: PR 3 – 1919 – 6 // PR1 – 1989 – 1; PR3 – 1992 – 4 Finland: Ind – 1863 – ?; PR3 – 1906 – 17 // 1945 – 14 France: Ind – 1789 – l; 1791 – 2; 1795 – 4; 1815 – 2; Maj1 – 1817 – 3; Maj3 – 1820 – 1; 1824 – 3; Maj1 – 1848 – 2; 1871 – 2; Maj3 – 1875 – 3; Maj1 – 1885 – 1; Maj3 – 1889 – 7 // PR1 – 1919 – 2; Maj3 – 1927 – 3 // PR1 – 1945 – 3; 1951 – 2; Maj3 – 1958 – 7; PR1 – 1986 – 1; Maj3 – 1988 – 4 Georgia: Mix – 1995 – 2 Germany: Maj3 – 1871 – 13; PR2 – 1919 – 1; 1920 – 8 // PR 3 – 1949 – 1; 1953 – 1; 1956 – 8; 1985 – 5 Ghana: Maj3 – 1956 – 2 // 1969 – 1 // 1979 – 4 Greece: Ind – 1822 – 3 // 1843 – 1; Maj1 – 1844 – 8; 1864 – 21 // 1920 – 2; PR1 – 1926 – 1; Maj3 – 1928 – 1; PR1 – 1932 – 1; Maj1 – 1933 – 1; PR1 – 1936 – 1 // 1946 – 3; Maj1 – 1952 – 1; Mix – 1956 – 1; PR 1 – 1958 – 1; 1961 – 2 // 1974 – 4; 1989 – 3; 1990 – 3 Guatemala: PR1 – l944 – 3 // 1984 – 4 Honduras: Maj1 – 1879 – 1 // 1954 – 1; PR1 – 1966 – 6 Hungary: Mix – 1922 – ?; 1990 – 4 India: Maj3 – 1950 – 2; 1961 – 3 // 1977 – 8 Indonesia: PR 3 – 1953 – 2 // PR2 – 1999 – 1

Strategy and History of Electoral System Choice

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

75

Ireland: PR1 – 1922 – 25 Israel: PR2 – 1949 – 1; 1951 – 6; 1973 – 5; 1992 – 3 Italy: Maj3 – 1879 – 4; Maj1 – 1882 – 3; Maj3 – 1892 – 7; PR 3 – 1919 – 3; Mix – 1923 – 1 // PR3 – 1946 – 2; Mix – 1953 – 1; PR3 – 1956 – 9; Mix – 1993 – 3 Jamaica: Maj3 – 1962 – 10 Japan: Maj2 – 1946 – 1; 1947 – 19; Mix – 1994 – 2 Kenya: Maj3 – 1963 – 10 Korea, South: Maj3 – 1950 – 1 // Mix – 1962 – 3 // 1987 – 4 Latvia: PR3 – 1922 – 4 // 1992 – 4 Lebanon: Maj1 – 1943 – 3; 1953 – 6 // 1992 – 3 Lithuania: Maj3 – 1920 – 1; PR2 – 1922 – 3 // Maj3 – 1990 – 1; Mix – 1992 – 2; 2000 – 1 Macedonia: Mix – 1998 – 2 Madagascar: PR1 – 1992 – 1; Mix – 1998 – 2 Malawi: Maj3 – 1994 – 2 Malaysia: Maj3 – 1957 – 3 // 1974 – 7 Mali: Maj1 – 1992 – 4 Mexico: Ind – 1824 – 5 // 1843 – 3 // 1857 – 2 // 1867 – 4 // Maj3 – 1917 – 17; Mix – 1963 – 5; 1977 – 6; 1996 – 3 Moldova: Maj3 – 1989 – 1 // PR2 – 1993 – 3 Mongolia: Maj3 – 1989 – 1; Maj1 – 1992 – 1; Maj3 – 1996 – 2 Mozambique: PR2 – 1993 – 2 Namibia: PR 2 – 1990 – 2 Nepal: Maj3 – 1959 – 1 // 1981 – 5 Netherlands: Ind – 1815 – ?; Maj1 – 1848 – ?; Maj3 – 1887 – 9; PR3 – 1918 – 5; 1933 – 1 // 1946 – 17 New Zealand: Maj3 – 1890 – 6; 1908 – 28; PR3 – 1993 – 3 Nicaragua: PR 1 – 1984 – 4 Nigeria: Maj3 – 1959 – 2 // 1979 – 2 // 1992 – 1 // 1999 – 1 // Norway: Ind – 1815 – 24; 1884 – 7; Maj3 – 1906 – 5; PR1 – 1921 – 6 // 1945 – 2; 1953 – 8; PR3 – 1985 – 5 Pakistan: Maj3 – 1970 – 2 // 1988 – 4 // 2002 – 1 Panama: Ind – 1916 – 2 // PR 1 – 1946 – 8 // Mix – 1983 – 1 // 1989 – 1 // 1994 – 2 Papua New Guinea: Maj3 – 1975 – 6 Paraguay: PR 1 – 1993 – 2 // Peru: Ind – 1828 – 1 // Maj2 – 1861 – 1 // 1868 – 4; PR1 – 1933 – 2 // 1950 – 3 // 1963 – 1 // 1979 – 3 // PR 3 – 1995 – 1 // 2002 – 2 Philippines: Maj1 – 1935 – 2; Maj3 – 1953 – 5 // 1987 – 3; Mix – 1995 – 2 Poland: PR 1 – 1918 – 3 // 1928 – 2 // PR 3 – 1991 – 1; 1993 – 2; PR2 – 2001 – 1 Portugal: Ind – 1820 – 1; Maj1 – 1822 – 1; Ind – 1826 – 1 // 1834 – 3; Maj1 – 1852 – 2; Maj3 – 1859 – ?; 1878 – 2; 1884 – 2; Maj1 – 1895 – ?; Mix – 1896 – ?; Maj2 – 1901 – ?; 1911 – 1; PR1 – 1915 – ? // PR 2 – 1975 – 11 Romania: PR 1 – 1990 – 4 Russia: Ind – 1917 – 1; Maj3 – 1990 – 1; Mix – 1993 – 3 Senegal: Mix – 1992 – 2 Slovakia: PR3 – 1992 – 4 Slovenia: PR3 – 1990 – 4 South Africa: PR 3 – 1994 – 2 Spain: Ind – 1810 – 1; 1812 – 1 // 1820 – 3 // 1834 – 1; Maj1 – 1836 – 1; Ind – 1836 – 1; Maj1 – 1837 – 6; Maj3 – 1846 – 4; Maj1 – 1854 – 1; Maj3 – 1857 – 4; Maj2 – 1865 – 2; 1868 – 6; Maj3 – 1878 – 2 // Maj2 – 1931 – 3 // PR1 – 1977 – 8 Sri Lanka: Maj3 – 1946 –7 // 1977 – 1; PR3 – 1978 – 3 Sweden: Ind – 1810 – ?; 1866 – 15; PR1 – 1909 – 13; 1952 – 6; PR3 – 1970 – 11 Switzerland: Ind – 1848 – 8; 1872 – 10; 1900 – 6; PR1 – 1918 – 22 Taiwan: Mix – 1991 – 4

76

Josep M. Colomer

Summary Table 1A • • • • • • • • •

(Continued)

Thailand: Maj1 – 1978 – 8; Mix – 1997 – 1 Trinidad-Tobago: Maj3 – 1962 – 10 Turkey: Maj1 – 1946 – 4 // PR1 – 1961 – 1; 1965 – 1; 1969 – 1; 1973 – 2 // 1982 – 2; 1991 – 1; 1995 – 3 Ukraine: Maj3 – 1990 – 1; Mix – 1994 – 3 United Kingdom: Maj1 – 1832 – 9; 1868 – 2; Maj3 – 1885 – 15; 1949–15 United States: Maj1 – 1789 – 26; Maj3 – 1842 – 63; 1968 – 18 Uruguay: PR 1 – 1918 – 6 // PR3 – 1934 – 10 // 1984 – 4 Venezuela: PR1 – 1946 – 1 // 1958 – 7; PR3 – 1989 – 2; PR1 – 1999 – 1 Yugoslavia: PR 1 – 1992 – 3

Notes This table summarizes the detailed five world regional tables included at the end of the opening chapters to each of the parts in this book. – –

–

Information for each country: type of electoral system – starting year – number of elections. Type of electoral system: Ind.: indirect elections; Maj1: multi-member districts, bloc ballot, majority rules; Maj2: multi-member districts, limited or cumulative ballots, majority rules; Maj 3: single-member districts, majority rules; Mix: mixed system; PR1: average district magnitude lower than nine, proportional representation; PR2: average district magnitude higher than eight, closed lists, proportional representation; PR 3: average district magnitude higher than eight, open lists, open ballot or double vote, proportional representation. Key: The sign // indicates a period without elections or with authoritarian fake elections. Present democratic periods are counted since the latest //.

Source: See Appendix: Notes and Sources for Summary Tables at the end of the book.

Strategy and History of Electoral System Choice Summary Table 1B • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

77

Presidential electoral systems

Argentina: Col – 1853 – 12 // 1931 – 2 // 1946 – 1; Plu – 1949 – 1 // Col – 1955 – 2 // Maj – 1972 – 2 // Col – 1983 – 2; QP – 1994 – 2 Armenia: Maj – 1991 – 3 Austria: Maj – 1945 – 10 Bangladesh: Plu – 1977 Belarus: Maj – 1996 – 2 Benin: Maj – 1991 – 2 Bolivia: Maj – 1899 – 4 // 1921 – 3 // 1940 – 1 // 1947 – 1 // 1951 – 1; Plu – 1956 – 3 // Maj – 1967 – 6 Bosnia-Hercegovina: Plu – 1998 – 2 Brazil: Col – 1834 – 2; Maj – 1892 – 12 // Plu – 1945 – 4 // Maj – 1986 – 4 Bulgaria: Maj – 1991 – 3 Cape Verde: Maj – 1992 – 2 Central African Rep.: Maj – 1993 – 2 Chile: Maj – 1925 – 11 // Maj – 1989 – 3 Colombia: Col – 1821 – l // Col – 1832 – 7; Plu – 1853 – 2 // Col – 1863 – 11 // Col – 1886 – 2 // 1904 – 1; Plu – 1910 – 10 // 1958 – 9; Maj – 1991 – 3 Costa Rica: Col – 1889 – 4 // 1909 – 1; Maj – 1913 – 2 // 1919 – 4; QP – 1936 – 4 // 1953 – 13 Croatia: Maj – 1990 – 3 Cuba: Col – 1901 – 5 // Col – 1940 – 3 // Dominican Republic: Plu – 1962 – 1 // 1966 – 8; Maj – 1995 – 2 Ecuador: Maj – 1895 – 7 // 4 // Maj – 1978 – 5; QP – 1997 – 2 El Salvador: Maj – 1886 – 8 // Maj – 1963 – 3 // Maj – 1983 – 4 Estonia: Maj – 1992 – 1 Finland: Col – 1919 – 10; Maj – 1988 – 3 France: Maj – 1848 – 1 // Maj – 1962 – 7 Georgia: Maj – 1995 – 2 Germany: Maj – 1918 – 2 // Ghana: Maj – 1979 – 1 // 1992 – 3 Guatemala: Maj – 1944 – 2 // Maj – 1984 – 4 Honduras: Maj – 1879 – 3 // 1954 – 3 // Plu – 1966 – 5 Ireland: STV – 1937 – 9 Israel: Maj – 1996 – 3 Kenya: QP – 1963 – 1 // 1992 – 3 Korea, South: Plu – 1950 – 3 // 1963 – 3 // 1987 – 4 Lithuania: Maj – 1992 – 3 Macedonia: Maj – 1994 – 2 Madagascar: Maj – 1992 – 3 Malawi: Plu – 1994 – 2 Mali: Maj – 1992 – 3 Mexico: Col – 1824 – 3 // 1847 – 2 // 1857 – 5; Col – 1874 – 10 // Plu – 1917 – 17 Moldova: Maj – 1991 – 2 Mongolia: Maj – 1993 – 3 Mozambique: Maj – 1994 – 2 Namibia: Maj – 1990 – 2 Nicaragua: Maj – 1911 – 2 // Plu – 1984 – 2; QP – 1995 – 1; QP – 1999 – 1 Nigeria: QP – 1979 – 2//QP – 1993 – 1 // 1999 – 1 Panama: Plu – 1916 – 4 // 1952 – 5 // 1984 – 1 // 1989 – 1 // 1994 – 2 Peru: Col – 1828 – 1 // Col – 1861 – 1 // 1866 – 1 // 1868 – 3 // Plu – 1896 – 5 // 1915 – 1 // QP – 1931 – 1; QP – 1933 – 2 // 1956 – 1 // 1962 – 1 // 1963 – 1 // Maj – 1978 – 7 Philippines: Plu – 1935 – 6 // 1987 – 2

78

Josep M. Colomer

Summary Table 1B • • • • • • • • • • •

(Continued)

Poland: Maj – 1990 – 3 Portugal: Maj – 1976 – 6 Romania: Maj – 1990 – 4 Russia: Maj – 1993 – 3 Slovenia: Plu – 1991 – 3 Sri Lanka: SV – 1978 – 4 Taiwan: Plu – 1996 – 2 Ukraine: Maj – 1994 – 2 United States: Col – 1789 – 54 Uruguay: Plu – 1899 – 12 // 1942 – 5 // 1984 – 3; Maj – 1996 – 1 Venezuela: Col – 1830 – 2 // 1847 – 3 // Col – 1874 – 1 // Plu – 1946 – 1 // 1958 – 10

Notes This table summarizes the detailed five world regional tables included at the end of the opening chapters to each of the parts in this book. – – –

Information for each country: type of electoral system – starting year – number of elections. Type of electoral system: Col: electoral college: Plu: Plurality rule; QP: qualified plurality rules; Maj: majority runoff; STV: single-transferable vote; SV: supplementary vote. Key: The sign // indicates a period without elections or with authoritarian fake elections. Present democratic periods are counted since the latest //.

Source: See Appendix: Notes and Sources for Summary Tables at the end of the book.

Part 2 The Americas

This page intentionally left blank

2 The Americas: General Overview Josep M. Colomer

The most distinguishing characteristic of the electoral and political systems traditionally and currently existing in the Americas is the separation of powers between the Congress and the Presidency. Nowadays, the only exceptions can be found in Canada and a number of small islands in the Caribbean which maintain the typically British-legated formula of parliamentary regime with a single parliamentary election in single-member districts by plurality rule. In about twenty countries in North, Central and South America the Congress and the Presidency are elected separately under two different sets of rules and procedures. As we will see in the following overview, in most countries the rules for the election of the lower or single chamber of Congress have typically been reformed over time from initially restrictive formulas based on small districts and plurality rule to more permissive devices which allow the inclusion of some minorities, up to the present dominance of proportional representation. Regarding the Presidency, the early diffusion of the United States formula of an electoral college with a second round in Congress gave way, in most countries, to direct elections which were held early on by plurality rule and, more recently, by qualifiedplurality or majority rules with a second-round runoff, whether directly by voters or by Congress.

The Congress The United States constitution of 1789 created a Congress with two chambers. For the upper chamber or Senate, two seats were allocated to each state, regardless of its population, while for the lower chamber or House of Representatives – which will be the focus of the present survey – seats were allocated to the states in proportion to the population. As discussed in Chapter 1 the basic mathematical formulas of proportional representation were invented at that time by American constitution-makers to allocate House seats to the states on the basis of population. Each state, however, was allowed to choose its own electoral system to elect its House members. Initially, most states chose multi-member districts with plurality rule to elect representatives to the federal House, thus producing a high number 81

82

Josep M. Colomer

of single-party state systems. This choice was in congruence with the rules used for their own local and state institutions. Actually, in the early twenty-first century, a majority of local officials in the United States are still elected in multi-member districts by plurality rule, while a majority of the seats on state legislatures were elected from such districts until the 1970s. Immediately after the approval of the 1789 constitution, a majority of 8 out of the 13 states organized their elections to the House of Representatives at large, that is in a single statewide multi-member district. The most common ballot procedure for both local, state and federal elections was bloc ballot, also called ‘general ticket’, in which each voter had to choose a number of candidates equal to the number of seats in the district and the winners were those obtaining the highest numbers of votes. As coordination among candidates developed and political parties began to be formed, a common consequence of this system was the exclusion from representation of all political groups but one in the state, however great the other groups could be or whatever the number of potentially smaller districts in which they might have obtained the largest popular support. Another consequence was that a few large states’ representatives obtained much decision-making power in the House thanks to the system of single-party state representation. The other five founding states, as well as most of the new states joining the Union in the further process, adopted, in contrast, systems based on single-member districts and plurality rule. However, during the first fifty years about one-third of the House Representatives came from states using at-large districts. Although the available evidence on institutional strategies at state level in this period is limited, it seems that a single dominant party tended to impose a single district at large with bloc ballot, while more pluralistic initial settings favoured the division of the state into smaller single-member districts able to produce different winners in each. The first attempt at general reform of the electoral rules for the United States House of Representatives was launched by the Whigs, one of the largest parties during the early nineteenth century, when they began to experience a significant electoral decay, mostly to the advantage of the Democrats – as analysed in the corresponding chapter by Erik Engstrom. Fearing that they could be eliminated from representation in some states by the at-large, bloc ballot electoral system, the Whigs promoted the adoption by Congress of homogeneous rules based on single-member districts for electing the members of the House across all United States territory. This formula, although still highly restrictive, would have allowed some state minorities to obtain federal representation on the basis of their larger support in some small districts. The corresponding provision was approved by Congress in 1842, but it was not actually enforced and was in fact dropped after the 1850 census. On the occasion of approving each of the following six decennial censuses, the Congress mandated once again that the elections to the House were to be organized in single-member districts with contiguity, equal number of inhabitants and compact territory, but many states continued to elect their representatives at large. With the support of the Supreme Court, the states have been allowed again to choose their own rules for the House elections since 1929. In total, at least 16 states used

The Americas: General Overview

83

at-large districts and the bloc ballot during some periods between the 1840s and the 1960s. The second, more successful, reform was triggered by certain reactions to the Voting Rights Act of 1965 enfranchising African-Americans in Southern states. Many of these states began implementing a single district at large with bloc ballot for their own state legislative elections, or developed new forms of gerrymandering, that is the art of designing irregular single-member districts with well-chosen territorial limits, in an effort to offset the African-Americans’ vote. Most incumbent federal Representatives, fearing that they would be threatened by this kind of state measure as well as perhaps by any corresponding Supreme Court reaction imposing some other common pattern for electoral rules, approved again a mandate for single-member districts in 1967. There were still some bloc ballot elections to the House in 1968, as well as significant pressure from some Southern states to reintroduce multi-member districts. But single-member districts have been enforced in all the United States without exception for elections to the House of Representatives since 1970. In comparison with large multi-member districts, the adoption of single-member districts as the nationwide procedure reduced the number of single-party state systems and increased the diversity of state representatives in the House. However, as discussed extensively in the introduction and several chapters of this book, plurality rule in single-member districts still tends to induce the formation of a low number of candidacies and distorts the representation of citizens’ preferences. Specifically, in 31 per cent of the elections (27 out of 88) to the US House of Representatives since 1828, a party has received a majority of seats in spite of having obtained only a minority of popular votes. In ten of those elections the majority party in seats had even obtained fewer popular votes than some other party that was reduced by the electoral system to a minority of seats. It is highly probable that the institutional devices analysed here are behind the frequently expressed fears that the United States House of Representatives will tend to produce unpredictable and unstable legislative and policy decisions. From the debates during the constituent process in the late eighteenth century to some of the most recent academic literature on the US political system, repeated claims have been made that the House tends to be dominated by impulses of rage, resentment, jealousy, avarice and other irregular and violent tendencies. Examples range from Hamilton and Madison’s ‘propensity [attributed to] all single and numerous assemblies to yield to the impulse of sudden and violent passions’ (Hamilton et al., 1788: ‘No. 62’) to the modern media’s obsession with congressmen corruption, and to certain textbooks on the American political system praising ‘gridlock’ or lack of congressional decisions for fear that they could foster permanent instability. It would be difficult, however, to attribute such vicious impulses only to human nature, as it would be hard to understand why a populous assembly should be more unreliable than a single individual invested with strong powers (as was the figure of the president emerging from that discussion), especially because many of the critics here alluded to have been especially diffident of ‘human ambition’. Our perspective suggests that so persistent a diffidence to the United States Congress

84

Josep M. Colomer

has been supported, at least to some extent, by the unfair and unbalanced procedures by which it is elected. If, as a result of the way they are elected, congressmen can form legislative majorities and make decisions not corresponding to the preferences supported by a majority of citizens and even contradicting those preferences in a blatant manner, it can be reasonable to favour a reduction of such an institution’s effective decision-making capability. However, while institutionally-induced legislative paralysis might be a medicine against biased representation, it is not a formula for good government. In this light, it becomes thus particularly interesting to examine alternative electoral rules for the election of Congress, such as those that have been adopted in other parts of the continent and are also being considered and experimented with in the United States – as closely examined in the corresponding chapter by Richard L. Engstrom. The choice of electoral rules for congressional elections in the many different countries of Latin America has implied, of course, more variation than in the United States. From a global perspective, we can find a pattern by which they have moved over time from old procedures of lots and indirect elections imported from Europe, through some intermediate stages closer to the experience of the US, up to the general adoption of proportional representation in more recent times. The very first electoral experiences in the Spanish and Portuguese colonies in the early nineteenth century leading to their independence were organized on the basis of metropolitan regulations. These included combinations of lots for nominating or selecting candidates and voting by plurality rule, as in the 1809 elections to the ‘Central Junta’ of Spain which was in charge of organizing the resistance against the Napoleonic troops and the subsequent 1810 elections to the Spanish ‘Cortes’ (assembly). Elections with these rules were held in more than one hundred cities on the American continent, including Buenos Aires and other provinces in the Vice-royalty of Río de la Plata (which became Argentina) and New Spain (which became Mexico). The same kind of rules were also used in Chile in 1822. At a second stage, indirect elections were called to the new ‘Cortes’ created by the 1812 Spanish constitution approved in Cádiz. In further periods the formula of indirect elections was also introduced according to the United States electoral college model, as well as the French experience from the medieval Estates General to the republican revolution and the restoration of the monarchy, implying between two and up to five levels of representation. Indirect elections were adopted in Latin America mostly, with the aim of permitting the socially dominant political elites to keep control of the electoral process and outcomes. This intention was revealed, for instance, by Argentine constitution-maker and politician Juan B. Alberdi who defended ‘the system of double and triple election’ as a better means of ‘purifying universal suffrage’ of the erratic impulses of the masses without reducing or suppressing it (Alberdi, 1852), while Colombian constitutionalist and politician Justo Arosemena expected that, in the absence of sufficient popular instruction, good citizens’ choices could be facilitated by ‘an intermediate elector, more apt than the voter and better known than the final candidates’ (Arosemena, 1878). Indirect elections were organized either with plurality rule, whether in multimember districts with bloc ballot as in Argentina in 1819–26, or in single-member

The Americas: General Overview

85

districts as in Mexico since 1857, or with majority rule with a second round by plurality as in Colombia in 1821 and Peru in 1823. In Brazil, which initially also adopted the electoral regulations established in the 1812 Spanish constitution, all these formulas as well as limited ballot were combined with indirect elections at different periods between 1822 and 1875. However, indirect elections – including those to presidential electoral colleges, to be discussed below – often produced the effect of a highly unpredictable electoral outcome that was unintended by the constitution-makers. Even if absolute majority rule is adopted an election in two stages can give victory to ‘a majority of the majority, who may be, and often are, but a minority of the whole’, as remarked by John Stuart Mill in a quote given in the general introduction to this book. In the words of a Latin American constitutionalist of about the same time, ‘there can be deputies hardly achieving to represent one sixth or one eighth or one ninth of the citizens of his constituency’ (Santisteban, 1874: 224). The point is that actual minority winners – including those produced by elections in several stages – are more varied and unpredictable than majority winners can be in a given electorate. If, in addition, as was the case in congressional elections in most Latin American countries during the nineteenth century, intermediate electors were not required to be more qualified than the voters, were lesser known than the candidates or were not submitted to imperative mandate, actually they could form shifting coalitions and make decisions unrelated to the voters’ preferences revealed at the first stage of voting. In fact frequent claims were raised against intermediate electors’ manoeuvrings and conspiracies. As remarked, for instance, by a Mexican liberal deputy, ‘intrigue, bribery and coercion are very easy in electoral colleges, as formed by a very small number of persons’, in contrast to the difficulty of seducing or corrupting thousands or millions of voters in direct elections (in Calero, 1989), while in Peru electoral colleges were considered ‘centers of intrigue betraying popular will’ (in Chiaromonte, 1995). Thus, at a third stage, direct congressional elections according to the US model were established in Latin American countries. This occurred for first time in Chile in 1833, but rather later in the remaining countries such as Argentina in 1853, Colombia in 1857, Brazil in 1881 and Mexico in 1911. As in the United States, most Latin American countries initially used multi-member districts with bloc ballot which, as discussed, tend to create single-party local systems. They later moved to smaller single-member districts permitting slightly more varied congressional representation, as was the case in Brazil in 1881, Colombia in 1886, Argentina in 1902 and Mexico in 1911. As had happened with indirect elections, Brazil was the country with the highest number of experiments, including not only the two procedures just mentioned, but also some other more pluralistic formulas, such as limited ballot giving typically two votes per voter to elect three seats in several periods between 1882 and 1930 (a procedure also adopted in Argentina in 1912, Ecuador in 1929, Peru in 1931 and Bolivia in 1937) and cumulative ballot in 1904 (following the experience in Chile and Peru after 1874). The latest stage has been the adoption of proportional representation. Early introductions (for direct elections) of this type of congressional rules permitting multiple

86

Josep M. Colomer

parties to exist, compete successfully in elections, receive seats and survive include Costa Rica since 1913, Uruguay since 1918, Chile since 1925, Colombia in 1931, Cuba in 1940, Brazil in 1945, Ecuador, Guatemala and Venezuela since 1946, Bolivia since 1952, Honduras since 1966, Argentina and El Salvador since 1963 and the Dominican Republic since 1966. However, as had happened before with many Latin American political regimes holding elections with electoral systems based on indirect elections or on direct elections by plurality rule, most of these experiences of electoral democracy also failed and were interrupted by violent conflicts and dictatorships, especially in the early and mid-1970s. It was during the more recent processes of democratization or re-democratization in virtually all Latin American countries since the early 1980s that proportional representation was re-established or adopted as the standard principle for electing the members of the lower or single chamber of Congress. As will be illustrated in several chapters in the following pages, during the most recent period the choice of electoral systems can be understood as a consequence of expectations, calculations and inventions by previously existing political leaders, organizations and parties. Typically, the establishment or the maintenance of plurality rule in single-member districts was the result of decisions made within dominant single-party or balanced two-party systems, that is in situations able to feed established leaders’ expectations of maintaining control of the political competition. In contrast, proportional representation systems, as well as other inclusive electoral rules such as qualified plurality or majority-runoff rules for presidential elections to be discussed below, were typically adopted in pluralistic constituent conventions or by political agreements among multiple political parties with expectations of competing in more uncertain elections or of sharing power in multiparty congresses and cabinets. The most durable experiences of proportional representation in congressional elections in Latin America are those started in Costa Rica in 1953 and in Colombia and Venezuela in 1958. More recently, they include those initiated with multi-party congressional elections in Bolivia and Ecuador in 1979, Honduras in 1981, Argentina in 1983, Uruguay in 1984, El Salvador in 1985, Brazil and Guatemala in 1986, Chile in 1989 and Nicaragua in 1990, while Mexico has used several mixed systems of plurality rule and proportional representation since 1977, as explained by Alberto Díaz-Cayeros and Beatriz Magaloni in their chapter below.

The President The first popular-based election of a president of a republic was organized in the United States according to the electoral college formula included in the 1789 constitution. The initial choice of this procedure – which, as we will see, was later diffused across the continent – required a complex process of bargaining and invention by the US constituents. During the Convention gathered in Philadelphia in the summer of 1787, three basic proposals were presented. First, there was the proposal of parliamentary regime, by which the chief executive would be elected by Congress, as included in

The Americas: General Overview

87

the initial federalist plan elaborated by delegates from Virginia. This proposal was congruent with the way most states in the then existing Confederation were organized, since in 8 out of 13 states the governor was elected by the legislature. However, as most of those state legislatures were elected with highly distorting electoral systems (typically including multi-member districts, bloc ballot and plurality rule, as reviewed above), they were highly vulnerable to criticisms of inefficiency and demagoguery. In the absence of proposals for a better electoral system for the new Congress, the federal parliamentary proposal was open to similar criticisms, which gave leverage to the defenders of the separation of powers between the legislative and the executive. In its aim to create a strong new central government, the parliamentary proposal was also resisted by the defenders of the founding states’ powers. The second proposal, which can be mainly associated with republicans from Pennsylvania, was direct election of the president by the people in order to establish a neat separation of powers from Congress. This proposal was particularly disadvantageous to the small states, whose influence in the election would have been minor, as well as to the South, where voting rights were allocated to significantly lower proportions of the population than in the North. Finally, the small states proposed that the executive be appointed by the state governments. This proposal was consistent with the way the Continental Congress was elected at the time under the Confederation rules. It was also in line with the previous constituent decision to form the Senate with equal numbers of seats per state, regardless of population, to be filled by the state legislatures. The small states’ proposal for the presidential election was supported directly by a few participants in the Convention, but they were determined in their position because their political existence depended in some sense on the consolidation of the states’ powers. It can also be assumed that without the support of a few even if small states, there would not have been a constitution at all. Its is possible that the constituents’ preferences might have produced a cycle in a hypothetical process of decisions open to all imaginable comparisons, by which the proposal of a parliamentary regime might have been preferred to direct election by a majority of federalists and small states’ delegates, direct election would have been preferred to the appointment of the president by the states by both republicans and federalists in favour of creating a strong federal government, and perhaps the appointment of the president by the states would have been preferred to parliamentarism by both the small states and the republicans in favour of a separation of powers. Even if it is not possible to identify clearly this cycle in retrospect, the fact is that the Convention voted in favour of different formulas during the process of elaborating, discussing, revising and approving proposals in a series of committees. The parliamentary proposal was approved initially and on two more occasions, although by introducing different procedures for choosing the president involving joint or separate sessions of the two chambers of Congress. Opposition to parliamentarism, however, developed strongly, especially using the type of arguments previously referred to regarding the electoral rules for the US Congress and indirect elections, specifically the fear that the choice of president by a populous assembly would be

88

Josep M. Colomer

the occasion of ‘intrigue’, ‘cabal and corruption’. This didn’t mean in practice much more than the formation of congressional multi-party majority coalitions, which is typical of parliamentary regimes, but this experience was alien to the American constituents of the time. The crucial move was that the Republicans in favour of separate powers lowered their support for their own formula of direct presidential election – aided also by fears over of people’s ignorance and manipulability – while giving priority to preventing the alternative of establishing a parliamentary regime. Perhaps led by the manoeuvring skills of delegate Gouverneur Morris from Pennsylvania, the republicans in favour of separation of powers formed a coalition with the small states’ delegates. During the last few days of the Convention, the matter was referred to a committee from which a new invention emerged, able to reunite all those opposed to parliamentarism. The new formula of indirect election was based on an electoral college. As has been shown in the general introduction of this book, precedents of indirect elections of electors can be found in Greek assemblies, the Roman comitia, Mediterranean cities and, for the election of a single high magistrate, in the college of cardinals electing the pope and the colleges of princes and bishops electing a number of central European kings and the German emperor. In the context of making the United States constitution, the formula of electoral college implied that the election of the president would be separated from Congress, while its specific procedure would likely give the small states a high influence in the decision. According to the rules which were established in the United States constitution, the electoral college is formed by electors chosen in each state in a number equal to the sum of federal representatives and senators from the state – so giving the small states some overrepresentation due to the malapportionment of Senate seats. The procedures to select the electors are decided by the states themselves – which initially allocated the choice, in most cases, to the state legislatures, but also permitted direct elections. Actually the college never meets. During the constitutional negotiations, it was agreed that the electors of each state would gather in the state’s capital to cast their votes, which, at that time, seemed a very contingent concession to peripheral states fearful of the high costs of travelling to Washington. This decision also eliminated the perils that a new version of parliamentary ‘intrigue’ would emerge in the college, but, at the same time, reduced the expectation that electors would use their presumed higher capabilities of ‘insight and reflection’ than voters, as had been argued by some republicans. In case no candidate obtained a majority support of electors in the college, the choice of president would be transferred to the Congress among the five (later three) candidates with most votes in the college. However, the congressional decision would be made on the basis of an equal number of votes per state, both in the preliminary version that allocated it to the Senate and in the finally approved version transferring it to the House but by the rule of one vote per state (to be decided by majority of each state’s representatives). In short, the procedure, first, confirmed that the presidential election would be separated from congressional elections and, second, gave overrepresentation to the small states, thus

The Americas: General Overview

89

making a majority coalition of constituent delegates opposed to parliamentarism possible. Apparently, the widest expectation at the time was that there would be a proliferation of states’ favourite sons as presidential candidates and, as a consequence, usually no majority would be formed in the college. In order to reduce the territorial dispersion of candidates, the initial regulations established that each elector in the college should vote for two candidates, one of them not an inhabitant of the same state, from which the president and the vice-president would be selected. However, the confusion between candidates for the two offices led to a modification of this rule in 1800. Basically, it was expected that the college would nominate a selection of candidates from which the House would choose one by forming a coalition of a majority of states. Actually, the first five presidents of the United States were nominated as candidates by congressional caucuses, implying thus some degree of ‘intrigue’. The sixth, lacking a majority in the college, was appointed by the House of Representatives in spite of having been second in both popular votes and among college electors, in 1824. Political leaders from large states soon realized that it was to their advantage to unite around presidential candidates able to obtain broad support and win in the college rather than to have to negotiate with the overrepresented small states in the House. The best way to reinforce the influence of the large states’ electors in the election of the president was to choose an appropriate electoral system of electors. From 1828 on, most large states introduced direct elections of electors to college by at-large districts, bloc ballot and plurality rule, that is the electoral system known as ‘general ticket’ that was also widely used at the time for the election of the House of Representatives as well as for local and statewide institutions, as has been discussed above. Most of the smaller states could not but adopt the same formula in order to reduce the corresponding larger states’ advantage. The consequence of these further electoral reforms at state level was that the college became a rubber stamp of the outcome of direct elections held under highly distorting rules. The procedure of at-large districts, bloc ballot and plurality rule, together with malapportionment of electors gave rise to a number of unintended consequences. In spite of the requirement for an absolute majority in the college, the president of the United States can be elected with the support of only simple pluralities in a few large states. In fact, many presidents have been elected with a minority support in popular votes since 1828 (17 out of 44, or 39 per cent) – including Abraham Lincoln, the president that triggered the Civil War after having been elected by the college in 1860 with the lowest proportion of popular votes ever. In at least three cases, the winning president in the college had even obtained fewer popular votes than some other candidate – including George W. Bush in 2000. The procedures to elect the president in most of the many countries of Latin America have also been more varied than those that have been used in the United States, as happened with congressional elections reviewed above. First, it should be remarked that parliamentary formulas by which the Congress elected the executive president, initially inspired by the 1812 Spanish constitution, were established in

90

Josep M. Colomer

Argentina in 1819, Chile in 1822 and again during the period from 1891 to 1925, Peru in 1823, Ecuador in 1830 and 1845, and Uruguay in 1839 and for a collegial executive from 1952 to 1966. Second, the US-inspired formula of electoral college was later taken up widely, including the regulations established in Venezuela in 1819, 1830 and 1874, Colombia in 1821, 1832, 1863 and 1892, Mexico in 1824 and 1857, Central America in 1824 and the countries that separated from the confederation thereafter, Costa Rica, Guatemala, Honduras and Nicaragua, Argentina in 1826, 1853 and 1983, Bolivia in 1828, Peru in 1828, 1860, 1874 and 1890, Chile in 1828, 1833 and 1874, Brazil in 1834 (for the Regent), the Dominican Republic in 1844 and Cuba in 1902 and 1940. Typically, however, the Latin American electoral assemblies (or juntas as they were usually called rather than ‘colleges’) introduced some innovations into the procedures. On the one hand, certain electoral assemblies in large, territorially diverse countries were organized on the basis of an equal number of votes (either one or two) per state or province, as in Venezuela in 1819, 1830 and 1874, Mexico in 1824 and 1857 and Colombia in 1863. This somehow reflected the strength of the territorial elites, which – as discussed above – had already been highly relevant in the constitutional choice in the United States, as well as the territorially inclusive role that the electoral formula pretended to play. With formulas like these, the electoral college formula for presidential elections even increased the disadvantages of indirect elections already identified for congressional elections – basically, the unpredictability of electoral decisions which were hardly related to voters’ preferences. On the other hand, in almost all countries the provision of a second-round runoff by Congress, usually between the two candidates with the most votes, was regulated on the basis of its existing composition – in contrast to the US formula of giving one vote to each state’s representatives. This probably reflected the tradition of parliamentary formulas and implying the will of a closer collaboration between the Congress and the Presidency. In fact, and in contrast to the United States, a significant number of presidents in Latin America were elected at the second round by Congress after not obtaining a majority in the college, including four presidents in Colombia in 1837, 1841, 1845 and 1849 (plus another two in 1909 and 1910 following the resignation of the incumbent), three in Bolivia after 1828, one in Mexico in 1874 and one in Venezuela in 1877. A consistent alternative to majority rule in the electoral college with a second round in Congress, in order to avoid the disadvantages of indirect elections, was direct elections by majority rule with a second round in Congress. This system creates incentives for a multi-party majority to be formed both in support of the president and for legislating in Congress, in this way making cooperation between the two institutions, governance and broad social support for policy decisions easier to achieve. Precedents include Bolivia in regulations introduced in 1871, in 1967 and again since 1979, Honduras since 1879, El Salvador since 1886 and again in 1963, Brazil in 1892, Nicaragua in 1911, Costa Rica from 1913 to 1932, Chile from 1925 to 1970 and Guatemala in 1944. A large number of Latin American presidents have been elected at the second round by Congress after not obtaining a majority of the popular vote in direct elections, including three in El

The Americas: General Overview

91

Salvador in 1903, 1907 and 1911, three in Costa Rica in 1914, 1924 and 1932, four in Chile in 1946, 1952, 1958 and 1970, and six in a row in Bolivia since 1982. However, most Latin American countries, guided in many cases by re-elected presidents supported by single-party dominance (as well as by some authoritarian rulers), eventually suppressed completely the role of Congress and established direct presidential elections by plurality rule. These were introduced in Colombia in 1853, 1910 and 1958, Uruguay in 1918 and 1966, Brazil in 1945, Dominican Republic in 1962 and Nicaragua in 1984. All these countries, nevertheless, moved to more inclusive formulas later on. Only five Latin American countries still keep plurality rule for presidential elections in the early twenty-first century: Mexico since 1917, Venezuela since 1958, Honduras since 1981 and Panama and Paraguay since 1989. As discussed, plurality rule tends to produce winners with a minority of the popular vote, as was actually the case in 70 per cent of the direct presidential elections (38 out of 54) in South America since 1945. On many occasions the winning president with minority support in popular votes was located at an extreme position in the ideological space and didn’t have the support of the median voter, which made the formation of an alternative political majority possible (in other words, the winner was the Condorcet-loser able to be defeated by absolute majority by some other plurality-loser candidate). A number of political crises, often leading to violent civil confrontations and military coups, were produced in Latin American countries during the twentieth century as a consequence of political and social opposition to minority, unpopular presidents elected by plurality rule. The most durable presidential democracies in Latin America during the twentieth century have been those in which – according to the typical constituents’ expectation – a single dominant party could secure an absolute majority support in popular votes for the elected, even if plurality or other rules not requiring an absolute majority to win were implemented. Cases include Costa Rica, with the national liberation party (PLN) having been in government most of the time since 1953, Colombia with the long-term dominant liberals (PLC), and Venezuela with a comparable role for the social-democrats (AD), both since 1958. Situations like these successfully supported the introduction and maintenance of plurality or less-than-majority rules for presidential elections. However, the establishment or re-establishment of democratic regimes in other countries, as well as the more recent political evolution of those just mentioned, has been characterized by more pluralistic flourishing of political parties. Multi-party constituent assemblies or multi-side political negotiations favoured the replacement of plurality rule for presidential elections with more inclusive rules, namely majority rule with a second-round runoff between the two candidates with the most votes – for similar reasons as proportional representation was introduced for congressional elections, as discussed above. This analytical framework of multi-party settings choosing majority-runoff rule for presidential elections can fit rather well the cases of Ecuador and Peru in 1978, El Salvador and Guatemala in 1984, Brazil in 1986, Colombia in 1991 (the two latter developed, respectively, by Jairo Marconi Nicolau and by Arturo Valenzuela and Miguel Ceballos in their corresponding

92

Josep M. Colomer

chapters below), Dominican Republic in 1995 and Uruguay in 1996. Interestingly, in some other cases the incumbent military rulers also introduced majority runoff in the fear that a large but minority extreme party could take over power in an open election, as was the case in Argentina in 1972 and in Chile in 1989. A variant of second-round procedures is the rule requiring some qualified-plurality short of a majority, such as 40 or 45 per cent of votes, to win the election, while a second-round runoff is held only if no candidate obtains such a support. This type of rule tends to be adopted in situations of moderate multi-partism, in which the incumbent party calculates or simply feels that it can obtain larger support than any of the opposition groups but it is less sure of its dominance than in other settings previously referred to. In relatively balanced negotiations, this proposal can also be accepted by the opposition if it considers itself able to challenge the incumbent party’s candidate to obtain such a threshold and force a second round. Apparently, the earliest experiences with this type of requirement took place in Peru, where a 25 per cent rule was established in 1931, and a one-third rule with a second round in Congress was used from 1933 to 1963. Present cases include Costa Rica with a 40 per cent rule since 1936, and Argentina with a rule requiring either 45 per cent or 40 per cent with 10 percentage points of advantage to the second candidate since 1994 – two cases that are analysed, in separate chapters, respectively by Fabrice Lehoucq and Gabriel Negretto, below – as well as Nicaragua, which first adopted 45 per cent in 1995 to be replaced by 40 per cent in 1998, and Ecuador with 40 per cent with 10 percentage points advantage since 1997. The election of president by inclusive rules with a second round has produced more efficient results than previous or still surviving experiences with plurality rule. For 88 presidential elections in nine countries (all South American except congressional Bolivia and non-democratic Paraguay, plus democratically stable Costa Rica) from 1945 to 2000, the comparative results can be summarized this way: less than 60 per cent (16/54) of presidents elected by plurality rule obtained the support of the median voter and were able to form a consistent political majority, while more than 85 per cent (29/34) of those elected with majority or qualified plurality rules with a second-round runoff obtained the support of the median voter at the first round. In all countries in which the electoral rule was changed – Argentina, Brazil, Chile, Ecuador, Peru and Uruguay – the results have been better from this point of view than with plurality rule in the previous period. Thus looking at institutional choices in Latin America from a long-term historical perspective, it can be argued that the development of multi-partism and the consequent adoption of inclusive electoral rules, especially proportional representation for Congress and more-than-simple-plurality rules with a second round for president, have decisively helped to make electoral results widely acceptable. In contrast to exclusive political and electoral regimes existing in previous periods – like the indirect electoral college or direct elections by simple plurality rule – the present rules in most countries are favouring wider acceptance and longer-lasting democratic regimes than ever before in the region.

The Americas: General Overview

93

References Alberdi, Juan B. ([1852] 1957) Bases y pensamientos de partida para la organización política de la República Argentina. Buenos Aires: Sopena. Arosemena, Justo (1878) Estudios constitucionales sobre los gobiernos de la América Latina, 2 vols. Paris: Libreria española y americana de E. Denne Calero, Manuel ([1908] 1989) ‘Cuestiones electorales’, in En torno a la democracia: El debate politico en México (1901–1916). Mexico: Instituto Nacional de Estudios Históricos de la Revolución Mexicana, pp. 181–242. Chiaramonte, José Carlos (1995) ‘Vieja y nueva representación: los procesos electorales en Buenos Aires, 1810–1820’, in Antonio Annino (ed.), Historia de las elecciones en Iberoamérica, siglo XIX. Mexico: Fondo de Cultura Económica, pp. 19–64. Hamilton, Alexander, James Madison and John Jay ([1788] 1961) The Federalist Papers. New York: New American Library. Santisteban, José Silva (1874) Curso de Derecho Constitucional, 3rd edn. Paris: A. Bouret (cit. in Jorge Basadre (1980) Elecciones y centralismo en Perú. Apuntes para un esquema histórico. Lima: Universidad del Pacífico).

Further reading • Annino, Antonio (ed.) (1995) Historia de las elecciones en Iberoamérica, siglo XIX. Mexico: Fondo de Cultura Económica.

• Bens, Walter (ed.) (1992) After the People Vote. A Guide to the Electoral College. Washington, DC: American Enterprise Institute.

• Botana, Natalio ([1977] 1998) El orden conservador. La política argentina entre 1880 y 1916. Buenos Aires: Sudamericana.

• Calabrese, Stephen (2000) ‘Multimember District Congressional Elections’, Legislative Studies Quarterly, 25, 4: 611–43.

• Calvin Jillson, (1979) ‘The Executive in Republican Government: the Case of American Founding’, Presidential Studies Quarterly, 9: 386–402.

• Chapell, Jr, Henry W. and William R. Keech (1989) ‘Electoral Institutions in the Federalist • • • • • • • • • •

Papers: A Contemporary Perspective’, in Bernard Grofman and Donald Wittman (eds), The Federalist Papers and the New Institutionalism. New York: Agathon, pp. 39–52. Flores, Nicolas (2000) A History of One-Winner Districts for Congress. Thesis, Stanford University. Glennon, Michael J. (1992) When No Majority Rules. The Electoral College and Presidential Succession. Washington, DC: Congressional Quarterly. Graham, Richard (1990) Patronage and Politics in Nineteenth-Century Brazil. Stanford, CA: Stanford University Press. Jones, Mark P. (1995) ‘A Guide to the Electoral Systems of the Americas’, Electoral Studies, 14: 15–21. Jones, Mark P. (1995) Electoral Laws and the Survival of Presidential Democracies. Notre Dame, IN and London: University of Notre Dame Press. Jones, Mark P. (1997) ‘A Guide to the Electoral Systems of the Americas: An Update’, Electoral Studies, 16: 13–15. Linz, Juan J. and Arturo Valenzuela (eds) (1994) The Failure of Presidential Democracy. Baltimore, MD: Johns Hopkins University Press. Mainwaring, Scott and Matthew S. Shugart (eds) (1997) Presidentialism and Democracy in Latin America. Cambridge and New York: Cambridge University Press. Malamud, Carlos (ed.) (2000) Legitimidad, representación y alternancia en España y América Latina: las reformas electorales (1880–1930). Mexico: Fondo de Cultura Económica. Milner, Henry (ed.) (1999) Making Every Vote Count: Reassessing Canada’s Electoral System. Peterborough, On.: Broadview Press.

94

Josep M. Colomer

• Nohlen, Dieter (1993) Enciclopedia electoral latinoamericana y del Caribe. San José: Instituto Interamericano de Derechos Humanos.

• Nohlen, Dieter, Sonia Picado and Daniel Zovatto (eds) (1998) Tratado de derecho electoral comparado de América Latina. Mexico: Fondo de Cultura Económica.

• Posada-Carbó, Eduardo (ed.) (1996) Elections Before Democracy: The History of Elections in • • • •

Europe and Latin America. London: Macmillan – now Palgrave/New York: St. Martin’s Press – now Palgrave. Riker, William H. (1984) ‘The Heresthetics of Constitution-Making: the Presidency in 1787, with Comments on Determinism and Rational Choice’, American Political Science Review, 79: 1–16. Shugart, Matthew S. and John M. Carey (1992) Presidents and Assemblies. Constitutional Design and Electoral Dynamics. Cambridge and New York: Cambridge University Press. Valenzuela, J. Samuel (1985) Democratización vía reforma: La expansión del sufragio en Chile. Buenos Aires: IDES. Weaver, R. Kent (2001) ‘Electoral Rules and Electoral Reform in Canada’, in Matthew S. Shugart and Martin P. Wattenberg (eds), Mixed-Member Electoral Systems: The Best of Both Worlds? New York and Oxford: Oxford University Press, pp. 542–70.

Summary Table 2A

The Americas: Assembly (lower or single chamber)

Country Law year Election years Argentina 1826: 1828 // 1853: 1862–1882, 1886, 1892, 1896, 1900, 1902, 1906, 1908 1912: Universal MS 1912, 1914, 1916, 1918, 1920, 1922, 1924, 1926, 1928, 1930 // 1946: 1946, 1951 // 1957, 1958, 1960, 1962 // 1963: 1963, 1965 // 1973 // 1983, 1985, 1987, 1989, 1991, 1993, 1995, 1997, 1999, 2001

Term

Districts

Ballot

Rule/Formula

Seats

No.

Magn.

4

44

15

>1

Public Bloc

Plurality

4

114–150

15

2–12

Bloc

Plurality

2

120–158

15

1–11

Secret Limited 2/3

Plurality

5

149–158

149

1

Single

Plurality

4

243–257

24

3–35

Closed list

Proportional – d’Hondt

1 ≥3

Single Limited

7–28

Closed list

Bolivia 1924:

1952: Universal MS 1956, 1958, 1960, 1962, 1964 // 1966 // 1980, 1985, 1989, 1993

Threshold

3% nat. 8% dis.

Plurality

4 5

102–130

8–9

Proportional – Hare 1 quota 95

96

Summary Table 2A

(Continued)

Country Law year Election years 1996: 1997, 2002

Brazil 1824: 1826, 1830, 1834, 1838 // 1860: 1860, 1864, 1867, 1869, 1872

Term

Districts

Ballot

Rule/Formula

Threshold

Single

Proportional – d’Hondt

3% nat.

Seats

No.

Magn.

5

130 62 68

9 9 68

5–31 5–31 1

4

102–111

19–20

1–20

Public Single/Multiple

Indirect Plurality

4

122

20

2–20

Bloc

Plurality

1875: 1876, 1878

4

122

20

2–20

Limited 2/3

Plurality

1881: 1881, 1885, 1886, 1889

4

122–125

122–125

1

Single

Direct Majority/2nd round runoff

1890: 1890

4

205

21

2–37

Bloc

Plurality

1892: 1894, 1896, 1899, 1903

3

205

63

2–37

Limited 2/3

Plurality

3

212

21

4–7

Limited M-1/ Cumulative

Plurality

2

214

22

2–37

1904: 1906, 1909, 1912, 1915, 1918 // 1927, 1930 1932: 1933 //

Secret Double: Open list Single

Parallel: Proportional – Hare Plurality

1945: 1945 1950: 1950, 1954, 1958, 1962 // 1982, 1986, Universal MS 1990, 1994, 1998, 2002 Canada 1867: 1867, 1872 1874: 1874, 1878, 1882, 1887, 1891, 1896, 1900, 1904, 1908, 1911, 1917, Universal MS 1921, 1925, 1926, 1930, 1935, 1940, 1945, 1949, 1953, 1957, 1958, 1963, 1965, 1968 1972: 1972, 1974, 1979, 1980, 1984, 1988, 1993, 1997, 2000

4

286

22

2–35

Open list

Proportional – Hare

4

326–513

25–27

1–75

Open list

Proportional – d’Hondt

5

181–200

181–200

1

Public Single

Plurality

4

206–265

206–301

Most 1 Some > 1

Secret Single Bloc

Plurality

4 5

264–301

264–301

1

Single

Plurality

Public Bloc

Plurality

Cumulative

Plurality

Chile 1833: 1879: 1879–1924 1925: 1925 // 1932, 1937, 1941, 1945, 1949, 1953, 1957, 1961, 1965, 1969, 1973 // 1980: Universal MS 1993, 1997, 2001

3

118

4

132–152

28

1–18

Open list

Proportional – d’Hondt

4

110–120

55–60

2

Secret Open list

Proportional – d’Hondt

1 quota

97

(Continued)

Country Law year Election years

98

Summary Table 2A

Term

Districts Seats

No.

Colombia 1857:

1931: 1931, 1933, 1935, 1937, Universal MS 1939, 1941, 1943, 1945, 1947, 1949 1951: 1951, 1953 //

1958: 1958, 1960, 1962, 1964, 1966, 1968 1970: 1970, 1974, 1978, 1982, 1986, 1990, 1991, 1994, 1998, 2002 Costa Rica 1893: 1894, 1896, 1898, 1900, 1902, 1904, 1906, 1908, 1910, 1912

Bloc

Plurality

Single

Plurality

>3

Limited 2/3

Plurality

Closed list

Proportional – Hare

Single

Two tiers: Plurality Proportional – Hare

1 4

Rule/Formula

Threshold

Magn.

>1

1886: 1910: 1913, 1917, 1921, 1925, 1929

Ballot

2

118–132

33

3–9

2

132 88 44

88

1

2

148–204

26

*7

Closed list

Two-party proportional – Hare

4

160–210

26–33

2–10

Closed list

Proportional – Hare

Public 4

41–42

7

Indirect Coexistence Proportional – Hare Majority/2nd round runoff

≥3 ≤2

Multiple Single

1913: Universal MS 1919, 1921, 1923, 1925, 1928, 1930, 1932, 1934, 1936,1938, 1940, 1942, 1944, 1946, 1948 // 1953: 1953, 1958, 1962, 1966, 1970, 1974, 1978, 1982, 1986, 1990, 1994, 1998, 2002 Cuba 1901: 1908, 1910, 1912, 1914, 1916, 1918, 1920, 1922, 1924, 1926 // 1940: 1940, 1942, 1944, 1946, 1948, 1950 // Dominican Republic 1966: Universal MS, 1966, 1970, 1974, 1978, 1982, 1986, 1990, 1994, 1998, 2002 Ecuador 1895: 1904, 1906, 1912, 1914, 1916, 1924 // 1928 //

Secret 4

43 32 11

7

4

45–57

4

Direct Coexistence

≥3 ≤2

Closed list Bloc

7

4–21

Closed list

Proportional – Hare

84

6

2–11

Closed list

Indirect Proportional – Hare

4

127–136

6

2–11

Closed list

Direct Proportional – Hare

4

91–150

30–33

2–31

Closed list

Proportional (2 tiers) – d’Hondt

Public Open

Plurality

Closed list

Proportional – Hare

1- quota -2

Two closed lists

Prop. (2 tiers) – Hare

1 --- quota 2

2

1946: 1950, 1952, 1954, 1956, 1958, 1960, 1962 //

2

1978: Universal MS 1979, 1984, 1986, 1988, 1990, 1992, 1994, 1996

2–4

21

72–121

21–23

1–20

1 quota Proportional – Hare Plurality 1- quota -2

1 quota

1- quota -2

99

100

Summary Table 2A

(Continued)

Country Law year Election years

Term

Districts Seats

No.

Ballot

Rule/Formula

Open ballot

Plurality

Magn.

1998: 1998

4

121

2000: 2002

4

121

23

1–20

Open list

Proportional – d’Hondt

42

14

3

Public Open

Plurality Proportional – Hare

El Salvador 1886: Universal MS 1886 // 1963: 1964, 1966, 1968, 1970, 1972, 1974, 1976, 1978 //

Threshold

2

52

14

*4

Secret Closed list

1984: 1985, 1988, 1991, 1994, 1997, 2000

3

60–84

14

3–20

Closed list

Proportional (2 tiers) – Hare

Guatemala 1944: Universal MS 1944, 1950, 1953

6

76

22

1–10

Open list

Proportional – Hare

1984: 1985, 1990, 1995, 1999

5

80–116

22

2–20

Two closed lists

Proportional (2 tiers) – d’Hondt

Honduras 1879: 1930 // 1954

4

59

18

*3

Public Multiple

Majority

4

82–134

18

1–16

Closed list

Proportional – Hare

1966: Universal MS 1981, 1985, 1989, 1993, 1997, 2001

Jamaica 1962: 1962, 1967, 1972, 1976, 1980, 1983, 1989, 1993, 1997, 2002 Mexico 1824: 1824, 1827, 1829, 1831, 1833 // 1843, 1845, 1850 // 1857: Universal MS 1857, 1861 // 1867, 1871, 1872, 1876 // 1917: 1918, 1920, 1922, 1924, 1926, 1928, 1932 1934, 1937, 1940, 1943, 1946, 1949, 1952, 1955, 1958, 1961 1963: 1964, 1967, 1970, 1973, 1976

1977: 1979, 1982, 1985

1986: 1988, 1991, 1994.1997, 2000, 2003

60

60

1

Single

Plurality

2

150

150

1

Public Single

Indirect Plurality

2

180

180

1

Secret Single

Plurality

2

136–263

136–263

1

Single

Direct Plurality

210–237 178 32–43

178 1 nat.

1 32–35

372–400 272–300 100

272–300 1(3–5 lists)

1 100

500 300 200

300 1(5 lists)

1 200

90 (+2–3)

9–18

1–25

3

3

3

3

6

Single

Two tiers: Plurality Proportional – Hare

1.5–2.5% nat.

Double: Single Closed list

Parallel: Plurality Proportional – Hare

1.5% nat.

Single

Two tiers: Plurality Proportional – Hare

2% nat.

Closed list

Proportional (2 tiers) – Hare

101

Nicaragua 1987: Universal MS 1987, 1990, 1995, 2001

5

(Continued)

Country Law year Election years Panama 1916: 1918, 1924 // 1946: 1948, 1952, 1956, 1960, 1952, 1956, 1960, 1964, 1968 // 1983: 1984 // 1989 // 1994, 1999

Term

Districts Seats

6

Ballot

Rule

No.

Magn.

10

>1

Public Closed list

Indirect Proportional – Hare

6

42–53

10

*5

Secret Closed list

Direct Proportional – Hare

5

67–71 39–43 28

40 12 28

1–6 2–5 1

Closed list Single

Coexistence Proportional – Hare Plurality

Paraguay 1993: Universal MS 1993, 1998

5

80

18

1–14

Closed list

Proportional – d’Hondt

Peru 1828: 1828 //

4

87

87

1

Public Single

Indirect Plurality

1861: 1866 // 1868, 1870, 1872, 1874 // 1896–1919

4

130–161

80–130

1–2

Cumulative

Direct Plurality

1933: 1939, 1945 // 1950, 1956, 1962 // 1963 //

6

139–145

24

*6

Secret Closed list

Proportional – d’Hondt

1979: Universal MS 1980, 1985, 1990 // 1995 // 2000, 2001

5

120–180 120

24 1 nat.

1–11 120

Open list

Proportional – d’Hondt

Threshold

102

Summary Table 2A

Trinidad-Tobago 1962: 1966, 1971, 1976, 1981, 1986, 1991, 1995, 2000, 2001, 2002 United States 1789: 1790, 1792, 1794, 1796, 1798, 1800, 1802, 1804, 1806, 1808, 1810, 1812, 1814, 1816, 1818, 1820, 1822, 1824, 1826, 1828, 1830, 1832, 1834, 1836, 1838, 1840 1842, 1844, 1846, 1848, 1850, 1852, 1854, 1856, 1858, 1860, 1862, 1864, 1866, 1868, 1870, 1872, 1874, 1876, 1878, 1880, 1882, 1884, 1886, 1888, 1890, 1892, 1894 1896, 1898, 1900, 1902, 1904, 1906, 1908, 1910, 1912, 1914, 1916, 1918, 1920, 1922, 1924, 1926, 1928, 1930, 1932, 1934, 1936, 1938, 1940, 1942, 1944, 1946, 1948, 1950, 1952, 1954, 1956, 1958, 1960, 1962, 1964, 1966, Universal MS 1968, 1970, 1972, 1974, 1976, 1978, 1980, 1982, 1984, 1986, 1988, 1990, 1992, 1994, 1996, 1998, 2000, 2002

36

36

1

Single

Plurality

2

106–391

Most Some

>1 1

Public Bloc Single

Plurality

Most Some

1 >1

Single Bloc

2

391–435

Secret

435

435

1

Single

3

123

19

4–30

Open list

4

99

1 nat. (19 lists)

99

5

Proportional – d’Hondt

103

Uruguay 1918: Universal MS 1919, 1922, 1925, 1928, 1930, 1931 // 1934, 1938, 1942, 1946, 1950, 1954, 1958, 1962, 1966, 1971 // 1984, 1989, 1994, 1999

5

104

Summary Table 2A

(Continued)

Country Law year Election years

Term

Districts

Ballot

Rule/Formula

Threshold

Seats

No.

Magn.

5

110–201

23

1–37

Closed list

Proportional – d’Hondt

1989: 1993, 1998

5

204–207 102–119 88–102

23–24 23–24 88–102

3–25 3–25 1

Double: Closed list Single

P-Proportional (2 tiers) – d’Hondt/Hare

1999: 2000

5

165

24

3–14

Closed list

Proportional – Hare

Venezuela 1946: Universal MS 1947 // 1958, 1963, 1968, 1973, 1978, 1983, 1988

Source: See Appendix: Notes and Sources for Summary Tables at the end of the book.

The Americas: Presidency

Country Law year

Election years

Term

Rule

1862, 1868, 1874, 1880, 1886, 1892, 1898, 1904, 1910, 1916, 1922, 1928 // 1931, 1937 // 1946 1951 // 1958, 1963 // 1973a, 1973b // 1983, 1989 1995, 1999

6

College majority/2nd round runoff by Congress

6+6 6 6 6 4+4

Plurality College majority/2nd round runoff by Congress Majority/2nd round runoff College majority/2nd round runoff by Congress Qualified plurality (either 45% or 40% + 10)/2nd round runoff

1904, 1909, 1913, 1917 // 1921, 1926, 1931 // 1940 // 1947 // 1951 1956, 1960, 1964 // 1980, 1985, 1989, 1993, 1997, 2002

5

Majority/2nd round runoff by Congress

4+4 4

Plurality Majority/2nd round runoff by Congress

1835, 1838 (Regent) 1894, 1898, 1902, 1903, 1906, 1910, 1914, 1918, 1919, 1922, 1926, 1930 // 1945, 1950, 1955, 1960 // 1989 1994 1998, 2002

3 4

College majority/2nd round runoff by Congress Majority/2nd round runoff by Congress

5 5 4 4+4

Plurality Majority/2nd round runoff

1925, 1927, 1931, 1932, 1938, 1942, 1946, 1952, 1958, 1964, 1970 // 1989, 1994, 1999

6

Majority/2nd round runoff by Congress

5

Majority/2nd round runoff

Argentina 1853: 1949: 1955: 1972: 1983: 1994:

Bolivia 1899: 1956: 1967: Brazil 1834: 1892: 1945: 1986: 1993: 1997: Chile 1925: 1989:

105

Summary Table 2B

(Continued)

Country Law year

Election years

Colombia 1821: 1832: 1853: 1863: 1886: 1910:

1991: Costa Rica 1889: 1913: 1926: 1936:

Cuba 1901: 1940: Dominican Republic 1962: 1995:

106

Summary Table 2B

Term

Rule

4 2

College 2--- /2nd round 2--- Congress/3rd round runoff by 3 3 Congress College majority/2nd round runoff by Congress Plurality College states-majority/2nd round runoff by Congress

6+6 4

College majority/2nd round runoff by Congress Plurality

4

Majority/2nd round runoff

1890, 1894, 1898, 1902 // 1909 1913, 1917, // 1919, 1923 1928, 1932 1936, 1940, 1944, 1948 // 1953, 1958, 1962, 1966, 1970, 1974, 1978, 1982, 1986, 1990, 1994, 1998, 2002

4 4 4 4

College majority/2nd round runoff by Congress Majority/2nd round runoff by Congress Majority/2nd round runoff 2 --- (40%)/2nd round runoff

1908, 1912, 1916, 1920, 1924 // 1940, 1944, 1948 //

4+4 4

College majority/2nd round runoff by Congress College plurality

1962 // 1966, 1970, 1974, 1978, 1982, 1986, 1990, 1994 1996, 2000

4

Plurality

4

Majority/2nd round runoff

1826 // 1833, 1837, 1841, 1841, 1845, 1849, 1853 1857, 1860 // 1864, 1866, 1868, 1870, 1872, 1874, 1876, 1878, 1880, 1882, 1884 // 1892, 1898, // 1904 1914, 1918, 1922, 1926, 1930, 1934, 1938, 1942, 1946, 1950, // 1958, 1962, 1966, 1970, 1974, 1978, 1982, 1986, 1990 1994, 1998, 2002

5

Ecuador 1895:

4

Majority/2nd round plurality

1978: 1983: 1998:

1901, 1905, 1911, 1912, 1916, 1920, 1924 // 1948, 1952, 1956, 1960 // 1978 1984, 1988, 1992, 1996 1998, 2002

5 4 4

Majority/2nd round runoff 40% + 10/2nd round runoff

El Salvador 1886: 1963: 1983:

1803, 1907, 1911, 1915, 1919, 1923, 1927, 1931 // 1967, 1972, 1977 // 1984, 1989, 1994, 1999

4 5 5

Majority/2nd round runoff 1--3- by Congress Majority/2nd round runoff by Congress Majority/2nd round runoff

Guatemala 1944: 1984: 1994:

1944, 1950 // 1985, 1990 1995, 1999

6 5 4

Majority/2nd round runoff by Congress Majority/2nd round runoff

Honduras 1879: 1966:

1924, 1928, 1932 // 1954, 1956, 1957 // 1981, 1985, 1989, 1994, 1999

4 4

Majority/2nd round runoff by Congress Plurality

1825, 1829, 1833 // 1847, 1851 // 1857, 1861, 1867, 1871, 1872 1876a, 1876b, 1880, 1884, 1888 1892, 1898, 1904, 1910, 1911 // 1917, 1920, 1924 1928, 1929 1934, 1940, 1946, 1952 1958, 1964, 1970, 1976, 1982, 1988, 1994, 2000

4 4+... 4+... 6+... 4 6+6 6

College states majority/2nd round runoff by Congress

1928, 1932 // 1984, 1990

4 6+6

Mexico 1824: 1857: 1874: 1890: 1917: 1927: 1933:

Plurality

Majority/2nd round runoff by Congress Plurality

107

Nicaragua 1911: 1984:

College majority/2nd round runoff by Chamber Deputies

108

Summary Table 2B

(Continued)

Country Law year

Election years

Term

Rule

1995: 1999:

1996 2002

6 6

Qualified plurality 45%/2nd round runoff Qualified plurality 2--- (40%)/2nd round runoff

4

Plurality

1983:

1916, 1920, 1924, 1928 // 1952, 1956, 1960, 1964, 1968 // 1984 // 1989 // 1994, 1999

Paraguay 1989:

1989, 1993, 1998

5

Plurality

Peru 1828: 1861: 1896: 1931: 1933: 1978: 1993: 2001:

1830 // 1862, // 1866 // 1868, 1872, 1876 // 1899, 1903, 1904, 1908, 1912 // 1915 // 1931 1939, 1945, // 1956 // 1962, // 1963 // 1980, 1985, 1990, 1995 1995, 2000 2001

4+4 4 4 5 6 5 5+5 5

College majority/2nd round --31- by Congress College majority/2nd round runoff by Congress Plurality Qualified plurality 1--4Qualified plurality --13- /2nd round 1--3- by Congress Majority/2nd round runoff

1788–9, 1792, 1796, 1800, 1804, 1808, 1812, 1816, 1820, 1824, 1828, 1832, 1836, 1840, 1844, 1848, 1852, 1856, 1860, 1864, 1868, 1872, 1876, 1880, 1884, 1888, 1892, 1896, 1900, 1904, 1908, 1912, 1916, 1920, 1924, 1928, 1932, 1936, 1940, 1944, 1948

4+4+...

College majority/2nd round runoff by Congress

Panama 1916:

United States 1789:

5

5

1952, 1956, 1960, 1964, 1968, 1972, 1976, 1980, 1984, 1988, 1992, 1996, 2000

4+4

4 4

1966: 1996:

1899, 1903, 1907, 1911, 1915 1920, 1925, 1926, 1928, 1930, 1932, 1938 // 1942, 1946, 1950 1966, 1971 // 1984, 1989, 1994 1999

Venezuela 1830:

1831, 1835 // 1847, 1851, 1855 //

4

1877 // 1947 // 1958, 1963, 1968, 1973, 1978, 1983, 1988, 1993, 1998 2000

2 5

1951:

Uruguay 1899: 1918:

1874: 1946: 1999:

5 5

Plurality

Majority/2nd round runoff College states 2--3- /2nd round 2--3- /3rd round runoff by Congress College states majority/2nd round runoff by Congress Plurality

6+6

Source: See Appendix: Notes and Sources for Summary Tables at the end of the book.

109

3 Argentina: Compromising on a Qualified Plurality System Gabriel L. Negretto

In 1994 Argentina shifted from the indirect election of presidents via an electoral college to a system of direct election by qualified plurality rule with few precedents. The new rule emerged as a byproduct of the short-term electoral expectations and relative strength of the main negotiators of the constitutional reform of 1994. The alternative thresholds of 45 per cent or 40 per cent with a difference of 10 points over the second most voted candidate were created as a compromise between the plurality system desired by the incumbent president of the Justicialist Party (PJ), Carlos Menem, and the majority runoff system demanded by Raúl Alfonsín, leader of the main party of the opposition, the Radical Civic Union (UCR). These rules were chosen with the aim of maintaining a two-party system with the PJ as the dominant party. Since 1995, however, the emergence of third forces has created a more uncertain scenario in which the outcome of elections is determined by the ability of incumbent parties to reach the minimum thresholds and the capacity of opposition parties to coordinate into a common alliance. In this context, the qualified plurality system created by self-interested actors is likely to lead to less arbitrary and more socially acceptable results than its main alternatives, plurality rule and majority runoff.

The initial rules: the Electoral College Like many other Latin American constitutions during the nineteenth century, the Argentine constitution of 1853 adopted a system of indirect election of presidents by an Electoral College inspired by the precedent of the American constitution. Except for the elections of 1951 (plurality) and 1973 (majority runoff), this method basically determined the election of all presidents in Argentina until the reform of 1994. The system established a three-stage process for the election of presidents. First, the citizens of each province and the capital (24 districts since 1983) would vote for electors for President and Vice-President in a number equal to double the number of deputies and senators each district sent to Congress. Second, the electors would meet in the capital of each province (meaning that the ‘Electoral College’ never met as a single collective body) to cast their votes for President and Vice-President 110

Argentina

111

on separate ballots. Third, the Congress would meet in joint session to certify the number of votes cast for each candidate and proclaim winners those who received an absolute majority of the vote. Failing a majority for either President, Vice-President or both, Congress in joint session – with a qualified quorum of three-fourths of the totality of the members of both chambers – would decide between the two most voted candidates. Two important features must be highlighted about the working of this system: first, the disproportionality it created between population size and number of electors per province; second, the way it worked as an effective plurality rule in spite of the majority principle that regulated the decisions of the Electoral College. The requirement that each province would select a number of electors equal to double the number of senators (two) and deputies (variable according to population) sent to Congress meant in practice a minimum of six electors per province. This rule, clearly beneficial for small provinces, was reinforced after 1951 with the establishment of a minimum number of deputies, which was initially two in 1951 and five since 1983. As a result, no district (with the exception of Tierra del Fuego) sent less than 14 electors from 1983 to 1989. Due to this rule, winning in the largest districts, like the provinces and cities of Buenos Aires, Córdoba and Santa Fe, provided a considerable advantage but it was not sufficient to reach a majority of electors. Although a majority of electors was required to win a presidential election, the system effectively worked under the effects of plurality rule. First, for most of the existence of the Electoral College until 1963, electors for President were selected by plurality rule from multi-member districts. The effect was that a candidate could obtain a majority of electors by simply winning a plurality of the total national vote. In 1922 and 1958, for instance, 49 per cent of the national vote was equivalent to 57 and 68 per cent of the total number of electors, as shown in Table 3.1. Second, when a plurality winner in popular votes did not get a majority of electors, the practice was to provide this candidate with a majority of votes in the Electoral College. This occurred twice using plurality rule and limited vote (1916 and 1931) and once (1963) using proportional representation (with d’Hondt formula) for the selection of electors. In particular, in 1916 additional support came from provincial parties that competed for electors but did not commit to vote for any particular candidate. Because of this practice, in no single instance was the intervention of Congress necessary to decide on the election of presidents. However, on two occasions (1854 and 1860) Congress did intervene to select a Vice-President. The majority or near majority support that on average presidents obtained in electors and (for the periods for which we have reliable data) in popular votes was only in part due to the incentives that the Electoral College may have provided to the formation of broad coalitions. Elections were hardly fair or competitive until the electoral reform of 1912. From then on, while the legal framework made possible an effective party competition, the political process prevented it, first, because the main democratic parties in the country, like the Radical Civic Union (UCR) from 1916 to 1928 and the Justicialist Party (PJ) from 1946 to 1952, emerged as dominant parties, and second, because electoral competition was repeatedly restricted by the

112

Gabriel L. Negretto

Table 3.1 Argentina: Electoral College presidential elections, 1854–1989 Year

1854 1860 1862 1868 1874 1880 1886 1892 1898 1904 1910 1916 1922 1928 1931 1937 1946 1958 1963 1983 1989

Winner

Runner-up

Popular votes %

Electors %

Votes in college %

Popular votes %

Votes in college %

N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A 47 49 62 32 56 54 49 32 52 49

73 56 88 52 64 68 72 91 73 81 88 44 57 66 36 66 81 68 35 53 53

89 58 100 60 65 69 79 95 85 81 99 51 63 65 63 65 80 68 57 57 54

N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A 13 8 10 31 42 44 32 21 40 37

7 37 – 20 35 31 15 2 15 11 0.4 35 16 19 32 34 18 29 18 43 39

Note: Author’s elaboration from data in Molinelli (1989) and Nohlen (1993). Popular votes are proportions of valid and positive votes (excluding null and blank ballots), while electors and votes in college are proportions of the total number of electors (including absentees, abstentions and blank ballots).

proscription of one of these parties – the UCR during the 1930s and the PJ during the 1950s and 1960s. From this perspective, it has been only since 1983 that the country has experienced an open electoral competition in the context of a seemingly solid two-party system.

Compromising between plurality and majority rule A failed attempt at constitutional reform during the presidency of Raúl Alfonsín of the UCR (1983–89) is the most immediate and important precursor of the reform that took place in 1994. In 1986 and 1987 a presidential commission produced a report on constitutional reform suggesting the convenience of replacing the existing presidential regime by a premier-presidential one. When in January 1988 president Alfonsín (UCR) initiated formal negotiations with the PJ, then the main opposition party, experts and political leaders from this party had already exposed their disagreement with central aspects of the proposal. Their main target of criticism

Argentina

113

was the attempt to create a Prime Minister politically responsible to the Chamber of Deputies and the transformation of the Senate (where the PJ had a plurality) into a mere revising chamber on issues other than the distribution of territorial power. But the report of the commission also included other areas of potential disagreement between the two main parties, such as the replacement of the Electoral College by a direct election by majority rule. The PJ rejected this method and favoured instead a popular election by plurality rule. According to one of the leading negotiators for the PJ, Alberto García Lema, the majority principle was inconvenient given the two-party system that prevailed in Argentina. In his view, in case neither the UCR nor the PJ reached the threshold of 50 per cent of the vote, a second round would give undue weight to small third parties, situated to the left or to the right of the two main contenders (García Lema, 1994: 184–5). The position of the PJ, however, is better understood by taking into account the electoral expectations of this party. Since its creation in the mid-1940s the PJ, led by Juan D. Perón, became a majority party, winning every presidential and congressional election where it was allowed to participate (the party was banned from 1955 to 1973). In this historical context, the success of the UCR in the presidential elections of 1983 and in the midterm legislative elections of 1985 was somewhat exceptional, related to the levels of political violence and mismanagement that most voters associated with the last Perón government. By 1987, however, the electoral trend was once again reversed in favour of the PJ. In the midst of a deep economic crisis, the PJ defeated the UCR in the midterm elections of August 1987, depriving the latter of the absolute majority it had in the Chamber of Deputies. In addition, the PJ won the elections for governor in 17 out of 22 provinces, making predictable a majority of this party in the Senate by 1989. In this electoral context it was obviously not in the interest of the negotiators of the PJ to accept an electoral reform that would damage its chances to win the presidential elections of 1989, much less when by June 1988 the PJ’s candidate, Carlos Menem, showed in most polls as the most likely winner. The current formula of presidential elections emerged in 1993 as a result of the second round of negotiations on constitutional reform that took place between the new president, Carlos Menem, and former president Alfonsín, still principal leader of the UCR. The process started in 1992, when in the context of a stabilized economy that led to rising levels of approval for his government, Menem proposed a constitutional reform whose main intention was to remove the existing proscription of immediate presidential re-election. The proposal included few other issues, but among them was the direct election of the President by plurality for four years. As was to be expected, the UCR rejected the reform. Given this opposition, Menem faced two serious obstacles to pass the reform. In the first place, he needed a two-thirds majority in both chambers and, at least in the Chamber of Deputies, it could not obtain that qualified majority without the support of the UCR.1 This meant that he could not risk a breakdown of negotiations without jeopardizing the possibility of being re-elected. Second, time was not in favour of Menem: to be re-elected in 1995, he had to conclude the reform early in

114

Gabriel L. Negretto

1994, before the PJ started its own internal competition among potential presidential candidates. As a consequence, the UCR could delay negotiations in order to frustrate Menem’s re-election or, exploiting his impatience, obtain greater concessions in exchange for it. But Menem was not powerless. He counted on a growing popular support that he could use to undermine the capacity of the opposition to hold out or eventually reject negotiations altogether. His bargaining strategy consisted then of making credible two complementary threats. The first was to call a plebiscite on the issue of his re-election, which he did right after winning the midterm elections of October 1993. Since the plebiscite was non-binding, Menem also issued another threat in case the plebiscite was successful: passing the reform in an irregular manner, without the support of the UCR. Specifically, the threat consisted of passing the reform by surprise, counting the required two-thirds majority over the present (rather than over the total) number of deputies. With polls showing more than 60 per cent of approval for Menem’s re-election, Alfonsín, as the representative of the UCR, decided to accept the re-election if he obtained something in exchange for his party. Along with other demands aimed at improving the institutional and political position of the UCR, Alfonsín required a reduction of the presidential term from six to four years and the direct election of the President by majority runoff. Menem accepted these demands only in part. In particular, he accepted the possibility of a second round of presidential elections only if the threshold to win in the first round was lower than 50 per cent. The latter was in fact one of the most controversial issues of the negotiation and the last to be defined. Both the agreement of 14 November (the so-called ‘Pacto de Olivos’) and the document signed on 1 December 1993, established that the President would be elected in two possible rounds but without defining a specific threshold to win in the first round. The negotiation about the threshold evolved as follows. Initially, the UCR demanded a threshold of 50 per cent, while the PJ stuck to a floor of no more than 40 per cent. Between late November and early December, negotiators from both parties agreed to ‘split the difference’. The UCR accepted a lowering of the minimum threshold from 50 to 45 per cent. The PJ agreed but only if an alternative rule was also established: that a candidate could still win with 40 per cent of popular votes if a significant difference separated the front-runner from the runner-up (La Nación, 28 November 1993). While there was some discussion about the exact percentage of votes that should separate both candidates, by 13 December a final document set the difference at 10 per cent. In addition, it was established that all percentages should be counted over ‘affirmative valid votes’, that is excluding not only null but also blank votes. The new formula thus emerged. A single ticket for President and Vice-President could win the election in the first round in two alternative ways: either by obtaining more than 45 per cent of affirmative valid votes, or at least 40 per cent with an advantage of more than 10 percentage points over the second most voted candidacy. In case no candidate reached these thresholds, a second round between the two most voted candidates would take place 30 days after the first. After the election

Argentina

115

of a Constituent Convention, which produced a new constitution in 1994, the new rules were incorporated in articles 94 to 98. The new formula was then a mid-way solution, a compromise between the plurality rule desired by Menem and the majority runoff supported by Alfonsín. These preferences did not derive from the influence of foreign models or expert advice but from the actors’ expectations on voters’ support, based on the recent and previous electoral performance of their parties. The minimum threshold of 40 per cent to win a presidential election certainly had some precedents, such as in Costa Rica in 1936 or in the proposals for constitutional reform of presidential elections considered in the United States in 1969 and 1977 (as mentioned by García Lema, 1994: 185–6). However, it is apparent that this threshold was primarily chosen by looking at the electoral performance of the PJ from 1983 to 1993. The average support for the PJ in the elections for deputies was 41 per cent. In the more polarized presidential elections the range of support for the PJ fluctuated between a minimum of 40 in 1983 to a maximum of 49 per cent in 1989. Obviously, 40 per cent was the safest choice because it was the ‘floor’ of votes for the party over ten years. Accepting a difference of 10 percentage points between the first and the second most voted candidacies was also a calculated choice, since in the last two legislative elections the difference between the PJ and UCR was, respectively, 11 and 12 per cent. Something similar could be said of the preferences of the UCR’s negotiators. While the majority formula was initially proposed by an impartial commission of experts, it is clear that Alfonsín’s insistence on a threshold of 50 per cent was also related to the electoral experience of the main parties. In both legislative and presidential elections the maximum of votes obtained by the PJ, including the 1989 presidential election, was below 50 per cent. With the PJ situated since 1983 in a centre-right position, the leader of the centrist UCR could expect that in a second round his party could obtain the support of left and centre-left parties and some centrist provincial parties closer to the UCR than to the PJ. To be sure, given the bargaining strength of the incumbent president during the negotiations, Alfonsín had no other choice but to accept a lower threshold of 45 per cent and the alternative rule of 40 per cent with a difference of 10 points. Nevertheless, it was still a better option than a mere 40 per cent without considering the difference with the runner-up. Given the experience of ten years of electoral competition both parties rationally expected to concentrate together (albeit unevenly) the largest share of votes and maintain themselves as the main political options in coming presidential elections. No third party had up till then obtained more than 7 per cent of the vote in a presidential election. But, as we will see, these expectations were based on certain characteristics of the party system that did not hold in the medium term.

The new formula The electoral formula created in 1994 presents two important features that are crucial to an understanding of the possible political consequences of its implementation.

116

Gabriel L. Negretto

The first, perhaps obvious, characteristic is that by replacing the Electoral College it eliminates the disproportionate weight that electorates and parties of small provinces had under the previous system. Now winning in large districts, such as the city of Buenos Aires and the provinces of Buenos Aires, Córdoba and Santa Fe, which gather 66 per cent of the registered voters in the country, becomes crucial to decide a national election. A second, most important feature of the new system is the way in which the alternative thresholds work. Just like pure plurality in a multi-candidate competition, it does not eliminate the possibility of a very narrow winner. Since the difference of 10 percentage points is only required when the front-runner obtains more than 40 but less than 45 per cent of popular votes, then, in the case of a polarized election between two blocs, it is possible to win by a margin of 45 to 44. The qualified plurality formula, however, is not likely to produce winners with a very small share of votes, as simple plurality systems might in a multi-party competition. Also similar to plurality, the new formula does not prevent a winner who is in fact rejected by a majority of voters. This could happen, for instance, if a front-runner with 40 per cent of the votes is the last preference of those who voted for second and third alternatives (each with, say, 30 per cent of support). But the minimum margin of 10 percentage points makes this outcome less likely than under simple plurality rule because it would require high fragmentation among politically close losing candidates. If no candidate reaches the minimum margin or the minimum threshold to win in the first round, the Argentine formula would prevent a ‘Condorcet-loser’ from becoming president. Apart from the obvious similarities between qualified plurality and majority runoff when a second round is needed to decide the election, the main difference between both formulas must be found in the different incentives they provide for coalition-making among parties before the election. In this sense, qualified plurality rule induces the formation of broad coalitions in support of a relatively low number of candidates less effectively than simple plurality but more effectively than majority runoff. Specifically, whenever one candidate (particularly from the incumbent party) is likely to get more than 40 per cent of the vote, opposing parties may have an incentive to coalesce before the election, either to reach the minimum share of 45 per cent and win in the first round or to force a runoff by reducing the winning margin of the front-runner. This would certainly occur less often with a threshold of 50 per cent which may induce different opposition parties to coalesce after the first round.

Political consequences The analysis of the actual consequences of the new formula is based on a rather small number of observations: the presidential elections of 1995 and 1999. But even if limited, these two cases provide enough material to understand the logic of the formula, particularly in comparison with the results of the presidential elections of 1983 and 1989 under the previous system. Table 3.2 shows this comparison.

Argentina

117

Table 3.2 Argentina: presidential elections, 1983–99 Year

Winner

1983

Alfonsín UCR Menem PJ Menem PJ De la Rúa UCR-Frepaso

1989 1995 1999

% votes 51 49 50 48

Runner-up Luder PJ Angeloz UCR Bordón Frepaso Duhalde PJ

% votes 40 37 29 38

Third Allende PI Alsogaray AC Massaccesi UCR Cavallo AR

% votes

Other %

2

7

7

7

17

4

10

4

Note: Author’s elaboration from data in Molinelli, Palanza and Sin (1999). Percentages of votes include all parties supporting the candidate. AC: Alliance of Center; AR: Action for the Republic; Frepaso: Front Solidarity Country; PI: Intransigent Party; PJ: Justicialist Party; UCR: Radical Civic Union.

An important aspect of the presidential election of 1995 revealing the structure of incentives created by the new formula is the fact that no major alliances were formed. Moreover, besides the three relevant candidacies of the PJ, the UCR and the Frepaso (Front Solidarity Country, a recently created centre-left coalition), there was an increase in the number of small parties (most of them from the city of Buenos Aires) with individual candidates obtaining less than 2 per cent of the vote each. In total, there were eleven such candidacies in 1995, compared to six in 1989 and nine in 1983. What explains this absence of alliance-building among opposition forces? As mentioned in the previous section, the new formula makes more likely than majority rule the coalescence of parties opposed to the candidacy of the expected front-runner but in no way guarantees this result. With simple plurality, leaders of opposition parties have a strong incentive to coalesce whenever the candidate of the incumbent party is expected to be the front-runner, whatever the proportion of votes expected. But with the Argentine formula this incentive may exist only as long as the candidate of the incumbent party is expected to obtain more than 40 per cent and maintain the margin required or else surpass the barrier of 45 per cent. Even if coalition-building among opposition parties becomes necessary to prevent a winner at the first round, those parties may still fail to coordinate as a single political force. This could happen when opposition parties are ideologically distant from each other. An ideologically disparate coalition is not likely to gather together the votes of the partners, which may also face an important loss in their future electoral base. Opposition parties must also be able to overcome the coordination problem represented in the selection of a single presidential candidate, particularly if each of these parties has a strong candidate for the election. The absence of an alliance between the centrist UCR and the centre-left Frepaso can be explained by the electoral expectations they had regarding the presidential

118

Gabriel L. Negretto

election of 1995. In the congressional elections of April 1994 the PJ and the UCR obtained, respectively, 38 and 20 per cent of the national vote, while the Grand Front (FG, antecessor of Frepaso) obtained 13 per cent. These results suggested that the dominance of the PJ was no longer assured, thus making the main opposition parties expect to obtain more benefits by running alone in the first round of the presidential election than by making a coalition with the other. By August 1994, when opposition parties initiated the process of candidate selection, polls were giving Menem 36 per cent of electoral preferences, followed by Carlos Alvarez, the potential presidential candidate of the Frepaso, with 24 per cent (Novaro y Palermo, 1996: 451). It was precisely during that month that leaders of the Frepaso and UCR met to explore the possibility of an alliance but, not surprisingly, this possibility was soon abandoned by both parties. Menem finally won in the first round by almost 50 per cent, followed by Bordón (Frepaso) with 29 per cent and Massaccesi (UCR) with 17 per cent of the vote. But no one could predict this result based on the most recent electoral performance of the PJ. Compared to the 1995 election, the most salient aspect of the 1999 election was the emergence of an Alliance among opposition parties recreating a pattern of competition between two blocs. In coincidence with this concentration of the vote between two main national candidates, there was also a decrease from eleven to six in the number of small parties (all of them from the city of Buenos Aires) presenting individual candidacies. No provincial parties ran alone this time. This realignment of the party system in two blocs began in 1997, when the UCR and Frepaso ran together at the legislative elections of that year. The unification of the two major opposition forces into a single coalition was no doubt induced by the results of the presidential election of 1995, which updated the beliefs of opposition parties about their slender chances of defeating the PJ without mutual support. The success of the alliance in the legislative elections of 1997 reinforced these beliefs and paved the way for the coordination of UCR and Frepaso in presenting a single candidacy to compete against the PJ in the presidential elections of 1999. One should observe, however, that the process of coalition-building that took place before the election did not lead to a perfect bi-party competition. A third candidate, Domingo Cavallo, Menem’s former Secretary of Economy who was now running for a recently created centre-right party, Action for the Republic (AR), won 10 per cent of the vote, thus playing a significant role in the election. While this party was ideologically close to the incumbent party, and at several times the candidate of the PJ, Eduardo Duhalde, invited its candidate to form an alliance before the election, Cavallo decided to run alone. Given that most polls persistently indicated the candidate of the Alliance UCR-Frepaso, Fernando De la Rúa, as the virtual winner in the first round, this decision was a dominant strategy for the candidate of AR. If, as expected, Duhalde were defeated in the first round, Cavallo could exploit the concurrence of legislative elections to increase the congressional slate of his party in the Chamber of Deputies. If, against forecasts, a runoff was necessary, he could then support Duhalde, but, of course, after obtaining political concessions for him and his party. In fact, De la Rúa won the presidential election

Argentina

119

in 1999 with more than 48 per cent of the vote, against 38 per cent for Duhalde and 10 per cent for Cavallo. Further developments include the forced resignation of president De la Rúa on December 2001 in the midst of violent protests against his government. The party system in Argentina initiated a process of accelerated fragmentation: the Alliance between Frepaso and UCR was dismantled, several candidates of the PJ ran separately for president in 2003 and new parties emerged. In a context of high party fragmentation the compromise reached by Menem and Alfonsín in 1994 was more likely to produce better and less arbitrary outcomes than any of the alternatives considered at the time. A majority runoff system, like the one supported by Alfonsín, would have contained less effectively the process of party fragmentation in the country. The simple plurality rule proposed by Menem could lead to the selection of presidents with only a small share of popular support. This shows that while neither efficiency nor legitimacy may constitute the primary goals of political actors devising institutions, inclusive agreements, even if self-interested, may lead to better results than rules imposed unilaterally by the stronger actor.

Note 1. The PJ was close to that majority in the Senate (it had 30 out of the 32 votes required) but not in the Chamber of Deputies. From a total of 257, the PJ had 117 deputies and the UCR 84. In other words, it needed exactly 54 votes to achieve the two-thirds (171) required by the existing constitution. Even after the incorporation of 9 deputies for the PJ in December 1993 and counting with the support of its most likely allies, centre-right parties like the provincial parties and the UCEDE and rightist parties like MODIN, the PJ would not reach the necessary majority.

References García Lema, Alberto (1994) La reforma por dentro. Buenos Aires: Planeta. Molinelli, Guillermo (1989) Collegios electorales y asambleas legislativas. Buenos Aires: Manantial. Molinelli, N. G., M. V. Palanza and G. Sin (1999) Congreso, Presidencia y Justicia en Argentina. Buenos Aires: Grupo T. Nohlen, Dieter (ed.) (1993) Enciclopedia electoral latinoamericana y del Caribe. Costa Rica: Instituto Interamericano de Derechos Humanos. Novaro, Marcos and Vicente Palermo (1996) Política y poder en el gobierno de Menem. Buenos Aires: Norma.

Further reading • Botana, Natalio and Ana María Mustapic (1991) ‘La reforma constitucional frente al régimen político argentino’, in Dieter Nohlen and Liliana De Riz (eds), Reforma institucional y cambio político. Buenos Aires: Legasa. • Castiglione, Franco (1995) ‘Ingeniería política, reforma constitucional y ballotage’, in Ricardo Sidicaro and Jorge Mayer (eds), Política y sociedad en los años del menemismo. Buenos Aires: Universidad de Buenos Aires. • García Lema, Alberto (1994) La reforma por dentro. Buenos Aires: Planeta.

120

Gabriel L. Negretto

• Mora y Araujo, Manuel (1991) ‘El cuadro político y electoral argentino’, in Dieter Nohlen and Liliana De Riz (eds), Reforma institucional y cambio político. Buenos Aires: Legasa.

• Negretto, Gabriel (1999) ‘Constitution-Making and Institutional Design: the Transformations of Presidentialism in Argentina’, European Journal of Sociology, XL: 2.

• Novaro, Marcos and Vicente Palermo (1996) Política y poder en el gobierno de Menem. Buenos Aires: Norma.

4 Brazil: Democratizing with Majority Runoff Jairo M. Nicolau

Two aspects of the Brazilian electoral system stand out: the first refers to a set of singular features in the operation of the proportional representation system for the election of the Chamber of Deputies. The other concerns the effects of the majority system at the presidential elections – contrary to expectations, the system has not provided an incentive for many parties to present candidates in the presidential elections, producing instead a concentration of the dispute among a reduced number of competitors. The purpose of this paper is to analyse both these dimensions of the Brazilian representative system: proportional representation for the Lower House and the two-round majority election for President. Presently, Brazil uses a majority system with second-round runoff to elect the President. The 513 representatives of the various states or electoral districts in the Chamber of Deputies are chosen by proportional representation – Brazil is the largest democracy in the world using proportional representation. Both the President and Representatives are elected for a four-year term. In addition, there is an Upper House (the Senate), with strong legislative powers, whose members are elected by plurality rule; each state elects three Senators alternately (two at one election and the third at the next) for an eight-year mandate. Voters elect by majority rule with a second-round runoff the governors of the 27 units of the Federation, that is the 26 states and the Federal District. The members of the state legislatures are elected under the same system, using the same rules as the Chamber of Deputies. The 5,559 Brazilian municipalities are also governed by elected mayors, but only cities with more than 200,000 voters use the two-round majority system. All other towns use plurality rule. Each municipality has a legislative branch (municipal chamber) whose members are chosen by proportional representation – 60,277 city-council representatives are elected. In the almost two decades of the present democratic cycle, there has been practically no conflict between the executive and the legislature in Brazil, notwithstanding the fact that the country uses a combination which had been condemned in certain literature on political institutions as the worst possible: proportional representation for the lower house and majority election for the president. The ‘efficient secret’ of the Brazilian representative system should perhaps be sought in the nature of the parties. With the exception of the Workers’ Party (PT), they are pragmatic, 121

122

Jairo M. Nicolau

vote-oriented and with low ideological commitment. Irrespective of his party, if the President is centre-oriented, the parties always seem ready to cooperate.

Proportional representation in Congress Brazil has a long track record of Lower House elections. The first elections for the Chamber of Deputies took place in 1824. From then on, 52 legislatures have been elected. During the Empire (1822–89), the representatives were elected by different forms of majority rule. Up to 1880, there were indirect elections with two levels: first, the voters elected the electors, which in turn chose their representatives in the Chamber of Deputies. Voters were required to be over 25 (21 if they were either married or military officers and of any age if they were clergymen or law school graduates). Slaves and women did not vote, but freed slaves were allowed to vote at the first level. In addition, a minimum income was required to qualify as a voter: 100,000 réis per annum to vote, and 200,000 réis to qualify as an elector – values that were updated in 1846 to 200,000 and 400,000 respectively. From 1881 on, elections for the Chamber of Deputies became direct. When the Republic was proclaimed in 1889, the income requirement for electors was abolished, but the illiterate were still barred from voting. During the first four decades of the Republican era three electoral systems were used (all of them variations of majority rule). The longer-lasting system (1904–30) divided the states into fivemember districts. Voters cast limited ballots for up to four candidates; in addition, they could vote for the same candidate, also known as cumulative vote. During this period, the elections for both President and the Chamber of Deputies were marked by large-scale frauds and low electoral participation. A new Electoral Code adopted in 1932 modernized the voting process and was a first step towards consolidating an electoral democracy in Brazil: women were entitled to vote; Electoral Justice was created and made responsible for organizing voter registration, the elections themselves, the ballot counting and proclamation of the elected officials; and new measures were adopted to ensure vote confidentiality. Up to the 1930s, no party or political movement of significance had defended the adoption of proportional representation in the country, and only a handful of intellectuals defended the idea. Some of them had participated in drafting the 1932 Electoral Code – which ultimately adopted a ‘parallel’ mixed system (part of the representatives elected by proportional representation and part by plurality rule) – the operation of which was extremely complex, particularly in the elections for the Chamber of Deputies. This system was applied in two elections only (1933 and 1934), since the coup led by Getúlio Vargas in 1937 suspended elections and closed both the parties and Congress. Elections were reintroduced in 1945, in the context of a democratization process. During this period, elections for the Chamber of Deputies adopted proportional representation which has remained practically unchanged since then.1 After 21 years under authoritarian rule (1964–85), Brazil returned to democracy. Three key alterations introduced into the Brazilian voting process should be highlighted. The first is the extension of voting rights to the illiterate in 1985.

Brazil

123

Historically, only a handful of democracies ever made requirements of either literacy or a certain degree of education to qualify voters. In Europe, only Portugal restricted voting rights to the literate, a restriction that was suppressed in 1974. In Latin America, the literacy requirement to qualify as a voter was waived in the following order: Uruguay (1918), Colombia (1936), Bolivia (1952), Chile (1970) and Peru (1980). Brazil was the last country to eliminate this restriction and extend the voting rights to the rather significant proportion of illiterates in the country’s population. The second key change was the creation, in 1986, of a computerized roll of all Brazilian electors, which eliminated the possibility of fraud in the registration process. Such fraud had been common at other times in Brazilian electoral history. Finally, the voting process went through a major alteration in 1996, when printed ballots were replaced by electronic balloting (a machine resembling bank computer terminals). The purpose of this change was to eliminate frauds in both the voting and ballot-counting processes, which were still common in the poorest regions of Brazil. Experimentally introduced in 1996, the electronic balloting machines were fully used by all voters in national elections only in 2002. The obligation to both register and vote has been in force since the 1934 Constitution. Presently, registration and vote are mandatory for citizens over 18 and optional for both youths between 16 and 17 and senior citizens over 70. Voters absent from their electoral district or sick may justify their absence either on election day or later at the electoral notary office. The penalty for defaulting voters that do not justify their absence is payment of a fine equivalent to one US dollar, an amount that the judge may increase tenfold in consideration of the economic condition of the violator. If any voter fails to present evidence that he/she either voted, justified his/her absence or paid the corresponding fine, he/she will be prevented, among other penalties, from enrolling in public contests, obtaining both passport and identification cards and, in the case of public employees, receiving their respective payments. The names of voters that fail to vote in three consecutive elections (or to either justify their absence or pay the corresponding fine) are deleted from the roll of voters. Today, when they cast ballots for their representatives at the Chamber of Deputies, voters may enter in the electronic ballot machine the number of either the candidate or the party they choose. The screen will then display either a photo of the candidate or a reference to the chosen party. The walls of all electoral stations are covered with lists with the names and numbers of all candidates competing in the election so that voters may clarify any last-minute doubts. In practice, this voting process gives voters a feeling that the elections operate as a broad competition among candidates rather than as a proportional representation list. Most electors have no idea of the complexity of the system of ballot aggregation and apportioning seats among the competing parties. In broad terms, the major mechanisms of the proportional representation system presently in force in Brazil are as discussed below. Malapportionment of Chamber of Deputies seats The Chamber of Deputies consists of 513 representatives elected for the 27 units of the Federation. At the moment, there is no proportional apportionment between

124

Jairo M. Nicolau

the population of each state and their seats in the Chamber of Deputies. The main reason for malapportionment derives from the constitutional regulation that stipulates a minimum (eight seats) and a maximum (70 seats) per state. Hence, a state that should have one seat apportioned to it receives eight, thus generating distortions. Although all states do reveal a certain degree of distortion, the most marked are, on the one hand, the over-representation of small states of the Northern Region (Roraima, Amapá, Acre, Tocantins), and on the other, the under-representation of São Paulo (presently entitled to 41 seats less than a strictly proportional system would apportion to it). As we can see hereunder, the malapportionment of seats in the Chamber of Deputies also affects partisan representation, since the parties with significant numbers of votes in over-represented states will also tend to be over-represented, while parties with high numbers of votes in under-represented states will tend to lose seats. However, due to the existence of mega-electoral districts (between 30 and 70 representatives), more than half the representatives are elected in five mega-districts: São Paulo, Minas Gerais, Rio de Janeiro, Bahia, Rio Grande do Sul and Paraná. Even the other districts can be considered comparatively medium and large sized (between 8 and 25 representatives). The parties that manage to concentrate their votes in some of the mega-districts ensure a reasonable number of seats in the Chamber of Deputies. Electoral formula and threshold The parties may compete either in separate lists or by forming an apparentement. In order to achieve representation, the party (or apparentement) must exceed the Hare electoral quota, which is computed by dividing the total number of votes cast for the parties and their candidates by the number of seats contended for. In terms of percentage, the electoral quota is the result of dividing 100 per cent by the number of seats contended for in the election, e.g. the state of Acre has eight representatives in the Chamber of Deputies, hence the quotient is 12.5 per cent (100/8). The quota operates as a threshold (in Acre any party or apparentement must receive more than 12.5 per cent of the votes to elect one representative). The computation for seat apportionment is done in two steps. In the first, the total number of votes for a party (or apparentement) is divided by the electoral quota; each party (apparentement) will be entitled to as many seats as the number of times it reaches the quota. In the second step, unoccupied seats are apportioned in accordance with the d’Hondt formula: the total number of votes for each party or apparentement is divided by the number of seats it has already received plus one; the parties with the highest averages receive the seats that were not apportioned in the first step. In allocating seats to the various candidates of a party (or apparentement), Brazil uses the open list system. Each party presents a list of candidates with no order of preference. The seats obtained by the individual parties or apparentements are then apportioned among the candidates that receive the highest number of votes. Hence, the ballots cast for a party, or party vote, only serve to apportion the seats, but do not affect the identification of the candidates to be approved in each list.

Brazil

125

The proportion of party votes was rather small in the last four elections for the Chamber of Deputies: 14 per cent (1986), 18 per cent (1990), 8 per cent (1994), 14 per cent (1998) and 10 per cent (2002). The literature addressing the effects of electoral systems highlights the fact that the open list system favours individuals’ reputations rather than the party’s. In Brazil there is strong evidence of personalized electoral campaigns. Each candidate organizes his/her campaign structure, including participation in events, production of material, collection of resources and accounting for the expenditure, practically independently from the party’s executive committee. Since the party’s performance derives, in great measure, from the aggregate votes obtained by successful candidates, when they organize the list of candidates, they have a strong motivation to include people who are popular in their respective fields of activity, but not necessarily with a track record of involvement in partisan activity (including artists, soccer players, radio speakers, religious leaders, etc.). The very mechanism of casting ballots for the Chamber of Deputies, by which the voters enter in the electronic ballot machine the number of their chosen candidate but receive no information whatsoever on the apparentements or the candidates on the list, reinforces voter perception that the election is a dispute between names, not parties. There is no research on the degree of information the voters receive about the ballot-counting mechanism. However, there is a widespread perception that the elected candidates would be the individuals that receive the highest number of votes, as it would be with single districts by majority rule. An IUPERJ survey of voters in the city of Rio de Janeiro during the 1994 electoral campaign showed the weight of individual reputations in elections for the legislature. Asked about the procedure to select candidates to the Chamber of Deputies, 74 per cent revealed they voted for the candidate, independent of his/her party; 14 per cent said they chose the party; and only 7 per cent said they voted exclusively for the party (Nicolau, 1994). In practice, as no new computation is made to apportion the total number of seats of an apparentement in proportion to the contribution of each party, the apparentements operate as true political parties. This procedure produces two important effects. The first is that any member party of an apparentement considers that the most important goal is to ensure that their candidates rank among the first on the list, independently of their specific individual contribution to the ultimate number of votes of the apparentement. Hence, some aberrations may occur: party A, a member of the AB apparentement, may elect one representative with less votes than party B, and the latter might end up electing no candidate; furthermore, party A may elect a representative even if it receives fewer votes than party C, which did not participate in any apparentement and did not reach the threshold. The second effect is that, when voters cast their ballots for a party that participates in an apparentement, their votes contribute to the ultimate result of the apparentement but are not specifically counted as the party’s. Hence, in the end, the voter’s intention of favouring a given party is ultimately violated by the mechanisms of the Brazilian electoral system. Table 4.1 shows results of this type for the Chamber of Deputies elections in 1998 in the electoral district of Minas Gerais. However, it

126

Jairo M. Nicolau

Table 4.1 Brazil: example of seat allocation – Minas Gerais, elections for the Chamber of Deputies, 1998 Parties/apparentements

% votes

No. seats

% seats

PT PDT PC do B Three other Total

11 4 1 1 17

7 1 1 0 10

13 2 2 0 17

PSDB

24

14

26

PMDB PST

16 2

9 1

17 2

Total

18

10

19

PFL PPB PTB

14 8 8

8 7 2

15 13 4

Total

30

17

32

5 1 6

3 0 3

6 0 6

5 100

0 54

0 100

PL Two other Total Other Total Total of valid votes (V): 7,211,593 Seats to be filled (C): 53 Threshold: V/C: 136,067 (1.8%)

Note: The numbers show how the apparentement mechanism can produce marked distortions: two parties, the PC do B (0.6%) and the PST (1.6%) elected representatives despite the fact that they received less votes that the electoral quota (1.9%); both the PDT (4.0%) and the PC do B (0.6%), with highly different numbers of votes, got one seat each; the PL (5.2%) got three seats, despite the fact that their total number of votes was smaller than that of the PTB (6.5%), which got only two seats; the PSB (0.8%) was a member of the same apparentement as the PC do B, but, even receiving more votes, the latter did not elect any candidate, while the former elected one representative.

is important to stress that, in view of its randomness, it is impossible to foresee the direction the distortions will take. The mechanism of apparentement can produce random results when we compare both votes and seats of each party in each electoral district. An important aspect is that, up to 1998, the parties were entirely free to enter into a different apparentement in each electoral district. Consequently, it was rather common to find parties that were opponents in the presidential election forming apparentements in certain electoral districts to vie for the Chamber of Deputies. In the 2002 elections, a decision of the Electoral Supreme Court (Tribunal Superior Eleitoral – TSE) stipulated that, in each electoral district, parties that form an apparentement in the presidential elections could only form apparentements between themselves or with parties not apparented in presidential elections in the elections for the Chamber of Deputies. It is important to stress that it is not mandatory that the apparentements cover all electoral districts. As an example,

Brazil

127

an apparentement between three parties (A, B and C) in the presidential elections may take on different configurations in different electoral districts: A and B apparented and C running alone; A and C apparented and B alone; BC apparented and A running alone; A, B and C forming an apparentement, thus reproducing the alliance made in the presidential elections. Another important aspect is that the parties that do not take part in apparentements in the presidential elections are free to form any apparentement they wish in the states. Thus party D (with no candidate running for President) may form an apparentement with parties A, B, C or E and F (with no presidential candidate) in another district. Today, Brazil has one of the most fragmented lower houses in the world. The average value of the effective number of parties for the decade from 1990 is 8.0. The Brazilian Chamber of Deputies is only less fragmented than Papua New Guinea’s (N = 10.8) and Belgium’s (N = 8.5) (Farrell, 2001). In the first elections that took place during the democratic period, in 1986, fragmentation was low (N = 2.9). But in the 1990 elections there was a marked increase in the effective number of parties (N = 8.7). Despite slight declines in 1994 (N = 8.2) and 1998 (7.1), it increased again in 2002 (N = 8.4) to a high level in comparison with other democracies. What could explain the high fragmentation of the Brazilian partisan system? Two electoral attributes play a key role. The first is the very high district magnitudes mentioned above. The second is the regulation that rules that apparentement votes should be computed. As far as I know, it is only used in Brazil. As shown in Table 4.2, this system enables parties that do not reach the electoral quota (or that would not reach it if the votes were apportioned out within the apparentement) to elect representatives. Several small parties owe their representation in the Chamber of Deputies to the apparentement regulation. The disproportion between a party’s votes and seats – measured by the LoosemoreHanby index (Index D) – is lower than could be expected in a representative system with electoral districts the size of Brazil’s. In the first five elections for the Chamber of Deputies the average was 9.4 per cent. Three aspects of the electoral system contribute to increase the disproportion between a party’s votes and seats. The first is the malapportionment of the seats in the Chamber of Deputies between the various states: the parties whose votes are concentrated in São Paulo (the under-represented state) lose seats, while the parties whose votes are concentrated in the small states of the Northern Region end up over-represented. The second factor is the apparentement regulation, which, as we have seen in the example above, produces rather unpredictable results in terms of party representation. Finally, the combination of the d’Hondt formula with the threshold provision affects party representation, especially in the states with a lower number of representatives.

Majority runoff for President From the very moment the Republic was proclaimed (1899), Brazil adopted presidentialism. The first republican Constitution (1981) established that both the

128

Table 4.2

Brazil: Chamber of Deputies electoral results, 1982–2002

Year

1982 1986 1990 1994 1998 2002

Centre

Left PCdoB/PCB v/s

PT v/s

PSB v/s

PDT v/s

PPS v/s

1/1 2/1 2/2 3/1 5/5

4/1 7/3 10/7 13/10 13/11 18/18

1/1 2/2 2/3 3/4 5/4

6/5 6/5 10/9 7/7 6/5 5/4

1/1 1/1 1/0 1/1 3/3

PV v/s

0/0 0/0 1/1

Right

Others

PSDB v/s

PMDB v/s

PPB/PPR/PDS v/s

PL v/s

PTB v/s

PFL v/s

9/8 14/12 18/19 14/14

43/42 48/53 19/21 20/21 15/16 13/14

43/49 8/7 9/8 9/10 11/12 8/10

3/1 4/3 4/3 2/2 4/5

4/3 4/4 6/8 5/6 6/6 5/5

18/24 12/17 13/17 17/20 13/16

PP v/s

7/7

v/s

3/1 16/15 3/2 5/3 6/1

Note: PCdoB: Communist Party of Brazil. PCB: Brazilian Communist Party. PT: Workers’ Party. PSB: Brazilian Socialist Party. PDT: Democratic Workers’ Party. PPS: Popular Union Party. PV: Green Party. PSDB: Brazilian Social Democratic Party. PMDB: Party of Brazilian Democratic Movement. PPB: Brazilian Progressive Party. PPR: Republican Progressive Party. PDS: Party of Social Democracy. PL: Liberal Party. PTB: Brazilian Workers’ Party. PFL: Party of the Liberal Front. v/s: percentage of votes/ percentage of seats.

Brazil

129

President and Vice-President would be elected directly by the people in independent ballots and should have a majority of votes. If no candidate obtained a majority, Congress should choose among the candidates with the most votes – a fact that never occurred, since all Presidents were elected by a large margin in poorly competitive elections. Both the President and Vice-President had four-year mandates and could not be re-elected for a subsequent term. Between 1891 and 1930 eleven Presidents were elected. Presidential elections were reinstated only after the country’s redemocratization in 1945. In the first democratic period (1945–64), presidential elections were held by plurality rule. The president’s term in office was five years. Elections for vice-president continued to be separate from the presidential election. Hence, it was possible, for instance, to elect a governmental candidate for President and one from the opposition for Vice-President. In three of the four presidential elections held in the period, the President-elect received less than 50 per cent of the votes: Getúlio Vargas (PTB) was elected with 49 per cent in 1950; Juscelino Kubitschek obtained only 36 per cent in 1955 and Jânio Quadros (PDC/UDN) 48 per cent of the votes. The only exception was Eurico Dutra (PSD), who obtained 55 per cent of the votes in 1945. Despite the fact that plurality rule was stipulated in the Constitution, these results never failed to produce criticism, as defeated parties questioned the legitimacy of the results. In 1950 and 1955 the UDN (a centre-right party) campaigned against the instatement of presidents elected, arguing that they had not obtained a majority in the ballots. In 1985, when Brazil returned to democracy, discussions over which rules to adopt for presidential elections became an important part of the political agenda. The 1988 Constitution opted for the majority rule with a second-round runoff between the two most voted candidates. The Constitution barred re-election and defined a five-year presidential term of office. The major argument in the defence of the two-round rule in Brazil was the guarantee that the elected President would have significant support among the people. The criticism faced by the plurality rule system in the 1946–64 period – especially the possible election of the President by a minority of the voters – was a strong motivation for the delegates who approved the Constitution to opt for the two-round system. Of the four presidential elections that took place in the period, two required a second round. In 1989, Fernando Collor (PRN) obtained only 30.5 per cent and Luiz Inácio Lula da Silva (PT) 17.2 per cent of the votes; in the second round, Collor received 53 per cent of the votes. In 1994 and 1998, Fernando Henrique Cardoso won the elections in the first round, with 54 per cent and 53 per cent of the votes, respectively. In the 2002 elections, candidate Luiz Inácio Lula da Silva (PT) almost won in the first round (46 per cent of the votes); in the second round he obtained 61 per cent of the votes. Two amendments to the 1988 Constitution approved in the 1990s changed the rules of Brazilian presidentialism. The first, sanctioned in June 1994, reduced the President’s term to four years. This measure was aimed at increasing the connection between the votes obtained by the President’s party (or apparentement) and the representation of the parties in the Chamber of Deputies. The experience of two

130

Jairo M. Nicolau

Presidents elected with a weak link to political parties (Jânio Quadros in 1960 and Fernando Collor in 1989) in non-concurrent presidential elections, both with weak support in parliament, was a strong motivation to reduce the presidential term of office. Since then, the presidential elections have been conducted simultaneously with the elections for the National Congress, state governors and state legislatures. A second Constitutional Amendment, sanctioned in June 1997, allowed for the re-election of chief executives (president, governors and mayors) for one consecutive term. This enabled President Fernando Henrique Cardoso and the governors elected in 1994 to compete again in the 1998 elections. In the 1989 presidential elections – the first after the new Constitution was promulgated – all relevant parties presented candidates. If we calculate the effective number of presidential candidates, we will see a broad dispersion of voters’ preference (Nc = 5.7). In the following three elections the dispute was concentrated around a lower number of candidates. Measured by the effective number of candidates, the dispute was clearly concentrated: 1994 (2.7), 1998 (2.5) and 2002 (3.2). Therefore, only the 1989 election presented a standard closer to what should be expected from a two-round election: in the first round, the larger parties presented candidates; after the first round alliances were built with the defeated parties around the candidates that were elected for the second round. In the following elections, the dispute was polarized between two apparentements, led by the centrist Party of Brazilian Social-Democracy (PSDB) and the leftist Workers’ Party (PT) – the latter competing in all disputes with the former blue-collar metallurgist Luiz Inácio Lula da Silva. Only more in-depth research could provide us with more definitive explanations of the reasons why presidential elections have taken on the aspect of a plurality system rather than a majority election. My suggestion is that, in addition to factors linked directly with the electoral system (mechanical and psychological effects), the number of competitors in the presidential election might be affected by exogenous factors (nature of the parties, federalism and campaign costs). We are still trying to find out why several significant parties have been refusing to present candidates for President, preferring to participate in apparentements with others or, in some cases, simply not to run for the office at all. Some hypotheses may be advanced. The first is the strongly decentralized characteristics of the Brazilian parties, derived from the country’s federalist structure. Many leaders of such parties have local significance but are incapable of projecting themselves as national politicians. Consequently, they prioritize political survival in their state of origin. The second is the high cost of the campaign. The Brazilian electoral campaign is one of the most expensive in the world. To give an idea, the 1998 campaign cost was – in reported values – $330 million. As such funds are obtained almost exclusively from the business community, only candidates with some electoral feasibility can obtain donations that enable them to make a minimally competitive campaign.

The ‘efficient secret’ The experience shows that the Brazilian representative system operates with two relatively autonomous spheres. On the one hand is the extremely high

Brazil

131

fragmentation of the Chamber of Deputies. On the other is the concentration of the presidential dispute around a reduced number of competitors. The autonomy between both spheres is so intense that, for instance, the Party of the Liberal Front (PFL), which obtained the largest number of seats in the Chamber of Deputies in 1998, did not even participate in an apparentement in the 2002 presidential elections. The obvious question is: what does this imply in terms of forming a parliamentary majority for the President? Due to the high fragmentation in the Chamber of Deputies, there is only a very small possibility of the President’s party obtaining a legislative majority. In the four presidential elections, the electoral apparentement that supported President Fernando Henrique Cardoso obtained such a majority only in 1998, at the time of his re-election, when it obtained 58 per cent of the seats. Both Fernando Collor’s (elected in 1989) and the first of Fernando Henrique Cardoso’s administrations (elected in 1994) had to enter into post-electoral alliances to approve their programmes in Congress. Collor faced greater difficulty because he competed in a non-concurrent presidential election, running for a small party that had only 4 per cent of the seats in the Chamber of Deputies. In 2002, the electoral apparentement that supported Luiz Inácio Lula da Silva obtained no more than 25 per cent of the seats in the Chamber of Deputies and had to incorporate other parties to support his administration. Thus, in spite of having what some authors have labelled the worst possible combination of electoral systems – congressional proportional representation and presidential majority runoff – the Brazilian representative system has been able to produce very low levels of inter-institutional conflict.

Note 1. Only two significant changes have been made since then: the criterion for apportioning seats not occupied in the first allocation (1950) and the exclusion of blank votes when computing the electoral quotient (1998).

References Nicolau, Jairo Marconi (1994) Multipartidarismo e Democracia: um estudo sobre o sistema partidário brasileiro (1985–1994). Rio de Janeiro: Fundaçao Getúlio Vargas Editora. Farrell, David (2001) Electoral Systems: A Comparative Introduction. London and New York: Macmillan – now Palgrave.

Further reading • Ames, Barry (2001) The Deadlock of Democracy in Brazil. Ann Arbor: University of Michigan Press.

• Carey, John and Matthew S. Shugart (1995) ‘Incentives to Cultivate a Personal Vote: A Rank Ordering of Electoral Formulas’, Electoral Studies, 14, 4: 417–39.

• Graham, Richard (1990) Patronage and Politics in Nineteenth-Century Brazil. Stanford, CA: Stanford University Press.

• Lavareda, Antonio (2000) A democracia nas Urnas: o Processo Eleitoral Brasileiro. Rio de Janeiro: Revan.

132

Jairo M. Nicolau

• Leal, Víctor Nunes ([1949] 1986) Coronelismo, Enxeda e Voto. São Paulo: Alfa-Ômega. • Mainwaring, Scott (1999) Rethinking Party Systems in the Third Wave Democratization: The Case of Brazil. Stanford, CA: Stanford University Press.

• Nicolau, Jairo Marconi (1999) Sistemas Eleitorais: Uma Introdução. Rio de Janeiro: Fundaçao Getúlio Vargas Editora.

• Nicolau, Jairo Marconi (2002) História do Voto no Brasil. Rio de Janeiro: Jorge Zahar. • Porto, Walter Costa (2002) O Voto no Brasil: da Colônia à Quinta República. Rio de Janeiro: Topbooks.

• Santos, Wanderley Guilherme dos (1987) Crime e castigo: partidos e generais na Política Brasileira. Rio de Janeiro: Vértice.

• Tavares, José Giusti Tavares (1998) ‘O Sistema Político Brasileiro’, in José Giusti Tavares and Raúl Rojo (eds), Instituições Políticas Comparadas dos Países do Mercosul. Rio de Janeiro: Fundaçao Getúlio Vargas Editora.

5 Costa Rica: Modifying Majoritarianism with 40 per cent Threshold Fabrice Lehoucq1

Allocating elected posts to candidates that get the most votes has intuitive appeal. Yet, several decades of research (e.g. Colomer, 2001; Powell, 2000) indicate that plurality elections can generate perverse results. In races featuring three or more candidates, a determined minority can triumph and exclude the median voter – the voter at the very centre of the electorate. Institutional engineers have experimented with alternative ways of selecting presidents in separation of powers systems (Shugart and Taagepera, 1994). A variant of the majority-runoff system so popular among reformers of presidential systems is the qualified-plurality system that Costa Ricans use to elect their presidents. Since 1936, citizens of this Central American republic have elected presidents that obtain a plurality of votes as well as 40 per cent of the valid votes (total votes minus annulled and blank ballots) in quadrennial elections held on the first Sunday in February. Should no candidate satisfy both conditions, the first two runners-up compete in a popular election on 1 April. Citizens and politicians alike agree that 40 per cent qualified-plurality with a second-round runoff is a great way to elect presidents. Conflicts about political succession in the 1940s and the 1950s never included criticism of this rule (Lehoucq, 1991; Bowman, 2002). The 1949 Constituent Assembly – the most comprehensive review of the country’s institutional architecture – ratified this reform of the 1871 constitution with virtually no debate (ANC, 1953–7). Subsequent institutional engineers have never proposed scrapping this system. Since its inception, the median voter has seen his candidate elected to the presidency four-fifths of the time. And, as Table 5.1 shows, only in 2002 did this system fail to select a president in the first round – the key and largely successful aim of institutional engineers. So, if stability and citizen satisfaction are important criteria, the 40 per cent threshold has been an unalloyed success. This chapter explains the development of the 40 per cent qualified-plurality with second-round runoff system. It makes the point that this system was not built overnight. It took at least a decade to perfect and is arguably the product of more than 30 years of experimentation and evolution about how to ensure the peaceful exchange of executive power. This article also suggests that each of its principal 133

134

Fabrice Lehoucq

Table 5.1

Costa Rica: presidential selection, 1882–2002

Election date

11–12/1889 2–4/1894 11–12/1897 (fraud) 12/1901, 4/1902 8/1905, 4/1906 8/1909, 4/1910 4/1914 (Congress) 1917 (coup) 12/1919 12/1923 2/1928 2/1932 2/1936 2/1940 2/1944 5/1948 (revolt) 11/1949 (2/1948) 7/1953 2/1958 2/1962 2/1966 2/1970 2/1974 2/1978 2/1982 2/1986 2/1990 2/1994 2/1998 2/2002

President

José-J. Rodríguez Rafael Iglesias Rafael Iglesias Ascención Esquivel Cleto González Ricardo Jiménez Alfredo González Federico Tinoco Julio Acosta Rafael Jiménez Cleto González Ricardo Jiménez León Cortés Rafael A. Calderón Teodoro Picado José M. Figueres Otilio Ulate José M. Figueres Mario Echandi Francisco Orlich José J. Trejos José M. Figueres Daniel Oduber Rodrigo Carazo Luis A. Monge Oscar Arias Rafael A. Calderón José M. Figueres Miguel A. Rodríguez Abel Pacheco

Party

PCon PC PC PUN PUN PR PR PP PC PR PUN PRN PRN PRN PRN PSD PUN PLN PUN PLN UN PLN PLN UC PLN PLN PUSC PLN PUSC PUSC

Votes % 1st

2nd

81 25 100 71 41 71 – – 89 42 59 47 60 84 65 N/A 55 65 46 50 50 55 44 51 59 52 51 50 47 39

89 53 100 79 92 91 – – – – – – – – – – – – – – – – – – – – – – – –

Median voter Yes Yes No Yes No Yes No No Yes Yes Yes Yes Yes Yes Yes N/A Yes Yes No Yes Yes Yes No Yes Yes Yes Yes Yes Yes No

No. of candidates 4 2 1 2 5 2 5 – 2 3 2 4 3 2 2 N/A 2 2 3 4 2 5 8 8 6 6 7 7 13 13

Key: PC: Civil Party; PCon: Constitutional Party; PLN: National Liberation Party; PR: Republican Party; PRN: National Republican Party; PSD: Social Democratic Party; PUSC: United Social Christian Party; PUN: National Union Party; UC: United Coalition; UN: National Unification Party. This table does not include the interim presidencies of Juan B. Quirós (1919), Francisco Aguilar (1919), León Santos (1948), each of which held office for no more than a week. Sources: Lehoucq (1992), Molina (2001), Obregón (2000) and Ocontrillo (1982).

components, a majority-runoff system (1926) and the 40 per cent threshold (1936), did not involve much or any political bargaining. Oddly enough, parties that typically supported reforms if they promoted their re-election interests backed two reforms to advance the stability of the republic. Indeed, the establishment of a 40 per cent qualified-plurality system with second-round runoff by 1936 was

Costa Rica

135

part of a much broader struggle to establish rules of political succession acceptable to a fluid and intensely competitive party system, one that gradually eliminated the role of intermediate, representative bodies in presidential selection. This finding suggests that durable institutions take time to construct and that they are not always the product of foresight or clever calculation.

Constitutional provisions and political realities The 1871 constitution established a majority rule in Electoral Assemblies (College) with runoff in the Congress system for electing the President. According to paragraph 2 of article 73, the candidate that obtained an ‘absolute majority’ of votes in the Assemblies would become President. Should no one satisfy this requirement, Congress would select the candidate from ‘among the two individuals that obtained the largest number of votes’ (Peralta, 1962: 472). Elections were held every four years; though men could run for re-election, they had to wait a term before trying to become chief executive. Who was a citizen? Who did Congress and the President represent? According to the 1871 constitution, the electorate consisted of all men, 20 years or older, that met a vaguely worded property requirement. In an export-oriented economy (one that produced coffee and bananas) that contained large numbers of small- and medium-sized property holders, virtually all adult men satisfied suffrage requirements. Indeed, by 1913, the numbers of men registered to vote equaled census-based estimates of the size of the adult male population (Molina, 2001). Until 1913, however, adult men did not vote directly for the President or deputies; the citizenry chose electors every four years through the public ballot. Meeting in provincial Electoral Assemblies, electors selected presidents as well as deputies. Qualifications for becoming an elector were stiffer: men had to be at least 25 years old, literate and own property. So, the voice of the people could be heard as filtered through a collection of intermediary bodies that could and, as we now know, did alter the people’s instructions on Electoral Assembly day. Until 1909, citizens voted publicly in November and their representatives selected executives and legislators in February. Congress consisted of deputies serving four-year terms that could stand for re-election in one of seven provinces. Half ran for office alongside the President (and on separate ballots). The other half ran in midterm elections. The provincial Electoral Assemblies elected two years earlier voted for this half. Until 1893, Assemblies selected deputies through plurality rule. Between 1893 and 1913, Assemblies used proportional representation in provinces sending three or more deputies to Congress or absolute majorities in provinces sending one or two deputies to Congress. This system remained in effect between 1913 and 1946 with one change: candidates only needed pluralities to become deputies. Between 1913 and 1948, proportional representation allocated three-fourths of all seats. Since 1948, closed-list proportional representation allocates all seats to parties. (Before 1948, seats left over went to the parties with the largest remainders and that had already

136

Fabrice Lehoucq

won a seat in the relevant province; since 1948, there has been a 50 per cent threshold for divvying up these seats.) Even with indirect elections, Congress could claim that it echoed the voices of the people. In addition to choosing the President if no candidate obtained an absolute majority of the vote, Congress also selected two Designates to the Presidency – vice-presidents, in effect, that could fill the presidency if it was left vacant. Despite the possibility that agents – electors and deputies – had ample room to interpret the people’s will, apparently this was far from a ‘bad’ constitutional blueprint. After all, it made the President dependent upon currying the favour of citizens, electors and legislators. However, in practice, the President came to dominate the Congress because executives manipulated elections to ensure that loyal majorities certified elections to advance his partisan interests. Like many other nineteenth-century constitutions, Congress certified election results, including that of the President. Presidents could also declare states of siege, which they did 21 times between 1882 and 1909, when Congress amended the constitution to allow executives to suspend individual rights and guarantees only with their approval. So, while the 1871 constitution created the basis for political accountability, it possessed flaws of institutional design that transformed the President into an arbitrary master of the political system. Between 1871 and 1886, incumbents chose their successors who typically ran unopposed – an unsurprising result given Tomás Guardia’s dictatorship (1871–82). In 1890, official candidates began to compete against opposition candidates organized in rival parties – a watershed for the republic. Though President Bernardo Soto (1886–90) reluctantly gave up power to the candidate that got more than 81 per cent of the popular vote, his successors typically manipulated provincial Electoral Assemblies and suspended the constitutional order to repress the opposition on election day. Indeed, President Rafael Iglesias Castro (1894–98) of the Civil Party (PC) got mayors to reform the constitution so that he could stay in office until 1902. Under the threat of opposition revolts, President Cleto González (1906–10) of the National Union Party (PUN) strengthened electoral safeguards (in the form of the 1907 and 1908 amendments to the 1893 electoral law) that led to the opposition Republican Party (PR) victory in the 1909 elections. By peacefully transferring power to the opposition, González restored a reputation sullied in 1906, when his predecessor, Ascención Esquivel (1902–6) of the PUN, jailed a third of all electors to rig his election to the presidency. The cycle of presidential imposition and opposition revolts – between 1882 and 1910 there were 14 insurrections – began to draw to a close with two important constitutional reforms, ones crucial for the subsequent establishment of a 40 per cent qualified-plurality system. In 1913, Republican Party (PR) factions – which held all legislative seats but one – settled their differences in Congress to pass two constitutional amendments. Following guidelines for reforming the constitution (articles 134–5), two-thirds of the total number of deputies voted in favour of establishing direct elections and restricting the executive’s ability to suspend the constitutional order in two different sessions of Congress. Then as now, presidents

Costa Rica

137

did not participate in this process; chief executives simply promulgated these constitutional amendments and could do little more than issue ‘observations’ that Congress could ignore. The solution to imperial presidents was to ban an intermediate, representative body. If the president could manipulate provincial Electoral Assemblies to engineer the imposition of his successors, constitutional reformers opted to empower the electorate – virtually all adult men – to select the chief executive directly. Though the Republican-dominated Congress also issued a new electoral law which did little more than consolidate the reforms of the previous six years, it left the classical constitutional theory of electoral governance intact. The executive remained responsible for the organization of elections and Congress still had to certify the results of executive and legislative elections.

Establishing a majority runoff system (and debating thresholds) in 1926 Thirteen years later, Congress made the electorate completely responsible for selecting the President, a reform that President Ricardo Jiménez promulgated in May 1927. Henceforth, the old Congress would simply convene, in an extraordinary session on 1 March, a runoff between the two runners-up. Why did Congress amend the constitution to remove itself from electing presidents? The legislative transcript indicates that it had nothing to do with re-election concerns. No one accused anyone of trying the tilt the electoral playing field in any party or candidate’s favour. No one proposed retaining the congressional runoff system because politicians associated it with instability and shady, backroom dealings. Even though Congress was embroiled in typically contentious negotiations with the President, a two-thirds majority took up his suggestion to dismantle the congressional runoff system because it seemed to be the most efficient and fairest way of rotating control of the presidency. Everyone was in favour of a popular runoff between the two candidates with the largest pluralities. Debate revolved around the nature of thresholds in a handful of sessions in August 1925 when the Agricultural Party (PA) delegation – holding 49 per cent of the seats – and the Republican contingent – holding 37 per cent of the seats – each presented reform bills. Republican Party (PR) Deputy (and former Congressional President) Arturo Volio recommended convening a popular runoff if no one obtained an absolute majority of the vote in the previous election. The presidency, according to this proposal, would then go to the candidate that got a plurality of the vote among all first-round contenders. PA Deputy Francisco Mayorga recommended awarding the executive to the candidate with a plurality and at least a third of the popular vote. Majorities defeated both proposals, in large part because each delegation could not get enough rival deputies to support its bill. Chi-square tests indicate the relationship between partisanship and the vote on the amendments was highly significant (p < 0.005). Consensus emerged around PR Deputy Juan Rafael Arias’s majority runoff proposal. With 33 deputies in favour and 5 against, Congress approved the constitutional

138

Fabrice Lehoucq

amendment in favour of convening a runoff election among the most popular candidates, a race that the candidate obtaining the most votes would win. Unlike the two other proposals, partisanship had nothing to do with support for the measure (p: 0.71). A year later, two-thirds of all Congressmen again backed the amendment before sending it to the president. Accepting President Jiménez’s suggestion, Congress also transferred election day from the first Sunday of December in the year before the president assumed office to the second Sunday of February. Experience encouraged deputies to remove themselves from the process to choose the President. If the people could not settle upon a president in the popular election, Congress decided to restrict its actions to holding a runoff between the most popular candidates. The political class seemed to agree that letting congressional factions negotiate the selection of the president only fuelled partisan rancour and endangered the stability of the republic. Two elections in the previous 13 years gave substance to these views. First, the inability of the system of majority rule with runoff in Congress to produce a legitimate and constitutionally acceptable President in the 1913 general elections triggered a series of events that badly scarred the body politic. None of the five candidates obtained an absolute majority of the vote in an election that featured a 90 per cent turnout rate. Máximo Fernández, the populist of the PR, obtained 42 per cent of the vote. Dr Carlos Durán of the PUN got 31 per cent of the vote, Iglesias got 28 per cent of the vote, and two other candidates got a total of 6 votes. According to a pre-election pact, the PUN and the PC agreed to have their deputies vote for the non-PR candidate that got the most votes. After rumours circulated that the PC was negotiating with the PR to send Iglesias to the presidency and Fernández had ‘resigned’ from his position as first runner-up, the PUN brokered a secret deal whereby Durán would also resign from his position and allow 5 PR and 17 PUN – 51 per cent of all – deputies to select Radical Republican Deputy Alfredo González as First Designate to the Presidency. Durán and his deputies then presented this agreement to President Ricardo Jiménez, who demonstrated his support for the plot by turning over control of the military barracks to González’s backers before Congress met to select the President. This was a decisive move and one that earned Jiménez some harsh criticism; according to time-honoured practice, outgoing presidents only relinquished control of the barracks to the new President after he had taken the oath of office in Congress. Deputy González became president amid accusations of a constitutional coup d’état and as a major economic and fiscal crisis was getting underway. The Allied naval embargo of Germany – the country’s principal purchaser of its coffee crop – during the First World War strangled the economy of export earnings and the government of revenue. Despite the opposition of bankers, merchants and coffee exporters, President González got Congress to approve the establishment of a national bank and an income tax to reduce inflation and increase government revenue. With rumours that he was planning to impose his successor in the 1917 general election (and widespread allegations of executive manipulation during the 1915 midterm elections), virtually the entire political class backed Defense Minister Frederico Tinoco’s coup.

Costa Rica

139

During Tinoco’s rule, a Constituent Assembly scrapped the constitution and produced one calling for a bicameral legislature. Interestingly, constitution-makers, which included some of the most distinguished politicians and jurists in the country, opted largely for staying with the system of majority rule with runoff in Congress for electing the President. According to article 55 of the 1917 constitution, an Electoral College consisting of senators, deputies and their alternates as well as municipal councilmen (regidores) would select presidents for six-year, non-renewable terms. As military expenditures rose to defeat an opposition-led rebellion, his regime became unpopular, especially since President-in-exile Alfredo González was leading a successful effort to prevent US recognition of Tinoco’s regime – and thus starving it of foreign bank loans. In 1919, Tinoco fled into exile after unknown assailants gunned down his brother, Joaquín, the Defense Minister; an ad hoc Committee of Notables then turned power over to ex-President González’s Second Designate to the Presidency, Francisco Aguilar Baquero, who reinstated the 1871 constitution. The second presidential election to go awry – and the immediate backdrop to the 1926 reforms – was that of 1923. Unlike the 1919 general elections, where the voters gave consensus candidate Julio Acosta 89 per cent of the vote, citizens again split their vote among three candidates. Republican Jiménez got 42 per cent of the vote, Alberto Echandi of the PA received 38 per cent of the vote and Jorge Volio of the Reform Party came in a distant third with 20 per cent of the vote. Initial returns indicated that the PA had won enough Congressional seats to catapult Echandi to the presidency. When added to the 12 PA deputies elected in 1921, the 10 seats it was going to win allowed it to assemble a bare majority of 22 deputies – 51 per cent – to elect its candidate to the presidency. During the official tally of the votes in two of the provincial electoral juntas conducted, however, the PA lost two legislative seats, allegedly because of faulty electoral registries. These electoral juntas then reallocated these seats, one to the PR and the other to the Reform Party, which had formed an alliance known as the ‘Fusion’ to thwart the PA. Though named by the President, these juntas contained party representatives that saw the PR and the Reform Party collude to catapult Jiménez to the presidency. Amid rumours of coups, politicians and newspaper columnists struggled to solve the rapidly developing political and constitutional crisis. President Acosta ignored the PA’s request to reverse these two juntas’ decision. He preferred to let Congress fulfil its constitutional responsibility to arbitrate conflicts about electoral law. Since Congress could only convene with two-thirds of its membership present, he might very well have expected the Fusion and the PA to reach an agreement by 1 May when the new Congress met to certify election results and choose the President. Despite efforts to select a compromise candidate, the Fusion relied upon a selfserving (and probably unconstitutional) reading of Congressional procedure to send Jiménez to the presidency. With the PA boycotting Congress, the Fusion used the presence of one PA deputy alternate as well as the PR and Reform Party’s deputies and alternates to establish quorum. Consisting only of Fusion deputies, the Credentials Committee certified the election of its deputies. With approximately half of its total membership missing from the chamber (and in line with the

140

Fabrice Lehoucq

classical theory of electoral governance), Congress then voted in favour of making Jiménez President for the second time in his life.

Establishing the 40 per cent threshold in 1936 While the 1928 presidential election between the PR and the PUN led to Cleto González’s victory (the National Union Party’s candidate, PUN) with 59 per cent of the vote, the 1932 general election ended in another political crisis. Understanding what happened in 1932 is key to understanding why legislators created the first, durable qualified-plurality rule system for electing presidents. In a repeat of the 1913 and 1924 general elections, the 1932 race involved three candidates because of a split within PR ranks. The National Republican Party (PRN) backed Jiménez in his third bid for the presidency. He ran against Manuel Castro of the Republican Union Party (PUR) and Carlos María Jiménez of what was left of the PR. Following new guidelines, Congress had to convene a runoff election between Jiménez, who had attracted 47 per cent of the vote, and Castro, who had the support of 29 per cent of the voters. Less than twelve hours after the closing of the polls, Castro and his followers captured the Bellavista barracks in their attempt to overthrow the government, in a rebellion known as the ‘Bellavistazo’. After several days of shooting and negotiations, the PUR surrendered control of the Bellavista barracks. Humiliated and defeated, Castro resigned his position as second runner-up. Though the coup failed, Congress was in a quandary. Since the constitution called upon the legislature to convene a runoff between the top two runners-up, it was not clear what Congress should do if one of the two most voted candidates walked away from his position, especially if he had been involved in an insurrection. It seemed pointless to convene a runoff with only Jiménez left in the race. So, when Congress met on 1 March, it declared the race for the presidency null and void. On 1 May, when the new Congress convened for the first time, deputies voted Jiménez as First Designate to the Presidency and, because of the absence of a popularly elected executive, it was he who became President. It was against this backdrop that President Jiménez recommended, in his May 1934 Message to Congress, that legislators reform the constitution to award the presidency to ‘the victorious candidate [with] a third, for example, of the vote . . .’ On 5 June 1934, 14 deputies presented a bill that endorsed the President’s proposal without much qualification in a move that seemed to be very focused on avoiding a repetition of 1932. They recommended that the first two runners-up be prohibited from resigning their positions in the race. They also endorsed the proposal that the oldest candidate would get the presidency if two candidates each got the same number of votes (and at least 30 per cent of the votes). After some debate, a Congressional Committee approved the measure, but raised the threshold to 40 per cent of the valid votes. With 35 deputies in favour and 6 against – more than the two-thirds of all deputies required for constitutional amendments – Congress sent the President the 40 per cent bill on 8 August 1934. Following constitutional guidelines, the President sent the bill back to Congress a year later

Costa Rica

141

with the recommendation to lower the threshold to 30 or 35 per cent. Deputies approved the measure, but kept the threshold at 40 per cent. In 1936, after deputies had voted in favour of the bill again, President Jiménez promulgated this innovative reform. Reform of the constitution had nothing to do with re-election concerns. Neither the legislative transcript nor the newspapers contain accusations that the President or any politician sought to extract advantage from a qualified-plurality system. The constitution prohibited presidents from running for consecutive re-election, a ban that Jiménez and virtually all other politicians respected and whose repeal was simply not on the table. Records also indicate that there is no reference to results of the recent midterm Congressional elections or to the next general elections, still two years away. The rise of the Communist Party might be interpreted has having prompted legislators to establish the threshold. Indeed, US State Department correspondence suggests that legislators enacted a constitutional provision to create mandatory voting – with which the 40 per cent threshold became bundled – to reduce the vote share of the Communists. As the share of non-Communist voters increased, the share of votes Communist could win would necessarily fall. Between 1934 and 1942, a respectable 12 per cent the electorate voted for the Communists, a significant minority in a staunchly petit bourgeois society. So, it is conceivable that preventing a runoff was part of the strategic calculations that PR factions made to prevent a determined minority from snatching the presidency in a popular runoff. While plausible, there is no written evidence that anti-communism motivated politicians to adopt the 40 per cent threshold. A third more plausible explanation argues that deputies were trying to maximize efficiency from behind a Rawlsian veil of ignorance. By efficiency, I mean a costless and transparent way of selecting the chief executive. Neither the elimination of Electoral Assemblies in 1913 nor the removal of Congress from the presidential runoff in 1926 had eliminated internecine conflict from presidential succession. So, in debates free of partisan rancour, deputies relied upon an interpretation of the past that blamed intermediate, representative bodies for the instability associated with electing presidents. In the words of the legislative sponsors of the qualified-plurality bill, ‘what matters is that the popular will not be frustrated by the bastard intervention of the powers created to serve it; the authority of the head of state should be sustained by public opinion, and that it is served by its devotion to serving national interests’ (emphasis my own). Most evidence with the veil of ignorance argument is consistent with legislative behaviour. No one challenged the theory of corrupt, representative bodies. The President and deputies endorsed it unanimously. Yet, the smooth Congressional selection of Jiménez as First Designate suggests that representative bodies could solve the impasses that the popular election of presidents could generate. How do we square this finding with the overwhelming consensus in favour of a less than 50 per cent threshold for presidential elections? Taking account of the role of partisan interests is the way to solve this conundrum. It is probably not accidental that virtually all deputies – 90 per cent

142

Fabrice Lehoucq

of them – belonged to PR factions, 64 per cent of which belonged to the PRN. While deputies were trying to maximize efficiency, they were undoubtedly trying to solve a Republican coordination problem. Even though President Jiménez and the Congressional Special Committee recommended a 30–35 per cent floor in 1934 (a suggestion the president repeated in his May 1935 Message to Congress), deputies quickly endorsed a 40 per cent threshold for two reasons. First, once PUR Deputy Chacón noted on 6 August 1934 that winning candidates had always obtained at least 40 per cent of the vote in three-man races, debate on this matter rapidly ended. Second, opposition Deputy Otilio Ulate (later President between 1953 and 1958) argued on 30 July 1934 that a 30 per cent threshold might let a minority – what Ulate called ‘a plutocratic group’ – win the presidency, especially if three or more candidates ran for the presidency. Solving Republican coordination problems also resolves the inconsistency between espousing the theory of corrupt, representative bodies and the legislative solution to the 1932 Presidential election. The political class seems to have interpreted the peaceful outcome of the election as an example of the virtues of popularly elected presidents. Jiménez, after all, had gotten within three points of an absolute majority of the vote and obtained more than 18 points more than his nearest rival. Having just become President for the third time in his life, Jiménez was arguably the embodiment of the statesman executive, the elected official who could rise above the partisan fray to represent the true interests of the republic. It was the electorate, many apparently reasoned, that had elected Jiménez; Congress simply ratified the popular will when it declared him President. So, in the end, deputies were not completely behind a veil of ignorance. Thanks to PUR Deputy Chacón’s quick calculations, Republicans figured out that lowering the bar could facilitate their triumph in presidential elections, the next of which was scheduled for 1936. And by not lowering it to a third of the popular vote, they could prevent influential minorities from depriving them of victory. To the extent that this had anything to do with the development of a Communist Party, as the second explanation for the 40 per cent rule contends, Republican deputies were simply trying to ensure that a determined minority could not prejudice their chances of coming to power.

Conclusions Over the course of 30 years, politicians gradually reformed the majoritarian procedures for electing their chief executive. They went from a system where Congress selected the president should no one obtain an absolute majority of the vote to one where the electorate single-handedly chooses the president. In 1926, legislators established a popular runoff for the presidency. In 1936, they decided to reform the constitution to award the executive to the candidate that obtains a plurality and at least 40 per cent of the vote. The establishment of the first, durable qualified-plurality system was therefore a product of institutional evolution within an intensely competitive party system. Unfamiliar with political theory, Costa Rican politicians tinkered with reforms

Costa Rica

143

that gradually removed the legislature from helping to select the chief executive. Impartiality seems to have motivated deputy behaviour most of the time; an interest in eliminating conflict from political succession – an outcome that benefited no party – prompted legislators to let the electorate choose the president. Yet partisan interests were not completely absent from their decision-making. Though surviving records do not speak on this issue, it is hard to avoid the conclusion that the dominant party at the time, the PR, was trying to solve a coordination problem when it gave the final push to establish a 40 per cent threshold in 1936. Only in 2000 did this system fail to elect a president in the first round when the right-of-centre Abel Pacheco of the United Social Christian Party (PUSC) was forced into a popular runoff with Rolando Araya of the National Liberation Party (PLN) because the PLN split and fielded two candidates in the presidential contest. A related and apparently unintended consequence of this reform has been to foment a bipartisan party (Shugart and Carey, 1992: 288–92). Indeed, for much of the post-1949 period, analysts agree that the Costa Rican party system has been polarized around two electoral blocs (Hernández, 1998; Lehoucq, 1997; Rovira Mas, 2001; Wilson, 1998), even as they disagree whether this party system is a strict two-party system. Though presidents have been elected with the support of the median voter 82 per cent of the time since 1936, the system has done less well on this criteria than on ensuring the smooth and transparent election of the chief executive. This should come as no surprise because maximizing the interests of the median voter was not foremost on legislator’s minds. Nevertheless, in their efforts to eradicate conflict from political succession and to solve partisan coordination problems, reforms stumbled upon a formula that maximizes both social welfare and political stability.

Note 1. Josep M. Colomer, Gabriel Negretto and Juan-Carlos Rodríguez provided incisive comments on earlier drafts, for which I am thankful. I am particularly grateful to Juan-Carlos for his encouragement and invaluable help with primary sources.

References ANC (Asamblea Nacional Constituyente) (1953–57) Actas de la Asamblea Nacional Constituyente, 3 vols. San José: Imprenta Nacional. Bowman, Kirk S. (2002) Militarization, Democracy and Development: The Perils of Praetorianism in Latin America. University Park, PA: Pennsylvania State University Press. Colomer, Josep M. (2001) Political Institutions. Democracy and Social Choice. New York and Oxford: Oxford University Press. Hernández Naranjo, Gerardo (1998) ‘El sistema de partidos en Costa Rica, 1982–1994, análisis sobre el bipartidismo’. MA Thesis, Central American Program in Sociology, University of Costa Rica. Lehoucq, Fabrice (1992) ‘The Origins of Democracy in Costa Rica in Comparative Perspective’. PhD dissertation, Duke University.

144

Fabrice Lehoucq

Lehoucq, Fabrice (1997) Lucha electoral y sistema político en Costa Rica, 1948–1998. San José: Editorial Porvenir. Molina, Iván (2001) Estadísticas electorales de Costa Rica (1897–1948): una contribución documental’, Revista Parlamentaria, 9: 354–67. Obregón, Clotilde (2000) El proceso electoral y el Poder Ejecutivo de Costa Rica. San José: Editorial de la Universidad de Costa Rica. Ocontrillo García, Eduardo (1982) Un siglo de política costarricense: crónica de 24 campañas presidenciales, 2nd edn. San José: Editorial Universidad Estatal a Distancia. Peralta, Hernán G. (1962) Las constituciones de Costa Rica. Madrid: Centro de Estudios Constitucionales. Powell, G. Bingham (2000) Elections as Instruments of Democracy. Majoritarian and Proportional Visions. New Haven, CT: Yale University Press. Rovira Mas, Jorge. (2001) ‘¿Se debilita el bipartidismo?’, in Jorge Rovira Mas (ed.), La democracia de Costa Rica ante el Siglo XXI. San José: Editorial de la Universidad de Costa Rica. Shugart, Matthew Soberg, and John M. Carey (1992) Presidents and Assemblies: Constitutional Design and Electoral Dynamics. Cambridge and New York: Cambridge University Press. Shugart, Matthew Soberg and Rein Taagepera (1994) ‘Plurality vs. Majority Election of Presidents: a Proposal for a “Double Complement Rule”’, Comparative Political Studies, 27: 323–48. Wilson, Bruce M. (1998) Costa Rica: Politics, Economics and Democracy. Boulder, CO: Lynne Rienner.

Further reading • Carey, John M. (1996) Term Limits and Legislative Representation. Cambridge and New York: Cambridge University Press.

• Lehoucq, Fabrice (1991) ‘Class Conflict, Political Crisis, and the Breakdown of Democratic • • • • • • • •

Practices in Costa Rica: Reassessing the Origins of the 1948 Civil War’, Journal of Latin American Studies, 23: 37–60. Lehoucq, Fabrice (1996) ‘The Institutional Foundations of Democratic Cooperation in Costa Rica’, Journal of Latin American Studies, 26: 329–55. Lehoucq, Fabrice (1997) Lucha electoral y sistema político en Costa Rica, 1948–1998. San José: Editorial Porvenir. Lehoucq, Fabrice (2000) ‘Institutionalizing Democracy: Constraint and Ambition in the Politics of Electoral Reform’, Comparative Politics, 32: 459–77. Lehoucq, Fabrice (2002) ‘Can Parties Police Themselves? Electoral Governance and Democratization’, International Political Science Review, 23: 29–46. Lehoucq, Fabrice and Iván Molina (2002) Stuffing the Ballot Box: Fraud, Electoral Reform, and Democratization in Costa Rica. Cambridge and New York: Cambridge University Press. Ocontrillo García, Eduardo (1982) Un siglo de política costarricense: crónica de 24 campañas presidenciales, 2nd edn. San José: Editorial Universidad Estatal a Distancia. Rodríguez, Juan Carlos (2002) Entre curules y estrados: la consulta preceptiva de las reformas constitucionales en Costa Rica. San José: Investigaciones Jurídicas. Salazar Mora, Orlando (1990) El apogeo de la república liberal en Costa Rica, 1870–1914. San José: Editorial de la Universidad de Costa Rica.

6 Mexico: Designing Electoral Rules by a Dominant Party Alberto Díaz-Cayeros and Beatriz Magaloni

The very long-term ruling party of Mexico, the Institutional Revolutionary Party (PRI), has developed a number of electoral reforms since the early 1960s by achieving successive compromises with the opposition parties. On the one hand, electoral fraud was reduced and electoral competition was increased through a very slow and gradual process of successive reforms. On the other hand, the PRI crafted electoral rules aimed at protecting its own dominance. Basically, plurality rule in single-member districts was kept in place while adding proportional representation in multi-member district races for a minority of total seats. The PRI followed a divide-and-rule strategy, offering short-term electoral benefits to a fraction of the opposition, in exchange for rules that discouraged coordination among the opposition parties and which would eventually make the incumbent party more difficult to dislodge. Opposition political actors abided by the new rules in the expectation of generating greater political openings. But it was not until the 1997 congressional elections that the PRI lost its absolute majority in seats and not until the 2000 presidential election that an alternative candidate finally won.

The PRI’s electoral rules The Institutional Revolutionary Party (PRI) has been in power in Mexico since the 1930s (with different names) thanks to tight restrictions on political freedom and very low levels of electoral competition. But from the early 1960s on the PRI developed a number of electoral reforms involving some political openness of the system by achieving successive compromises with the opposition parties, namely the National Action Party (PAN) and the Party of Democratic Revolution (PRD). The general trend in Mexico’s electoral reforms was to keep plurality rule in single-member districts in place while adding proportional representation in multi-member district races for a minority of seats. Until the early 1960s the system was purely majoritarian. The lower chamber was elected by plurality rule in a variable number of single-member districts, 145

146

Alberto Díaz-Cayeros and Beatriz Magaloni

increased according to population estimates with each census. Two senators from each of the states were elected by plurality rule and limited vote in statewide races where voters could cast only one vote. Governors, local legislators and municipal presidents were all elected by plurality rule in single-member districts. In the years of greatest party hegemony, the electoral system created an insurmountable threshold for minority opposition parties. Virtually all the elective offices in the country were controlled by the PRI. The consecutive electoral reforms implemented from the 1960s to the 1990s gradually added multi-member district races without eliminating the majoritarian nature of the system. The resulting set of rules created strong incentives for opposition parties to compete against each other as long as they could not seriously attempt to win in single-member districts. Over the years, the PRI maintained a logic of institutional design that can be summarized as threefold: 1. To preserve the majoritarian nature of the original single-member district races such that the system tends to disproportionately reward the ruling party. 2. To reduce entry costs to opposition parties by adding proportional representation in a few multi-member districts. 3. To minimize the risk of losing the PRI’s majority control of both chambers by establishing ‘safety vote thresholds’ above which a majority of seats was artificially fabricated. These safety thresholds, however, were never used in practice. It was not until the 1997 midterm election that the PRI officially lost the absolute majority of the vote, but then it didn’t achieve the required threshold to be given over-representation. Overall, electoral reforms tended to preserve the existing balance of forces emerging from previous elections: protecting the PRI’s dominance while accommodating a slowly growing but divided opposition camp.

The Chamber of Deputies The Chamber of Deputies in Mexico has been formed by a set of rules that retained majoritarian districts in order to sustain the advantage of the incumbent PRI, while at the same time opening up spaces to opposition parties unable to win in the single-member districts. The opening of such spaces, however, prevented the coordination of opposition parties around single challengers in single-member districts, so that the PRI kept on winning most of the majoritarian races. Even more importantly, the lack of coordination in legislative races by the opposition contaminated the concurrent presidential race. Changes to the rules for the composition of the Chamber of Deputies were publicly justified on the grounds of enhancing political pluralism. While it is clear that new parties were able to enter the Chamber thanks to those rules, they also weakened in the long run the prospects for established opposition parties to win in singlemember districts.

Mexico

147

The rules for the translation of votes into seats were first changed in 1963, when so-called ‘party deputies’ were first introduced: opposition parties beyond a threshold of 2.5 per cent of votes could receive up to 20 seats. A more formal mixed system was established with the 1977 electoral reform, which reduced the barriers to entry of new parties into the system. The reform drew up 300 singlemember districts by plurality rule plus 100 compensatory seats, the latter distributed in a rather proportional manner by the Hare quota formula to blocked lists of minority parties only, namely those winning less than 60 single-member districts. The 1986 reform increased the number of seats allocated by proportional representation in multi-member districts to 200 and these were to be assigned to any party beyond a lower threshold of 1.5 per cent and less than 50 per cent of the vote. In addition, the 1986 reform established a so-called ‘governance clause’, which automatically gave the largest party the majority of seats if its vote was above 35 per cent but below 51 per cent. The rule also established an upper bound of 350 seats to the largest party. But the most long-lasting change of the 1986 reform supporting PRI dominance was to establish that voters could cast only one vote in the single-member district races that would automatically be counted for the allocation of seats coming from the regional multi-member districts. Thus, split-ticket voting was no longer possible, which had momentous implications for the coordination of the opposition. Rules for the translation of votes into seats were changed again in 1989 and 1993. In 1996 the independence of the electoral authority, the Electoral Federal Institute (IFE), was guaranteed, making the electoral results openly accepted by all party leaders for the first time. Campaign finances and access by opposition leaders to the media were regulated in a rather fair manner for first time. As an exchange, however, the PRI consolidated electoral rules that were custom-made for itself. The mixed system for the Chamber of Deputies was maintained with 300 singlemember districts and 200 deputies by party lists. The ‘governance clause’ now became a maximum limit for over-representation of eight percentage points, that is the majority of seats would be given to the largest party with more than 42 per cent of votes, which, in spite of the PRI’s decay, was within its expected reach. It should be noted that the PAN voted, together with the PRI, to approve these last electoral reforms. This decision can be explained by the strategic calculation, in a dominant party system, that settling on short-term concessions might eventually lead to a greater political opening. Three issues deserve attention. First, notwithstanding the controversies they generated at the time they were established, the rules to fabricate majorities were actually not used because the PRI had the necessary majority in both votes and single-member district victories to retain control of the Chamber of Deputies. Second, although the overall degree of disproportionality in the system was not significantly different from that of democratic systems, it systematically benefited the party more capable of winning in single-member races, which happened to be the PRI, and punished the second largest party in those races, namely the PAN. Third, the system created strong incentives for opposition parties not to coordinate to credibly challenge the PRI in the single-member races. In this respect, the

148

Alberto Díaz-Cayeros and Beatriz Magaloni

prohibition to divide votes among the single-member and multi-member races was particularly important. Through the prohibition of split-ticketing, the rules significantly increased the coordination costs among the opposition since electors could no longer strategically cast a vote for the strongest contender to the PRI in single-member district races without, at the same time, sacrificing a vote for their most preferred alternative in the multi-member races. The result was an utter lack of opposition coordination in the single-member district races. The effects of the electoral rules on the composition of the lower Chamber of Deputies are presented in Table 6.1. The third column presents a measure of disproportionality: the over-representation obtained by the PRI, namely the difference between seat share and vote share. In 1960, its over-representation was 6 per cent, a rather low number considering that the system was purely majoritarian. This reflects the degree to which the system was hegemonic at the time. Through the introduction of party deputies, the PRI was actually ‘giving away’ seats to some parties that, at least with the official vote tally, could not surpass the vote thresholds required by the legislation, let alone win single-member races. In fact, between 1964 and 1976, the PRI was slightly under-represented in three out of the five elections despite controlling virtually all the seats in the Chamber. After the mixed system was introduced in 1977, the tendency was to over-represent the PRI, though not by a large margin. But as elections became more competitive, the electoral rules tended to over-represent this party more. In the 1994, 1997 and 2000 elections, over-representation became closer to the two-digit range. Another important point is that the second largest party, the right-wing PAN, was generally under-represented. However, as rules for multi-member seat allocation Table 6.1 Mexico Chamber of Deputies: electoral results of the PRI in single-member districts 1961–2000 Year

PRI %v

PRI Seats/total

Disproportionality %s – %v

1961 1964 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

91 88 84 85 78 88 74 68 68 51 61 49 42 37

172/178 178/178 175/177 178/178 192/194 193/194 296/300 299/300 289/300 223/300 290/300 274/300 164/300 132/300

6.0 −1.6 −0.6 0.2 5.5 −3.0 −6.6 5.4 4.2 0.9 2.5 9.9 5.6 7.1

Note: The second column indicates the percentage of PRI votes; the third, the total number of seats won by the PRI in single-member races; and the fourth the PRI over-representation (seat share minus vote share). Sources: From data in Gómez Tagle (1997), correcting double accounting in Cuadro 8. For 1997 and 2000, Instituto Federal Electoral.

Mexico

149

changed and elections became more competitive, under-representation of the PAN was reduced or even reversed. This implies that between 1960 and 1985, the electoral rules originally benefited the PRI and small parties at the expense of the second largest party, increasing the incentives for the opposition to fragment. But after 1988, when it was the left that united and the elections became highly competitive, the PAN fared much better in the system and it was not systematically punished to the benefit of the tiny parties. Despite the PRI’s over-representation, the existing mixed system was better for the opposition than a purely majoritarian system. Clearly, all of the reforms significantly reduced entry cost to opposition parties. The reason is that the opposition was extremely unsuccessful at winning single-member races. Throughout this period, the PRI won the overwhelming majority of plurality races, the only exceptions being 1988 and 1997, when it won 75 per cent and 68 per cent of the seats coming from the single-member districts. Thus, in the short term, the mixed electoral system significantly reduced entry costs to opposition parties. But the existing system created perverse incentives in the longer term, since it discouraged coordination among these parties. The reasons why opposition parties have failed to coordinate have varied across time, depending on the parties’ electoral strength and the rules of the game. Originally, the most important issue for opposition parties was not so much winning control of the chamber, which was not really an imaginable prospect at the time, but to be players in the political arena, to enter the system. So the first priority was to obtain enough votes so as to be above the electoral threshold and hence retain the official registry. The resulting equilibrium was one where opposition parties were extremely unsuccessful at winning single-member races, which counted the most in the distribution of assembly seats. Winning more single-member races would require coordination. If opposition parties were regionally specialized in such a way that they did not have overlapping areas of support, an arrangement of coordination could presumably had been possible to achieve – although with high transaction costs. But this was not the case since there was considerable regional overlap between the PAN and the PRD vote. While each party tended to specialize in some regions, the former being more popular in the North and the Bajío region, and the latter in the South, both were strong in the more populous states of the Center, in the metropolitan area around Mexico City, and in the relatively urban and modern districts. Existing rules thus facilitated the entry of opposition parties, but made it hard for the PRI’s dominance to be seriously challenged. The opposition parties entered and established themselves as relevant institutional players; they failed for a long time to produce an alternative majority alliance to defeat the PRI. Actually the PRI did not need to use the ‘safety thresholds’ previously mentioned to fabricate its majority control of the Chamber. The ruling party’s share of the vote and its total number of single-member districts were above 50 per cent until 1994 – in fact, much above such a threshold, the only exception being 1988. Not until the 1997 midterm elections did the PRI’s vote share fall below 50 per cent and then, by a narrow margin, the rules failed to fabricate a majority for this party.

150

Alberto Díaz-Cayeros and Beatriz Magaloni

The Senate Even in well-established democracies, upper chambers are often used for the entrenchment of traditional or authoritarian traits inherited from the past. During the period of domination by the PRI, Mexico’s Senate was no exception. Reforms for the composition of the Senate highlight, more than any other development, the logic of institutional design explicitly directed at dividing opposition parties. Originally, two senators were elected from each state and the Federal District by plurality rule with limited vote in two-member districts with only one vote per voter. The rule disproportionately rewarded majority parties, namely the PRI. In fact, the ruling party had virtually won 100 per cent of senatorial seats until 1988, when the PRD won four seats in Mexico City and Michoacán. The 1993 reform doubled the size of the assembly, creating four seats for each state, three coming from a trinomial formula in statewide plurality races with only one vote cast and the fourth seat being given to the first minority party in the state. The electoral reform was specifically designed to benefit the PAN over the PRD, opening up a political space that had been virtually closed but without risking the PRI’s control of the assembly. Despite its increasing electoral strength, the PAN had only won one senatorial seat in the 1991 elections. The reform gave 24 senators to this party in the 1994 race. This meant that the formula ensured, in each state, 75 per cent of the seats for the party coming first place and 25 per cent for that coming second. This was consistent with the idea of ‘governance’ similar to that previously mentioned for the Chamber of Deputies, in which the winner was over-represented, although not as much as under a purely majoritarian system. The rules implied an unambiguous punishment for the third party and a reward to parties winning a state with less than 75 per cent of votes. For the second-place party, the reward or punishment depended upon whether they were above or below 25 per cent of the votes. Thus the formula generated strong incentives for the two major opposition parties to compete against each other for the second place, especially in states where they were unable to win a plurality of the vote. A second reform took place in 1996. It returned to the binomial formula, kept the first minority seat and established 32 seats by principle of proportional representation. This reform meant that the thresholds for over- or under-representation were now 66.6 and 33.3 per cent for the first- and second-place party respectively, rather than 75 and 25 per cent. Hence, the system was meant to become more proportional. The electoral reform was intended to benefit the PRD by introducing the principle of proportional representation, although only for one-quarter of the Chamber. This meant that, in the short run, the majority of the PRI was safe while the PRD was sure to have a significant presence in the Senate, even if it was weaker than the PRI or the PAN. The formula not only discouraged coordination, it further encouraged competition between the two major opposition parties. The electoral rule for the Senate implies that voters should mainly focus on electing two slots, a binomial formula and a third senator belonging to the second largest.

Mexico

151

It would hence get ever more difficult for the opposition to dislodge the PRI from the Senate: if the opposition failed to coordinate, the PRI would very likely win the overwhelming majority of the plurality binomial races and majority control of the upper Chamber. However, the PAN and the PRD approved the new rules because they significantly reduced entry costs compared to the former purely majoritarian system. Thus the logic of institutional design: reward the existing incumbent majority, reduce entry costs to the opposition and, in doing so, inhibit opposition coordination.

Political consequences What is distinct about the PRI’s performance in government for so long was not its position in the political spectrum, but its ability to craft electoral rules and repeatedly change them without threatening its dominance. Electoral rules translating votes into seats played a crucial role in dividing the opposition. Table 6.2 presents the results of presidential elections in Mexico from 1964 until 2000. This table shows the percentage of the votes received by the PRI in each election and votes received by the first and second loser. While the margin of victory for the PRI was narrowing throughout those decades, it is clear that there was no opposition coordination. In fact, even in the 2000 presidential elections, when the PRI finally lost office, there was no coordination either. The third party in the race still obtained 17 per cent of the popular vote (which was, quite surprisingly, almost the same percentage received by the second losers in both 1988 and 1994). One could object that the rules for the composition of the assemblies should not significantly discourage coordination in a presidential race. All Latin American countries have a combination of majoritarian presidential races with different forms of proportional representation in congressional elections that tend to produce different political majorities in the two institutions. While such logic holds in Table 6.2

Mexico: presidency electoral results, 1964–2000

Year

Winner % vote

First loser % vote

Second loser % vote

1964 1970 1976 1982 1988 1994 2000

88 85 88 68 51 49 43

11 14 4 16 31 26 36

<www.mtholyoke.edu/acad/polit/damy/prlib>

Part 3 Western Europe

This page intentionally left blank

9 Western Europe: General Overview Josep M. Colomer

Most countries in Western Europe use proportional representation rules for the election of their parliament. The two major exceptions are Britain and France in which plurality and majority rules, respectively, dominate – although both countries also use proportional representation systems for choosing their members of the European Parliament and in other second-rank elections. As will be shown in the following overview, a general pattern in the evolution and change of electoral systems in Western Europe since the early nineteenth century can be sketched in several stages. First, most countries held indirect elections for traditional estates-based assemblies, as well as for a number of more democratic parliaments. Second, direct parliamentary elections in a number of countries were organized in multi-member districts with multiple ballot allowing some degree of political pluralism, but they usually evolved into restrictive results at a political party’s induction of the bloc ballot. Third, further openness included, in a few cases, the introduction of a limited ballot giving some minorities representation or, more generally, the establishment of single-member districts, whether with plurality or majority rules. Fourth, proportional representation was introduced for the election of the national parliament in about fifteen countries in Western Europe at different points in time. In the early twentieth century, a number of traditional ruling parties – basically conservatives or liberals – introduced proportional representation when they felt they were under threat from pressure to expand suffrage rights and from the emergence of new parties, especially christians and socialists. Later on, similar electoral system choices were made in processes of democratization or redemocratization developed in multi-party contexts and driven by uncertain expectations.

The survival of majoritarian rules From the formation of the English Parliament in 1265 there were direct elections in rural shires and urban boroughs. The vote was open, by show of hands or acclamation. The majority principle was explicitly introduced by a statute in 1430, referring, however, to ‘the greatest number’ of voters, which consolidated the 179

180

Josep M. Colomer

practice of declaring winners by relative majority or simple plurality rule. Elections were held in multi-member districts with multiple ballot, which, while political parties were weak or non-existent, permitted some varied representation of different groups of voters. By the early nineteenth century, five-sixths of seats were still elected in England in two-member districts, although single-member districts had already become the norm in Ireland, Scotland and Wales. A series of gradual reforms starting in 1832 enlarged the number of voters, introduced a more homogeneous distribution of seats across the territory to correspond with the population, and made elections more fair and reliable. Gradually, this process favoured the emergence of new political leaders and groups. A crucial development was the replacement of independent local candidates with nationwide political organizations. The centralization of legislative power in the Cabinet induced more cohesive party voting in the House of Commons, thus increasing party coordination and discipline. On the basis of their parliamentary power, national leaders became able to nominate party candidates in most districts, which, in multi-member districts, induced voters to vote in bloc for the several candidates of a single party. The Liberals were most advantaged by their ability to coordinate and the corresponding over-representation. Seen in this perspective, the gradual replacement of two-member districts with smaller single-member districts can be considered to have been a measure able to delay the creation of single-party majorities supported by only a minority of popular votes. More than one-fifth of seats were elected in single-member districts after 1832, more than a quarter after 1867, and more than 90 per cent after 1885. Indeed over-representation was restrained: less than half (6 out of 14) single-party Cabinets enjoyed a majority of seats based on a minority of popular votes during the period from 1885 to the Second World War. Another decision oriented to prevent a high concentration of seats for a single party was the introduction of a limited ballot in certain multi-member districts. This formula had been proposed by Conservative members of Parliament in 1831, 1854 and 1860, and was implemented after 1867, in parallel with the establishment of larger household suffrage rights and secret ballot. In twelve three-member districts each voter was given two votes, while in a four-member district in London voters were given three votes each, in this way creating the expectation that at least a second party would obtain representation in each of those districts. The British experience of limited ballot was, however, very modest, since it was established to elect, in total, less than one per cent of parliamentary seats. In addition, in a few districts the Liberal Party excelled in coordinating the votes of its supporters among its several individual candidates in order to maximize the number of candidates elected by both plurality and second minority, thus producing again single-party victories. As a consequence of its extremely bounded political effects, the experiment of limited ballot was abandoned in Britain after 1885, while single-member districts were then adopted as the general rule. The debate on the electoral system reached a peak in the late nineteenth and early twentieth centuries, reflecting the increasingly visible political effects of the enlargement of the electorate. Minority radical and labour candidates traditionally

Western Europe: General Overview

181

running in liberal-backed candidacies eventually formed, with the help of trade unions, a new Labour Party. For a while, Conservative and Liberal leaders continued to defend the electoral system based on single-member districts by simple plurality rule, while the minority Labourites promoted proportional representation. However, the continuing rise of Labour candidates and the split of the Liberal Party (by Lloyd George’s faction) changed electoral expectations and the corresponding electoral system preferences. By the early 1920s, Labour leaders forecasted a possible party victory by relative majority against mutually rival Conservatives and Liberals under the existing electoral rules. As a consequence, they turned against their former proposals for proportional representation. Simultaneously, the Liberals, suddenly finding themselves in a minority likely to be defeated, turned in favour of proportional representation. A number of bills proposing alternative ballot or proportional representation by means of single transferable ballot were introduced in Parliament in 1916, 1920, 1924 and 1930, but they failed to obtain majority support, mainly due to the opposition of the Conservatives and divisions within the other two parties. The Labour Party did receive a plurality of seats in Parliament for first time in 1929, in spite of having obtained fewer votes than the Conservatives, and was able to form the government, although with minority parliamentary support and only for a brief period. Later on, in the elections after the Second World War, in spite of obtaining only a minority of popular votes, the Labourites achieved huge over-representation, were able to form a government with majority parliamentary support and stayed faithful to the existing electoral system. In this way, the previous party system, which was polarized around an over-representation of Conservatives and Liberals, was thus replaced with frequent alternations in government of over-representations of Conservatives and Labourites, while the Liberals, now dramatically under-represented in Parliament by the rule of relative majority, remained out of government for several decades. The electoral system based on single-member districts and simple plurality rule – usually known in Britain as ‘first past the post’ – found thus strong endogenous support from the parties benefited by the system itself. A new disturbance in the rhythm of party alternation then led the Labour Party temporarily to shift again its electoral system preferences. From 1979 on, the Conservatives won four elections in a row and stayed in government for 18 years, a period almost three times longer than the average in the previous three decades. This moved the Labour leaders to reconsider again the advantages of proportional representation. When they won the election and returned to government in 1997, they fulfilled some of the electoral promises formulated during that long period in opposition and introduced proportional representation systems for the election of newly created assemblies in London, Scotland and Wales, as well as for the British members of the European Parliament. However, the Labour Party’s new optimistic electoral prospects, which materialized in their re-election in 2001, led the government to freeze its own project of electoral reform for the House of Commons – as explained in the relevant chapter by Patrick Dunleavy and Helen Margetts.

182

Josep M. Colomer

As repeatedly discussed in this book, plurality rule tends to produce a single absolute winner supported by only a minority of popular votes. In Britain, the current two larger parties are advantaged by their relatively concentrated support in a number of local constituencies, in recent times focalized in South-East England for the Conservatives and in West and North England, as well as in Scotland and Wales, for Labour. In contrast, the Liberals have relatively strong national support but spread more evenly. With sufficient strength in a number of districts, a party can receive a majority of seats, even if it has not obtained a majority of popular votes either in the district or at national level. Actually, no single British party has obtained a national majority of popular votes since 1906 but, as a consequence of the electoral system based on single-member districts and simple plurality rule, a single party received an absolute majority of seats in the House of Commons and was able to form a single-party government in 15 out of 16 elections since 1945. Since all governments in this period have been formed by a single, minority and relatively extreme party, none of them has enjoyed the support of the median voter in the electorate. Even the second party in national votes became the winner in seats in six parliamentary elections since 1885, including in 1951 in favour of the Conservatives and in 1974 in favour of Labour. Along with Britain, France has also used majoritarian electoral systems throughout most of its contemporary history. The most usual formulas have been based on majority rule with a second-round runoff, which, in contrast to plurality rule, has permitted multiple political parties to exist and participate in democratic institutions. However, these majoritarian formulas have also produced frequent big distortions between the distribution of popular votes and the allocation of seats to the different parties, including huge shifts in government party composition as a consequence of a few voters’ swings. A very high number of changes to the electoral rules in France suggests that majoritarian systems may often produce large numbers of unsatisfied losers, as well as unintended consequences against the previous promoters of a specific set of rules, thus feeding new preferences and strategies in favour of innovative formulas. Following a medieval tradition started in the fourteenth century, indirect elections in multi-member districts were irregularly held in the kingdom of France for provincial estates and the Estates General. Voters met and expressed their oral vote; every seat was filled separately by voting in several rounds, if necessary, until a candidate obtained majority support. In the third estate election for the 1789 Estates General an absolute majority was required only in two rounds, the third round being decided by plurality. On that occasion, these rules, together with broad popular participation, apparently permitted representation of different categories of interests and the formation of large majorities with widely shared objectives of individual rights and economic modernization. Indirect elections by oral vote were also held in a later period, always in multi-member districts by majority rule and three rounds, the third round either requiring only plurality or a runoff between the two most voted candidates. However, as previously discussed in this book, elections in several stages can distort voters’ preferences and produce unforeseeable results as a consequence of shifting coalitions among intermediate electors. Actually, very different

Western Europe: General Overview

183

winners emerged from successive indirect elections in late eighteenth- and early nineteenth-century France, thus giving room for high political instability throughout a parliamentary monarchy, a republican convention, an empire and a restored traditional monarchy. Direct elections were introduced in France by the restored Bourbon monarchy in 1817, but representativeness was limited in this period to a very small minority of voters fulfilling strong economic requirements. During this period, elections were held in a mix of single- and multi-member districts, again by majority rule with a second round either by plurality or runoff. Universal men’s suffrage was enacted for the first time in France – and in the world – in 1848. The election of a constituent assembly was held again in multimember districts, now with open lists by which every voter could vote for as many candidates as desired up to the number of seats to be filled in the district. The electoral rule was majority with a second round by plurality, which was able to produce an over-represented single winner in each district. Also a first in the world was the direct election of the President of the Republic by universal men’s suffrage introduced in 1851 by using majority rule with a second round in the Assembly. Shifting and surprising results were produced by these electoral rules, including two different parliamentary majorities within a year, one in favour of the Republic, the other in favour of the Monarchy, and the election of Prince Louis-Napoleon Bonaparte as president, which shortly preceded his authoritarian dissolution of the assembly and the re-establishment of the empire. The longest democratic period in modern France was the Third Republic inaugurated in 1871. Initially, there was a dominant party, the conservative republicans, which confidently re-established a system based on multi-member districts by plurality rule, in the expectation of benefiting from the corresponding overrepresentation. However, electoral competition increased very soon between both the right-wing monarchists and the left-wing republicans, which led the intermediate conservatives to establish less risky, smaller single-member districts with the rule of majority with a second round by plurality after 1875. Yet the enforcement of plurality rule at the second round still encouraged the tactics of ‘divide (the rivals) and win’ absolutely, even with a minority of votes. In the subsequent confidence of obtaining single-party absolute victories, the left-republicans, initially led by Léon Gambetta, promoted the re-establishment of the Second Republic system of multi-member districts. The system was enforced in 1885, but soon threatened to backfire against its initial promoters when the conservatives appeared to be favoured in electoral expectations. Accordingly, a new shift in electoral system preferences emerged by which optimistic conservatives adhered again to multimember districts while the republicans under threat now promoted single-member districts – the latter being re-established in 1889. The long-lived electoral system based on single-member districts with the rule of majority and second-round plurality was thus less risky than the previous system based on multi-member districts. But it still induced a high degree of polarization in French politics, especially because of the anticipated advantage of the larger parties at the second round. While monarchists on the right and radicals on the

184

Josep M. Colomer

left were intent on maintaining the system, the declining conservatives and the new minority socialists favoured its replacement with some formula of proportional representation. However, they succeeded in establishing an ephemeral mixed system only in 1919 by which majority rule was applied at the first round and proportional representation was used to fill the remaining seats. Yet the majoritarian component of the system encouraged the formation of coalitions able to win by majority at the first round, so moving polarization from the second round to the first. Actually, polarization kept growing during the following period in which single-member districts were re-established, reaching a peak in 1936 when the election was won by the leftist popular front formed by the radicals, the socialists and the communists. In France, political polarization only came to be replaced by the dominance of broad centrist coalitions with the establishment of an electoral system by proportional representation with the Fourth Republic initiated in 1945. From 1951, electoral procedures included multiple ballot (‘panachage’) by which every voter could choose candidates from all parties, as well as a majority premium given to those parties that, although running separately, established apparentements and obtained together an absolute majority of popular votes. These rules favoured the formation of broad electoral coalitions prefiguring further parliamentary and cabinet multi-party coalitions. They were always organized around the centre, basically including the socialists, the radicals, the christians and the conservatives (with prime ministers from the centrist radicals and christians), to the exclusion of the communists on the left and the Gaullists on the right. However, a coup d’état organized by General Charles de Gaulle ended the experience in 1958 and established a new Fifth Republic. Again, assembly elections were held in single-member districts by the rule of majority with a second round by plurality, and again, as on the occasion of the Second Republic, the direct election of the President was introduced, now by majority rule with a second-round runoff between the two most voted candidates – as explained by Gerard Alexander in the corresponding chapter. A high degree of polarization resumed and developed again in French politics in the following period, this time around the socialists and the communists on the left and the Gaullists and the conservative republicans on the right. Only the prospect of electoral decline by the socialists in government moved them to introduce proportional representation for the election of 1986, but subsequently the winning Gaullists soon re-established the previous majoritarian formula. As discussed in several sections of this book, majority rule with a second round can produce highly distorting electoral results depending on asymmetric coalitional skills at forming candidacies, the corresponding elimination of potentially popular candidates, and the over-representation received by winners at the district level. Actually, in most assembly elections in the French Fifth Republic, the potentially winning parties in votes became losers in seats. Specifically, the socialists and the communists, running separately, obtained higher numbers of votes, if counted together, than the Gaullists and their republican allies, but the latter received a majority of seats in 1958, 1962, 1967, 1973 and 1978. After the socialists and the communists began to form electoral coalitions, the electoral system backfired

Western Europe: General Overview

185

and gave a majority of seats to the two leftist parties, in spite of their having obtained fewer votes than the two rightist parties previously mentioned, in 1981 and 1988. With regard to presidential elections, four of the first seven winners were minority, relatively extreme candidates at the first round who benefited from the elimination of more moderate, median voter candidates who would likely have defeated them by majority in a pair-wise comparison in 1965, 1988, 1995 and 2002.

The diffusion of proportional representation A more complete evolution from indirect elections to majoritarian formulas and to proportional representation than in the two cases discussed above can be identified in most countries of Western Europe from the early nineteenth century. In the first stage, medieval traditions of indirect elections for estates-based assemblies were maintained in France until 1789 (as mentioned above), but they were also used in Sweden until 1864 and in Finland until 1906. Indirect elections were also organized for more modern parliaments based on individual suffrage in France until 1816 (as mentioned), Denmark until 1834, Spain until 1836, Greece until 1843, the Netherlands and Switzerland until 1848 and Norway until 1903. As has been discussed in the general introduction, indirect elections can produce outcomes unrelated to voters’ preferences through uncommitted intermediate electors forming shifting coalitions. They can thus produce a high degree of electoral unpredictability and political instability, even if – as was the case in the countries and periods mentioned – suffrage rights were significantly restricted. At the second stage, direct parliamentary elections were organized in many countries with multi-member districts and multiple ballot. Instead of the typically British simple plurality rule, most countries used a rule of majority and a second round, whether by plurality or in runoff between the two most voted candidates. This was the case, in addition to France until 1885, of Belgium from 1831, Spain from 1836, Greece from 1844, the Netherlands and Switzerland from 1848 and Portugal from 1852. As previously discussed, while multiple ballot may permit the representation of a variety of interests and opinions when individual candidates run independently, the formation of more compact candidacies and the emergence of disciplined political parties tend to induce voters to vote in bloc for candidates of a single party. Thus multi-member districts with a majoritarian rule may become a highly restrictive formula giving one party huge over-representation. Accordingly, in a third stage of the evolution of electoral systems in Western Europe, two alternatives oriented to reduce the concentration of power in a single party and permit the representation of large minorities were devised. These were, first, the introduction of limited ballot in multi-member districts and, second, the establishment of smaller single-member districts. The formula of limited ballot was seen as a device to open representation for a minority of seats to opposition parties to the extent that it reduces the overrepresentation usually received by the larger party in a multi-member district with bloc ballot. With this intention, it was used in several European monarchies

186

Josep M. Colomer

dominated by conservatives and liberals but threatened by strong pressures from republican and leftist parties. In Spain, limited ballot was used in three different periods. It was first used after 1865 – a couple of years earlier than in Britain, as explained above – in multi-member districts in which each voter had a single (non-transferable) ballot. Second, it was adopted by the restored monarchy after 1878 for a few districts with magnitudes from three to eight seats, giving each voter a number of votes amounting to around two-thirds of the seats in the district, to elect less than a quarter of the total seats in parliament. In this way, the conservative and the liberal parties could maintain their domination while also permitting the republican opposition to achieve a small minority representation. Third, when the republican parties eventually won local elections and the Spanish Second Republic was established in 1931, the self-confident republicans adopted, in reaction to the restrictions of the previous monarchical system and in the expectation of remaining politically dominant, limited ballot as the general formula for all parliamentary seats. New districts were drawn with magnitudes between 2 and 18 seats using again the rule of two-thirds. Limited ballot as a general formula with high-magnitude districts permitted multiple parties to exist, but also promoted a high degree of polarization, as well as several electoral results in which a multi-party bloc obtained more votes but received fewer seats than the parties of the opposite bloc – as explained in Chapter 13. A mixed system was also adopted in Italy, but only for a short period after 1882. The political dominance of the liberals was maintained through the election of about two-thirds of the total number of seats in parliament in multi-member districts with magnitudes between two and four seats, in which each voter had as many votes as seats to be filled. But, at the same time, a limited ballot giving each voter four votes was established in five-member districts for about one-third of the total seats in parliament, in this way giving some small representation to the radicals, democrats and reformists. The experience, however, was cancelled in 1890. Finally, limited ballot was also used in Portugal in three different periods. After 1884 it was introduced in urban districts with up to six seats, while rural areas maintained single-member districts. After 1901, still under the monarchy, limited ballot was adopted for the election of all seats in parliament. While in opposition, the republicans were in favour of proportional representation. But the Portuguese First Republic inaugurated in 1910 maintained limited ballot with three votes per voter in four-member districts, with the exception only of four urban districts in Lisbon and Oporto in which proportional representation was used for a short period. The choice of limited ballot was made – as in Spain 21 years later – in the expectation that the republicans would enjoy permanent popular majority support, but the exclusiveness of the system also contributed to an authoritarian reaction from the right. In contrast to these experiments, a reduction in electoral risk and an openness to the representation of minorities were attempted in other countries by moving from indirect elections or multi-member districts to single-member districts. This was the case in Denmark in 1848, in Sweden (which also kept some multi-member districts) in 1866, in Britain in 1885 (as mentioned) and in the Netherlands in 1887, all

Western Europe: General Overview

187

countries that still had restricted suffrage rights, while Germany in 1871 and Norway in 1906 moved in that direction at the same time as they introduced universal male suffrage. With the new framework of single-member districts, Britain and the three Scandinavian countries mentioned, Denmark, Sweden and Norway, used plurality rule, while the other two adopted majority rule with a second round to be held between the two most voted candidates in Germany and by plurality in the Netherlands. As discussed, single-member districts may permit the representation of some minorities under the condition that they are sufficiently strong in some districts, so typically are unevenly distributed across the territory. Yet if, in contrast, political minorities are spread more evenly across districts, they can be excluded from representation while more concentrated parties can receive much higher proportions of seats than proportions of votes obtained. Finally, the adoption of proportional representation became a fourth stage in this evolution driven by decisions made under pressure for the enlargement of the electorate and the emergence of new political groups and parties. The opportunity had first appeared in Britain with the peculiar proposal of a single transferable ballot put forward by Thomas Hare in the late 1850s. But the principle of proportional representation was more easily understood and found to be feasible for large mass electorates when more simplified voting and counting procedures were devised, especially facilitated by the formulas of the Belgian Victor d’Hondt and the Swiss Eduard Hagenbach-Bischoff in the 1880s, as well as the French André Sainte-Laguë in the early twentieth century. In most countries in Western Europe, proportional representation was devised by incumbent rulers and traditional parties as an institutional safeguard against strong pressures promoting the adoption of universal suffrage. The traditional parties sensed that the formation and rise of new parties, especially the socialists, supported by industrial workers and other likely new voters, threatened to turn the previous political equilibrium upside down if electoral contests continued to be run by majoritarian rules. The threat could be considered stronger the higher the levels of industrialization, as in Germany and the Scandinavian countries, and the more divided the ruling groups, as in territories with religious or language cleavages such as Belgium, the Netherlands and Switzerland. The ruling parties in these countries could accept the broadening of the franchise if the protective device of proportional representation were also introduced. With this, they expected to prevent the formation of a new single-party political majority without social majority support and to guarantee fair representation and influence to traditional groups and interests. Belgium was the first country to introduce simultaneously for national elections universal men’s suffrage (although with more than one vote for voters fulfilling some economic conditions) and proportional representation with party lists using the d’Hondt formula in 1899. With this change, the country replaced the previous, highly risky system based on multi-member districts. The move was led by the liberals, mostly localized in French-speaking Wallonia, who had alternated or shared power with the christians who were stronger in Dutch-speaking Flanders, when they felt seriously challenged by the rising socialists.

188

Josep M. Colomer

More dramatically, moving away from indirect estates-based elections, proportional representation was introduced in Finland together with the establishment of universal suffrage for men and – for first time in Europe – also women in 1906. In Sweden proportional representation was also introduced suddenly by the conservatives, simultaneously with the adoption of universal men’s suffrage in 1910, in the face of resistance from both liberals and socialists who expected to be able to win by the existing plurality rule system – as explained by Leif Lewin in the relevant chapter. In other countries, similar strategic motivations were latent until a major wave of proportional representation swept over Europe at the end of the First World War. In Switzerland the electoral dominance of the radicals for several decades after their victory in the civil war in 1848, a dominance which also benefited from a majoritarian system enforced in multi-member districts, began to be challenged by the emergence of new christian and socialist parties in the late nineteenth and early twentieth centuries. This challenge motivated the adoption of proportional representation in 1918, again as a safeguard device – as explained by Georg Lutz in the corresponding chapter. At about the same time, several other countries adopted almost simultaneously electoral systems of proportional representation and universal men’s suffrage, including the old kingdoms of Denmark, Italy, the Netherlands and Norway, and the new republics of Austria, Germany and Ireland (the latter with single transferable ballot). In all these cases, multi-party systems already existed, typically including both traditional conservatives and liberals (sometimes after undergoing significant recycling) and new radicals, christians and socialists. After the introduction of proportional representation, in general no party received a majority of seats, which gave room for the formation of multi-party coalition governments with broad electoral and parliamentary support which were able to foster wide popular allegiance to the new democratic regimes. Further occasions to establish proportional representation electoral systems were created by the emergence of multi-party systems in opposition to new authoritarian regimes, as happened at the end of the Second World War in most of the countries mentioned above. Also in the 1970s proportional representation systems were adopted in the processes of redemocratization in Greece (which had experimented with proportional representation after 1926, but had later moved back and forth), Portugal and Spain. This overview illustrates our previous hypothesis that electoral system preferences and choices are shaped by a political party’s electoral expectations, as well as by the availability of sufficiently innovative formulas, rather than by a party’s ideology or policy programmes. As discussed above, with the formation of disciplined electoral and parliamentary parties, traditional multi-member districts decided under majoritarian rules became a highly risky choice, only betted on by a few parties with optimistic expectations to remain or become dominant – such as the French left-republicans in 1848 and again in 1885 and the Spanish progressive liberals in 1854, but on each of these occasions for only one election. Alternative formulas, namely the establishment of limited ballot in the

Western Europe: General Overview

189

same type of multi-member districts or the adoption of single-member districts, reduced risk and were able to maintain the likely dominance of two alternating parties, such as the traditional conservatives and liberals, while giving some large minorities small representation. These were the most usual choices in most countries in Western Europe in the second half of the nineteenth century. However, the formation of more pluralistic political settings after the late nineteenth century, typically including new parties such as the christians and the socialists, precipitated the adoption of proportional representation systems. In contrast to the restrictive and polarizing political consequences of majoritarian systems, proportional representation permits multiple parties to exist, participate regularly in democratic institutions and share power in parliamentary coalition governments. In fact, the choice of proportional representation was favoured, at different moments in time and under specific circumstances, by three types of parties. First, its promotion was pioneered in late nineteenth and early twentieth centuries by conservatives in decline or facing strong challenges, as in France and Sweden, as well as by liberals subjected to similar threats, as in Britain, Belgium, Italy and Norway. Second, proportional representation was also promoted by new parties emerging during the first decades of the twentieth century, when they were still in a minority, as was the case of the christians in France, the Netherlands, Spain and Switzerland as well as the socialists (and the communists, if electorally relevant) in Austria, France, Germany, Italy and Switzerland. Third, other permanent minority parties, such as those based on certain ethnic groups, also supported proportional representation. This has been traditionally the case for parties representing the Germans in Denmark, the Swedes in Finland, the Protestants in Ireland, the Basques and the Catalans in Spain, and the Frenchand Italian-speaking groups in Switzerland. In contrast, some other left-wing, radical or republican parties maintained risky majoritarian systems for multi-party elections by universal suffrage – even if adopting slightly open formulas such as limited ballot, as in Portugal in the 1910s and Spain in the 1930s, but these were not able to prevent a high degree of polarization and political and social conflicts. Also, when some of the new, emerging parties just mentioned became sufficiently strong, they reversed their electoral system preference in favour of majoritarian systems able to give them an absolute victory, even if they only obtained a relative majority of popular support. This was the case for Labour in Britain after the 1920s as well as the French socialists which temporarily allied with the radicals in the late 1920s. Even when proportional representation had already been established, attempts to re-establish electoral systems based on majoritarian rules were launched by certain new would-be dominant parties, such as the christians in Italy in the 1950s and in Germany in the 1960s, the nationalists in Ireland in the 1960s and the socialists in the Netherlands in the 1970s. But all these latter attempts were promoted in contexts where a high degree of multi-partism, already well-established thanks to previously existing proportional representation rules, had found strong resistance from other parties and failed.

190

Josep M. Colomer

The political consequences of proportional representation in Western Europe can be summarized by the fact that most multi-party parliamentary governments – and in most countries all of them – have been supported by a majority of voters, including, of course, the one at the median position. This is in vivid contrast to the usual outcomes in Britain and France where – as discussed – an often extreme electoral minority over-represented in Parliament tends to rule in exclusivity. The Federal Republic of Germany, as organized after the defeat of Nazism, has probably obtained one of the best results – as explained by Marcus Kreuzer in his corresponding chapter. Since the election of 1953, each voter has had two votes: one for a party list in a very large multi-member district and one for an individual candidate in a single-member district. Virtually all seats in parliament are allocated to the parties on the basis of their votes by a formula of proportional representation (initially d’Hondt but since 1985 Hare). About half of the seats are filled by candidates in the party lists and the other half by individual candidates who have obtained the highest number of votes in the single-member districts. This system of personalized proportional representation may be considered to reunite the best of both worlds: first, proportionality in the allocation of seats and multi-partism, but without the rigidity of party closed lists; second, closeness and exchange between voters and individual candidates, but without loss of minority representation. From Germany, the system has been imported by a number of democracies in other parts of the world at the end of the twentieth century, as explained in the relevant chapters of this book. All German parliamentary governments after the Second World War have been formed by coalitions of two or more parties. From the election in 1949 to that in 1957, the christians obtained the support of the median voter and – thanks to proportionality in representation – also the median seat, so that they were able to form majorities in parliament and support governments together with some minor allies. From 1961 to 1994, after the rightist and nationalist parties surviving from the previous period had disappeared, the median voter and the median seat were captured by the liberals, who began to participate in government as a minor partner with either the christians or the socialists for longer periods than any of the two larger parties just mentioned. After 1998, perhaps as a somewhat delayed effect of the reunification with the eastern part of Germany, the socialists became the median party in votes and seats and could form majority governments in coalition with the greens. Comparable developments can be found in Belgium, the Netherlands and Luxembourg, all of them governed by multi-party coalitions dominated most of the time by the christians, but also including some periods of socialist-led governments. In Sweden, Norway and Denmark, in contrast, political dominance has rather leaned to the left, led by the socialists, although conservative governments have also been formed, usually in coalition with the centrist agrarian party which has tipped the balance between left and right on a number of occasions. The pivotal role of the agrarians has been stronger in Finland, where they have led many

Western Europe: General Overview

191

multi-party governments. In Switzerland, broader multi-party coalitions, including both the christians and the socialists, as well as the liberals and the agrarians, have been formed over several decades. Supermajority governments result in this case from proportional representation and from the threat of cantonal referendums able to block federal legislation if its political support is not sufficiently widespread in the territory. In general, proportional representation electoral systems create strong self-reinforcing mechanisms: the multiple political parties that they permit to exist and, on the basis of their votes, receive seats and government offices tend to oppose and resist any attempt to replace those rules with more restrictive alternative formulas. A somewhat comparable but in the end different development took place in Italy after the end of the Second World War. During eight elections from 1946 to 1979, the Italian christian party, located at the political centre, obtained the support of the median voter and the median seat and thus was able to form the government in parliament, either in coalition with small partners or as a singleparty minority not vulnerable to a joint challenge from its right and its left. In the following period, as a result of increasing electoral support for the communists, the median voter and seat lay in the hands of the socialists and other small centre-left ‘lay’ parties, which obtained a large proportion of the ministries in government – although still under the primacy of the christians. However, the long dominance of the christian party led the opposition to provoke the split of that party and to replace the existing electoral system of proportional representation with a parallel mixed system, in the expectation of producing clearer alternations of parties in government – as explained by Diego Gambetta in the relevant chapter. Although multiple parties managed to survive, two polarized coalitions were indeed formed which have alternated quickly in government since the second half of the 1990s. Finally, the establishment of direct elections to the European Parliament since 1979 also created new occasions to introduce proportional representation electoral systems in the 15 member states of the European Union, including Britain and France in spite of the maintenance of majoritarian formulas in most of their domestic elections. In most countries, European elections are held in a single nationwide district with open lists. Although nation-based electoral campaigns have been run by up to more than one hundred national parties in total, their proportional representation in the European Parliament has fostered the formation of broad, encompassing European political groups. In fact, after five direct elections, the European political groups of the populars (formed by christians and conservatives) on the centre-right, the socialists on the centre-left with the liberals in between have always brought together between two-thirds and three-quarters of the members of the European Parliament, in this way facilitating the formation of political majorities and sheer effectiveness in parliamentary decision-making. Its pluralistic performance has given the European Parliament an increasingly significant role within inter-institutional relations in the European Union, making the latter move away from being merely a diplomatic intergovernmental organization towards more democratic formulas.

192

Josep M. Colomer

Further reading • Bartolini, Stefano (1984) ‘Sistema partitico ed elezione diretta del capo dello stato in Europa’, Rivista Italiana di Scienza Politica, 14, 2: 223–43.

• Boix, Charles (1999) ‘Setting the Rules of the Game: the Choice of Electoral Systems in Advanced Democracies’, American Political Science Review, 93: 609–24.

• Caramani, Daniele (2000) Elections in Western Europe since 1815. London: Macmillan – now Palgrave Macmillan.

• Carstairs, Andrew McLaren (1980) A Short History of Electoral Systems in Western Europe. London: George Allen & Unwin.

• Cole, Alistair and Peter Campbell (1989) French Electoral Systems and Elections since 1789. Aldershot: Gower.

• Colomer, Josep M. (ed.) (2002) Political Institutions in Europe, 2nd edn. London: Routledge. • Cotteret, Jean-Marie and Claude Emeri (1973) Les systèmes électoraux. Paris: Presses Universitaires de France.

• Hoag, Clarence and George Hallett (1926) Proportional Representation. New York: Macmillan. • Lakeman, Enid (1974) How Democracies Vote: A Study of Electoral Systems. London: Faber & Faber.

• Lijphart, Arend (1994) Electoral Systems and Party Systems. A Study of Twenty-Seven Democracies, 1945–1990. New York and Oxford: Oxford University Press.

• Mackie, Thomas T. and Richard Rose (1991) The International Almanac of Electoral History, 3rd edn. London and Washington, DC: Macmillan-Congressional Quarterly.

• Noiret, Serge (ed.) (1990) Stratégies politiques et réformes électorales: Aux origins des modes de scrutin en Europe aux XIX et XX siècles. Baden-Baden: Nomos (English trans. Political Strategies and Electoral Reform. Origins of Voting Systems in Europe in the 19th and 20th Centuries. Baden-Baden: Nomos). • Rokkan, Stein, and Jean Meyriat (eds) (1969) National Elections in Western Europe. The Hague and Paris: Mouton.

Summary Table 3A

Western Europe: Assembly (lower or single chamber)

Country Law year Election years

Term

Districts

Ballot

Rule/Formula

Threshold

1 quota

Seats

No.

Magn.

3 4

165–183

25

*7

Open list

Prop. (2 tiers) – d’Hondt

1970: 1971, 1975, 1979, 1983, 1986, 1990, 1994, 1995, 1999

4

183–251

9–43

*42

Open list

Prop. (2–3 tiers) – Hare/d’Hondt 4% nat.

Belgium 1831: 1847, 1848, 1850, 1852, 1854, 1856, 1857, 1859, 1861, 1863, 1864, 1866, 1868, 1870, 1872, 1874, 1876

3

108–124

41

26–30

Public Multiple

Majority/2nd round plurality

1877: 1878, 1880, 1882, 1884, 1886, 1888, 1890, 1892, 1894, 1896, 1898

4

132–152

41

32–37

Secret Multiple

Majority/2nd round plurality

1899: 1900, 1902, 1904, 1906, 1908, 1910, 1912, 1914

4

152–186

30

51–62

Closed list

Proportional – d’Hondt

4

150–212

21–39

2–33

Open list

Proportional (2 tiers) – d’Hondt/Hare

Austria 1918: Universal MS 1920 1923, 1927, 1930 // 1945, 1949, 1953, 1956, 1959, 1962, 1966, 1970

1919: Universal MS 1919, 1921, 1925, 1929, 1932, 1936, 1939 // 1946, 1949, 1950, 1954, 1958, 1961, 1965, 1968, 1971, 1974, 1977, 1978, 1981, 1985, 1987, 1991, 1995, 1999

193

(Continued)

Country Law year Election years

194

Summary Table 3A

Term

Districts Seats

No.

1920: 1920, 1924, 1926, 1929, 1932, 1935, 1939, 1943, 1945, 1947, 1950 1953: 1953, 1957, 1960, 1964, 1966, 1968, 1971, 1973, 1975, 1977, 1979, 1981, 1984, 1987, 1988, 1990, 1997, 2001 Finland 1863:

Public Single

Indirect Plurality

Threshold

100–113

100–113 1

Single

Direct Plurality

114

114

1

Secret Single

Plurality

2

140 93 47

140 93 3

16 1 *16

Single

P-Proportional – Hare/d’Hondt

4

139–148

23–24

5–31

Open list

Proportional (2 tiers) – Hare/d’Hondt

4

175

17–23

2–40

Open list

Prop. (2 tiers) – m. St-Laguë/ Hare

198–316

4 chambers

Public Single

Indirect Plurality

1901: 1901, 1903, 1906, 1909, 1910, 1913, 1915 1915: Universal MS 1918

Rule/Formula

Magn.

Denmark 1834:

1848: 1848, 1849, 1852, 1853, 1854, 1855, 1858, 1861, 1864, 1866, 1869, 1872, 1873, 1876, 1879, 1881, 1884, 1887, 1890, 1892, 1895, 1898

Ballot

2% nat.

1906: Universal MS 1907, 1908, 1909, 1910, 1911, 1913, 1916, 1917, 1919, 1922, 1924, 1927, 1929, 1930, 1933, 1936, 1939 // 1945, 1948, 1951 1954, 1958, 1962, 1966, 1970, 1972, 1975, 1979, 1983, 1995, 1999

3

192–200

14–16

1–30

Secret Open list

Direct Proportional – d’Hondt

>1

Public Multiple

Indirect Majority/3rd round plurality

4

France 1789: 1789

1791: 1791, 1792

2

745

249 dep.

1–3

Multiple

Majority/3rd round runoff

1795: 1795, 1796, 1797, 1798

3

750

dep.

*3

Multiple

Majority/3rd round plurality

1815: 1815, 1816

5

258–402

dep. & arr.

>1

Multiple

Majority/3rd round runoff

1817: 1817, 1818, 1820

5

256

dep.

>1

Multiple

Direct Majority/3rd round plurality

1820: 1820

5

430 258 172

258 arr. dep.

1 >1

Single Indirect

Majority/3rd round runoff

195

196

Summary Table 3A

(Continued)

Country Law year Election years

Term

Districts

Ballot

Rule/Formula

Seats

No.

Magn.

430 258 172

258 arr. dep.

1 >1

Single Indirect

Majority/3rd round plurality

1848 Universal MS 1848, 1849 //

750–900

dep.

>1

Open list

Majority/2nd round plurality

1871: 1871, 1872

750

dep.

>1

Open list

Plurality

1875: 1876, 1877, 1881

533–542

533–542

Single

Majority/2nd round plurality

1885: 1885

569

dep.

Open list

Majority/2nd round plurality

541–592

541–592

Single

Majority/2nd round plurality

1919: 1919, 1924

574–616

90 dep.

Secret Multiple

Majority/proportional – Hare

1927: 1928, 1932, 1936 //

602–608

602–608

Single

Majority/2nd round plurality

1824: 1824, 1827, 1830

1889: 1889, 1893, 1898, 1902, 1906, 1910, 1914 //

5

1

>1

1

*7

1

Threshold

1945: 1945, 1946, 1946

1

522–544

90 dep.

2–9

Closed/open lists

Proportional – Hare

1951: 1951, 1956

5

544

90 dep.

2–9

Multiple

Majority/proportional – Hare

5

465–474

465–474 1

Single

Majority/2nd round plurality

1986: 1986

5

556

96 dep.

*6

Closed list

Proportional – Hare

1988: 1988, 1993, 1997, 2002

5

555

555

1

Single

Majority/2nd round plurality

3

382–397

382–397 1

Single

Majority/2nd round runoff

1

421

36

*12

Closed list

Proportional – Fix quota

4

459–647

52

*11

Closed list

Proportional (3 tiers) – Hare

4

400 (+2) 158 242

11 11 242

3–65 3–65 1

Single

P-Proportional – d’Hondt

1958: 1958, 1962, 1967, 1968, 1973, 1978, 1981

Germany Universal MS 1871, 1874, 1877, 1878, 1881, 1884, 1887, 1890 1893, 1898, 1903, 1907, 1912 1919: 1919 1920: 1920, 1924, 1924, 1928, 1930, 1932, 1932, 1933 // 1949: 1949

5% dis.

5% dis.

5 60,000 v.

5% dis.

197

198

Summary Table 3A

(Continued)

Country Law year Election years 1953: 1953

1956: 1957, 1961, 1965, 1969, 1972, 1976, 1980, 1983 1985: 1987, 1990, 1994, 1998, 2002

Term

Districts

Rule/Formula

Threshold

Seats

No.

Magn.

4

484 242 242

9 9 242

3–65 3–65 1

Double: Closed list Single

P-Prop. – d’Hondt

5% or 1 dis.

4

494–496 (+3) 247–248 242

1 9–10 242

494–496 3–65 1

Double: Closed list Single

P-Prop. – d’Hondt

5% or 3 dis.

4

496–656 (+5) 248–328 248–328

1 496–656 10–16 3–65 248–328 1

Double: Closed list Single

P-Prop. – Hare

5% or 3 dis.

>1

Public

Indirect

>1

Secret Multiple

Direct Majority/further rounds

≥1

Single/Multiple

Plurality

Greece 1822: Universal MS 1823, 1825, 1829 // 1843 1844: 1844, 1847, 1850, 1853, 1856, 1859, 1861, 1862 1864: 1865, 1868, 1869, 1872, 1873, 1874, 1875, 1879, 1881, 1885, 1887, 1890, 1892, 1895, 1899, 1902, 1905, 1906, 1910, 1912, 1915 // 1920, 1923

Ballot

34–98

1926: 1926

286

41

*7

Open list

Proportional (3 tiers) – d’Hondt/Hare

1928: 1928

250

250

1

Single

Plurality

Open list

Proportional (3 tiers) – d’Hondt/Hare

Single/multiple

Plurality

Open list

Proportional (3 tiers) – d’Hondt/Hare

Single/multiple

Plurality

Single

Coexistence Proportional for two parties Plurality

65

Open list

Proportional (3 tiers) – Hare/d’Hondt

65

Open list

Prop. (3 tiers) – d’Hondt/Hare 10–25% nat.

65

Open list

Proportional (2 tiers) – Hare

Open list

Prop. (3 tiers) – d’Hondt/Hare 0–3% nat.

1932: 1932

44

1933: 1933

98

1936: 1936 // 1946, 1950, 1951

39–42

1952: 1952

300

1955: 1956

1958: 1958 1961: 1961, 1963 // 1974, 1977, 1981, 1985 1989: 1989, 1989, 1990 1990: 1993, 1996, 2000

288

>1

99

1–10

41 32 9

4–25 2–3

80

1–36

199

(Continued)

Country Law year Election years

200

Summary Table 3A

Term

Districts

Ballot

Rule/Formula

Seats

No.

Magn.

128–166

28–42

3–5

Single transferable

Proportional Droop

508

508

1

Single

Majority/2nd round runoff

508 333 175

100 35

2–4 5

Single Limited 4/5

1892: 1892, 1895, 1897, 1900, 1904, 1909, 1913

508

508

1

Single

Majority/2nd round runoff

1919: Universal MS 1919, 1921 1921

508–535

40–54

1–28

Closed list

Proportional – d’Hondt

535 365 170

Closed list 1 nat. 15

365 *11

Q. plurality/Proportional Q. plurality 25%/ Proportional – Hare

Ireland 1922: Universal MS 1922, 1923, 1927, 1932, 1933, 1937, 1938, 1943, 1944, 1948, 1951, 1954, 1957, 1961, 1965, 1969, 1973, 1977, 1981, 1982, 1987, 1989 1992, 1997, 2002 Italy 1870: 1870, 1874, 1876, 1880 1882: 1882, 1886, 1890

1923: 1924 //

4

5

Plurality

Threshold

1946: 1946, 1948

5

556–574

1953: 1953

5

590 381 209

31

5

596–630

31–32

5

630 476 154

476 1 nat.

1956: 1956, 1963, 1968, 1972, 1976, 1979, 1983, 1987, 1992 1993: 1994, 1996, 2001

Netherlands 1815: every year 1848: every two years 1887: 1888, 1891, 1894, 1897, 1901, 1905, 1909, 1913, 1917 1918: Universal MS 1918, 1922, 1925, 1929, 1933

31

*19

Open list

Prop. (2 tiers) – Hare/Imperiali

Closed list

Majority/Proportional Majority Proportional

*20

Open list

Prop. (2 tiers) – Hare/Imperiali

1 154

Double: Single Closed list

Parallel: Plurality Proportionality – Hare

Public

Indirect

2

Secret Multiple

Direct Majority/2nd round plurality

3 estates

3 4 4

100

100

1

Single

Majority/2nd round plurality

4

100–150

1 nat. (18 lists)

100–150

Open list

Proportional – Hare

0.5–0.75% nat.

201

202

Summary Table 3A

(Continued)

Country Law year Election years 1933: 1937 // 1946, 1948, 1952, 1956, 1959, 1963, 1967, 1971, 1972, 1977, 1981, 1982, 1986, 1989, 1994, 1998, 2002, 2003 Norway 1815: 1815, 1817, 1820, 1823, 1826, 1829, 1832, 1835, 1838, 1841, 1844, 1847, 1850, 1853, 1856, 1859, 1862, 1865, 1868, 1870, 1873, 1876, 1879, 1882. 1885, 1888, 1891, 1894, Universal MS 1897, 1900, 1903 1906: 1906, 1909, 1912, 1915, 1918 1919: 1921, 1924, 1927, 1930, 1933, 1936 // 1945 1949 1953: 1953, 1957, 1961, 1965, 1969, 1973, 1977, 1981, 1985 1985: 1989, 1993, 1997, 2001

Term

Districts

Ballot

Rule/Formula

Threshold

Seats

No.

Magn.

4

100–150

1 nat. (18–19 lists)

100–150

Open list

Prop. (2 tiers) – Hare/d’Hondt 1.67% nat.

3

114–117

33–45

*3

Public Multiple

Indirect Plurality

Secret

3

123–126

123–126 1

Single

Direct Majority/2nd round plurality

3

150

29

*6

Open list

Proportional – d’Hondt

4

155

19–20

*8

Open list

Proportional – m. St-Lagüe

4

165

19

4–15

Open list

Prop. (2 tiers) – m. St-Laguë

4

4% nat.

Portugal 1820: 1820:

Public

1822: 1822 1826: 1828 // 1834, 1838, 1840 1852: 1852, 1856

Direct Plurality 4

>1

Multiple

Indirect Plurality Direct Q. plurality 25%/2nd round plurality

1

Single

Majority/2nd round plurality

1

Single

Plurality

1 3–6

Single Limited 2/3

>1

Multiple

Plurality

1 >1

Single Closed list

Coexistence Plurality Proportional

4

Limited 3/4

Plurality

111

4

1859: 1859 1878: 1878, 1881 1884: 1884–1895

Plurality

151 79 72

1895: 1895 1896: 1896 2 1901: 1901–1911

Indirect Plurality

149–155

203

204

Summary Table 3A

(Continued)

Country Law year Election years

Term

Districts Seats

No.

Ballot

Rule/Formula

Coexistence Proportional Plurality

Magn.

1911: 1915 // 16

4

4

Closed list Limited 3/4

230–247

22

1–55

Secret Closed list

Proportional – d’Hondt

Spain 1810: 1810

291

34

>20/7/3/1

Public Multiple

Indirect Plurality/majority/lots

1812: 1812 // 1820, 1822, 1823 //

149

33

≥1

Single/multiple

Plurality/majority

1834: 1834

188

49

≥1

Single/multiple

Majority/3rd round plurality

1836: 1836

258

49

>1

Multiple

Direct Majority/2nd round plurality

1836: 1836

337

49

>1

Multiple

Indirect Plurality/majority

1837: 1837, 1839, 1840, 1841, 1843, 1844

375

49

>1

Multiple

Direct Majority/2nd round plurality

1975: Universal MS 1975, 1976, 1979, 1980, 1983, 1985, 1987, 1991, 1995, 1998, 2002

4

Threshold

1846: 1846, 1850, 1851, 1853

349

349

1

Single

Majority/2nd round plurality

1854: 1854

375

52

>1

Multiple

Majority/2nd round plurality

1857: 1857, 1858, 1863, 1864

349

349

1

Single

Majority/2nd round plurality

1865: 1865, 1867

381

82

1–7

Limited (single)

Majority/2nd round plurality

1868: 1869, 1869, 1871, 1872, 1873, 1876

379–391

384

1–6

Limited (single)

Plurality

392–409 307–311 95–98

307–311 1 26–28 3–8

Single Limited 2/3

473

63–66

1–18

Secret Limited 2/3

Q. plurality 20–40%/2nd round plurality

350

52

1–43

Closed list

Proportional – d’Hondt

Public

Indirect

1878: 1879, 1881, 1884, 1886 Universal MS 1891, 1893, 1896, 1898, 1899, 1901, 1903, 1905, 1907, 1910, 1914, 1916, 1918, 1919, 1920, 1923 // 1931: 1931, 1933, 1936 // 1977: 1977, 1979, 1982, 1986, 1989, 1993, 1996, 2000

Plurality

3% dis.

Sweden 1810: every 5 years 1844: every 3 years

5

4 estates

205

206

Summary Table 3A

(Continued)

Country Law year Election years 1866: 1866, 1872, 1875, 1878, 1881, 1884, 1887, 1887, 1890, 1893, 1896, 1899, 1902, 1905, 1908 1909: Universal MS 1911, 1914, 1914, 1917, 1920 1921, 1924, 1928, 1932, 1936, 1940, 1944, 1948 1952: 1952, 1956, 1958, 1960, 1964, 1968 1970: 1970, 1973, 1976, 1979, 1982, 1985, 1988, 1991, 1994, 1998, 2002 Switzerland 1848: Universal MS 1848, 1851, 1854, 1857, 1860, 1863, 1866, 1869 1872: 1872, 1875, 1878, 1881, 1884, 1887, 1890, 1893, 1896, 1899 1900: 1902, 1905, 1908, 1911, 1914, 1917

Term

Districts

Ballot

Rule/Formula

Indirect (rural), direct (urban) Plurality

Seats

No.

3

214–230

173–201 1 2

Secret Single/ multiple

3

230

28–56

Closed list

Direct Proportional – d’Hondt

4

Threshold

Magn.

*4 *8

4

230–233

28

*8

Closed list

Proportional – m. St-Laguë

4

349–350

28–29

2–40

Open list

Prop. (2 tiers) – m. St-Laguë

3

111–128

47–49

≥1

Public Single/multiple

Indirect Majority/3rd round plurality

3

135–147

48–52

≥1

Secret Single/multiple

Majority/3rd round plurality

3

167–189

49

≥1

Single/multiple

Majority/2nd round plurality

4% nat. or 12% dis.

1918: 1919, 1922, 1925, 1928 1931, 1935, 1939, 1943, 1947, 1951, 1955, 1959, 1963, 1967, 1971, 1975, 1979, 1983, 1987, 1991, 1995, 1999

3 4

United Kingdom 1832: 1835, 1837, 1841, 1847, 1852, 1857, 1859, 1866, 1868

25–26

1–34

Multiple, cumulative

Direct Proportional – d’Hondt

Public 658 153 505

1868: 1874, 1880

Plurality 153 248

1 2–4

Single Multiple Secret

1885: 1885, 1886, 1892, 1895, 1900, 1906, 1910, 1910, Universal MS 1918, 1922, 1923, 1924, 1929, 1931, 1935 1949: 1950, 1951, 1955, 1959, 1964, 1966, 1970, 1974, 1974, 1979, 1983, 1987, 1992, 1997, 2001

189–200

5

Plurality

658 196 422 40

196 211 13

1 2 2–4

Single Multiple Limited

615–670 567–670 37–54

567–670 1 18–27 2–3

Single Multiple

625–659

625–659 1

Single

Plurality

Plurality

Source: See Appendix: Notes and Sources for Summary Tables at the end of the book.

207

208

Summary Table 3B Country Law year

Western Europe: Presidency Election years

Term

Rule

1951, 1957, 1963, 1965, 1971, 1974, 1980, 1986, 1992, 1998

6+6

Majority/2nd round runoff

1919, 1925, 1931, 1937 // 1950, 1956, 1962, 1968, 1978, 1982

6+6

College majority/3rd round runoff

1988:

1988, 1994, 2000

6+6

Majority/2nd round runoff in College

France 1848: 1962:

1848 // 1965, 1969, 1974, 1981, 1988, 1995

7+7

Majority/2nd round runoff in Assembly Majority/2nd round runoff

2001:

2002

5+5

Germany 1918:

1925, 1932 //

7+7

Majority/2nd round plurality

1945, 1959, 1966, 1973, 1974, 1976, 1983, 1990, 1997

7+7

Majority/single transferable vote

1976, 1980, 1986, 1991, 1996, 2001

5+5

Majority/2nd round runoff

Austria 1945:

Finland 1919:

Ireland 1937:

Portugal 1976:

Source: See Appendix: Notes and Sources for Summary Tables at the end of the book.

10 France: Reform-mongering Between Majority Runoff and Proportionality Gerard Alexander

For perhaps no other country do we have so much ‘evidence of the politician’s belief in the importance of electoral systems as political weapons . . . Between 1870 and 1940 the parliamentary electoral system was changed seven times, and between 1945 and 1988 five times’ (Criddle, 1992: 110). As this suggests, France stands out in many analyses for its unusually frequent alteration of electoral rules. Change at the national level has also been shadowed in recent decades by change in the rules governing subnational elections, both municipal and, later, regional. These changes, too, were typically related to partisan strategizing. This chapter argues that a pattern can easily be discerned: most consistently, parties in a position to design the rules governing the next elections were, in the French case, unembarrassed to change the existing ones if others substantially better served their purposes. This is true of the parties of the postwar Fourth Republic (1946–58) who imposed first proportional representation and then undermined it when it suited their purposes. It was true of the Gaullists who sponsored the sweeping institutional changes of 1958, including a greatly strengthened presidency and a return to France’s historic (pre-1945) procedure of majority rule with a second-round runoff. And it was true of both the socialists and centre-right parties who favoured, and successively imposed, opposing electoral rules in the 1980s. The majority runoff rules which have predominated since 1958 have been associated with a dramatic shift from record levels of cabinet instability to substantial cabinet stability. To an extent, we can use this record of change as a form of experimentation with which to scrutinize that correlation. This chapter also argues that the record suggests, instead, that electoral rules tout court had only limited effects even on the dimensions of the party system, and may have mattered more only when operating in conjunction with other rules of the political game. What this record emphasizes above all is that institutions are particularly vulnerable to revision when two conditions are met: ‘losers’ from existing rules remain within the system and volatility periodically endows them with both the power and the motive to revise those rules. Finally, this chapter argues that we should consider the modern 209

210

Gerard Alexander

French case a cautionary tale for those approaches which assume that formal decision rules become highly ‘sticky’ or path dependent once implemented. Given the French record of change, it is not obvious what constitutes the ‘initial rules’ for a brief analysis; accordingly, the discussion below arbitrarily begins with the re-establishment of French democracy in 1944–45.

The initial rules France, by the time it was liberated from Nazi occupation in 1944, had a history of electoral politics characterized by at least two prominent features. First, the seven decades preceding war and occupation had witnessed durable contestation over electoral rules. Second, and as a result of this, both majority runoff and (nominally) proportional representation rules had become familiar options. Most common before 1940 were majority rules, with either single- or multi-member districts and with an absolute majority required for victory in the first round but only a plurality (relative majority) in the second. In 1919–24, France had a set of rules adopted as proportionalist but which had decidedly non-proportional results and which are too complex and internally conflicted to describe briefly. The resulting shift in seats produced a centre-left majority in 1924 which restored majority rule. This was not the limit of electoral-rule contestation: the conservatives, better able to form first-round pacts, favoured a full-scale shift to single-round plurality rules, but they proved unable to impose them on the other political sectors. As a result, majority runoff in single-member districts remained in effect for the duration of the interwar period.

The choices: initiatives and processes of reform This backdrop may have helped inform the resumption of the unusually frequent pace of electoral rule change after the Second World War. For heuristic purposes, we can identify three broad postwar electoral rule ‘regimes’: initial proportional representation (1945), highly manipulated ‘proportionalism’ (1951–56), and majority with a second round (1958–86, 1988–). The first round of contention, in 1944–45, was the least extended when it came to electoral rules (as opposed to constitutional features). In the wake of liberation, the centrist radicals, rich in local notables and ambidextrous alliance options, were sympathetic to a restoration of the majority rule which had benefited them historically so greatly. But political processes in the war’s immediate aftermath were rapidly dominated by two sectors which favoured PR. The christian-democratic MRP, conservatives and many of those around Charles de Gaulle did so because they perceived proportional representation as the best defence against a communist-led majority in the assembly. The large, national and organized parties of the left – the socialist SFIO and the communist PCF – like the MRP, favoured proportional representation because it would logically undermine local notables and strengthen organized parties, moving political debate and competition onto their preferred terrain of programmatic discipline within parties and differences between parties. Given their organizational preponderance, these diverse forces were able relatively easily to impose closed-list

France 211

proportional representation in multi-member districts (d’Hondt). This system governed the three general elections of 1945–46. But seeds of change were already present. ‘Gaullism’ was a force powerful in the Resistance but only beginning to be organized politically as such. Central to it was a project of institutional refashioning, first laid out in some detail in de Gaulle’s 1946 speech at Bayeux, ‘which the General took two months to write, twenty-seven minutes to deliver, twelve years to think over and ten years to put into practice’ (Lacouture, 1993: 129). This project was broadly majoritarian, not surprising for a force which had a reasonable expectation of being a leading party: de Gaulle had unique draw as a national figure and his first party (the RPF) became a leading one in the country almost immediately after its founding. After 1946 the Gaullists’ agenda was initially focused not on electoral rules but on executive– legislative relations, particularly the call for a strengthened presidency. But in so far as any force was gestating an interest in a wholesale shift to more majoritarian, even plurality, electoral rules, it was the Gaullists.

The post-Second World War rules The 1945–46 rules facilitated the emergence of a meaningfully multi-party parliament. In 1946, for example, the PCF won 30 per cent of seats, the SFIO 17 per cent, the MRP 29 per cent and the (disorganized) centre-right 24 per cent. This arithmetic was joined to an institutional matrix characterized by a weak president, a cabinet with minimal powers to discipline the assembly and a weak upper house. The goal for many institutional architects emerging from the Resistance had been to avoid concentrations of power, and this overall design accomplished that admirably – so much so that the centre/centre-left forces which had helped champion the new order soon became sympathetic to a major revision of the electoral rules in particular. In a series of episodes stretching from the war’s end through, as we will see, to the 1980s, the stability of this particular institution was crucially influenced by the actions of those forces ‘losing’ under the existing electoral rules (Clemens and Cook, 1999). By 1947, the break was complete between a strong communist PCF and all other parties, and a Gaullist formation sudden rose which was bent on a major constitutional revision. The perceived threat from the left and right led the centrist parties in between to use their parliamentary majority to revise the electoral system in a way designed to advantage parties in alliances (themselves) and weaken isolated parties near each pole. This purpose was made perfectly explicit by a prominent complexity in the new system: it created two electoral regimes within France. In Paris, where the centrist parties were weak, proportional representation (Hare) was implemented to limit their defeat. In all other areas of the country, the new rules permitted apparentement (the linking of party lists) and ‘gave an immense advantage to those parties which could join in an electoral alliance likely to win an absolute majority of votes’ in a district – what Williams terms ‘bastard proportional representation’ (Cole and Campbell, 1989: 80; Williams, 1966: 324–31, 533–7).

212

Gerard Alexander

Outcomes were thus to be decisively affected not only by the rules but also by a non-institutional factor (in addition to voting preferences): the pattern of alliances among centrist parties. While not systematic, alliances were sufficient in 1951 to ensure a majority of seats from which centrists could plausibly form coalitions. As was intended, the parties worst positioned to ally were punished: the 47.6 per cent of votes earned by the PCF and the Gaullists won them a total of 37.4 per cent of seats. Not surprisingly, the PCF maintained its preference for proportional representation and now increasingly the Gaullists were sympathetic as well. In this context, however, it was the centrist parties which had the power to change the rules and substantial interest in doing so, though not in the direction the PCF and RPF wished. But progress toward yet another set of rules was aborted by the 1956 elections, in which centrist alliances were even narrower, diluting the new rules’ disproportional effects and resulting in a severely fractured assembly. The Fourth Republic’s mixture of multi-partism, weak executive powers and multiple dimensions of conflict resulted in one of the highest levels of cabinet instability known to modern democracies (an average six-month duration). While the dance of governments and cabinet members was frustrating and even farcical at times, it is not always clear it had profound policy consequences. These same years were ones of wide-ranging legislation on postwar reconstruction, the expansion of France’s then-modest welfare state, the reconstruction of its colonial apparatus, Cold War defences and cooperation, and the founding steps of European integration. Nor did cabinet instability result in any perceptible decline in support for democracy as the preferred regime type; there seems little reason to conceive of the events of 1958 as a case of ‘re-equilibration’ during a ‘democratic breakdown’. What is perhaps most striking was the widespread and durable conviction that key problems could be solved not through a different regime type but through institutional redesign. From the viewpoint of this conviction, what was regrettable was, in René Coty’s words, ‘the habit of shooting at the pianist when what is wrong is that the piano is out of tune’ (Pickles, 1971: 267). Beginning in the early 1950s and intensifying as the decade progressed, parties and leaders advanced diverse reform proposals aiming at change not only in electoral laws but in almost every arena of the national institutional structure. These agendas coincided with the worsening of the politics surrounding the Algerian war. The result was the second round of major institutional revision since 1946, one substantially more far-reaching than that of 1951. The pace in this revision effort was set by de Gaulle and his supporters, whose proposed reforms concerned a dramatically strengthened president and, in the end, a return to majority runoff elections. Innovation thus occurred along two dimensions of institutional design – both electoral rules and the powers of the offices filled via those rules – which the Gaullists crucially believed were intimately related. It is true that de Gaulle had for years typically been most explicit on the need for a horizontal separation of powers. But increasingly he and especially his main constitutional adviser, Michel Debré, argued that reform had to address both issues at once. Trade-offs on one dimension required trade-offs on the other, in order to approximate the desired outcome: governments as stable and effective as those in certain other

France 213

democracies. In their eyes, a French cabinet’s accountability to parliament was acceptable if that parliament contained parties capable of producing stable majorities, such as Britain’s. They believed this outcome hopelessly unlikely under PR, but they also knew a full-scale shift to British-style plurality rules was unfeasible (at least immediately). So if the new order had electoral rules not quite as majorityinducing as Britain’s, it had to compensate for that with a stronger executive than Britain’s. The Gaullist project for electoral rules was distinctive in being part of a much broader and quite integrated package. The more immediate Gaullist proposals concerned executive–legislative relations and powers, which, unlike the electoral rules, were to be addressed in the constitution itself. The revision proposal voted on in the 1958 referendum was hammered out in a matter of weeks within a committee including Gaullists, socialists and christiandemocrats, as well as Félix Houphouet-Boigny, a parliamentary deputy from the Côte d’Ivoire. The rapid, chaotic method of drafting and consultation does not always make it easy to disentangle each party sector’s views on each provision, but in most key ways the final product corresponded closely with the agenda de Gaulle had been advocating since 1946. This proposal gained adherents reaching through the right and the centre and deep into the centre-left. As a result, it attracted more debate than real contestation. It was bitterly opposed by the PCF, by small radical left and new left splinter groups, by several socialist leaders (including François Mitterrand) and by a rightist fringe. But these forces were entirely unable to promote an alternative institutional design and several essentially did not actively campaign at all. The proposal won nearly 80 per cent of votes in every département. The issue of the electoral law was only taken up afterward, by the cabinet and in a political setting awed by de Gaulle’s massive apparent electoral mandate. Preferences over electoral rules did not involve a rigorous consistency on the part of the Gaullists. As Williams notes, ‘In 1948, when the [Gaullist] RPF was the strongest single party in the electorate [in surveys], it favoured a large-constituency majority system. When its strength declined [in the following years], however, PR . . . was probably the private choice of most of the leaders’ (Williams, 1970: 102 n.1). By 1958, the Gaullists were once again in a position to reasonably expect they could organize a plurality of the electorate. Whereas proportional representation would return a strong PCF to the parliament – and the PCF favoured such a system – more majoritarian rules would magnify Gaullist strengths. Thus Gaullists championed either of two possible rules: the majority runoff or a multi-member district system by plurality rule designed to reward even more heavily the leading list. Some Gaullists who supported the latter perceived it as a stepping stone to plurality rules. Debré later insisted that, regarding the electoral rules as much as anything, the 1958 ‘reform was not as strong as I would have wished’. His preference was for ‘an election method as close as possible to the British-American system’, which would force parties to make the ‘necessary fusions to win in a brutal but clear struggle’ (Debré, 1981: 7–8). But he was required to make concessions to the forces that were simultaneously non-communist and non-Gaullist. These parties concluded that of the two likely options, the majority runoff was the less likely to

214

Gerard Alexander

result in a thorough national Gaullist sweep. This emerged as the non-communist consensus and was adopted.

The 1958 rules and after The new institutional design launched in 1958 thus included a strong presidency (s)elected independently of parliament (by a limited electoral college), a slew of key changes in parliamentary procedure and a restoration of majority rule in singlemember districts for the parliament’s key lower chamber. The threshold for advancing to the second round evolved: it was set at 5 per cent of first-round votes in 1958 and was raised to 10 per cent and 12.5 per cent of registered voters in 1967 and 1976 respectively. Even without the shift to outright plurality rules, this overall transformation stands out in at least two ways. It remains one of the most dramatic institutional reforms in any postwar democracy. And 1958, as with France’s 1951 electoral-rule change, represents one of the relatively rarer institutional revisions in the direction of substantially reduced, rather than greater, proportionality (Lijphart, 1994: 56). The consequences of this new set of the rules of the political game are considered in the next section. But it is worth noting that even then electoral rules were not a settled matter. In 1962 de Gaulle led a successful referendum campaign to make the presidency directly elected. And in the 1980s the lower chamber’s electoral rules were once again contested. In 1981 the socialists (by now the PSF) won the presidency and an absolute majority of National Assembly seats, and found themselves quite comfortable with the Fifth Republic’s broad institutional design. But by 1985, the PSF faced general elections in a more pessimistic mood having declined sharply in the polls, suffered a string of by-election defeats and seen a centre-right alliance win a plurality of the votes in the 1984 Euro-elections. It was in this context that the PSF enacted a shift to PR, a set of rules that could be expected to have three main effects. First, it would minimize rather than magnify the decline of the PSF. Second, it would facilitate the entry to parliament (and therefore possibly the electoral appeal) of the rising extreme-right National Front, which had scored 10 per cent of votes in the Euro-elections and which at this stage was understood to take votes otherwise destined for the centre-right. Finally, for both these reasons, proportional representation might well deprive the ascendant centre-right of an absolute majority in the Assembly. Not surprisingly, the centre-right opposed the shift. President Mitterrand’s gambit exerted an effect in the intended direction – diminishing what the centre-right would otherwise have won – but not enough of one. The conservatives were able to form a stable majority government nonetheless, and promptly imposed a return to the majority runoff rules which stood to favour them by magnifying the governing bloc’s plurality status and excluding the far-right National Front. As with the 1986 shift to PR, ‘the government made no excuses for tailoring the electoral system to suit its own purposes’ (Cole and Campbell, 1989: 141). Just as self-servingly, the likely losers from the new shift bitterly opposed it, foremost among them the minoritarian and ally-less National Front. Since 1988,

France 215

majority runoff rules have been in place, and their most vocal and consistent critics have been the parties most punished by them: the National Front and other small parties lacking large election-time allies.

Political consequences In his 1946 speech on institutional design, de Gaulle cited the Greek thinker Solon’s view of the need for a proper ‘fit’ between a constitution and a specific society (‘for which people and in which era?’). One might understand France’s unusually frequent revision of the rules of the game as a notably uncertain search for the rules that best fitted their people in their era, a search for the rules that would best serve the public or collective interest. But it is more tempting, reviewing this record, to see it as a pattern of contestation between competing political forces over the rules which offered them the greatest sectoral advantages. This is an easy case to make regarding many of the smaller-scale changes. But even the Gaullist’s sweeping majoritarian and presidentialist project of 1958 was, it must be remembered, one carried out by a grouping poised at least since 1947 to organize a party that could contend for the electorate and led by the most obvious or natural candidate for the national presidency. This is not to deny broader agendas and deeper values. But it is to observe that France’s shifts in rules have been designed to serve their designers first and foremost. This discussion has highlighted two patterns. First, parties anticipating a substantial decline in their fortunes have repeatedly favoured more proportionalist rules, that is loss- (and hence win-) minimizing rules. This describes, for example, the calculations of many Gaullists during their decline after 1947 and the socialists in 1986. This also corresponds to a common interpretation of the broad European shift in the direction of proportionalism (see Chapter 9 introducing this part, and also Boix, 1999). Second, parties which are reasonable contenders for plurality status repeatedly favour more majoritarian rules, that is win- (and hence loss-) concentrating rules. Examples include the centrist alliance in 1951, the Gaullists in 1945–47 and 1958 onward, including in 1987–88, and – in a sharp break with their history – the socialists since the 1970s, with the brief exception of 1985–86. This association, too, can be seen in comparative perspective (Alexander, 2001: 263–4). Inevitably, France also offers possible exceptions to this self-interested pattern.1 It is thus tempting to trace the frequency of electoral rule change to the volatility of party size, organization and geographic distribution of appeal in the French party system. Various electoral sectors might suffer from specific electoral rules: the rising Gaullists from the 1951 rules, the PCF from the majority runoff and the socialists from the majority runoff in 1985–86. Sufficient volatility in party organizations, alliances and votes meant that at times these ‘losers’ became ‘winners’ long enough to change the rules. Volatility also ensured new winners could not confidently expect to stay winners indefinitely, and so had a motive as well as the capacity to change to rules more favourable to their more predictable levels of performance. Thus rising and declining forces changed the rules to their advantage

216

Gerard Alexander

when they could: the centrist alliance in 1951, the Gaullists in 1958, the socialists in 1986 and the centre-right in 1988. Did the electoral rules created in 1958 have the intended effects? Majority runoff rules are understood to exert a pressure toward the emergence of two blocs in the party system. In a technically accurate use of the word, Duverger and others label this trend ‘polarization’, which unfortunately easily elides into connotations of mounting inter-party hostility (Duverger, 1963: 226; see also Reed, 2001). A better term for the trend might be ‘bipolarization’, and it was this that Debré himself hoped for, arguing that ‘bipolarization’ is the sine qua non of effective governments backed by stable majorities. Sartori is among many who have concluded that the new electoral rules ‘reshap[ed] the French polity in a bipolar mold’ after 1958 in just this way (Andrews and Hoffmann, 1981: 334; Sartori, 1994: 11). But in fact it remains quite unclear to what extent this outcome can persuasively be attributed to this category of formal institution. By 1958, many people wanted to overcome the French Fourth Republic’s cabinet instability. Change of the electoral rules seemed one way to do that. But Colomer quite rightly reminds us that the cabinet instability of the Fourth Republic is not fairly ‘attributed to the electoral system of proportional representation . . . since a similar degree of Cabinet instability had existed in the previous French Third Republic’ under majority rules (Colomer, 2001: 128). The same could be said of the Fourth Republic’s multi-partism. Colomer is reminding us that different electoral rules can be associated with some of the same outcomes. For related reasons (concerning the interaction between formal rules and their social context), the same rules can be also associated with different outcomes. By the same reasoning, it is not at all clear why we should attribute the more stable electoral alliances and governing coalitions which have indeed emerged since 1958 to the Fifth Republic’s electoral rules, since they are basically the same rules as those of the Third Republic, a period of frequent fragmentation and instability. We must remember that Gaullists were just as critical of the Third Republic’s cabinet instability as of the Fourth’s. After all, de Gaulle’s 1946 Bayeux speech was organized around a critique of the weak pre-1940 governments which had failed to effectively confront Nazi power; that critique was the justification for his plea to the architects and citizens of the Fourth Republic not to repeat this mistake. Was it then an act of astonishing historical myopia to reinstall the Third Republic’s electoral rules? When these rules are considered in isolation, it is not clear why Gaullists would have expected them to exert the effect they desired. Cole and Campbell emphasize that double ballot rules are compatible with diverse outcomes, because those outcomes are also influenced, among other extra-institutional factors, by the extent to which parties form wide alliances on either the first or second round (Cole and Campbell, 1989: 20–1). When parties did not do so, as in many stretches of the Third Republic, contests in many districts had to go to the second round and resulted in the election of candidates with narrower appeal, preserving a role for greater diversity in the party system (as opposed to within parties). The result was multi-partism, coalition governments and, in this case, periods with average cabinet durations as short as those of the Fourth Republic: six months or

France 217

so. As Criddle describes some of these periods, ‘Bipolar in appearance at the second ballot, party politics remained multi-polar and fragmented in fact’ (Criddle, 1992: 109). In contrast, in Third Republic phases dominated by wide alliances, elections in many districts were determined on the first round, and the coalition governments formed by the alliances’ members were much more stable. Majority runoff rules were, so to speak, agnostic as to which of these patterns obtained. A comparable variation in outcomes can be detected in the Fifth Republic itself. Whereas the trend from 1958 into the 1980s was toward a party system with fewer ‘effective’ parties, the pattern since then has been toward party system fragmentation. The National Front has survived election-rule adversity, the greens emerged and consolidated a modest base, far-left vehicles of various kinds emerged, portions of the centre-right UDF split away and dissident socialists launched several major splinters. Many of these entities appeared not only at the ballot box but also in parliament. The latter was typically made possible by nationwide, election-time agreements between the larger and these smaller parties to ‘scratch each other’s backs’, district by district. This has permitted smaller entities such as the greens, dissident socialists, independent conservatives and centrist christian-democrats, for example, to enter the Assembly, provincial governments and the European parliament since the 1990s. Considered in admittedly somewhat melodramatic terms, the spectre this raises is one in which the French party system, under majority runoff rules, might converge with, say, Israel’s since the 1980s, in which traditionally hegemonic parties of the centre-right and centre-left gradually shed votes and seats to more and stronger small parties and sometime allies. There, proportional representation rules facilitated the small parties’ emergence and persistence. In France, backscratching agreements make the same outcome possible even under rules which allegedly encourage consolidation. But whatever its destination, the alreadyexisting trend in France can be used as evidence of the profound limitations of institutional engineering. Under diverse rules in several periods across the twentieth century – including relatively stable majority runoff rules since 1958 – social divisions consistently recover expression within the party system as actors learn to subvert what might be considered the spirit of the rules while operating within their letter. That interpretation may be appropriate if we focus on that one particular outcome: the number of effective parties in the system. But it is misplaced if we consider another, perhaps more important one: politics in France since 1958 has generated relative stability in left and right party blocs and in coalition governments, to a degree unprecedented in the Third and Fourth Republics (the centre-right bloc gelled immediately in 1958; the left bloc did so only in the late 1970s). Even governments formed from the increasingly populated Assemblies of the 1990s and early 2000s stably served out their terms.2 (Of course, it is another matter whether greater cabinet stability had any further consequences, such as substantially different policy outcomes.) A distinct possibility is that greater cabinet stability is explained by a formal rule of the game, just not the rules governing parliamentary elections. It can plausibly be traced instead to the Fifth Republic’s

218

Gerard Alexander

powerful and directly elected presidency. The left and right blocs which have provided stable governments since 1958 can usefully be seen as alliances of those who ally to vote together on the second round for the high-stakes presidency created by the 1958 constitution (this argument is made by Criddle, 1992: 109, among others). The new electoral rules of 1958, after all, were part of a package of changes, including presidential powers, cabinet powers, National Assembly procedures, the upper chamber’s role and a constitutional court, in addition to the rules governing legislative elections. The transformation of 1958 created a novel institutional situation. Since the presidency’s powers were unprecedented in the country’s democratic history, so, too, was this particular combination or package of rules. In this crucial sense, it is not odd that de Gaulle accepted a return to the electoral rules of a Third Republic which he otherwise criticized; the Third Republic’s electoral rules were designed to operate in a very different overall institutional context than existed before 1940. This suggests that a given decision rule (say, a country’s legislative electoral rules) may not be interdependent with all ‘surrounding’ decision rules in one sense: it may well not be characterized by the institutional interrelatedness that theorists like Douglass North (1990: 95) argue contributes to path dependence. But it is interdependent with those other rules in another sense: in terms of the effects it exerts. This raises the distinct possibility that studies which examine a single decision rule as their independent variable – a decision rule operating within a more complex institutional matrix – may be crucially limited. One reason that these rules proved not to be path dependent is that these specific rules do not appear to require participants to amass the ‘specific assets’ which some path dependency accounts suggest can be key causes of persistence in institutions after their adoption (Pierson and Trowbridge, 2002). That is, it would not appear that learning how to operate within one set of rules constituted an expensive enough investment in knowledge and skills to lead them to strongly prefer not to shift to a different set of rules. Instead, French politicians felt comfortable operating within multiple electoral rules not only sequentially but even simultaneously. By the turn of the millennium, the protagonists of France’s political life confronted as complicated a combination of decision rules as any on offer in advanced liberal democracies. Consider electoral rules alone: national parliamentary and presidential elections are held under majority runoff rules, European parliament elections under PR, cantonal votes under majority runoff, regional assemblies under proportional representation with a significant threshold, and most mayoralties under majority runoff with a supermajority-manufacturing bonus to the winning slate. And at subnational levels, as at the national level, self-interested revision of rules has also occurred periodically. In this way as in the others highlighted, it is clear that the French case challenges a number of the major current theories about formal rules. What remains unclear is the extent to which these rules, and changes in them, can be demonstrated more systematically to have had substantial effects for political and social actors other than the politicians who care about them so greatly.

Table 10.1

France: National Assembly electoral results, 1945–2002

Year

Communist PCF v/s

Socialist SFIO/PS v/s

Radicals

1945 1946 I 1946 II 1951 1956

26/28 26/28 29/30 27/18 26/27

1958 1962 1967 1968 1973 1978 1981 1986 1988 1993 1997 2002

19/2 22/9 22/15 20/7 21/15 21/18 16/9 10/6 11/5 9/4 10/7 5/4

Centre MRP/CD v/s

Republican I/UDF v/s

Gaullist RPR v/s

Conserv.

v/s

Green V v/s

24/26 21/22 18/17 15/17 15/16

11/7 11/7 12/10 10/14 15/13

– – – – –

25/27 28/31 26/29 13/15 11/13

– – – – –

– – 2/1 22/20 4/3

13/12 13/12 13/13 14/16 15/17

– – – – 12/9

15/9 13/14 19/25 16/12 19/19 25/21 38/56 32/37 36/48 19/12 23/42 24/24

7/0 8/0 – – – – – – – – 1/2 1/1

– – – – – 2/0 1/0 1/0 1/0 11/0 7/1 4/0.1

11/12 9/8 13/0 10/0 12/6 – – – – – – –

– 4/7 6/9 – 10/11 22/26 19/13 – 18/22 – 14/19 5/5

20/43 32/42 32/41 44/74 24/37 22/30 21/17 42/47 19/22 40/78 16/23 34/62

22/29 9/10 4/2 4/2 3/3 3/2 3/1 3/2 3/2 4/6 – 4/2

1/0 – – – – – – 10/6 10/0 12/0 15/0.2 11/0

v/s

Right UFF/FN v/s

Key: PCF: Communist Party of France; SFIO: French Section of the Workers’ International; PS: Socialist Party; Radicals: RGR, UDSR, MRG, Radical; V: The Greens; MRP: People’s Republican Movement; CD: Democratic Centre; I: Independent Republicans; UDF: Union of French Democracy; RPR: Rally for the Republic, UG, ARS, RPF; Conservatives: Moderates, PRL, Peasants; UFF: Poujadists; FN: National Front; v/s: percentage of votes/percentage of seats.

219

220

Gerard Alexander

Notes 1. The most obvious may be the Communist PCF. Arguably, as a leading party it would have benefited from win-concentrating rules after Liberation. On the other hand, it could not initially be sure of its exact electoral appeal. And even a steady quarter of the vote would benefit from, say, plurality type rules only if PCF voters were highly concentrated territorially. This was by no means always the case. As such, even the PCF needed reliable allies, and prominently lacked them during its years of demonstrated strength from 1947 to the 1970s. This is why the reintroduction of majority runoff rules in 1958 led to a decimation of the PCF’s parliamentary delegation. So even this large party has most consistently benefited under PR. 2. The most notable exception – Alain Juppé’s, elected in 1995 and terminated in 1997 – ended for electoralist and other reasons, not because it lacked either a parliamentary majority or cohesion among the coalition partners.

References Alexander, Gerard (2001). ‘Institutions, Path Dependence, and Democratic Consolidation’, Journal of Theoretical Politics, 13: 249–70. Andrews, William, and Stanley Hoffmann (eds) (1981) Appendix: ‘Dialogue: Michel Debré and Conference Participants’, in The Impact of the Fifth Republic on France. Albany, NY: SUNY Press. Boix, Charles (1999) ‘Setting the Rules of the Game: the Choice of Electoral Systems in Advanced Democracies’, American Political Science Review, 93: 609–24. Clemens, Elisabeth S. and James M. Cook (1999) ‘Politics and Institutionalism: Explaining Durability and Change’, Annual Reviews: Sociology, 25: 441–66. Cole, Alistair and Peter Cambell (1989) French Electoral Systems and Elections Since 1789. Aldershot: Gower. Colomer, Josep (2001) Political Institutions. Oxford: Oxford University Press. Criddle, Byron (1992) ‘Electoral Systems in France’, Parliamentary Affairs, 45: 108–16. Debré, Michel (1981) ‘The Constitution of 1958, Its Raison d’Être and How It Evolved’, in William Andrews and Stanley Hoffmann (eds), The Impact of the Fifth Republic on France. Albany, NY: SUNY Press. Duverger, Maurice (1963) Political Parties. New York: John Wiley & Sons. Lacouture, Jean (1993) De Gaulle: The Ruler, 1945–1970. New York: W. W. Norton. Lijphart, Arend (1994) Electoral Systems and Party Systems. Oxford: Oxford University Press. North, Douglass (1990) Institutions, Institutional Change, and Economic Performance. Cambridge: Cambridge University Press. Pickles, Dorothy (1971) French Politics: The First Years of the Fourth Republic. New York: Russell & Russell. Pierson, Paul and Shannon O’Neil Trowbridge (2002) Asset Specificity and Institutional Development. Paper presented at the American Political Science Association annual conference, Boston. Reed, Bruce (2001) ‘Duverger’s Law is Working in Italy’, Comparative Political Studies, 34: 312–27. Sartori, Giovanni (1994) Comparative Constitutional Engineering. New York: New York University Press. Williams, Philip M. (1966) Crisis and Compromise: Politics in the Fourth Republic. Garden City, NY: Anchor Books. Williams, Philip M. (1970) French Politicians and Elections, 1951–1969. Cambridge: Cambridge University Press.

Further reading • Alexander, Gerard (2001) ‘Institutions, Path Dependence, and Democratic Consolidation’, Journal of Theoretical Politics, 13: 249–70.

• Alexander, Gerard (2002) Democratic Consolidation. Ithaca, NY: Cornell University Press.

France 221

• Andrews, William and Stanley Hoffmann (eds) (1981) Appendix: ‘Dialogue: Michel Debré • • • • • • • • • • •

and Conference Participants’, in The Impact of the Fifth Republic on France. Albany, NY: SUNY Press. Cole, Alistair and Peter Cambell (1989) French Electoral Systems and Elections Since 1789. Aldershot, UK: Gower. Comité National Chargé de la Publication des Travaux Préparatoires des Institutions de la V e République (1987) Documents pour Servir à l’Histoire de l’Élaboration de la Constitution du 4 Octobre 1958, 3 vols. Paris: Documentation Française. Criddle, Byron (1992) ‘Electoral Systems in France’, Parliamentary Affairs, 45: 108–16. Debré, Michel (1981) ‘The Constitution of 1958, Its Raison d’Être and How It Evolved’, in William Andrews and Stanley Hoffman (eds), The Impact of the Fifth Republic on France. Albany, NY: SUNY Press. Hoffmann, Stanley (1959) ‘The French Constitution of 1958: I. The Final Text and Its Prospects’, American Political Science Review, 53: 332–57. Lacouture, Jean (1993) De Gaulle: The Ruler, 1945–1970. New York: W. W. Norton. Pickles, Dorothy (1971) French Politics: The First Years of the Fourth Republic. New York: Russell & Russell. Wahl, Nicholas (1959) ‘The French Constitution of 1958: II. The Initial Draft and Its Origins’, American Political Science Review, 53: 358–82. Williams, Philip M. (1966) Crisis and Compromise: Politics in the Fourth Republic. Garden City, NY: Anchor Books. Williams, Philip M. (1970) French Politicians and Elections, 1951–1969. Cambridge: Cambridge University Press. Wright, Gordon (1950) The Reshaping of French Democracy. London: Methuen.

11 Germany: Partisan Engineering of Personalized Proportional Representation Marcus Kreuzer

This chapter assesses how the shadow which Weimar cast on the constitutional deliberations of postwar Germany shaped actors’ institutional preferences and thereby constrained partisan motivations. It proceeds in two sections. The first section exclusively focuses on the formal, institutional evolution of Germany’s electoral law from Weimar through the three post-1945 electoral laws. It sketches out the historical dynamic of these institutional changes and describes the key features of the 1949 electoral law and its 1953 and 1957 modifications. The second section chronicles the intra-party deliberations and inter-party bargaining that accompanied the three postwar electoral laws. By matching the actors’ institutional proposals with historical accounts about their intentions, it tries to disentangle to what extent historical learning constrained partisan considerations about maximizing seats and executive control.

The evolution of the German electoral system In 1945, the Germans and Allies alike were all too aware of the paradox that the Weimar Republic was created and destroyed by democratically elected parties. The parties’ ideological polarization, their ossified governance structures, their co-optation by interest groups and, above all, their fragmentation all decisively contributed to Weimar’s collapse. The Germans and Allies were equally aware that these deficiencies of Weimar party politics owed much – but by no means everything – to the perverse incentives created by the electoral system and, to a lesser extent, the governing institutions. Against this historical background, three things stand out about the evolution of Germany’s postwar electoral system. First, the 1949 electoral law and its successors are striking in the closeness with which their seat-related procedures – that is the elements directly affecting the translation of votes into seats – followed the same proportional representation logic that contributed so significantly to Weimar’s fragmentation. At first, the introduction of 242 single-member districts may seem like grafting an important majoritarian element onto the postwar proportional representation system. Yet, the 1949 electoral system has little in common with the recent proliferation of mixed electoral 222

Germany

223

systems. It worked like a regular proportional representation system since seats were proportionally awarded by a party’s state-level (Land) vote aggregates. This vote total was calculated by pooling all the single-member district votes cast for a party in a given state. The actual seat shares were awarded through a virtual, state-level district even though voters cast a single ballot for an actual single-member district’s candidate. The number of candidates elected from these state-level district lists was determined after deducting a party’s single-member districts seats from its overall seat total. Single-member districts thus mattered in the distribution of seats within the party while the virtual multi-member district determined a party’s overall seat share. Overall then, the underlying proportional logic of the 1949 electoral law meant that the majoritarian looking single-member districts did not in any way address the party fragmentation which plagued the Weimar Republic. Nevertheless, the 1949 electoral law addressed fragmentation in two ways. First, it introduced an electoral threshold in the form of requiring a minimum 5 per cent of vote share in any given state or the winning of at least one single-member district. The later district threshold requirement proved an important loophole through which smaller parties could slip which had a regional vote concentration or which were capable of coordinating strategic candidate withdrawals with larger parties (Kitzinger, 1960: 38–56). Second, the elimination of various vote pooling mechanisms – like apparentements and national adjustment seats – served as another attempt to discourage party fragmentation. In 1949 then, the seat allocation was just as proportional as in the Weimar Republic, especially given the new electoral system’s weak entry barriers and the very high magnitude of its virtual multi-member districts. The 1953 and 1957 electoral laws only slightly modified the basic functioning of the 1949 electoral system. The widening of the electoral threshold from the Land to the federal level in 1953 and increasing the one single-member district requirement to three in 1957 considerably raised the entry barriers for smaller parties, especially those with a regional vote concentration. However, the allocation of seats among parties unaffected by entry barriers became even more proportional in 1957 as parties were now able to nationally pool their state list votes. This vote pooling provision differed from conventional apparentements by being strictly limited to state branches of one and the same party. The Bavarian Christian Social Union (CSU), for example, was already considered too separate to pool votes with its sister party, the Christian Democratic Union (CDU) (Kitzinger, 1960: 23–4). The 1957 vote pooling provision thus turned Germany into a de facto single, national district with a magnitude of 516 (Shugart and Wattenberg, 2001: 19). The introduction in 1953 of a separate vote for party lists allowed voters to split their tickets but it had no discernible effect on how proportionally votes would be translated into seats. So, the seat-related electoral procedures closely followed the logic of a proportional representation system despite the majoritarian-looking adornments. Second, the new electoral law’s effort to personalize representation constituted a significant departure from the Weimar electoral system, arguably more significant than the changes in the seat-related electoral procedures. The most striking, personalizing element was single-member districts which, while being mere adornments

224

Marcus Kreuzer

with respect to the translation of votes into seats, allowed voters to refine their choices and had significant organizational consequences. In 1949, the single vote severely limited the single-member district’s personalizing effect since, given the aforementioned vote aggregation procedures, formal votes cast for individual candidates were first and foremost party votes. It made it impossible for voters to differentiate between their candidate and party preferences. The introduction of a second vote in 1953 permitted such ticket splitting and allowed voters to effectively reward and punish the performance of individual candidates irrespective of party considerations. Compared to the Weimar Republic, there is no question that the single-member districts personalized political representation (Kitzinger, 1960: 62; Kreuzer, 2000). The fact that this has not led to the same the degree of personal voteseeking as in single-member district systems may be attributable to the district’s large physical size (about twice the size as in France and the United Kingdom) and the possibility for candidates to hedge their career bets by simultaneously running on party lists. The abolition of Weimar’s apparentements and adjustment seats further personalized representation by making voters’ choices more transparent. It did so by protecting voters from casting their votes for one party and thereby unwittingly supporting candidates of another party with which it had formed an electoral alliance which frequently was undisclosed on the ballot. Similarly, national adjustment seats meant that voters elected candidates which they did not even know about. Adjustment seats were a controversial element of Weimar’s electoral system and were constitutionally challenged for violating the right to direct elections (Pollock, 1932: 59–60; Schanbacher, 1982: ch. 4, 8). Such vote-pooling mechanisms thus depersonalize representation by obscuring the transparency of electoral choices. Third, the new electoral system significantly departed from that of the Weimar by regulating pre-electoral, party-internal candidate recruitment. This recruitment involves nomination after an initial screening and shortlisting of candidates, selection from these nominees of an official candidate or candidates and their ranking on a party list (Kreuzer, 2001: 45–55). The candidate recruitment process is structured by various elements of the electoral system as well as by party law. The same electoral procedures meant to personalize electoral choices also served to decentralize and to a certain degree also ‘democratize’ the candidate recruitment process. The single-member districts democratized candidate recruitment in two ways. First, they multiplied the number of recruitment sites thus making it more difficult for central party leadership to control candidate selection. Second, it made candidate recruitment less hierarchical since it deprived party leaders of lists on which it could rank candidates. Such ranking conferred significant party-internal control since it directly affected a candidate’s career prospects (Kreuzer, 2000: 55–6). Furthermore, the abolition of the national adjustment seats that played a prominent role in Weimar eliminated another important leverage over candidate recruitment. It meant that no candidates could be selected and ranked at the national level (leaving it the prerogative of state parties). Such national adjustment seats were particularly attractive because, being based on the remaining votes from regular electoral districts, they virtually guaranteed election. Being a by-product

Germany

225

of the electoral formula, these remaining votes were constant across elections and unaffected by a party’s overall performance (Kitzinger, 1960: 23–6; Kreuzer, 2001: 57). These democratizing effects of the regular electoral mechanism were reinforced in further ways. Article 21 of the Basic Law (constitution) broadly required that parties have democratic governance structures while the 1949 electoral law stipulated that candidates be selected by secret ballot cast by party members in single-member or multi-member districts. In 1967, a separate party law reinforced and elaborated these democratic governance requirements in greater detail (Roberts, 1988: 97–8). Overall then, Germany’s postwar electoral system differed significantly, albeit to varying degrees, from Weimar’s electoral system. These departures were more significant for the candidate recruitment process and the nature of voters’ choices than for seat-related electoral procedures. Moreover, the 1953 and 1957 revisions of the original electoral law strengthened the changes the 1949 electoral law had undertaken vis-à-vis the Weimar Republic. The 1953 federalization of the electoral threshold and the 1957 three single-member district seats increased the entry barriers for new parties; the introduction of the second ballot in 1953 further personalized the voters’ choices, and the 1953 electoral law and 1967 party law stipulated more clearly the democratization of the candidate recruitment process.

The politics of institutional design The preceding bird’s eye perspective gives a rough indication of the extent to which postwar Germany’s electoral systems differed from their Weimar precursor. As far as actors’ institutional preferences are concerned, this macro perspective leaves the impression of considerable historical learning with respect to electoral rules affecting the transparency of voter choices and the democratization of candidate recruitment and a more modest learning with respect to seat-related electoral procedures. Corroborating this impression requires a more microscopic look at the politics of institutional design. In other words, drawing valid motivational inferences from the existing electoral laws requires a detailed look at the political circumstances and the extent to which they shaped individual actors’ institutional preferences along partisan or non-partisan lines. The following sections therefore analyse the politics behind Germany’s three postwar electoral laws. The 1949 electoral system will receive a disproportionate amount of attention since it comprised roughly 85 per cent of all the procedures used in the latter two electoral laws. This chronological survey largely demonstrates the considerable fluidity in parties’ institutional preferences and a gradual shift from non-partisan to partisan considerations. The principal driving force behind this shift was the decline in the overall political uncertainty under which actors operated and the resulting increase in available information politicians needed to calculate the particular costs and benefits associated with different electoral mechanisms. German politicians thus were no more disinterested than their counterparts elsewhere. They instead found themselves in a temporary political limbo during which they were unable to positively identify what electoral procedures would help

226

Marcus Kreuzer

them maximize seats. By default, they were therefore reduced to defining their institutional preferences in terms of redressing the deficiencies of Weimar’s electoral system. 1949 electoral law The 1949 electoral law evolved in two phases which distinguished themselves by their very different strategic environments. First, parties internally deliberated about their electoral system preferences. These deliberations began roughly in 1946 when the Allied occupational authorities started licensing new parties. These partyinternal deliberations came to a close in September 1948 with the formation of the Parliamentary Council which was authorized by the Allies to draft a new constitution. During this second phase, parties directly bargained with each other until the final law was adopted in June 1949. As we will see, the shift from the unstructured party-internal discussions to the highly formalized interactions in the Parliamentary Council significantly reshaped the parties’ institutional preferences. Intra-party deliberations: 1946 – July 1948 Partisan considerations certainly were on politicians’ minds when they began to articulate their parties’ original institutional preferences. The Social Democratic Party (SPD) leader Kurt Schumacher, for example, admonished his party not to let ideological principles get in the way of electoral exigencies. Both parties also used state election results to simulate the effects of different electoral systems. Still, partisan considerations played a subordinate role in shaping parties’ first institutional preferences. (This section primarily draws on Erhard Lange’s excellent and very thorough study in 1975.) The CDU/CSU institutional preferences were very fluid before 1948 and for obvious reasons. The party was just starting to build an organizational infrastructure, at both the state and especially the federal levels, and also tried to integrate different bourgeois constituencies with very diverging interests. The lack of a central leadership meant that the party’s preferences varied across the different occupational zones as well as grassroots constituencies. Party-internal debates consequently were wide open and first and foremost driven by considerations of finding institutional solutions to two of Weimar’s deficiencies. Erhard Lange (1975: 189) reports German politicians had a vested interest in seeking constitutional solutions as it implied that Weimar collapsed because of poor institutional design rather than, as Allied propaganda emphasized, authoritarian political culture. Concentrating on institutional engineering thus became an explicit strategy for politicians to affirm their own importance and underscore the possibility of a political fresh start. First, CDU/CSU politicians were greatly concerned with Weimar’s party fragmentation and the resulting ungovernability which had undermined public trust in liberal democracy and provided cheap political ammunition to anti-democratic parties. Second, CDU/CSU politicians were greatly concerned with Weimar’s polarization. One party representative pointedly noted that ‘the cancer of our parliamentarism was the over boarding class warfare and materialism’ (Lange, 1975: 143–4). In many ways, the very self-definition of the CDU/CSU as an interdenominational people’s

Germany

227

party sought to address these two problems. But party leaders also were clearly aware that Weimar’s excessive proportional representation system directly contributed to polarization and fragmentation. They questioned therefore whether ‘the mindless rule of numbers, the counting of every single vote, each political splinter group is the ideal of democracy’ (Lange, 1975: 143–4). Yet, by the same token, the CDU/CSU’s enthusiasm for majoritarian single-member district electoral systems was tempered by Germany’s negative experiences with the double ballot system prior to 1918 when the party system was also polarized and fragmented. If anything, the CDU/CSU tried to mitigate the perceived shortcomings of proportional representation and plurality systems by grafting elements of the former onto the latter. Its somewhat vague 1946 proposal expressed a preference for single-member districts but tempered its majoritarian effects with compensatory proportional representation elements. The CDU/CSU preferences remained equally vague when it formally constituted itself as a national party in February 1947. Weimar also cast its shadow over the SPD’s deliberation. Weimar’s party fragmentation and the Anglo-American exile of prominent SPD leaders dampened the party’s long-standing enthusiasm for a highly proportional electoral system. The SPD also proposed a mixed electoral system but one that sought to personalize a proportional representation system rather than proportionalize a majoritarian system. It anticipated that such a move would reduce party fragmentation. The addition of single-member districts was strongly supported by various Southern states which anticipated that such a modification would weaken the traditionally Northern German dominated national leadership over candidate recruitment. The SPD was less interested in the moderating effects of single-member districts since it still was a highly programmatic party viewing elections as battles of ideas rather than non-ideological leadership selection contests. They thought that the programmatic thrust of elections would be particularly important to ideologically capture the large number of Eastern refugees, calculating that, having lost all their property and wealth, these refugees would become a new underclass available for the left’s ideological encapsulation (Lange, 1975: 236, 248). The party’s 1948 proposal, far more specific than that of the CDU/CSU, reflected the lessons the SPD sought to learn from Weimar. It closely mirrored the ultimate 1949 electoral law with candidates elected from single-member and multi-member districts with the latter determining the overall distribution of seats. The proposal also contained a 5 per cent state-based electoral threshold provision. In the immediate postwar years, the legacies of Weimar’s ungovernability, polarization and party ossification clearly informed parties’ dialogue about their institutional preferences. Yet it would be incorrect to assume that historical learning had reshaped parties’ institutional preferences along collective, non-partisan lines. We already saw that party leaders exhorted their followers to place electoral expediency ahead of ideology, let alone considerations for the common good. On closer analysis, actors’ stated goals to learn from the institutional lessons taught by Weimar owed less to civic-mindedness or non-partisanship than it did to the uncertain political environment under which they operated. Various factors contributed to this uncertainty but their net effect was to prevent any meaningful

228

Marcus Kreuzer

assessment of the different electoral procedures’ political pay-offs. Lange reports that ‘after 1945, it was extremely difficult to gain a clear sense of one party’s possibility as well as that of others, so that great uncertainty existed over which electoral procedures would advance parties’ self-interest’ (Lange, 1975: 767). This uncertainty had many different sources. Parties, for example, did not know whether the Soviet zone would become part of a new Germany nor could they easily keep track of the massive population migration fuelled by Eastern refugees. After the first local and state elections were held, parties used these results to simulate the pay-offs of different electoral systems. But the results were so inconclusive that the same electoral system which proved favourable to a party in one state did not in another. The states simply varied too much in their socio-economic make-ups to be useful laboratories. Parties, especially the CDU/CSU, had great difficulties ranking institutional preferences for reasons that went beyond the difficulty of assessing the political pay-offs of different electoral procedures. State parties were less interested in such ramifications than they were in the organizational consequences of electoral procedures, especially the control over candidate recruitment. The federal party, in turn, tried to estimate how electoral procedures would affect the electoral integration of the different bourgeois constituencies to which it tried to appeal. This integration of the centre-right political spectrum was a paramount consideration for the CDU/ CSU since it wanted to avoid at all costs the hyperfragmentation of the Weimar right (Kreuzer, 2001: 91–132). Finally, the CDU/CSU and SPD were already thinking about how different electoral procedures would affect smaller parties which could potentially be crucial coalition partners after the election. Given the parties’ unstable preference ranking and the lack of important information about voters, it became exceedingly difficult for parties to base their institutional preferences on narrow partisan considerations. In such an environment of uncertainty, historical legacies shaped, albeit more by default than by collective intentions, actors’ institutional preferences. Inter-party bargaining: September 1948 – June 1949 With the Parliamentary Council’s appointment in September 1948, the political playing field became far more clearly defined. Actors now directly bargained with each other, operated under clear decision-making rules and precisely knew each other’s voting strength. Seats in the Council were awarded proportionally according to party strength in the state assemblies giving the CDU/CSU 27 seats, the SPD 27, the Free Democratic Party (FDP/LDP/DVP) 5, the German Party (DP) 2, the Centre Party (Zentrum) 2 and the Communist Party of Germany (KPD) 2. Institutional preferences now became subjected to partisan forces. The resulting three bargaining rounds led to the adoption of the first electoral law in June 1949. Table 11.1 summarizes the result of these bargaining rounds as well as the parties’ opening bargaining positions. A closer look at the bargaining over the 1949 electoral law shows that the more clearly defined strategic context greatly increased the importance of partisan considerations compared to the earlier party-internal deliberations. The June 1949 electoral law constituted a clear victory for the SPD over the CDU/CSU. Recognizing the crucial voting role that smaller parties assumed in the

Germany

Table 11.1

229

Germany: bargaining rounds over the 1949 electoral law

Opening positions, September 1948 SPD: Multi-member districts with 6 seats. Candidates with highest votes directly elected. Unused remaining votes recycled through federal adjustment seats. CDU/CSU: 300 single-member districts by plurality vote. Wasted single-member district seats recycled trough 50 compensatory list seats. First draft, April 1949 Provisions: 200 single-member districts and 200 multi-member district seats with voter casting a single ballot for single-member district candidates. Seats awarded at state level, proportionally to aggregate single-member district vote total. Single-member district seats deducted from overall seat share to determine the number of list seats. Support: All parties but CDU/CSU. Second draft, May 1949 Provisions: Virtually unchanged from April draft. Support: All parties but CDU/CSU. Electoral law, June 1949 Provisions: 60% single-member district and 40% multi-member district seats with compensatory allocation; 5% state electoral threshold or one single-member district. Support: State governments, tacitly CDU in Parliamentary Council. Opposed by backers of first draft, but their opposition overridden by Allied Powers. Source: Adapted from Lange (1975).

Parliamentary Council, the SPD quickly moved from its earlier proposal of a mixed electoral system to a more exclusively proportional one. It dropped the electoral threshold and sought to personalize the electoral system by advocating smaller multi-member districts rather than outright single-member districts. These changes clearly were meant to appeal to the small swing parties. The SPD never considered cooperating with the CDU/CSU in adopting a majoritarian electoral system that would have eliminated the smaller parties. It rejected this option for two reasons. First, party-internal simulations indicated that a majoritarian system would give the CDU/CSU a disproportionate vote share. Second, the threatened elimination of smaller bourgeois parties by a majoritarian electoral system would have provided additional votes for the CDU/CSU as well as deprived the SPD of future possible coalition partners to drive the CDU/CSU from office (Bawn, 1993: 987). The SPD was willing to incorporate a mixed, less proportional system in the Parliamentary Council’s first and second draft to win the liberal FDP’s support, which had been advocating a strictly mixed electoral system without a compensatory proportional representation element. The SPD’s continued opposition to an electoral threshold, on the other hand, was the price for assuring the support of the Zentrum and the communist KPD. The CDU/CSU badly misplayed its hand. Its opening position pushed for an electoral system which was far more majoritarian than the alternatives it entertained during its party-internal deliberations. This drove the smaller parties into the hands of the SPD even though they were ideologically much closer to the CDU/CSU. Moreover, the FDP’s preference for a non-compensatory mixed system

230

Marcus Kreuzer

actually was closer to the CDU/CSU’s ideal point than to that of the SPD thus making it a potential ally. So, when the Parliamentary Council got around to voting on the first draft, the CDU/CSU was in the minority, receiving support only from the DP. It is difficult to understand why the CDU/CSU’s opening position was more majoritarian than what it entertained during its party-internal deliberations. Its proposal clearly did not recognize the pivotal role played by the smaller parties in the Parliamentary Council and the SPD’s preference for PR. There are indications that the CDU/CSU was appealing to actors outside the Parliamentary Council. It was trying to exploit the vague instructions the Allied occupational authorities had issued over whether the Parliamentary Council or individual state governments had the authority to draft an electoral law. The CDU/CSU hoped to overcome its minority position in the Council by shifting the decision-making arena elsewhere. This strategy worked briefly when the Allied occupational authorities rejected the First Draft on grounds that electoral laws should be a state responsibility. Surprisingly enough, they raised no objections to any of the proposed electoral procedures. The state governments, however, passed up the Allied Powers’ invitation to draft their own electoral law and instead instructed the Parliamentary Council to reconsider the CDU’s objections and pass the next draft by a two-thirds majority. The Parliamentary Council ignored these instructions and re-passed the same draft bill with the same bare majority. After this rebuff, the state governments reconvened and specifically demanded a 5 per cent electoral threshold at the state level and shifting the 50/50 single- to multi-member district seat ratio to 60/40. These modifications accommodated some of the CDU/CSU’s demands for a less proportional electoral system. However, state governments did not accept CDU/CSU demands for eliminating the requirement to democratize candidate recruitment and its request for permitting apparentements. Chancellor (Prime Minister) Konrad Adenauer (CDU/ CSU) believed that such apparentements would facilitate electoral cooperation with other small bourgeois parties and contribute in the long term to their organizational integration into the CDU/CSU. His opposition to local candidate recruitment in turn was motivated by a desire to strengthen party discipline. But still insisting on its exclusive authority for drafting the electoral law, the Parliamentary Council rejected the modifications demanded by the state governments. At this point, the Allied occupational powers intervened directly and declared the May 1949 draft bill together with the state government proposals as the new electoral law. All parties abided by this verdict. In retrospect, the shift in the parties’ institutional preferences after the formation of the Parliamentary Council demonstrates that the greater clarity about the rules of the game and the distribution of power led actors to shift their institutional preferences from non-partisan to partisan considerations. This shift, however, was not equally pronounced across all electoral procedures. It was particularly evident in preferences over seat-related electoral procedures. The SPD was willing to sacrifice the electoral threshold and the personalizing effect of single-member districts to win the support of the smaller parties. The CDU/CSU, in turn, first backed off its promise to add significant compensatory elements to a plurality system and then,

Germany

231

after the prospects for such an electoral system were nil, pushed at the last-minute for a vote-pooling mechanism. Ultimately, the last-minute intervention of state governments and Allied occupational powers reinserted the 5 per cent electoral threshold and made sure that the power of small parties would not produce a party system almost as fragmented as that of Weimar. Electoral procedures affecting the transparency of voters’ choices and candidate recruitment were less affected by partisan considerations than those related to seats. The collective desire for democratizing candidate recruitment was threatened only by Adenauer’s last-minute demand to leave candidate selection unregulated. However, he was easily overridden by the other parties and state governments since nobody had a desire to return to the centralized and autocratic Weimar parties. The attempt to personalize representation by introducing single-member districts faced a more significant challenge after the SPD backed down from its original commitment. The goal of such personalization was to make representation more transparent, limit polarization and decentralize candidate selection. Ultimately however, the FDP, DP, CDU/CSU, the state governments and the Allied occupational powers’ commitment to personalizing representation won out over the SPD’s plan to adopt a pure proportional representation system. 1953 and 1957 modifications Soon after the 1949 election, parties were back drafting a new electoral law since the earlier one was only to be in effect for the first election. The new constitution, the first federal election and various additional state elections had provided actors by now with additional information about their strategic environment. This further decline in uncertainty made the bargaining over the 1953 and 1957 electoral laws more partisan than ever before. The opening move for a new electoral law came from the CDU/CSU-led coalition government in January 1953. Its proposal had a dual thrust. First, it continued the CDU/CSU’s quest to make the electoral system more majoritarian by proposing a 5 per cent national threshold and by removing any compensatory links between single-member districts and multi-member district seats. Fifty per cent of seats were to be directly elected from single-member districts and 50 per cent from state lists. The second reform thrust was far more overtly partisan. The CDU/CSU proposed once again national apparentements and a complicated, second-preference vote. This preference vote was different from the eventual second, separate list vote in that it was to be cast for another single-member district’s candidate and could be transferred in case the candidate of the voter’s primary vote had more votes than he/she required. Overall, the CDU/CSU’s 1953 proposal sought to block the entry of new, startup parties and compensate smaller established ones with vote-pooling provisions. Like earlier CDU/CSU proposals, this one was striking for its political naivety. The vote pooling provisions were so overtly partisan and bound to blur political accountability, just as they had in Weimar, that they elicited widespread criticism in the media and consequent public rejection. Moreover, the smaller parties clearly rejected Adenauer’s attempt to closely tie their electoral survival to the good graces of the CDU/CSU. Not surprisingly, the proposal

232

Marcus Kreuzer

barely passed the first reading and ultimately was defeated on the floor, 175 against 202 with 7 abstentions. Moreover, the Bundesrat (upper federal chamber) overwhelmingly opposed the government’s plan. Not only had the CDU/CSU lost its majority with the merger of three Southern states but even the CDU/CSU state governments rejected the newly proposed electoral law and the consequent decrease in Bundesrat representation. They did so because of its centralizing organizational consequences and because smaller state parties, afraid that their survival would be jeopardized by the demise of their federal branches, threatened to terminate their coalition arrangements with the CDU/CSU. Quick to exploit the CDU/CSU blundering, the SPD, together with the CDU/CSU coalition partner FDP, proposed that the compensatory link between single-member districts and multi-member districts be retained, that voters get a separate, second list vote and that the electoral threshold be lowered to 3 per cent. On the defensive by now, the CDU/CSU dropped its original proposal and went on damage control by insisting that the threshold remain at 5 per cent and be applied to parties’ national vote. In this, it was backed by the FDP. Eventually, the SPD’s demand for a second list vote and the CDU’s demand for a 5 per cent national threshold were added to the original 1949 electoral law. The balance between non-partisan and partisan considerations tipped during the bargaining over the 1953 electoral law further in the direction of the latter. Lange is correct in concluding ‘how strongly the intention was with all parties to design an electoral system to advance their own interest and harm those of others. This [institutional] micro engineering extended into the smaller details’ (Lange: 1975: 560). The electoral features personalizing representation and affecting candidate recruitment remained unchanged in the 1953 law. Their partisan effect, being limited to the party-internal distribution of power as opposed to the legislative distribution of seats, seems to have been small enough not to seriously threaten the lessons learned from Weimar. Moreover, attempts by federal parties to recentralize candidate selection would have produced an almost certain Bundesrat veto since the state party leaders made it clear that they would oppose any diminution of their party-internal powers (Kitzinger, 1960: 25; Lange, 1975: 783–5). As in 1949, the seat-related electoral procedures were subject to the most evident partisan tug of war. It is impossible to clearly differentiate to what extent the CDU/ CSU’s push for more majoritarian features was motivated by collective, governability concerns or narrow partisan considerations. The CDU/CSU was genuinely concerned about the prospect of increasing party fragmentation. Yet, by the same token, its own internal simulations showed that majoritarian features would award it with a disproportionate seat share. Furthermore, Adenauer anticipated that the higher electoral threshold and the move to a mixed electoral system would undermine the long-term electoral viability of its smaller, bourgeois competitor parties. The CDU/CSU’s push for various vote-pooling mechanisms were more unambiguously partisan. The apparentement and secondary vote served to offer smaller parties sufficient vote-pooling possibilities so that they would not object to the CDU/CSU’s more majoritarian demands. The two provisions also had the more sinister goal of tying the other small bourgeois parties firmly into pre-electoral

Germany

233

coalitions with the CDU/CSU thereby promoting their long-term organizational integration. Defending the status quo, the SPD and FDP did not have to reveal their partisanship as overtly as the CDU/CSU. The SPD proposed lowering the electoral threshold in an attempt to pry away the CDU/CSU’s smaller coalition partners. Its proposal for a separate list vote, in turn, was meant to reduce the vote boost the CDU received by generally having more attractive candidates and by being the incumbent party. The second list vote meant that the CDU would no longer benefit from SPD or FDP leaning voters who preferred and hence voted for CDU/CSU candidates. The second vote reduced this risk by permitting such voters to express their party and candidate preferences separately (Bawn, 1993: 978–82). Finally, the FDP backed the 5 per cent national threshold because it was interested in eliminating the other small parties in order to enhance its coalition leverage as the only remaining swing party. Buoyed by the parliamentary majority it won in 1953, the CDU/CSU took another stab at making the electoral system more majoritarian. It reintroduced its 1953 proposal minus the vote-pooling provisions. This brought on an immediate coalition crisis since the FDP would have lost over half of its seats under a genuine mixed electoral system. The CDU/CSU thus had to withdraw its proposal because it required the FDP vote to secure the two-thirds majority required to pass various crucial constitutional amendments. The two parties temporarily patched up their differences and ultimately agreed to a minor logroll. The CDU/CSU agreed to the FDP’s demand for increasing the electoral threshold by requiring now three instead of one single-member district seats. The FDP, in turn, supported the CDU/CSU’s request for a highly watered down apparentement provision. The CDU/CSU dropped its original demand for vote-pooling among different parties and accepted a new, second, federal-tier district in which a party’s state list votes would be pooled to determine that party’s overall seat share. The seats themselves would continue to be awarded to candidates from state lists and proportionally according to the state lists’ vote shares. Ultimately, the federal-tier district had the effect of recycling left-over state votes (Kitzinger, 1960: 23–5). The final version did not meet Adenauer’s initial intent of awarding the federally recycled party votes through a federal party list which he could use to control the recruitment of personal confidants. Lange probably is correct in arguing that Adenauer was not sincere in his quixotic-looking demand for a mixed electoral system, but strategically used it to extract concessions from the smaller coalition parties on important economic and foreign policy issues (Lange, 1975: 798). The SPD’s only foray into the 1957 electoral law debate consisted of demanding a return to the single vote used in 1949. It was concerned that the two separate votes could facilitate strategic withdrawals among bourgeois parties. It entertained the somewhat hypothetical example that in any given state the FDP would withdraw all single-member district candidates, thus favouring CDU/CSU candidates, in return for the CDU/CSU not fielding a party list. This fear proved exaggerated since in 1953 the CDU/CSU had only 42 withdrawal agreements with other bourgeois parties (Bawn, 1993: 982n).

234

Marcus Kreuzer

There is little change in the level of partisanship that can be observed between the 1953 and 1957 electoral law debates. The SPD championed the annulment of the second vote for which it pushed four years earlier because it recognized potential risks. The CDU/CSU, in turn, took institutional engineering into new partisan directions by returning to the politically unrealizable mixed system as a strategic move to extract political concessions from the FDP in other unrelated policy arenas. In general, the institutional manoeuvring surrounding the 1957 electoral law was so specific to the political circumstance of the time that the Weimar legacy hardly figured anymore. But also the changes that ultimately were implemented were so minor that they did not significantly alter whatever impact Weimar’s shadow had cast on the earlier electoral laws.

Conclusion This chronological survey shows that political partisanship was indeed a key driving force and the German politicians were no different from their counterparts elsewhere. But partisanship and the Allies’ influence were tempered by two additional interacting factors, namely historical learning and political uncertainty. Certain lessons learned from the Weimar Republic were a key factor shaping parties’ preferences early on, committing them to grafting proportional and majoritarian elements together in ways no other electoral system had done heretofore. We also saw that the preferences that these historical legacies shaped about electoral Table 11.2

Germany: Bundestag electoral results, 1949–2002

Year

Communists DKP/PDS v/s

Green Grünne v/s

Socialist SPD v/s

Liberal FDP v/s

Christians CDU/CSU, Z, B v/s

Others right v/s

1949 1953 1957 1961 1965 1969 1972 1976 1980 1983 1987 1990 1994 1998 2002

6/4 2/0 – – – – – – – – – 2/2 4/4 5/5 4/1

– – – – – – – – – 5/5 8/8 5/1 7/7 7/7 9/9

29/33 29/31 32/34 36/38 39/41 43/45 46/47 43/44 43/44 38/39 37/38 35/36 37/38 41/45 39/42

10/13 9/10 7/8 13/13 9/9 5/6 8/8 7/7 10/10 7/6 9/9 7/12 6/6 6/6 7/8

38/41 48/51 50/55 46/49 48/50 47/49 45/45 49/49 45/46 48/50 44/45 45/49 43/45 36/37 38/40

17/9 12/8 11/3 5/0 3/0 5/0 1/0 1/0 2/0 2/0 2/0 6/0 3/0 5/0 3/0

Key: DKP: German Communist Party; PDS: Party of Democratic Socialism; Grünne: Greens and Bündnis 90; SPD: Social-democratic Party of Germany; FDP: Free Democratic Party; CDU: Christian Democratic Union; CSU: Christian Social Union; Z: German Centre Party; B: Bavarian Party; v/s: percentage of votes/ percentage of seats.

Germany

235

procedures related to candidate recruitment and transparency of voter choice and remained largely unaffected by partisan considerations throughout the postwar era. When it came to seat-related electoral procedures, partisanship was held at bay for a while due to parties’ lack of information which prevented them computing the pay-offs of different electoral procedures. This changed, however, as soon as the decision-making moved from non-strategic party-internal deliberations to rulebound, inter-party bargaining in the Constitutional Council, and as soon as the first state and local elections revealed crucial information about the parties’ electoral strengths. By the time the veil of uncertainty was being lifted, the historically induced path towards mixing majoritarian and proportional elements, democratizing candidate recruitment and increasing the transparency of voters’ choice had already advanced too far to leave partisan considerations unconstrained. In sum, the political actors clearly used whatever room they had for partisan engineering, but Weimar’s shadow, aided by immediate postwar uncertainty, imposed important limits on this bargaining space.

References Bawn, Kathleen (1993) ‘The Logic of Institutional Preferences: German Electoral Law as a Social Choice Outcome’, American Journal of Political Science, 37: 965–89. Kitzinger, U. W. (1960) German Electoral Politics. Oxford: Clarendon Press. Kreuzer, Marcus (2000) ‘Electoral Mechanisms and Electioneering Incentives’, Party Politics, 6/4: 487–504. Kreuzer, Marcus (2001) Institutions and Innovation: Voters, Parties, and Interest Groups in the Consolidation of Democracy: France and Germany, 1870–1939. Ann Arbor: University of Michigan Press. Lange, Erhard (1975) Wahlrecht und Innenpolitik. Meisenheim: Anton Hain. Pollock, James (1932) Money and Politics Abroad. New York: Knopf. Roberts, Geoffrey (1988) ‘German Federal Republic: Two-Lane Route to Bonn’, in M. Gallagher and M. Marsh (eds), Candidate Selection in Comparative Perspective: the Secret Garden of Politics. London: Sage. Scarrow, Susan (2001) ‘Germany: the Mixed Member System as a Political Compromise’, in Matthew S. Shugart and Martin P. Wattenberg (eds), Mixed Electoral Systems. The Best of Both Worlds? New York and Oxford: Oxford University Press. Schanbacher, Eberhard (1982) Parliamentarische Wahlen und Wahlsystem in der Weimarer Republic. Düsseldorf: Droste Verlag. Shugart, Matthew Soberg and Martin P. Wattenberg (2001) ‘Mixed-Member Electoral Systems: a Definition and Typology’, in Matthew S. Shugart and Martin P. Wattenberg (eds), Mixed Electoral Systems. The Best of Both Worlds? New York and Oxford: Oxford University Press.

Further reading • Bawn, Kathleen (1993) ‘The Logic of Institutional Preferences: German Electoral Law as a Social Choice Outcome’, American Journal of Political Science, 37: 965–89.

• Kitzinger, U. W. (1960) German Electoral Politics. Oxford: Clarendon Press. • Kreuzer, Marcus (2001) Institutions and Innovation: Voters, Parties, and Interest Groups in the Consolidation of Democracy: France and Germany, 1870–1939. Ann Arbor: University of Michigan Press. • Lange, Erhard (1975) Wahlrecht und Innenpolitik. Meisenheim: Anton Hain.

236

Marcus Kreuzer

• Merkl, Peter (1963) The Origin of the West German Republic. New York and Oxford: Oxford University Press.

• Peters, Hans (1956) ‘Kandidatenaufstellung für freie demokratische Wahlen’, in T. Maunz • • • •

and H. Nawiasky (eds), Vom Bonner Grundgesetz zur gesamtdeutschen Verfassung. Munich: Isar Verlag. Pollock, James (1932) Money and Politics Abroad. New York: Knopf. Roberts, Geoffrey (1988) ‘German Federal Republic: Two-Lane Route to Bonn’, in Michael Gallagher and Michael Marsh (eds), Candidate Selection in Comparative Perspective: the Secret Garden of Politics. London: Sage. Scarrow, Susan (2001) ‘Germany: the Mixed Member System as a Political Compromise’, in Matthew S. Shugart and Martin P. Wattenberg (eds), Mixed Electoral Systems. the Best of Both Worlds? New York and Oxford: Oxford University Press. Schanbacher, Eberhard (1982) Parliamentarische Wahlen und Wahlsystem in der Weimarer Republic. Düsseldorf: Droste Verlag.

12 Italy: Lofty Ambitions and Unintended Consequences Diego Gambetta and Steven Warner1

The system of proportional representation by which all Italian governments have been elected since the ratification of the present constitution dates from 1946. It predated the constitution itself and was never included in that document, so the reform of the electoral system requires no constitutional amendment. There were some minor alterations in 1948, and a short-lived attempt at imposing a high majority premium in 1953 (the so-called ‘tricky law’ or legge truffa). This led to a bitter parliamentary fight between the christian-democrats (DC) and the communist party (PCI), which eventually resulted in the full reinstatement of the proportional system in 1956. However, this was not the first proportional system to be adopted in the country. In 1918, when Italy conducted its elections under a majority system, Vittorio Emanuele Orlando proposed both the extension of the male suffrage and the introduction of a proportional electoral system. These proposals were welcomed by all the major parties of the time: by the liberals because they feared for their future under the majority system in the face of the rise of mass parties of the right and left; and by both the socialists and the catholic Popular party, once it was formed in 1919, who each saw in such a system the means by which to become a credible political force. With such a broad range of support, it is not surprising that ‘an act implementing [proportional representation] was passed in August 1919, and the suffrage was extended at the same time’. The result of the reform was not, however, a success. The liberals remained the largest party, but they were internally divided and their majority had been cut by the advances made by the socialists and the Popular party, necessitating a coalition. ‘However, the other two parties refused to form a coalition either with the Liberals or with each other, and in these circumstances parliamentary government broke down’ (Carstairs, 1980: 155). Proportional representation did not have an auspicious debut in Italy, and was soon abandoned during the years of fascist dominance. Following the Second World War, Italy was faced with the challenge of reconstituting itself as a democratic polity. In 1946, the king appointed the National 237

238

Diego Gambetta and Steven Warner

Consult, a body charged with the responsibility of drawing up the rules under which the Constituent Assembly would be elected and the constitution framed. The question of the electoral system was discussed between 11 and 22 February. The result of the delegates’ deliberations was Law no. 74, a decree issued by the king on 10 March 1946. The law detailed a proportional system based on party lists and a separate ‘preference vote’ for individual candidates. In the elections to the Constituent Assembly, this system delivered 37 per cent of the seats to the christian-democrats (DC). This result clearly justified the support that the DC group had expressed for the system in the National Consult. Sufficient other groups at the Constituent Assembly found the result congenial to approve its use for normal parliamentary elections. There was some slight alteration made to the number of seats in some constituencies, and the number of preference votes permitted was increased to four in the larger ones. Apart from these changes, the 1948 parliamentary elections were conducted under the same system that had governed the selection of the Constituent Assembly. The only other change of note in the electoral system before the 1990s was the reduction of the divisor used in calculating the quota.2

Proportionality rejected The move away from proportionality in Italian elections began with the referendum of 1991. While this referendum left the proportional representation system essentially intact, it did reduce the number of preference votes to one per voter in every constituency, and in doing so raised the question of the electoral system more generally. Electoral reform was finally introduced in August 1993 after another referendum, the resulting system being presented, especially in the Italian press, as being a unique hybrid that could have dangerous and unpredictable consequences. This inference obtains to some extent but relied on a misguided premise, for all electoral systems can generate unpredictable consequences. While the effects of that electoral reform were indeed uncertain, this is due to the very nature of electoral systems whose effects are generated through a complex interaction with the wider political context. Moreover, the new Italian system was not even such an extravagant invention. In fact, the new system is a variant of the large class of mixed systems, which all allocate seats by some combination of both plurality and proportionality rules. They differ mostly in terms of the proportion of seats allocated by either rule, by the proportional corrections imposed on the plurality allocation and by the structure of voting options open to the voters. Another well-known variant is the additional-member system used in Germany. Other variants, some of which are very similar to the new Italian system, have been recently adopted by other countries. In this text mixed systems like the new Italian system are called ‘parallel’ or ‘mixed-member majoritarian’ systems. The Italian version varies slightly between elections to the Chamber of Deputies and those to the Senate. In the former case, 75 per cent of the seats (472 deputies) are to be elected by the plurality system in single-member districts. The logic of such a contest means that each party will field only one candidate: only one

Italy

239

person can win, and so fielding more than this number risks dividing the loyalties of the party’s supporters and reducing the likelihood that the party will carry a plurality of the votes. The remaining 25 per cent (158 deputies) are to be elected on the basis of a regional vote, which is cast separately but concurrently with that for the single-member district election. The region is a multi-member district, and the election is carried out under the system of party-list proportional representation to which Italians had became accustomed over the previous fifty years. The main difference between this stage of the electoral process and that which was in place prior to the reform is that the threshold that must be passed before a party becomes eligible for consideration in the allocation of seats has been raised to 4 per cent of the total votes cast in the region. In the case of the Senate, 75 per cent of the seats are again awarded in singlemember districts on the basis of a plurality election (237 senators). The remaining 25 per cent (78) senators are elected in multi-member districts on the basis of a separate vote. This differs from the equivalent stage in elections for the Chamber in that there is no threshold imposed: every candidate who stands is eligible for consideration in the allocation of seats. The reason for this omission is that the number of seats available in each multi-member region is small enough to prevent the election of anyone gaining less than around 10 per cent of the votes cast, thus ensuring that each successful candidate has a certain level of support without resorting to artificial restrictions such as a threshold.

Paradoxes of the system’s origins The electoral system was approved in the Chamber with 287 ‘ayes’, 78 ‘noes’ and 153 abstentions. In the Senate, the result was 128 in favour, 29 against and 59 abstentions. While this indicates that the reform was approved with 55.4 per cent of votes in the Chamber being favourable and a 59.3 per cent level of support in the Senate, 112 deputies and 99 senators were entirely absent for the vote. If we take this staggeringly high number of absentees into account, we find that only 45.6 per cent of deputies voted for the reform, and only 40.7 per cent of their colleagues in the Senate voted likewise. Most of the support, paradoxically, came from those largely discredited parties which held the reins of power throughout the Republic’s history. The reform that emerged from the Chamber and the Senate in early August neatly exhibited the received idea that Italian politics are paradoxical. The reform was often held to involve the creation of a new republic, the sweeping away of what had gone before, and there is no doubt that popular pressure to this end existed. However, there was, in the end, no contest between democracy and party-cracy, between the people and the ‘regime’. Rather, the electoral reform was the product of the political elite, who presided over a national debate characterized by superficiality: there was neither a systematic examination of all available alternatives, nor of the possible consequences of those that were considered. There was little opportunity for the public to play an informed role in the generation of the reform, nor to convey democratic legitimacy upon one or other of the alternatives

240

Diego Gambetta and Steven Warner

by signalling support. The reform was definitively the preserve of parliament, a parliament, moreover, which had been elected in 1992 prior to the emergence of most of the details of scandals concerning corruption known as ‘Kickbacksville’ (Tangentopoli) and the composition of which was no longer reflective of public support for the parties.3 The first paradox of the reform is thus that it was under the control of the establishment, of those against whom it was directed. A further paradox concerns the distribution of support for the selected reform within that establishment. When the new system was first approved by the Chamber of Deputies in June, those in favour were the christian-democratic DC, the Northern League, the socialist PSI and the social-democratic PSDI, in other words, with the exception of the League, those parties which had held power during the past fifty years. Opposition came from the post-communist PDS, the republican PRI, the liberal PLI and the anti-mafia Network (Rete), with the fascist MSI, the Greens, Panella List and Communist Refoundation, together with some dissenting socialists, taking the indecisive step of abstaining. Thus the reform lacked the support of, or received the opposition of, all major groups within Italian politics that might be deemed ‘progressive’, or at least ‘alternative’ – a truly paradoxical state of affairs for a reform of this nature. While several criticisms continue to be voiced on the effects of the new system – there were even a couple of aborted attempts at changing it again in the 1990s – the criticism of its somewhat tainted origins is not among them. The final paradox was that those previously dominant parties, such as the DC and the PSI as well as the PSDI, which voted for the new system were eventually swept away by the new political dynamics that the reform contributed to trigger.

The sources of the enthusiasm for reform What kind of reform attracts such an unusual distribution of support? Before we consider how this particular reform was decided upon, it is worth examining the question of why the pressure for reform emerged, for it is undeniable that the Italian public was truly enthusiastic for some kind of electoral reform. First, many Italians wanted a political system closer to that of Anglo-Saxon democracies, and, through that, to remove those features of the Italian system that they perceived as highly undesirable. They wanted fewer parties, government alternation, more governmental stability and effectiveness, closer links between parliamentarians and their constituency voters, as well as reduced opportunities for corruption in the electoral process. Elsewhere (Warner and Gambetta, 1994; Gambetta and Warner, 1996) we have examined in detail how those aims came to be associated with the pressure for electoral reform, and why it was wrongly thought that the reform could bring about those changes. Contrary to a simplistic interpretation of Duverger’s Law, these are only feebly associated with the electoral system per se. We will return to this below. There is, however, a further element that caused these aims to be so widely embraced. Until 1992, the governing coalition routinely captured around 60 per cent of the votes cast. These governments were broad churches of interests, were

Italy

241

formed by a process of inter-party bargaining and were consequently weak and unstable. However, their shares of the vote held reasonably constant, suggesting that dissatisfaction was not endemic; for whatever reason, voters were willing to reward the DC’s coalitions with the means to form governments for decades on end. The year 1992 saw a change to this behaviour. Although the same parties were returned to office, they received a total of only 48.8 per cent of the vote. In a highly proportional system, this was a significant reduction.4 The coalition of the DC, the PSI, the PSDI and the PLI gained 52.5 per cent of the seats, enabling them to form a government. While the characteristics of weakness and a breadth of interests continued in government, the result was a clear indication that voters had begun to become dissatisfied with the status quo. The climate of ‘Kickbacksville’ no doubt increased this feeling of discontent, but the 1992 election result demonstrated that the pressure for change predates the scandals. At this stage, of course, this desire for change was not attached to the reform of the electoral system; rather it was expressed through the mechanism of the proportional electoral system, and concerned the result sought from that system. It is worth noting that the electoral ‘earthquake’ of 1992 would not have been registered as more than a slight tremor under a plurality system. The question to be addressed is that of why voters’ choices changed in this manner. One of the key facts about the 1992 election is that it was the first conducted since the fall of the Berlin Wall and the end of the Cold War. The collision between the capitalist and communist ideologies had been a central feature of Italian politics since the close of the Second World War; the international order created a domestic cleavage of considerable power. The Marxist and Catholic subcultures, each alien to the traditions of liberal democracy but representative of East European communism and Western values respectively, were significant both in the sense that they determined voting behaviour for large numbers of Italians, and in the further sense that they exerted a direct influence upon the formation of governments. This is to say that the effect of the ideological competition on the composition of government was not simply mediated through the effects they had upon the votes cast for each party. The PCI was successfully characterized as an anti-democratic party, and could be presented as a significant threat because the general opinion following the experience of fascism was that democracy was weak and fragile in Italy. This stigmatization of the PCI survived despite successive moves by that party towards autonomy from the Moscow line. Perceived thus, the party could be precluded from inclusion in any government coalition. This fact is crucial to the explanation of the lack of alternation throughout the Republic’s history: DC dominance was produced by a party system that excluded the second-largest party from participation in government, while DC hegemony was actually prevented by the presence of a highly proportional electoral system that calculated representation in such a way as to ensure that few votes were ever ‘wasted’.5 For those who bemoan the absence of alternation in Italy, the presence of a system of proportional representation should be a cause for relief, not a focus for blame. Without this system, and given the exclusion of the PCI, the DC would probably have had

242

Diego Gambetta and Steven Warner

a far greater hold on the reins of power than was actually the case. The result of such a situation may well have been a deeply divided and tension-ridden nation. The communists had been well aware of this and had consistently supported proportional representation. A majority system or one with a strong majority premium would have undermined the very foundations of consociativism, confining the left, at least at the parliamentary level, to a marginal role. The battle over the so-called (by the communists) ‘tricky law’ of 1953 – one of the few battles in which antagonism won over consociativism – is evidence of that awareness. Thus by 1992 the political context had changed and voters felt able to express their choices over a range of very different opportunities. The electoral results clearly indicated that the DC could no longer take for granted voters’ support.

The illusion of a compromise While we can understand the different kinds of pressure for reform that built up and peaked in 1992, the question of why that reform rather than another remains to be addressed. Why, more precisely, did enough politicians find the new electoral system compatible with their different interests? The source of this assumed compatibility may lie in a misappreciation of the electoral system. Rather than being seen as a specific variant of the mixed system generally, in the manner in which we have been discussing it, the reform proposal that was eventually adopted tended to be seen as a compromise arrangement between the pure plurality system and the old proportional representation. Robert Newland of the British Electoral Reform Society clearly warned that the Italian reform ‘may be said to exhibit all the defects of the systems combined, together with the new defects arising from the mix’ (quoted in Walston, 1993: 7; emphasis added). Conceived thus, the new system may also be interpreted as likely to exhibit two entirely separate properties, with one aspect seen as likely to act in the same manner as the British system, the archetype of all majority systems, and the other believed likely to act as did the Italian proportional system prior to reform. Our suggestion that the reform was seen as a two-part compromise can explain how same-party politicians came to support the law, despite the fact that in some areas a party such as the DC would perform better under a proportional system and in others would clearly benefit from a plurality system.6 Under this explanation, northern christian-democrats would view the proportional stage as effectively distinct from the plurality stage and conclude that the system would act to produce the same type of results as the old Italian system. Likewise, southern christiandemocrats would examine only the plurality aspect of the system and conclude that the system would act to produce British-style election results in which they were clear contenders for achieving a plurality. Given the need to produce a reform, we may further postulate that northern DC deputies would have preferred to see a greater emphasis on proportionality in the resulting system – a 50–50 allocation is the most plausible given the range of options considered during the Italian debate – while the Northern League and the southern wing of the DC would have preferred a pure plurality system, believing that such a system would

Italy

243

guarantee them regional hegemony. Acting upon the belief that the effects of each part of the mixed system are effectively separate, these two factions could then engineer reform with the appearance of being a compromise. The final 75–25 per cent balance in favour of plurality elections would thus be taken to reflect the relatively greater strength of the Southern DC–Northern League axis. In fact, the new system does not reproduce the effects of either of its ‘component’ systems – to look for an imitation of either the United Kingdom, the Italy of the past or some simple mixture of the two is unproductive. The type of results yielded by the Italian mixed system is likely to change as both parties and voters adjust their behaviour to the constraints of the new system. However, it is clear that at no point will the electorate approach the two ballots as distinct votes, for although the two votes are separate, they will be cast with a view to the production of a single desired outcome. Parties and voters face a learning process with regard to how to use the system. The possibility of tactical voting will be new to voters, and the effective use of this tool relies in part upon being able to extrapolate from previous results in order to gauge the likely level of support for the various parties within the constituency and within the region and to estimate the range of viable winners. The type of result provided by the voters could be expected, and in fact was subject to considerable evolution. Voters most definitely will not behave in the one ballot in the manner of an electorate accustomed to a plurality system and in the other as an electorate familiar with proportional representation. The reform of the electoral system not only takes place within a unique context, but also generates changes to that context that must be taken into account in the calculation of its likely effects. We suggest that politicians in Italy have fallen foul of both the ‘appearance of universality’ and the ‘appearance of compromise’. As a result, they have come to believe that the new system will imitate two distinct systems. This is a dual error: it is only pseudo-imitation that they take to be a true likeness, and the new system, despite the presence of two separate ballots, is most definitely a single system, different from both the plurality system and proportional representation. The particular electoral law that Italians now have is the result of persuasion, and of mistake on the part of politicians, acting on the discontent of their electorate. But how successful an exercise in manipulation and control has this been? The fact that political calculations were made on the basis of pseudo-imitation of Anglo-Saxon systems may explain why, as we predicted in 1994, some politicians found that they have not been as successful in forwarding their aims as they supposed. It is not the case that by importing some other nation’s electoral system the shape of one’s own political system would come to mirror that of the model country. Rather, it is the case that no matter what reform is introduced, the effects of a new system cannot be predicted merely by reference to the politics of another country. There is furthermore the question of instability – the reform was introduced into a context in which voters’ loyalties were undoubtedly shaken. The uncertainty regarding voters’ behaviour when confronted with a new type of ballot was compounded by the post-Kickbacksville environment. Under such conditions, the potential existed both for apparent change to be greater than real change and for

244

Diego Gambetta and Steven Warner

the emergence of genuine change of an unexpected kind. Which scenario would be actualized could not, contrary to widespread expectations, be a product of the selected electoral system, but instead only a product of voters’ political choices conducted under it. While the public was excluded from the process of reform at a relatively early stage, their decisions proved central in shaping the effect of the reform. As we have emphasized, it is unlikely that any particular electoral system could prove by itself to be a panacea capable of creating a stable political regime, and the question of stability as much as that of change was answered by the electorate.

The effects of the reform The first three elections under the new electoral system were held in 1994, 1996 and 2001 and, in the process, the Italian political system changed dramatically. Let us now briefly examine the main changes and whether they live up to the ambitions originally associated with the desire for reform. Two main changes affected the party political landscape. One was the collapse, under the pressure of the corruption scandals, of the two main dominant parties, the DC and the PSI, which disappeared together with their satellite parties, the PSDI, the PRI and the PLI. The other was of course the emergence and success of Let’s Go Italy (Forza Italia). Berlusconi’s party became the alternative of choice for the disenfranchised centre and right-of-centre voters, even though new small parties were formed from splinters of the old parties and quite a few of the politicians from the old regime still haunt today’s political scene. In 1994 and in 2001, the Pole of Freedom (subsequently renamed House of Freedom), the right-of-centre coalition put together by Silvio Berlusconi around Let’s Go Italy, won the contest, while the centre-left coalition, the Olive Tree, won in 1996. It thus appeared that Italians have at last obtained one of their ambitions for a well functioning political system: alternation between the parties in government. While some have hailed the occurrence of alternation as a success of the new system (e.g. Pasquino, 2002), others, like Giovanni Sartori, claim that alternation would have occurred under other systems too (2002: 107), something that yet others deny (Bartolini, Chiaromonte and D’Alimonte, 2002: 368). A counterfactual of this kind can hardly be empirically proven to be correct. What can be proved, however, is that voters did generate alternation, paradoxically without appreciably changing their collective preferences which have remained very stable over the left–right divide. In 2001 the parties in Berlusconi’s coalition even lost some votes relative to the sum of their votes in 1996 and nonetheless gained a majority of seats in both houses of parliament. A crucial reason for that preference insensitivity of the electoral results is due to electoral alliances. In both 1994 and again in 2001 Berlusconi put together the coalition that most effectively exploited the electoral system’s properties and voters’ dispositions, while he failed to do so in 1996. In 1994, Let’s Go Italy formed a two-pronged alliance with the League in the north and National Alliance in the

Italy

245

south, and won; by contrast the left and the centre (Pact for Italy, Segni List and others with about 16 per cent of votes) ran separately and lost. In 1996 the situation was reversed: Berlusconi failed to keep the League and the Tri-color Flame, a small right-wing party, in his coalition. The defection of the former caused the loss of single-member constituencies for Let’s Go Italy in the north, and the latter contributed to the defeat of National Alliance in a number of districts in the south. By contrast, in 1996, on the opposite side of the political spectrum, a new coalition, the Olive Tree, emerged. It included both the centre and the left-wing parties, and further it was crucially able to rely on a so-called ‘desistance pact’ with the Communist Refoundation, whereby the two groups refrain from presenting candidates in the same constituency in order not to split their votes. The Olive Tree won the election. In 2001 the situation was once more reversed. The centre-left coalition this time was unable to reconstitute the whole alliance and related pacts, while Berlusconi, mindful of the lesson of 1996, succeeded again to form an alliance that comprised both the League and the Tri-color Flame, and won a comfortable majority in both houses. Between 1996 and 2001, as we said, the distribution of preferences among the electorate remained largely stable. The most notable shifts occurred within each coalition with a reallocation of votes in favour of the moderate parties – the Daisy (Margherita) at the expense of Democrats of the Left (ex-PDS and ex-PCI) and Let’s Go Italy at the expense of National Alliance. Those outcomes were not of course the effect of the electoral system alone nor of the options that parties and coalitions supplied to voters, but of what voters chose to do with them. They depended in particular on the fact that not enough voters in crucially contested constituencies chose to use tactical voting but stuck to their first-choice party. For instance, in 2001, in several constituencies the centre-left coalition lost the single-member seat to the candidate of the House of Freedom because of a split vote: left-wing voters preferred to give a useless vote to the candidates of Communist Refoundation rather than to support the centre-left candidates. They thus de facto promoted the victory of more right-wing candidates. Similarly, in 1996, right-wing voters in the north chose to vote for the League, which was competing outside Berlusconi’s coalition in that election, even though by so doing they helped the centre-left coalition to win more seats in singlemember constituencies. Slowly, however, voters have been progressively concentrating their choices on the two main coalitions. If one considers the votes in the single-member constituencies for the Chamber of Deputies, the two main alliances totalled 80.1 per cent in 1994 and 89.7 per cent in 2001. In terms of the total number of seats in both houses, the coalitions obtained around 90 per cent in both 1994 and 1996, and shot up to 97.6 per cent in 2001 (see Bartolini, Chiaromonte and D’Alimonte, 2002, for a detailed analysis of this process of concentration). Two centre-left small parties (one of which was led by the famous ex-judge turned politician Antonio Di Pietro), after exhausting and turbulent negotiations, stubbornly chose to stay out of the Olive Tree and were sorely left seat-less, as they failed to make the 4 per cent threshold. (There is even evidence that some voters misunderstood

246

Diego Gambetta and Steven Warner

the new system for they gave more votes to these two parties in the single-member constituencies than in the proportional phase of the election in 2001 (Sartori, 2002: 113).) Still, in so far as in marginally contested constituencies a sufficiently high number of voters’ decisions are subject to ideological, religious, ethnic or other lexicographic preferences, the pressure on parties to form as broad a coalition as possible before the election remains extremely high. ‘United we win, divided we perish.’ The formation of two main opposing coalitions has somewhat simplified the political landscape and the choice for voters, bringing about essentially a form of bipolarism which was one of the aims that motivated the pressure for reform. However, this state of affairs has had undesirable consequences in other respects. First, there has been an increase in the number of parties rather than the expected and desired decrease. While old parties were dying out new parties mushroomed. If we use the party rather than the coalition as the unit, the seats in the Chamber in 2001 were split between 22 parties (20 in 1994 and 24 in 1996) and 23 in the Senate. If one takes into account not only the number of parties but also their relative strength, however, there is a decrease in terms of fragmentation, which still remains very high by international standards (Bartolini, Chiaromonte and D’Alimonte, 2002: 374–5). The current interpretation blames the persistent party fragmentation on the proportional part of the electoral system. Yet it is not only this that keeps the small parties alive. In fact, there are even parties that hold single-member constituency seats which would not have made the 4 per cent quota. The main reason is that in so far as a party, however motley, has (or makes the bigger parties believe that it has) the control of a crucial number of constituency votes, it has bargaining power over the larger parties. It can thus obtain singlemember constituencies for its candidates in return for not contesting other constituencies and disturbing the vote for bigger parties – just as mobsters ‘persuade’ night club or restaurant owners to pay protection money in return for not upsetting their patrons. The higher the number of hotly contested constituencies the lower the number of votes required for a party to have that power over at least one of the two main contenders in that constituency. Even tiny political creatures stand a chance in this situation, and creating a new party has become a way to make a living. This point, which was made by other commentators such as Sartori (2002: 110) and Bartolini, Chiaromonte and D’Alimonte (2002: 375), cannot, however, be taken to the extreme of denying any role to the proportional part of the electoral system. This acts indirectly on the fragmentation by allowing smaller parties to count their real electoral force independent of the coalition they join, and helps the bigger parties to refute exaggerated demands and weed out bargaining claims orchestrated by entirely phony parties. Before the elections the turbulence of negotiations to arrive at ‘desistance pacts’ – ‘I shall not have any candidate to stand in A if yours withdraws from B’ – had reached monumental proportions. There have even been some attempts at codifying the rules of this ‘trade’ (Di Virgilio, 2002). In this situation smaller

Italy

247

parties have been able to bargain even for the ‘safer’ constituencies and ended up with far more seats than their vote would warrant and, equally, their elected representatives can end up with better positions in government than their vote would justify. The most successful at this game has been the ChristianDemocratic Centre (CCD): in 2001 with only 3.2 per cent of the votes they obtained 40 seats in the Chamber and 29 in the Senate (Sartori, 2002: 115) and its leaders have been appointed to key institutional and ministerial positions. This two-tier political system that combines parties and electoral coalitions looks firmly set to stay. Next, one declared objective of the reform was to achieve a situation in which the electorate is encouraged to vote more with reference to the qualities of the individual candidate than to the party to which he or she belongs. This expectation was unwarranted all along, for as we have explained elsewhere (Warner and Gambetta, 1994; Gambetta and Warner, 1996) this ambition could not be realized with any electoral system that involves single-member constituencies. This expectation, which is still there judging by the surprise that this failure has generated, was based on a misunderstanding of how a majority system, notably the British system, really works in this respect. Under any such system – whether based on plurality rule, majority runoff or a mixed system – the parties have considerable leverage over their candidates which derives from the discretion they have in selecting which one of their members is to contest the seat. The clearest guide to a candidate’s political profile is thus clearly the party label or banner under which they appear on the ballot paper. The new Italian system, by relying so heavily on coalition pacts, has made matters even worse than other majority-based systems would have. Candidates are centrally chosen and parachuted into constituencies in which they have never resided and of which they have no prior knowledge or contacts with their potential local voters. They may belong to an unknown little party alien to local party preferences and traditions. Not even such a thing as ‘our Member of Parliament (MP)’, as it exists in Britain for instance, has emerged, as parliamentarians may or may not be selected at the next election regardless of whether they took any notice of the interests of their constituencies. 7 In Britain the local party members can at least partially contribute to select and deselect candidates. Thus the Italian electorate have cast their votes far more on the basis of party labels, alliance banners and their opinions of the parties’ national leaders than on the basis of any appreciation of the local candidates themselves. The candidates belonging to the Socialist Party (PSI), for example, fared badly in the proportional representation stage of the 1994 contest, in which they fought as members of the PSI. But other PSI candidates, contesting the plurality stage of the election, did so under the banner of the left-wing alliance and were elected. It seems clear that the key to this apparent discrepancy in the fortunes of socialist candidates was that PSI was a vote-losing label, while the Progressive, as the left-wing alliance was then called at the time, was a vote-attracting banner. An analogous example can be found in the opposite coalition: in 1994 both the Northern League and Let’s Go Italy selected candidates from the widely despised ‘old regime’. And despite

248

Diego Gambetta and Steven Warner

the animosity felt towards those regarded as complicit in the corrupt politics of the previous fifty years, many of these figures achieved success in their constituencies. This is a result which can plausibly be accounted for by presuming that, for their electors, their current party labels counted for more than did their personal past and character. Third, the pressure to form a coalition, even if it is now felt before rather than after the elections as it did under proportional representation, dilutes the focus of political manifestos and thus keeps transparency low. Elections, it was hoped, would be transparent in these related senses: the connection between the casting of votes and the formation of a government would be clear and obvious, and the electorate would, when casting their votes, be able to select the manifesto to which they felt most attracted. They would thus know which goods they were ‘buying’ with their votes if their preferred party was successful. With the introduction of the new system, the incentive to make alliances prior to the election was increased; it might seem, then, that the electoral reform has made a contribution to this ambition of removing the opaqueness that post-election coalitions displayed under proportional representation. However, both left- and right-wing alliances have been characterized by disputes over priorities and policies. Furthermore, they both exist largely for the purpose of contesting the single-member constituencies, while parties still have an incentive to differentiate themselves to contest the proportional part of the elections. As a result, whichever alliance proves successful, the legislative programme of the resultant government is the result of a bargain worked out between its member parties once the balance of power between them is determined by the election results. In other words, as before, the electorate can predict who the members of a government would be if their preferred party is successful, but cannot predict what exactly their legislative programme is going to be. For instance, the brilliance of the Berlusconi coalition in both 1994 and 2001 was keeping together with Let’s Go Italy two parties which in many ways are at odds with each other. He induced voters in the south to vote for an alliance that includes the League, whose interests and programme are as opposed to southerners as they can get. And in parallel he induced voters in the north to vote for an alliance that includes National Alliance, whose proclivities towards public spending and subsidies in favour of the south can be seen as opposed to the interests ‘of the north’. Did the southern voters who chose to support Berlusconi’s coalition in 2001 really choose to have the devolution legislation that his government brought about in 2002 to appease the demands of the League? How many votes for the tri-partite coalition can be seen as votes for the government programme that eventually emerged? Not only do coalitions thus formed reduce transparency, but substantive legitimacy as well. Matters, it is said, are improving and the system is beginning to work (e.g. Pasquino, 2002, and, more cautiously, Bartolini, Chiaromonte and D’Alimonte, 2002) and yield governmental stability and efficacy, both of which were among the ambitions for reform. The first Berlusconi government fell in less than a year because the League decided to exit from the government, which thus lost its

Italy

249

majority in parliament (an event known as ‘il ribaltone’ or alternation). And the left-wing coalition elected in 1996 went through tormented internal changes – the prime ministerial baton shifted hands three times from Prodi to D’Alema (twice) to Amato – and the composition of the government coalition was also readjusted three times (Pasquino, 2002: 14). For the first time in the 2001 election the coalition that won coincided with that which went on to form a government (while in 1996 the coalition included Communist Refoundation, which stayed out of government). Moreover, this time the majority of the coalition was large enough for it not to be hostage to the smaller members such as the League, something that should lodge Berlusconi firmly in the saddle, and increased both the stability and the efficacy of the government. Furthermore, the House of Freedom was an altogether more cohesive coalition, with ‘only’ four parties forming it and Let’s Go Italy in a clearly dominant position, than the Olive Tree, with eight parties, none of which was significantly in control. The internal stability between the parties in government, which were heterogeneous in terms of their general aims, remained rather brittle. Conflicts over, for example, justice reform or state media organisation were the order of the day. Yet there is no doubt that this government enjoyed an overall stability that has eluded all its predecessors since 1946. The enthusiasts of strong government come what may see this as a reason to rejoice, and some of the social engineers of electoral reform feel some satisfaction on ‘technical’ grounds regardless of their personal preferences. The many Italians who think of Berlusconi as an unmitigated curse on the country’s political, social and cultural life probably think otherwise. They may now regret the abandonment of proportionality, one result of which can be to enhance the power of the executive – an outcome welcome only if one happens to like the incumbents, or, more generally, to trust politicians to be of high quality (Gambetta, 1988). Governments structured in such as way as to be able to do little can also do less wrong. A more general lesson one can draw from the Italian reform is that several of the effects of introducing new electoral systems are seldom predictable. In fact, one can predict that there will be unpredictable and unintended consequences, as there were in this case. This is partly because electoral systems come to life only via the strategies of adaptation of the players – by the way in which these strategies play out both between parties and between parties and voters. It is further due to the fact that electoral systems are hardly ever changed ‘in cold blood’, as it were. The inertia of electoral systems is a well-known phenomenon: those who win have the means but no reason to change the system and the opposition may have the reasons but lack the means. More generally, it is extremely difficult, given the varied distribution of party political interests, to reach an even simple majority in favour of any one particular new system. Thus electoral reforms, when they occur, come almost invariably accompanied by some crisis – wars, corruption scandals, deepening of cleavages – and because of this, their effects on the political system are even more likely to have unintended consequences.

250

Table 12.1 Year

1992 Year

1994 1996 2001

Italy: Chamber of Deputies electoral results, 1992–2002

Communist

Left

Green

Socialists

RC v/s

PDS v/s

FV v/s

PSI v/s

5/6

16/17

3/2

14/15

Republican

Northern

Christian

Liberal

Fascist

Other

PSDI v/s

PRI v/s

LN v/s

DC v/s

PLI v/s

MSI v/s

v/s

3/3

4/4

9/9

30/33

3/3

5/5

8/3

Communists RC/PdCI v/s

Left DS v/s

Sunflower FV/PSI v/s

Daisy PPI/RI v/s

Northern LN v/s

Whiteflower CCD/CDU v/s

Pole FI/AN v/s

Other

6/6 9/6 7/2

20/27 21/25 17/22

5/0 4/6 2/3

16/5 11/13 15/12

8/18 10/9 4/5

21/4 6/5 3/6

14/35 37/34 41/45

10/5 2/2 11/5

v/s

Key: RC: Communist Refoundation; PdCI: Party of the Italian Communists; PDS/DS: Party of Left Democrats/Progressives/Olive Tree; FV: Federation of Greens; PSI: Italian Socialist Party; PSDI: Italian Social-Democratic Party; PRI: Italian Republican Party; LN: Northern League; DC: Christian Democracy; PLI: Italian Liberal Party; MSI: Italian Social Movement; PPI: Italian People’s Party; RI: Italian Renewal; CCD: Christian-Democratic Centre; CDU: United Christian-Democrats; FI: Let’s Go Italy; AN: National Alliance; v/s: percentage of votes/percentage of seats.

Italy

251

Notes 1. The opening sections of this chapter reproduce with minor changes and updating parts of Warner and Gambetta (1994) except on the effects of the electoral reform. 2. In 1948, the divisor was the number of seats plus three, generating an extraordinarily low quota to the benefit of the larger parties. In 1956 this divisor was reduced to the number of seats plus two. This quota remained lower than the Droop quota, which involves a divisor of seats plus one. 3. The 1993 mayoral elections provided ample proof of this fact: whereas the PSI, for example, took 13.6 per cent of the votes in 1992, they were reduced to a humiliating 2.2 per cent in Milan in 1993, previously one of their most supportive areas. 4. In the United Kingdom, of course, a result for the government of 48 per cent is almost unheard of: a single party can govern with much less than this, thus generating no incentive to form a coalition that has majority support. 5. Had the First Republic been born with a plurality system, the DC’s majority would almost certainly have been secure for many years. By 1992, of course, the DC vote was below 30 per cent, and the party would have been forced into coalitions throughout the 1980s and early 1990s, no matter what the electoral system, if they were to remain in power. 6. One need not subscribe to the idea that electoral systems are active to appreciate that this is clearly the case. Where a party lacks a concentration of support, it stands a far greater chance of having its candidates elected if the system is based around multi-member constituencies with seats allocated proportionately than it does if seats are allocated in single-member constituencies under the plurality rule. The effects of ‘Kickbacksville’ on voters’ loyalties created disparate situations for the DC, exaggerating the ‘southernization’ of their vote which had been proceeding for some time. 7. The turnover of candidates ‘rooted’ in constituencies has been very high: between 1994 and 2001 only 21.1 per cent in constituencies for the Chamber and 18.1 per cent in constituencies for the Senate have remained the same (Bartolini, Chiaromonte and D’Alimonte, 2002: 376).

References Bartolini, Stefano, Alessandro Chiaromonte and Roberto D’Alimonte (2002) ‘Maggioritario finalmente? Il bilancio di tre prove’, in Roberto D’Alimonte and Stefano Bartolini (eds), Maggioritario finalmente? Bologna: Il Mulino, pp. 363–80. Di Virgilio, A. (2002) ‘L’offerta elettorale: la politica delle alleanze si istituzionalizza’, in Roberto D’Alimonte and Stefano Bartolini (eds), Maggioritario finalmente? Bologna: II Mulino, pp. 79–130. Gambetta, Diego (1988) ‘Apologia della proporzionale’, Biblioteca della Libertà, 23: 95–106. Gambetta, Diego and Steven Warner (1996) ‘The Rhetoric of Reform Revealed (or If You Bite the Ballot It May Bite Back)’, Journal of Modern Italian Studies, 1, 3: 357–76. Carstairs, Andrew McLaren (1980) A Short History of Electoral Systems in Western Europe. London: Allen & Unwin. Pasquino, Gianfranco (2002) ‘Un’elezione non come le altre’, in Gianfranco Pasquino (ed.), Dall’Ulivo al governo Berlusconi. Bologna: Il Mulino, pp. 11–22. Sartori, Giovanni (2002) ‘Il sistema elettorale resta cattivo’, in Gianfranco Pasquino (ed.), Dall’Ulivo al governo Berlusconi. Bologna: Il Mulino, pp. 107–15. Walston, J. (1993) ‘Mixing the Systems: Winners and Losers’, Wanted in Rome, 7 July. Warner, Steven and Diego Gambetta (1994) La retorica della riforma. Fine del sistema proporzionale in Italia. Torino: Einaudi.

252

Diego Gambetta and Steven Warner

Further reading • D’Alimonte, Roberto and Stefano Bartolini (eds) (1998) Special issue on Italian political change, European Journal of Political Research, 34.

• D’Alimonte, Roberto and Stefano Bartolini (eds) (2002) Maggioritario finalmente? Bologna: Il Mulino.

• Donovan, Mark (1995) ‘The Politics of Electoral Reform in Italy’, International Political Science Review, 16, 1: 47–64.

• Gambetta, Diego and Steven Warner (1996) ‘The Rhetoric of Reform Revealed (or If You Bite the Ballot It May Bite Back)’, Journal of Modern Italian Studies, 1, 3: 357–76.

• Katz, Richard S. (2001) ‘Reforming the Italian Electoral Law’, in Matthew S. Shugart and Martin P. Wattenberg (eds), Mixed-Member Electoral Systems. The Best of Both Worlds? New York and Oxford: Oxford University Press, pp. 96–122. • Pasquino, Gianfranco (ed.) (2002) Dall’Ulivo al governo Berlusconi. Bologna: Il Mulino. • Warner, Steven and Diego Gambetta (1994) La retorica della riforma. Fine del sistema proporzionale in Italia. Torino: Einaudi.

13 Spain: from Civil War to Proportional Representation Josep M. Colomer

The choice of a majoritarian electoral system by the leaders of the Spanish Second Republic established in 1931 can explain the high and increasing levels of political polarization which eventually led to revolution, coup d’état and ultimately civil war. The provisional government of the Republic chose an electoral system with the aim of putting an end to the electoral domination that local bosses (‘caciques’) had been able to exert during the previous monarchy on the basis of a high number of single-member districts by plurality rule. The new government kept the majoritarian principle, but generalized now the system of multi-member districts with limited vote that had been used in a few districts during the monarchy. The new republican rulers relied upon the broad popular support that had obtained at local elections in April 1931 and expected to be the beneficiaries of the corresponding over-representation. But their institutional choice was based on a miscalculation regarding the likely electoral support of the republican parties in open competition with new candidacies supported by traditionally monarchical electors. The over-representation effect of the electoral system backfired: it produced oversized majorities for both the left republicans and the rightist monarchist parties at successive elections and in general very rapidly increasing political polarization which led to confrontation and violent conflict, culminating in civil war in 1936–39. In contrast, the following, much more successful, democratic experience in Spain, which started in 1977, has been based on elections by proportional representation.

The initial rules The most durable electoral system in the history of Spain had been established in 1878, at the beginning of the restoration of the monarchy, by the conservative government led by Antonio Cánovas. At that time very limited suffrage rights encompassing about only five per cent of the total population were conceded. The electoral system, based on plurality rule, remained in place until the election of 1920, in spite of the introduction of universal male suffrage by a liberal government in 1890 and a significant reduction in electoral fraud 253

254

Josep M. Colomer

thanks to reforms introduced by another conservative government from 1907 onwards. The electoral system was based on plurality rule in a mix of single-member districts with single ballot and a few small multi-member districts with limited ballot, as inspired by the model introduced in Great Britain between 1867 and 1885, and also imitated in other Latin European and Latin American countries in the following years. Specifically, there were, first, 307 single-member districts, mostly in rural areas, in which the winner was usually either a conservative or a liberal candidate, and second, 26 multi-member districts with between three and eight seats, up to 90 seats in total, mostly in urban areas, in which, besides the conservative and liberal candidates, some republican, Catalan regionalist and other candidates won representation in parallel with the expansion of suffrage rights and the reduction of fraud. In the multi-member districts, each voter had a limited vote, that is he could vote for several candidates but fewer in number than the seats in the district (the rates between the number of votes per voter and the seats in the district were 2/3, 3/4, 3/5, 5/7 and 6/8). The seats in each district were allocated to the most voted candidacy which received the number of seats for which the voter could vote, and to the second candidacy in votes which received the remaining seat or seats. Thus about 75 per cent of the seats of Parliament were chosen in singlemember districts, while up to 93 per cent were given to the candidacies with a plurality of votes in each district. When the opposition parties won first place in multi-member districts, they usually proclaimed a ‘moral victory’, in spite of receiving only a minority of seats in Parliament. (The foregoing is based on the author’s calculations with data in Martínez, 1969, 1973; see also Fernández, 1992; Rueda, 1998.) Proportional representation was considered for introduction in Spain in the years immediately after the First World War, that is about the time when most parliaments in Europe began to be elected under this new type of rules. As discussed, the new European regimes tended to establish rules of the game that made the formation of a single winner difficult, thus making them acceptable by both the traditional conservative and liberal groups and the newly emerging parties, especially socialists and Catholics. Likewise, the Spanish liberals, with the support of the at that time minority republicans, socialists and Catalan regionalists, introduced in parliament a number of proposals for adopting proportional representation in 1919, 1921 and 1923. On the first two occasions, the conservative government of the moment began to draft the corresponding bills that would have established a more pluralistic and inclusive framework and would perhaps have saved a parliamentary monarchy along the lines of other European countries. But the corresponding legislative processes were delayed by futile reasons and, finally, the initiative of electoral reform failed when the king and the military reacted to social turmoil and anarchist terrorism with a coup d’état and the establishment of a dictatorship in 1923 (Carreras, 1973, 1983; Ull Pont, 1976; early support for proportional representation by the left republican Catalan nationalists can be found in Rovira Virgili, 1910, and by the conservative Catalan regionalists in Vidal Guardiola et al., 1919).

Spain

255

The republicans’ choice The Spanish Second Republic was established in April 1931, in reaction to the monarch’s support for military dictatorship during the period 1923–29, in which no competitive elections had been called. The provisional government of the Republic, formed by liberals, republicans and socialists, decreed new rules to elect a Constituent Assembly. They didn’t resume their own attempt to establish proportional representation but, in contrast, basically kept the criteria of the existing, rather exclusive rules of the game. In the new electoral system, all the single-member districts, which had been bases of traditional support for the monarchist parties, were abolished. Small and medium multi-member districts with limited ballot, similar to those that had traditionally allowed the republican and opposition parties to obtain representation and ‘moral victories’, were established across the entire country in the expectation that they would produce the same effects in favour of the newly ruling parties. Although it probably did not seem so, seen from the perspective presented in the general introduction to this book the new system was even more restrictive than the old one since it was more likely to create artificial parliamentary majorities with a minority voting support than the system based on small single-member districts. The republican rulers expected to eliminate what they considered to be the falsification of the real Spain produced by local clientelism and electoral fraud during the old regime and to keep broad social support indefinitely. However, they overestimated the coercive component of traditional local clientelism and underestimated the monarchist parties’ social support. They thus slightly reformed the existing electoral rules in the expectation that they would again give overrepresentation to the most voted candidacies, but now on the republican side. As we will see, the parties that were actually advantaged turned out to be both the left republicans and the right monarchists, on different elections. In two contests the electoral system even produced non-monotonic results by which the bloc of parties united in a coalition candidacy obtained more seats than the other bloc of parties with a higher number of total votes but running in separate candidacies The discussion in the Constituent Assembly of the electoral bill introduced by the republican government was highly revealing. The Prime Minister, Manuel Azaña, put forward a poor defence, saying that, in contrast to the majoritarian system under discussion, a system of proportional representation such as that proposed by some rightist opposition parties would force minority groups to form electoral coalitions with other parties. He showed thus a significant lack of knowledge of the issue, up to the point of confusing electoral coalitions, which are indeed encouraged by majoritarian systems such as the one under discussion, with post-electoral coalitions typically formed in parliaments elected by proportional representation to support a cabinet. Azaña dismissed in a few words the danger that the electoral system would fabricate an artificial parliamentary majority on account of his ‘indestructible conviction that republicans and socialists were the majority of the country’. When some Catholic and regionalist deputies protested

256

Josep M. Colomer

that minorities would be crushed, Azaña, however, admitted that he ‘didn’t know yet whom we are going to crush, or even who will be the crushers and crushed.’ (The parliamentary debates are collected in Mori, 1933, vol. XIII: 345ff.) The basic elements of the electoral system were confirmed by the Law of Electoral Reform, July 1933. The most obvious alternative was a project of proportional representation elaborated in 1932 by a Consulting Legal Commission appointed by the government with the mission of elaborating a Constitutional project and other major regulations. This project of proportional representation was transformed into a bill by further governments supported by the Catholics and the Catalan regionalists on three occasions during 1934 and 1935, but it didn’t gain approval in parliament (Carreras, 1983).

The new rules The electoral system improvised in 1931 replaced the 333 small districts of the old monarchical period with 63 multi-member districts with magnitudes up to 18 seats. In each district the voter could vote for several individual candidates, but fewer in number than the seats in the district, that is there was limited vote with preferential ballot, similar to the one that had been used during the monarchy in a few districts. The rates between the number of votes per voter and the seats in the district were the same as those of the old regime for the smaller districts, enlarged now to 7/9, 8/10, 8/11, 9/12, 9/13, 11/14, 12/15, 12/16 and 14/18. The seats in each district were distributed between the two most voted lists, which came to be called ‘the majority’ and ‘the minority’. The list with a qualified plurality of at least 20 per cent of votes (40 per cent after 1933) received the number of seats for which the voter could vote (that is between 67 and 80 per cent), while the second most voted list received the remaining seats. By these rules, the second candidacy in votes, called ‘the minority’, received only between 33 and 20 per cent of votes in each district even if it had obtained higher proportions of votes. In total, ‘the majorities’ received 75 per cent of the seats in parliament (355 out of 473). If there was not a sufficient number of individual candidates with the required threshold, there was a second round by plurality rule for the seats remaining to be allocated. Let’s look at a very simple example: a district with eight seats in which each voter could vote for six candidates (which was very close to the average district). If three candidacies obtained 41, 30 and 29 per cent of votes, respectively, they were allocated 6, 2 and 0 seats respectively. Thus a candidacy with slightly more than 40 per cent of votes could receive 75 per cent of seats (6/8), while another candidacy with almost 30 per cent of votes would not obtain representation. The threshold of exclusion, or maximum share of votes that could award no seat, can be calculated with the following formula: Excl = k/k + m, where Excl is threshold of exclusion, k the number of votes per voter and m the district magnitude in seats. Accordingly, in the about average district with eight seats and six votes per voter mentioned above, a candidacy was only guaranteed some representation if it obtained more votes than the threshold of exclusion, that is about 43 per cent

Spain

257

of votes (Excl = 6/(6 + 8) = 42.85). This implies that if a large party or coalition ran in two separate lists, with about 57 per cent of popular votes (100 − 42.85) appropriately distributed, it could receive all the seats, by both ‘the majority’ and ‘the minority’ (57.15 × 6/8 = 42.86). The strategy of competing for candidates of both ‘the majority’ and ‘the minority’ was developed in some districts by the republican-socialist coalition in 1931 and by the Catholic CEDA and its allies in 1933. To sum up, a candidacy with about 57 per cent of votes and optimum distribution in the territory could receive 100 per cent of the seats; a candidacy with more than 40 per cent, if it were the most voted one because its rivals were sufficiently divided in several candidacies, could obtain 75 per cent of seats (all ‘the majorities’); but if its rivals were sufficiently united, with almost 43 per cent of votes it could receive no seats. The strategic consequence of these mechanical effects of the electoral system was the formation of large multi-party coalitions able to guarantee representation and to achieve the over-representation given to the most voted candidacy in each district. These strategies generated a high degree of political polarization. The analysis of the electoral results presented here (see Tables 13.1 and 13.2) is based on a method of counting votes that gives the same value to every individual vote, in spite of the fact that voters in different districts could vote for different numbers of candidates, as explained above. In order to weight the votes to achieve such equality, that is in order to discover the electoral preferences revealed by the voters, for each district the total number of votes obtained by each candidacy has been divided by the number of candidates each voter could vote for. 1

Political consequences Six multi-party blocs can be distinguished as shown in Table 13.1 from left to right: (1) Communists; (2) Socialists; (3) Republicans (including radical-socialists); (4) Liberals (including radicals, liberals and centrists, as well as regionalists and most independent candidates); (5) Catholics (including agrarians); and (6) Monarchists (including fascists). It would not have been impossible for a centrist coalition of republicans, radical-socialists, radicals and liberals, which were the bulk of the initial support for the new Republic, to have obtained a plurality of votes and the corresponding oversized parliamentary majority at successive elections. But the political centre, although it was quite popular among voters, was very weakly organized and extremely divided into a large number of small parties and personalistic groupings. The two best organized groups were the socialists (PSOE) on the left and, from the second election on, the Catholics (CEDA) on the right. The numerous and small intermediate parties very rapidly tended to join one of these two larger parties, forming thus two broad coalitions according to the incentives created by the electoral system discussed above. The two competing coalitions also included the most extreme small parties, the communists and the monarchists, respectively, which increased polarization. The electoral rules induced thus a high degree of political confrontation between the two blocs, much higher than could

258

Table 13.1 Year

1931 1933 1936

Spain: electoral results, 1931–36

Communists Votes v/s

Socialists Votes v/s

43,242 1/0 190,244 2/0 [294,608

[880,274

+

1,685,318 20/13 1,625,404 47/60

+

+

1,516,835 84/89 1,199,976 14/8 2,540,146]

Catholics Votes

Liberals Votes v/s

Republicans Votes v/s +

Others Votes

117,633]

52,002

777,254]

32,259

783,974]

–

v/s

1,245,492] [2,548,939 30/37 2,091,710 22/15

Monarchists Votes

[230,718 +

2,059,290 [2,237,066

+ 15/11 + 34/42 + 31/25

Key: Communists: PCE, BOC, POUM, PCP, trade unionists; Socialists: PSOE, USC; Republicans: AR, RS, IR, UR, ERC, PCR, AC, PNCR, UR, ANV, IV, ORGA, EV, federalists, independents; Liberals: PRR, DL, PC, LC, PNV, ASR, PLD, PC, PP, autonomists, independents; Catholics: CEDA, DRV, PNA, AN, social-christians; Monarchists: RE, CT, FE, AO. Note that coalitions [+] are over-represented. Votes: numbers of votes; v/s: percentage of votes/percentage of seats. Sources: Author’s calculations with data from Venegas (1942), Tusell (1971, 1982), Linz and De Miguel (1977), Linz (1978), Irwin (1991).

Table 13.2

Spain: electoral results, 1977–2000

Year

Communist PCE/IU v/s

Socialist PSOE v/s

Basque PNV v/s

Catalan CiU v/s

Centre UCD/CDS v/s

Conservative AP/PP v/s

Others v/s

1977 1979 1982 1986 1989 1993 1996 2000

9/3 11/7 4/1 5/2 9/5 10/5 11/6 6/2

29/34 30/35 48/58 44/53 40/50 39/45 36/40 35/36

2/1 2/2 2/3 2/2 1/1 1/1 1/2 2/3

3/2 3/2 4/3 5/5 5/5 5/4 5/4 4/4

35/48 35/49 10/2 9/6 8/4 – – –

8/5 6/4 26/30 26/30 26/31 35/40 38/45 45/52

14/7 13/1 8/3 9/2 11/4 10/5 9/3 8/3

Key: PCE: Communist Party of Spain; IU: United Left; PSOE: Spanish Workers’ Socialist Party; PNV: Basque Nationalist Party; CiU: Convergence and Union; UCD: Union of Democratic Centre; CDS: Democratic and Social Centre; AP: People’s Alliance; PP: People’s Party. Votes: numbers of votes; v/s: percentage of votes/percentage of seats.

259

260

Josep M. Colomer

have been expected with a proportional representation system giving each party the corresponding share of seats. More specifically, in the first election in June 1931, the parties in favour of the new Republic (including left-republicans, radical-socialists, radicals and liberals), running together with the socialists, obtained a very large, rather centrist majority over the monarchists. The republicans were given over-representation and received a large majority of seats on their own, even without the socialists. In contrast, the disunited groups that had supported the monarchy (including Catholics, agrarians, and Basque and Catalan regionalists) were under-represented. In the second election in November 1933, the republican centre split into two factions. While left-republicans and radical-socialists remained on the centre-left, the radicals and the liberals were now ready to collaborate with recovered monarchists. They reached irregularly distributed, sometimes informal, pre-electoral arrangements in most districts with a reinforced right which was able to organize a new large party, the CEDA, with the Catholics and the agrarians. Both the candidacies of the centre-right and those of the right obtained fewer votes but were given more seats than the more separated centre-left and left parties if the latter are counted together, as shown in Table 13.1. Specifically, while the separated centre-left and left parties obtained (22 + 14) 36 per cent of votes, they were given only (13 + 8) 21 per cent of seats. This was because the parties in the right half of the spectrum attained narrow majorities or pluralities in a high number of districts and were rewarded with oversized representation. In contrast, the centre-left republicans and the left socialists, now running separately, became frequent district ‘minorities’ and were badly affected by the electoral system favouring the larger candidacies. After the 1933 election, a rather moderate cabinet led by the centre-right radical party was formed, thanks to the parliamentary support of the right. The largest party in votes and seats, the Catholic CEDA, did not enter the cabinet directly. Despite this, and in the belief that the CEDA would eventually gain more influence, in October 1934 the extreme left of the socialists and the anarchists organized an armed rebellion to which they were able to attract some centre-left republicans. The rebellion was fought by the army and resulted in more than 1,500 deaths (on both sides) and about 15,000 prisoners. The election of February 1936 was a kind of revenge. The centre-left republicans together with the socialists and the communists formed the Popular Front. This time it was the centre-right that was weaker than the large coalition of the centre-left and left parties. In addition, fear of wasting their votes moved a significant number of citizens to vote strategically either for the Popular Front or for the rightist National Front formed by the Catholics, the agrarians and the monarchists. The increasing polarization promoted by the incentives supplied by the electoral system did not correspond to the degree of polarization that could be found among voters’ sincere preferences. Using the preferential vote available on the ballot, the citizens gave more votes to the moderate candidates within each coalition or bloc. Specifically, the centre-left republicans obtained more support than did the socialists or the communists in the same lists, while the Catholics received more votes than the monarchists. In fact, some of the moderate candidates running in opposite

Spain

261

coalitions had been members of the same electoral candidacy (or even the same party) just a few years earlier (Jackson, 1965: 518–25; Linz, 1978). This time it was the Popular Front that obtained fewer votes but was given more seats than the now less unified centre-right and right parties, if the latter are counted together, as shown in Table 13.1. Specifically, while the separated centreright and right parties obtained (23 + 31) 54 per cent of votes, they were given only (15 + 24) 40 per cent of the seats. If the centre-right and right parties had run in perfect unity and had obtained the same amount of votes that they received separately, they would have collected a majority in votes but still fewer seats than the Popular Front, according to district-by-district calculations (Tusell, 1976: 111). This indicates that the perverse results were only increased by party leaders’ asymmetrical skills in forming electoral coalitions, but they were basically a mechanical result of majoritarian rules and unfair apportionment. As a consequence of both the electoral rules and the coalition strategies, relatively minor changes in voters’ choices at successive elections produced upheavals in institutional representation. Following the 1936 election, a rather moderate cabinet led by centre-left republicans was formed without the direct participation of the socialists or other leftist parties. Yet a military uprising, initially supported by the Catholics and agrarians, was carried out less than five months after the election. The division of Spain into two camps supporting the Republic and the rebel military roughly conformed to the territorial division of votes in the previous election. The two blocs, however, were soon dominated by the the communists and the fascists, respectively. During an almost three-year civil war, several hundred thousand people were killed and others went into exile. The outburst of a polarized conflict at the induction of electoral rules rather than from social unrest greatly surprised many Spaniards. The lack of understanding why civil war had occurred fed a great fear that a similar conflict might reappear in the future. The subsequent military dictatorship lasted for almost forty years.

Re-democratizing with proportional representation At the following, much more successful democratization attempt, which started in 1977, a proportional representation electoral system was established for the lower chamber of parliament as well as for regional and local elections. As has been repeatedly remarked, the memory of the civil war and the subsequent dictatorship was an important factor in inducing cooperative strategies and agreements between incumbent rulers from General Franco’s regime and the democratic opposition. Approving an electoral system able to prevent some of the worse consequences of the one implemented in the 1930s was also a concern of most political actors in the 1970s (Colomer, 1995). The electoral system for the first election after Franco’s death, which would be later confirmed by further democratic governments, was negotiated by a non-elected government appointed by King Juan Carlos and led by Adolfo Suárez, later leader of the centrist party, the UCD. Centrist Suárez faced both hard-line Francoists and the opposition. On the one hand, Suárez negotiated with the bulk of the members

262

Josep M. Colomer

of the authoritarian assembly, the ‘Cortes’, for the approval of a law for political reform that included the call for a multi-party election – the first to be held in more than forty years. The pivotal collection of assembly members, grouped as the ‘Popular Alliance’ (AP) which would become one of the major political parties in the future, proposed a majoritarian electoral system for a two-chamber parliament to be approved by the incumbent, non-democratic ‘Cortes’. On the other hand, the opposition, formed mainly by socialists, communists, Catalan and Basque nationalists and some christian-democrats, which were coordinated on a multiparty platform, announced that they would not participate in the election if a system of proportional representation were not introduced. In November 1976, Prime Minister Suárez in a compromise with the AP produced a formula for a two-chamber parliament in which the upper house, which was going to have very weak powers, would be elected by a majoritarian system, while the lower chamber would be based on proportional representation. For the latter, however, the 50 administrative provinces would become electoral districts with a minimum of deputies each, independently of the population, and other ‘corrective devices’. Actually the electoral rules were decreed by the Suárez government in April 1977 after the dissolution of the former authoritarian assembly. It included a relatively low number of 350 seats in the lower chamber, called the Congress of Deputies, and the apportionment of a minimum of two deputies per province. This gave over-representation to the 38 less populated provinces, all with district magnitudes lower than 9, and together able to elect a majority of seats in spite of gathering barely 40 per cent of voters. With an optimum distribution of votes in these provinces, a party could receive an absolute majority of seats in the Congress of Deputies (176 out of 350 seats) with 33 per cent of national popular votes. The law was custom-made for the centrist UCD, and organized by the incumbent government within the restriction of permitting some multi-party representation in order to obtain the electoral participation of opposition parties. However, at the first election in June 1977 the UCD obtained 34.6 per cent of popular votes, not optimally distributed, and remained short of a few deputies from an absolute majority. This led the UCD leaders to negotiate a political reform more encompassing than planned, including an agreement on a new democratic constitution with the socialists, communists and Catalan nationalists (Colomer, 1990: 93–102; 1995, 2001). Furthermore, the 1977 electoral system has not only given over-representation to the UCD and its replacement on the right side of the political spectrum since the late 1980s, the People’s Party (PP, formed from the previously mentioned AP), but also to the larger party on the left side, the socialist PSOE. Disproportionality between party votes and seats has oscillated between 5 and 15 per cent, achieving the highest average values among proportional representation systems in Western Europe. In four of the first eight elections up to 2000, a single party received a majority of seats in the Congress of Deputies with a minority of popular votes (between 39 and 44 per cent), while in the other four elections no party achieved a seat majority, thus requiring the negotiation of larger parliamentary support for the corresponding cabinets. In 1985, while having an absolute majority of seats

Spain

263

and forming a single-party cabinet, the PSOE introduced in parliament a new electoral bill which basically confirmed, with greater legal status, the provisions of the 1977 decree (Colomer, 2001). As a whole, the electoral system used in Spanish democracy since 1977 has favoured the two larger Spain-wide parties as well as the large regional parties in the corresponding districts while hurting minor Spain-wide parties (all those below 20 per cent of votes). This has induced bipolar competition for leading the formation of the national government. However, in contrast to the electoral system in the 1930s, minor parties running separately have also achieved parliamentary representation. The electoral system has also favoured the flourishing of multiple regional parties and the corresponding pressures in favour of political decentralization and strengthening of regional governments. Thus bipolarization has not developed by way of two distinct coalitions which include the most extreme parties but by centripetal competition for the centrist voters. The principle of proportional representation has not produced broad multi-partism, but its ‘correctives’ have not reduced political competition to only two parties or blocs either. The degree of inclusiveness of the existing system, although much lower than other proportional representation systems in well-established parliamentary democracies, has prevented the reproduction of the fatal confrontations and reactions against the system of the past and has fostered the general acceptance of electoral results.

Note 1. This method had been used by different authors for the elections of 1936 (Linz and De Miguel, 1977) and of 1933 (Irwin, 1991), and has been completed by the author for the 1931 election with the help of Francesca Vassallo and Eric Langenbacher. The corresponding electoral results are summarized in Table 13.1. Other studies have presented either the sum or the average of all preferential votes obtained by the candidates, or the number of votes obtained by the first candidate or the most voted one in the list (the mode), but these procedures distort the voters’ preferences seriously because they give higher weight to voters in large districts with the possibility of voting for higher numbers of candidates.

References Carreras, Francesc de (1973) Legislación electoral en la Segunda República española. Doctoral dissertation, Autonomous University of Barcelona. Carreras, Francesc de (1983) ‘Los intentos de reforma electoral durante la Segunda República’, Revista de Estudios Políticos, 31–2: 165–97. Colomer, Josep M. (1990) El arte de la manipulación política. Votaciones y teoría de juegos en la política española. Barcelona: Anagrama. Colomer, Josep M. (1995) Game Theory and the Transition to Democracy: The Spanish Model. Aldershot: Edward Elgar. Colomer, Josep M. (2001) ‘The 2000 General Election in Spain’, Electoral Studies, 20, 3: 490–95. Fernández Domínguez, Arturo (1992) Leyes electorales españolas de diputados a Cortes en el siglo XIX. Estudio histórico y jurídico-político. Madrid: Civitas. Irwin, William J. (1991) The 1993 ‘Cortes’ Elections. Origin of the ‘Bienio Negro’. New York: Garland.

264

Josep M. Colomer

Jackson, Gabriel (1965) The Spanish Republic and the Civil War. Princeton, NJ: Princeton University Press. Linz, Juan J. (1978) ‘From Great Hopes to Civil War: the Breakdown of Democracy in Spain’, in Juan J. Linz and Alfred Stepan (eds), The Breakdown of Democratic Regimes. Baltimore, MD: Johns Hopkins University Press, vol. 2, pp. 142–215. Linz, Juan J. and Jesús De Miguel (1977) ‘Hacia un análisis regional de las elecciones de 1936 en España’, Revista Española de la Opinión Pública, 48: 27–67. Martínez Cuadrado, Rafael (1969) Elecciones y partidos políticos en España, 1868–1931. Madrid: Taurus. Martínez Cuadrado, Rafael (1973) La burguesía conservadora (1874–1931). Madrid: Alianza. Mori, Arturo (1933) Crónica de las Cortes Constituyentes de la Segunda República Española. Madrid: Aguilar. Rovira Virgili, Antoni (1910) La representación proporcional en el sufragi universal. Barcelona: Centre Nacionalista Republicà. Rueda, José-Carlos (ed.) (1998) Legislación electoral española (1808–1977). Barcelona: Ariel. Tusell, Javier (1971) Las elecciones del Frente Popular. Madrid: Cuadernos para el diálogo. Tusell, Javier (1976) ‘The Popular Front Elections in Spain, 1936’, in Stanley G. Payne (ed.), Politics and Society in Twentieth Century Spain. New York: New Viewpoints. Tusell, Javier (1982) Las constituyentes de 1931: Unas elecciones de transición. Madrid: Centro de Investigaciones Sociológicas. Ull Pont, Eugenio J. (1967) ‘El sufragio universal en España (1890–1936)’, Revista de Estudios Políticos, 208–9: 105–30. Venegas, José (1942) Las elecciones del Frente Popular. Buenos Aires: Patronato HispanoArgentino de Cultura. Vidal Guardiola, Miquel, José M. Bassols, Carlos Jordà, José M. Tallada, Fernando Sanz Buigas and Luis Puig de la Bellacasa (1910) La representación proporcional. Barcelona: Joventut Nacionalista.

Further reading • Colomer, Josep M. (1990) El arte de la manipulación política. Votaciones y teoría de juegos en la política española. Barclona: Anagrama.

• Colomer, Josep M. (1995) Game Theory and the Transition to Democracy: The Spanish Model. Aldershot: Edward Elgar.

• Colomer, Josep M. (2002) ‘Spain and Portugal: Rule of Party Leadership’, in Josep M. Colomer (ed.), Political Institutions in Europe, 2nd edn. London: Routledge.

• Gunther, Richard (1989) ‘Electoral Laws, Party System, and Elites: the Case of Spain’, American Political Science Review, 83: 3.

• Jackson, Gabriel (1965) The Spanish Republic and the Civil War. Princeton, NJ: Princeton University Press.

• Linz, Juan J. (1978) ‘From Great Hopes to Civil War: The Breakdown of Democracy in Spain’, in • • • • •

Juan J. Linz and Alfred Stepan (eds), The Breakdown of Democratic Regimes. Baltimore, MD: Johns Hopkins University Press, vol. 2. Linz, Juan J. and José R. Montero (eds) (1986) Crisis y cambio: Electores y partidos en la España de los años ochenta. Madrid: Centro de Estudios Constitucionales. Montabes, Juan (ed.) (1998) El sistema electoral a debate: Veinte años de rendimientos del sistema electoral español (1977–1997). Madrid: Centro de Investigaciones Sociológicas. Tusell, Javier (1971) Las elecciones del Frente Popular. Madrid: Cuadernos para el diálogo. Tusell, Javier (1976) ‘The Popular Front Elections in Spain, 1936’, in Stanley G. Payne (ed.), Politics and Society in Twentieth Century Spain. New York: New Viewpoints. Tusell, Javier (1982) Las constituyentes de 1931: Unas elecciones de transición. Madrid: Centro de Investigaciones Sociológicas.

14 Sweden: Introducing Proportional Representation from Above Leif Lewin1

In Sweden at the beginning of the twentieth century, universal suffrage was the most important goal for the left. Yet it was introduced by a Conservative government which was an ideological opponent to this demand. How can this strange decision be explained? Rational choice is an excellent instrument if we wish to understand decisions like this. From this theory we learn that individual rationality is no guarantee for having one’s way in politics. Even if a politician well knows his preferences and adequately chooses in accordance with these preferences, the outcome may be different from what he expects. The result may be suboptimal. The Prisoner’s Dilemma is just one example of a situation when miscalculation of the opponent’s strategy leads to unintended consequences.

The arguments In the middle of the nineteenth century the right to vote was very limited in Sweden. Members of the First Chamber were elected indirectly by county councils; local voting rights, in turn, were graded by income and wealth. The right to vote to the Second Chamber was open to men aged 21 or older also of a certain income or wealth. The method applied was majority rule in single-member districts. These rules gave the franchise to just about 20 per cent of all legally competent men. Only a few more men gained the right to vote during the next few decades as a result of rising salaries. Limited suffrage was best for the nation, according to the Prime Minister Louis De Geer: ‘In order to ensure a calm social order, it may be important to restrict the right to vote to persons with living conditions that may be regarded as leaving some time and inclination free for independent participation in political life.’ Wealth was regarded as the best basis for qualification. ‘Ownership of property may generally be assumed to indicate a more secure and independent position than that of an ordinary wage earner, even if the latter has a good income for the moment, and to result in a great interest in the improvement and welfare of society.’2 The property requirement was thus set relatively low, the income requirement relatively high. 265

266

Leif Lewin

Two lines of reasoning can be identified. The right to vote should be granted on the basis of wealth and ability. Members of the Estate of the Clergy denied that wealth was a sign of ability. The ‘golden pale of money’ – not education, experience or maturity – was now being granted political power. 3 But in the industrial society that was now beginning to emerge, victory went to capital’s claims to guarantee conceptual values. It has been aptly stated that capital ‘was not only mobilized during the first golden age of European liberalism; it was also nobilized.’4 The first main feature of the Conservatives’ opposition to broader voting rights was thus a defence of what were regarded as the legitimate interests of wealth. Power was in the hands of the landowners, especially the large landowners. Their political position was threatened by demands for extending the franchise. Without being concealed in any ideological reasoning, the class interests of rural property owners were often presented as arguments in the suffrage debate. The Constitution Committee of parliament justified its rejection of suffrage demands with the following words: ‘The committee wished to prevent the representatives of landowners’ important interests from being reduced to an insignificant minority in the Second Chamber in order to make room for an overwhelming majority of representatives of the so-called working class.’ 5 A university professor and member of parliament explained that the ruling classes were not willing to ‘abdicate completely’.6 Another MP warned the landowners against ‘the downfall of their class’.7 The conservatism of wealth was regarded as guaranteeing that the country would develop calmly. Landowners could not depart from the country in times of danger as could ‘the loose classes’.8 In words that hark back to Edmund Burke farmers were portrayed as especially calm and quiet, in sharp contrast to the supporters of voting rights, ‘these popular speakers, these agitators and these chairmen of demonstrations . . . Can it be proper and right to let these people come to power? Power now placed themselves under the shelter and protection of laws and religion’.9 Tax policy motivated over-representation of the upper class. Because the wealthy paid more taxes, they should also have more of a say. 10 Others pointed out that constitutional changes were unnecessary because inflation and wage hikes continuously pushed new groups over the qualification regulations, admittedly albeit slowly.11 The best should rule was, in short, the ideology of the Conservatives. It is not primarily the rights of the individual that should determine the issue of broadening political voting rights, but the best interests of the state. It is thus both the right and duty of the state to ensure that suffrage regulations are designed not to accord a preponderant influence to those members of society who have the smallest prerequisites for attaining the independence, the maturity of political judgment and the experience that should be found among those to whom influence on matters of state should be entrusted.12 Now, of course, those who had acquired their wealth through inheritance also enjoyed the right to vote. Could an inherited fortune also be a sign of ability to serve the public? Yes, this was the case: ‘If it is inherited money, it often means

Sweden

267

education and an interest in public affairs; if it is earned money, it often means thrift, industriousness and energy.’13 A wealthy person had more time and interest to devote to political issues. The same concept could also be expressed negatively: a poor person had to devote all his time to working, had no time to form an independent opinion on social issues and was thus not ripe for the suffrage. The hard work with which our workers are occupied, insolvency, shortages, poverty, the uncertainty of their whole existence, in short all these weighty concerns that every man with a heart would very much like to lift from them if he could, must necessarily limit their political range of vision to comparatively narrow conditions, but on the other hand make their senses open and receptive to the teachings of those social improvers who single out the existing social order as the root and origin of all evil. It is thus also to be feared that if they ever came to power, they would exercise it without recoiling from the violent overthrow of social conditions, which overthrow they naturally feel they have nothing to lose from, but everything to gain.14 This fear of the leaders of the working class ran like a red thread through the Conservative opposition to broadening the suffrage,15 and as a warning lesson they pointed out ‘the promised country of universal suffrage, America’ where corruption has spread and the label politician ‘which should be a title of honor had become a term of abuse . . . It is a general experience that the more voting rights are extended, the lower the level of representation.’16 The notion of giving greater influence to the talented was developed into an advanced, so-called organic theory. According to this theory, the state was an organism whose various body parts corresponded to different social classes. The monarch and the upper classes corresponded to its head and higher functions, the working classes to its lower functions. Progress had to follow the laws of this social organism, assuming the form of the quiet growth of the organism without sudden changes. 17 This doctrine was extremely influential because it was taught as part of the curriculum in philosophy and political science at the universities and became something of a bureaucratic philosophy for Sweden’s higher civil servants. The mutual duties and rights of the various social classes were postulated. It was not necessary to investigate whether such a harmony of interests among the classes existed. ‘In this way,’ writes the Swedish political scientist Herbert Tingsten, ‘the prevailing philosophy became a coulisse of suggestive concepts fitted together into an artistic pattern that concealed personal interests and social struggles.’18 The class interests of farmers would seem to have little in common with this subtle academic metaphysics. But the consequences of both arguments were the same: suffrage rules should be designed in such a way as to assure the members of the ruling class greater political influence than others. Wealth had its own interests that deserved to be represented. It was also a sign of the ability to form an independent opinion.

268

Leif Lewin

The left rallied around the demand for universal male suffrage. ‘One man – one vote’ was the slogan. A citizen’s rights should not depend on his economic position. Everyone must personally participate in electing the representatives of the people. As time went on, this view was expressed more radically: not merely a lowering of all barriers but their elimination, not only an equal franchise to the Second Chamber but also in local elections, not only universal and equal voting rights for men but also for women. The left justified its opinion mainly by citing the so-called personality principle. It had pretensions of representing a morally higher view than the right’s ‘worship of money’. Left politicians called for the introduction of universal male suffrage in Sweden with the personality principle as the chief motivation.19 Outside parliament there was a suffrage movement that gathered several hundred thousand signatures.20 In 1893 and 1896, elections to a so-called People’s Parliament took place on the basis of universal suffrage for both men and women. The people demanded, ‘in the name of justice and civil equality’, a system of representation based on personality, not money. This is ‘our holy, just cause: the people’s right of self-determination over their own fate.’21 More and more bills were presented in parliament asking for universal suffrage22 and the suffrage issue dominated the parliamentary party system. In 1895 a liberal party emerged with the suffrage issue as the main point of its programme. In 1900 the modern Liberal Party was formed with Karl Staaff as the party leader, immediately seizing the initiative on the suffrage issue.23 Under the firm leadership of Hjalmar Branting, the Social Democratic Party shifted from a revolutionary to a reformist policy, whose main expression was the struggle for universal suffrage. During the 1902 parliamentary debate on voting rights, the Social Democrats underlined their demands with a three-day demonstration strike. During the parliamentary debates on the subject, Branting sometimes spoke with surprising moderation for universal suffrage in a way that differed little from Liberal speakers; he based it on the personality principle and the value of ‘the equal political rights of all members of society’.24 The Liberals and Social Democrats thus fought side by side to introduce universal male suffrage. In addition to the personality principle the left attached a number of other, partly contradictory arguments. The first concerned the impact of suffrage reform on the social order. One of the chief rightist arguments was that broadening the franchise would disturb social stability. The leftists replied somewhat ambiguously. On the one hand, they tried to calm the right. Universal suffrage would eliminate the discontent in society and consequenly the threat of revolution.25 On the other hand, it was stated almost as self-evident that universal suffrage would, in fact, have an impact on the social order. When Hjalmar Branting was addressing the People’s Parliament, not the ordinary Parliament, his tone was different. The purpose of the proposed reform was to abolish the glaring differences between social classes and bring about an economic levelling. 26 A second argument was the principle that equal duties should be balanced by equal rights. But unlike the Conservatives, by duties the left did not mean the tax

Sweden

269

burden but compulsory military service. A person who was required in times of danger to ‘wear a bloody shirt for his native land’ should also be entitled to participate in elections to parliament,27 a statement that did not exclude heavy criticism of militarism.28 A third contradictory argument emerges in the debate on the ability of people to form an independent opinion on political issues. On the one hand, the left, eloquently supported by the novelist and dramatist August Strindberg, rejected the Conservative proposal that a person showed his abilities by creating a fortune. The richest people were not the best to rule. Wealth was rather a sign of selfishness. The new age, that worshiped economic success at any price, was an age of humbug.29 On the other, the left sometimes admitted that the mass of people were actually not very capable of taking a political stand. But this could also be cited as an argument for broadening the suffrage. In keeping with John Stuart Mill’s doctrine of democracy as an educational process, they argued that individuals would develop a public spirit if they could take part in politics.30 Finally, there was a fourth constellation of arguments based on natural law and utilitarianism. When the Left invoked the personality principle, it was a human right argumentation. Liberals and Social Democrats also maintained, however, that a reform would be to the benefit of society.31 But if suffrage was a natural right, it did not have to be justified by pointing to its usefulness; a right should be respected regardless of its consequences. And if it was useful to society, it did not have to be justified by arguing that it was a natural right, because there would be reasons for the government to implement it anyway. But in the world of suffrage supporters, all good things coincided. It was not only ‘fair’ but also ‘wise’ to introduce universal male suffrage. No matter what arguments the left resorted to, it was unable, however, to overcome the right’s resistance to broader suffrage. A solid conservative majority in the First Chamber resisted every reform proposal, ‘wisely slow in action but firm and strong in opposition’. These words from the 1890 constitutional committee described a reality that leftist demands for suffrage reform were unable to overcome for decades.

The strategies When the Conservative Party leader Arvid Lindman became prime minister in May 1906, he made it one of his top priorities to resolve the suffrage issue. No matter how rigid the positions of the right and left, there was one definite trend from one election to the next: economic growth would in the long run help more and more industrial workers to fulfil the economic requirements to vote, and there was no doubt as to which side these new voters supported. The Conservatives were the losing party. In the prevailing system of majority rule in single-member districts one district after another switched from a Conservative representative to a member of parliament for the left. In the wave of industrialization that was now sweeping across the country, the political map was being repainted in red – this was the disturbing perspective that the Conservatives had to face.

270

Leif Lewin

Could anything be done to change this trend? Could a skilful Conservative politician take any step to prevent the party from becoming the victim of what appeared to be a major long-term disaster? Much can be done in a situation like this – if only the politician is imaginative enough. As a matter of fact, the very notion of politics as rational action is most clearly illustrated when a politician faces a defeat; when he is a potential loser, that is, an actor who, according to the majority rule, is about to lose if nothing is done about the course of events. If we put ourselves in the situation of a potential loser, the task is to work out alternatives or what in rational choice and game theory language is called strategies. As a potential loser, Lindman faced a hostile left who had the majority position and consequently the power to introduce a system of universal suffrage, which would mean that an overwhelming number of electoral districts would shift from the Conservative incumbent to the left candidate. What strategy could Lindman use to stop this unfortunate development? The 44-year-old Arvid Lindman, a naval officer and industrialist, has been described as ‘the daring young Prime Minister’,32 ‘a strong, energetic force’,33 one of ‘the young conservative activists, accustomed to fighting their way and running things’.34 He was an extremely gifted tactician and he was not ready to give in just because he had a minority position. He decided to apply a strategy that could be called ‘amending the agenda’. What was necessary, Lindman argued, was to add some aspect to the suffrage issue that would split the majority left. Obviously, universal suffrage was unavoidable. Sweden would soon be the only country in Europe that did not allow all adult men to vote. But could not this process take place in a way less harmful to the Conservatives than by embracing the suffrage ideas of the left? When we try to reconstruct Lindman’s thinking on the suffrage issue, we see that the application of rational choice theory by no means compels the researcher to ignore structural causes. The difference between rational choice theory and structural-oriented theories is only that in the former the researcher tries to state the role that the structural circumstances play in a decision-maker’s calculation. In explaining Lindman’s suffrage proposal by means of rational choice theory, we very much have to take into account the structural conditions represented by industrialization, the general rise in wage levels and the resulting increase in the number of industrial workers entitled to vote. But we present the issue as Arvid Lindman saw it. We would like to demonstrate how he perceived the problem and how he reasoned in order to counteract the threat to the position of the Conservative Party that these trends represented. The election method turned out to be the solution. Conservatives had to accept universal male suffrage. But majority rule should be substituted by proportional representation. This change would at least guarantee the Conservatives an influence equivalent to their strength in the electorate.35 Lindman was not the inventor of this strategy. Among the Conservatives there had over the years been much talk about ‘universal suffrage with guarantees’ 36 and this issue was also the cause of the cabinet crisis that made Lindman prime minister. At the paliamentary session in spring 1906 a members’ bill suggesting

Sweden

271

proportional representation had been rejected by parliament by a rather thin margin. The starting point of Lindman’s calculations was the number of votes cast for and against this suffrage bill. The figures indicated that there was already a majority in favour of proportional representation in the First Chamber. In the Second Chamber at least twenty MPs would have to be persuaded to switch their allegiance from majority rule to proportional representation. In other words, it was necessary to phrase the amendment on proportional representation in such a way that it attracted new supporters in the Second Chamber without thereby losing its majority in the First Chamber. By the late summer and early autumn of 1906, there was already speculation as to what manoeuvre Lindman was planning on the suffrage issue.37 On 2 February 1907 the government submitted its proposal to parliament. 38 Its text represented the perspective of the potential loser. The warning was that the ruling majority would not respect the just interests of the minority. Therefore, to strengthen the position of the minority, the government proposed that the introduction of universal male suffrage be accompanied by a switch from majority rule to proportional representation in both chambers. 39 To win a majority in the Second Chamber, the government had added three radical elements: a lowering of the voting age from 25 to 24; the extension of proportional representation to the county councils and big-city municipal councils which served as the electors to the First Chamber; and a limit of 40 on the number of votes one person could cast in local elections, where the franchise was linked to tax liability. Lindman was hoping that the last rule in particular would split the left and attract the necessary twenty or so new votes. The proposal was specially worded to take into account the farmers from the forested counties of northern Sweden. Because of bad experiences with forest product companies they had a negative attitude to ‘Big Business Conservatives’ and voted more leftist than most farmers.40 They would probably be receptive to proposals that limited the influence of wealth on local government. Lindman did not hesitate to hold out this bait to the small farmers in the text of the government proposal. These farmers were said to be squeezed between the big companies and the labour movement. Limiting local voting rights to a maximum of 40 votes per person would improve their standing in relation to the companies. Introducing proportional representation would protect them from the outrage of the left. Industrialization was now advancing into traditional agricultural districts. ‘With universal suffrage and elections in single-member districts, the time will not be distant when the interests of farmers here and there in the country are not well represented in elected bodies.’ But because of the proposal on proportional representation ‘the danger of a shift in the direction now indicated is very greatly diminished. Even if farmers are no longer in a majority, agriculture should nevertheless have a chance to enjoy its fair share of representation.’ 41 The proposal was discussed in both chambers. The First Chamber’s reaction was severe. Sweden had a Conservative government and yet a proposal was presented on the favourite issue of the left: universal suffrage! The proposal was criticized by

272

Leif Lewin

two of its most conservative members, and no one defended it.42 In the Second Chamber the reception was more mixed. Because of the uncertain parliamentary situation members of parliament spoke very cautiously and it was not possible to draw any conclusions as to the proposal’s prospects.43 The main alternative to the government proposal was a Liberal bill co-sponsored by a total of 73 MPs headed by Staaff. This bill proposed universal male suffrage while retaining majority rule.44 One name was missing from the Liberals’ main bill: Daniel Persson i Tällberg, a farmer from northern Sweden. The explanation came as late as 16 April when Persson i Tällberg submitted his own bill on the suffrage issue, co-sponsored by four other Liberal farmers. This brief bill, characterized as ‘a purely tactical move’, 45 began like many other leftist statements with a criticism of the long-time Conservative opposition to suffrage reform. ‘Only as a last resort these days, forced by circumstances, do they seem to have realized the serious importance and implications of this issue.’ Tällberg now wanted to make an effort to save the suffrage issue from ‘the foggy waters of mistrust, where the various parties cruise around the issue and, constantly colliding with each other, nearly forget their real task: trying to tow the issue to a safe harbor.’ He could accept the governmental proposal if two more conditions were fulfilled, both concerning the First Chamber: looser election requirements and a salary for the elected.46 The defectors had stated their price. It was higher than Lindman had expected. But he was also willing to pay this price. The Prime Minister threw his support behind the Tällberg bill. To their dismay, Liberal Party leaders saw the ranks of the left in the Second Chamber divided over Tällberg’s bill.47 In this way, the left split in the Second Chamber exactly as Lindman had hoped, and in this chamber the government won the vote by a large majority. The situation was much more difficult in the First Chamber, which had already strongly opposed the original government proposal. And, to make it even worse, now Lindman had taken an additional step in a radical direction to please Tällberg! Lindman said that he was fully aware of what a sacrifice the Conservative Party in the First Chamber, which he also belonged to, would be making if it supported him. But what would happen if the First Chamber remained unchanged while the Second Chamber was now being reformed? Conflicts between the chambers would increase, perhaps becoming so great that the whole bicameral system would be damaged. There were, of course, those who wished to undermine or even abolish the First Chamber – the members certainly understood this allusion to the detested Staaff.48 After the Prime Minister finished speaking, a storm of criticism against him broke out. With disarming honesty the Prime Minister confessed his tactical methods. As every military man knew, Lindman reminded the chamber, there was sometimes reason ‘to give up one position, without losing sight of your goal, in order to gain another.’ Everyone knew that the suffrage issue could not be resolved in the way outlined in the original government proposal. Should the government alone then stick to its old views? Should it sit as a silent witness and watch the chambers arguing about the proper solution? Among those who now condemned

Sweden

273

the government, were there men willing to form a new government that better understood the art of governing? Perhaps that art would not be so difficult if, as the speakers seemed to be saying, it only meant sitting, with arms crossed, waiting for Parliament to decide matters.49 The threat of an imminent government crisis was in the air. In a bold fighting mood, the following day Lindman announced a vote of confidence – not on his own government proposal but on the Tällberg bill. 50 Some people regarded this as going much too far. It was ‘a new theory not only in Sweden but, I believe, in the rest of the parliamentary world, that a government uses the threat of resignation to persuade a chamber that wishes to approve that government’s proposal to vote for a more far-reaching reform.’51 But after the Prime Minister’s message, the mood in the First Chamber began to shift. Immediately after the announcement of a vote of confidence, one member declared that he ‘had not heard a government called weak when it wishes to maintain leadership up to the very end.’52 The debate increasingly came to resemble an open vote and speaker after speaker declared his support for Lindman. But it was a conversion in the face of the seemingly unavoidable. Finally the First Chamber approved the payment of salaries to its own members while maintaining strict conditions for election to the First Chamber. The versions approved by the two chambers were not too far apart to permit a compromise to be reached. They agreed to lower the wealth requirement for election to the First Chamber to 50,000 kronor and the income requirement to 3,000. Parliament approved the suffrage reform. According to the Constitution, it had to pass the identical bill again after an intervening election. Final approval consequently came in 1909. The suffrage issue was thus finally settled. With great strategic skill, though at a high price, Arvid Lindman had split the left and won a prestigious victory on the left’s own pet issue. There was great bitterness among the leaders of the left. Both the Liberal leader Karl Staaff and Social Democratic leader Hjalmar Branting voted no when universal suffrage was introduced in Sweden.53 Staaff was especially disappointed. As his biographer remarks, he was caught off guard by Lindman’s resolute behaviour during the many twists and turns of the suffrage issue. And of his friends, ‘several became so downhearted that they had a hard time maintaining their balance.’54 But for Arvid Lindman, the suffrage reform was a parliamentary triumph that he for the rest of his life remembered with great satisfaction.

Explaining the decision Why was universal male suffrage introduced by a Conservative government ideologically opposed to this demand? Game theory can help us understand. We will try to explain the decision by summarizing our reasoning in a game matrix. This matrix is not a voting record that states all the proposals and bills introduced; it is not even the case that a vote took place exactly between the alternatives listed in the matrix. The function of the game matrix is, instead, to capture the essence of the issue at hand and to simplify and clarify it so that the fundamental problem becomes evident.

274

Leif Lewin

For this reason we will simply pose two issues against each other: on the one hand, universal male suffrage versus limited suffrage; on the other, majority rule versus proportional representation. The actors are the right and the left, modelled on Lindman and Staaff respectively. Left’s demand for universal male suffrage was what started the decision process. The right’s preference was limited suffrage, the left’s universal male suffrage. As the left’s position among voters was becoming stronger, it became increasingly likely that its demand for universal male suffrage would be realized. To soften the effects of this reform, the right amended the agenda to include a shift to proportional representation. The new question posed by the right was a choice between two alternatives, majority rule or proportional representation. The result was two intersecting questions, more complex preference orderings by the actors and a more interesting decision-making situation. In this reconstruction, the right’s preference ordering is first, limited suffrage with proportional representation; second, limited suffrage using majority rule; third, universal male suffrage with proportional representation; and last, universal male suffrage with majority rule. The left’s ranking is first, universal male suffrage with majority rule; second, universal male suffrage with proportional representation; third, limited suffrage with majority rule; and last, limited suffrage with proportional representation. In the game matrix, we assign the best alternative for each actor a utility of 4, the second best 3, the second worst 2 and the worst 1. This gives us the matrix in Figure 14.1. Are there any problems in summarizing the preferences of the right and the left on the suffrage issue as we have done? As far as the left is concerned, the reconstruction is fairly unproblematic: there is no doubt that the left preferred universal suffrage to limited suffrage and majority rule to proportional representation. As far as the right goes, it is likewise obvious that the broadening of suffrage was the most important litmus test and that in terms of election methods, it preferred proportional representation to majority rule. But is it correct and fair to say that even some years into the twentieth century, the right preferred a restricted suffrage to universal male suffrage? Yes, our analysis of the ideological debate leads to the conclusion that this was indeed the case. There were, admittedly, a few firm believers in democracy among the Conservatives. But the dominant opinion in the Conservative Party was that a limitation of voting rights was still preferable, given the First Chamber’s task of representing maturity of judgement and taking

The left

The right

Figure 14.1

Universal male suffrage

Limited suffrage

Majority rule

1, 4

3, 2

Proportional representation

2, 3

4, 1

Sweden: the suffrage game.

Sweden

275

responsibility for the preservation of social order. As risky as it may be to make statements about what an actor ‘sincerely wanted’ – as opposed to the position he adopted – there is sufficient evidence in the proceedings of parliament, including Lindman’s own speeches, to justify the opinion expressed in the Tällberg bill that it was the coercive force of circumstances, not any profound ideological change, that now persuaded the Conservative government to propose universal male suffrage. If we wish to list the true preferences of the actors, it seems proper to let the right put limited male suffrage ahead of universal male suffrage. In the matrix both players have dominant strategies. That means that for each of them one alternative is always to be preferred instead of the other regardless of the choice of the other player. The right should always choose ‘proportional representation’, which will give them a value of 2 or 4 instead of 1 or 3. In the same way the left should always choose ‘universal suffrage’, which will give them 4 or 3 instead of 2 or 1. If both actors are rational, it is then obvious how they should choose. The quadrant thus singled out is the solution of the game. From this no one could move, given the choice of the other player, without his utility deteriorating. As we can easily see, the combination of the rational choices of right and left leads to the lower left quadrant. ‘Universal suffrage with proportional representation’ was also the decision that was actually made. But this solution did not represent the first preference of any actor. The suffrage reform of 1907–9 is an outstanding example of one of the main findings of rational choice theory: there is no decision rule that guarantees that a collective decision will be rational, even though the individual actors behave rationally. There is no guarantee that a democratic decision reflects the true preferences of a majority. No one was really pleased with the suffrage decision; no one’s first preference was adopted. But no one was left out completely, either. The left’s demands for broader voting rights were implemented, but the majority rule had to be sacrificed. The Conservatives finally accepted universal male suffrage but won acceptance for proportional representation, which they regarded as a guarantee against majority oppression. Finally, the game matrix also gives us more precise information on the distribution of gains and losses between actors. Surprisingly enough, it turns out that Karl Staaff, who was very disappointed, received 3 while the triumphant Arvid Lindman received only 2. Yet political successes are not judged according to what is ideologically desirable, but according to what is strategically possible. Even though Staaff got his way on the main issue of expanding the suffrage, he was deeply disappointed because he had seemed so likely to achieve his very first preference. And then he was defeated by his main opponent on the issue he felt most strongly about. Arvid Lindman’s greatness consisted of the fact that from the viewpoint of a potential loser, he intuitively understood the decision-making situation he was in, he knew what room for manoeuvring he had, and he was able to manipulate the game in such a way that the outcome, in our terms, was 2 instead of 1. The former value here stands for a considerable ideological sacrifice, as Lindman himself was very much aware. The Tällberg bill made the price of proportional elections high. The conservatism of wealth and an independent view of society were values that

276

Leif Lewin

Table 14.1 Sweden: transitional electoral results, 1905–14 Year

Social Democrat v/s

Liberal v/s

Conservative v/s

1905 1908

10/6 15/15

45/47 47/46

45/47 38/39

1911 1914a 1914b

29/28 30/32 36/38

40/44 32/31 27/25

31/28 38/37 37/37

Key: v/s: percentage of votes/percentage of seats.

the Conservatives now had to sacrifice, after decades of struggle, in order to prevent the creation of a Parliament totally dominated by the left. The age of upper-class rule was past. But the party that represented the upper-class interest had been given a chance to survive.

Notes 1. The research for this article was conducted in the framework of the research programme ‘Politics as Rational Action’, Department of Government, Uppsala University, sponsored by the Bank of Sweden Tercentenary Fund. A longer version was published in one chapter in my book Ideology and Strategy: A Century of Swedish Politics. Cambridge and New York: Cambridge University Press. 2. Proceedings from the Swedish Parliament: Prop 1862/63: 61, p. 41. 3. KU 1862/63: 7, p. 31. 4. Georg Andrén, Tvåkammarsystemets tillkomst och utveckling. Sveriges Riksdag II: 9. Stockholm, 1937, p. 65. 5. KU 1902: 6, p. 99. 6. AK 1894: 30, p. 49 (Boethius). 7. AK 1906: 54, p. 47 (Hermelin). 8. AK 1890: 28, p. 32 (Redelius); FK 1888: 21, p. 36 (Casparsson); FK 1902: 35, p. 24 (Rodhe). 9. AK 1890: 28, pp. 2–23 (Jansson). 10. FK 1893: 32, p. 45 (Billing); AK 1894: 30, p. 51 (Boethius); AK 1899: 34, p. 29 (Ivar Månsson); FK 1902: 35, p. 18 (Billing); FK 1906: 48, p. 73 (Nyström). 11. FK 1899: 27, pp. 25–6 (Åkerhielm); FK 1888: 21, p. 36 (Casparsson); FK 1893: 32, pp. 58–9 (Nyström); KU 1899: 14, p. 9; FK 1896: 23, pp. 45–6 (Treffenberg). 12. KU 1899: 14, p. 5. 13. FK 1902: 35, p. 65 (Palmstierna). Also KU 1896: 11, p. 19. 14. FK 1896: 23, p. 40 (Säve). 15. FK 1896: 23, p. 30 (Nyström); AK 1899: 34, 30 (Ivar Månsson); AK 1899: 34, p. 42 (Lindblad); FK 1902: 35, p. 24 (Rodhe). 16. FK 1893: 32, pp. 51–2 (Casparsson). 17. Rudolf Kjellén, Om uppgiften för ett fosterländskt förbund. Göteborg, 1896, p. 7; FK 1893: 32, 57 (Nyström). 18. Herbert Tingsten, Den svenska socialdemokratiens idéutveckling I. Stockholm: Tiden, 1941, p. 29. 19. Adolf Hedin, ‘Hvad folk väntar af den nya representationen. Femton brev från en demokrati till svenska riksdagens medlemmar’, 1867, in Tal och skrifter, ed. Valfrid

Sweden

20. 21. 22. 23. 24. 25. 26.

27. 28. 29.

30. 31. 32. 33. 34. 35.

36. 37. 38. 39. 40.

41. 42. 43. 44. 45. 46. 47.

277

Spångberg. Stockholm: Bonniers, 1904, pp. 15, 32 and 54; Mot AK 1880: 146; Mot FK 1890: 46. On the suffrage movement, see Torbjörn Vallinder, I kamp för demokratin. Rösträttsrörelsen i Sverige 1886–1900. Stockholm: Natur och kultur, 1962. ‘Folkriksdagens manifest till svenska folket’, in 1893 års folkriksdag. Handlingar och beslut. Stockholm, 1893, pp. 31ff. Mot AK 1897: 130; AK 1902: 53, pp. 58–9. Vallinder, 1962, pp. 132–3 and 211. Tingsten, 1941, II, p. 61. Mot FK 1890: 46, p. 7 (Mankell); Mot AK 1899: 226, p. 2 (Bergström et al.); AK 1904: 65, p. 63 (Lindhagen); AK 1890: 28, p. 16 (Bratt); AK 1899: 34; AK 1902: 52. Mot AK 1902: 197, p. 8 (von Friesen et al.); Mot AK 1899: 226 (Bergström et al.). Karl Staaff, Fridtjuv Berg, David Bergström, Oskar Eklund, J. Byström, Edvard Wavrinsky, Magnus Höjer, M. F. Nyström and Curt Wallis co-sponsored both bills; 1893 års folkriksdagar. Handlingar och beslut, pp. 55–6. Vallinder, 1962, pp. 268–9. Tingsten, 1941, II, pp. 147–85. Johan Bergman, according to Vallinder, 1962, p. 263; AK 1894: 30, p. 53 (Schönbeck); minority statement in KU 1896: 11, p. 21 (Elowson); AK 1899: 33, p. 20 (von Frisen); Andrén, 1937, p. 282. Mot FK 1890: 46, p. 7 (Mankell); minority statement in KU 1902: 6, p. 114 (Moberg). FK 1890: 24, p. 38 (Mankell). Ivar Andersson, Arvid Lindman och hans tid. Stockholm: Norstedts, 1956, p. 106. Erik Timelin, Ministären Lindman och representationsreformen 1907–1909. Karlskrona, 1928, p. 89. Leif Kihlberg, Karl Staaff II. Regeringschef, oppositionsledare 1905–1915. Stockholm: Bonniers, 1963, p. 97. On the background of the suffrage reform and its implementation, see especially the extensive account in Timelin, 1928; also Andrén, 1937, pp. 450–85; Axel Brusewitz, Kungamakt, herremakt, folkmakt. Stockholm: Prisma, 1964, pp. 17–32 (first published in 1951); Anderson, 1956, pp. 99–110; Kihlberg, 1963, pp. 104–18. Unfortunately Lindman did not begin his diary, which has now been published, until after the suffrage reform. Kjellén, 1896. Anderson, 1956, p. 106. It was divided into three government bills: no. 28 (political voting rights), no. 29 (municipalities) and no. 30 (county councils). Prop 1907: 28, p. 11. Leif Lewin, Bo Jansson and Dag Sörbom, The Swedish Eelctorate 1866–1968. Stockhlm: Almqvist & Wiksell, 1972, pp. 38 and 79; Sten Carlsson, Lantmannapolitiken och industrialismen. Partigruppering och opinionsförskjutningar i svensk politik 1890–1902. Stockholm: Lantbruksförbundets tidskriftsaktiebolag, 1953, pp. 16ff. Prop 1907: 28, p. 13. FK 1907: 6, pp. 6–18. AK 1907: 8, pp. 4–33. Mot AK 1907: 203. Ibid., p. 346. Mot AK 1907: 252, quotation from pp. 1–2. SäU (proceedings of the special committee) 1907: 3 and 4. In the First Chamber, the government supported a members’ bill that had been worked out by some First Chamber members headed by Olof Jonsson i Hov, which merely stipulated that chamber members would be paid a salary. In the Second Chamber, where there was a bill by Hans Andersson et al. incorporating the Tällberg demands for both a salary and a lowering of property requirements, the government merely appealed for a decision that would pave the way for a final agreement.

278

Leif Lewin

48. 49. 50. 51. 52. 53.

FK 1907: 38, pp. 48ff. FK 1907: 39, pp. 38–9. FK 1907: 40, p. 20. FK 1907: 40, p. 37 (Trygger). FK 1907: 40, p. 21 (Tamm). When the bill came up for reaffirmation in 1909, Branting again voted no, while Staaff – whose party now aimed for women’s suffrage and the complete democratization of local voting rights – voted in favour. 54. Kihlberg, 1963, pp. 115ff.

References Andersson, Ivar (1956) Arven Lindman och hans tid. Stockholm: Norstedts. Andrén, Georg (1937) Tvåkammarsystemets tillkomst och utveckling. Sveriges Riksdag II: 9. Stockholm. Brusewitz, Axel ([1951] 1964) Kungamakt, herremakt, folkmakt. Stockholm: Prisma. Carlsson, Sten (1953) Lantmannapolitiken och industrialismen. Partigrupperingar och opinionsförskjutnignar i svensk politik 1890–1902. Stockholm: Lantbruksförbundets tidskriftsaktiebolag. Hedin, Adolf ([1867] 1904) ‘Hvad folk väntar af den nya representationen. Femton brev från en demokrati till svenska riksdagens medlemmar’, in Tal och skrifter, ed. Valfrid Spångberg. Stockholm: Bonniers. Kihlberg, Leif (1963) Karl Staaff II. Regeringschef, l oppositionsledare 1905–1915. Stockholm: Bonniers. Kjellén, Rudolf (1896) ‘Om uppgiften för ett forsterländskt förbund’, in Proceedings from the Swedish Parliament. Göteborg. Timelin, Erik (1928) Ministrären Lindman och representationsreformen 1907–1909. Karlskrona. Tingsten, Herbert (1941) Den svenska socialdemokratiens idéutveckling I och II. Stockholm: Tiden. Vallinder, Torbjörn (1962) I kamp för demokratin. Röisträttsrörelsen i Sverige 1886–1900. Stockholm: Natur och kultur. 1893 års folkriksdagar. Handlingar och beslut.

Further reading • Grofman, Bernard and Arend Lijphart (eds) (2002) The Evolution of Electoral and Party Systems in the Nordic Countries. New York: Agathon.

• Lewin, Leif (1988) Ideology and Strategy. A Century of Swedish Politics. Cambridge: Cambridge University Press.

• Lewin, Leif, Bo Jansson and Dag Sörbom (1972) The Swedish Electorate 1866–1968. Stockholm: Almqvist & Wiksell.

15 Switzerland: Introducing Proportional Representation from Below Georg Lutz

Introduction Proportional representation (PR) was introduced in Switzerland at the end of the First World War in 1918, through a popular vote. Earlier attempts in parliament to introduce such a type of electoral rules had been unsuccessful, and two popular votes in 1900 and 1910 also failed. The introduction of proportional representation in Switzerland largely follows a general pattern to the introduction of proportional representation elsewhere. With industrialization and population growth, social problems increased and social democrat parties grew stronger. The socialist party formed an electoral reform alliance with the other large opposition party, the catholic-conservative party. In 1918, when the social democrats, together with the catholic-conservative party, were strong enough to end the almost 70-year long dominance of the radical democratic party, proportional representation was finally introduced. However, the Swiss case has some particularities. Institutional elements First, the cantons’ independence to determine their own political systems enabled them to introduce proportional representation and experience this system in cantonal parliament elections. Several cantons did this before it was introduced at the national level. In addition, the government (Federal Council) had already introduced proportional representation in the canton of Tessin in 1891 to accommodate a violent political conflict there. The cantonal experiences of proportional representation made it increasingly difficult for the ruling radical democratic party to make principled arguments against it. Secondly, the introduction of the popular initiative as a political ‘tool’ in 1891 facilitated electoral reform because it bypassed the need for a parliamentary majority, requiring instead a majority of people to vote in favour of proportional representation. The national parliament, dominated by the ruling radical democrats, opposed its introduction. 279

280

Georg Lutz

Window of opportunity At the end of the First World War social problems were severe and socialist groups were growing stronger. The possibility that social democrats could gain a majority threatened many citizens and the ruling radical democratic party stopped campaigning as heavily against proportional representation for the last campaign as it had done previously. Most of the leading media and many citizens saw proportional representation as a way to accommodate social and political conflicts at the national level, just as it did in some cantons. This explains the very high yes vote in the third popular initiative about PR which took place in 1918. The introduction of proportional representation had immediate consequences: the radical democratic party’s seat share dropped dramatically in the 1919 election, but the catholic-conservatives did not win any extra seats and the social democrats’ gains were smaller than expected. Instead, the number of parties increased: the formation of the Farmers’ and Peoples’ Party is directly linked to the introduction of PR. Even in the first proportional representation election they won a significant number of seats and from then on they were the fourth strongest party for a long time. The radical democratic party’s 70-year parliamentary majority came to an end and the radical democrats had to enter a government coalition with the catholic-conservative party in 1919. (Ten years later the Farmers’ and Peoples’ Party was integrated into this coalition as well.) Proportional representation with four strong parties but no one dominant party and the need to have strong parliamentary consensus in order to avoid or defeat a referendum generated a longlasting and very broad government coalition. From 1959 it included the four major parties in government according to their size.

The initial rules The modern Swiss state was formed in 1848: a short civil war where the liberal and radical democratic forces defeated the catholic-conservative movement ended the ‘ancient regime’ of a loose consociation of politically rather independent cantons. The radical democratic winners formed a unified state under the roof of a common constitution. The constitution was negotiated in a very short amount of time and adopted in a popular vote in 1848. It introduced general (male) suffrage and created political institutions which to a very large extent have remained unchanged until today. The parliament has two chambers, one representing the people (National Council – Nationalrat) and the other representing the cantons (Council of the States – Ständerat). The constitution defined that the National Council was to be elected through majoritarian elections, the details of which were to be approved in a national electoral law. In the Council of States each canton has two representatives – two each for the 19 cantons, and one each for the 6 half-cantons. The decision of how the cantons were to choose their representatives was left to the cantons. This is still the case, and has brought a variety of systems for electing representatives. Some cantons

Switzerland

281

had indirect elections through the cantonal parliament while other cantons had direct elections. The two chambers of parliament elect the government jointly after each general election to the National Council. The government is a body of seven members. The constitutional setting was federal and the influence of the cantons was of great importance in several ways: the two chambers have equal rights, and all laws and amendments have to pass the two chambers independently. These basic political institutions have remained unchanged to a large extent. The only two major changes were the change from majoritarian runoff elections to proportional representation for National Council elections in 1918, and the introduction of female suffrage in 1971 (which was very late compared to many other states in the western world). In 1848 a majoritarian electoral system was beyond question. Proportional representation was not very well known, and parties were not yet strong national political mass organizations. However, in the commission which debated the new constitution after the civil war, the form of electoral districts was a hotly contested topic. The range of suggestions went from letting the entire country form a single electoral district to the partition of the country into numerous singlemember districts. In the end electoral districts were not allowed to overlap cantonal borders, which guaranteed at least 25 electoral districts. Every canton received one seat per 20,000 inhabitants (and an additional seat if the remaining number of inhabitants was greater than 10,000). Each canton was guaranteed at least one seat, although a canton could be divided into several districts. For the first elections which took place in 1848 the regulation was left to the cantons as to how to elect their parliamentary members. This led to a great variety of mechanisms but had the advantage that elections could take place very shortly after the constitution was approved in a popular vote. From 1850 the magnitude and borders of the electoral districts were big issues in debates about changes to the electoral law. It was clear that the bigger cantons should be divided into more than one electoral district following different criteria, such as historical borders and existing administrative units, geographic formation and religious divisions. In 1850 the national government proposed electoral districts with a magnitude of four. However, in practice this proposal was unrealistic because most of the cantons had a number of seats smaller than, or indivisible by, four. In addition, the attempts to solve the problem of districting through technical criteria were heavily undermined by party political calculation to maximize support. The dominant radical democratic party was openly trying to form electoral districts whose magnitudes and borders favoured them in the first parliamentary debate. In the end, the country was divided into 49 electoral districts: 12 single-member districts, 13 two-member districts, 14 three-member districts and 10 four-member districts. According to the electoral law, district magnitude and borders had to be revised after every census. Six revisions took place between 1851 and 1918 which always provoked intense parliamentary debate. The number of districts was rather stable between 1848 and 1919 (the number varied from 47 to 52), but the district magnitudes

282

Georg Lutz

varied more. Often it was easier to reach consensus about enlarging district magnitude (e.g. increasing the number of seats in a given district) than to change district borders. In 1911, when the last redistricting took place, there were only 6 single-member districts left, 7 two-member districts, 10 three- and four-seat districts, 6 five-seaters, 3 six-seaters, 5 seven-seaters and 2 with eight seats. In all six revisions of the electoral districts between 1850 and 1918 the parties tried to influence the magnitude and borders in their favour. The ruling radical democratic party was the most effective and managed to make itself over-represented, especially after the revision of 1881. However, given the spatial distribution of votes, the catholic-conservative party (which had been the main opposition party since the civil war in 1847) was quite well represented as well. It was mainly smaller parties or new, emerging parties like the social democratic party that suffered because they did not have clear geographical strongholds. A three-round runoff electoral system was introduced. To be elected in the first two rounds required an absolute majority, while in the third round a simple majority was sufficient. In 1901 the electoral system was changed to a two-round runoff system. Second and third rounds were rare, since a large proportion of seats were filled in the first round and in the main only some big-city districts necessitated second and third rounds. Usually about 50–60 per cent of the districts – including larger districts – were single-party monopoly districts, where a single party won all the seats. This number decreased slightly and fell below 50 per cent in 1911 which reflected the changes in the party system over time.

The long way to proportional representation The first signs of the idea of proportionality in Switzerland are to be found in some local and regional elected bodies where seats were already divided among the two major religious groups, Catholics and Protestants, in the first half of the nineteenth century. The 1815 constitution of the canton of St Gallen set the composition of parliament to 84 Catholics and 66 Protestants. In 1831 a quota of Catholics and Protestants was set for every constituency in St Gallen. There were similar regulations about the representation of religious groups in other cantons (Glarus, Thurgau). These regulations disappeared with the Swiss constitution of 1848. The rather small districts which did not overlap cantonal borders secured a reasonably proportional representation of the two main political groups, the radical democrats and the catholic-conservatives, which were strongly divided territorially.1 However, proportional representation of parties was not an issue at that time. Parties were loose groupings of individuals, and national or cantonal party structures were almost non-existent. Although parties formed parliamentary groups party discipline was weak and MPs frequently moved from one party to another. A wider and more public debate about proportional representation only started in Switzerland towards the end of the nineteenth century. The introduction of proportional representation can be divided into two phases: an unsuccessful parliamentary phase in which members of various opposition parties tried to

Switzerland

283

introduce proportional representation through parliamentary proposals, and a second phase in which they tried to introduce proportional representation through popular initiatives. Even when a parliamentary decision had been made, it would have required a popular vote because on all constitutional changes a referendum needs to take place. Parliamentary phase In the constitutional debate of 1872 a member of parliament proposed the introduction of proportional representation in constituencies that had more than one seat. This proposal failed. The ruling radical democrat party’s main argument against proportional representation was that they saw no need for minority representation because it is justified in a democracy to have an electoral system that strengthens the main political force in a country (e.g. the radical democrats themselves). The radical democrats also referred to their own historic importance in the formation and unification of the country. These arguments were used until 1918 in debates about proportional representation as well as in debates about redistricting. In the 1870s socialist parties started to form in Switzerland as a consequence of population growth and industrialization. The existing electoral system left these new forces strongly under-represented politically. With the emergence of new political groups the number of proposals in parliament asking for the introduction of proportional representation increased and the issue was almost constantly on the political agenda. As in many other countries it was mainly the ruling party that argued against proportional representation while the under-represented and opposition parties argued for PR. There was a great variety of claims for stronger proportionality, either asking for a limited vote (which would have increased proportionality slightly without the need for a constitutional change) or for a proportional representation system of some other type. The pro-PR argument was that only proportional representation brings equality and real democracy and guarantees fair representation. The ideological argument against proportional representation remained that proportional representation threatened a strong, unifying government which was seen as a sine qua non to keeping the country together; accordingly, there was no perceived need for minority representation (and it was perhaps better if some radical political groups were not represented at all); proportional representation should be introduced at the cantonal level first to gain some experience of its effects. In the early 1890s some new elements started to influence the debate. The Federal Council was faced with severe problems in the canton of Tessin: as in many other cantons, tensions were high between the radical democratic party and the catholicconservatives, which were divided into almost equally strong camps in this canton. However, in other cantons the conflict remained peaceful whereas in Tessin it turned violent. The national government was forced to intervene, and among other things the Federal Council introduced proportional representation in Tessin as a way to accommodate the conflict. Tessin was in other words the first canton where proportional representation was used in cantonal parliament elections.

284

Georg Lutz

At the same time as the national government introduced proportional representation in Tessin, it opposed the introduction of proportional representation at the national level. However, the introduction of proportional representation in Tessin undermined the national government’s position because proportional representation did successfully accommodate the conflict there. Five other cantons also introduced proportional representation in cantonal parliament elections before 1900 (Schwyz, Zug, Solothurn, Neuchátel and Geneva). However, the parliamentary path to proportional representation at the national level was blocked: party majorities remained very stable in the national parliament because new groups were so disadvantaged through the majoritarian electoral system, and thus the radical democrats successfully defended their ruling position even in strongly contested urban areas. Popular vote phase A constitutional change in 1891 opened a new path for proportional representation supporters: the popular initiative was introduced. The popular initiative meant that any group that could collect 50,000 signatures could demand a constitutional change from then on. This made it possible to introduce proportional representation through popular vote. A vote on a constitutional change requires a majority of the people, and a majority of the cantons, in order to pass, and to bypass a majority in parliament which was by nature biased against the introduction of proportional representation because the ruling party had an absolute majority due to the existing system. The first popular initiative to introduce proportional representation was launched in 1898, by which time it was clear that the parliamentary way to achieve this change was ineffective: the ruling radical democrat party rejected even small changes (such as redistricting) that would have a proportionalizing effect. In the 1898 initiative parties of the left (a number of socialist and social democrat groups) and the right (the catholic-conservative party) formed an electoral reform alliance. This was not easy. On the left, where the initiative was formulated, some groups took the view that it would be preferable to win a parliamentary majority under the majoritarian system that was in place, and that therefore a system change was not desirable. Some other left-wing groups favoured means other than the introduction of proportional representation to increase their influence, such as direct elections to the Federal Council (which was indirectly elected by the parliament) assuming that the left would have a chance to win government seats in a direct election. The eventual compromise between left-wing groups was to launch two initiatives at the same time, one to introduce proportional representation and the other for direct governmental elections. On the right, the catholic-conservative party was not unified on the question of proportional representation because in some cantons the majoritarian system favoured the catholic-conservatives. Many partisans of the catholic-conservative party were sceptical about an alliance with the social democrats, fearing that that could benefit the social democrats more than it would benefit the catholic-conservative party. However, in the end the catholic-conservative party’s national convention voted strongly in favour of

Switzerland

285

supporting the proportional representation initiative but remained undecided on the question of direct governmental elections. A total of 64,685 signatures were gathered for the proportional representation initiative and 56,350 signatures were gathered for the direct governmental elections initiative. In the parliamentary debate that followed in 1900, and during the campaign, the topic became a major political issue with two clear camps. The radical democrats and the media they controlled campaigned strongly against proportional representation in parliament as well as in public debate. The radical democrat dominated parliament recommended voters by a two-thirds majority to vote no. The result of the popular vote was decisive: 169,008 votes were cast in favour of PR, and 244,666 votes were cast against it. Turnout was 58.8 per cent which was about average for this time period. However, surprisingly enough, almost the same number of cantons were in favour as against the proposal: 10.5 cantonal votes were in favour, 11.5 were against. For a constitutional change to pass, a majority of the people and a majority of the cantons were required; for the cantonal majority the six half-cantons only count half. It was mainly the cantons where the catholic-conservative party dominated that were in favour of the proposal, whereas the radical-dominated cantons were strongly against it. The range of votes in the cantons was very large and went from 10 per cent yes votes (Appenzell A. Rh.) to over 75 per cent yes votes (Schwyz, Fribourg, Valais). Some cantons were almost split in half. It is worth giving an example: in the canton of Lucerne the divide on the cantonal level was opposite to the divide at the national level. The catholic-conservative party, which was dominant in this canton, was strongly against proportional representation. In contrast, the radical democrats favoured proportional representation because they would benefit from it. However, they were pressured to follow the national party lines and did this with little enthusiasm. The unclear party positions were reflected in the result: only 50.6 per cent of votes in a canton where the catholic-conservative party had a very clear majority were in favour of PR and 49.4 per cent were against it. After the initiative’s defeat the introduction of proportional representation issue calmed down at the national level for a while, but the idea developed further at the cantonal and local levels. Some new cantons and several bigger cities introduced PR. Eight years after the defeat of the first popular initiative some opposition parties on the right and on the left made a new effort and

Table 15.1 Year

1911 1917 1919

Transitional elections in Switzerland, 1911–19

Social Democrat v/s /8 /11 22/23

Radical v/s

Catholic v/s

Farmers v/s

Liberal v/s

Social groups v/s

Other v/s

/61 /54 32/29

/20 /22 22/21

– /2 15/15

/7 /6 5/4

/3 /4 4/5

/1 /1 1/3

Key: v/s: percentage of votes/percentage of seats. Sources: Gruner (1977), BSF (1929), data set from the Bundesamt für Statistik at Neuchâtel.

286

Georg Lutz

launched a new initiative. The social democrats were once again the driving force behind the proportional representation initiative. As in the first attempt, some cantonal catholic-conservative parties were sceptical because of their own dominant situation at cantonal level, but the national catholic-conservative party supported the proportional representation initiative. Compared to the first initiative the collection of signatures was easier and a great success: 143,152 signatures were collected although only 50,000 were required. The parliamentary debate and the campaigns were as intense as they had been in 1900. This initiative was lost by a very narrow margin: 265,194 votes against, and 240,305 in favour (turnout: 62 per cent). A majority of cantons voted for PR and almost all cantons showed an increase in their support for it. One exception was the canton of Fribourg, which had been a catholic-conservative stronghold for a very long time. While in 1900 the catholic-conservative party there was in favour of PR, it changed its opinion in 1910 and campaigned strongly against PR. Introducing proportional representation on the cantonal level was on the political agenda at that time, and supporting proportional representation at the national level would have undermined the anti-PR campaign on the cantonal level, where the majoritarian system worked strongly in favour of the catholicconservative party. The change in the outcome in this canton due to the catholicconservative party’s changed position was remarkable: in 1900 77 per cent of the voters voted yes, whereas in the subsequent voting in 1910 the yes vote fell to 25 per cent. Three years later a group of people from different opposition parties (again dominated by the social democrat party) organized the launch of a new initiative. They collected the necessary signatures in a very short time in 1913. The radical democrat government opposed the introduction of proportional representation as usual, and now used the outbreak of the First World War to freeze the initiative and not hold a public vote. The parliament (which, as mentioned, was dominated by the radical democrats) supported this decision which was of questionable legal standing since the competency to delay a popular vote had not been handed over to the government explicitly. During the First World War social divergence increased in Switzerland as in many other countries because of the lack of a social security system which pushed many families of soldiers and farmers into severe poverty. As a consequence, support for the social democrats increased considerably. Estimates from Gruner (1978: 533) give an increase in support for the social democrats from 80,000 votes in 1911 to 158,000 votes in 1917. This boosted the pro-PR camp. The introduction of proportional representation in more and more cantons had the same effect (most importantly, the canton of Zurich introduced proportional representation in 1916). At the same time the government majority of the radical democrats tried to delay the initiative further by not putting it on the parliamentary agenda. A parliamentary debate was a necessary step before a popular vote could take place and not holding a parliamentary debate was a way of holding back a popular vote. At this time the radical democrats knew that they would very likely lose the next popular vote on proportional representation and this would lead to a big loss of seats in any

Switzerland

287

subsequent election. In the upcoming election of 1917 the socialist party increased its number of seats from 15 to 20. So the party could now show that it was strong enough to win many seats with the current system. The pressure on the government from the catholic-conservative opposition as well as from the socialist opposition increased in parliament and in the media. The pressure became very intense when the socialist movement called for a general strike in 1918 where the introduction of proportional representation was one of the main demands. Eventually the radical democrats stopped trying to avoid a popular vote and realized that they had lost the battle against PR. The success of the third initiative and the swing of public opinion in 1918 was impressive: 299,550 votes (67 per cent) were in favour and 149,035 votes (33 per cent) were against the introduction of PR. Only two and half cantons voted against the initiative this time round. Turnout was rather low (49.6 per cent), which reflected that the battle had been won even before the voting day and that many people were assuming that the initiative would pass anyway and therefore did not go to the polls.

The new rule and its political consequences After the constitutional change in 1918 the preparation of a new law developed very quickly. That every canton would form a constituency was already fixed in the constitution. This created a huge variation in the magnitude of electoral districts: five cantons became single-member constituencies (with plurality rule elections) while the biggest canton (Berne) had 32 seats. However, the advantage was that the long and tiring discussion about districting within cantons that had been a constant battleground since 1848 was avoided and the space for manipulation in the districting process was eliminated. The new law introduced the HagenbachBischoff formula for translating votes into seats. Seats were distributed within each canton. Moreover, it was clear that an open ballot system would be used, with the possibility of panachage and accumulation. The law was passed very quickly and parliament called elections immediately rather than wait until 1920 when elections were due to take place. After the electoral reform, the alliance between the catholic-conservatives and the social democrats collapsed and the catholic-conservatives and the radical democrats formed an alliance against the left. In the media there were expectations that the electoral reform would benefit the social democrats more than anyone else and that the absolute majority of the ruling radical democrat party would end. One of the remarkable aspects of the first proportional representation election, which took place in October 1919, was the increase in turnout from 60 to 80 per cent. This election mobilized many voters who had abstained in the 1917 election. The number of invalid votes remained rather small (below 1 per cent) although the electoral system with open ballot and different options (panachage, cumul, scratching candidates) was relatively complicated. The results were unexpected in some ways. The radical democrats’ loss of seats was very large: their number of seats went down from 103 (which still gave them

288

Georg Lutz

an absolute majority in parliament in 1917) to 60 – just over 30 per cent of the seats. The change in the electoral system immediately ended the dominance of the radical democrats who had been in power for almost 70 years. However, it was assumed that the social democrats would gain more from the electoral reform. Although the social democrats doubled their number of seats from 20 to 41, the seats and vote share of just above 20 per cent was much lower than the party itself and its opponents had expected. The number of seats for the catholic-conservative party remained almost the same as it was under the majoritarian electoral system. Moreover, the success of the Farmers’ and Peoples’ Party was a surprise. The Farmers’ and Peoples’ Party was a conservative split from the radical democrats, mainly in cantons where the catholic-conservative party was rather weak such as Zurich, Berne and Aargau. The first cantonal party branch formed in Zurich in 1917 as an immediate reaction to the introduction of proportional representation at cantonal level there in 1916. The formation of other cantonal party organizations followed in Bern and Schaffhausen in 1918 and in Aargau 1920. The Farmers’ and Peoples’ Party had its original stronghold in the canton of Berne, and half of the party’s seats in the 1919 election (16 out of 32) came from this canton alone. That the electoral success of the Farmers’ and People’s Party was limited to liberal-democratic strongholds has to do with the fact that in the catholic cantons the electorate was already divided between the catholic-conservative party which was rather rural oriented, and the liberal party which was the urban party. In the protestant cantons, in contrast, the liberal democrats covered, in the absence of the catholic-conservative party, a broader range during majoritarian election times, a gap which was filled by the new party after the introduction of PR. One of the main arguments against proportional representation had always been that it would lead to a large increase in party fragmentation. This turned out to be only partly the case after 1919. Smaller parties did win seats but some of them lost seats as well after the electoral reform. However, the fragmentation has always been limited: votes have been highly concentrated in four larger parties. Several elements hindered the formation of new parties. The variance in numbers of seats between the cantons was quite high. Only five cantons had more than 10 seats and only two cantons had more than 20 seats. Consequently, the threshold for a party to gain representation had been higher than 10 per cent in most cantons, and only in two cantons was the threshold below 5 per cent. In fact only Berne with 32 seats and Zurich with 25 seats had an electoral threshold below 5 per cent in 1919. However, due to the allowance of apparentement which means that lists could be put together in order to get a better chance for extra seats, the threshold could be lower than the threshold for single parties would have been. In addition, many of the larger cantons with lower thresholds were rural cantons which disadvantaged the formation of new parties because it was easier to form a party and campaign in urban areas. In sum, the change of the electoral system to proportional representation led to a significant restructuring of the party system and of the political forces in Switzerland. Under the majoritarian electoral system a single party controlled policy-making in parliament and the government for almost 70 years. The first

Switzerland

289

proportional representation elections changed this immediately, leaving four large parties with less than 30 per cent of the votes each from 1919 onwards. Until the mid-1990s the four major parties remained very even and similar to the situation in 1919 when the first proportional representation election took place. The radical democrats’ vote share decreased constantly in the 1920s, and in 1935 their vote share fell below 25 per cent and never recovered. Since then, the radical democrats as well as the catholic-conservative party have had a vote share of 20–25 per cent. The social democrats managed to increase their vote share in successive elections and in the 1935 election became the biggest party for the first time. However, the social democrats’ vote share never surpassed 30 per cent, and fell below 25 per cent in the late 1960s again. The Farmers’ and Peoples’ Party has been the smallest of the ‘big four’ for a long time, with a vote share ranging between 11 and 16 per cent. The change in the seat distribution made new government coalitions necessary because the radical democrats lost their parliamentary majority. In 1891 the catholicconservative party received a seat in the seven-member government for the first time. This was basically a (successful) move of the radical democrats to ‘divide and conquer’ the opposition against further centralization on several political issues, such as the creation of a national state railway company. From 1919 the parties agreed that the catholic-conservative party should obtain two seats in the government coalition to guarantee against a left-wing majority. Only ten years later, in 1929, the Farmers’ and People’s Party was integrated into government as well. The integration of these two parties into government was necessary for two reasons. First, none of the parties was very much larger than the others. The government majority was rather small, party discipline quite weak and broadening the government a necessary step in order to pass bills in parliament. Second, forming larger majority (consensus) government was necessary because of direct democracy: the absence of consensus increased the ‘risk’ of having to hold a referendum, thereby handing power to the people, increasing the uncertainty about an outcome and delaying policy-making.

Conclusion The Swiss case partly contradicts the findings of Stein Rokkan (1970), who argued that ruling political elites switch to proportional representation when they fear losing their majority to the opposition. In other words, the switch to proportional representation is stipulated to be based on elites’ estimations of their own future performance. In the Swiss case such estimates about future performance were clearly evident in debates and they strongly influenced party positions. However, not all parties correctly estimated their future performance. It was clear that proportional representation would bring to an immediate end the dominance of the radical democratic party. The radical democratic party never favoured PR, and never took steps to introduce it. The catholic-conservatives believed that they would do better in proportional representation elections than under the majoritarian electoral system, and this turned out to be a miscalculation. Boix (1999)

290

Georg Lutz

extended Rokkan’s explanation by saying that proportional representation was introduced when there was a strong socialist camp and a divided non-socialist camp. The Swiss case does not conform to that hypothesis either. In Switzerland left parties never exceeded 30 per cent of the vote and before the introduction of proportional representation there were only two parties in the right camp, both of which were better or equally well off under the majoritarian system as under PR. It is not possible to reduce the introduction of proportional representation to a single factor. There are political elements which explain the transition due to the strength and position of political actors, institutional elements such as federalism and direct democracy which favoured the introduction of proportional representation, and a window of opportunity at the end of the First World War which helped to push through the third initiative without any further major resistance. Political elements After the foundation of the modern Swiss state in 1848, the strength of the two main parties remained unchanged for almost 50 years: the radical democrats had a clear majority in both chambers of the national parliament and they had an absolute majority in the government. However, the catholic-conservatives had a large seat share as well. The electoral system favoured the radical democrats but the over-representation was not very large because of the spatial distribution of votes (which enabled the catholic-conservatives to win seats in the areas they dominated). The strengths of the political parties coincided with social changes taking place at the end of the nineteenth century. The industrial revolution and population growth created an urban working class which saw itself as not being properly represented by the traditional parties. This was the opening for the creation of socialist parties, which became stronger and stronger over time. However, these parties found it very difficult to win seats in urban areas under the electoral system that was in operation at the time, because they were in direct competition mainly with the dominant radical democrats. This situation changed during the First World War. At that time, the socialist parties started to win seats even under the majoritarian system. The social democrats were less of a threat to the catholicconservatives than they were to the radical democrats, who suffered more losses as a result of the social democrats’ progress. The catholic-conservatives tended to be based in rural areas, and the attachment of catholic voters to the catholicconservative party was very strong. In addition the radical democrats overestimated (as did everybody else) the strength of the socialist movement and feared that they could be overruled to a large extend in urban areas. This explains why the radical democrats stopped putting a lot of effort into defeating the third initiative at the end; however, the radical democratic party’s opposition to proportional representation remained unchanged until proportional representation was introduced in 1919. Because of the strong bias of the previous electoral system in favour of single-party government, the vote in parliament in favour of proportional representation in the third and final debate about the initiative in 1918 was as small as in all the previous debates in the preceding 30 years.

Switzerland

291

However, in 1918 the clear anti-PR majority in parliament no longer reflected public opinion. The coalition of social democrats, catholic-conservatives and some other small groups had become strong enough to convince the voters to vote in favour of PR. Institutional elements Two different factors have to be taken into account: the strong federalism and the role of direct democracy. (1) Due to the federal nature of the Swiss state, cantons could introduce proportional representation for elections to the cantonal parliament. By 1918 six cantons had introduced PR, including the canton of Zurich (the second largest at that time). The radical democrat dominated Federal Council introduced proportional representation in the canton of Tessin, which had several consequences. First, it became increasingly difficult for the ruling party to argue in principle against proportional representation at the national level. Moreover, it was no longer possible to argue that proportional representation should be introduced on the cantonal level first because many cantons had done that already. There was no evidence at the cantonal level that proportional representation led to ungovernability or very large party fragmentation. On the contrary, in several cases proportional representation helped to resolve political obstruction or to successfully accommodate political and even violent conflicts. In addition, in some cantons the radical democrats were in a minority position and therefore in favour of PR. In the same cantons the reverse was true of the catholic-conservative party: it had a comfortable majority partly because of the majoritarian electoral system and therefore the cantonal party branches opposed the introduction of PR. In all these cantons the catholic-conservatives feared that proportional representation at the national level would strengthen demands for proportional representation in their cantons, too. This made the political battle over proportional representation more complex than it would otherwise have been, and partly explains why the second initiative was unsuccessful. The interplay between the national level and the cantonal level is especially important because the cantonal party structures and organisations at that time were in many cases much stronger than they were at the national level. (2) The popular initiative was introduced in 1891 and became an effective ‘tool’ for the opposition parties. Through the popular initiative it became possible for the opposition parties to place proportional representation firmly on the political agenda without requiring a parliamentary majority. Because of the strong bias in the seat distribution in favour of the ruling party, it would have been much more difficult to introduce proportional representation through a parliamentary vote and the introduction of proportional representation would have taken much longer without the popular initiative. Window of opportunity There was clearly a window of opportunity at the end of the First World War. The social consequences of the war were severe, given the absence of a welfare state.

292

Georg Lutz

Socialist groups were in ascendancy as a direct consequence of this, which threatened not only the ruling radical democratic party and eventually made it abandon its strong resistance to PR, but it might also have convinced many people that giving the social democrats better representation would help accommodate social conflicts. In 1917–18 there was a national general strike which turned violent in some cities. This mobilization led to overestimations of the support for the social democrats, which turned out to be weaker than expected in the first proportional representation elections in 1919. The radical democrats and significant elements within the radical democratic media ended their resistance to proportional representation when they knew they were beaten. These external circumstances help explain the large shift in favour of proportional representation at the time of the third popular initiative. Yet, the national strike and the radicalization of the social democrats challenged their alliance partners, the catholic-conservative party. So one can say that proportional representation was introduced just before the catholic-conservatives were close to abandoning their alliance with the social democrats and go into coalition with the radical democrats against the socialists. Such a move might also have reversed the catholic-conservatives’ position on PR.

Note 1. It is a simplification to speak about two parties, because parties were not compact blocks at all but rather loose groups until the end of the nineteenth century. The two main camps were split into several factions and party discipline was low anyway. However, dividing the party system into two camps reflects the main political division including the division about proportional representation at this time and until the introduction of PR.

References BFS, Eid. Statistisches Amt (1929) ‘Die Wahlen in den Schweizer Nationalrat 1919, 1922, 1925, 1928’, Schweizerische Statistische Mitteilungen, XI. Jahrgang, 1. Heft. Boix, Charles (1999) ‘Setting the Rules of the Game: the Choice of Electoral Systems in Advanced Democracies’, American Political Science Review, 93: 609–24. Gruner, Erich (1977) Die Parteien der Schweiz. Berne: Francke. Gruner, Erich (1978) Die Wahlen in den Schweizerischen Nationalrat 1848–1919. Berne: Francke (French trans. Les élections au Conseil national suisse 1848–1919. Berne: Francke). Rokkan, Stein (1970) Citizens, Elections, Parties: Approaches to the Comparative Study of Political Development. Oslo: Universitetsforlaget.

Further reading • Linder, Wolf (1994) Swiss Democracy. Possible Solutions to Conflict in Multicultural Societies. New York: St Martin’s Press – now Palgrave Macmillan.

• Lutz, Georg and Dirk Strohmann (1998) Wahl- und Abstimmungsrecht in den Kantonen. Bern: Haupt.

Switzerland

293

• Natsch, Rudolf (1967) ‘Die Auseinandersetzung um das gerechte Wahlverfahren im 19. Jahrhundert’, in E. Walder, P. Gilg, U. Im Hof and B. Mesmer (eds), Festgabe Hans von Greyerz zum sechzigsten Geburtstag. Bern: Herbert Lang, pp. 535–58. • Natsch, Rudolf (1972) ‘Die Einführung des Proporzwahlrechts für die Wahl des Schweizerischen Nationalrates (1900–1919)’, in R. Ruffieux, H. Mesmer, A. Lasserre and R. Natsch (eds), Die schweizerische Referendumsdemokratie im XX. Jahrhundert. Fribourg: Éditions universitaires Fribourg Suisse, pp. 119–82.

16 The United Kingdom: Reforming the Westminster Model Patrick Dunleavy and Helen Margetts

Plurality rule voting has historically lain at the heart of the Westminster system and of British ‘exceptionalism’ from its European counterparts. Throughout most of the twentieth century, the fundamental British conceptions of how liberal democracy should link citizens to their governments and political representatives have been based on customary ways of counting votes, have sought only very limited information about people’s preference structures, and have been strikingly unresponsive in many different ways to wider patterns of social and political change. Political representation and accountability to citizens have been interpreted in a minimal way, within a long-lived and arguably ‘primitive’ conception of the scope of accountability and the meaning of citizens’ representation. This approach survived intact for virtually the whole century – despite the transformative impacts of two world wars, massive economic and social class changes and the shift from Fordist and patriarchal social structures to ‘postmodern’ patterns of social life. Although the importance of public opinion and voting in conditioning government policies was always central to the Westminster model, there was no idea that any fine-grained control of public policies should be vested with citizens, nor any notion of constraining the discretionary behaviour of political parties by imposing procedural controls on them. Voters had strictly limited choices and options determined by party elites, and many citizens’ ballots were effectively screwed up and discounted from influencing the selection of representatives as soon as they were cast. Election results could be severely disproportional at a national level, typically with the largest party in terms of votes receiving a strongly exaggerated share of Commons seats (the ‘leader’s bias effect’), while third- or fourth-placed parties nationally accumulated significant vote shares and won few or no seats. Across whole regions of the country it was common for the leading party to win virtually all seats, creating ‘electoral deserts’ for its opponents, a trend that grew sharply worse after the mid-1970s with increasing regionalization of alignments (Jenkins Commission, 1998). The information environment and political finance aspects of elections could become severely unbalanced, always with the Conservatives far outspending their rivals or enjoying a highly favourable print media advantage, with only the restrictions on broadcasting to act as counterweight. 294

The United Kingdom

295

Under the Westminster model the electorate could (normally) expect that the less unpopular of the two major parties would form a single majority government with a Commons majority. But the combination of plurality rule elections and the manifesto doctrine meant that there was no guarantee that the virtually unscrutinized policies enacted by the controlling party’s MPs on auto-pilot would enjoy majority support in any individual policy area. Citizens could have more differentiated local or individual influence with MPs only on issues outside the scope of main party commitments, or on administrative and implementation concerns. Executive action by the government could commit the country to wars and foreign engagements without prior approval by Parliament or citizens, and legislative and public consultation processes were controlled throughout by the government. Interactions in public policy-making were severely limited and there were few effective checks and balances. With no decentralized centres of legislative or executive power above a weak local government level, policy-making in mainland Britain operated on a much larger scale than in most other liberal democracies. Completely missing from this picture were some of the ‘normal’ postulates of either European liberal democracies (notably that seat shares in the legislature should accurately reflect the balance of preferences expressed by voters) or of the United States’ approach (notably that citizens’ rights should be specified in an invariant form and protected by strong institutions insulated from control by this year’s legislative majority). This version of the Westminster model lasted into the last quarter of the twentieth century with only minor changes and adaptations, despite being subject to some stresses and strains. We turn now to consider how much of this picture has altered since 1997.

Electoral system change since 1997 The origins of contemporary electoral system changes can be traced back to the mid-1980s defeats for Labour which by 1987 raised serious doubts about whether Labour could regain power, especially in Scotland which lost out on devolution in 1979. Most of the already implemented new electoral systems originated with the work of the Scottish Constitutional Convention (SCC) on defining devolution proposals in Scotland. Dominated by Labour and the Liberal Democrats, plus churches, unions and other civil society interests, the SCC was always boycotted by the Tories and attracted no corporate participation from the Scottish National Party. The electoral arrangements included in the abortive 1970s proposals for a devolved Scottish Assembly maintained plurality rule and created widespread suspicions in other regions of Scotland that their government would be permanently dominated by lowlands, urban Labour, which boosted the anti-devolution result. So the SCC had to come up with a Lab–Lib compromise on a new voting system, which would also be acceptable to the Scottish National Party. While conceding the need for a proportional system, Labour completely blocked the Liberal Democrats’ top preference for the single transferrable vote (STV), fearing that it would produce a collapse of party discipline and erode their vote in safe areas. These considerations

296

Patrick Dunleavy and Helen Margetts

pushed the SCC towards an additional member system (AMS) along the German model. But Labour successfully insisted on the importance of having a majority of local members of the Scottish Parliament and of choosing a system which maintained constituency contacts even for those seats elected under top-up arrangements. Table 16.1 shows the salient features of the Scottish system which was implemented by the devolution legislation in 1997–98, almost unchanged from the outcome of the SCC’s deliberations. The mixed voting system ‘north of the border’ had a profound impact in Wales, where linked proposals for a lesser degree of devolution stalled for a long time because Labour MPs there insisted on maintaining an explicit defence of plurality rule. However, late in the 1992–97 Parliament the Shadow Welsh Secretary Ron Davies persuaded his reluctant colleagues in the Welsh Parliamentary Labour Party (PLP) to accept a modified version of the proposed Scottish system. The new Welsh AMS variant had a higher ratio of locally elected Assembly members

Table 16.1 1997–98

United Kingdom: key features of electoral systems enacted or proposed,

Scottish Parliament

Welsh Assembly

London Assembly

Jenkins Commisson ‘AV+’ system

Assembly size Mix of local : top-up seats members (%) Elected by

129

60

25

659

57 : 43 Plurality rule in 73 Westminster constituencies

67 : 33 Plurality rule in 40 Westminster constituencies

Top-up areas

8 former Euro constituencies

5 former Euro constituencies

57 : 43 Plurality rule in 14 double or triple borough areas Greater London

83 : 17 Supplementary vote in 543 redrawn Westminster constituencies 80 counties in England; top-up areas in Scotland and Wales; 2 new areas in N. Ireland

7

4

11

1 or 2

15 to 17

11 to 13

25

5 to 11

4.8

5.9

3.6

7.1

5.6

7.1

3.8

8.3

Top-up seats per area Effective district magnitudes Inclusion threshold Exclusion threshold

Note: The inclusion threshold is the minimum level of the vote on which a party could win a seat if conditions such as the fragmentation of the vote were optimal. The exclusion threshold is the level a party needs to win in order to be guaranteed a seat (Rae et al., 1971). Both thresholds given here are the minimum feasible levels. In the Jenkins system (AV+: alternative vote with additional members) these figures assume a basically proportional allocation of local seats between major parties. Where this condition does not hold, the threshold for winning top-up seats could be much higher. For a detailed consultants’ report on why and how the areas were defined as they are for the London system and the Jenkins report, see Dunleavy and Margetts (1998a, 1998b).

The United Kingdom

297

and fewer proportionally allocated seats per top-up area (see Table 16.1). Its sole rationale was to maintain Labour’s ability to win an outright majority in the Assembly in most years, while avoiding the accusations of an automatic or built-in dominance by the South Wales Labour ‘Taffia’ which helped to fatally undermine support for the late 1970s devolution bill (convincingly rejected then by Welsh voters). In March 1997 the Labour and Liberal Democrats headquarters in London publicized a joint concordat on constitutional issues, negotiated by Robin Cook and Robert Maclennan. The Cook–Maclennan pact ratified the two devolution schemes, presaged a similar scheme for a new London metropolitan government, included agreement to introduce list PR for Britain’s European elections (confirmed in the autumn after the general election) and promised to hold a referendum on switching from plurality rule for the House of Commons to an alternative system, which would be defined by a small Commission. In the 1997 general election six weeks later unprecedented tactical voting between supporters of the two parties produced a very strong Commons majority for Labour and doubled the Liberal Democrats’ representation to 46 seats, their highest total for more than half a century. Liberal Democrat voters’ second preferences in 1992 favoured the Tories nationwide, but in 1997 they shifted to split more than five to two in Labour’s favour (Dunleavy et al., 1997: 14). In the autumn of 1997 the confirmation of both Scottish and Welsh devolution in two referenda coincided with a speech by Tony Blair to the Labour Party annual conference which included an explicit call for the next century of British politics to be dominated by ‘radical’ forces instead of the Tory political predominance in the twentieth century. Shortly thereafter the Jenkins Commission was set up to decide with what system to implement the manifesto pledge for a referendum on voting reform for the House of Commons. Its brief included two ‘Labour’ criteria (keeping the link to constituencies and stable government) and two ‘Liberal Democrat’ criteria (proportionality – modified to ‘broad proportionality’ by Labour – and extending voter choice). The definition of London’s electoral arrangements also took place at this time, with the scheme announced in February 1998 again using a mixed system for the London Assembly (see Table 16.1). Blair personally reinstated a provision for the London Mayor to be elected using the supplementary vote (a simplified form of the alternative vote), which had been deleted in favour of plurality rule by a Cabinet committee under Lord Irvine concerned about the multiplication of electoral systems. In the early spring of 1998 Jenkins reportedly met Blair and recommended an AMS solution for electing the House of Commons using a 67 : 33 mix of local to top-up seats. However, the proposal was vetoed and the Commission encouraged instead to work on a mixed system with a much lower proportion of top-up members. In the light of political science research (Dunleavy et al., 1997; Dunleavy, Hix and Margetts, 1998; Dunleavy and Margetts, 1998a) the Commission decided that a mixed electoral system with a remarkably high share of locally elected MPs, and many fewer top-up MPs, could achieve ‘broad proportionality’. The key features of the Commission’s scheme were published in late October 1998 and are shown in Table 16.1. Like the implemented systems in the

298

Patrick Dunleavy and Helen Margetts

table, Jenkins’ solution was a mixed electoral system, but the Commission scheme differed in not adopting straight AMS. Instead it proposed that the alternative vote should be used to elect local members, and that a very small proportion of the Commons (only around a sixth to a fifth), would be elected in semi-local top-up areas to provide a measure of proportionality (see Dunleavy and Margetts, 1998a, 1998b, 1999a for detailed analyses). The Jenkins proposals ran into very strong opposition within the Cabinet, among Labour MPs and in the ranks of some big trade unions and constituency Labour parties in ‘heartland’ areas. The pledge to hold a referendum on voting systems also posed huge problems for the government given the electoral timetable between 1999 and early summer 2001 when the next general election was expected. Blair decided to kick the Commission’s proposals into touch for the time being and to live with the damage this did to the cooperation from the Liberal Democrats. The most notable of these was Paddy Ashdown’s exit from being Liberal Democrat leader under a small cloud in mid-1999. The framework of Lab–Lib cooperation in the end survived, with the new Liberal Democrat leader (Charles Kennedy) making his constitutional priority to achieve electoral reform at local government level (an element completely missed out of the Cook–MacLennan pact). Debate continued about whether the Jenkins proposals were now dead completely or only sleeping, until autumn 2000 when Blair persuaded Labour opponents to agree a pledge that voting reform would be ‘re-examined’ in 2003 in Labour’s manifesto for the 2001 general election. The PM’s stance reflected a conviction that achieving a Labour second term with a secure majority could hinge on sustaining Lab–Lib cooperation to repeat at least some of the vote coordination effects achieved in 1997. It should be clear from Table 16.1 and this quick historical sketch that the already enacted systems of ‘British AMS’ share numerous features. All elect a majority of representatives in local constituencies. They elect top-up or additional members for relatively small areas accessible to or already well recognized by citizens. And they all allocate top-up seats using the d’Hondt rule (also used in the European elections held under regional list PR in June 1999). Neither the Scottish nor the Welsh schemes specified a legal minimum threshold of votes needed to win seats, instead keeping down the effective district magnitude so as to keep inclusion and exclusion thresholds fairly high, an approach which the Jenkins proposals followed closely. However, in London, while the government’s initial proposals omitted any threshold for parties to win seats in the Greater London Assembly, groups representing ethnic minorities, Labour Party activists and academic advisors to the government all argued for a minimum legal threshold as a safeguard against racist parties possibly being able to win seats if party competition for the Assembly should become more fragmented. Ministers eventually decided upon a 5 per cent list vote as the legal threshold for winning top-up seats – the first time any British election has had this feature. The Jenkins Commission scheme shares most of the key features of British AMS but is distinctive in the much smaller proportion of top-up MPs, which means that the system will not guarantee proportionality as the other systems basically do.

The United Kingdom

299

A final system to take note of is the closed list PR method for electing Britain’s Members of the European Parliament (MEPs) which was finally adopted in spring 1998, despite strong calls in the Commons for an open list system to be used instead. In 1999 MEPs were elected in 12 government standard regions, with between 4 and 11 members in each, again using a d’Hondt seat allocation rule. Critics argued that the Labour government’s scheme was an unwise fix, designed solely to smooth the transition from its huge majority of MEPs elected under plurality rule in 1994 to a necessarily reduced seat share under proportional representation. Closed list voting was also argued to be offputting for voters, especially after Labour adopted a thoroughly undemocratic method for selecting and ordering its lists of candidates. Historically Euro election turnouts in Britain have been very low at 34 per cent. Under the new system, with ‘Europe’ unpopular with British public opinion and with a deliberately low-key Labour campaign designed as a damage-limitation exercise, turnout plunged to just 24 per cent. Labour’s vote share was heavily eroded by defections to fourth and fifth parties and by abstentions among its core voters. In July 1999 the Labour National Executive Committee voted to scrap closed list regional PR and to explore instead a version of AMS for the next Euro elections due in 2004. A general feature of all the 1999 and 2000 elections is that turnout levels were lower – and in most cases markedly lower – than those in general elections: 58 per cent in Scotland, just 46 per cent in Wales, only 24 per cent in the European elections and 35 per cent in London. Turnout levels for plurality rule local government elections also slumped to around 29 per cent in 1998, 1999 and 2000, and one Westminster by-election reached a low of 20 per cent in June 1999. Some commentators warned of the dangers of a ‘turnout time bomb’ now that the fierce factional politics of the Thatcher years have been superseded by Blairite consensualism, a lack of strong mobilizing issues and with an apparently weakened opposition in national opinion polls (Dunleavy and Margetts, 1998b). The most important features are the considerable dispersion of the vote evident in the actual 1999 and 2000 results compared with the vote shares shown in the 1997 general election projections. In Scotland the effective number of parties (ENP) score for votes pushed above four for the first time ever in any British mainland election, and in London the list vote under AMS saw 4.7 parties attracting support. In Wales and the European elections the ENP also came close to the four level. And while all the 1997 projections assumed that no fourth parties in England or fifth parties in Scotland and Wales could win seats, in fact three such parties won seats in Scotland’s parliament (one admittedly a deselected Labour rebel MP). The Greens won a seat in Edinburgh, one seat in each of southern England and London in the Euro election, and three seats in the London Assembly. And the UK Independence Party won three MEPs. The nationalist plus other vote reached 38 per cent in Scotland and 36 per cent in Wales, and in the Euro elections it touched 25 per cent (excluding Northern Ireland votes). In London over a quarter of the total list vote did not go to the three main parties. As a result all the 1999 and 2000 elections also saw the ENP score for seats push above three – which should mark a decisive end to talk of Britain as a ‘two-party system’.

300

Patrick Dunleavy and Helen Margetts

All the midterm elections resulted in no overall legislative majority for any party. In the Scottish Parliament it was always expected that Labour could miss gaining an overall majority, and a Labour–Liberal Democrat coalition was rather painfully constructed after the election result. This arrangement survived well the death of Labour’s first minister (Donald Dewar) and the election of a new Labour leader in 2000. It seems likely to predominate in future Parliaments, given the political isolation of the Conservatives and the difficulties for the three other parties in doing any deal with the Scottish National Party for fear of boosting Scottish independence chances. However, the Liberal Democrats fared poorly in the extra competition which proportional elections ushered in, and a more substantial Green presence in Scotland is possible in future. In Wales Labour was expected to get a majority, but instead saw its vote collapse sharply, partly in protest at the party’s mishandling of the Welsh Labour leadership. This issue became enmeshed first in a fierce personal scandal over the previous Labour leader and then in a furore over Blair’s (in the end successful) efforts to impose his own nominee (Alun Michael) on the Welsh Labour Party instead of a rival contender backed by most Labour members in Wales (Rhodri Morgan). Because Michael was mistrusted by the other parties in Cardiff as a Blairite clone, Labour initially opted to form a minority government in the Welsh Assembly and to proceed in an ad hoc way without a coalition deal with anyone. After barely a year, however, Michael lost an Assembly vote of no-confidence and had to resign. His replacement was Morgan, the person whom 10 Downing Street had gone to such lengths to prevent becoming leader. After several months’ negotiations, in autumn 2000 Morgan set up a formal Labour–Liberal Democrat government in Wales with an Assembly majority. In the London Assembly elections Labour similarly fared poorly, and not only did not get its expected majority, but actually gained fewer votes than the Tories. Immediately after the election a Labour–Liberal Democrat coalition, supported more conditionally by the Green assembly members, steered the Assembly’s business. The 2000 London mayoral election operated at the same time as the assembly poll using a supplementary vote system where voters could mark a first and a second preference for mayor by placing an X in two columns. Labour in government picked up the system as suggested by consultants for fear that on the one hand plurality rule would produce a weak mayor commanding only minority support, while the rival alternative vote system might decrease the barriers to factional candidates standing from within Labour’s ranks. What Downing Street wanted to avoid was any win for Ken Livingstone, the left-wing former leader of the Greater London Council abolished in 1986 by Margaret Thatcher’s government. In the party’s internal selection process to produce the Labour candidate, Blair and the Labour hierarchy intervened heavily against Livingstone and in favour of the Labour Cabinet member Frank Dobson. As in Wales they created a complex electoral college, weighting very heavily the votes of London Labour MPs and the ultraloyalist candidates whom the national party allowed to contest assembly seats. The popular and very well-known Livingstone lost in the electoral college despite winning majorities of the votes from Labour’s London membership and the trade

The United Kingdom

301

unions. But he then stood as an independent in the mayoral election against his party, collecting the top share of 38 per cent of the first preference votes in an 11-candidate race (compared to Dobson’s 13 per cent) and going on to win 58 per cent of the second-round runoff vote. This almost exact repeat of Blair’s Wales fiasco a year later undoubtedly depressed Labour’s London support in the assembly election also. Both events demonstrated in the most publicly possible way the traumatic difficulties which the major party machines had in adapting to the changed realities of proportional and multi-preference electoral systems, where voters were no longer constrained into supporting Labour or Conservatives for fear that their votes would otherwise be wasted.

Explaining institutional modernization in Britain There are different possible ways of explaining the apparent volte-face in Britain away from complete reliance upon plurality rule and towards proportional systems and we sketch their outlines only: Why not earlier? Why now? Cross-national trends include ‘democratic reappraisal’ and policy learning trends, while UK-specific factors should take into account long-run dynamics, short-term dynamics and explanations undermining ‘British political exceptionalism’. There has been a widespread move in the aftermath of the Cold War for liberal democracies to re-examine electoral systems which have been fixed for long periods of time. First, the end of the Cold War era has allowed criticisms of previous democratic inadequacies to be more legitimately expressed in all the established democratic systems. Second, there has been enhanced cross-national policy learning, with decision-makers ready to look at the experience of other countries in search of solutions. The German AMS system has been influential internationally in creating the shift towards mixed systems. Third, within Europe there has been a specially intense level of policy learning, with the British government under pressure to introduce PR to avoid the strong partisan distortions which the previous plurality system introduced into the composition of the European Parliament, and with policy learning about PR in the form of STV in Ireland and Northern Ireland also having a diffuse impact. The short-termist approach to electoral reform focuses on the immediate causal factors which led the Blair government to implement its sweeping constitutional changes, and persuaded the otherwise conservative Parliamentary Labour Party and the unions to accept them (Linton and Southcott, 1998). To get back to power in the 1990s Labour needed Liberal Democrat supporters’ help, and narrowly failed to get it in the botched and last-minute efforts of Kinnock’s 1992 campaign. By contrast Blair engineered his victory with every available tool, achieving an unprecedented seats/votes ratio for Labour in 1997, accompanied by an unprecedented increase in the Liberal Democrats’ seats/votes ratio. Via tactical voting the two parties’ voters effectively operated as an integrated bloc of 61 per cent of the electorate. On the eve of his victory Blair still contemplated giving the Liberal Democrat leader a Cabinet post, an idea that was dropped only on election morning as the scale of Labour’s overall majority emerged. Instead he established a machinery

302

Patrick Dunleavy and Helen Margetts

of Lab–Lib cooperation, a joint Cabinet committee in which constitutional and later other issues were concerted. Subsequently this line of analysis argues that electoral reform initiatives beyond the Cook–Maclennan pact have responded most immediately to the ebbs and flows of Blair’s dominance over his party, and to Labour’s perceived need to maintain a broader coalition of support. With the party riding high in the opinion polls, 1998 was a bad year for the Jenkins proposals to progress against vested Labour interests. But the over-reaching of Blair’s centralizing ambitions were charted in huge voter disaffection and internal morale problems in the mid-term 1999 and 2000 elections. The onset of the next general election, and the collapse of Labour’s opinion poll lead over the Conservatives in autumn 2000, both created some renewed recognition among Labour activists and MPs of the fragility of their party’s grip on power, hence the survival of a scaled-down manifesto commitment to look again at electoral systems reform. But substantial further extension of electoral reform to encompass Westminster elections might still need to await a switch from current large Labour majorities to something closer to a hung Parliament in the 2005 or 2006 election. The electoral system changes undoubtedly form a key element in the enigma that ‘New Labour’ has become under Tony Blair, although this remains a much wider and hard to analyse process (Heffernan, 2000). The broader elements of this change include: a remorseless repositioning of Labour as a centre party on economic and fiscal issues; a determination to distance the parliamentary party and leadership from traditional (‘Old Labour’) limiting constituencies and core support mechanisms; a pragmatic effort to reappraise welfare policies and economic strategies in a microeconomic way and to adapt them to modern conditions; and a thorough-going effort to re-energize democratic pluralism. Labour’s constitutional agenda sometimes seems motivated by principled commitments which have nothing to do with traditional ‘class party’ ways of looking at things, and sometimes appear more as a way of allowing central government to disengage from issues it would like to export to subnational bodies in order to concentrate on economic and welfare ‘fundamentals’. Underlying the whole constitutional agenda there are two rival Blairite projects for the party system, between which the Prime Minister is famously unable to decide. One vision involves the progressive assimilation of the Liberal Democrats into a new centrist Labour party (possibly with the explicit loss of the old Labour left) – hence the importance of the proto-coalition with the Liberal Democrats. The short-term concessions involved are justified by the long-term gains in prospect if the relationship can be managed successfully. The other involves a less extensive drawing together of the two parties, aided by electoral reform for Westminster in the next Parliament. If either project is to work, New Labour will have to accomplish a transition which some analysts have effectively declared impossible for a social democratic party (Przeworski and Sprague, 1986). In the end, neither scenario for major change may happen, and Labour may settle for a sequence of short-term goals, beginning with just looking at the following election. But the changes already wrought in pursuit of the longer-term strategies, and the constitutional change already enacted to prepare the way for possible electoral reform at Westminster,

The United Kingdom

303

have both been extensive enough to rank as fundamental shifts in the trajectories of the British party system.

Conclusions: future prospects for Britain’s electoral systems The next few years are certain to demonstrate again that changes in voting systems can make important alterations in structures of political power and systematic ‘luck’. Three possible long-term outcomes can be identified. The first would involve either an outright Conservative Party victory at a future general election or a widespread disillusionment in Labour’s ranks sufficient to trigger a wholesale abandonment of the new electoral systems and a reversion to plurality rule inside Scotland, Wales and London, as well as for European elections. This prospect still seems overwhelmingly unlikely at the present time. For the Tories the AMS systems have given them back a substantial political basis in Scotland and Wales while their Westminster representation in both countries has been eliminated. For Labour the devolved elections delivered somewhat lower turnout levels in Scotland and markedly lower in Wales and London – but still a great deal higher than those experienced simultaneously under plurality rule for local council elections and for some Westminster by-elections. A second scenario would see the current status quo stabilized indefinitely, with proportional systems accepted by all parties for all the new bodies and for European elections, but without the established Westminster or local council systems being altered. In this case the Westminster elections might be expected to remain the ‘primary’ elections in terms of high political interest and turnouts. Some analysts have already suggested that all the devolved bodies in Scotland, Wales and London would then be treated by most people as ‘secondary’ elections, influencing less important issues, and hence a situation where if they vote at all they feel free to vote a little more erratically and wildly than for Westminster. In the long run the only impact of more fragmented voting patterns under PR systems would be to damage principally the Liberal Democrats, without fundamentally weakening the coalitional and governing grip of the two largest parties. A third scenario sees the period of coexistence between proportional and plurality rule elections as necessarily temporary, with the entrenchment of PR systems progressively eroding the political and ideological support for plurality rule. By winning seats under PR arrangements new political forces like the Greens and the anti-European UK Independence Party have already found it easier to consolidate a political presence and to erode the vote bases of the majors. Recent sophisticated studies of split-ticket voting in established AMS systems in Germany suggest that it is a complex phenomenon with strong implications for political change (Thurner and Pappi, 1998, 1999). And preliminary work in Britain points up its importance here too (Rallings and Thrasher, 1998). Some observers argue that in Scotland at least turnout and interest levels for future Westminster elections could still fall below those for the Scottish Parliament and Executive, which might become the more politically salient focus of attention for Scotland’s voters and separate media system, with lesser effects in Wales. New fronts for electoral reform energies will

304

Patrick Dunleavy and Helen Margetts

Table 16.2 Year

1945 1950 1951 1955 1959 1964 1966 1970 1974a 1974b 1979 1983 1987 1992 1997 2001

United Kingdom: House of Commons electoral results, 1945–2001 Labour v/s

Liberal v/s

Conservative v/s

Scottish & Welsh v/s

Irish v/s

Others v/s

48/62 46/51 49/47 46/44 44/41 44/51 48/58 43/45 37/47 39/50 37/42 28/32 31/35 34/42 43/64 42/63

9/2 9/1 3/1 3/1 6/1 11/1 9/1 8/1 19/2 18/3 14/1 25/3 22/3 18/3 17/7 19/8

40/32 44/48 48/52 50/55 49/58 43/48 42/51 46/53 38/46 36/43 44/53 42/61 42/58 42/51 31/25 33/25

– – – – – – – – – 4/1 2/2 2/2 2/2 3/2 3/2 3/2

– – – – – – – 1/1 2/2 2/2 2/2 3/2 2/2 2/2 3/3 2/2

3/1 1/0 0/0 1/0 1/0 2/0 1/0 2/0 2/1 1/1 1/0 0/0 1/0 1/0 3/0 1/0

Key: Scottish: Scottish National Party; Welsh: Plaid Cymru/Party of Wales; Irish: Ulster Unionist Party, Democratic Unionist Party, Social Democratic and Labour Party, Sinn Fein, Alliance Party of Northern Ireland; v/s: percentage of votes/percentage of seats.

also open up, as they did during 1999 with a strong campaign by some Labour councillors and the Liberal Democrats to democratize local government elections (see Dunleavy and Margetts, 1999c; Leach and Game, 1999), and with the Royal Commission on Reform of the House of Lords (2000) acknowledging PR elections for part of the upper chamber’s membership (see Dunleavy and Margetts, 1999b). The diffusion of elected regional governments in England outside London may occur, and if they happen such bodies are certain to be constituted via PR. In this perspective, which we still see as most plausible, action on changing the Westminster electoral system will occur much later than for other levels of elections. The long-term future still rests on the holding of a referendum in which the public could yet bind reluctant MPs to accept a radical change of Westminster elections. But with the referendum now decisively abandoned by Labour, the trajectory for possible reform lengthens appreciably. The method of electing the House of Commons was the ‘source code’ for Britishinfluenced plurality rule voting systems worldwide. If the genetic forerunner of almost all other plurality systems eventually is modified, the symbolic and policy influence will spread widely through electoral system debates.

References Dunleavy Patrick and Helen Margetts (1998a) Making Votes Count: Mixed Electoral Systems, Democratic Audit Paper No. 14. Colchester, Essex: University of Essex.

The United Kingdom

305

Dunleavy, Patrick and Helen Margetts (1998b) The Performance of the Commission’s Schemes for a New Electoral System: Report to the Independent Commission on the Voting System. London: LSE Public Policy Group and Birkbeck Public Policy Centre. Dunleavy, Patrick and Helen Margetts (1999a) ‘Mixed Electoral Systems in Britain and the Jenkins Commission on Electoral Reform’, British Journal of Politics and International Relations, 1, 1: 12–38. Dunleavy, Patrick and Helen Margetts (1999b) ‘Electing Members of the Lords (or Senate)’, included on the CD-Rom, Royal Commission on the Reform of the House of Lords (1999). London: LSE Public Policy Group. Dunleavy, Patrick, and Helen Margetts (1999c) Proportional Representation for Local Government: An Analysis. York: Joseph Rowntree Foundation. Dunleavy, Patrick, Simon Hix and Helen Margetts (1998) Counting on Europe: Proportional Representation and the June 1999 Elections to the European Parliament. London: LSE Public Policy Group. Dunleavy, Patrick, Helen Margetts, B. O’Duffy and Stuart Weir (1997) Making Votes Count: Replaying the 1990s General Elections Under Alternative Electoral Systems, Democratic Audit Paper No. 12. Colchester, Essex: University of Essex. Heffernan, R. (2000) New Labour and Thatcherism: Political Change in Britain. Basingstoke: Macmillan – now Palgrave Macmillan. Jenkins Commission (1998) The Report of the Independent Commission on the Voting System, Cmnd 4090–1. London: Stationary Office. Leach, S. and C. Game (1999) Hung Authorities, Elected Mayors and Cabinet Government: Political Behaviour under Proportional Representation. York: Joseph Rowntree Foundation. Linton, M. and M. Southcott (1998) Making Votes Count: The Case for Electoral Reform. London: Profile Books. Przeworski, Adam and John Sprague (1986) Paper Stones: A History of Electoral Socialism. Chicago: University of Chicago Press. Rae, Douglas, Victor Hanby and John Loosemore (1971) ‘Thresholds of Representation and Thresholds of Exclusion: an Analytic Note on Electoral Systems’, Comparative Political Studies, 3: 479–88. Rallings, C. and M. Thrasher (1998) ‘Split-ticket Voting at the 1997 British General and Local Elections: an Aggregate Analysis’, in British Elections and Parties Review. Royal Commission on the Reform of the House of Lords (2000) A House for the Future, Cmnd 4534. London: Stationery Office. Thurner, P. W. and F. U. Pappi (1998) ‘Measuring and Explaining Strategic Voting in the German Electoral System’, Arbeitbereich, II, 21. Mannheim: Mannheim Zentrum für Europäische Sozialforschung. Thurner, P. W. and F. U. Pappi (1999) ‘Causes and Effects of Coalition Preferences in a Mixed-Member Proportional System’, Arbeitbereich. Mannheim: Mannheim Zentrum für Europäische Sozialforschung.

Further reading • Dunleavy, Patrick and Helen Margetts (1995) ‘Understanding the Dynamics of Electoral Reform’, International Political Science Review, 16, 1.

• Dunleavy, Patrick and Helen Margetts with Stuart Weir (1998) The Politico’s Guide to Electoral Reform in Britain. London: Politicos.

• Farrell, David M. (2001) ‘The United Kingdom Comes of Age: the British Electoral Reform “Revolution” of the 1990s’, in Matthew S. Shugart and Martin P. Wattenberg (eds), MixedMember Electoral Systems: The Best of Both Worlds? New York and Oxford: Oxford University Press, pp. 521–41. • Farrell, David M. and Michael Gallagher (1999) ‘British Voters and Their Criteria for Evaluating Electoral Systems’, British Journal of Politics and International Relations, 1: 293–316.

306

Patrick Dunleavy and Helen Margetts

• Hart, Jeniffer (1992) Proportional Representation: Critics of the British Electoral System, 1820–1945. Oxford: Clarendon Press.

• Jenkins’ Commission (1998) The Report of the Independent Commission on the Voting System, Cmnd 4090-1. London: Stationery Office.

• Linton, M. and M. Southcott (1998) Making Votes Count: The Case for Electoral Reform. London: Profile Books.

Part 4 Eastern Europe

This page intentionally left blank

17 Eastern Europe: General Overview Carlos Flores Juberías

At the time when the democratic revolutions of 1989 started to gain momentum in Eastern Europe, all the so-called ‘popular democracies’ were using the typical Soviet system of majority rule with a second-round runoff in single-member districts. Though there were some differences – both in legal and in practical terms – among the several nations of the Soviet Bloc, and despite some degree of evolution throughout the different stages of the communist period, in general terms we could talk about a homogeneous electoral model – which was also the only one compatible with the Marxist idea of what elections ought to be in a Socialist society: if pluralism was not a goal to be achieved, nor a value to be guaranteed, then one single party and one predetermined candidate per seat was all that voters could realistically expect.

Electoral system design in times of regime change As democratization started to develop in the different countries of Eastern Europe, the reform of the existing electoral systems began to be seen as a crucial feature in the process of change and, moreover, as an essential step if democracy was to be effectively established. This was due not only to the fact that conditions limiting pluralism had to be done away with, but also because it was essential to replace the old majoritarian formulas with new electoral systems in which the nascent social pluralism could find a better accommodation. Throughout this imposing process of systemic transformation, it was possible to realize the existence of a close relationship between, on the one hand, the model of transition towards democracy followed by each of the countries we are analysing and, on the other hand, the electoral system used, at least in their first free, founding elections. This relationship could be summarized in the following three patterns: 1. In countries whose democratic transitions were characterized by a sudden and revolutionary breakaway with the past and which therefore held their first free elections under the supervision of the former oppositional organisations now in power thanks to their ability to disallow communist party officials from 309

310

Carlos Flores Juberías

office, a basically proportional representation electoral system was chosen. This was the case in countries such as the former (East) German Democratic Republic, the former Czechoslovak Federation and Romania. 2. In countries where the transition to democracy was carried out by means of negotiation, usually between the opposition forces and the more liberal sectors of the single Party – neither of whom were able to impose their points of view on their opponents – the option was a mixed electoral system, combining both proportional and majoritarian elements. In these countries – namely Hungary, Bulgaria and Slovenia – elections took place under Communist apparatus control, although the democratic opposition was strong enough to impose its point of view on a significant number of aspects and to somehow scrutinize the entire process. 3. In countries where opposition was practically non-existent or very weak and disorganized until the actual moment of the elections, the organisation of these fell undisputedly into the hands of the Communist Party apparatus and therefore they were held using the traditional majority system – which most Communist officials felt was best suited for them. In spite of this, electoral campaigning allowed the democratic opposition forces to engage in a process of rapid growth, and in some cases to win substantial parliamentary representation, which subsequently made it possible to overthrow the Communist Party and bring about a change of regime. In this third tier, the USSR and most of its Republics, Albania and also some of the countries of the former Yugoslavia should be included. Let us look at all these cases with some detail. For the June 1990 elections Czechoslovakia opted for a party list proportional system. As in East Germany, the main argument in favour was that proportional representation was best suited to produce a more realistic picture of the diversity of existing opinions, and that the choice of a party list system would probably help strengthen the party structures and popularize them among voters. Furthermore there was the country’s previous experience, since it was the system the short-lived Czechoslovak Republic had used in the interwar years. It was even the simplest option, an argument of vital importance if we take into consideration that elections to a bicameral Federal Assembly and two separate parliaments for the Czech and Slovak Republics had been simultaneously called. What’s more, even the Communist Party seemed to be in favour, since it was quite able to foresee its astounding defeat at the hands of the combined democratic parties. Nevertheless, the proportional system favoured by the Czechoslovak Electoral Law of 27 February 1990 was not exempt from correctives aimed at preventing an excessively fragmented legislature and at limiting the amount of power in the hands of party officials. In order to achieve the first of these goals, parties intending to run for election were required either to prove a 10,000-strong membership or to gather a similar number of signatures supporting its candidacies. Besides, the system introduced a significant 5 per cent legal threshold, to be calculated at the republican – i.e. not federal or district – level. Preferential voting was also introduced in an attempt to achieve the second of the

Eastern Europe: General Overview

311

two objectives, so voters could indicate up to four preferences, which would be taken into account only when at least 10 per cent of those voting for a particular list used this mechanism. The elections to the National Councils of the two Republics were held using an almost identical procedure, with the only noticeable difference being a lower 3 per cent threshold for the Slovak National Council. In Romania, the system applied in the May 1990 election was contained in a Decree issued on 14 March by interim President Iliescu, leader of the now ruling National Salvation Front (FSN). The Decree introduced proportional representation for both houses of parliament, on the basis on the pre-existing districts: for the Assembly proportionality was achieved applying the Hare formula in each constituency, and distributing the surplus votes at the national level according to the d’Hondt formula; in the case of the Senate, with no redistribution of surplus votes, proportionality played a part only at the district level. In both cases, legal thresholds were absent. The system also contained a clause granting parliamentary representation for each of the eleven officially-recognized ethnic minorities, providing a seat in the Assembly for each one even in the case that none of their candidates managed to be elected. With the PCR disbanded, the framing of the 14 March Decree resulted from negotiations between the now ruling FSN and the more radically democratic parties. For the FSN, proportional rule had the advantage of allowing its local candidates to benefit from the huge popularity of President Iliescu and Prime Minister Roman, while it also made more likely the dispersion of the oppositional vote among the numerous parties which had already begun to spring up. For the latter, the system seemed to offer a better guarantee of access to parliament than any other, bearing in mind that when elections were called a convincing FSN victory seemed unavoidable. The results would very soon demonstrate how right these calculations had been. As we can see, in Czechoslovakia and Romania, a rapid process of political change – not entirely devoid of obscure aspects in the latter case – allowed democratic forces to rapidly seize power, despite the fact that sectors of the communist nomenklatura were still keeping control of important political institutions. When elections were called, these democratic forces were able to apply the electoral system which appeared to be more inclusive and more conducive to the rapid consolidation of the nascent political pluralism, a pattern also followed in East Germany. On the other hand, in Hungary, Bulgaria and Slovenia democratization started with a series of internal disputes within the Communist Party, by means of which the more liberal sectors of the ruling elite managed to topple the old guard and gain control of the party apparatus in order to advance the most urgent political changes from within. Since the democratic opposition saw them as acceptable negotiating partners, their strategy became that of progressively bargaining larger and larger levels of freedom, finally producing a democratic transition via a much slower but less tension-ridden path. The Electoral Law which governed the parliamentary elections of March and April 1990 in Hungary was passed by the National Assembly in October 1989, thus becoming the first unquestionably democratic electoral law in Eastern Europe – and, to be sure, one of the most praised and thoroughly analysed by the

312

Carlos Flores Juberías

academic literature. Its drafting resulted from the painstaking negotiations which took place from mid-1989 at the National Round Table. The position of the ruling Communist Party, the MSzMP, favoured a majority system in single-member districts, with one-seventh of the seats elected by means of a national list. This proposal was not acceptable for the democratic opposition parties sitting at the Round Table, who generally favoured a proportional party list system. Finally, agreement took the shape of a mixed electoral system: 176 seats would be elected on the basis of single-member districts and majority rule with runoff when needed; 154 would be elected by proportional rule on the basis of party lists in each of the country’s counties; and finally, the residual votes obtained both by individual candidates and county lists from the same party would be transferred to the party’s national list, in order to have the remaining 58 seats distributed. Apart from this, the agreement also contained a series of correctives designed to avoid an excessive fragmentation of the National Assembly. Thus candidates in single-member districts had to collect at least 750 signatures of local residents; parties interested in presenting their lists in a county had to put up candidates in at least a quarter of the county’s SMDs; and in order to set up a national list parties were required to put up at least seven county lists. Finally, no party could get any seat from its county or national lists unless the overall vote for its county lists surpassed a 4 per cent legal threshold. In short all this led to an electoral system which was unquestionably democratic, but which clearly favoured a bipolarization which seemed to satisfy everyone. Similarly the Bulgarian Electoral Law of 3 April 1990 was the result of a long series of negotiations between representatives of the democratic opposition – gathered around the Union of Democratic Forces (SDS) – and the most liberal sectors of the still ruling Bulgarian Socialist Party (BSP), the successor of the old Communist Party. As in other cases, the opposition tried to postpone elections at least until November of 1990, thus gaining extra time to organize itself. At the same time they demanded a purely proportional system and a large enough legislature, which would enable them to achieve a sufficiently large representation. In contrast, former communists – well aware of the fact that their organizational advantage might melt away if elections were not held any time soon – favoured the immediate election of a smaller house by the existing majority system. In the end, a compromise was reached on the size of the Assembly and the electoral system, but not on the date of the elections which by imposition of the BSP took place in June 1990. Consequently half of the 400-seat Grand National Assembly would be elected on the basis of single-member districts and majority rule with runoff, and the other half would be chosen by a closed-list proportional system in 28 multimember districts. To secure parliamentary representation in the latter case, parties had to win at least 4 per cent of the vote, a significant feature which in the end would reduce the number of parties with a significant presence in the House to just four. Thus the Bulgarian system proved to be the least proportional of all those mentioned so far, a natural consequence of the fact that it had been drafted in a political context where almost all the cards were still in the hands of former communists.

Eastern Europe: General Overview

313

In the case of Slovenia, negotiations between the ruling party – already embarked on a process of deep internal debate – and the emerging oppositional movements prior to the April 1990 elections took place basically along the same lines already described for Hungary and Bulgaria, with the opposition pushing for proportional representation and the reformed communists sticking to the existing majority system in single-member districts. However, these elections presented a rather specific appearance, since they were to take place within the existing constitutional framework, which still featured a tricameral legislature, composed of a Chamber of Associated Labour, a Chamber of the Communes and a Socio-Political Chamber. This need to accommodate the basic features of a democratic electoral system within the framework of the titoist constitution led to the election of the Socio-Political Chamber – the most significant of the three – under a proportional system in 16 multi-member districts, the other two being chosen under majority and plurality systems. Therefore, the hybrid nature of the 1990 Electoral Law was doubled, as it combined different kinds of electoral systems and was built on principles and modelled on patterns pertaining both to the old and the new political structures. All in all, the April 1990 elections provided the democratic opposition, gathered in an umbrella coalition aptly named DEMOS, with a significant victory – which propelled them into a power-sharing situation with the simultaneously elected Slovenian president, reformed communist Milan Kucan. Finally, in those countries in which transition was slower or took longer to start, in which democratic opposition parties proved too weak to draw significant concessions from the ruling communist elite at the key moments, the traditional Soviet-style majority rule with runoff in single member districts was the most widely used system. As said above, this was the case for most of the elections which took place in the former Soviet Republics throughout 1990, but this was the case in a number of former Yugoslav Republics too, as well as in Albania. The latter will provide an excellent illustration of the point. When in March/April 1991, traditionally isolated Albania held its first multi-party elections since the arrival of communism in power, the huge democratizing processes which had already shaken political structures everywhere in Eastern Europe had provoked only superficial scratches in the Albanian Workers’ Party’s grip on power. Therefore the 1991 elections took place in a context of theoretical – but not real – political competition, where the supremacy of the ruling party on all grounds was key in explaining the final outcome. The applicable electoral rules were set by the 13 November 1990 Law which was little more than a reproduction of the 1966 Law – itself a classical example of communist electoral systems based on a majority vote with runoff in 250 single-member districts – only modified to allow the nomination of candidates from parties other than the Communist and its satellites: quite tellingly, the law required the compulsory presence of at least two candidates in each district. As expected, the system and its conditioning factors gave the Workers’ Party an easy victory, fundamentally sustained by a massive rural support. Nonetheless, the Democratic Party – the largest in the opposition – also got substantial support from the urban districts, constituting – with 38.7 per cent and 75 seats – a notable oppositional counterbalance, which somehow anticipated

314

Carlos Flores Juberías

bipartism and alternation, two of the most characteristic features of Albanian politics to day. Curiously, the exception to this rule came from Estonia via an intelligent compromise and an unusual electoral system. As Evald Mikkel and Vello Pettai explain in the relevant chapter, while anti-independence communist forces wanted to keep the standard Soviet single-member districts – more by habit than accurate perception of self-interest – the nationalist Popular Front proposed a simple quota and largest remainders with an average district magnitude of three, which implied the use of party lists. Since the communists knew they would do better as individual candidates than under their despised party label, a compromise was reached in order to introduce STV, which is effectively a non-list proportional representation system. In this particular case the debate was not so much between supporters of proportional representation and defenders of majority rule, but between supporters of a party list system – quite likely conducive to a multi-party system – and defenders of a non-party system based on individual and independent candidacies. Hence, the resulting compromise led to a highly proportional system, but one which restricted strongly the role of parties. None of the other former Soviet Republics, except Georgia, was to achieve such a compromise, so elections ended up being held according to the traditional Soviet majority system. A fairly different case which merits specific analysis – anticipating the one provided by Marek Kaminski in the following pages – was that of Poland. Here, compromise between the ruling communist party – the PZPR – and the fast-growing opposition forces led by Solidarity arrived at a rather early stage in the transitional process, when even minor concessions from the communist rulers were hard to obtain and looked like major, historic breakthroughs. This fact led to the drafting of a truly transitional Electoral Law which combined not just proportional and majoritarian features, but a set of electoral rules that were half way between those of the old communist system and those of the democratic world. Hence, the June 1989 Polish elections took place by means of an electoral system which was only partially competitive – or in which competition was ‘compartmentalized’ – and in a context in which party competition was severely constrained – a fact which, nevertheless, could not prevent electoral outcomes overtly different from those expected by the communist rulers. Following the Round Table Agreements signed between Solidarity and the communist-led government, the 7 April 1989 Electoral Law provided for the formation of a 100-member Senate, with two senators elected by majority rule in each voivodship (three in the case of Warsaw and Katowice), and a 460-seat lower house – or Sejm – to be elected on the basis of 108 multi-member constituencies, with a district magnitude ranging from 2 to 5. Despite this appearance of limited proportionality in the lower house, the system required candidates to run not for any but for a particular seat in the constituency, so the election was in fact carried out also by majority rule with runoff. However, the most defining feature of the 1989 Polish Electoral Law was the set of restrictions plaguing the candidate nomination process, which in the end supposed that only 161 of the 460 Sejm seats – one, or in some cases two, per district – would be freely contested by candidates not

Eastern Europe: General Overview

315

linked with the Communist Party, provided they managed to gather the 3,000 voters’ signatures required to actually be nominated. The remaining seats were reserved exclusively for candidates nominated by the PZPR and its satellite organizations, either running in the districts or forming part of the 35-member national list, made up of the most important figures in the aforementioned parties, who were to run for election without opposition. These carefully calculated proportions guaranteed the communists and their allies that, even in the worst-case scenario – the one that in fact happened after the June elections, when Solidarity-backed candidates won every disputed seat in both houses but one – they would enjoy a sufficient majority in the Sejm and thus control over the composition of the government, plus they could also outvote the non-communist deputies and senators in the election of the President, which was awarded to both houses voting together as a National Assembly. Only on a few important issues such as amending the Constitution or annulling the Senate veto, where a two-thirds majority in the Sejm was necessary, could the communist-led coalition theoretically be opposed by non-communist forces.

Electoral system change and democratic consolidation Throughout the last decade, the mutability of electoral systems has been the rule in Eastern Europe. Although the extent to which these systems have changed and the frequency of these changes has varied dramatically, the political debate on this topic has been a permanent feature in all of them and it is still unlikely to be over. Electoral codes have not been viewed as fixed rules that have to be respected, channelling political competition in a neutral fashion, but as merely provisional guidelines resulting from political bargaining and bound to be changed as soon the actors or the political scenario change. While this assessment may very well be applicable to many other countries, in no other region of the world it is possible to find such a widely acknowledged disregard for the benefits – in terms of institutional and political stability – of keeping the rules of the electoral game constant. While bargaining over the electoral system is certainly not a new phenomenon, doing so in so many countries and in such a short period of time represents quite an unmatched case – which, unfortunately, cannot be assessed but as a clear testimony that in these countries democratic consolidation is more a goal than a reality. Leaving aside the cases of the Czech and the Slovak Republics, Hungary and Poland – which are analysed in full detail by Petr Kopecký, John Schiemann, and Marek Kaminski and Monica Nalepa in their own country chapters – and reducing to a minimum the references to the Baltic countries and Russia – which are addressed by Evald Mikkel and Vello Pettai and Olga Shvetsova – our argument will be sustained on the remaining cases. Albania Strengthened by its relative electoral success at the March/April 1991 elections and by the favourable international environment, the Democratic Party (DP) forced the Socialist government to enter negotiations on a long list of legislative measures,

316

Carlos Flores Juberías

among them a new electoral law that would satisfy the most basic requirements of a democratic system. But during the successive debates, the strategies of both parties changed as radically as their immediate electoral expectations. The Democratic Party’s initial defence of proportional representation gave way to support for a majority system or, at least, for a mixed system that would provide substantial advantages to the major parties and prevent the access of smaller parties to parliament. At the same time, the Socialist Party quickly realized that the existing majority system was quite likely to damage its expectations in the case of a predictable opposition victory and therefore began advocating for the most proportional system available. Finally, Electoral Law No. 7556 was passed by Parliament on 4 February 1992, introducing an apparently mixed but in fact proportional system, not very different to the German model. The 140-member Popular Assembly would have 100 deputies directly elected in single-member districts by majority rule with runoff, while the remaining 40 seats would be allocated proportionally among candidates comprising the national lists. Each voter would cast a vote for the individual candidate of his/her choice, plus a second one for his/her preferred party list. Then, each party would receive as many seats as districts carried, plus a fraction of the 40 seats reserved for the national lists sufficient to adjust its actual presence in parliament to the percentage of votes effectively obtained. The Law also established ways to prevent the representation of minor parties, first by requiring that each individual candidate present 400 signatures from voters from his/her district in order to formalize his/her candidacy; second, by forbidding the submission of national lists by parties proposing candidates in less than 33 circumscriptions of 9 different districts; and lastly, by introducing a 4 per cent threshold for the allocation of national list seats. Once applied, in the March 1992 elections, the system worked just as expected, providing the winning Democratic Party a vast advantage in the single-member district seat allocations, but then levelling this advantage in the overall result with the distribution of the national list seats, and providing those who did not surpass the 4 per cent threshold at least a testimonial presence in parliament via the single-member districts. However, in one of the most self-explanatory examples of the winding path of electoral system reform in Eastern Europe, the 1992 Albanian electoral law was substantially reformed on 1 February 1996. Pushed by the still ruling Democratic Party with the obvious intention of maximizing its support in the immediate May/June election, the new Law basically enhanced its majoritarian features, raising thresholds and reducing by 15 the number of seats proportionally allotted from the national list as a compensatory mechanism. Though not a radical change, the 1996 reform enhanced the majoritarian features of the existing system. As a consequence, Albania moved from majority rule to an essentially proportional system, only to diminish its degree of proportionality by the selective introduction of a series of majoritarian devices in the rather brief period of five years. Bulgaria Giving continuity to its well-established tradition of electoral bickering – a practice that from 1879 to 1939 had given rise to no less than eight different electoral laws

Eastern Europe: General Overview

317

and 27 additional amendments – post-communist Bulgaria decided to engage in electoral system reform as soon as the new Constitution was adopted by the Grand National Assembly on 12 July 1991. Agreement on the basis of a new electoral system was easier to achieve then than it had been a year before; however, for the still ruling BSP, the majoritarian component of the 1990 system was no longer a desirable feature, as it was quite clear that the democratic opposition had a fair chance of being the most voted option, and for the SDS, a majoritarian system seemed too risky at a point in time when internal divisions were too frequent. Since proportionality seemed to be the most convenient option for all those concerned, the 22 August 1991 Electoral Law supressed the majoritarian half of the 1990 system and established that the election of the 240 National Assembly deputies – 160 less than in 1990 – would take place by proportional rule in 31 multi-member districts on the basis of closed party lists, translating party votes into seats at a national level and allocating them among the different constituencies – once parties under 4 per cent were left aside – using the d’Hondt formula. In the case of independent candidates the Law required an endorsement by 2,000 constituents – by only 500 in 1990 – but it relieved them of the 4 per cent requisite, considering them elected once they achieved the Hare quota. The key issue of the colour of the ballots, which in Bulgaria has traditionally been a major means of partisan identification, had to be addressed by a further reform of the Law on 12 September which provided more flexibility in the ballot design so that parties could be identified with their traditional colours despite the multiplication of new organizations. But perhaps the most contentious aspect of the 1991 reforms concerned none of these issues, but that of ethnic minority representation. Since the new Constitution prohibited the formation of political parties organized on ethnic, racial or religious lines, and the Electoral Law restricted campaigning only to political parties, the Turkish-based Movement for Rights and Freedoms (DPS) had to go through an endless succession of legal procedures in order to have its candidates duly registered. Although the Constitutional Court upheld the Movement’s position, its decision came only after DPS candidates were forced to run as independents in the 1991 election. Despite this clear move towards proportional rule, the fact is that the 1991 elections marked an additional step towards the bipolarization of the Bulgarian party system, which until the very recent victory of Simeon II’s National Movement in the 2001 election was characterized by the alternation between the SDS (winner of the 1991 and 1997 elections) and the BSP (winner in 1990 and 1994), with the DPS and a bunch of agrarian and centrist parties performing as rather secondary actors. The 4 per cent threshold, which in 1991 left 24.9 per cent of the voters unrepresented, was quite a conducive element in this purpose. As far as Romania is concerned, the system created by the 14 March 1990 Decree, which was applied at the first democratic elections in May that year, was foreseen as a merely provisional solution. The passing of the 1991 Constitution, which mandated new elections to be held within a year, reignited the debate over the electoral system. Though the ruling FSN had been advocating for the introduction

318

Carlos Flores Juberías

of several clauses aimed at reducing the high level of proportionality of the existing law in a clear attempt to benefit from the opposition’s division, internal dissent within the Front itself advised some precaution. So, when the new legislation was passed on 15 July 1992, some of the predicted changes had been significantly watered down. Nevertheless, the new Electoral Law introduced two sets of measures: one aimed at keeping minoritarian forces out of parliament (reducing the average district magnitude and introducing a 3 per cent threshold to be calculated at the national level); and a second aimed at discouraging coalition formation, establishing that electoral alliances would have to face a variable threshold that was to be 1 per cent higher for every new member up to a maximum of 8 per cent, and prohibiting parties running on their own in some constituencies to join a coalition in some others. Although the 3 per cent threshold could be considered relatively benign by international standards, it should not be forgotten that in the previous elections only three of the 18 parties that were able to win seats – the FSN, the Hungarian Democratic Union (UDMR) and the National Liberal Party (PNL) – had achieved voting percentages above that figure, and that given the high level of political instability in which the country seemed embroiled, the consequences of such a decision were truly unpredictable. In fact, when the new rules were put into operation in the September 1992 elections, their effects proved to be critical, since only eight parties were able to remain in the Senate, and just seven in the Assembly, with approximately 19 per cent of the voters becoming unrepresented in the latter and 15 per cent in the former. Following this precedent, additional restrictions on proportionality and coalition formation in the Romanian electoral system were introduced in July 2000, when an amendment to the Electoral Law raised the existing 3 per cent threshold up to 5 per cent for parties running on their own and to 8 per cent for two-member coalitions, keeping the clause requiring multi-member coalitions to obtain an extra 1 per cent of the vote for every additional party they comprised, with no limit this time. The most immediate consequence of this measure was very soon perceived as in the 26 November elections the still ruling but deeply discredited Democratic Convention of Romania 2000 was only able to gather 5.0 per cent of the vote – far from the threshold that, as a four-member coalition, was imposed on it – and the Alliance for Romania got a mere 4.1 per cent, both of them being forced to stay out of parliament. Slovenia In Slovenia, the adoption of the 23 December 1991 Constitution implied a complete transformation of the Slovene legislature, which in turn forced a substantial reform in the electoral rules applied for the 1990 elections. The new legislature was to be a rather peculiar one, with a popularly elected 90-member lower house (the National Assembly or Drzavni Zbor) and a sort of neocorporatist upper house (the National Council), allegedly shaped after the Bavarian Senate, intended to represent the various local, social, professional and economic interests of the citizens. Consequently enough, the electoral system created for the National Assembly by the December 1992 Electoral Code was also a truly odd one, since despite its majoritarian façade,

Eastern Europe: General Overview

319

it was designed to operate in a fairly proportional fashion. In a nutshell: the country was divided into eight constituencies, each one subsequently divided into 11 singlemember districts, the remaining two seats being reserved by constitutional mandate for the Hungarian and Italian minorities. Votes cast in favour of each individual candidate were added at the constituency level to those of other candidates belonging to the same party, seats were allocated to parties applying the Hare quota, and then they were awarded to candidates according to their share in the overall number of votes, with the surplus votes aggregated again at the national level in order to distribute the remaining mandates applying the d’Hondt formula among parties having obtained at least three seats. In this way, the system appeared to be a majoritarian and candidate-centred one, while in fact it operated in a mostly proportional and very party-centred fashion, as the 1992 and 1996 elections have clearly shown. But debate about electoral rules was far from coming to an end with the 1992 Code. Promoted by the Social Democratic Party, a referendum held in December 1996 let Slovenes express their preferences on the issue, with 44 per cent supporting a two-round majority system, 26 per cent an amended proportional system and 14 per cent a mixed system. However, subsequent legal and political disagreements – with the government considering the referendum invalid due to low turnout, the Constitutional Court supporting the binding nature of the referendum’s outcome and the National Assembly dodging the issue by means of a constitutional reform – led to the introduction of a modified proportional system, the most remarkable feature of which was a 4 per cent threshold, shortly before the October 2000 election. As far as the National Council was concerned, the Constitution and the Electoral Code established that four seats would be awarded to representatives of the employers; four to the employees; four to farmers, craftsmen and independent professions; six to representatives of non-economic activities (a category comprising universities, cultural institutions, educators and sportsmen, among others); and 22 to representatives of local interests – the election of which would depend on each electing body’s internal rules. Russian and Yugoslav Republics As mentioned above, between January and October 1990 all 15 former Soviet Republics held competitive – or, in some cases, just semi-competitive – elections for the renewal of their Republican legislatures, in almost every case using the traditional Soviet-style majority system with runoff. But with the collapse of the Soviet Union, these now independent republics also rushed to debate which electoral system would best suit the election of their newly created parliaments. The option of a mixed system was followed in Lithuania and, more remarkably, in the Russian Federation and in Ukraine, while different types of proportional electoral systems were introduced in Estonia, Latvia and Moldova. Therefore only Belarus among the European former Soviet Republics has remained attached to the previous majority system. In a similar fashion, the Republics of the former Yugoslavia also gained independence, embarking on constitution-making and a process

320

Carlos Flores Juberías

of electoral system reform, which in Croatia and Macedonia has followed similar paths, moving both countries from majority to proportional rule, with an intermediate stop at a mixed formula. In particular, the Ukrainian Verkhovna Rad, in striking similarity – mimetism, one might be led to believe – to the Russian Duma, is composed of 450 members, of which 225 are elected in single-member districts by a first-past-the post system, while the other 225 are distributed by proportional rule from blocked national lists belonging to parties getting at least 4 per cent of the national vote. However, the Ukrainian mixed electoral system was introduced at a comparatively later stage, in 1998. For the previous election, in April 1994, the 18 November 1993 Law had kept the existing majoritarian system, so all 450 deputies were elected in singlemember districts by majority rule with runoff. As expected – or maybe as intended – the 1993 system proved not to be very helpful for the consolidation of a stable party system or an efficient parliament in Ukraine: elections had to be repeated twice and even more times due to insufficient turnout, independent deputies ended up being a majority in the house, party fragmentation became a permanent feature and regional divisions did not take long to surface. Unfortunately, the new electoral system was only able to change some of these consequences, but not to put an end to the periodical allegations of vote fraud, and what is euphemistically known as ‘use of administrative resources’. Ten years after independence, free and fair elections in Ukraine are still dubious. In Moldova, the adoption of a new Electoral Code for the 1994 election took place in a rather bizarre fashion. The 14 October 1993 Law did away with the Soviet-era majority system used in the 1990 elections and introduced proportional representation via the d’Hondt formula in multi-member districts, with a 4 per cent threshold at the national level. However, the Law itself did not draw the electoral districts, and when Moldovan officials got to the task only a few days later, they were confronted with a complex dilemma: if districts were drawn on ethnic lines – which basically meant the creation of specific constituencies in the conflict-prone areas of the Transdniester and Gagauzia – the boycott announced by pro-independence forces in these territories could be presented as proof that the new parliament was not representative of these communities; but if electoral districts deliberately ignored ethnic realities, placing ethnic-Russian and Gagauze populations in ethnic Moldovan-dominated districts, the measure was bound to be seen as a downright provocation. In the end, the circle was squared by the creation of a single national district by an additional Act of Parliament, adopted on 19 October 1993. The measure was intended to dilute the impact of a boycott and to break any possible legal connection between specific parties or deputies and concrete regions of the country which might foster separatism. The Moldovan system thus turned out to be one of the most proportional and also one of the most ethnically blind of the whole of Eastern Europe. The 18 May 2000 amendments to the Electoral Code restricted the participation in elections to parties registered at least two years earlier but kept most of the previous traits of the system intact. Passed by means of a highly controversial referendum, and loaded with a full set of hardly democratic provisions, the 1996 Belarussian Constitution suppressed

Eastern Europe: General Overview

321

the existing unicameral Supreme Soviet and established a bicameral legislature, the National Assembly, comprising a House of Representatives (or Palata Predstaviteley) and a Council of the Republic (Soviet Respubliki). The former has 110 members elected in single-member constituencies for a four-year term, while the latter – dubbed the ‘House of Territorial Representation’ – has eight members elected by the local Soviets from each of the six regions of the Republic and the city of Minsk, and an additional eight directly appointed by the President of the Republic. Hence, Belarus is the only East European country whose lower house is still elected through the traditional Soviet-style majoritarian system. However, the issue of which electoral system Belarus does apply is virtually meaningless, since the National Assembly is no more than a rubber-stamp legislature playing a completely secondary role in the country’s political life which is totally dominated by its authoritarian President Alyaksandr Lukashenka, and parliamentary – as well as presidential – elections are regularly rigged. In fact, the 15 October 2000 elections were boycotted by all major opposition parties in protest over repeated harassment of their leaders and militants, over lack of transparency in the electoral process and over permanent government interference. As a consequence, pro-Lukashenka candidates were able to garner 108 seats in the 110-member House of Representatives, amid allegations of forged turnout figures (the law requires a 50 per cent turnout for elections to be valid) from national as well as international monitoring organizations, a fact which provides Belarus with yet another element of distinction among East European countries, since it is the most clearly authoritarian regime in the region. In Croatia, after the 21 December 1990 Constitution suppressing the tricameral legislature and creating a new bicameral Sabor, a new Electoral Code was needed. For the August 1992 election, the HDZ-dominated legislature provided for a mixed electoral system, in which 60 seats were to be proportionally distributed among party lists in an at-large national district with a 3 per cent threshold, and the remaining 60 seats were allocated through plurality vote to individual candidates running in single-member districts. The remaining four seats would be elected in special electoral units by citizens belonging to the different ethnic minorities, although the Serbs appeared entitled to be represented in the parliament proportionally to their participation in the overall population of the country – which at that time was around 11.5 per cent. Again, the choice for a mixed electoral system proved tremendously positive for HDZ, since the party was able to maximize its share of the vote more than any other force, getting 61 per cent of the seats with 43 per cent of the national vote. But this was far from being the last change in the Croatian electoral system, as a new reform altered the basic traits of the existing system on the eve of the 1995 elections, and yet another one moved Croatia into proportional rule for the postponed January 2000 elections. The 1995 reform maintained the mixed nature of the system, but the number of seats elected under proportional rule was increased from 60 to 80 while the number of seats elected through plurality rule was reduced from 60 to 28, the remaining 12 seats being set aside for the Croatian diaspora – a constituency traditionally faithful to HDZ – voting in a special unit. This move towards a more substantial degree of proportionality

322

Carlos Flores Juberías

was counterbalanced by the introduction of higher thresholds, which were set at 5 per cent for parties, 8 per cent for two-member coalitions and 11 per cent for larger ones. As far as ethnic minorities were concerned, the new system was much more restrictive than the previous one, since only seven seats – three for the Serbian community and four for all the rest – were set aside to be elected in special districts. Conversely, the 29 October 1999 Electoral Code abandoned the mixed formula and introduced an entirely proportional system with 140 deputies elected in multi-member constituencies, up to six members chosen by proportional representation representing Croatians residing abroad, and five members representing ethnic minorities – out of which the Serbs, in a further restriction to their previous privileges, were entitled to just one seat, their right to vote as common citizens notwithstanding. In Macedonia, electoral system reform has been following a rather similar path, although the leading force here was not the desire of the ruling party to maximize its electoral results but the need to find an inclusive enough electoral formula able to meet the demands of the opposition and the national minorities as well. The 1990 Electoral Code kept the existing majority system with runoff and applied it in the first two democratic elections in 1990 and 1994, amid heavy criticism due to its many technical flaws, its lack of guarantees, its gross malapportionment and its consequent inability to fairly represent the nascent Macedonian pluralism. In fact, the two major opposition parties, the right-wing VMRO-DPMNE and the Democratic Party, boycotted the second round of the 1994 election after having failed to obtain a single seat in the first round, despite polling 29 per cent of the vote altogether. The claim for some degree of proportionality in the electoral system led to a substantial reform in 1998, which introduced a mixed system in which 35 seats would be distributed proportionally using the d’Hondt formula among national lists of parties with a 5 per cent legal threshold, while the remaining 85 seats would be elected as usual in newly drawn single-member districts by a majority system with runoff. The system was applied in the subsequent 1998 elections which led to the victory of the right-wing opposition. However, the new system proved short-lived. As a part of the Ohrid Framework Peace Agreement which tried to put an end to the outbursts of ethnic violence which plagued the country during the first half of 2001, a new electoral code was drafted under the close scrutiny of the international community, very much aware of the need to stop ethnic conflict before it was too late. The 18 June 2001 Electoral Law moved Macedonia into the domain of proportional electoral systems, creating six multi-member districts, each electing 20 deputies. The rationale for the reform was multiple: on the one hand, the new Code put an end to the potential for conflict shown on previous occasions by second rounds; it ruled out a well established tradition of gerrymandering; it increased the guarantees for a fair electoral process; and, of course, it introduced a higher degree of proportionality at a point in time at which inclusiveness and power-sharing were much in need.

Eastern Europe: General Overview

323

Conclusions Is there a pattern of evolution in these seemingly endless processes of electoral system change which have for so many years plagued the political debate in postcommunist democracies? Contradictory as it may seem, after a detailed overview we can conclude that these changes have been oriented by a rather common goal: to find a convenient path, equally distant from the unfair, anti-partisan, majoritarian systems used in Soviet times and subsequently favoured by communist successor parties as from the negative consequences of extreme proportionality when it comes to forming a workable parliament and a stable government. As we have been able to see, those countries which held their first, founding elections under majority rule have since then moved towards more proportional formulas, either by enacting mixed systems in which the proportional and the majoritarian components are strictly balanced – like Lithuania (1992), Croatia (1992), Russia (1993) and Ukraine (1998) – or by introducing proportional correctives, like a national list, to an otherwise basically majoritarian system – like Albania (1992) or Macedonia (1998) – or directly moving towards proportional rule, like Moldova (1994), Latvia (1995), Poland (1991), Croatia (2002) or Macedonia (2002). Bulgaria, too, could well fit into this pattern, since its originally mixed electoral system was stripped of its majoritarian component and turned into a proportional system in time for the 1991 election, so the only exception to this rule would be Belarus’s persistent adherence to majority rule. In an opposite direction, most countries which in their first, founding elections tested entirely – in some cases even extremely – proportional formulas have moved for their subsequent elections to more moderate means of achieving proportionality, not abandoning the basic traits of their original systems but introducing in them a sometimes complex set of filters and thresholds aimed at enhancing governability. While still proportional, the electoral systems resulting from the reforms introduced in Poland (1993 and 2001), Romania (1992 and 2000), the former Czechoslovakia (1992) and its successor countries (in 1998 in Slovakia and in 2002 in the Czech Republic) do show a tendency towards moderating the effects of proportionality. Not surprisingly, the country which from the very beginning enjoyed the more eclectic electoral system – Hungary – has also been the less inclined to introduce radical changes in it: the mere increase of one percentage point in the legal threshold seems, amid this sea of changes, a mere drop in the ocean.

Further reading • Ágh, Attila (ed.) (1994) The Emergence of East Central European Parliaments: The First Steps. Budapest: HCDS.

• Colomer, Josep M. (2000) Strategic Transitions: Game Theory and Democratization. Baltimore, MD: Johns Hopkins University Press.

• Elster, Jon (1996) The Roundtable Talks and the Breakdown of Communism. Chicago: University of Chicago Press.

• Elster, Jon, Claus Offe and Ulrich K. Preuss (eds) (1998) Institutional Design in Post-Communist Societies: Re-building the Ship at Sea. New York and Cambridge: Cambridge University Press.

324

Carlos Flores Juberías

• Flores Juberías, Carlos (1995) ‘The Transformation of Electoral Systems in Eastern Europe •

• • • • • • • • • • • • •

and Its Political Consequences’, Journal of Constitutional Law in Eastern and Central Europe, 2, 1: 5–66. Flores Juberías, Carlos (2000) ‘Post-Communist Electoral Systems and National Minorities: a Dilemma in Five Paradigms’, in Jonathan P. Stein (ed.), The Politics of National Minority Participation in Post-Communist Europe: State-Building, Democracy, and Ethnic Mobilization. Armonk, NY: East-West Institute/M. E. Sharpe, pp. 31–64. Furtak, Robert K. (ed.) (1990) Elections in Socialist States. Hemel Hempstead: Harvester Wheatsheaf. Geddes, Barbara (1995) ‘A Comparative Perspective on the Leninist Legacy in Eastern Europe’, Comparative Political Studies, 28: 230–74. Hermet, Guy, Richard Rose and Alain Rouquie (eds) (1978) Elections Without Choice. London: Macmillan. Ishiyama, John (1997) ‘Transitional Electoral Systems in Post-Communist Eastern Europe’, Political Science Quarterly, 112, 1: 95–115. Lanchester, Fulco (ed.) (1995) La legislazione elettorale degli Stati dell’Europa centro-orientale. Milan: A. Giuffrè. Leng, Hermann Otto (1973) Die allgemeine Wahl im bolschewistischen Staat. Theorie, Praxis, Genesis: Meisenheim an Glan. Lijphart, Arend (1992) ‘Democratization and Constitutional Choices in Czechoslovakia, Hungary and Poland, 1989–91’, Journal of Theoretical Politics, 4, 2: 207–23. Longley, Lawrence D. and Drago Zajc (eds) (1998) The New Democratic Parliaments: The First Years. Appleton, WI: IPSA Research Committee of Legislative Specialists. Moser, Robert (1999) ‘Electoral Laws and the Number of Parties in Post-communist States’, World Politics, 51, 3: 359–84. Nix, Stephen B. (ed.) (1995) Election Law Compendium of Central and Eastern Europe. Kiev: IFES/ACEEEO. Nohlen, Dieter and Mirjana Kasapovic (1996) Wahlsysteme und Systemwechsel in Osteuropa. Opladen: Leske & Budrich. Olson, David M. (ed.) (1996) The New Parliaments of Central and Eastern Europe. London: Frank Cass. Shvetsova, Olga (1999) ‘A Survey of Post-communist Electoral Institutions: 1990–1998’, Electoral Studies, 18: 197–409.

Summary Table 4A

Eastern Europe: Assembly (lower or single chamber)

Country Law year Election years

Term

Districts Seats

No.

Magn.

1

Albania 1991: 1991

4

250

250

1992: 1992, 1997, 2001

4

140–155 100–115 40

100–115 1 nat.

1 40

Ballot

Rule/Formula

Single

Majority/2nd round runoff

Double: Single Closed list

Parallel: Majority/2nd round runoff Proportional – Hare

Belarus 1996: 1996, 2000

4

260

260

1

Single

Majority/2nd round runoff

Bosnia-Hercegovina 1996: 1996, 1998, 2000, 2002

2

42

8

*5

Closed list

Proportional – St-Laguë

273

1

273

Closed list

Proportional – d’Hondt

400 200 200

200 28

1 3–26

Double: Single Closed list

Parallel: Majority/2nd round runoff Proportional – d’Hondt

Bulgaria 1911–1931 // 1990: 1990

4

Threshold

4–2% nat.

4% nat.

325

(Continued)

Country Law year Election years

326

Summary Table 4A

Term

Districts Seats

No.

Magn.

1 nat. (31 lists)

240

Ballot

Rule/Formula

Threshold

Closed list

Proportional – d’Hondt

4% nat.

Single

Two tiers: Proportional – d’Hondt Plurality

11% nat.

1991: 1991, 1994, 1997, 2001

4

240

Croatia 1995: 1995

5

108 80 28

1 nat. 28

80 1

2000: 2000

4

151

11

11–14

Closed list

Proportional (2 tiers) – d’Hondt

5% nat.

Czechoslovakia 1920: 1920, 1925, 1929, 1935 //

6

281–300

12

6–24

Closed list

Proportional (2 tiers) – Hare/Droop

20,000 v./dis.

1990: 1990

4

150

12

5–18

Open list

Proportional – Droop

5% nat.

Czech Republic 1992: 1992, 1996, 1998

4

200

8

15–41

Open list

Proportional – Droop

5% nat.

2002: 2002

4

200

14

5–25

Open list

Proportional – d’Hondt

5% nat.

3

100–120

11

5–20

Open list

Proportional – d’Hondt

Estonia 1919: 1919, 1920, 1923, 1926, 1929, 1932 //

1989: 1990

4

101

46

1–5

Single transferable

Proportional – Hare

5% nat.

1992: 1992 1995, 1999, 2003

3 4

101

11–12

5–13

Open list

Proportional (3 tiers) – Hare/ d’Hondt

5% nat.

245 199–215 30–46

199–215 4–10

1 2–10

Double Single Closed list

Parallel Majority/2nd round runoff Proportional

386 210 176

21 176

4–58 1

Double: Closed list Single

Parallel: Proportional (2 tiers) – Droop/d’Hondt Majority/2nd round plurality 4–5% nat.

*20

Open list

Proportional – St-Lagüe

1

Single

Majority/2nd round runoff

14–28

Open list

Proportional – St-Lagüe

1

Single

Majority/2nd round runoff

Hungary 1922: 1922 1926, 1931, 1935 // 1990: 1990, 1994, 1998, 2002

4

4

Latvia 1922: 1922, 1925, 1928, 1931 //

3

100

5

1979: 1990

5

201

201

1992: 1993 1995, 1998, 2002

2 4

100

5

112

112

Lithuania 1919: 1919

2%

4–5% nat.

327

328

Summary Table 4A

(Continued)

Country Law year Election years

Term

1920: 1920, 1922, 1923, 1926 //

Districts Seats

No.

Magn.

78–85

6–7

*12

Ballot

Rule/Formula

Closed list

Proportional – Hare

Threshold

1978: 1990

5

141

141

1

Single

Majority/2nd round runoff

1992: 1992, 1996

4

141 71

71

1

Double: Single

70

1 nat.

70

Open list

Parallel: Majority/2nd round runoff Proportional – Hare

5% nat.

141 71 70

71 1 nat.

1 70

Double: Single Open list

Parallel: Plurality Proportional – Hare

5% nat.

120 85 35

85 1 nat.

1 35

Double: Single Closed list

Parallel: Majority/2nd round runoff Proportional – d’Hondt

5% nat.

380

380

1

Single

Majority/2nd round runoff

2000: 2000

Macedonia 1998: 1998, 2002

Moldova 1989: 1990 //

4

4

4

1993: 1994, 1998, 2001

4

Poland 1918: 1919, 1920, 1922 // 1928, 1930 // 1989: 1989

101–104

1 nat.

101–104

Closed list

Proportional – d’Hondt

372–380

52–64

4–16

Closed list

Proportional – d’Hondt

460

460

1

Single

Majority/2nd round runoff

4–6% nat.

1991: 1991

4

460

37

7–69

Open list

Proportional (2 tiers ) – Hare/m. St-Laguë

1993: 1993, 1997

4

460

52

3–69

Open list

Proportional (2 tiers) – d’Hondt

5% nat.

2001: 2001

4

460

41

7–19

Closed list

Proportional – m. St-Laguë

5% nat.

Romania 1990: 1990, 1992, 1996, 2000

4

328–346

42

4–39

Closed list

Proportional – Hare

0–3% nat.

524

1–15

Approval

Indirect Majority/3rd round plurality

1068

1068

1

Single

Direct Majority/2nd round runoff

450 225 225

225 1 nat.

1 225

Double: Single Closed list

Parallel: Plurality Proportional – Hare

Russia 1917: 1917 // 1990: 1990 1993: 1993, 1995, 1999

2 4

5% nat. 329

330

Summary Table 4A

(Continued)

Country Law year Election years

Term

Districts Seats

No.

Magn.

Ballot

Rule/Formula

Threshold

Slovakia 1992: 1992, 1994, 1998, 2002

4

51–150

1–4

5–150

Open list

Proportional – Droop

5% nat.

Slovenia 1990: 1990, 1992, 1996, 2000

4

80–90

8–14

1–11

Open list

Proportional – Hare

2.5–4% nat.

Ukraine 1990: 1990

5

450

450

1

Single

Majority/2nd round runoff

1993: 1994, 1998, 2002

4

450 225 225

225 1 nat.

1 225

Double: Single Closed list

Parallel: Majority/2nd round runoff Proportional – Hare

4% nat.

138

27–43

1–5

Closed list

Proportional – d’Hondt

5% dis.

Yugoslavia 1992: 1992, 1996, 2000

4

Source: See Appendix: Notes and Sources for Summary Tables at the end of the book.

Eastern Europe: General Overview Summary Table 4B Country Law year

331

Eastern Europe: Presidency Election years

Term

Rule

Belarus 1996:

1996, 2001

5+5

Majority/2nd round runoff

Bosnia-Hercegovina 1998:

1998, 2002

4+4

Plurality

Bulgaria 1991: 1996:

1992 1996, 2001

4+4 5+5

Majority/2nd round runoff

Croatia 1990:

1992, 1995, 2000

5+5

Majority/2nd round runoff

Estonia 1992:

1992

4

Majority/2nd round runoff in Assembly

Lithuania 1993:

1993, 1997, 2002

5+5

Majority/2nd round runoff

Macedonia 1994:

1994, 1999

5+5

Majority/2nd round runoff

Moldova 1991:

1991, 1996

4+4

Majority/2nd round runoff

Poland 1990:

1990, 1995, 2000

5+5

Majority/2nd round runoff

Romania 1990: 1992:

1990 1992, 1996, 2000

2 4

Majority/2nd round runoff

Russia 1991: 1993:

1991 1996, 2000

5+5 4+4

Majority/2nd round runoff

Slovenia 1991:

1992, 1997, 2002

5+5

Plurality

Ukraine 1994:

1994, 1999

5+5

Majority/2nd round runoff

Source: See Appendix: Notes and Sources for Summary Tables at the end of the book.

18 The Baltics: Independence with Divergent Electoral Systems Evald Mikkel and Vello Pettai

Although Estonia, Latvia and Lithuania share a common recent past as Soviet-era republics and as independent states before that, their electoral systems in the contemporary era differ significantly. These divergences span almost every electoral category, including assembly size, district magnitude, ballot structure, vote counting formulas and legal thresholds. Since the early 1990s, each country has travelled a separate path both in terms of the creation of electoral institutions as well as their reform later during the decade. They illustrate diverse patterns of conflicting political interests, limited knowledge about electoral systems, rushed institutional bargains, controversial compromises and post-hoc decision-making. As a result, the three states also represent an ideal laboratory to test for the effect of electoral institutions on parties, party systems and democratic governance (Pettai and Kreuzer, 1999). In this chapter, we will concentrate on the main tenets of the three Baltic electoral systems. But in our conclusion we will aim also to contribute to broader knowledge about institutional design by addressing these important after-effects.

Electoral system elements From a comparative perspective, the basic elements of the electoral systems in the three Baltic republics can be analysed as follows. Assembly size Latvia, Estonia and Lithuania have kept their respective assembly sizes at the same level of 100, 101 and 141 seats for the entire post-Soviet period. For the Estonian Riigikogu the size of 101 had already been introduced in 1990 when the last Supreme Soviet of the Estonian Socialist Soviet Republic was elected. Interestingly enough, this was on the recommendation of the Estonian-American political scientist and renowned expert on electoral systems, Rein Taagepera, who applied his cube root law regarding the optimal size of national assemblies. In Latvia’s Saeima, this number was derived from the size of the country’s parliament during the interwar era. Although the number is an even 100, Latvians have so far avoided having problems with a 50–50 split. The largest Baltic state, Lithuania, also has 332

The Baltics

333

the largest national parliament, 141 seats in the Seimas, although it is actually comprised of two separate electoral systems. This number was also derived from the size of the last Soviet-era parliament which served from 1990 to 1992. District magnitude Estonia has assigned 101 seats in either 11 or 12 districts depending on the election. The variation in district size has ranged between 5 and 13 seats, with an average magnitude of 9.2 and 8.4, respectively. Latvia elects 100 seats from five districts. The magnitude here has ranged since independence from 14 to 28 with an average of 20. Lithuania combines two parallel systems where 71 seats are allocated via 71 single-member districts and 70 seats by one nationwide district. This combination of two different magnitudes of 1 and 70 has been in use for all elections with no changes at all. Ballot Estonia has a complex categorical ballot structure because of its three tiers of vote aggregation. The first tier involving so-called personal mandates resembles an open list ballot as voters cast a single vote for one candidate. At the second tier a party’s vote in the district is amalgamated via its individual candidates. Finally, the third tier of national party lists is completely closed. Indeed, as will be seen below, a significant share of seats was actually distributed during the 1990s via the closed national list, leaving voters with no choice at all. Latvia has the most permissive ballot in the Baltic states for although voters can cast their vote for only one party list, they are allowed to mark preference votes within that ballot and for as many candidates on that ballot as they choose. These preference votes take the form of either a plus sign next to or a line drawn through the candidate’s name. The ranking of the candidates on the ballot paper is only a hint from the parties as to their most preferred candidates. It is the voter who makes the final choice, as any mandate that the party wins in the district goes to the candidate with the highest net score of ‘positive’ and ‘negative’ votes in that district. Lithuania’s dual system combines a standard party list vote in the nationwide multi-member district and an individual candidate vote within the single-member districts. Within the party list vote, the ballot is semi-ordinal, since Lithuanian voters, too, since 1996 have been given the option of casting preference votes within their chosen party list. The effect of this voting, however, has been circumscribed by three factors. First, party ranking is also used in determining a candidate’s final placement. Thus, in reality, the original ranking and final placement of candidates has rarely shifted more than 1–2 spots. Second, parties can actually avoid preferential voting entirely if at the time of their electoral registration they submit a request to that effect. Third, voters are allowed to express their preference only in relation to five candidates on their chosen ballot. Within the single-member district voting, Lithuania held its first two post-independence elections under majority rule with second-round runoff. In 2000, the country shifted to a single-round with simple plurality rule, marking a clear move towards a more categorical vote.

334

Evald Mikkel and Vello Pettai

Rules and formulas Estonia and Latvia use different forms of proportional representation while the two parts of Lithuania’s mixed parallel system can be classified as semi-proportional for the multi-member district and plurality for the single-member districts. In Estonia, deviation from proportionality between votes and seats has averaged 13.3 per cent because of the complicated multi-tiered nature of seat allocation. Not only is this figure slightly higher than in Latvia, but it also masks a number of other inequities, which have in turn become precisely the object of electoral system reform. For 2003, the Estonian system was as follows. In the first tier of seat allocation, so-called personal mandates are awarded to those candidates who have surpassed the Hare quota for the district. In the second tier, additional mandates are allocated by adding up all of the votes cast for members of a single party in the district and seeing if the party as a whole has surpassed the Hare quota. If it has, then the highest vote-getter in that party receives a mandate. Moreover, among these fullquota parties, additional mandates are then awarded for each 0.75 additional Hare quotas the party has received. For example, if three candidates received 0.8, 0.7 and 0.3 Hare quotas respectively, the party would receive two district mandates for the first two candidates, since in total the party had received 1.8 quotas. This procedure is restricted by two provisos: first, the party itself must have received at least 5 per cent of the national vote; and second, the candidate slated to receive the mandate must have received at least 10 per cent of the Hare quota. Lastly, at the third tier, adjustment seats are awarded based on the nationwide total of votes cast for each party via its individual candidates. For this, a modified d’Hondt formula (with divisors raised to the power of 0.9) is used, which slightly lowers the level of proportionality. Mandates are awarded to individual candidates based on separate party lists, which means that at this level winners are selected primarily according to predetermined party rankings. The only restriction here is that in order to win such a seat the candidate must have received at least 5 per cent of the Hare quota for his district. If he/she has not, then the next candidate on the national list receives the seat. Latvia’s highest-averages method based on the Sainte-Laguë formula provides for the highest proportionality. During its four parliamentary elections from 1993 to 2002, the deviation between votes and seats has averaged 12.6 per cent. Lithuania, in a mixed parallel system, combines the Hare quota with largestremainders for the party list voting, with a plurality system for single-member districts. The modification to the Hare quota involves the fact that the denominator in the equation is the number of valid votes cast only for those parties exceeding the legal threshold and not the total number of valid votes. As a result, the division of seats via the multi-member district is less proportional. From 1992 to 2000, the average deviation from proportionality in the party list voting was 18.0 per cent. Legal threshold One of the few things that is similar for all three Baltic states is the electoral threshold – 5 per cent. Still, this uniformity emerged only over time. Estonia had

The Baltics

335

already settled on this threshold during its first post-independence elections in 1992. It has been modified only through the introduction of the individual subthresholds mentioned above regarding the percentage of a Hare quota a candidate must have received in order to be eligible for an actual mandate. In Latvia, the first post-independence elections in 1993 featured a 4 per cent threshold. Later in 1995, the Saeima switched to a 5 per cent threshold, and this has remained so through 2002. In Lithuania, voting for the multi-member district was subject in 1992 to a dual framework involving a 4 per cent threshold for standard parties and a special 2 per cent threshold for ethnic minority parties. In 1996, this was scrapped in favour of a 5 per cent threshold for all parties; in addition, a special 7 per cent threshold was introduced for electoral coalitions (Krupavicius, 1998). Lastly, Lithuania is the only Baltic country with a veiled threshold in terms of turnout. The multi-member district voting must include at least a 25 per cent turnout, while single-member district voting requires 40 per cent. The latter caused repeated problems during the mid-1990s, when the second round of singlemember district voting often did not muster the required turnout and elections had to be re-held. Eligibility The issue of eligibility plays an important role in the electoral process and its outcomes in the Baltic states, in particular in the cases of Latvia and Estonia. If during the Soviet period all permanent residents (and even temporarily resident Soviet Army conscripts) were allowed to vote in Baltic elections, then following the restoration of independence citizenship in Estonia and Latvia was limited only to those who had been citizens before the Soviet takeover in 1940 and their descendants. Although in both Estonia and Latvia this move ended up disenfranchising up to a third of the population (since this had been the magnitude of postwar immigrants from other parts of the Soviet Union), the decision rested on the principle that because the Baltic states had been illegally occupied by the Soviet Union, power had to go back directly to those from whom it had been taken. The consequence was a significant narrowing of the eligible electorate between the last Soviet-era and the first post-Soviet elections: from 1,164,495 to 661,074 in Estonia and from 1,960,639 to 1,243,956 in Latvia (Central Electoral Commission of Latvia, 2002; Eesti Vabariigi Valimiskomisjon, 1992; ENSV Valimiskomisjoni sekretariaat, 1990). Over time these numbers have risen due to the naturalization of Soviet-era immigrants, but nevertheless sizeable non-citizen populations remain in both countries – 20.0 per cent in Estonia and 23.7 per cent in Latvia (Statistical Office of Estonia, 2001; Pabriks, 2002). The situation in Estonia is eased by the fact that permanent residents are allowed to vote in local elections. A second important restriction concerns language requirements. Again, in both Estonia and Latvia, rules were in place during the late-1990s which required that candidates for elected office must have advanced proficiency in the state language. This was viewed as discriminatory by the large Russian minorities in these states;

336

Evald Mikkel and Vello Pettai

the rules were also criticized by international organizations as a violation of international human rights conventions (OSCE/ODIHR, 1998, 1999). In the event, the rules were never extensively enforced and eventually in late 2001 they were abolished in Estonia. Latvia followed suit in mid-2002. Thirdly, all three Baltic states had special electoral restrictions associated with post-communist lustration. In Estonia, the lustration rules were relatively mild, requiring only that candidates sign a statement that they had not actively persecuted others during the Soviet era. While in theory this statement could be contested in court and a candidate or even deputy could be disqualified, no such cases were ever brought. Moreover, in 2000 this requirement lapsed. In Latvia, an electoral ban was placed on all those who had belonged to Soviet-era security services as well as those who after 13 January 1991 were affiliated with the Communist Party or other anti-Latvian organizations. This had a notable impact on one party, the social-democrats, whose major leader, Juris Bojars, was never allowed to run for office since he had been an officer in the KGB. In the late 1990s this also applied to the former communist leader Alfreds Rubiks, who spent six years in prison for having collaborated with the instigators of the anti-Gorbachev coup in Moscow in August 1991, but upon his release he resumed a prominent role in the renamed Socialist Party Lastly, it is worth noting that in Estonia eligibility is also limited to lists organized by single political parties. That is to say, if in 1992 and 1995 parties could form loose electoral coalitions and field joint electoral lists in order to surpass the electoral threshold more easily, in 1998 these were banned in an effort to force parties to merge and to consolidate the party system. In the event, the strategy did not prove immediately successful, since in 1999 a number of coalitions continued to exist; they simply fielded their candidates under the name of the strongest party. Yet by 2003, the result was that only single parties fielded lists. In Latvia and Lithuania, meanwhile, coalitions continued to be permitted.

Reform strategies Given the fact that little more than a decade has passed since the Baltic states regained their independence, one could argue that the issue of electoral system reform actually extends all the way back to when the three nations began dismantling the bogus electoral systems of the Soviet era. It was here that the first understandings of electoral systems began to be formed, and in many cases these experiences would have an important effect on future electoral system choices. First, therefore, we will begin by looking briefly at the politics behind the genesis of democratic elections in the Baltic states before going on to examine the patterns of post-independence electoral reform. Pre-independence reform strategies As in any context of democratic transition and political reform, there are two main factors which can affect institutional choice. The first is lessons learned from past historical experience and the second is more narrow competition

The Baltics

337

between power-holders and opposition forces over how to engineer the system to their respective advantage. In the Baltic states, both of these factors were at work, although the second was by far the stronger. Compared to many other democratizing states, the Baltic states did not have an overarching legacy of past historical experience determining their institutional choices. This was mostly because of the length of the Soviet period (fifty years) and the total destruction of all previous institutional frameworks. At most, the lessons came from the Baltic states’ twenty-year period of independence between the two world wars, which in each state ended with the eventual collapse of democratic government. Therefore one of the lessons to be learned involved the dangers of an excessively permissive electoral system. From the immediate period of Soviet rule, the main lesson was to avoid too high thresholds of electoral participation. According to the Soviets’ majority electoral rules, a winner had to receive at least 50 per cent of the vote, while a turnout of at least 50 per cent was required. While this was not a problem during the communist era (when both figures always surpassed 95 per cent), under democratic rules and multiple candidates this stipulation would turn out to be very tedious. Elections in some districts would be declared invalid and new elections would be mandated. At the same time, in most cases turnout would only drop and in the end no one might be elected. Prior to independence, there were semi-free elections in the Baltic states, including the USSR Congress of People’s Deputies in March 1989, municipal elections in the autumn of 1989, and grassroots-based elections for special citizens’ congresses in Estonia and Latvia held in the spring of 1990. While all of these provided Baltic politicians and voters with electoral experience, the Supreme Soviet elections were viewed as the most important for democratic transition. Still, the Soviet legacy of single-member districts favoured the ruling elites since they could then campaign as individuals and project an image as reform communists. By contrast, a list-based system would have forced them to run as a party and be tarnished by their communist name. It was for this reason that this overall system was retained during the first fully free elections in the Baltic states, i.e. the re-election of the republican Supreme Soviets in early 1990. In Estonia, a change was made in terms of electoral formula, and instead the relatively rare single transferable vote system was used. This was in part because of the influence of Rein Taagepera, who had already begun writing about electoral systems in Estonian newspapers in 1988. (Moreover, his ground-breaking book with Matthew Shugart, Seats and Votes, was translated into Estonian the same year it came out, 1989.) Taagepera had recommended open list proportional representation with simple quota and largest remainders ‘. . . as the simplest method that struck some balance between party representation and voters’ inputs regarding particular candidates’ (Grofman, Mikkel and Taagepera, 1999: 235–6). The reform communists, however, opposed any form of list elections which would betray their highly unpopular party affiliation. Instead, they called for the single non-transferable vote without any party labels at all. A compromise was finally reached based on the STV proposed by the Popular Front. Thus in Estonia it is interesting to note

338

Evald Mikkel and Vello Pettai

that there emerged early on a number of people fairly knowledgeable about electoral systems. At the same time, this would soon create a situation where ‘too many cooks spoiled the broth’, since eventually the post-independence system would become a jumble of many different methods. In both Latvia and Lithuania, the single-member districts were maintained along with the minimum electoral participation requirements. True to the drawbacks mentioned above, repeat elections were later required for several districts until both parliaments were fully elected. Still, the system worked by and large. Following independence in August 1991, the three Baltic states first had to address the issue of constitutional frameworks. In Estonia, amid the tension and uncertainty of the attempted Soviet coup, the country’s politicians were able to agree on the creation of a special Constitutional Assembly which would undertake the writing of a new basic law. These efforts resulted in a parliamentary system being approved in a referendum in June 1992. In Latvia, this question was resolved almost immediately, as the country decided to reinstate its last democratic constitution from 1922. Although this system also restored some of the weaknesses of parliamentarism that the country had suffered during the first period of independence, it was nevertheless a reasonable compromise, since at the time Latvian politicians were too divided between nationalist radicals, moderates and communists in order to embark on a full-scale constitution-making process. Finally, in Lithuania the choice in favour of writing a new constitution was delayed until May 1992 while the nationalist leader Vytautas Landsbergis first attempted to amend the existing constitution in favour of a presidential republic. When this effort failed in a referendum, a special expert committee was formed which would go on to draft the semi-presidential system which exists currently. A similar pattern soon emerged with respect to electoral law. In Estonia another test of forces emerged between ex-communists who wanted to retain personalitybased voting, and nationalists who favoured party-list voting. This time, however, an even more distended outcome would result, since the new system would attempt to literally mesh both preferences by having candidate-based voting at the district level, but party-based vote aggregation at the nationwide level. Moreover, this was done without the kind of double vote system in Germany or even the mixed variant in Lithuania. As Grofman, Mikkel and Taagepera have described the result, ‘all of its components were borrowed or adapted from the West, but their combination was Byzantine’ (1999: 235). This is also probably why popular dissatisfaction with the electoral system remains the highest in Estonia. Latvia once again reverted to its 1922 electoral system for simplicity’s sake (although it raised the old threshold of 2 per cent to 4 per cent). Moreover, political jockeying eventually delayed the holding of founding parliamentary elections until June 1993. In Lithuania, the decision to go with a fully mixed electoral system was also reflective of a desire to have the best of both worlds. Lithuania’s parties at this time were unquestionably the strongest of any in the Baltic states. They were led by the continued strength of the ex-communist Lithuanian Democratic-Labour

The Baltics

339

Party and its chairman Algirdas Brazauskas. But likewise the nationalist Sajudis movement represented a strong player in the party system, along with smaller parties such as the christian democrats and the social democrats. These forces were therefore amenable both to proportional representation and to majority voting systems. The mixed system was seen as an easy compromise (Lukošaitis, 1998; Girnius, 1992; Gelazis, 2001). Post-independence reform strategies After the first post-independence parliamentary elections were held during 1992 and 1993, initial lessons could be drawn about each country’s choice of electoral institutions. First, in Estonia there were attempts to modify and improve the existing systems as a response to complaints about the system’s performance. Second, in Lithuania there were attempts to more directly manipulate the electoral system in order to improve a party’s expected electoral performance. In Latvia, no substantive changes to the electoral system were made during the decade following independence. In Estonia, the politicians’ hodge-podge of an electoral system generated its first surprises on election night in 1992. As the returns came in, it suddenly became apparent that some candidates who had amassed as many as 3,000 votes at the district level would not get into parliament because they fell short of a Hare quota in their district and were ranked too low to win via the national list. At the same time, a number of candidates sailed into the legislature with less than 100 votes thanks to their high ranking on the national list. Put in more institutional terms, the system had created a potentially conflicting logic between voters’ preferences at the district level and the more powerful role of party list rankings at the national level. During each election this disjunction could easily be measured by comparing the biggest ‘losing candidate’ with the smallest ‘winning candidate’. This difference was key to causing popular consternation about – and in some cases alienation from – the Estonian electoral system. Still, the politicians were loath to radically alter the system. They argued that if more voters began concentrating their votes on a single candidate within their district, more seats would be allocated at the district level and thus people would feel that the right candidates got into parliament. To their credit, this trend did begin to materialize over time. At the same time, the parties did little themselves to contribute to this trend. Rather, between 1992 and 1999 they did the exact opposite by continuing to field more and more candidates, thus diluting the voting pool even further. For example, the effective number of candidates within any district party list (meaning the degree of dispersion of votes among candidates in a district list) rose from 2 in 1992 to 3.75 in 1999. This would only be changed in 2003, when a limit of 125 candidates per party would be introduced and the respective dispersion measure fell to 3. The politicians’ only substantive measure of reform involved minor changes intended to soften the harsh effects of the system but not to change it fundamentally. Thus in 1995 an amendment was put into effect, which required candidates to

340

Evald Mikkel and Vello Pettai

have received at least 10 per cent of the Hare quota of their district in order to be eligible for a mandate at the district level. In 2003, this was followed up by a similar Hare quota proviso of 5 per cent for any candidate slated to receive a seat from a national list. For previous elections, the latter proviso especially would have altered the allocation of up to 12 out of 101 seats (Ideon, 2002). Nevertheless, even under these rules, some MPs would have still got in with less than 300 votes. To alleviate this problem further, another change was made for 2003 which would lower the number of votes a party in its entirety would have to receive at the district level in order to qualify for a second district-level mandate. As described above, parties that had exceeded the full Hare quota for their district through the combined votes of their candidates would receive their next district mandate with just 0.75 Hare quotas. This, too, caused many more mandates to be determined at the district level, thus bringing the electoral outcome closer to the voters’ preferences. While in 1992 only 41 out of 101 seats were allocated at the district level, in 2003 this rose to 74. Ultimately, Estonian politicians were reluctant to do away with the national lists entirely because they claimed that such lists would allow them to promote more young people in politics (who might not otherwise get into parliament because of their inexperience). Likewise, the parties wanted to be able to keep certain specialists within their ranks, who knew, say, budgetary affairs very well, but who were not very well known to the public. At the same time, it was clear that the national lists also allowed the parties to retain a certain category of safe seats which they could then use to protect certain politicians from popular wrath. This dilemma between democracy and political interest has yet to be solved. A second type of electoral reform in the Baltic states has involved institutional manipulation aimed at improving prospective electoral performance. As mentioned above, Lithuania in 1996 raised its electoral threshold to 5 per cent, and in addition introduced a special 7 per cent threshold for electoral coalitions. The official reason for this change was to help consolidate the kaleidoscopic Lithuanian party system and to limit the advantages of (and hence the temptations to form) broad electoral coalitions in comparison to single-party lists. Yet according to some political observers, another reason behind the move involved the ruling Democratic Labour Party’s desire to attract votes from one of its expected rivals in the upcoming elections, the Lithuanian Social-Democratic Party (LSDP). Since the LSDP was hovering at the time around 4 per cent in the polls, the hope appeared to be that the new threshold would prompt voters to act strategically and abandon the LSDP in favour of the LDLP. In the event, the LDLP still lost the election by a wide margin and the LSDP topped the new threshold. Four years later, suspicions were raised again when the new ruling party, the Homeland Union-Lithuanian Conservatives (HU-LC), moved to do away with the two rounds of voting for single-mandate districts previously used in 1992 and 1996, and to switch to a single-round plurality system. Again, the intention seemed to be that this would help stem the losses the ruling party was expecting to suffer as a result of declining popularity over its term in office. In particular, the calculation was that the HU-LC would be able to retain more seats as a plurality

The Baltics

341

winner in many districts rather than as a finalist in a second-round contest where other parties would probably gang up on it. Yet, as Clark and Prekevicius (2001) have shown, the HU-LC and its Christian Democratic allies achieved a net gain of zero seats as a result of their reform, while in fact it was the opposition Social Democratic Coalition which won the most – a full nine extra seats.

Consequences The experiences of the Baltic states both in terms of creating as well as reforming electoral systems show that politicians do not always achieve their expected outcomes, and that in fact the result is often quite the opposite. Where politicians have tried to over-engineer the system (as in Estonia), the outcome has been growing disenchantment with elections and a decline in voter turnout. Where politicians have tried to tinker with the system for electoral gain (as in Lithuania), here too the opinion of voters has proved stronger. To be sure, from the standpoint of a number of important indicators of electoral and party system consolidation, the Baltics have improved over the years. Yet in many cases there has been regression, for instance in electoral turnout as well as deviation from proportionality. In the Estonian case, it is difficult to say whether this morass was anticipated by the original designers. Grofman, Mikkel and Taagepera (1999: 240) speculate that ‘. . . even if the importance for electoral outcomes of the national closed lists was correctly anticipated, it appears unlikely that any Estonian politician anticipated the potential delegitimizing effects of mixing open list and closed list components.’ At the same time, once popular protest had died down after the first 1992 election, there followed little discussion of more fundamental reform. Most voters seemed to reconcile themselves to the situation, and the main parties certainly had no direct interest in endangering the rules by providing safe seats for their leaders. Moreover, one can say that there is a gradual process of cartelization going on, where the dominant parties have introduced new means of limiting political and electoral competition. Other elements of the Estonian electoral and party system contributing to it include a relatively high deposit fee for electoral candidates (two minimal wages, approximately 250 euros), the lack of any limits on political advertising and campaign expenditures, a high minimum-member requirement for new political parties (1,000 citizens), state subsidies totalling over 1.2 million euro only for political parties represented in parliament, a ban on electoral coalitions at the national level and lastly the skewed ethnic composition of the citizenry. In Lithuania, the challenge was to avoid new manipulations of the electoral system in advance of the 2004 parliamentary elections. With only two elections conducted under the original system and one election under single-round plurality, there was no overriding trend in one direction or the other. The country may yet see another reform, especially since the presidential electoral system continues to have two rounds. At the same time, such a change would only affect 71 out of 141 seats; at least the multi-member national district by proportional representation appears settled.

342

Evald Mikkel and Vello Pettai

Finally, even in Latvia, where no significant electoral changes have been made, the country continues to suffer from one of the highest effective number of parliamentary parties and the shortest cabinet duration of any of the three Baltic states. These problems may also prompt electoral reform at some stage.

Conclusions Among analysts of post-communist politics, there appears to be no common agreement concerning the optimality of repeated institutional and electoral change in these societies. One approach points to the advantages of flexibility gained via continuous engineering. According to Zielonka (1994), it would seem even desirable for rules to be adaptable amid the multiple social, economic and political transitions affecting these nations. By contrast, a second school, characterized by Taagepera (1997), counsels against change in institutions, since more often than not such reforms are driven by personal interest rather than by public good. Moreover, such changes provide precedents and temptations for future successors to alter the rules again and again, thus making them even more complex and confusing for all players involved. Still, both approaches agree on the need, for the process of democratization to be successful, to introduce new rules as soon as possible. They also both recommend adopting relatively simple rules in order to avoid possible misunderstandings and endless conflicts over divergent interpretations. In the Baltic states one could well find support for both of these schools. In Latvia, where the system has been stable, popular satisfaction (at least in terms of electoral turnout) has likewise remained high. In Estonia and Lithuania, reform has been more endemic and turnout also more erratic. Although in 1992 this flexibility was to be welcomed as part of a determination to get away from the failed past, after the mid-1990s electoral reform began to show a more cynical edge. Which school will ultimately be vindicated will take a number of more elections to confirm.

Table 18.1

Electoral performance in the Baltic states, 1992–2003 Estonia

Turnout (%) Wasted votes (%) Deviation from proportionality (%) Effective number of parties (votes) Effective number of parties (seats) Average cabinet duration

Lithuania*

Latvia

1992

2003

1993

2002

1992

2000

67.8 14.7 17.7

58.2 5.0 6.5

89.9 10.6 10.6

74.0 16.1 15.9

75.2 14.1 12.3

58.6 23.4 15.7

8.9

5.4

6.2

6.8

3.8

5.6

5.9

4.6

5.1

5.0

2.9

4.2

15.7 months

11.6 months

14.7 months

* For wasted votes, deviation and parties in votes, only data from the multi-member district are used. Source: Authors’ calculations.

Table 18.2 Year

Estonia: State Council (Riigikogu) electoral results, 1992–2003 Left

ESDTP EÜRP v/s v/s 1992 1995 1999 2003

2/0 2/0 – 0/0

– 6/6 6/6 2/0

Centre-Left KE v/s

RL v/s

12/15 – 14/16 – 23/28 7/7 25/28 13/13

Centre

M v/s

R v/s

EE v/s

KMÜ v/s

10/12 6/6 15/17 7/6

3/1 – – –

2/1 – – –

14/17 32/41 8/7 –

Centre-Right RE v/s

RP v/s

– – 16/19 – 16/18 – 18/19 25/28

EKRP v/s – – 2/0 1/0

Right

Others

EK ERSP v/s v/s

SK v/s

P v/s

7/8 4/0 – –

7/8 1/0 – –

– 22/29 5/5 8/8 – 16/18 – 7/7

9/10 – – –

IL v/s

v/s 13/0 6/0 7/0 2/0

Key: ESDTP: Estonian Social Democratic Labour Party; EÜRP: Estonian United People’s Party; KE: Estonian Center Party; RL: Estonian People’s Union; M: People’s Party Moderates; R: Estonian Greens; EE: Estonian Entrepeneurs Party; KMÜ: Coalition Party and Country People’s Union; RE: Estonian Reform Party; RP: Union for the Republic – Res Publica; EKRP: Estonian Christian People’s Party; EK: Estonian Citizen; ERSP: Estonian National Independence Party; SK: Independent Royalists; P: Right-Wingers’ Party; IL: Party Fatherland Union; v/s: percentage of votes/percentage of seats.

343

344

Table 18.3

Latvia: Parliament (Saeima) electoral results, 1993–2002

Year

Left Saskana LSP L v/s v/s v/s

1993 1995 1998 2002

12/13 6/6 14/16 19/25

– 6/7 6/5 – – – – –

Centre-Left

Centre

Centre-Right

LSDSP v/s

Vieniba v/s

LZS v/s

DCP/Saimnieks v/s

JL v/s

LC v/s

1/0 5/0 13/14 4/0

0/0 7/8 0/0 –

11/12 6/8 2/0 9/12

5/5 15/18 2/0 –

– – – 24/26

32/36 15/17 18/21 5/0

Right TP v/s

LPP v/s

TB v/s

LNNK v/s

Others TKL v/s

v/s

– – 5/6 13/15 – – – 12/14 6/8 15/16 21/24 – 15/17 – 2/0 17/20 10/10 5/7 – –

15/6 7/0 13/8 7/0

Key: Saskana: People’s Harmony Party; LSP: Latvian Socialist Party; L: Equal Rights; LSDSP: Latvian Social Democratic Workers Party; Vieniba: Unity Party; LZS: Farmers Union; DCP/Saimnieks: Democratic Party Saimnieks; JL: New Era; LC: Latvia’s Way; TP: People’s Party; LPP: First Party; TB: For Fatherland and Freedom; LNNK: Latvian National Independence Movement; TKL: People’s Movement for Latvia-Siegerists; v/s: percentage of votes/percentage of seats.

Table 18.4

Lithuania: Diet (Seimas) electoral results, 1992–2000

Year

1992 1996 2000

Centre-Left

Centre

Centre-Right

Right

Others

LVP v/s

LDDP/SDK v/s

LSDP v/s

LRS v/s

LLRA v/s

NS(SL) v/s

LCS v/s

LLS v/s

NKS v/s

LKDP v/s

KDS v/s

TS(LK) v/s

JL/PKS v/s

LTS v/s

v/s

– 2/1 4/3

44/52 10/8 31/36

6/6 7/7 –

– 2/0 –

2/3 3/1 2/3

– – 20/21

3/1 8/9 3/1

2/0 2/1 17/24

– – 2/1

13/13 10/11 3/1

– 3/1 4/1

21/19 30/50 9/6

4/1 4/1 1/1

2/3 2/2 1/0

3/2 17/8 3/2

Key: LVP: Lithuanian Peasant Party; LDDP/SDK: Lithuanian Democratic Labour Party/Social Democratic Coalition; LSDP: Lithuanian Social Democratic Party; LRS: Lithuanian Russian Union; LLRA: Polish Electoral Action; NS(SL): New Union (Social Liberals); LCS: Lithuanian Centre Union; LLS: Lithuanian Liberal Union; NKS: Moderate Conservative Union; LKDP: Lithuanian Christian Democratic Party; KDS: Christian Democratic Union; TS(LK): Homeland Union (Lithuanian Conservatives); JL: Young Lithuania; PKS: Union of Political Prisoners; LTS: Lithuanian National Union; v/s: percentage of votes/percentage of seats.

345

346

Evald Mikkel and Vello Pettai

References Central Electoral Commission of Latvia (2002) available at . Clark, Terry D. and Nerijus Prekevicius (2001) ‘The Effect of Changes to the Electoral Law in Premier-Presidential Systems: The Lithuanian Case’, in Algimantas Jankauskas (ed.), Lithuanian Political Science Yearbook, 2000. Vilnius: Institute of International Relations and Political Science, pp. 124–37. Eesti Vabariigi Valimiskomisjon (1992) Vabariigi Presidendi ja Riigikogu valimised 1992: Dokumente ja materjale. Tallinn: Eesti Vabariigi Valimiskomisjon. ENSV Valimiskomisjon sekretariaat (1990) ENSV Ülemnõukogu valimised 18. mästdil 1990 aastal. Gelazis, Nida (2001) ‘Institutional Engineering in Lithuania: Stability Through Compromise’, in Jan Zielonka (ed.), Democratic Consolidation in Eastern Europe, Volume I: Institutional Engineering. Oxford: Oxford University Press, pp. 165–85. Girnius, Saulius (1992) The Parliamentary Elections in Lithuania, RFE/RL Research Report 1/48, 4 December, pp. 6–12. Grofman, Bernard, Evald Mikkel and Rein Taagepera (1999) ‘Electoral Systems Change in Estonia, 1989–1993’, Journal of Baltic Studies, 30, 3: 227–49. Ideon, Argo (2002) ‘Uued valimisreeglid soosivad Toompeal häältemagneteid’, Postimees, 31 October 2002. Krupavicius, Algis (1998) ‘The Post-Communist Transition and Institutionalization of Lithuania’s Parties’, Political Studies, 46: 465–91. Lukošaitis, Alvidas (1998) ‘Parlamentas ir parlamentarizmas nepriklausomoje Lietuvoja. 1918–1940 ir 1990–1997 m’, in Algis Krupavicius (ed.), Seimo Rinkimai ‘96, treciasis ‘atmetimas’. Kaunas: Tverme, pp. 1–39. OSCE/ODIHR Election Observation (1998) Republic of Latvia Parliamentary Elections and National Referendum, 3 October 1998. OSCE Office for Democratic Institutions and Human Rights, Assessment Mission. OSCE/ODIHR Election Observation (1999) Republic of Estonia Parliamentary Elections, 7 March 1999. OSCE Office for Democratic Institutions and Human Rights. Pabriks, Artis (2002) Occupational Representation and Ethnic Discrimination in Latvia. Public Policy. Riga: Nordic. Pettai, Vello and Marcus Kreuzer (1999) ‘Party Politics in the Baltic States: Social Bases and Institutional Context’, East European Politics and Societies, 13, 1: 150–90. Statistical Office of Estonia (2001) 2000 Population and Housing Census II. Tallinn. Taagepera, Rein (1997) ‘The Tailor of Marrakesh: Western Electoral Systems Advice to Emerging Democracies’, in Jürgen Elklit (ed.), Electoral Systems for Emerging Democracies. Experiences and Suggestions. Copenhagen: Danida, pp. 49–62. Taagepera, Rein and Matthew Shugart (1989) Seats and Votes. The Effects and Determinations of Electoral Systems. New Haven, CT: Yale University Press. Zielonka, Jan (1994) ‘New Institutions in the Old East Bloc’, Journal of Democracy, 5, 2: 87–104.

Further reading • Huang, Mel (2000) ‘Changing the Rules at the Half’, Central Europe Review, 2/27, 10 July. • Mikkel, Evald (1999) ‘1999 aasta Riigikogu valimised ja kandidaadid’, in Rein Toomla (ed.), Riigikogu valimised. Tartu: Tartu Ülikooli Kirjastus, pp. 84–133.

• Pettai, Vello (2001) ‘Estonia: Positive and Negative Institutional Engineering’, in Jan Zielonka (ed.), Democratic Consolidation in Eastern Europe, Volume 1: Institutional Engineering. Oxford: Oxford University Press, pp. 111–38. • Sprudzs, Adolf (2001) ‘Rebuilding Democracy in Latvia: Overcoming a Dual Legacy’, in Jan Zielonka (ed.), Democratic Consolidation in Eastern Europe, Volume 1: Institutional Engineering. Oxford: Oxford University Press, pp. 139–64.

19 The Czech Republic: Entrenching Proportional Representation Petr Kopecký

After the collapse of the communist regime in Czechoslovakia, the first free elections – to the then Czechoslovak Federal Assembly and the Czech and Slovak National Councils – were held simultaneously in June 1990. The same set of parliamentary elections was repeated in June 1992. After the split of the Czechoslovak Federation and the emergence of the independent Czech Republic and Slovakia in early 1993, there were separate elections to the two chambers of the Czech parliament and the Slovak assembly. Common to all post-communist elections held in the territory of the present Czech Republic is that they were conducted under similar electoral systems – a proportional representation list system in the case of the lower house and a majority system in the case of the upper house. Indeed, electoral systems proved to be rather resilient in the face of several far-reaching reforms that were attempted in the first decade of post-communist politics, as did in fact the entire institutional system of the Czech Republic enshrined in the compromise constitution adopted at the time of the peaceful partition of Czechoslovakia. This chapter focuses mainly on the electoral system to the lower house of the Czech Republic parliament – constitutionally superior to the upper house – and starts by exploring the bargaining process of political actors during the transition period where the current system originated. I sketch out key elements of original electoral design and show that, while in principle similar to the system used in the pre-communist Czechoslovakia, it departs from it in several important aspects that were instituted mainly because of the strong anti-party sentiments of the early 1990s. I then proceed to analyse several reforms of the original electoral system, especially those that led to its change in the run up to the 2002 elections. I show that while most of the reforms were geared to strengthen the position of bigger and established political parties at the expense of smaller parties and the newcomers, the existing balance of political power, together with prohibitive constitutional provisions, have made the task of willing electoral engineers nearly impossible. In the concluding section, I assess the consequences of proportional representation on several aspects of the political process in the Czech Republic. 347

348

Petr Kopecký

Choosing proportional representation during the transition The origins of the current Czech electoral system date back to the early 1990s. On 17 November 1989, the police brutally repressed a student demonstration on the streets of Prague. The popular reaction was immediate and, within the course of few days, the communist regime collapsed. Following several round table meetings between the Communist Party and the anti-communist opposition, organized into two umbrella movements – Civic Forum (OF) in the Czech lands and Public Against Violence (VPN) in Slovakia – the opposition representatives were co-opted to both the Federal Assembly and the Czech and Slovak National Councils. At the same time, the former dissident and playwright Václav Havel became the Federal President and both the Federal Government, as well as the two respective regional governments, were recomposed to reflect the changed balance of power after the breakdown of the communist regime. For various reasons, and despite its dominant position following the implosion of the communist power monopoly, the opposition decided to postpone major constitutional reforms until after the first free elections. These were scheduled for June 1990 and, therefore, the one and only major institutional decision to be taken before the first free elections was that concerning the electoral law. Following lengthy negotiations, the decision was to change the majority electoral system then in place for (nominal) elections under the communist regime to a system based on proportional representation which had been used in the pre-war First Republic (1918–39): indeed this was the first time that a reference to the First Republic was explicitly made in reshaping the institutional structure of the country. On one side, well calculating the advantages it offered to smaller parties, both Slovak representation in the Federal Assembly (VPN), as well as the decimated Communist Party and its former satellite parties, all favoured a proportional representation system directly from the beginning. The Civic Forum (OF), on the other side, was heavily split on the issue. Havel and some of his associates in particular argued for a majority system allowing for the selection of independent candidates and the relative marginalization of the role played by political parties. In their view, the First Republic was a negative example of political consequences brought about by a proportional representation electoral system: it produced a fragmented parliament, it empowered political parties and, at the same time, it reduced the chance for independent candidates to compete.1 However, Havel and his followers within the OF were eventually persuaded that proportional representation was a better option at that time, and this for several reasons. Firstly, it was agreed that a proportional representation system provided more opportunities for a complete spectrum of parties to be formed – something generally deemed desirable at a time when new party political organizations were only beginning to spring up. Under a majority system, it was argued, OF (and VPN in Slovakia) would very likely exploit their then dominant position by winning all the parliamentary seats. Secondly, the new electoral law was considered only temporary (i.e. covering just the first free elections) – a fact which, as we shall see below, fostered Havel’s later attempts to push through an alternative electoral

The Czech Republic

349

system. Thirdly, both OF and VPN realized the perils of restructuring the centrally planned economy and thus tried to avoid sole responsibility for the difficulties lying ahead by institutionally fostering a broad coalition government. Proportional representation was accepted nearly unanimously by the Federal Assembly on 27 February 1990 and, shortly afterwards, the Czech National Council – later to become the supreme legislative body in the newly emerged Czech Republic – approved its own electoral law, which nevertheless reflected the basic conception of the federal one. However, while the proponents of a majority formula were not quite able, or even willing, to block the choice of a proportional representation system, they were at least able to preclude the simple adoption of the pre-war electoral system – the solution proposed for example by the Socialist Party which, at that time, had a sizeable representation in both the Federal Assembly and the Czech National Council. They also exerted pressure on the designers of the electoral law who, consequently, felt obliged to take into account some of the grievances against proportional representation closed list systems, including those used in the First Republic. This can be seen most clearly in three key aspects of the electoral system adopted in 1990: the formula for seat allocation, the legal threshold and the ballot structure. While the electoral system for the lower chamber in the prewar republic used the Hare quota in (23) districts for the first-tier distribution of seats and the Droop quota for the national tier, the 1990 electoral system for the 200-member Chamber of Deputies (Czech National Council between 1989 and 1992) adopted the Droop quota and largest remainders. The reason for this change was to reduce the possibility that the much unloved centralized party executives could influence candidate selection at the latter stage of the electoral process since, as research shows (see Krejcí, 1994: 144–5; Lijphart, 1994: 23), the Droop quota, 1/M + 1 (rounded up), helps to achieve more complete distribution of seats at the district level than the Hare quota, 1/M. Indeed, recalculations of the 1992 electoral results using the Hare quota instead of the Droop quota showed that the number of unallocated seats after first scrutiny would have been considerably higher (Kopecký, 2001). The second innovation to the electoral system concerned the legal threshold. While the legal threshold in the First Republic was only for the national redistribution (at least one mandate or 20,000 votes in any district), fostering a fragmented multi-party system, it was set at 5 per cent of the votes for the 1990 elections to the Czech National Council (and also to the Federal Assembly). Behind the introduction of the threshold was obviously an attempt to avoid excessive fragmentation of parliament, perceived as a real possibility in 1990 when dozens of political parties were created amid the euphoria of freshly acquired freedoms of association. The actual percentage was probably influenced by foreign examples, specifically that of neighbouring Germany and Austria, and by the fact that 4 or 5 per cent is now widely seen as a judicious level at which the smallest parties are eliminated while the proportional representation principle remains basically intact (see Lijphart, 1992). The third important innovation of the electoral system concerned the ballot structure. Compared to the system used in the First Republic, which utilized closed

350

Petr Kopecký

lists of candidates ordered by political parties, the 1990 electoral law provided voters with the possibility of altering the proposed lists by casting preferential votes. This change was a clear concession to those arguing for giving voters the choice of parties as well as personalities, and against the possible over-concentration of power in the hands of intra-party organs responsible for the selection of candidates for national office. Here again, the strong anti-party sentiments played a role in the initial electoral design – the design which, as we shall see now, subsequently turned out to be very difficult to change.

Attempts to reform the electoral system Not surprisingly, given the compromises reached during the period of transition, Havel was the first to challenge the electoral system that was used for the first free elections in 1990. As the Federal president between 1990 and 1992, he proposed a new electoral law for the coming 1992 elections (the first legislative term was deliberately set up for only two years). The draft electoral law was part of a broader package of constitutional amendments, also including proposals to increase presidential powers and change the rules on referenda, presented to the Federal Assembly in December 1991. Havel’s electoral bill aimed at dividing the country into singlemember districts in which voters would cast their votes for individual candidates rather than for political parties, using supplementary vote to determine the winning candidate. The president defended his proposal in the 1992 New Year’s speech, arguing that the new electoral law provided for ‘deepening the link between voters and MPs, and for providing stability to the future parliament, government, and the whole political system’ (Havel, 1992: 139). However, the proposal, and indeed the whole package of reforms, was turned down by the Federal Assembly, principally because Havel could not, and in fact even did not sufficiently try to, persuade political parties that his proposals were not threatening their rapidly emerging dominant position within the political system. Havel was nevertheless more successful with his pursuit of a majority electoral system during the negotiations on the new Czech constitution at the end of 1992, when the Czechoslovak Federation came to the negotiated end. It was then decided that the new Czech parliament would be bicameral, with the Czech National Council (now renamed the Chamber of Deputies) acting as its lower chamber. In order to ensure a slightly different make-up of the two chambers, the upper house was to be elected under a different electoral system than that used for the lower chamber. Indeed, the Czech constitution explicitly stipulates that a majority system must be used to elect the Senate. Interestingly, party squabbling over exactly which majority formula to adopt for Senate elections delayed the establishment of the upper house for another three years. The compromised law finally adopted in 1995, and used ever since, is majority runoff. A candidate is elected in a singlemember constituency (each representing about 100,000 inhabitants) if he or she receives more than 50 per cent of votes. If no one in a given constituency receives over 50 per cent in the first round, the second round is held a week later, consisting of runoffs with the two most successful candidates.

The Czech Republic

351

While the electoral law for the Senate has not yet been subjected to reforms, the proportional representation system used for the Chamber of Deputies has been (slightly) changed before almost all elections. However, it is important to emphasize at the outset that that the task of electoral engineers is complicated by another important provision in the Czech Constitution which stipulates that the elections to the lower house of the parliament are to be conducted under a proportional representation formula. Any amendment of the Constitution requires three-fifths of all deputies of the Chamber of Deputies and three-fifths of the Senators present. This, together with the existence of smaller parties in (and out) of parliament, have made a wholesale change of the electoral (and constitutional) system always unlikely. And indeed, the more serious debates and changes, relatively modest until 1998, have always focused on rehearsing the various measures, such as the further raising of legal thresholds or decreasing the number and size of electoral districts, which would create the desired effect without a major constitutional overhaul. However, the results of the 1996 elections produced a constellation of parties that made it difficult to establish a majority controlling government coalition. The coalition of the Civic Democratic Party (ODS), the Christian Democratic Union (KDU-CSL) and the Civic Democratic Alliance (ODA) formed a cabinet again (as in between 1992 and 1996), but were two seats short of a parliamentary majority. When the ODS split at the end of 1997 amid allegations of illicit party financing and the fragile government fell as a result, new elections were called for the spring of 1998. These elections saw a victory for the Social Democratic Party (CSSD), but because of the tensions between political parties at that time, no majority winning coalition was formed despite the mathematical possibility to do so. Instead, hitherto major political arch-rivals – the conservative ODS and the social democrat CSSD – decided that the former would provide support for a single-party minority government of the CSSD in return for several important policy concessions. Among these was an agreement, enshrined in the so-called ‘Tolerance Pact’, to change the electoral system, not surprisingly in the way that would benefit the two largest parties. The protracted negotiations on so far the most far-reaching electoral reform thus started. The ODS, with a sizeable contingent of politicians (including its leader Václav Klaus) long-time favouring a majority electoral system, took the lead. Both the ODS and CSSD had at that time sufficient strength in both houses of parliament to force a change in the Constitution to allow for a majority system, but because of disunity within the ODS, the party proposed a radical modification of the proportional representation system then in use rather than a wholesale electoral (and constitutional) change. With the twin aim to manufacture a government majority of one political party and to reduce parliamentary representation of smaller parties (chiefly their former coalition allies plus the pariah Communist Party), they suggested, in spring 1999, to decrease the size of the Chamber of Deputies from 200 to 162 members elected in small 35 multi-member districts using the Imperiali formula and a 5 per cent legal threshold. The CSSD, now alarmed by their poor performance in the opinion polls, wavered in the face of

352

Petr Kopecký

this drastic proposal and suggested a system that would essentially create a lower effective threshold: a 200-member Chamber of Deputies elected in 35 multi-member districts using the d’Hondt formula. A major political fight broke out between the two parties, with the ODS threatening to terminate the ‘Tolerance Pact’ and bring down the government. Both parties had nevertheless come to a compromise and, in May and June 2000, pushed a new electoral law through both houses of parliament. This law was based on a 200-member Chamber of Deputies elected in 35 multi-member districts (each between four and eight members) using a modified d’Hondt formula in districts. Furthermore, while the 5 per cent electoral threshold for single parties was to be maintained, the threshold for coalitions of two, three and four plus parties was to be increased to 10, 15 and 20 per cent respectively. While obviously still a proportional representation system, the electoral law would have created a higher effective threshold than the existing system, thus increasing the disproportionality of electoral results. (An excellent analysis of the likely effects of the proposed electoral law can be found in Crawford (2001) and Williams (2002); see also Koudelka (2000) and Kostelecký (2000).) The President Havel, unhappy with the blatant electoral engineering of the two largest parties, used his constitutional prerogatives and vetoed the law. His veto was nevertheless overturned by the ODS and CSSD dominated parliament and so, also in accordance with the Constitution, he decided to send this law, together with the new and similarly engineered law on campaign finances, to the Constitutional Court. Somewhat ironically given his previous views on the electoral system, Havel argued that both laws were unconstitutional because they hindered the formation of coalitions and transformed a proportional representation system into a majority one; he also argued that both laws restricted political competition. The members of parliament from the christian democratic KDU-CSL and the Freedom Union (US), now united in the so-called Quad-Coalition (4K), signed the petition as well, even though, on the basis of opinion polls at that time, they were actually more likely to benefit from the new electoral law than its authors. In January 2001, the Constitutional Court struck down all but one of the controversial provisions of the new electoral law, arguing that, together with the campaign finances law, it constituted an attack on the smaller opposition parties by both the ODS and the CSSD. The only provision to survive the Court’s ruling was the increased threshold for the entry of coalition parties to parliament. With the elections fast approaching, it was now the CSSD that took the initiative and proposed new electoral rules in a compromise version in which the number of electoral districts was adjusted to equal the number of regions (14) that were established in past years, and the distribution of seats calculated on the basis of d’Hondt system (see Table 19.1). This law was approved by both houses of parliament in January 2002, despite the opposition of (some of the) Quad-Coalition deputies and senators to the increased threshold for party alliances to enter the parliament. The regular elections in May 2002 were conducted under this new electoral system.

The Czech Republic

Table 19.1 Year

1992 1996 1998 2002

353

Czech Republic: Chamber of Representatives electoral results, 1992–2002

Communists

Socialists

Centre

Conservatives

Right

Others

LB/KSCM v/s

CSSD v/s

LSU v/s

KDU-CSL/US v/s

ODS/ODA v/s

SPR-RSC v/s

v/s

14/19 11/11 11/12 19/21

8/10 27/31 32/37 30/35

6/7 – – –

6/7 8/9 29/19 14/15

39/49 36/40 28/32 25/29

7/8 8/9 4/0 –

20/0 10/0 6/0 12/0

Key: LB: Left Block; KSCM: Communist Party of Bohemia and Moravia; CSSD: Czech Social Democratic Party; LSU: Liberal Social Union; KDU-CSL: Christian and Democratic Union-Czech People’s Party; US: Freedom Union-Democratic Union; ODS: Civic Democratic Party; ODS/ODA: Civic Democratic Alliance; SPR-RSC: Rally for the Republic-Republican Party of Czechoslovakia; v/s: percentage of votes/percentage of seats.

Democratizing with proportional representation The collapse of communism in the Czech Republic and elsewhere in the region did not only open the question about the likelihood of a successful transition and consolidation of democracy, but also the question about what type of democracy was to be instituted. The early 1990s saw such debates flourishing on the Czech domestic political scene, as we have seen above in relation to negotiations concerning the form of electoral system. Broadly speaking, one cross-section of the political elite, situated mainly on the emerging right of the political spectrum and later concentrated in the ODS, argued essentially for a parliamentary democracy with a strong role for political parties. Another group, including many of the dissidents (notably Havel) and situated mainly on the emerging liberal/left part of the political spectrum, argued for an alternative vision of politics based on broad and diffuse interest representation. The idea of ‘party’ and ‘party government’, as well as the politics of formal institutional mechanisms, encountered a strong normative opposition within this section of the political elite. The former group eventually prevailed and, with the benefit of hindsight, it is possible to say that the contingent decision to adopt proportional representation in early 1990 actually marks the beginning of the formation of party government in the Czech Republic – the political arrangement which the proportional representation electoral system has fostered in several ways. Firstly, representation at the national level is dominated by political parties; it is the party rather than individual MP which provides the crucial link between citizens and the state. Research shows that deputies see their roles of representing local constituencies or professional and social interests as largely subordinated to the roles associated with the execution of the party mandate. This is not that surprising given that MPs owe their careers to political parties which, if officially registered and in possession of sufficient money to pay election deposits, are the only institutions that can actually put forward candidates for elections. The

354

Petr Kopecký

shape of the electoral system, before and after the reforms, provides structural incentives underpinning this form of representation. For the Czech deputies, most of whom lack the tradition of and experience with constituency work, are elected in relatively large districts under a system where the initial position on the list of candidates is more important in the electoral process than popularity with the local electorate. Indeed, preferential voting or open list has had a relatively limited impact on the electoral process. True, the 2002 elections produced an intriguing twist of fate for several prominent politicians from the electoral coalition of the KDU-CSL and the US: their very hard fought-over list containing a carefully balanced mix of candidates from both parties was dramatically altered by the voters, who used preferential votes in such a way as to reduce the representation of the US to the advantage of the KDU-CSL. However, this surely owed far more to the fact that the KDU-CSL voters are very disciplined and loyal party followers, who voted en bloc for their party men (and women) rather than for any better qualities and reputations among individual candidates. A similar twist had occurred during the 1990 elections. Voters were then largely unaware of the candidates of the democratic opposition. In order to promote their campaign, both the OF and VPN placed a number of well-known popular artists and actors on their lists but, because of their unwillingness to serve as members of parliament, located them low down. To the surprise of many, voters allocated them sufficient preferential votes as to guarantee them seats in parliament. Very shortly after the election, a number of such ‘accidentally’ elected MPs resigned their mandate and were replaced by those more willing to serve but less popular. Overall, preferential voting is now more a test of personal prestige for the various leading politicians rather than, as was initially hoped, the means to induce (local) candidate-centred electoral choice. Thus, while not necessarily owing their election to the personalized system of voting, a rare occurrence indeed, politicians occasionally use obtained preferential votes to enhance their bargaining position in the post-election selection of ministerial and top parliamentary posts. Secondly, the work of the parliament is almost entirely dependent on the organization of the political parties that operate within it. The individual deputies have become, partly in institutional and almost fully in behavioural terms, the bearers and executors of the party mandate. This can be seen in high levels of party cohesion during floor votes in the parliament; also, the behaviour of deputies in parliamentary committees and the distribution of committee and parliamentary chairs reflect the subservience of the parliamentary machinery to party politics. The party cohesion has increased, in comparison with the first and second legislative terms, because initially high levels of party fragmentation have sorted out many unlikely friendships and political alliances, thus making political parties internally more homogeneous. We should also note the increased ability, and willingness, of parties to use patronage to reward party loyalty. However, the candidate selection process, with its direct relation to the form of the proportional representation electoral system used in the Czech Republic, has played an important role too. For the disobedient and other out-of-favour members of parliament can hardly count

The Czech Republic

355

on securing nomination or placement sufficiently high on the list in the generally much controlled intra-party processes of candidate selection. The increased levels of parliamentary party cohesion have been paralleled by decreasing levels of parliamentary party fragmentation. Again, parliamentary rules, geared towards instituting the system where party plays a crucial role, are important in this respect. For example, at least ten members of Parliament are needed to form a parliamentary party; nevertheless, if such a group is a splinter from a regularly elected party, it is not entitled to the same material and financial benefits which all parties get from the parliamentary budget. This has made ‘faction hopping’ a distinctly difficult phenomenon and, indeed, while dozens of MPs changed party affiliation between 1990 and 1996, only two deputies (one of them expelled) changed party benches between 1998 and 2002. However, the decreased levels of parliamentary party fragmentation also relate to the electoral system (and the voting behaviour under it). For both new parties and splinter groups usually suffer badly in elections, which are now more than ever a race between established and stable party organizations. This has of course had an effect on both the disproportionality of electoral results and the number of parliamentary parties, both decreasing over the years. While in 1990, 18.8 per cent of votes (as a percentage of total votes cast) were given to parties that did not pass the threshold, the same figure for 2002 was 12.4 per cent (rates of disproportionality, measured with the least-squares index, were decreasing: 11.6 in 1990, 8.6 in 1992, 5.6 in 1996 and 5.8 in 1998 – Williams, 2002). And while eight parties (or coalitions) were elected to the parliament in 1992, it was only five parties and four parties (or coalitions) in 1998 and 2002 respectively. Note, however, that the so-called Two-Coalition (of the KDU-CSL and US) immediately formed two separate parliamentary parties after the 2002 elections. As Sarah Birch (2001) shows, the effective number of parliamentary parties in the Czech Republic (in 1998) was 3.70, somewhat below the Eastern European average of 4.08 Thirdly, given the shape of the electoral system, any government controlling the majority of seats in the parliament requires at least two parties to join forces and form a multi-party coalition. The two largest parties shared between them 129 seats (out of 200) after the 1996 elections, and 137 and 128 after the 1998 and 2002 elections respectively. However, the two largest parties – (since 1996) the ODS and CSSD – are on opposite sides of a left–right dimension around which the party competition centres, representing two major rivals within the Czech party spectrum. Therefore, minimum-winning coalitions normally require either the CSSD or ODS, in coalition with the KDU-CSL and US (ODA until 1996), since the Communist Party (KSCM) have so far been universally considered without coalition potential, despite their stable parliamentary presence. In practice, this has meant an ODS/KDU–CSL/ ODA government following the 1992 and 1996 elections, and a CSSD/US/KDUCSL government after the 2002 elections, always commanding, at least initially, a tiny majority (105 seats in 1992, 100 seats in 1996 and 101 seats in 2002). On the face of this data, the 1998 coalition deal, which has meant the support of the ODS for a CSSD single-party minority cabinet in exchange for support for an electoral reform that would possibly manufacture larger shares of votes for

356

Petr Kopecký

larger parties, is not that surprising. The government by coalition has, under these circumstances, proved to be far more contentious on the elite level than the party government model itself, mainly because two major political parties started to blame the electoral system for their inability to form and maintain government coalitions. The truth is that even in 1998 there was a mathematical possibility for a minimum-winning (108 seats) centre-right coalition. This suggests that the underlying problem is not necessarily proportional representation, but rather the inter-elite and inter-party relations, plagued, as was apparent in 1998, not only by ideological and policy differences, but also by several personal antipathies. In any case, the tiny majorities on which the Czech governments tend to hang inevitably created pressures to maintain parliamentary party cohesion which, as indicated above, has led to several institutional and political measures restricting the freedom of individual deputies. And in this respect, a great deal of governmental stability in the Czech Republic speaks volumes against many academic accounts that associate East European party systems only with fragmentation and instability. In the final analysis, it is important to emphasize the political character of all attempts to reform the electoral system in the Czech Republic. The 1990 decision to adopt a specific form of proportional representation was certainly a measure then perceived as serving the public good rather than only narrow party interests. However, the reforms that have been attempted ever since, perhaps with the exception of President Havel’s constitutional activities between 1990 and 1992, have always been horse-trading between political parties, each pursuing their powerrelated preferences, however dressed up and perceived. Nevertheless, the limited success of such reforms allows one to speak about entrenched proportional representation. And one also wonders whether it is only a matter of coincidence that any wholesale change of the electoral system in Czechoslovakia or the Czech Republic came about only with a major regime change – from a majority system to proportional representation in 1918, to a majority system in 1948 and back to proportional representation in 1990 – and that it was a proportional representation system that has always gone together with full democracy.

Note 1. The prewar Czechoslovak Republic was primarily a party democracy, or, as some have it, ‘the dictatorship of political parties’ (Broklová, 1992: 74; see also Rothschild, 1992; Skalnik-Leff, 1988). Party system stability descended from the mutually reinforcing blend of institutional factors and the fabric of Czechoslovak society. In addition to a very stable, albeit numerically restricted electoral base (no party was able to master more than a quarter of the popular vote), the rules of the game actively aided the emergence of strong party politics too. The centrality of parties, especially the traditional Czech parties, was already apparent during the constitutional bargaining in 1918, and it is thus no surprise that a number of constitutional provisions provided for both strengthening parties’ position within the polity as well as for the position of leadership within parties. Among these was, first of all, the proportional representation based electoral system, providing for compulsory voting and closed party-ordered lists of candidates. Together with the rules that a deputy or senator must give up his seat should he leave the party during the legislative

The Czech Republic

357

term, and the deputies’ and senators’ compulsory membership in a parliamentary party (club), this endowed party leaderships with a formidable system of discipline to ensure unity within their respective party organizations.

References Birch, Sarah (2001) ‘Electoral Systems and Party Systems East and West’, Perspectives on European Politics and Society, 2, 3: 355–77. Broklová, Eva (1992) Deskoslovenská Demokracie: Politický Systém DSR 1918–1938. Prague: Slon. Havel, Václav (1992) Vázení obd ané: Projevy dervenec 1990 – dervenec 1992. Prague: Lidové noviny. Crawford, Keith (2001) ‘A System of Disproportional Representation in the Czech Republic: the Proposed Electoral Law for the Czech Republic’, Representation, 38, 1: 46–58. Kopecký, Petr (2001) Parliaments in the Czech and Slovak Republics: Party Competition and Parliamentary Institutionalization. Aldershot: Ashgate. Kostelecký, Tomáš (2000) ‘Proposed Changes to the Electoral Law Resulting from Amendments to the “Opposition Agreement” in 2000 and Possible Consequences’, Czech Sociological Review, 36, 3: 299–305. Koudelka, Zdenèk (2000) ‘Electoral Reform: Simulation of the Elections According to the Proposed New System’, Parlamentní zpravodaj, 6, 3: 2–44. Krejdí, Oskar (1994) Kniha o volbách. Prague: Victoria Publishing. Lijphart, Arend (1992) ‘Democratization and Constitutional Choices in Czechoslovakia, Hungary and Poland: 1989–1991’, Journal of Theoretical Politics, 4, 2: 207–23. Lijphart, Arend (1994) Electoral Systems and Party Systems: A Study of Twenty-Seven Democracies 1945–1990. New York and Oxford: Oxford University Press. Rothschild, Joseph (1992) East Central Europe between the Two World Wars, 7th edn. Seattle: University of Washington Press. Skalnik-Leff, Carol (1988) National Conflict in Czechoslovakia. Princeton, NJ: Princeton University Press. Stanger, Allison K. (2000) ‘The Price of Velvet: Constitutional Politics and the Demise of the Czechoslovak Federation’, in Michael Kraus and Allison K. Stanger (eds), Irreconcilable Differences? Explaining the Dissolution of Czechoslovakia. Lanham, MD: Rowman & Littlefield, pp. 137–59. Williams, Kieran (2002) ‘The Czech and Slovak Republics: the Surprising Resilience of Proportional Representation’, in Sarah Birch, Frances Millard, Marina Popescu and Kieran Williams, Embodying Democracy: Electoral System Design in Post-Communist Europe. Basingstoke: Palgrave – now Palgrave Macmillan.

Further reading • Elster, Jon (1995) ‘Transition, Constitution-Making and Separation of Czechoslovakia’, European Journal of Sociology, 36: 105–34.

• Havel, Václav (1992) Vázení obcané: Projevy cervenec 1990 – cervenec 1992. Prague: Lidové noviny.

• Jicínský, Zdenek (1993) Ceskoslovenský parlament v polistopadovém vývoji. Prague: NADASAFGH.

• Kopecký, Petr (2000) ‘From Velvet Revolution to Velvet Split: Consociational Institutions and the Dissolution of Czechoslovakia’, in Michael Kraus and Allison K. Stanger (eds), Irreconcilable Differences? Explaining Czechoslovakia’s Dissolution. Lanham, MD: Rowman & Littlefield, pp. 69–86. • Kopecký, Petr (2001) Parliaments in the Czech and Slovak Republics: Party Competition and Parliamentary Institutionalization. Aldershot: Ashgate.

358

Petr Kopecký

• Kopecký, Petr (2003) ‘Building Party Government: Political Parties in the Czech and Slovak Republics’, in Paul Webb and Stephen White (eds), Political Parties in New Democracies. Oxford: Oxford University Press. • Lijphart, Arend (1992) ‘Democratization and Constitutional Choices in Czechoslovakia, Hungary and Poland: 1989–1991’, Journal of Theoretical Politics, 4, 2: 207–23. • Stanger, Allison K. (2000) ‘The Price of Velvet: Constitutional Politics and the Demise of the Czechoslovak Federation’, in Michael Kraus and Allison K. Stanger (eds), Irreconcilable Differences? Explaining the Dissolution of Czechoslovakia. Rowman & Littlefield, pp. 137–59. • Williams, Kieran (2002) ‘The Czech and Slovak Republics: the Surprising Resilience of Proportional Representation’, in Sarah Birch, Frances Millard, Marina Popescu and Kieran Williams, Embodying Democracy: Electoral System Design in Post-Communist Europe. Basingstoke: Palgrave – now Palgrave Macmillan.

20 Hungary: Compromising Midway on a Mixed System John W. Schiemann1

In Hungary’s mixed electoral system voters elect 386 members of the national assembly from three districts using two ballots and four decision rules. This system was designed during Hungary’s national round table talks between the ruling Hungarian Socialist Workers’ (communist) Party (MSzMP) and a coalition of opposition parties during the Warsaw Pact country’s transition to democracy in 1989. Looking ahead to the parliamentary elections that would be conducted under the rules they would choose, the parties attempted to secure rules that would assist their – and/or hinder their opponents’ – entry into parliament, resulting in the patchwork of electoral rules little altered since. The four elections with this system have witnessed an alternation in power between conservative-populist and socialist parties. In the first elections of 1990 the Hungarian Democratic Forum formed a governing coalition with smaller parties tracing their origins to Hungary’s fleeting postwar democracy. In 1994 the majoritarian effects of the electoral system gave the Socialists an absolute majority of the seats but they choose to form a coalition government with the Association of Free Democrats, their one-time opponents under communism. In 1998, a strategic, ideological shift by the former liberal youth party, the Association of Young Democrats, to populist conservatism paid off and swept them and their allies to power. The 2002 elections saw a consolidation of the party system into two large blocs, the Socialists and their Free Democratic allies on the centre-left and Fidesz on the centre-right, with the former taking power from the latter.

Hungary’s mixed electoral system Mixed electoral systems are nothing new to Hungary, though earlier versions were employed for a restricted set of voters and designed to keep the conservative authoritarian regime in power. From 1922 to 1939 urban voters chose between party lists in multi-member districts while rural voters cast their ballots for individuals in single member districts (Földes and Hubai, 1994). In 1939 rural voters received an additional ballot, choosing both an individual from the local single-member district and a party running a list competing for multiple seats in the county. These mixed systems were dropped in favour of a purely proportional party list 359

360

John W. Schiemann

system for Hungary’s first experiment with free elections following the Second World War. With communism relegating elections to political ritual from 1947 to 1985, the question of electoral system design was of little significance. This all changed when the ruling communists (MSzMP) agreed in February 1989 to a multi-party system, something that would require a new election law. With the stakes high for both the regime and multiple opposition groups, they sat down to design Hungary’s new electoral system. The rest of this chapter examines the system they chose and the minor changes made since, explains why they chose it, and identifies some of the more important consequences it has had on politics in Hungary. The electoral system in Hungary is a mixed ‘parallel’ system of majority rule (with a second-round runoff) and proportional representation (with two-tier allocation). There are 386 members in the National Assembly. Of these 176 are elected in single-member districts and the remaining 210 are elected from party lists at both the territorial (county) and national level. Unlike the German system, seats won in the single-member districts tier do not count toward a party’s overall seat share. The electoral law calls for approximately 60,000 residents in each single-member district; in practice single-member districts have approximately 45,000 eligible voters, with some having fewer than 28,000. No doubt a result of uncertainty about the geographic distribution of voter support in 1989, there appears to have been little gerrymandering in the original apportionment of the single-member districts. The one instance in which the opposition suspected an attempt by the MSzMP to gerrymander backfired badly. The MSzMP – still viewing itself as a party of the working class – assigned two single-member districts to a working-class section of Budapest at the expense of a neighbouring wealthier district.2 After the first round of voting, the MSzMP was crushed in the two working-class districts while it did better (though still lost) in the wealthier section of Southern Buda – populated by cadres supportive of the party (Horváth, 1990: 436). Of the 210 list seats, 152 are distributed among Hungary’s 19 counties and Budapest while 58 come from a national list. District magnitude in the territorial tier ranges from 4 in sparsely populated counties to 28 in Budapest, with an average of 7.6. The desire for a minimal level of proportionality in the low-density counties has resulted in a wide variance in the territorial population to list seat ratio, from 59,611 in Zala county to 77,197 in Pest county, with an average of 67,086. The 152–58 distribution is, however, only a nominal starting point; due to a restriction on largest remainder allocation for the territories (discussed below), many unfilled territorial list seats are added to the 58 on the national list for allocation so that in practice district magnitude is larger for the national list and smaller for some of the territories. In 1990, 32 territorial seats were transferred to the national list, in 1994 it was 27, in 1998 24, and in 2002 12 seats were added to the national list, with an average transfer of nearly 21 seats per election. An unusual feature permits a single candidate to run simultaneously in a single-member district, on a party’s territorial list and on the party’s national list. Although nominally a dual ballot system, in practice most Hungarians cast three votes. In the first round of voting, Hungarians receive two ballots, one for their

Hungary

361

single-member districts and one for their multi-member county district. Placement on the single-member district ballot requires 750 signatures and both independents and party-backed candidates may run. A party wishing to establish a list for the territorial district voting must run a candidate in one-quarter of the county’s singlemember district races, with a minimum of two for low district magnitude counties. To establish a national list, a party must run party lists in seven territories/counties. In order for a candidate to win a seat in the first-round voting for single-member districts, he or she must exceed an absolute majority. Given the large number of candidates contesting the first round, this is rare and most Hungarians return to the polls two weeks later for a runoff round similar to the French system. In Hungary, the top three candidates and any candidate with more than 15 per cent of the vote in round one advance. The winner is decided by plurality. The decision rules for list voting are perhaps the most unusual features of the Hungarian system. At the territorial level seats are allocated according to the Droop quota, 1/M + 1 (rounded up), with largest remainders submitted to some restriction. In order to obtain a seat in the largest remainder allocation at the territorial level, a party must have two-thirds of the quota; if no party meets this threshold and there are still seats to distribute, those seats are added to the national list for compensatory allocation based on ‘surplus votes’ (explained below). Parties whose remainders fall between the two-thirds limit and the full quota and that receive a seat in the largest remainder allocation must make up the difference from their nationally pooled surplus votes, i.e. there is a negative vote transfer to the party’s national pool. The 58 national district seats as well as any territorial district seats are allocated to parties according to the d’Hondt highest average formula based on parties’ surplus votes. Surplus votes are: (1) votes for party candidates from the first round of single-member district voting that do not elect a party candidate in either round, and (2) remainder votes cast for parties in the territorial list voting (i.e. votes that did not count toward obtaining a seat in either the quota or largest remainder allocation). The legal threshold for allocation of seats on both the county and national lists is 5 per cent of the national vote (4 per cent in 1990). Any party that falls below 5 per cent of the national vote is ineligible to receive seats in the county list voting irrespective of their vote totals there. Running a joint list for small parties helps but little: joint lists still require 5 per cent per party, to a maximum of 15 per cent for more than three parties. Finally, turnout provisions require that 50 per cent of eligible voters vote in the first round; turnout below 50 per cent renders the balloting invalid and the election is repeated in the second round, where the turnout requirement is one-quarter of eligible voters.

The negotiated origins of Hungary’s electoral system This patchwork of different electoral system components has its origins in Hungary’s negotiated transition to democracy in 1989. With the agreement to a multi-party system in February of that year, the MSzMP committed itself to reforming the

362

John W. Schiemann

electoral system. Indeed, in anticipation of electoral threats from opposition groups, the MSzMP had already begun work on electoral ‘reform’ in late 1988 that actually represented a retreat from the reforms of 1985 by restricting the nomination process (Vass, 1994: 453). Protests from opposition groups, however, forced the MSzMP to jettison the plan in early 1989 and draft a completely new law. When the MSzMP accepted a multi-party system shortly thereafter, the opposition demanded a more formal role in the creation of the laws necessary for free elections and a peaceful transition, including the electoral law. Having failed in its initial attempt to impose a new electoral law and other institutions, the ruling party came to perceive the necessity of formal talks with opposition parties, now joined in a coalition against the regime, the Opposition Round Table (EKA). To get the EKA to the table, the MSzMP promised that it would withdraw its own draft electoral law from parliament and use it as the basis for negotiations in the National Round Table Talks (NKA) between the MSzMP and the EKA. Despite the fact that the MSzMP held the ultima ratio of any regime – coercive power – and the fact that most of the opposition parties were little known to the public, there was real bargaining in the National Round Table Talks over the provisions of the electoral law. That is, while the electoral power of many of the opposition parties was low, their bargaining power was not because international pressure (Western loans) meant that the MSzMP had to reach a modus vivendi with the opposition forces. This bargaining was further complicated by the fact that the seven opposition parties within the EKA had to reach agreement before negotiating as a bloc with the ruling party. While some of the opposition parties appear to have been motivated by non-partisan concerns over some elements of the electoral system, all the parties pursued their electoral self-interest in the design of some, and most parties of most, electoral rules. The Opposition Round Table was divided into three camps. First, there were four historical parties that had competed in the postwar elections – the Christian Democratic Peoples’ Party (KDNP), the Hungarian Peoples’ Party (MNP), the Hungarian Social Democratic Party (MSzDP) and the Independent Smallholders (FKgP). Second, the Association of Free Democrats (SzDSz) and the Association of Young Democrats (Fidesz) were two Western-leaning parties, the former having emerged from the dissident intelligentsia and the latter from a student movement. Finally, the Hungarian Democratic Forum (MDF) was the earliest independent group to form, growing out of the populist wing of the intelligentsia with support from reformist members of the MSzMP. Revived by some of their original members, the historical parties possessed famous names and political symbols but little in the way of known leaders, organization or recent political experience. As a result, they invoked the 1945 system and advocated a proportional system of direct voting for multi-member party lists based in the counties, a smaller national compensation list and a very low, perhaps no, threshold. In the KDNP’s proposal, the formula for the county allocation would distribute mandates to parties for every multiple of 20,000 votes (ARF 2000, II. fn. 69, pp. 447–9). Its weak name recognition vis-à-vis both the historical parties and the MDF but a strong stable of individual dissidents active against the regime (and concerns over governability) led the SzDSz to advocate

Hungary

363

a predominantly single-member district system (200 seats) with an additional 50 seats to be allocated proportionally based on remainder votes via party lists. Fidesz was in a position similar to the SzDSz, though they hoped their origins as a social movement of youth might make them more competitive in list voting. They consequently came out in favour of the SzDSz-type system but with proportions closer to an even division of single-member districts and list mandates. They also floated the possibility of an open list, but the idea was quickly dropped. Finally, the MDF was the most popular opposition party by far, with members and local chapters across Hungary. They not only stood to do well in competition between parties, but perhaps in head-to-head races with the MSzMP and other opposition candidates as well. In the absence of opinion polls they couldn’t be certain of this and so favoured a dual ballot system that combined single-member districts and list elements, a proposal they also knew would be necessary to bridge the divide within the EKA between the historical parties on the one hand and the SzDSz and Fidesz on the other. The EKA compromise presented to the MSzMP was just such a system: a mixed parallel system with dual ballot; an even 150–150 seat division between a singlemember district tier and a directly elected national party list tier with a 3 per cent threshold. For the single-member district tier, the EKA adopted the government’s own draft plan calling for a runoff round with the top two candidates if no candidate secured an absolute majority of the votes in the first round. The opposition also proposed, in an attempt to secure party leaders a seat in the new parliament, that an individual be permitted to run in a single-member district, on two territorial lists and on the national list simultaneously. The MSzMP favoured a system close to the SzDSz: 300 mandates chosen in single-member districts supplemented by a compensatory 50-seat national list allocating seats with a Hare quota (V/M). These positions changed over the course of July and August 1989 under the twin forces of bargaining compromises and (particularly for the MSzMP) increasing uncertainty about electoral prospects. The ruling party’s stunning defeat in three by-elections in late July using essentially their proposed single-member district majority runoff system, for example, caused it to demand a change to the rule to require at least three candidates in advance of the runoff. It also revived the idea of a national compensation list to utilize votes cast for losing MSzMP candidates in the single-member districts. Compromises between the two sides also contributed to the final system, including the nomination requirements for single-member district candidacy, basing the direct list on the counties, the requirements to run territorial and national lists and the threshold (from 3 and 5 per cent to 4 per cent). At the urging of an opposition expert who designed the two-thirds limit on largest remainder allocation in the county lists, the Droop quota replaced the Hare quota. The Hare quota was also replaced on the national list by the d’Hondt quota after the round table talks at the prompting of the MSzMP, who believed it would favour them as the largest party. The round table agreement submitted to parliament for ratification was met with resistance by MSzMP incumbents who had no role in the drafting process and who feared for their own political futures. MSzMP pressure on them limited their amendments to altering the relative balance between

364

John W. Schiemann

the single-member districts/county list/national list tiers (from 152/152/70 to 176/152/58) and changing the simultaneous candidacy rule (from one single-member district, two territorial lists and the national list to one single-member district, one territorial list and the national list). The law has changed little since 1989, the only significant amendment being the raising of the threshold for list allocation to 5 per cent in 1993.

Consequences As complicated as the system is, its consequences have become well known both to voters and to the political elites who contest power according to its rules. The general mechanical effects of the election law – relatively high disproportionality favouring larger parties or blocs of well-coordinated parties – have generated a variety of psychological effects, or voter and party strategies that anticipate those mechanical effects. The result in 2002 is a consolidation of the party system around two large blocs, one centre-left and the other centre-right. Although the list voting in the territories and the compensation mechanism drawing on unused single-member district votes would seem to provide a proportional counterbalance to the disproportionality of the single-member district tier, the Hungarian system has been marked by relatively high disproportionality, particularly in the first two elections of 1990 and 1994. To be sure, the list tiers do limit the disproportionality in the single-member district tier. In 1994, for example, the single-member district tier was four times more disproportional than the list tier. Nevertheless, disproportionality has decreased markedly beginning with the third elections in 1998. Thus while disproportionality for the system as a whole in 1990 and 1994 was 15.2 and 15.1 per cent respectively, it dropped to 6.5 per cent in 1998 and to 6.1 per cent in 2002 (calculation based on list votes). The populist MDF benefited from the system’s disproportionality in the first elections, when nearly 25 per cent of the list vote secured almost 43 per cent of the seats. The most dramatic instance of the system’s disproportional effects occurred in 1994, when the Socialists won 33 per cent of the party list vote but took 54 per cent of the seats in the National Assembly. Despite the fact that the electoral system favours larger parties, six parties passed the 5 per cent threshold and secured seats in parliament in the first three elections. For the first two of these elections, in 1990 and 1994, all six parties were participants at the National Round Table Talks (NKA) in 1989. The people’s MNP was the only political party at the NKA that has failed to win any seats in parliament. In 1998, the christian-democrat KDNP failed to pass the 5 per cent legal threshold and was replaced by the extreme right-wing party, the Hungarian Justice and Life Party (MIÉP). The MDF also failed to pass the legal threshold in 1998, but won 17 single-member district seats, 4.4 per cent of the total. MIÉP was unable to repeat its success in 2002, however, failing to win any single-member district races and falling under the legal threshold with 4.4 per cent of the party list vote. Indeed, the number of parties passing the legal threshold has declined from six in both 1990 and 1994 to five in 1998 and to only three in 2002. The 2002 elections saw

Hungary

365

a sharp break from the six-party parliament pattern with a consolidation of the party system around two blocs, the left-of-centre Socialists and their Free Democrat allies on the one hand and the centre-right Young Democrats Fidesz and populist MDF on the other hand (Fidesz added Hungarian Civic Party, MPP, to its name beginning with the 1998 elections). The importance of the single-member district tier to party fortunes is in promoting a shift from a multi-party system to a twoparty system. The effective number of parties captures this trend, with 3.8 in 1990, 2.9 in 1994, back up to 3.5 in 1998, and a sharp decrease to 2.2 in 2002. It is the psychological effects of the law on both party elites and the electorate that account for what might otherwise appear an odd combination of a decrease in the effective number of parties and a decrease in disproportionality. Strategic coordination by parties and voters has resulted in fewer parties and fewer votes ‘wasted’. Although the first elections of 1990 were free, few opposition parties save for those that participated in the National Round Table Talks were prepared for national elections, with 19 parties running territorial lists and 12 parties able to establish national lists, for a total of 231 lists. As result, the number of parties participating in the elections rose in the next cycle: 19 parties ran territorial lists, 15 parties stood a national list and the total list count was 271. Since 1994, however, the number of parties in the electorate has dropped steadily. The total number of lists run dropped to 218 in 1998 and to 155 in 2002. Whereas 15 parties ran territorial lists in 1998, 13 ran such lists in 2002. For national lists the number dropped from 12 parties in 1998 to only eight parties in 2002. Moreover, whereas the seven largest parties in 2002 established lists in every territorial district (20), four of the remaining six ran only one list and another ran just four. There is a similar trend in the single-member district tier. Despite the added incentives to stand candidates (the necessary requirement to establish party lists and the surplus votes pooled nationally), the number of parties running candidates shrank from 27 in 1998 to 20 in 2002 (18 if those parties only running joint candidates with a major party are excluded). This has not, however, opened up opportunities for independents: the number of independents contesting singlemember districts declined sharply from 199 in 1990, 103 in 1994 and 53 in 1998 to 40 in 2002, with six winning in 1990, one in 1998 and none in either 1994 or 2002 (Benoit, 2001: 490). Instead it is the large parties that dominate the singlemember district tier; the largest handful of parties that run 20 territorial lists and a national list also run candidates in virtually every single-member district race, except where there are joint candidates or agreements with an allied party not to contest. Voters too have adapted to the complex incentives of the electoral system. There is, for example, some evidence for split ticket voting, in which a voter casts a ballot for a candidate in his or her single-member district race that is not sponsored by the party choosen on the list ballot. Although the compensation mechanism drawing on unused single-member district votes does give some incentive for sincere voting in the single-member district ballot in the first round, such votes may still be perceived as wasted if the candidate’s party appears unlikely to make the legal threshold of 5 per cent (in which case losing single-member district votes do not

366

John W. Schiemann

contribute to the compensatory list and hence are irrelevant). The 5 per cent threshold also provides an incentive for supporters of smaller parties to vote for their secondary or tertiary choices on the party list, suggesting possible strategic voting here as well. The 2002 elections also represented a departure for voter turnout. Turnout was low in international comparison for the first three elections, including the founding election of 1990, and so low in 1998 that several list and single-member districts did not meet turnout requirements and had to repeat the election in round two. In contrast and exceeding all expectations, in 2002 71 per cent participated in the first round and 73 per cent participated in round two. In what is perhaps the most explicit evidence for voter and party adaptation to the electoral rules, the 1998 and 2002 elections witnessed widespread strategic coordination among parties and voters. In 1998, the Socialists led after the first round of voting in both the single-member districts and list balloting. Nevertheless, all but one of the single-member district races required a runoff round and prior to the runoff three of the smaller right-wing parties withdrew their candidates, urging their supporters to line up behind the largest two rightist parties, Young Fidesz and the populist MDF. There was less cooperation on the left, between the Socialists and the Free Democrats, and the result was that Fidesz was able to form a coalition government with the MDF and the smallholders FKgP. In 2002, much of the coordination on the part of the rightist parties took place before the balloting began, with the MDF and Fidesz running joint candidates in the first round. With fewer parties contesting the single-member districts, 45 races were decided in the first round, up sharply from only one in 1998. Perhaps having learned from their right-wing counterparts, the left coordinated fully in between rounds in 2002, with the Free Democrats SzDSz withdrawing the vast majority of its candidates from the runoff round in favour of a more powerful social-democratic MSzP candidate. In total, of the 131 runoff races, at least one candidate withdrew from 124 of them, reducing the contest to a choice between the two major blocs, FideszMDF on the right and the MSZP or SzDSz on the left. The importance of the single-member district races to the parties and the increasing dominance by the larger parties of the single-member districts has solidified party discipline in what was an already party-centric system. Moreover, whereas the first two parliaments witnessed party fragmentation on the right, as the Independent Smallholders split into rival factions and a group of MDF deputies split off and formed the Hungarian Justice and Life Party, parliaments since then have seen the consolidation of fragment parties into the two major blocs, auguring well for governing stability if not the range of alternatives available to Hungarian voters. The German-modelled constructive vote of no confidence also contributes to governing stability, to which the stability of multi-party coalition governments in 1990 and 1998 attests.

Conclusion The system’s increasingly predictable consequences and the difficulty of changing the electoral law combine to make it one of the most stable in Eastern Europe.

Hungary

Table 20.1 Year

1990 1994 1998 2002

367

Hungary: National Assembly electoral results, 1990–2002

Left

Centre-left

Centre

Right

Others

MP v/s

MSzP v/s

ASz v/s

Fidesz v/s

MDF v/s

SzDSz KDNP v/s v/s

FKgP v/s

MIÉP v/s

v/s

3/0 4/0 4/0 3/0

11/9 31/54 30/34 42/46

3/1 3/1 – –

5/6 7/5 26/44 –

24/43 12/10 3/1 41/49

22/24 18/18 10/6 5/5

11/12 7/7 13/12 1/0

– 1/0 6/3 4/0

15/0 10/0 5/0 4/0

6/5 7/5 3/0 –

Key: MP: Worker’s Party; MSzP: Hungarian Socialist Party; Asz: Agrarian Alliance; Fidesz: Federation of Young Democrats/Hungarian Civic Party; MDF: Hungarian Democratic Forum/Fidesz; SzDSz: Alliance of Free Democrats; KDNP: Christian Democratic People’s Party; FKgP: Independent Party of Smallholders, Agrarian Workers and Citizens; MIÉP: Hungarian Justice and Life Party; v/s: percentage of votes/percentage of seats.

The two-thirds supermajority requirement to amend the law makes it difficult to change its provisions. The parties that have an interest in changing the system usually do not have the votes to do so, while (at least some of) the parties that have the votes, do not have an interest in changing the system. Thus while there has been talk of reform at various points, most recently before the 2002 elections, the major political parties could not reach an agreement. Given the stability of the electoral rules and the appearance of effective coordination between voters and party elites, the Hungarian electoral and party system may be converging to a stable two-bloc equilibrium.

Notes 1. I thank Ken Benoit for our many conversations and our previous work together on the Hungarian electoral system as well as for his comments on this paper. Any errors are my own. 2. Interview with EKA expert Péter Tölgyessy, 3 July 1995. The total number of registered voters in these districts was 31,505 (District 30) and 31,746 (District 31), some ten to twenty thousand fewer than almost every other district in Budapest (Szoboszlai, 1990: 465–6).

References Benoit, Kenneth (2001) ‘Evaluating Hungary’s Mixed-Member Electoral System’, in Matthew S. Shugart and Martin P. Wattenberg (eds), Mixed-Member Electoral Systems: The Best of Both Worlds? New York and Oxford: Oxford University Press, pp. 477–93. Földes, György and László Hubai (eds) (1994) Parlamenti Képviseloválasztások, 1920–1990 [Parliamentary Represenative Elections, 1920–1990]. Budapest: Politikatörténeti Alapítvány. Horváth, Tamás. M. (1990) ‘Pesties választásikiskörkép [Electoral Overview of Pest]’ in Gyorgy Szoboszlai (ed.), Parlamenti Választások 1990 (Parliamentary Elections 1990). Budapest: MTA Társadalomtudományi Intézet, pp. 431–9. Szoboszlai, Gyorgy (1990) Parlamenti Választások 1990 [Parliamentary Elections 1990]. Budapest: MTA Társadalomtudományi Intézet.

368

John W. Schiemann

Vass, Henrik (ed.) (1994) A Magyar Szocialista Munkáspárt Határozatai és Dokumentumai, 1985–1989 [Resolutions and Documents of the Hungarian Socialist Workers’ Party, 1985–1989]. Budapest: Interart Stúdió.

Further reading • Benoit, Kenneth (2001) ‘Evaluating Hungary’s Mixed-Member Electoral System’, in Matthew S. Shugart and Martin P. Wattenberg (eds), Mixed-Member Electoral Systems: The Best of Both Worlds? New York and Oxford: Oxford University Press, pp. 477–93. • Benoit, Kenneth and John W. Schiemann (2001) ‘Institutional Choice in New Democracies: Bargaining Over Hungary’s 1989 Electoral Law’, Journal of Theoretical Politics, 13, 2: 153–82. • Schiemann, John W. (1999) Risk, Radicals, and Regime Change: Institutional Choice in Hungary, 1989. PhD thesis, Columbia University. • Schiemann, John W. (2001) ‘Hedging Against Uncertainty: Regime Change and the Origins of Hungary’s Mixed-Member System’, in Matthew S. Shugart and Martin P. Wattenberg (eds), Mixed-Member Electoral Systems: The Best of Both Worlds? New York and Oxford: Oxford University Press, pp. 231–54.

21 Poland: Learning to Manipulate Electoral Rules Marek M. Kaminski and Monika A. Nalepa1

When given the opportunity, politicians try to manipulate democratic institutions to their own advantage (Riker, 1986). In mature democracies, institutional constraints that prevent manipulation have evolved over time. In contrast, the political systems of young democracies are less restricted in this area and encourage electoral heresthetics. The most vulnerable parameters of parliamentary elections are district magnitudes and boundaries, the formulas for seat allocation, the ballot structure, the thresholds of exclusion and the timing of elections. This chapter describes the background, strategic motivations and political consequences of the 2001 reform of the electoral law to the Sejm (Lower House) in Poland. It was implemented merely a few months before the following elections. Prior to this reform, the electoral law had undergone three transformations. The first took place in 1989 during the round table bargaining between the communists and Solidarity over the procedures to be used in the first semi-free elections. Two subsequent reforms of the law took place in 1993 and 1997. The political consequences in all three cases were striking. We first describe the 1989 bargaining and the 1993 and 1997 reforms. The next section gives a background for the 2001 reform. It is followed by a detailed analysis of the reform and an evaluation of its political consequences. The final section considers the prospects for subsequent reforms.

Electoral engineering in Poland 1989–2000 The 1989 round table negotiations started seven years after the communist regime had crushed the Solidarity movement and forced it underground. The Martial Law economy was bleeding and the repeated attempts at its revitalization had failed. Experiencing a gradual economic slump accompanied by a series of Solidarity-led mass strikes in the late 1980s, the communist rulers decided to invite the opposition to take responsibility for the sinking ship of the economy. The opposition, however, resisted co-optation. Partially free elections – the first in the Soviet bloc – contestable on conditions mutually agreed to in round table negotiations were the price the communists had to pay to avoid a political disaster. Among the institutions of the future polity that were at stake during the 1989 negotiations, the electoral 369

370

Marek M. Kaminski and Monika A. Nalepa

law was crucial. The major actors at the bargaining table were the communist coalition and the Solidarity trade union. The compromised electoral law divided the Sejm’s seats into a 65 per cent part reserved for the communist coalition and a 35 per cent part contestable in free elections, using a simple majority runoff system in single-member districts. All 35 seats in the nationwide district were divided among the partners of the communist coalition. The size of the Sejm was set at 460 seats and has remained unchanged to this day. The 100-seat Senate elections were completely free. The voting rule employed was the limited vote with runoff, very close in its effects to majority runoff. Having the president elected by both houses was a final part of the round table arrangements. The 35 per cent of the Sejm’s seats open to free elections resulted from long and painful negotiations. No significantly lower or higher quota was acceptable by Solidarity or the communists, respectively. For both sides, the 331/3 per cent value was critical. It derived its importance from the qualified majorities necessary for overriding the Senate’s veto. With more than two-thirds of the Sejm’s seats reserved for the communist coalition, the latter would have won complete decision-making power regardless of the election results. It would have been able to choose a cabinet, elect a president, change the constitution and rule effectively even in the worst-case scenario of Solidarity’s winning all free-quota Sejm seats and all Senate seats. Such an arrangement was not acceptable to Solidarity since it would grant it merely token status in the future parliament and could lead to loss of support to more radical opposition groups. As Solidarity’s chief negotiator Geremek declared, ‘we needed [a slight margin above the qualified majority] as oxygen’ (Skorzynski, 1995: 244). The final compromise on 35 per cent of seats open to competition, with a two-thirds qualified majority rule for overriding the Senate’s veto, gave Solidarity a chance at winning the power to block constitutional change and to veto bills. Thus, it legitimized the trade union’s stamp on the round table agreements. Surprisingly, the second important feature of the electoral law, the rules of majority runoff for the Sejm and limited vote for the Senate instead of proportional representation, was neglected by Solidarity’s negotiators. However, various confidential documents of the Communist Party (PUWP) as well as strong statements from top communist politicians indicate that they approached this issue with the utmost scrutiny, giving it due attention at Politburo and Central Committee meetings (Reykowski, 1997; Urban, 1997). It is clear from the arguments presented that the communists found no proportional representation system acceptable. First, a prerequisite for list-based PR was the legalization of opposition parties, a move that they resisted vigorously (Kurski and Semka, 1993: 18). Recall that majority runoff elections are held between individuals not necessarily affiliated with parties, while party-list proportional representation elections are held by parties or party-like entities. Second, the communists believed their ‘secret weapon’ to be the candidates for the free quota recruited among popular actors, writers and TV celebrities who were ‘non-partisan friends of the communists’. During their private meetings the communist leaders would refer to these celebrity candidates as ‘monkeys’ (Urban, 1997: 48). The

Poland

371

communists expected their ‘monkeys’ to score considerably higher under the personal vote in majority runoff or limited vote laws than under a proportional representation list system. Reykowski, the chief communist electoral law designer, reports that his last-minute emergency project of a proportional representation electoral law for the Senate, which was presented at a secret meeting to the PUWP’s leader General Jaruzelski, was rejected for precisely the reasons given above. At that time, Reykowski correctly estimated that a limited vote would produce a disaster for the communists (Reykowski, 1997: 42–3). Solidarity masterfully used the tiny window of opportunity created by the 35 per cent quota and won all unreserved Sejm seats and 99 out of 100 Senate seats. The blocking power acquired by the trade union in parliament resulted in a ten-week stalemate, the breakdown of the communist coalition and the formation of the first non-communist cabinet, and helped to initiate the snowballing revolutions of 1989 in other Eastern European countries. Under an alternative proportional representation electoral law, the communists would have scored much higher. Most importantly, they would have preserved the winning power in all major parliamentary games, thus saving the system – at least for some time – from collapse (Kaminski, 1999). The round table compact specified that the new electoral law was valid only for the 1989 elections. Bargaining over new laws started in early 1990, soon after it became evident to the political actors that the communist order in Eastern Europe had been dismantled. Both the Sejm and the Senate electoral laws were at stake, though the final Senate law was only slightly modified from the 1989 version. While the debates over the new Senate electoral law rarely left the floor of the parliament, the Sejm law attracted the attention of a frontline player, President Walesa. His focus was on the extent of presidential control over the timing of elections and the National Electoral Commission. The way in which votes were translated into seats was of lesser importance. At some point, the President offered two options to parliament: one of them was less proportional than the status quo and the other was more proportional. The fight for the electoral law resulted in a major crisis, a series of presidential vetoes and, finally, a desperate attempt to change the electoral law via constitutional amendment. Threats to dissolve parliament and even to seek the aid of the military could be heard. The final vote overruling Walesa’s veto was taken under strong pressure to pass some sort of law at all. Between 10 May and 11 July, five major voting sessions dealt with new projects or presidential vetoes, and the sixth and last session dealt with the presidential project of a constitutional change affecting the new law. The final electoral law employed the Hare-Niemeyer quota method in regional districts and the SainteLaguë divisor method for the 69 seats elected in the nationwide district. The average district magnitude was relatively large and equal to 12.1. The barriers for entry and for seat allocation were low. There was no threshold for regional districts and the threshold for the nationwide district was equal to 5 per cent. The most important political consequence of the 1991 electoral law was the fragmentation of the Sejm. The new law did not encourage the parties who had

372

Marek M. Kaminski and Monika A. Nalepa

considered consolidation in early 1991 to coalesce. The effective number of parliamentary parties in the Sejm reached a record high of 10.7. The two years before the next parliamentary elections witnessed four cabinets constantly balancing on the brink of breakdown. A coalition of reformers quickly assembled. The first project of the Democratic Union (UD), the largest party, revived an early 1990 idea for a mixed proportionalmajoritarian electoral law, but was opposed by all remaining parties. The next, more realistic project attracted a larger group of prospective gainers. Finally, a coalition of seven parties emerged following a long period of fine-tuning and bargaining over various parameters of the law. It included both post-communists and fierce anti-communists, indicating that preferences for electoral laws are independent of ideology. The law was accepted by a comfortable majority in the Sejm on 15 April 1993. It was booed by smaller parties as a ‘jerky solution’, ‘electoral swindle’ and the ‘dictatorship of large parties’ (Chrusciak, 1999: 107–8). The Sejm’s coalition lacked winning power in the Senate and the project was returned with substantial amendments. Fortunately for the seven-party coalition, the Sejm overruled the Senate’s amendments and essentially restored the original proposal. The final vote took place on 28 May 1993, in the midst of significant turmoil. The fourth cabinet in two years had just lost a confidence vote, inviting President Walesa to dissolve the parliament the following day. Unexpectedly, the electoral law that was supposed to serve the remote 1995 elections was tested early, after just four months, in September 1993. The new electoral law was still proportional but all three major changes strongly favoured larger parties. Stricter eligibility requirements and thresholds ranging from 5 per cent for parties to 8 per cent for coalitions of parties excluded small players from the final seat allocation. The new electoral formula, d’Hondt, was more favourable to larger parties. Finally, the average district magnitude significantly decreased from 12.5 to 8.8 seats. Splits, mergers, frequent reshuffles of political identities and failed electoral coalition-building marked the period between May and September 1993. When the 1993 electoral law was voted in, there were 18 parliamentary caucuses or circles (small caucuses). The main parliamentary parties went to the elections under their established brand names. A few parties formed electoral coalitions or changed their names. Among the 15 electoral competitors, there was one new party, one trade union and two non-parliamentary parties. The short-term political consequences of the new law compared to the old were tremendous. In the Sejm, two post-communist parties commanded a comfortable majority of 65.9 per cent of the seats while winning only 35.8 per cent of the votes. The fragmented post-Solidarity right was miserably defeated. The distribution of seats under the previous law would have been dramatically different: the postcommunist parties would have fallen short of the Sejm majority and the cabinet would most likely have been drawn from the pool of post-Solidarity parties (Kaminski, Lissowski and Swistak 1998). Over the next years political parties actively tried to affect the electoral law but no successful electoral reform followed. The new Constitution of 1997 introduced

Poland

373

the clause of proportional representation in the Sejm. Despite some support among politicians and political commentators for a majoritarian formula, the constitutional clause made such a reform practically impossible to implement. The parties’ activity focused on tweaking the specific parameters of proportional representation. In 1997, an attempt to introduce a more proportional electoral law was blocked by an ad hoc coalition of the post-communist Alliance of the Democratic Left (SLD) and the anti-communist post-Solidarity coalition AWS. The next reform was introduced in 2001.

The electoral reform in Poland in 2001: background After a decade of hands-on experience, politicians in Poland and other Eastern European countries have learned a great deal about tailoring electoral laws to their own advantage. While their early attempts were not so successful, with time, their manipulative skills improved considerably (Kaminski, 2002). As one Polish electoral expert bluntly stated: There has been a phenomenal leap in terms of knowledge [between 1989 and 2000]. All the parties have experts now. There are many books published on these matters now . . . We have had ten years of experience. The [political] elite enjoys the rules of the game – electoral geography is a passion for some deputies. They are professional and they understand the game. (Raciborski, quoted in Benoit and Hayden, 2001) It had been anticipated that the elections of 2001 would bring another sharp turn in Polish politics. The post-communist SLD, the senior partner in the 1993–97 cabinet that lost the 1997 elections to the united post-Solidarity coalition AWS was expected to return to power again. During their tenure, the AWS and its cabinet coalition partner, the Freedom Union (UW), initiated four fundamental institutional reforms. The reforms dealt with education, the pension system, local administration and health care. The reforms were co-supervised by UW’s chairman Balcerowicz. He had been the celebrity mastermind of the Polish ‘Big Bang’ which initiated the economic transitions in Eastern Europe. While his early macroeconomic reforms had been successful, he was unable to repeat this success at the institutional level. The overwhelming short-run social costs of the ‘second step of transition’, coupled with the striking incompetence with which the reforms were implemented, heavily contributed to the decline in the AWS’s ratings. The ratings of the ruling AWS began their free-fall soon after the reform of local government had changed the constituency boundaries on 1 January 1999. The percentage of voters who were willing to cast their vote on the AWS gradually declined from over 30 at the end of 1998 to the low teens and even single-digits in early 2001 (PBS, 2002). While politically disastrous, this reform brought about an important consequence for the electoral reform that later turned out to the advantage of the rulers. Since the Constitution demanded that the district structure be reconciled

374

Marek M. Kaminski and Monika A. Nalepa

with administrative boundaries, the administrative reform implied a reform of the electoral law. Following the administrative reform, in mid-1999, the parties prepared five proposals to change the electoral law for the Sejm and three to alter the Senate electoral law. A Special Committee for Electoral Reform was quickly appointed in the Sejm to work out a compromise. Between October 1999 and April 2001, the Committee considered 16 different proposals to amend the electoral laws for both houses. The final proposal was similar to the failed project of 1997, when three small parties, the liberal UW, the peasants’ PSL and the labour union UP, had proposed an electoral reform beneficial for themselves; at that time, two parties with higher poll scores, the AWS and SLD, opted for the status quo formula and the status quo prevailed by one vote. Except for campaign finance reform, all major changes in the 2001 law concerned the votes-to-seats properties of the electoral law. The average district magnitude increased from 8.7 seats to 11.2 seats. The old d’Hondt formula of seat allocation in districts that favoured large parties was substituted by the modified Sainte-Laguë formula that promotes smaller parties. Finally, the nationwide list was eliminated (see Table 21.1).

Electoral heresthetics in 2001 Four parties played a major role in designing the new electoral law of 2001: the post-Solidarity AWS, the post-communist SLD, the peasants’ PSL and the liberal UW. The UW, PSL and SLD’s positions were consistent from the beginning of the electoral debate in 1999. The UW and PSL supported making the electoral law more favourable to smaller parties, while the SLD opposed the changes. The initial project of the AWS was d’Hondt-based and similar to the SLD’s. In late 2000 and early 2001, after steadily declining polls and the formation of a new party, Citizen

Table 21.1

Changes in the Polish Sejm electoral laws, 1993–2001 Nationwide district

Number of districts Seats in districts (total) Ballot Formula Nationwide thresholds: – Single party or committee – Coalition of parties – Minority group*

Regional districts

1993

1993

2001

1 69 Closed list d’Hondt

52 391 Open list d’Hondt

41 460 Open list Mod. Sainte-Laguë

7% 7% 0 or 7%

5% 8% 0 or 5%

5% 8% 0 or 5%

* A minority group could choose one threshold to be waived.

Poland

375

Platform (a spin-off from the AWS and UW) (PO), the AWS’s scores slipped to the levels of the PSL and UW and oscillated around 10 per cent (PBS, 2002). The new electoral perspective provided the AWS with incentives to join the coalition of reformers. A heated debate of the electoral reform committee in January made clear that party experts had strong opinions on the mechanical consequences of electoral laws: Krzysztof Janik (SLD): ‘We support d’Hondt . . . I know that [the reason for] your support for Sainte-Laguë is in the polls . . .’ Grzegorz Schreiber (AWS): ‘d’Hondt and small districts favor the strongest party . . . SLD.’ Olga Krzyzanowska (UW): ‘d’Hondt will make the opposition in the Sejm very weak.’ Ludwik Dorn (Movement for Defense of Respublica ROP-PC): ‘[Our simulations show that the modified Sainte-Laguë] gives the SLD nine seats more [than the original Sainte-Laguë]. Why are we giving the SLD such a gift of extra seats?’ (Wielowieyska, 2001) The votes cast for and against the new electoral law were neatly aligned with the simulated consequences of the new law. Among the four parties, the SLD opposed the electoral reform. Its expected loss from the reform was estimated at 7.3 per cent of all Sejm seats. The AWS, PSL and UW supported the reform and their mean estimated gains were equal to 3 per cent, 1.2 per cent and 1.1 per cent, respectively. The fifth major party, the PO, was launched around the time of parliamentary voting and both the size of its support and the content of its parliamentary representation were unclear. While the simulations do not report confidence intervals for the results, the direction of gains versus losses was identical in all three cases. In two scenarios, the UW did not pass the 5 per cent threshold and won no seats under both electoral laws. The discipline within the four parties during the voting session was remarkably high (see Table 21.2). The last move belonged to President Kwasniewski, the former chairman of the SLD. While the coalition of reformers did not have enough votes to override a possible presidential veto, it resorted to another trick. The leaders of the coalition petitioned the Constitutional Court, stating that the electoral law currently in force was unconstitutional since it did not match the new administrative boundaries. Since the hypothetical presidential veto would leave the old law as the only alternative, a constitutional stalemate would follow. In such a case, the parliament would not be able to pass a new law before the election deadline. Since the SLD-UP coalition’s rankings were considered to be at their peak, an electoral delay would have likely harmed it even more than the new electoral law. The president endorsed the new electoral law on 28 May, making clear, however, that ‘various procedural reasons could be used by the anti-SLD coalition to contest the election results under the incumbent electoral law’. He criticized the last-minute manipulation by the parties and the fact that the law was so influenced by party

376

Marek M. Kaminski and Monika A. Nalepa

Table 21.2

Poland: simulated and actual seat shares and electoral reform

Party

Estimated seat shares May 2001 Under 1993 law

Post-Comm. (SLD-UP) Citizen Platform (PO) Peasants (PSL) Solidarity (AWS) Liberal (UW) Labor Union (SA) Law and Justice (PiS) League Families (LPR)

54 21 14 6 0.4 – – –

Election September 2001

Under new 2001 law

Support for new law

Actual seat share

Under A28H

46 21 15 9 1 – – –

No (0–0–155) n. a. Yes (21–2–1) Yes (155–3–4) Yes (39–0–0) – – –

47 14 9 0 0 12 10 8

55 13 7 0 0 10 8 6

Note: Columns do not sum up to 100 per cent because smaller parties are not included. Support: numbers indicate votes (for-abstain-against) within a given party during the session on 7 March 2001. A28H refers to hypothetical electoral law with actual electoral vote shares. Sources: Estimated seats: two separate polls under various assumptions about coalition structures (N1 = 990, N2 = 1,071) in OBOP (2001a, 2001b); Support: Wyniki glosowania (2001); Seats under A28H: Kot (2003).

support in the polls. He reminded the voters that all three sponsors of the reform, the AWS, the UW and the PSL, used to support the d’Hondt method precisely for the same reasons for which they now opposed it. He also vowed that soon after the elections he would re-initiate d’Hondt as well as a provision banning changes in the electoral law within a year of elections. This provision ‘would prevent parties from writing the electoral law according to survey data’ (Pilczynski and Zdort, 2001; Dom, 2001). The Senate electoral reform generated much cooler emotions. The new law made few changes to the incumbent law beyond what was required by the administrative reform. The essentially majoritarian character of the old law was preserved.

Political consequences of electoral reform Before one considers the political consequences of a particular electoral law, one must juxtapose it against an alternative law or set of laws. Usually, finding a ‘closest competitor’ of the given law is not difficult. When the challenger electoral law wins, the closest competitor is the incumbent electoral law. If the challenger loses, it is considered the closest competitor. In the 2001 electoral reform, however, there is no unambiguous ‘closest competitor’, since following the administrative reform the incumbent law was perceived as unconstitutional. Thus the competing electoral law is defined somewhat arbitrarily here as being based on the so-called Amendment 28 and the d’Hondt electoral formula (A28H). Essentially, A28H reconciled the district structure under the old law with the new constitution and the administrative reform, while keeping the district magnitude small. Both Amendment 28 and d’Hondt were supported only by the SLD.

Poland

377

The usual caveat that must be repeated here is that, for the purpose of simulations, both the list of parties and the election vote shares are assumed unchanged under the A28H electoral law. Keeping this assumption in mind, the ‘consequences’ of the electoral reform mentioned below will be a shorthand for the ‘consequences of the actual electoral law as evaluated against the A28H electoral law under the described assumptions (see Table 21.2). The most important consequence of the electoral reform was that the SLD was pushed below the 50 per cent threshold of all Sejm seats. The simulated results for the A28H law show the SLD’s seat share to be 55.2 per cent. Kot (2003) shows that the effect of redistricting according to Amendment 28 alone was weaker than introducing d’Hondt alone. Under Amendment 28 alone, the SLD would have received 48.7 per cent of seats; under d’Hondt alone, it would have received 53.3 per cent. Other alternative components of the electoral law, such as lowering the threshold for coalitions to 5 per cent or preserving the nationwide list, were relatively less important. The seat share of roughly 55 per cent is of sufficient magnitude to ensure that minor coalition changes or small shifts in voter preferences would not disrupt the SLD’s majority. Consequently, the SLD formed the cabinet in coalition with the peasants’ party, the PSL. The post-election policies were affected by the agrarian agenda of the smaller partner. The PSL’s impact proved especially important for Poland’s accession negotiations with the European Union. After many stalemates, the tough position in the negotiations succeeded. The benefits granted to Polish and other Eastern European farmers by the European Union negotiators were greater than most political commentators had expected. The second important consequence of the new law was the fragmentation of the Sejm. First, the new law weakened the bonds holding the AWS coalition together and provided incentives both for splits and for new entries. The earliest entrant, the Citizens’ Platform (PO), had formed at the last stage of electoral reform. The other rightist entrants, Law and Justice (PiS) and the League of Polish Families (LPR), formed after the electoral law had been introduced. All three parties took many votes from the AWS and, to a smaller extent, from the UW. As a result, the two parties that jointly held a majority of Sejm seats in the previous elections, the AWS and the UW, won no seats in the 2001 elections. The increase in fragmentation can be assessed quantitatively only under the assumption that the votes and the list of parties would not change under an alternative law A28H. The effective number of parties under the A28H law versus the actual law increases from 2.88 to 3.6. Since 2.88 represents only the mechanical effect, it does not capture the impact of potential consolidation that could have occurred under the less proportional A28H law. Arguably, coalition-building would have reinforced the mechanical effect. The effective number of parties under A28H could have been closer to 2.5. The relative fragmentation of the Sejm can be interestingly contrasted with the consolidated Senate. Facing an unchanged, essentially majoritarian, electoral law and a single dominant party, the many post-Solidarity parties formed a ‘grand Solidarity coalition’. This coalition won 15 seats in the 100-seat Senate against

378

Marek M. Kaminski and Monika A. Nalepa

75 seats for the SLD. The corresponding effective number of Senate parties was equal to 1.7. An electoral reform may affect the disproportionality of elections. Again, only the mechanical effect of the new law can be assessed precisely. The disproportionality under the actual law was equal to 9.4 per cent. It was estimated to be 14.6 per cent for A28H. Both numbers are typical for democracies (for example, see Taagepera and Shugart, 1989: 106–7). Finally, a question can be asked whether the parties involved in the electoral reform were successful in reaching their goals. This problem of the ‘effectiveness of manipulation’ is too complex to be analysed at length here (see Kaminski, 2002, for a discussion). Kot (2003) showed that the parties very effectively identified the seat-maximizing option when the actual law was compared against the A28H competitor. Surprisingly, the parties failed dismally to identify their ‘ideal’ electoral law, i.e. the law that would give them the most seats. Only in exactly half of the cases would the parties’ ‘most preferred’ electoral law, according to preferences identified from party documents and Sejm transcripts, have given the party more seats than the actual law (see Kot, 2003, for a detailed discussion).

Conclusion: the prospects for subsequent reforms When Leszek Miller, the chairman of the triumphant SLD, retrospectively analysed all of the opportunities his party had for winning more seats, he focused on the electoral reform. Miller: [. . .] I doubt whether the SLD could win more seats [with a different campaign]. [The only way to increase our seat share] would be to restore the old electoral law. Under the old rules of converting votes into seats, there would be no problem [for the SLD to win the majority of seats]. Question: Will you try to change the electoral law then? Miller: If we only could, we would. But I wonder whether we could find a partner in the Sejm who would support us. While on other issues we can find support, in this case I see little hope. It’s a shame. Question: Is it nice to change the electoral law every four years? Miller: Was it nice to change the old law? It was pristine [. . .]. Miller’s scepticism was founded on the distribution of support at the election time. Except for the SLD, no party would gain seats under the A28H law. Moreover, no party had policy motivations to support the SLD. The party perceived to be closest to the SLD, the PSL, had strong incentives to oppose a potential reform. By supporting such a reform, the PSL would give a majority to the SLD and would therefore lose a chance of joining a cabinet. In fact, the PSL became the SLD’s partner in the coalition cabinet. Surprisingly, Miller’s scepticism proved unfounded. The polls showed some movement and the small populist Self-Defence (SA) steadily increased its ratings above its initial electoral vote share of 10.2 per cent. When in July 2002 the SA’s

Table 21.3

Poland: Sejm (Lower Chamber) electoral results, 1991–2001

Year

Socialist

1991 1993 1997 2001

Agrarians

Liberals

Solidarity

Right

Others

SLD/UP v/s

SA v/s

PSL v/s

UW v/s

UD v/s

KLD v/s

BBWR v/s

PO v/s

AWSP v/s

POC v/s

KPN v/s

UPR v/s

LPR v/s

PiS v/s

v/s

14/14 27/46 32/35 41/47

– – – 10/12

9/11 15/28 7/6 9/9

– – 13/13 3/0

12/14 11/16 – –

8/8 – – –

– 5/4 – –

– – – 13/14

16/18 5/0 34/44 6/0

9/10 – – –

8/10 6/4 – –

3/1 – – –

– – – 8/8

– – – 9/9

21/14 31/2 14/2 1/1

Key: SLD: Alliance of Democratic Left. UP: Union of Labour. SA: Self-Defence. PSL: Polish Peasants’ Party. UW: Freedom Union. UD: Democratic Union. KLD: Liberal Democratic Congress. BBWR: Non-Party Reform. PO: Citizens’ Platform. AWSP: Electoral Coalition Solidarity Action. POC: Civic Centre Alliance. KPN: Confederation of Independent Poland. UPR: Union of Political Realism. LPR: League of Polish Families. PiS: Law and Justice. v/s: percentage of votes/percentage of seats.

379

380

Marek M. Kaminski and Monika A. Nalepa

support jumped to a record 17 per cent, it joined the SLD to modify the new electoral law by reinstating the d’Hondt algorithm. Since then, the SA’s support has fallen back to 12–14 per cent in August–November 2002. At the end of 2002, there was again one large party, the SLD. There were five smaller parties ranging in support from 9 to 13 per cent that jointly held a majority of the seats in the Sejm (Frydrykiewicz, 2002; CBOS, 2002). After the most recent modification, the electoral law seems to be locked in for a while. It seems unlikely that any major anti-SLD modification could be passed before the following parliamentary elections. The SLD has both the blocking power in the Senate and a sympathetic president. No potential coalition of other parties has enough power to override a veto of any of these bodies. Only a large shift in voter preferences could bring about change.

Note 1. Acknowledgements are due to Josep Colomer, Matt Golder and an anonymous referee who provided helpful comments. The usual disclaimers apply.

References Benoit, Kenneth and Jacqueline Hayden (2001) Institutional Change and Persistence: the Origins and Evolution of Poland’s Electoral System 1989–2001. Paper presented at First General Conference, European Consortium for Political Research, University of Kent, Canterbury, 6–8 September. CBOS (2002) Preferencje partyjne w listopadzie [Party Preferences in November]. Warsaw: CBOS. Chrusciak, Ryszard (1999) System wyborczy i wybory w Polsce 1989–1998. Parlamentarne spory i dyskusje [Electoral System and Elections in Poland 1989–1998. Parliamentary Conflicts and Discussions]. Warsaw: Dom Wydawniczy Elipsa. Dom (2001) ‘Prezydent podpisal nowa ordynacje wyborcza’ [‘The President Signed the New Electoral Law]’, Gazeta Wyborcza, 4, 30. Frydrykiewicz, Filip (2002) ‘Prawo dla duzych’ [‘Law for the Big’], Rzeczpospolita, 20–21 July, p. 168. Kaminski, Marek M. (1999) ‘How Communism Could Have Been Saved. Formal Analysis of Electoral Bargaining in Poland in 1989’, Public Choice, 98, 1–2: 83–109. Kaminski, Marek M. (2001) ‘Coalitional Stability of Multi-Party Systems: Evidence from Poland’, American Journal of Political Science, 45, 2: 294–312. Kaminski, Marek M. (2002) ‘Do Parties Benefit from Electoral Manipulation? Electoral Laws and Heresthetics in Poland, 1989–1993’, Journal of Theoretical Politics, 14, 3: 325–59. Kaminski, Marek M., Grzegorz Lissowski and Piotr Swistak (1998) ‘The “Revival of Communism” or the Effect of Institutions? The 1993 Polish Parliamentary Elections’, Public Choice, 97, 3: 429–49. Kot, Michal (2003) ‘Uczenie sie demokracji. Ocena skutecznosci partii politycznych w trakcie gry o ordynacje wyborcza w 2001 roku’ [‘Learning Democracy. The Effectiveness of Political Parties in the Electoral Law Game’], Studia Socjologiczne, forthcoming . Kurski, Jaroslaw and Piotr Semka (1993) Lewy czenucowy [Left Stroke in June]. Warsaw: Editions Spotkania. OBOP (2001a) Preferencje partyjne Polakowpod koniec maja 2001 r [Party Preferences of the Poles in late May 2001]. Warsaw: OBOP. OBOP (2001b) Preferencje partyjne Polakow w polowie maja 2001 r [Party Preferences of the Poles in mid-May 2001]. Warsaw: OBOP.

Poland

381

PBS (2002) Archiwum Badan. Sopot: PBS. Pilczynski, Jerzy and Marcin Dominik Zdort (2001) ‘Wybory 23 wrzesnia’ [‘The Elections of 23 September’], Rzeczpospolita, 5 January. Reykowski, Janusz (1997) ‘Komentarz’ [‘Comment’], Studia Socjologiczne, 145, 2: 340–5. Riker, William (1986) The Art of Political Manipulation. New Haven, CT: Yale University Press. Skorzynski, Jan (1995) Wladza i opozycja 1985–1989 [Government and Opposition, 1985–1989]. Warsaw: Presspublica. Taagapera, Rein and Matthew Shugart (1989) Seats and Votes. The Effects and Determinants of Electoral Systems. New Haven, CT: Yale University Press. Urban, Jerzy (1997) ‘Komentarz’ [‘Comment’], Studia Socjologiczne, 145, 2: 46–50. Wielowieyska, Dominika (2001) ‘Kazdy ciagnie ordynacje w swoja strone’ [‘Everybody Pushes the Electoral Law to His Own Advantage’], Gazeta Wyborcza, 2, 16. Wyniki glosowania nr 75 – posiedzenie 103 (2001) [The Results of Voting nr. 75 – Session 103]. Warsaw: Sejm.

Further reading • Benoit, Kenneth and Jacqueline Hayden (2001) Institutional Change and Persistence: the Origins

• • • • •

and Evolution of Poland’s Electoral System 1989–2001. Paper presented at First General Conference, European Consortium for Political Research, University of Kent, Canterbury, 6–8 September. Kaminski, Marek M. (1999) ‘How Communism Could Have Been Saved. Formal Analysis of Electoral Bargaining in Poland in 1989’, Public Choice, 98, 1–2: 83–109. Kaminski, Marek M. (2001) ‘Coalitional Stability of Multi-Party Systems: Evidence from Poland’, American Journal of Political Science, 45, 2: 294–312. Kaminski, Marek M. (2002). ‘Do Parties Benefit from Electoral Manipulation? Electoral Laws and Heresthetics in Poland, 1989–1993’, Journal of Theoretical Politics, 14, 3: 325–59. Lewis, Paul (1990). ‘Non-competitive Elections and Regime Change: Poland 1989’, Parliamentary Affairs, 1: 90–107. Pelczynski, Zbigniew and Sergiusz Kowalski (1990) ‘Poland’, Electoral Studies, 9, 4: 346–54.

22 Russia: Compromising a Long-Lasting Transitional Formula Olga V. Shvetsova

It has been argued by many that when it comes to making decisions about political institutions, post-communist politicians act like politicians anywhere else and choose institutions strategically. A growing number of accounts indicate that the strategic calculus of the parties involved in institution-making determined the choice of post-communist electoral laws. As a rule, communists favoured singlemember districts, while pro-democracy opposition to the regime mostly preferred proportional representation. Where former totalitarian parties collapsed or were marginalized by the time of electoral design, new elites dominated the institutional process and selected proportional representation. Where the popular opposition was weaker or balanced by the residually strong communist parties, the outcome was one of compromise – mixed systems, with some parliamentary seats filled by proportional representation and others in single-member districts. Compared to this regional standard, Russian election law is an extreme case of ‘endogenous’ design, since it did not result from open institutional bargaining, but was unilaterally decreed into existence by President Yeltsin shortly after he dissolved the legislature by force in September 1993. Most observers concur that, at that moment, Yeltsin was not only free to impose any electoral regulation he deemed appropriate, but was even free to change his mind at any juncture and did so three times before the actual ballot took place in December of 1993. (The month following the decree of 21 September 1993, dissolving parliament and calling for the new elections, Yeltsin issued around 200 decrees (Remington, 2002; see also Belin and Orttung 1997).) Thus, among other things, the initial version of the Presidential Decree No. 1400 from 21 September 1993 provided for 270 single-member districts and 130 proportional representation (PR) seats, which in a few days was changed to 225–225, increasing not only the number of PR slots but also the size of parliament. While the basic outline of the electoral system has remained unchanged since 1993, this stability has been attained despite the unceasing efforts of the major players to reshape it to their momentary needs. The price of such stability has 382

Russia

383

been the achievement of compromises by yielding to the specific demands of interested players in the form of minor modifications of the rules. The consequence of this ongoing tactical meddling with election laws has been that most of it has turned into obstacles to the institutionalization of the party system and has encouraged continued political party fragmentation.

Invention of the original design for the Duma, 1993 From the Soviet period, Russia inherited the electoral system of single-member districts and majority rule with a second round, combined with the district turnout requirement to validate district election results. This was the mechanism applied in the 4 and 17–18 March 1990 elections to the 1,068-seat Congress of People’s Deputies of the Russian Federation, still within the USSR. Compared to Soviet elections, the novelty was the two-tier district system, where 168 seats were assigned to larger, but unequal in size, constituencies according to the country’s administrative division – constitutionally, Russia was an asymmetric federal republic: the 16 Russian Autonomous Republics were entitled to 5 territorial seats each, five Russian Autonomous Oblasts 2 seats each, and the other 94 units were granted 1 seat each. The remaining 900 seats were allocated in smaller territorial constituencies drawn to be approximately equal by population size. The Congress had unlimited legislative prerogatives but in practice performed only partial legislative functions, its most important purpose being the selection of the 376-member Supreme Soviet, a permanently convened professional body which effectively served as the country’s legislature until the stand-off with the president in September 1993. As alternative drafts of the new constitution were debated between 1991 and 1993, the need for having a directly elected professional legislature was a point of general consensus, but the subcommittee to the Constitutional Committee of the Supreme Soviet, where most if not all discussions of election law were conducted, was divided on the specifics of the electoral rule. The first attempt to introduce a new election law dates back to July 1992, but the draft was quickly rejected by the Supreme Soviet as poorly prepared (Levchik, 1994). With the next election not scheduled until 1995, electoral reform was not seen as a priority by the Supreme Soviet. Yet as the conflict between the Supreme Soviet and the president intensified in the beginning of 1993, Yeltsin’s supporters began to advocate for an election in 1994 or even 1993. Majorities of the Congress and Supreme Soviet rejected such calls and the electoral law became a point of political bargaining: a new election could not take place until there was the law. With time, both interest in the issue and disagreement within the committee of the Supreme Soviet responsible for the drafting were growing. Its original 1992 draft contained the mixed electoral system and by 1993 the members and experts of the committee were sharply divided over such a choice (Gelman, 1997). Those allied with the conservative majority of the Supreme Soviet advocated majoritarian elections, while the pro-Yeltsin group, prominently represented by Victor Sheinis, held onto the preference for a mixed system. The split was reflected in the two major drafts prepared by the committee in June of 1993. ‘Draft A’ provided for

384

Olga V. Shvetsova

single-member districts or two-member districts by plurality rule. ‘Draft B’ applied the options available through ‘Draft A’ to 70 per cent of the parliamentary seats, reserving 30 per cent of the seats to be filled by at-large proportional representation. Members and experts of the committee alike were arguing the merits of the two systems in terms of the need for elected representatives to advocate the ‘downto-earth’ parochial interests of their small constituencies versus their ability to rise above those interests and serve the entire nation. No arguments were publicly offered at the time about the possible implications of the election law for party system consolidation. There was little attention paid to the debates over election law in the media and many members of the Supreme Soviet seemed preoccupied with other matters. The law was viewed or presented as a technical matter for which an obvious solution should exist. As the communist deputy Sergei Baburin stated, in comparing elections with driving on the road, ‘We cannot have different views on traffic regulations, after all’ (Sevodnya, 1 June 1993, p. 2). As Yeltsin sought justifications for dissolving the legislature elected when Russia was still a part of Gorbachev’s Soviet Union, one of the popular slogans that surfaced conveyed that the independent Russia needed a ‘new’ parliament, elected according to ‘new’ democratic rules. For that reason alone, the departure from the old Soviet single-member district model was expected from Yeltsin. In addition, according to the proponents of the mixed electoral system, voting for candidates in constituencies could give too much weight to the local leaders, many of whom were former communists and opponents of the president (Victor Sheinis, interview in The Guardian, 8 October 1993, p. 12). On the other side of the argument, Yeltsin’s aides believed that a parliament elected in single-member districts would be easier to control and thus more supportive of the president. Ever since the 1990 election, it also became very clear that the single-member districts in the capital cities of Moscow and St Petersburg were overcrowded with big-name politicians and party lists could provide relief as far as the readmission of Moscow-based incumbents into the new parliament was pursued. The parallel mixed system was a ‘safe’ reform, since politicians could run, as before, in districts and also on party lists. Indeed, 96 per cent of the single-member district candidates nominated by the parties in 1993 were also included in their parties’ national proportional representation lists (Colton and Hough, 1998). The standoff in September–October of 1993 between Yeltsin and the Supreme Soviet finally put an end to the sluggish deliberation on election law and the new constitution. Upon prevailing by force over the legislature, Yeltsin unilaterally suspended the old constitution and decreed that the new parliament was to be elected in less than three months’ time (Presidential Decree No. 1400, 21 September 1993). The protracted bargaining on the subject that took place prior to the constitutional crisis supplied information about the preferences of other players in the political system but had no binding influence on Yeltsin’s decision. According to Remington and Smith (1996), the president’s decree reflected the provisions of the draft favoured by pro-presidential parties because Victor Sheinis was advising him on this matter. On Sheinis’s recommendation, the share of proportional representation seats was set even higher than in the most ‘proportional’ committee draft.

Russia

385

Another manipulation that took place was in regard to the electoral rule in territorial districts. On this subject, it was not Sheinis but the leaders of the right parties, especially of the presidential favourite, Russia’s Choice, that insisted on replacing majority runoff with simple plurality, believing that the double ballot in singlemember districts would advantage centrist and moderate candidates rather than more extreme pro-reformist parties such as Russia’s Choice. Yet another opportunistically introduced technical detail was that, in order to reduce the singlemember district chances of the communists, Yeltsin’s decree prohibited the display of candidates’ party affiliations on the single-member district ballot. These two measures, plurality elections and candidates running without party labels, have greatly contributed to the electoral fragmentation which was set in motion by the mixed rule.

Unintended effects on party development The parallel mixed system had the effect of lifting the constraints on party proliferation that each of its two component systems could impose. In proportional representation, the legal threshold deters excessive fragmentation since anything less than 5 per cent of the vote translates into zero seats for any of the party’s candidates, whether high or low on its list. According to Article 70 of the electoral law, proportional representation mandates are distributed by the largest remainders system, using the Hare quota. The Central Electoral Commission adds up the votes cast for parties crossing the 5 per cent barrier and divides that number by 225 to calculate the quota. Unclaimed mandates go to the parties with the largest remainders. Coalitions and even mergers among small organizations are therefore encouraged. In single-member districts, the campaigns of individual candidates receive a boost if organizations form that can combine such candidates across districts to undertake a nationwide campaign, which in theory at least leads to the forming of two national coalitions of candidates. In a parallel mixed system, however, the threshold loss can be circumvented for the party leadership by a district victory of a personal seat. Election after election, additional party labels succeeded in entering the Duma via single-member districts, where their leaders won a seat each while they sank below the 5 per cent barrier in PR.1 Single ballot in single-member districts became another fragmenting influence, since it meant that non-communist voters were under pressure to instantly coordinate in their districts support for the same ‘democratic’ candidate in order to defeat the front-runners among the left who were more clearly identified. District nominations resembled the game of chicken, where numerous democratic candidates expected others to withdraw from the race for the sake of preventing the communist victory. Failure of democratic coordination was near universal and supporters’ votes were simply wasted in many cases. Not only did candidates of different reformist parties compete with each other, both in the 1993 and 1995 elections, there were districts where two candidates of Russia’s Choice ran against each other. On average, there were 6.8 candidates per district in 1993, 12.1 in 1995

386

Olga V. Shvetsova

and 10 in 1999. While it is not that unusual to see so many contestants in district races in any country that uses single-member districts, the specifics of transitional Russia were that most candidates received meaningful shares of the district vote. The average number of ‘effective candidates’ per district was 5.5 in 1993, 6.6 in 1995 and 5.6 in 1999. Such fragmentation of single-member district ballots accounts for the fact that the average vote for the winner in single-member districts was around 30 per cent. Furthermore, listed on the ballot without party affiliations, single-member district candidates were both allowed and encouraged to run as ‘independent’ from the proportional representation parties, while fragmentation of the district ballot by party candidates made it easier for independents (and ranking members of minor parties) to win. In 1993, 667 single-member district candidates were nominated by parties, while another 822 ran as independents and won 136 out of 225 seats. Overall, the parallel mixed system seems to encourage party consolidation much less than either of its ‘pure’ parts. At the same time, the prior level of institutionalization of individual parties modified institutional influence. The persistent and stable support of the ‘old’ Communist Party, in contrast with great fluidity in both labelling and voting among parties of the right and centre-right, led to the opposite dynamics with regard to the party system development. The communists were equally successful in both parts of the ballot, which was not true for any other political organization. The advantage of prior institutionalization has also protected the Communist Party from any noteworthy challenge or internal split and prevented the emergence of a credible alternative left party (or parties). With regard to the political left of modern-day Russia, the mixed electoral system has served to preserve the status quo, a one-party electoral and political dominance. The election law’s impact on the political right and centre was different, leading to contrary results. The key distinction involves the starting point of no institutionalized parties, allowing for a rapid introduction of many new labels. This subset of the party system was characterized by easy entry with new entrants achieving ‘victory’ in each of the three elections, as did the Liberal Democratic Party of Russia (LDPR) with 23 per cent of the PR vote in 1993, Our Home Is Russia (NDR) with 10 per cent of PR in 1995 and Yedinstvo (Unity) with 23 per cent of PR in 1999. Another typical feature was the inability of new labels to hold onto their initial success, as was the case with the decline of the LDPR and the practical disappearance of the NDR. The diverse effect of the same electoral procedure on the two flanks of the party system is noteworthy and highlights the theoretical significance of party institutionalization for resolving the voter coordination problem, not just in single-member districts but in proportional representation as well, that was resolved for the left but not for the right. In light of the voter coordination problem, control of national television and press (central and regional newspapers) was essential in determining the winner in the centre-right.2 This is why strong financial supporters and/or incumbent regional officeholders were behind the success of Our Home Is Russia and Unity. Empowered to swing the vote due to the oversupply of parties

Russia

387

and the problem of unresolved voter coordination, these financial and governmental backers, in turn, further perpetuated the weaknesses of the party system.

Institutional strength of the president and revision of election laws It is a sign of the prominence in Russia of the institutional conflict between the strong presidency and a weak parliament that all cases of implemented changes in the parliamentary election law in the period 1993–2003 were initiated by the presidents. Even though the main features of the Duma election law remained unchanged since 1993, this is not for lack of trying on the part of the presidential administration. At times, it seems only miracles have prevented major alterations, while many ‘technical’ adjustments in the registration and seat allocation requirements have been adopted and have noticeably modified the incentives for the political actors. Elections to the Duma, in reality, are regulated by several complementary legal acts, of which the most important are the Federal Law on the Election of Deputies of the State Duma of the Russian Federation, the Federal Law on the Basic Guarantees of the Electoral Rights and the Federal Law on Political Parties. When one of these is changed, the other two need to be adjusted as well, giving the president the leverage in pushing for his preferred amendments. This is so because, once modifications in one of these interrelated documents invalidate the others, the president can issue a decree on the subject unless the new Duma laws are passed. Meanwhile, he can also veto any such Duma law, and his veto cannot be overridden by the Duma alone but also requires a two-thirds vote in the Federation Council. The president’s wishes, then, are hard to ignore and the adoption of new parliamentary versions of the law tends to be postponed until it becomes absolutely necessary to compromise. The ‘usual’ scenario is as follows: the old version of the electoral law becomes outdated due to a conflict with other legislative acts when those are amended for ‘technical’ reasons; the president gets a bargaining advantage; there is always a legislative deadlock over acceptable revisions and thus an opportunity for the president to use the threat of a decree. Through this mechanism, new versions of the election law were introduced and voted before each Duma election in 1995, 1999 and 2003. In all three cases, new election laws were adopted late in the electoral cycle – less than a year before the next election (in 1995 and 1999, three months in advance of the scheduled ballot).

The fight for the single-member districts and the ‘federalization’ of party lists When the president actively manipulates the legislative agenda to gain the upper hand in determining the election law, and the parliament responds by strategizing and manoeuvring, ‘technical’ compromises may have far-reaching and unforeseen consequences. The ‘recognition’ of the federal character of elections and an associated requirement where party lists must reflect that federal character is a case in point. It emerged as a by-product of the president’s heresthetical manipulation

388

Olga V. Shvetsova

intended to re-institute single-member districts, and it has proved to have profound consequences for party system development. In 1994, several Russian politicians and academics close to the president suggested that Russia should move toward a two-party system and, therefore, toward a more a majoritarian electoral system. On 1 November 1994 Yeltsin officially introduced a draft of a new electoral law, which aimed to increase the number of single-member district seats to 300 and reduce the number of proportional representation seats to 150. This bill also prohibited placing single-member district candidates on proportional representation lists, thus making party affiliation even less useful for local politicians. Practically all incumbent parliamentary parties viewed these two proposals as unacceptable, and the Duma passed a version of the law that preserved the equal division between proportional representation and single-member district seats and safeguarded the opportunity for parties to run the same candidates on both ballots. Yeltsin, however, had a powerful and interested institutional ally, the Council of the Federation, which went on to veto the Duma bill several times in a row, showing its backing of the majoritarian system as giving the regions a greater voice in selecting the representatives, while proportional representation lists were dominated by Moscow-based parties and Moscow politicians. As long as both Yeltsin and the Council of the Federation insisted on rejecting the 1993 electoral system, there would be no new election law. Meanwhile, without a new law, Yeltsin had the right to issue a decree regulating the forthcoming election. A solution to the problem was found when the incumbent Duma parties proposed, as a means of getting support from the Council of the Federation, to recognize the federal character of the election by requiring that party lists reflect that federal character. Apparently, this satisfied the upper chamber, since five days after the Duma vote on these revisions, it voted in support. There was also a promise made by the Duma leadership to the Federation Council to pass a law that would preserve direct elections to the upper chamber. This law was passed but later vetoed by the president. According to the new provisions, each party on the proportional representation ballot had to register at least two separate ‘territorial’ lists which would divide party seats in proportion to their regional contribution to the party vote. As a nod to Moscow politicians, each party was also allowed a ‘federal’ list of 12 names which had precedence over territorial lists in receiving seats. This internal ‘technical’ change to the party list compilation requirements drastically altered the seat allocation incentives to proportional representation candidates. For the top 12 positions, it guaranteed seats if a party got any seats at all: when the 5 per cent threshold requirement was applied in the 225-seat proportional representation district, the clearing of the threshold by a party automatically meant at least 12 seats – enough for those on its federal list. Beyond the federal list, there was a great discontinuity in the prospects of candidates. The near-certainty of winning for the number 12 was followed by the substantially lowered chances for the former candidate number 13, now at the top of one of the territorial lists. Should there be seats left in the party’s coffer in excess of 12, ‘territorial’ lists would receive them in proportion to

Russia

389

their regional vote, using the Hare largest remainders method. In practice, even with a small number of lists but few seats left to be divided, much if not most of this allocation would be based on the remainders, not the quotas. Thus, while the 12th position on the national list got a seat if his party had passed the 5 per cent threshold, it was difficult to predict who would get the next seats, such as the 13th and so on, because this would be determined by the exact distribution of a party’s vote across its self-defined regions. Therefore the chances of getting into the Duma had to be very severely discounted even for top positions on the regional lists. In undermining the incentives for local members, the federal list arrangement cemented the initial bias in favour of party elites based in capital cities and promoted continued political fragmentation. The immediate consequences of this new ‘technical’ provision consisted of squabbles in Moscow for placement onto ‘federal’ lists, organizational splits (which immediately allowed for the number of valuable ‘federal’ slots to multiply) that increased the supply of parties of the right and centre, and the universal adoption of the campaign strategy of mass advertising at the expense of attempting to develop grassroots organizations and local infrastructures. As a result, party fragmentation in the election of 1995 led to an unprecedented level of vote loss in PR, where 49.5 per cent of all valid votes were given to parties that failed to meet threshold requirements. Nevertheless, 21 party labels were represented in the new parliament as the leaders of defeated parties won seats for themselves in territorial constituencies. The federal list provision was amended going into the 1999 election. Far from being abolished, federal lists were expanded to 18 positions. This change could be expected to curb party splits, but not to promote increased organizational presence and cadre development at the subnational and local levels.3 In 1997, the Duma approved the revised version of the Basic Guarantees Law proposed by the Presidential Administration (to address the shortcomings revealed during the 1995 parliamentary and the 1996 presidential campaigns), with very little deliberation. But changes to that Law automatically invalidated the old Law on the Election of Duma Deputies, and Yeltsin asked the Duma again to use the occasion of the (unavoidable) revision of the electoral law for changing the mixed system into majority runoff in single-member districts. The Duma could do nothing but stall for time, and the reform did not go through because, by 1999, Yeltsin has changed his mind on the issue, now fearing the powers of regional leaders rather than looking to increase them. The next, third round was initiated in December 2000, when president Putin started by introducing the new Law on Political Parties (enacted in the summer of 2001). On the basis of that law, it became necessary to change the Law on the Basic Guarantees (in the summer of 2002),4 and that, in turn, required that the Law on the Election would be revised (in December of 2002).

Federalism and political incentives at the subnational level Until 1999, the Kremlin’s goals for electoral reform were to strengthen the pro-presidential parties and at the same time to weaken and fragment political parties

390

Olga V. Shvetsova

in general. The pure single-member district format that it pursued was likely to produce an even more fragmented and, most importantly, non-partisan parliament, with weaker factional discipline. Yet in early 1999, the Kremlin concerns shifted from the Duma to the growing political influence of the regional leaders assembled in the Council of the Federation, who started forming competing political coalitions for the forthcoming parliamentary and presidential elections. Naturally, it was easier for such coalitions to win parliamentary seats through single-member districts than through the nationwide PR, and the Council of the Federation, at that time made up of the regional bosses, consistently advocated majoritarian reform. Yet, precisely because a majority electoral system was attractive for the regional leaders, the Kremlin ceased supporting it, and to reduce the political influence of regional governments became an institutional reform priority for President Putin. First, regional leaders lost their seats in the Council of the Federation. Second, Putin insisted on formally restricting access to the ballot, leaving only nationwide parties eligible to run in all levels of elections (excepting only local self-government). Thus no parties can be established at the subnational level. Moreover, not only did the Kremlin turn to supporting proportional representation for federal elections, but it also pushed for the same in the regions, so that now, by a federal law, regional elections must be either by a mixed or a proportional representation system, while only nationwide parties can participate in PR. This undermines the control of regional incumbents over the regional elections. At the same time, the revised electoral regulations preserve and even strengthen incentives for individual politicians to continue to fragment the party system. Thus the (2001) Law on Political Parties, by instituting a complex and lengthy registration process, increased the cost of entry into electoral competition and thus the value of existing parties as legal entities, pushing politicians to preserve party organizations they control and further avoid risky mergers. And the Law made it practically compulsory for all existing political parties to participate in the Duma elections, since if they fail to do so, they lose their party status and eligibility for future elections.5 Furthermore, the Law on Guarantees limits to three the number of parties and other organizations that can form an electoral bloc.

Interactive electoral effects with the presidency The logic of electoral calculations has kept both Russian presidents to date (Yeltsin, 1991–99, and Putin, since 1999 as acting president and elected since 2000) from doing more than merely endorsing parties in Duma elections. Institutionally generated incentives force a president to often ignore and act against partisanship. With a party system fragmented, it is unlikely that any one party can hold a majority of votes and seats. Close association with a party that is far from representing a majority in the electorate, in turn, erodes the claim to national legitimacy that the majoritarian presidential election otherwise establishes for its winner. Furthermore, the non-concurrent timing of elections for the parliament and presidency offers the strong Russian president an opportunity to strategically choose and change his allies in the Duma without making any commitments.

Russia

391

Finally, the structure of the electorate’s preferences favours party-less presidential candidates. With about one-third of the voters customarily supporting the left (communist) candidate, the presidency is de facto decided in the first round by plurality among the non-communist contenders. This is so because: (a) with a third of the vote, a communist candidate is assured a place in the run-off and only one other spot is left for the rest to contest; and (b) in the runoff, non-communist voters will vote for any non-communist participant. Since in the past two presidential elections it was the near certainty that in a presidential run-off there would be a communist and a reformer, and that the reformer would most probably win, the party system became particularly sensitive to the interaction between presidential and parliamentary electoral incentives. An important presidency-related parameter, which the literature finds influences party development alongside the laws of parliamentary election, is the timing of the presidential race (Shugart and Carey, 1992). In Russia, elections for the Duma and the presidency were staggered six months apart (now three months apart as the resignation of Yeltsin shifted the schedule), with elections for parliament ending just as the campaign for the presidency commences. With parliamentary election rules promoting fragmentation and fluidity in the party system and no party clearly a leader in the reform bloc, the choice of election timing has turned the campaigns for parliament into primaries for reformist frontrunners in the presidential race. The situation with the presidency leads to the expectation that whoever is the top reformer in the Duma race will have a superior chance of becoming the next president. The Duma election of 1995 fragmented the right too much and failed to establish a clear leader. Essentially, all parties of the right have ‘lost’. It is in part for this reason that Yeltsin’s flailing incumbency (at 5 per cent popularity immediately after the parliamentary election of 1995) had to be revamped for the 1996 presidential race where he won in the second round. In 1999, with Yeltsin’s last term coming to an end, the stakes were particularly high and the ‘primaries’ effect was particularly damaging to the party system. Pushing aside all previously existing parties of the right, two newcomers dominated the centre and centre-right, Fatherland-All Russia (OVR) and Unity (see Table 22.1). Both parties were primarily identified by those they would have backed for the presidency and not by their policy programmes.

The Council of the Federation The upper chamber of the Russian parliament, the Council of the Federation, was elected directly only once, in 1993, and the president vetoed the 1995 law that would have made it elected directly on a permanent basis. In 1993, each of the 89 Russian Federation regions elected two members to the Council of the Federation in two-member districts, and to hold such elections was simply another strategic move by Yeltsin who was looking to increase the stakes for the local bosses in the campaign of 1993. Strong local leaders were assured at least a second place in a two-seat local vote, but being in the race forced them to actively participate in the organization of elections, which were linked that year with the constitutional

Russian Assembly elections, 1993–99

392

Table 22.1

1993 Party

Communist KPRF Zhirinovsky LDPR Yabloko Women of Russia Agrarian Russia’s Choice PRES DPR RDDR Civic Union Future of Russia KEDR Dignity Compassion Home Russia NDR Communities KRO Communist for USSR Workers’ Self-Government Democratic Choice Derzhava Forward, Russia! Power to the People! Pamfilova-Gurov-Lysenko Trade Unions Industrialists Kedr Ivan Rybkin Bloc S. Govorukhin Bloc

1995

Party list

SMD

1999

Party list

SMD

Party list

SMD

% vote

Seats

% vote

Seats

% vote

Seats

% vote

Seats

% vote

Seats

% vote

Seats

12 23 8 8 8 15

32 59 20 21 21 40 7 5 4 2 – 1 1

4 4 4 1 6 8 18 14 – – 1 – –

9 5 7 2 16 25 3 2 2 4 1 1 1

22 11 7 5 4

99 50 31 – –

14 6 4 1 7

58 1 14 3 20

24 6 6 2

67 17 16 –

13 2 5

47 0 4 –

45 – – – – – – – – – – – –

6 3 2 1 3 – 2 2 1 1 – 2 1

10 5 1 1 9 – 3 9 2 1 – 3 1

1 1 2

3 – –

8 21 –

1

3 1 5 6 – 3 10 4 4 4 4 3 2 2 2 2 1 1 1

My Fatherland Common Cause Transformation Fatherland Unity and Accord Economic Freedom 89 Regions Inter-Regional Fatherland OVR Union of Right SPS Stalinists for USSR Support of the Army All People’s Union (ROS) Pensioners Nikolayev & Fyodorov Spiritual Heritage (DN) Socialist Party RSP Civil Dignity Peace, Labor, May Peace and Unity Protection Women Russian Cause RD Conservative People’s Socialist SPR Social-Democrats Other lists

1 1 0.5 0.4 0.1 0.1

4

– – – – – –

1 0.2 0.4 0.5 0.3 0.3

1 1 1 1 1 1 23 13 9 1 1 0.4 2 1 0.1 0.2 0.6 0.6 0.4 1 0.2 0.1 0.1 0.1 0.1 2

64 37 24 – – – – – – – – – – – – – – – – 0

2 4 3

9 29 5

1 1 1 1 1 3

2 2 1 1 1 1 – – – – – – – – –

393

394

Table 22.1

(Continued) 1993

Party

Party list % vote

Against all lists All parties Independents Number of lists: – petitioned – registered – winning PR seats Registered voters Turnout Valid votes

4 412

Seats

83 58

1995 SMD

% vote

136

35 13 8 106,170,835 57,679,698 53,751,696

1999

Party list Seats

% vote

SMD

Party list

Seats

% vote

Seats

3

–

–

–

3

–

0.1

–

0.6

1

43

105

107,496,856 69,204,819 67,884,200

Seats

% vote

Seats –

35 26 6

69 43 4 (54.3%)

% vote

SMD

(64.4 %)

108,073,956 66,667,682 65,375,690

(61.7 %)

Sources: PR 1993: Central Electoral Commission (TsIK) Bulletin No.1 (12), 1996. PR 1995 and 1999: Website of the Central Electoral Commission <www.fci.ru>. SMD 1993: Colton and Hough (1998), based on unpublished final listing by the Central Electoral Commission. These results are based on candidate nomination data and not on the party affiliation adopted by elected representatives after the ballot. The list of winners and their initial party affiliation was published in Rossiiskaya Gazeta, 28 December 1993, pp. 2–6 (reproduced in Lester, 1995). SMD 1995 and 1999: Website of the Central Electoral Commission <www.fci.ru>.

Russia

395

referendum. In order to legitimize his constitution, Yeltsin needed for the turnout in the referendum/election to exceed 50 per cent and mobilization efforts by local bosses were supposed to contribute to that goal. According to the constitution, from 1995 onwards, the parliament’s upper chamber was to be formed by legislative and executive branches of the 89 federation subjects – the regions and republics. However, this provision is open to interpretation. The Council of the Federation has the right to veto all legislation approved by the state Duma, and a two-thirds supermajority in the Duma is required to override its veto. The upper chamber must also participate in the overriding of all presidential vetoes with a two-thirds majority. Consequently, conciliatory commissions between the lower and upper legislative chambers have become a routine stage in developing legislation. The Council of the Federation also approves presidential nominations to the Constitutional Court and participates in the constitutional amendment process. Given the significance of the Federation Council in the legislative process and in passing constitutional amendments, differences in interpretation could become politically profitable. The interpretation that prevailed during the Yeltsin presidency was that heads of the regional executive and legislative branches automatically held regional seats in the Council. The onset of Putin’s administration (March 2000) produced a major shift in the interpretation, where appointees of regional officeholders, rather than the elected officeholders themselves, became members of the upper chamber and served for the duration of the term of the person or body that appointed them. Formally justified by the high time demands precluding concurrent service in two offices, this interpretation established a peculiar agency relationship between the Council members and those in the regions who appointed them. In the Putin interpretation, the Federation Council threatens to become a body of professional lobbyists retained by regional governments providing very little in the way of a two-way link connecting federal and regional political processes, with no promise of contributing anything to the development of integrated national political parties and strongly influenced by the executive. Both Yeltsin and Putin, through the design of institutions and by raising direct objections, stood in the way of any partisan developments being implemented in the upper chamber. Such a behavioural pattern by presidents is easy to rationalize within the constitutional framework where the strength of the executive is largely contingent upon the ability to sustain its vetoes over legislative decisions. Although not too powerful a player in the legislative process overall, the Federation Council must participate in any veto override. For this reason, the weaker its partisan links with Duma majorities, the stronger is the legislative prerogative of the Russian president.

Notes 1. In November of 1998, the Constitutional Court ruled that the threshold requirement, though it did not violate the Constitution, was inapplicable if the electoral associations which overcame this barrier together did not receive an absolute majority of the vote or if

396

2.

3.

4.

5.

Olga V. Shvetsova

all seats were allocated to one party. Thus, regardless of the vote fragmentation, parties must be given proportional representation seats up to the point where seat recipients together account for over a half of the vote. An unofficial translation of the decision is available at . In fulfilment of this court decision, the 1999 version of the electoral legislation made the threshold ‘flexible’ – if it cuts off too many parties, it must be lowered until the 50 per cent and more than one party conditions are both met. The 2002 election law, furthermore, mandates that after the election of 2003 the threshold of 7 per cent will be imposed but will be similarly flexible, with at least four parties with at least 50 per cent of the popular vote to be given proportional representation seats, regardless of whether they meet the threshold (except if all fall below 7 per cent, in which case the election is invalidated). The initial success of the LDPR is an outlier in this pattern. In the brief campaign of 1993, advertisements and other campaign resources were not mobilized fast enough, and the focal effect was attained by the LDPR because the rhetoric of its founder and charismatic leader Vladimir Zhirinovsky was calculated to be both so outrageous and so irrelevant for any possible political implementation that it was naturally adopted as the newsworthy subject by the media smarting from the parliamentary–presidential stand-off. For the election of 2003, the size of the federal list remained limited to 18 places, while the minimum required number of regional lists that a party or electoral bloc must present is seven. It is worth noting that the first (1994) Basic Guarantees Law simply set the general principles of elections and was a ‘framework’ law for other legislative acts. Yet its latest version covers every possible detail of the electoral process and is 270 double-spaced pages long. A party which did not participate in elections for five years is subject to dissolution. Parties must participate in one of the following types of elections: run a proportional representation list or at least 12 single-member district candidates in the parliamentary election; nominate a presidential candidate; nominate at least 9 candidates for regional governorships or participate in at least 18 regional legislative elections.

References Belin, Laura and Robert W. Orttung (eds) (1997) The Russian Parliamentary Elections of 1995: The Battle for the Duma. Armonk: M. E. Sharpe. Colton, Timothy J. and Jerry F. Hough (eds) (1998) Growing Pains: Russian Democracy and the Election of 1993. Washington, DC: Brookings Institution. Gelman, V. (1997) ‘Sozdavaja Pravila Igry: Rossiskoe Izbiratelnoe Zakonodatelstvo Perehodnogo Perioda’, Polis, 4: 125–47. Levchik, D. (1994) ‘Osnovnye Elementy Elektoralnogo Prava i ih Primenenie v Rossiskom Zakonodatelstve 1992–1994’, Kentavr, 5: 105–14. Remington, Thomas F. (2002) Democratization and the Problem of Governance. Paper presented at Havighurst Symposium in Economics ‘Russia Ten Years Later: Taking Stock’, University of Miami, 7–8 April 2002. Remington, Thomas F. and Steven S. Smith (1996) ‘Political Goals, Institutional Context, and the Choice of an Electoral System: the Russian Parliamentary Election Law’, American Journal of Political Science, 40, 4: 1253–79. Shugart, Matthew S. and John M. Carey (1992) Presidents and Assemblies. Constitutional Design and Electoral Dynamics. Cambridge and New York: Cambridge University Press.

Further reading • Dawisha, Karen and Stephen Deets (2001) Political Learning in Postcommunist Elections. Paper presented at the 2001 American Political Science Association, Annual Meeting, San Francisco.

Russia

397

• Dunlop, John B. (1993) The Rise of Russia and the Fall of the Soviet Empire. Princeton, NJ: Princeton University Press.

• McFaul, Michael (2001) Russia’s Unfinished Revolution: Political Change from Gorbachev to Putin. Ithaca, NY: Cornell University Press.

• Moser, Robert G. (1997) ‘The Impact of Parliamentary Electoral Systems in Russia’, Post-Soviet Affairs, 13: 284–302.

• Moser, Robert G. (2001a) ‘Compromise Amidst Political Conflict: the Origins of Russia’s

• • • • • •

Mixed-Member System’, in Matthew S. Shugart and Martin P. Wattenberg (eds), MixedMember Electoral Systems: The Best of Both Worlds? New York and Oxford: Oxford University Press, pp. 255–77. Moser, Robert (2001b) Unexpected Outcomes: Electoral Systems, Political Parties, and Representation in Russia. Pittsburgh: University of Pittsburgh Press. Remington, Thomas F. and Steven S. Smith (1996) ‘Political Goals, Institutional Context, and the Choice of an Electoral System: the Russian Parliamentary Election Law’, American Journal of Political Science, 40, 4: 1253–79. Remington, Thomas F. and Steven S. Smith (2001) The Politics of Institutional Choice: The Formation of the Russian State Duma. Princeton, NJ: Princeton University Press. Shvetsova, Olga (2002) ‘Gaining Legislative Control Through Strategic District Nomination: the Case of the Russian Left in 1995’, Legislative Studies Quarterly, 27, 4: 635–57. Shvetsova, Olga (2003) ‘Resolving the Problem of Pre-Election Coordination: the Parliamentary Election as an Elite Presidential Primary’, in V. Hesli and W. Reisinger (eds), Elections, Parties and the Future of Russia. Cambridge and New York: Cambridge University Press. White, Stephen, Richard Rose and Ian McAllister (1997) How Russia Votes. Chatham, NJ: Chatham House.

This page intentionally left blank

Part 5 Africa

This page intentionally left blank

23 Africa: Dictatorial and Democratic Electoral Systems since 1946 Matt Golder and Leonard Wantchekon1

Elections have been an integral part of African politics since independence (Nohlen, Krennerich and Thibaut, 1999; Ellis, 2000). There have been 321 legislative and 167 presidential elections in Africa between 1946 (or independence) and 1996. These elections have occurred in both democratic and authoritarian periods. Although elections have been relatively common in Africa, very little scholarly attention has been paid to them. The dearth of electoral studies focused on Africa compared to other regions of the world can partly be explained by the widespread establishment of single-party regimes in the 1960s. It was not until the re-emergence of democratic multi-party elections following the third wave of democratization in the 1990s that African electoral studies began to grow (Bratton and van de Walle, 1997; Cowen and Laakso, 1997; Wiseman, 1992; Barkan, 1995; Sisk and Reynolds, 1998; Manning, 2002). This chapter represents an addition to this growing literature by providing an overview of elections and electoral systems in Africa. While focusing primarily on the electoral institutions employed in democratic elections, we also consider the role of elections in authoritarian periods. We describe the particular electoral rules employed, investigate why they were chosen and examine their impact on African party systems.

An overview of African elections Authoritarian government has dominated the postwar history of independent Africa. In fact, there have only been 189 country-years of democracy in Africa compared to 1,823 country-years of dictatorship between 1946 and 2000. Moreover, dictatorships still outstrip the number of democracies in Africa by a considerable margin despite the transitions to democracy that occurred in the early 1990s. Table 23.1 indicates the number of legislative and presidential elections that have occurred in democratic and authoritarian periods in each African country. It is interesting to note that only Eritrea and Somaliland have failed to experience a legislative or presidential election in the time period considered here. This suggests that elections have played an important role in African politics during both authoritarian and democratic periods. 401

402

Matt Golder and Leonard Wantchekon

Table 23.1

Africa: dictatorial and democratic elections, 1946–2002

Country

Democratic periods

Democratic elections

Dictatorial elections

Legislative Presidential Legislative Presidential Algeria Angola Benin Botswana Burkina Faso Burundi Cameroon Cape Verde Central A. R. Chad Comoros Congo Djibouti Egypt Equat. Guinea Eritrea Ethiopia Ethiopia 2 Gabon Gambia Ghana Guinea Guinea-Bissau Ivory Coast Kenya Lesotho Liberia Libya Madagascar Malawi Mali Mauritania Mauritius Morocco Mozambique Namibia Niger Nigeria Rwanda São Tomé Senegal Seychelles Sierra Leone Somalia Somaliland South Africa

Never Never 1991– Never Never Never Never 1991– 1993– Never 1990–4 1960–2, 1992–6 Never Never Never Never Never Never Never Never 1970–1, 1979–80 Never Never Never Never Never Never Never 1993– 1994– 1992– Never 1968– Never Never 1990– 1993–5 1960–5, 1979–82, 1999– Never 1991– Never Never 1961–6, 1996– 1960–8 Never 1994–

0 0 3 0 0 0 0 2 2 0 2 3 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 2 2 2 0 7 0 0 2 2 3 0 4 0 0 3 2 0 2

0 0 2 0 0 0 0 2 2 0 1 2 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 2 3 0 0 0 0 2 1 2 0 2 0 0 1 0 0 0

6 3 5 6 4 3 8 3 3 3 5 4 4 13 6 0 6 1 9 6 4 6 4 7 7 2 9 5 8 7 5 5 0 5 4 0 6 4 5 2 7 6 5 2 0 11

6 1 3 0 3 2 6 0 2 1 3 0 2 8 4 0 0 0 4 4 3 4 1 8 6 0 8 0 7 0 2 3 0 0 1 0 4 2 3 0 6 3 2 1 0 0

Africa: Electoral Systems since 1946

Sudan Swaziland Tanzania Togo Tunisia Uganda Zaire Zambia Zimbabwe

1956–7, 1965–8, 1986–8 Never Never Never Never 1980–4 Never 1991– Never Total

2 0 0 0 0 1 0 3 0 50

0 0 0 0 0 0 0 2 0 26

6 5 7 6 10 3 7 7 9 284

403

4 0 7 4 6 2 4 5 2 147

Note: Data are taken from Przeworski et al. (2000) and Golder (2003). The legislative elections of 1958 in Sudan, 1963 in Congo and 1969 in Somalia are treated as democratic. Although Przeworski et al. code these years as dictatorships, these elections actually occurred prior to the transition to democracy (Gandhi, 2003). Ethiopia 2 is Ethiopia after Eritrea’s secession. (Editor’s note: This table includes several democratic elections that do not appear in the General Tables, which are based, more restrictively, on a minimum of one million inhabitants and the Freedom House reports since 1972.)

Much of the recent research on African elections fails to provide a consistent definition of which elections should be considered democratic. In fact, the elections that are treated as democratic often vary from author to author. Clearly, this is problematic if one wants to develop testable and generalizable conclusions regarding African democratic elections. In this paper, a regime is classified as a dictatorship if the chief executive is not elected, the legislature is not elected, there is no more than one party or there has been no alternation in power (Przeworski et al., 2000). In other words, a regime is democratic if those who govern are selected through contested elections. Countries are coded based on the regime that existed at the end of the given year. This coding of democracy correlates highly with other attempts to classify democracy (Bollen, 1990; Coppedge and Reinicke, 1990; Gastil, 1990; Gurr, 1990). The coding results in 50 legislative and 26 presidential elections being classified as democratic. Thus the vast majority of elections in Africa have occurred under dictatorship. The number of elections considered democratic here is much lower than that often seen in other studies of African elections. For example, Mozaffar (2001) is not unusual in classifying as many as 91 legislative elections as being democratic between 1946 and 2000. The classification used here omits elections such as those in Botswana that are typically considered democratic. The problem with the elections in Botswana is that the same party has been ruling since independence. Thus there is some uncertainty as to whether elections are held in Botswana only because the ruling party is certain to win them and whether the ruling party would yield office if it ever lost. Holm (1988) notes that ‘the resulting conflict could well force the BDP [Botswana Democratic Party] to choose between losing in parliamentary elections and abandoning elections as a method of leadership selection. Given the paternalistic attitude of the BDP from President Masire down, the latter choice would not be surprising.’ If there is no alternation in power, regimes are treated as

404

Matt Golder and Leonard Wantchekon

dictatorships. The benefit of our classification of African regimes compared to those used elsewhere is that it is consistent, stated clearly and based entirely on observables rather than subjective judgements.

Electoral institutions under dictatorship The renewed interest in African elections is obviously related to the transitions to democracy and multi-party systems that occurred in the 1990s. However, it is important to remember that the vast majority of African elections have occurred under authoritarian rule. Thus far, we have accumulated little systematic knowledge concerning the role that elections play under dictatorship. African dictators often held elections, had legislatures and organized political parties. It remains an open research question as to why these seemingly democratic electoral institutions were chosen given that ‘parties do not compete, elections do not select, and legislatures do not decide’ in dictatorships (Gandhi, 2003). We do not pretend to provide an answer to this question here. Instead, we provide a brief historical description of the electoral institutions chosen during authoritarian periods and pinpoint some of the implicit explanations that exist in the literature in the hope that other scholars will conduct more detailed theoretical and empirical analyses in this area. Electoral experience on the part of Africans was largely absent during the colonial period (Wiseman, 1990). The few notable exceptions tended to be found in francophone Africa, where the ideology of cultural assimilation occasionally permitted African electoral participation. For example, French colonial settlements in four Senegalese communes were allowed to vote for a deputy in the French National Assembly after 1848 (Ellis, 2000; Cowen and Laakso, 1997). However, it was only in the final years of colonialism following the Second World War that African electoral participation became more widespread as the French and British attempted to mollify emerging nationalist movements (Ellis, 2000; Nohlen, Krennerich and Thibaut, 1999). The British began establishing fledgling parliamentary systems that would eventually form the basis for independence. The French ordinances of 1945 established electoral colleges by which Africans could elect representatives to the Constituent Assembly so as to participate in the writing of the Fourth Republic’s constitution. This was followed by the introduction of universal suffrage and a high degree of internal autonomy in francophone Africa in the Loi Cadre of 1956. The French ultimately hoped to maintain their influence in Africa by establishing a French-dominated federal community. The first African elections based on universal suffrage and unrestricted party formation were held in the late 1950s just before the formal date of independence for most countries (Cowen and Laakso, 1997). Similar electoral reforms to those instituted by the French and British occurred much later in the Belgian and Portuguese colonies. It is clear to see that the colonists only brought democracy to Africa as they left (Adejumobi, 2000). Multiple parties competed in the early elections following independence and voters often had a considerable range of choices. For example, Wiseman (1990)

Africa: Electoral Systems since 1946

405

states that there were 130 different parties in Somalia at one stage following decolonization in 1960. However, these multi-party systems did not last long and were soon replaced by single-party rule in the 1960s. In countries such as Togo, Benin, Sierra Leone, Kenya, Zambia, Tanzania, Angola and Sudan, one-party rule was introduced de jure. Tanzania and Kenya both used acts of parliament to make opposition parties illegal. In other African states, a de facto one-party system emerged after smaller parties dissolved voluntarily and their leaders and voters became absorbed into the ruling parties (Wanyande, 2000; Adejumobi, 2000). This was the case in the Gambia, Botswana, Mauritius, Zimbabwe and Senegal. Many African dictators continued to hold periodic elections throughout these authoritarian periods. Legislatures were often retained and parliamentary systems became increasingly presidential in nature. A handful of countries did temporarily return to multi-party politics (Ghana 1969, 1979; Nigeria 1979, 1983; Burkina Faso 1970, 1978; Uganda 1980; Central African Republic 1981; and Sudan 1986) (Nohlen, Krennerich and Thibaut, 1999). However, multi-party systems did not re-emerge with any real frequency in Africa until the 1990s. African dictators attempted to justify the shift towards single-party rule in several ways. For example, some claimed that single-party systems were more suitable for the nation-building projects that became necessary following decolonization; multi-party politics would only lead to ethnic conflict and division (Adejumobi, 2000). Others stated that single-party rule was required to help economic development. This notion was supported at the time by scholars who thought that dictatorships were necessary to generate development and that ‘political participation must be held down, at least temporarily, in order to promote economic development’ (Huntington and Nelson, 1976; Galenson, 1959; De Schweinitz, 1959). Still others simply claimed that multi-party politics was foreign to Africa and was an unnatural imposition by the former colonial powers (Hanna, 1964). In many cases, these justifications simply hide the obvious fact that ruling elites and dictators wanted to retain power and saw multi-party elections as too risky. The question is why these dictators chose to have a single party, hold elections and maintain legislatures at all; they could have simply relied on the use of force. After all, other dictators did eschew these institutions. For example, Sultan Sa’id ibn Taymur al Sa’id ruled Oman for 38 years without any parties or legislatures (Gandhi and Przeworski, 2001). A scholarly consensus on the answer to this question has not yet emerged. What is clear is that dictators saw elections as a tool to control the population. Milton Obote, a former Ugandan president, clearly made this point when he stated that elections are a way of controlling the people rather than being a means through which they could control him (Cohen, 1983). Wanyande (2000) is correct in stating that single-party elections are meaningless as measures of popularity and legitimacy. Instead, it is more useful to see elections in African dictatorships as a means for recruiting the political elite or as ceremonial performances that help enforce citizen obedience, induce complicity and socialize the electorate (Chazan, 1979; Cliffe, 1967). The maintenance of legislatures allows outsiders to achieve policy compromises and share the privileges of power, while parties act as a stable system of patronage to reward their own clientele and co-opt

406

Matt Golder and Leonard Wantchekon

political opponents (Nohlen, Krennerich and Thibaut, 1999; Gandhi and Przeworski, 2001). While these claims about the role of electoral institutions under dictatorship are often case-specific and anecdotal, they generate the testable hypothesis that dictatorial survival should be positively related to the presence of these institutions. Gandhi (2003) uses a duration model to examine the impact of elections, parties and legislatures on the survival of 512 dictators between 1946 and 1996 in 138 countries worldwide. She finds that the ideal institutional arrangement for dictatorial survival is to have a single party with periodic elections; having a legislature also helps. It seems that in many cases, African dictators chose the ideal institutional arrangement for maintaining their power. The introduction of one-party rule has had several important consequences for the development of African politics. The most obvious is that there was little alternation among parties and government during the authoritarian period.2 In fact, there was no alternation in national government in Africa until the electoral victory of the Mauritian socialists in 1982 and 1983. The lack of real choice often led to low electoral participation and turnout in elections during this period (Adejumobi, 2000; Cohen, 1983). More significantly, the absence of formal competition was often used to justify military coups. For example, the 1966 army coup in Ghana was justified on the grounds that the army and police had used the only means available for removing a dictator; only undemocratic methods could be used in the pursuit of the ultimate goal of democracy. Single-party rule has also had much longer-lasting consequences. Many ruling parties have been able to maintain their control on power despite the re-emergence of multi-party systems in the 1990s. Opposition parties have often been too weak and fractionalized from decades of suppression and intimidation to be electorally competitive. In fact, ruling parties have continued to alter electoral rules to maintain their hold on power (Adejumobi, 2000). This helps to explain why there were opposition boycotts of some sort in 11 of the 15 founding elections that occurred between 1995 and 1997 (Bratton, 1998). Ruling parties have also been able to use state resources and exploit the patronage networks that had been developed to co-opt opposition groups during the period of single-party rule. This is not to say that there has been no alternation in power following the emergence of multi-party elections. After all, regime change did follow the founding elections in Cape Verde, São Tomé & Príncipe, Benin, Zambia, Mali, Congo (Brazzaville), Madagascar, Niger, Lesotho, Burundi, Central African Republic, South Africa and Sierra Leone. However, second elections have not been marked by leadership alternation and ruling parties have been able to retain and consolidate their grip on power (Bratton, 1998; Ellis, 2000). In fact, clear presidential turnover occurred in only two cases (Benin 1996; Madagascar 1997). In both cases, members of the elite in the former dictatorship were returned to power. Former ruling parties continue to enjoy legislative majorities in most of sub-Saharan Africa (Manning, 2002). This is quite intriguing and leads some to wonder whether in many cases multi-party elections are simply another form of manipulation by political elites foisted upon reluctant incumbent regimes by donor governments and financial institutions (Cowen and Laakso, 1997; Adejumobi, 2000). More systematic research is required to fully

Africa: Electoral Systems since 1946

407

understand why some former dictatorial elites have been able to return to power in Africa (and Eastern Europe).

Electoral institutions under democracy Having examined the electoral institutions employed under dictatorship, we now turn to the choice and impact of electoral systems used for the 50 legislative and 26 presidential elections that took place during democratic periods in Africa. These elections occurred in just 20 African countries (see Table 23.1). Choice of electoral institutions One feature that distinguishes democratic electoral systems is whether the regime is presidential or parliamentary. Several different criteria have been proposed for classifying these regimes (Shugart and Carey, 1992; Lijphart, 1992). We follow a fairly minimalist definition, where a presidential regime is one in which the government serves at the pleasure of the elected president. The president may be directly or indirectly elected; the important feature is that the president selects and determines the survival of the government. A parliamentary system is one in which the government serves so long as it maintains the confidence of the legislature. A system in which the government must respond both to the legislative assembly and to an elected president is classified as mixed. Typically, these mixed systems are characterized by a president who is elected for a fixed term with some executive powers and a government that serves at the discretion of the legislature. This classification scheme follows the recommendations of Przeworski et al. (2000). The regime adopted by African countries on independence was primarily parliamentary in former British colonies and presidential or mixed in the colonies of other countries. Many of the former French colonies have adopted the mixed system that has characterized the French Fifth Republic since 1958. The only country to have a democratic parliamentary regime that was not a former British colony is Cape Verde. Many of the parliamentary regimes in Anglophone countries became presidential during the authoritarian period of single-party rule. By the 1990s, presidential regimes had come to dominate African democracies (Southall, 1999; Wanyande, 2000); the only democratic African countries that were parliamentary as of 2000 were Mauritius and Cape Verde. The predominance of presidentialism raises concerns about the survivability of Africa’s democratic regimes given the strong empirical evidence that parliamentary systems survive longer than presidential ones (Cheibub, 2002; Mainwaring, 1993; Linz, 1990; Stepan and Skach, 1993). The fact that the longest surviving democracy in Africa (Mauritius) is parliamentary may perhaps be no coincidence. Although Botswana is not considered a democracy in this analysis, it is interesting to note that it too has retained a parliamentary regime and been able to maintain a multi-party system since independence (Southall, 1999). The vast majority of African presidents are elected using absolute majority rule. In other words, a candidate must win over 50 per cent of the popular vote to become president. If no candidate overcomes this threshold then there is a runoff

408

Matt Golder and Leonard Wantchekon

between the two candidates who received the most votes in the first round. In many cases, a second round has not been necessary with one candidate winning an overwhelming majority. A qualified majority system was used in the 1996 presidential elections in Sierra Leone. In these elections a candidate had to win 55 per cent of the vote in order to be elected in the first round (Nohlen, Krennerich and Thibaut, 1999). The only democratic presidential elections that employed plurality rule between 1946 and 2000 occurred in the Congo (1961), Malawi (1994, 1999), Nigeria (1979, 1999) and Zambia (1996). Most states that had used plurality rule to elect presidents during the authoritarian period adopted absolute majority rule following their transition to democracy. Only Zambia has actually switched from using an absolute majority requirement (in 1991) to using plurality rule (in 1996). Having focused on presidential elections, we now examine the electoral institutions that characterize the democratic legislative elections in Africa. Traditionally, legislative elections have been distinguished by whether they use majoritarian or proportional electoral systems (Duverger, 1954; Taagepera and Shugart, 1989; Lijphart, 1994). The problematic nature of this simple dichotomy has become increasingly clear over time with the emergence of numerous countries using more complex electoral systems that employ multiple tiers and/or a combination of electoral formulas (Massicotte and Blais, 1999; Shugart and Wattenberg, 2001). We classify legislative electoral systems into three main types: majoritarian, proportional and mixed. Majoritarian and proportional systems employ a single electoral formula in one or more electoral tiers. A mixed system combines a majoritarian and proportional formula. Table 23.2 illustrates the electoral system used in democratic legislative elections in Africa by former colony. It is clear that the particular electoral system adopted by each country is heavily influenced by its former colonial ruler. For example, all of the majoritarian systems that employed plurality rule are former British colonies except for the Congo in 1963. Mauritius is somewhat unusual for a British colony since it employs plurality rule in multi-member districts and allocates up to eight ‘best-loser’ seats in order to ensure a fair representation for each community. The democratic elections in Comoros, the Central African Republic and Mali all employed the absolute majority runoff formula used in Fifth Republic France. The electoral system in Mali differed slightly from that in France since it used multi-member districts. The former colonies of Portugal and Italy introduced proportional systems. Cape Verde and São Tomé & Príncipe both adopted the d’Hondt quota that was being used in the Portuguese metropole. All of the other proportional systems in Africa employ the Hare quota except for South Africa, which allocates 200 seats using the Droop quota at the national level and 200 seats using the same formula in nine regional constituencies. Two African countries have used mixed systems that combine majoritarian and proportional electoral formulas. Madagascar allocated 82 seats by plurality rule and a further 78 by the Hare quota in 1998. Niger employed plurality rule to allocate eight seats and the Hare quota to distribute 75 additional seats in 1993 and 1995. Why did African countries choose the electoral institutions that they did? Clearly, the answer relates to the colonial heritage of each country. However, this

Africa: Electoral Systems since 1946

Table 23.2

409

Africa: colonial legacy and democratic electoral systems Electoral system Majoritarian

Proportional

British:

Ghana (1979) Malawi (1994, 1999) Mauritius (1976, 1982, 1983, 1987, 1991, 1995, 2000) Nigeria (1964, 1979, 1999) Sierra Leone (1962) Sudan (1958, 1986) Uganda (1980) Zambia (1991, 1996)

Sierra Leone (1996)

French:

Central A. R. (1993, 1998) Comoros (1992, 1993) Niger (1993, 1995) Mali (1992, 1997) Congo (1963, 1992, 1993)

Benin (1991, 1995, 1999) Madagascar (1993)

Portuguese:

Cape Verde (1991, 1995) São Tomé & Príncipe (1991, 1994, 1998)

Italian:

Somalia (1964, 1969)

Other:

Namibia (1994, 1999)

Mixed

Madagascar (1998)

South Africa (1994, 1999)

Note: Parentheses indicate election years. Data from Golder (2003).

answer is only somewhat informative. After all, electoral rules typically represent negotiated settlements between conflicting parties over institutional design (Kaminski, 1999).3 Mozaffar (1998) argues that Anglophone countries employed plurality rule for the multi-party elections of the 1990s because of the institutional incentive structure that had developed under colonial rule and that continued to exist during the authoritarian period of single-party rule. The British had been relatively tolerant to the formation of autonomous associational life compared to other colonial powers and this led to the creation of localized associations dealing with agriculture, welfare and industry. This, combined with the introduction of plurality rule elections at independence, encouraged the elites to develop strong links with the constituencies based on patronage and pork-barrel servicing. Plurality rule was maintained following the reintroduction of multi-party elections in the 1990s because authoritarian incumbents and opposition groups expected to be able to retain their respective local power bases. Mozaffar claims that electoral institutions were different in the former French colonies because associational life had typically been organized into state-sponsored peak associations. This organizational structure meant that there was little incentive for elites to develop strong constituency ties. The inclusion of these associations in choosing the new electoral institutions in the 1990s ‘led to a strategic convergence on proportional

410

Matt Golder and Leonard Wantchekon

representation’. It is important to note, though, that former French colonies such as the Central African Republic, Comoros, Mali and Congo retained majoritarian electoral systems. The conclusion that electoral systems result from strategic negotiations between conflicting parties can most clearly be seen in those countries where democratic multi-party elections emerged following civil war or unrest.4 For example, groups with conflicting interests in Namibia and South Africa were forced to make compromises during constitutional negotiations in order to achieve transitions to democracy. Both countries eventually introduced proportional electoral institutions. Consider the South African case. The National Party that held economic and political power recognized that it did not have a realistic hope of governing under a majoritarian system due to the presence of a permanent black majority. As a result, it sought a power-sharing formula during the negotiations prior to the first democratic election in 1994. In particular, it wanted to replace the current plurality rule electoral system with a proportional one that would guarantee representation to the white, coloured and Indian minorities that it hoped to represent (Southall, 1999). In contrast, the ANC was committed to majoritarian rule through a one-person, one-vote electoral system in an attempt to entrench the power of the majority. In order to ease the transition to democracy, both sides were ultimately forced to compromise. The National Party accepted that it would not have a veto in the post-election cabinet and scrapped its insistence that an upper house with veto power be mandated by the constitution. In return, the ANC agreed that the National Party would be part of the government for five years, that cabinet decisions would be made in the spirit of consensus, that 60 per cent of the Parliament would be needed to approve the final constitution and that cabinet seats would be guaranteed to minority parties that won more than 5 per cent of the seats in the lower house (Wood, 2000). Most importantly, the ANC agreed to a proportional representation electoral system. 5 Table 23.2 suggests that electoral systems chosen by African leaders have been relatively stable. Majoritarian systems employing plurality rule have been maintained in most of anglophone Africa. Only Sierra Leone, Namibia and South Africa have introduced proportional formulas. There has been little experimentation with alternative electoral systems in other regions of Africa as well. In fact, electoral system reform has been on the table in very few African countries (Mali, Benin and South Africa). This is somewhat surprising if one compares African democracies to those in Eastern Europe. Although the third wave of democracy swept across both regions at roughly the same time, East European elites have consistently and frequently experimented with their electoral systems (Golder, 2003; Kaminski, 2002a). This might suggest that African democracies have somehow managed to reach an equilibrium in multi-party competition (Kaminski, 2002b). However, it is important to remember that some authoritarian elites that failed to achieve satisfactory results at the polls simply ignored the election results. For example, the military nullified the Nigerian presidential elections of 1993. Losing parties also failed to accept the results of the elections in Angola (1992), Burundi (1993) and Lesotho (1997). In Algeria, the military interrupted the electoral process

Africa: Electoral Systems since 1946

411

prior to the second round of the 1991 elections when it appeared that the Islamic Salvation Front (FIS) was going to win an overall majority. These examples suggest that many African countries failed to reach an equilibrium in multi-party competition. It may be the case that Table 23.2 gives a false sense of stability in that it excludes precisely those cases where the losers were unwilling to abide by the electorate’s decision. It may also be premature to suggest that a long-term equilibrium has been successfully achieved in those countries where the losers have accepted the election results. To accept the electoral returns suggests that losing parties prefer the current institutional arrangements to open opposition and attempting to overthrow the government by extra-constitutional methods (Przeworski, 1991). However, it is unclear whether this situation will last in all of these countries. The specific combination of ethnically based parties with plurality rule and geographically located ethnic groups that characterizes certain African countries has generated situations in which losing parties may have little or no chance of ever holding executive power. Winning parties or alliances have been elected to power with overwhelming legislative majorities in these circumstances and the number of competitive districts is almost non-existent in certain countries (Sisk and Reynolds, 1998). For example, Frederick Chiluba’s Movement for Multiparty Democracy won 125 of the 150 parliamentary seats in the Zambian elections of 1991. In effect, several African political systems have become de facto one-party states with opposition parties ‘frozen’ in permanent opposition. If the costs of rebellion ever decrease or the benefits from holding office increase, it may well become preferable for these parties to rebel rather than fight a futile battle on the electoral front. Impact of electoral institutions This section briefly examines the impact of electoral institutions on African party systems. In particular, it investigates their impact on the number of parties in a given country. While there is an extensive empirical literature analysing the link between various aspects of electoral systems and the number of parties, this research has focused almost exclusively on Western Europe or OECD countries (Sartori, 1976; Taagepera and Shugart, 1989; Lijphart, 1994). Little systematic analysis has been conducted on the impact of electoral institutions on the number of parties in the African context.6 This is important since there is other empirical research to suggest that the number of parties has a significant impact on stability (Mainwaring, 1993), political violence (Powell, 1982) and party system extremism (Cox, 1990). The literature that focuses on electoral institutions suggests that greater electoral system permissiveness should be associated with a larger number of parties. This is because voters have less incentive to vote strategically when electoral systems are permissive. Equally, political elites have less incentive to withdraw in favour of better placed parties (Duverger, 1954; Cox, 1997). Thus disproportional systems are likely to be associated with few parties and proportional systems with many. The proportionality of an electoral system is typically measured in terms of its district magnitude (the number of seats allocated in a constituency). Table 23.3 indicates

412

Matt Golder and Leonard Wantchekon

Table 23.3

Africa: effective number of parties Average district magnitude

Effective parties in votes Effective parties in seats

1

2–5

6–10

>10

4.54 (12) 3.67 (18)

3.84 (14) 2.92 (17)

3.96 (6) 3.60 (7)

2.50 (5) 2.31 (5)

Electoral formula Plurality Majority 2.91 (19) 2.61 (23)

6.73 (4) 4.57 (8)

PR-Hondt PR-Hare 2.49 (5) 2.10 (5)

6.47 (7) 4.55 (9)

Note: Parentheses indicate number of elections. Data from Golder (2003).

the average effective number of electoral and legislative parties by average district magnitude.7 Empirical research in other regions of the world suggests that disproportionality also seems to vary systematically across electoral formulas. For example, it appears to decrease as one moves from plurality rule to absolute majority rule to the d’Hondt quota to the Hare quota. Table 23.3 indicates how the effective number of electoral and legislative parties varies with these electoral formulas. The theory would lead one to expect that the number of parties should increase as one moves from the left-most column to the right-most column. It is fairly clear that there is no monotonic increase in the number of electoral and legislative parties as district magnitude increases. Similarly, the number of parties does not increase as the electoral formula becomes more permissive. Thus it seems that electoral institutions do not (yet) have the same impact on the number of parties in Africa as they do in other regions of the world. There are at least two explanations for this result. Each explanation has contrasting implications for the distinctiveness of African party systems. It may simply be the case that there has been insufficient time for the strategic incentives generated by electoral institutions to be fully felt. After all, there is only a handful of countries that have held more than two consecutive elections. The fact that political parties and voters have little experience with electoral politics serves to weaken the incentive structure of the electoral system. Voters can only act strategically when they know who the trailing parties are and when they have public information about the preferences and intentions of other voters (Cox, 1997). This is unlikely to be the case in new democracies since voters have no past electoral performance by which to judge the electoral viability of particular candidates. Moreover, opinion polls are often inherently uncertain and unstable in such an environment. This uncertainty may explain why in some African countries so many opposition parties put up candidates in the founding elections. The point is that this uncertainty should decline as the number of elections increases. This should lead to a ‘shake-down’ process in the party system by which voters desert parties that are not electorally viable and party elites begin to better coordinate their electoral strategies. Thus one should expect to see the smaller African parties disappear as they are absorbed by larger parties or merge to form more electorally attractive organizations. This line of reasoning suggests that the strategic incentives

Africa: Electoral Systems since 1946

413

created by electoral institutions will be more fully felt over time and that African party systems will eventually come to resemble those in other regions of the world. This is the same process that seems to be occurring in the new democracies of Eastern Europe. The second explanation suggests that electoral institutions fail to have an effect on the number of parties because African party systems are rarely nationalized.8 Electoral institutions primarily create incentives for strategic behaviour at the district level (Cox, 1997, 1999). Thus one would expect to see few parties in a single-member district; the number of parties should be higher in constituencies with larger district magnitudes. One would only expect to see this empirical prediction at the national level if the party system was nationalized (Chiber and Kollman, 1998). Thus the empirical evidence linking electoral system permissiveness to the number of parties in West European and OECD countries implicitly indicates that these particular party systems are nationalized. In contrast, many African parties have regional fiefdoms in which the party wins almost all of the seats. Consider the elections of 1994 in Malawi where single-member districts and plurality rule were employed. The three main political parties were overwhelmingly dominant in their core regions and exceedingly weak in the remaining two. The Alliance for Democracy won every seat in the northern region, the Malawi Congress Party won 75 per cent of the seats in the central region and the United Democratic Front won 95 per cent of the votes in the southern region (Reynolds, 1995b). Although there was almost exactly one electorally viable party in each district, there appears to be three national parties once these districts are aggregated up to the national level. This example helps to explain why it is difficult to evaluate the impact of electoral institutions by focusing solely on the number of African parties at the national level. The fact that political parties are typically based on ethnic groups that are often geographically concentrated helps to explain why many African party systems are not nationalized. These party systems are unlikely to become nationalized until ethnic mobilization stops being the most stable source of material benefit and security for most voters (Cohen, 1983). This explanation suggests that African party politics may well remain distinct compared to other regions of the world.

Conclusion Elections in Africa have evolved considerably from the colonial period when only European settlers in a few countries were able to participate. The limited electoral opportunities for natives that had been introduced by the colonial powers in an attempt to undermine and co-opt nationalist movements between 1946 and 1957 were eventually expanded to include multi-party elections with universal suffrage in the independence years, 1957–65. However, the democratic institutions that were introduced after independence soon broke down under the combined weight of regional and ethnic conflict, military coups and the increased political influence of the former Communist bloc. Single-party regimes became the norm across Africa and elections resulted in a concentration of political power and limited

414

Matt Golder and Leonard Wantchekon

political representation. Elections during this period essentially served as a mechanism for co-opting political rivals and endowing authoritarian regimes with a public sense of legitimacy (Collier, 1982). It was only with the end of the Cold War and the emergence of domestic protest movements that multi-party democracy was restored in several African countries in the 1990s. The resurgence of democracy in Francophone Africa can also be traced to the Baule summit in 1989. It was here that the Mitterrand government switched from a policy of intervention in African internal politics to one of disengagement. While African electoral systems appear to be greatly influenced by colonial legacies, it would be wrong to underestimate the extent of strategic negotiation that goes on between political elites when determining electoral system choice. This is most obviously the case in countries that have experienced civil war, internal unrest or rising ethnic tensions. The uncertainty generated by the rapid return to multi-party elections in the 1990s has led to a high degree of political fragmentation in many countries. This uncertainty, along with the need to build bases of electoral support quickly, has typically resulted in a strategy of ethnic mobilization (Posner, 1999). The combination of ethnic mobilization and geographically concentrated ethnic groups has often led to the emergence of regional parties. This development is encouraged in those countries that have adopted plurality rule electoral systems. The fact that electoral institutions have often created incentives to place ethnic identity at the centre of African politics is clearly significant. This is because there is good reason to believe that ethnic politics lies at the heart of Africa’s growth tragedy (Easterly and Levine, 1997).

Notes 1. We would like to thank Michael Nest for his useful comments. 2. Some have argued that competitive politics simply shifted to within the party during authoritarian periods. For example, Hayward (1987) claimed that in ‘Kenya, Tanzania, Nigeria and Sierra Leone, surprisingly large numbers of incumbents have been defeated by voters unhappy with their performance.’ In some estimations, as many as one half of the incumbent deputies and at least one quarter of the government’s ministers were removed in some elections in Kenya and Tanzania. 3. The debate in the literature as to whether African countries should employ proportional representation or majoritarian rule ignores this point (Reynolds, 1995a, 1995b; Barkan, 1995). Particular institutions may simply not be sustainable as an equilibrium in certain societies. Instead, it is better to begin by examining the type of electoral institutions that might be acceptable to the relevant parties. 4. The importance of strategic negotiations in choosing electoral rules can even be seen in authoritarian periods. For example, regional leaders in Benin adopted a rotating presidency in 1969 in an attempt to keep a lid on rising ethnic tension. The ‘Conseil Présidentiel’ was essentially a power-sharing arrangement similar to the one introduced in the former Yugoslavia after Tito’s death in 1980. 5. A similar compromise occurred in Namibia. The 1989 elections were supposed to elect a constituent assembly that would write a constitution for an independent Namibia. These elections were conducted under the authority of the South African administration as de facto rulers and the UN Technical Assistance Group as representatives of the international community. The South African administration employed a proportional electoral system in an attempt to restrict SWAPO to less than two-thirds of the National Assembly.

Africa: Electoral Systems since 1946

415

This would prevent SWAPO from being able to write a constitution without the agreement of any other parties (Southall, 1999). 6. Mozaffar (1998, 2001) represents important exceptions. 2 7. The effective number of electoral parties is calculated as 1/ Σ v i where vi is the percentage th of votes won by the i party. The effective number of electoral parties substitutes the number of seats won by a party for the number of votes it won (Laakso and Taagepera, 1979). Data are missing for the number of electoral parties in ten cases: Central African Republic (1993), Comoros (1993), Congo (1992, 1993), Madagascar (1993, 1998), Niger (1995), Somalia (1969) and Sudan (1958, 1986). 8. We consider a nationalized party system to be one in which the parties are relatively competitive throughout the country.

References Adejumobi, Said (2000) ‘Elections in Africa: a Fading Shadow of Democracy’, in Okwudiba Nnoli (ed.), Government and Politics in Africa: A Reader. Harare, Zimbabwe: AAPS Books, pp. 242–61. Barkan, Joel (1995) ‘Elections in Agrarian Societies’, Journal of Democracy, 6: 106–16. Bollen, Kenneth (1990) ‘Issues in the Comparative Measurement of Political Democracy’, American Sociological Review, 45: 370–90. Bratton, Michael (1998) ‘Second Elections in Africa’, Journal of Democracy, 9: 51–66. Bratton, Michael and Nicolas van de Walle (1997) Democratic Experiments in Africa: Regime Transitions in Comparative Perspective. New York: Cambridge University Press. Chazan, Naomi (1979) ‘African Voters at the Polls: a Re-examination of the Role of Elections in African Politics’, Journal of Commonwealth and Comparative Political Studies, 17: 136–58. Cheibub, José (2002) ‘Minority Governments, Deadlock Situations, and the Survival of Presidential Democracies’, Comparative Political Studies, 35: 284–312. Chibes, Pradeep and Ken Kollman (1998) ‘Party Aggregation and the Number of Parties in India and the United States’, American Political Science Review, 92: 329–42. Cliffe, Lionel (1967) One Party Democracy: The 1965 Tanzanian General Elections. Nairobi: East African Publishing House. Cohen, Denis (1983) ‘Election and Election Studies in Africa’, in Yolamu Barongo (ed.), Political Science in Africa: A Critical Review. London: Zed Books, pp. 72–93. Collier, Ruth (1982) Regimes in Tropical Africa: Changing Forms of Supremacy, 1945–1975. Berkeley: University of California Press. Coppedge, Michael and Wolfgang Reinicke (1990) ‘Measuring Polyarchy’, Studies in Comparative International Development, 25: 51–72. Cowen, Michael and Liisa Laakso (1997) ‘An Overview of Election Studies in Africa’, Journal of Modern African Studies, 35: 717–44. Cox, Gary (1990) ‘Centripetal and Centrifugal Incentives in Electoral Systems’, American Journal of Political Science, 34: 903–35. Cox, Gary (1997) Making Votes Count: Strategic Coordination in the World’s Electoral Systems. New York: Cambridge University Press. Cox, Gary (1999) ‘Electoral Rules and Electoral Coordination’, Annual Review of Political Science, pp. 145–61. De Schweinitz, Karl (1959) ‘Industrialization, Labor Controls and Democracy’, Economic Development and Cultural Change, 7: 385–404. Duverger, Maurice (1954) Political Parties. New York: Wiley. Easterly, William and Ross Levine (1997) ‘Africa Growth Tragedy. Policies and Ethnic Divisions’, Quarterly Journal of Economics, 112: 1203–50. Ellis, S. (2000) ‘Elections in Africa in Historical Context’, in Jon Abbink and Gerti Hesseling (eds), Election Observation and Democratization in Africa. New York: St. Martin’s Press – now Palgrave Macmillan.

416

Matt Golder and Leonard Wantchekon

Galenson, Waiter (1959) Labor and Economic Development. New York: Wiley. Gandhi, Jennifer (2003) Political Institutions Under Dictatorship. PhD dissertation, Department of Politics, New York University. Gandhi, Jennifer and Adam Przeworski (2001) Dictatorial Institutions and the Survival of Dictators. Paper presented at the 2001 Annual Meeting of the American Political Science Association, San Francisco. Gastil, Raymond (1990) Freedom in the World: Political Rights and Civil Liberties. New York: Freedom House. Golder, Matt (2003) Democratic Electoral Systems Around the World, 1946–2000. Paper presented at the Department of Political Science, University of Montreal. Gurr, Ted (1990) Polity II: Political Structures and Regime Change, 1800–1986. Ann Arbor, MI: Inter-University Consortium for Political and Social Research. Hanna, William (1964) Independent Black Africa: The Politics of Freedom. Chicago: Rand McNally. Hayward, Fred (ed.) (1987) Elections in Independent Africa. London: Westview Press. Holm, John (1988) ‘Botswana: a Paternalistic Democracy’, in L. Diamond, J. Linz and S. Lipset (eds), Democracy in Developing Countries: Africa. Boulder, CO: Lynne Riener, pp. 179–216. Huntington, Samuel and Joan Nelson (1976) No Easy Choice: Political Participation in Developing Countries. Cambridge, MA: Harvard University Press. Kaminski, Marek (1999) ‘How Communism Could Have Been Saved: Formal Analysis of Electoral Bargaining in Poland in 1989’, Public Choice, 98: 83–109. Kaminski, Marek (2002a) ‘Do Parties Benefit from Electoral Manipulation? Electoral Laws and Heresthetics in Poland, 1989–93’, Journal of Theoretical Politics, 14: 325–59. Kaminski, Marek (2002b) A General Equilibrium Model of Multi-Party Elections. Mimeo, Department of Politics, New York University. Laakso, Markku and Rein Taagepera (1979) ‘Effective Number of Parties: a Measure with Application to Western Europe’, Comparative Political Studies, 12: 3–27. Lijphart, Arendt (1992) Parliamentary versus Presidential Government. Oxford: Oxford University Press. Lijphart, Arendt (1994) Electoral Systems and Party Systems. A Study of Twenty-Seven Democracies, 1945–90. Oxford: Oxford University Press. Linz, Juan (1990) ‘The Perils of Presidentialism’, Journal of Democracy, 1: 51–69. Mainwaring, Scott (1993) ‘Presidentialism, Multipartism and Democracy: the Difficult Combination’, Comparative Political Studies, 26: 198–228. Manning, Carrie (2002) Assessing African Party Systems after the Third Wave. Paper presented at the University of North Texas Conference on Parties and Elections in New Democracies. Massicotte, Louis and André Blais (1999) ‘Mixed Electoral Systems: a Conceptual and Empirical Survey’, Electoral Studies, 18: 341–66. Mozaffar, Shaheen (1998) ‘Electoral Systems and Conflict Management in Africa: a TwentyEight State Comparison’, in Timothy Sisk and Andrew Reynolds (eds), Elections and Conflict Management in Africa. Washington, DC: United States Institute of Peace Press, pp. 81–98. Mozaffar, Shaheen (2001) Institutions, Context and Outcomes in Africa’s Emerging Democracies. Paper presented at the 2001 Annual Meeting of the American Political Science Association, San Francisco. Nohlen, Dieter, Michael Krennerich and Bernard Thibaut (1999) Elections in Africa. A Data Handbook. Oxford: Oxford University Press. Posner, Daniel (1999) The Institutional Origins of Ethnic Politics in Zambia. Cambridge, MA: Harvard University Press. Powell, G. Bingham (1982) Contemporary Democracies: Participation, Stability and Violence. Cambridge, MA: Harvard University Press. Przeworski, Adam (1991) Democracy and the Market: Political and Economic Reforms in Eastern Europe and Latin America. New York: Cambridge University Press. Przeworski, Adam, Michael Alvarez, José Cheibub and Fernando Limongi (2000) Democracy and Development: Political Institutions and Material Well-Being in the World, 1950–1990. New York: Cambridge University Press.

Africa: Electoral Systems since 1946

417

Reynolds, Andrew (1995a) ‘The Case for Proportionality’, Journal of Democracy, 6: 117–24. Reynolds, Andrew (1995b) ‘Constitutional Engineering in Southern Africa’, Journal of Democracy, 6: 86–99. Sartori, Giovanni (1976) Parties and Party Systems: A Framework for Analysis. New York: Cambridge University Press. Shugart, Matthew and John Carey (1992) Presidents and Assemblies: Constitutional Design and Electoral Dynamics. New York: Cambridge University Press. Shugart, Matthew and Martin Wattenberg (eds) (2001) Mixed-Member Electoral Systems. The Best of Both Worlds? New York: Oxford University Press. Sisk, Timothy and Andrew Reynolds (eds) (1998) Elections and Conflict Management in Africa. Washington, DC: United States Institute of Peace Press. Southall, Roger (1999) ‘Electoral Systems and Democratization in Africa’, in John Daniel, Roger Southall and Morris Szeftel (eds), Voting for Democracy: Watershed Elections in Contemporary Anglophone Africa. Brookfield, VT: Ashgate, pp. 19–36. Stepan, Alfred and Cindy Skach (1993) ‘Constitutional Frameworks and Democratic Consolidation: Parliamentarism and Presidentialism’, World Politics, 46: 1–22. Taagepera, Rein and Matthew Shugart (1989) Seats and Votes: The Effects and Determinants of Electoral Systems. New Haven, CT: Yale University Press. Wanyande, Peter (2000) ‘Democracy and the One-Party State: the African Experience’, in Okwudiba Nnoli (ed.), Government and Politics in Africa: A Reader. Harare, Zimbabwe: AAPS Books, pp. 107–21. Wiseman, John (1990) Democracy in Black Africa: Survival and Revival. New York: Paragon House. Wiseman, John (1992) ‘Early Post-Redemocratization Elections in Africa’, Electoral Studies, 11: 279–91. Wood, Elisabeth (2000) Forging Democracy From Below: Insurgent Transitions in South Africa and El Salvador. New York: Cambridge University Press.

Further reading • Adejumobi, Said (2000) ‘Elections in Africa: a Fading Shadow of Democracy’, in Okwudiba • • • • • • • • • • •

Nnoli (ed.), Government and Politics in Africa: A Reader. Harare, Zimbabwe: AAPS Books, pp. 242–61. Bratton, Michael (1998) ‘Second Elections in Africa’, Journal of Democracy, 9: 51–66. Bratton, Michael and Nicolas van de Walle (1997) Democratic Experiments in Africa: Regime Transitions in Comparative Perspective. New York: Cambridge University Press. Chazan, Naomi (1979) ‘African Voters at the Polls: a Re-examination of the Role of Elections in African Politics’, Journal of Commonwealth and Comparative Political Studies, 17: 136–58. Cohen, Denis (1983) ‘Election and Election Studies in Africa’, in Yolamu Barongo (ed.), Political Science in Africa: A Critical Review. London: Zed Books, pp. 72–93. Cowen, Michael and Liisa Laakso (1997) ‘An Overview of Election Studies in Africa’, Journal of Modern African Studies, 35: 717–44. Ellis, S. (2000) ‘Elections in Africa in Historical Context’, in Jon Abbink and Gerti Hesseling (eds), Election Observation and Democratization in Africa. New York: St. Martin’s Press – now Palgrave Macmillan. Hanna, William (1964) Independent Black Africa: The Politics of Freedom. Chicago: Rand McNally. Hayward, Fred (ed.) (1987) Elections in Independent Africa. London: Westview Press. Manning, Carrie (2002) Assessing African Party Systems after the Third Wave. Paper presented at the University of North Texas Conference on Parties and Elections in New Democracies. Reynolds, Andrew (1995) ‘Constitutional Engineering in Southern Africa’, Journal of Democracy, 6: 86–99. Southall, Roger (1999) Electoral Systems and Democratization in Africa’, in John Daniel, Roger Southall and Morris Szeftel (eds), Voting for Democracy: Watershed Elections in Contemporary Anglophone Africa. Brookfield, VT: Ashgate, pp. 19–36.

418

Matt Golder and Leonard Wantchekon

• Wanyande, Peter (2000) ‘Democracy and the One-Party State: the African Experience’, in Okwudiba Nnoli (ed.), Government and Politics in Africa: A Reader. Harare, Zimbabwe: AAPS Books, pp. 107–21. • Wiseman, John (1990) Democracy in Black Africa: Survival and Revival. New York: Paragon House. • Wiseman, John (1992) ‘Early Post-Redemocratization Elections in Africa’, Electoral Studies, 11: 279–91.

24 Africa: Electoral Systems in Emerging Democracies Shaheen Mozaffar1

Electoral systems for electing the lower chamber of national legislatures in Africa’s emerging democracies are the outcomes of strategic choice structured by a combination of contextual and contingent variables. As elsewhere, political actors’ calculations about their prospects of winning and losing in future competitive elections were the principal sources of electoral systems choice in Africa’s emerging democracies. But contextual variables related to the political legacies of varied authoritarian regimes, as well as contingent variables related to the process of democratic transitions, influenced these calculations, informed political actors’ institutional preferences and shaped their choice of alternative electoral systems. Electoral systems, in the aggregate, produce their expected mechanical effects on the disproportionality between votes and seats and on the structure (fragmentation or concentration) of party systems as measured by the effective number of electoral and legislative parties. Closer analysis of disaggregated data, however, shows the counterintuitive effects of electoral systems, especially on the structure of party systems. These effects can be explained by the mediating influence of three contextual variables: 1. the characteristic information deficit of political actors about the incentive structures of new democratic institutions and the extent of their electoral support in democracies established after protracted authoritarian rule; 2. the institutional framework of presidential regimes adopted by most emerging democracies in Africa; 3. two dimensions of ethno-political cleavages – ethno-political fragmentation and ethno-political concentration – that independently and interactively with each other and with legislative and presidential electoral institutions limit the expansion of the number of electoral and legislative parties. As a result, party systems tend to be concentrated, which has important and rather counterintuitive implications for the stability of Africa’s emerging democracies. 419

420

Shaheen Mozaffar

Context, contingency and choice Elections, while commonplace, have a chequered history in sub-Saharan Africa. In the colonial period, very few elections were held and these were restricted to the European population and selected African elites. After 1945, the independence bargain negotiated between the departing European rulers and the emerging African nationalist elites led to the rapid extension of the franchise, as a result of which most African countries gained independence in the 1960s via competitive elections (Collier, 1982; Khadiagala and Rothchild, forthcoming). Since then, competitive elections have been held regularly in long-standing democracies (Botswana, Mauritius, Gambia until 1994, and Senegal since 1979) and intermittently in countries that experienced alternating democratic governments and military rule (e.g. Burkina Faso, Ghana, Nigeria, Sierra Leone). As authoritarian regimes steadily replaced democratic governments across the continent in the 1970s, single-party regimes held regular semi-competitive elections as occasions of policy debates and instruments of regime legitimization and elite recruitment (e.g. Côte d’Ivoire, Kenya, Tanzania, Zambia), while some military regimes organized tightly controlled plebiscitary elections to secure nominal legitimacy (Chazan, 1979; Hayward, 1987; Nohlen, Krennerich and Thibaut, 1999). Most electoral systems governing these varied elections were majoritarian and they remained largely unchanged until the onset of democracy’s third wave. Plurality rule was negotiated in virtually all African countries at independence in the 1960s, except for the former Belgian colonies of Rwanda and Burundi, both of which adopted proportional representation with preferential voting in multi-member districts. British colonies typically adopted plurality formulas in single-member districts, except Mauritius, where consideration of proportional representation and the single transferable vote in response to growing ethno-political tensions preceded the eventual adoption of three-member constituencies with candidate-centred open party lists and block voting and a unique ‘best-loser’ system to ensure ethnopolitical group parity and party balance in the national legislature after each election (Mathur, 1991: 54–71; Mozaffar, forthcoming). French colonies typically adopted plurality rule with closed party lists and bloc voting in large multimember constituencies that corresponded to the existing colonial administrative districts and facilitated representation in the territorial legislatures that preceded the creation of national legislatures at independence. These inherited electoral systems were retained largely intact for both competitive elections in short-lived democracies and for restricted elections in single-party regimes in the 1970s and 1980s. Benin and Burkina Faso were the only exceptions in replacing the inherited majoritarian systems with proportional representation formulas, the former with closed national party lists for the 1964 competitive elections and the latter with closed party lists in multi-member regional constituencies for competitive elections in 1970 and 1978. Beginning in the early 1980s, African authoritarian rulers confronted a combination of domestic and international pressures engendered by growing economic crises, diminished state capacities, declining regime legitimacy, politically unpalatable

Africa: Electoral Systems in Emerging Democracies

421

donor-imposed economic reforms, the demonstration effects of democracy’s third wave and widespread popular protests and demands for regime change. They responded to these pressures with varying degrees of political liberalization. In the ensuing process of democratization, the choice of electoral system became an important focus of political negotiations. Three factors jointly affected the nature, process and outcomes of the negotiations: (1) the electoral rules for converting votes into seats; (2) the power relations and electoral expectations of contending political actors associated with the exigencies of democratization; and (3) the structural-historical context that shaped these relations and expectations, informed political actors’ institutional preferences and constrained their bargaining strategies over alternative electoral systems (Mozaffar, 1998, 1999). Conflicts over the choice of electoral systems are conflicts over political representation measured by maximum seat shares in legislative assemblies. Electoral rules affect the distribution of seat shares and are therefore the object of strategic bargaining. The comparative advantage of different electoral rules in maximizing seat shares can be determined with reasonable accuracy from their mathematical properties, increasing the cost of bargaining over new rules. In African countries, moreover, colonial history and past electoral experiences (Mackenzie, 1960), as well as the ready availability of foreign models and experts, also helped to improve the predictive accuracy of electoral rules (Elklit, 2002; Horowitz, 1991: 163–203; Lijphart, 1990). The predictive accuracy of alternative electoral rules, however, can indicate the general pattern of seat distribution, but not whether seats will be allocated to individual actors. Actual seat allocations depend on strategic coordination and the interdependent electoral strategies of voters, candidates and parties, and these cannot be anticipated uniquely (Cox, 1997; Przeworski, 1991: 12). Hence, strategically rational actors do not bargain over seat shares but over rules that will structure their mutual electoral strategies that they expect will produce preferred seat allocations. However, even as the uncertainty that typically attends democratization increased their autonomy in shaping institutional outcomes, African political actors involved in crafting new democratic institutions did not enjoy unfettered choice over alternative electoral systems. The political legacies of authoritarian regimes and the exigencies of democratic transitions jointly constrained it. The political legacies of authoritarian regimes involved patterns of inter-group and state-group relations that resulted from the various attempts of these varied regimes to restructure the political space and reorganize state–society relations. The exigencies of democratic transition involved power asymmetries and associated differences in the electoral expectations, institutional preferences and bargaining strengths of groups comprising the pro-democracy and pro-incumbent coalitions. Variations in the combination of these contextual and contingent variables resulted in five distinct patterns of electoral system choice (see Table 24.1). Pattern 1 This pattern reflected the political legacies of relatively autonomous interest organization permitted by authoritarian regimes in virtually all Anglophone countries,

422

Shaheen Mozaffar

Table 24.1

Africa: patterns of electoral system choice in third-wave democracies

Rule

Pattern 1

Pattern 2

Plurality

Gambia Ghana Kenya Lesotho Malawi Nigeria Tanzania Zambia

Côte d’Ivoire Djibouti

Majority runoff

Proportional representation

Mixed parallel

Pattern 3

Pattern 4

Benin Burkina Faso Burundi Cape Verde Equat. Guinea Guinea-Bissau Madagascar Niger São Tomé & Príncipe

Angola Liberia Mozambique Namibia Sierra Leone South Africa Zimbabwe

Pattern 5

Central A. R. Chad Comoros Congo Gabon Mali Mauritania Togo

Cameroon Guinea Senegal Seychelles

Note: Botswana and Mauritius are excluded from this table because they are not considered to be third-wave democracies. In Niger, 75 out of 83 seats are allocated in 8 multi-member constituencies with district magnitudes ranging from 4 to 14, and 8 seats are allocated by plurality formula in 8 single-member districts that are especially designated for national minorities.

of restricted competitive elections permitted by single-party regimes in countries such as Kenya, Malawi, Tanzania and Zambia, and of fully competitive elections in short-lived democracies such as Ghana, Lesotho and Nigeria. Electoral opportunities in these elections were organized through multi-candidate electoral competition based on plurality formula in single-member districts. These opportunities generally encouraged the development of strong constituency linkages based on pork-barrel servicing, while the autonomy of interest organization enabled the development of extensive grassroots networks (soccer clubs in Ghana, for example) that helped lower the cost of potential opposition candidates mounting a serious challenge to authoritarian incumbents. While this challenge increased the incumbents’ risk of losing office, reducing their incentives to introduce multi-party democracies,

Africa: Electoral Systems in Emerging Democracies

423

the very territorial fragmentation of single-member districts also increased the cost of initiating and sustaining a united national campaign for potential opponents. In choosing new democratic institutions, both authoritarian incumbents and opposition groups, expecting to preserve their respective established local power bases while dividing their opponents, converged on single-member districts and plurality rule as the preferred formulas for the new electoral systems. For authoritarian incumbents, small territorial districts preserved their traditional strongholds and plurality rule helped their candidates over the typically numerous opposition candidates while fragmenting opposition parties, preventing them from forming legislative majorities and enhancing the ability of the now democratically elected authoritarian rulers to co-opt individual opposition members to form a government should the ruling parties fail to win legislative majorities. Similarly, for opposition groups, plurality formula and single-member districts preserved their traditional local power bases, but also enabled them to form a government by co-opting individual ruling and other opposition party members should the fragmented opposition parties fail to win legislative majorities. These calculations were the principal reason why, in Ghana’s democratic transition, for example, opposition as well as government-appointed members of the Consultative Assembly charged with devising a new constitution roundly rejected the German-style multi-member proportionality system that was proposed by the Committee of Experts appointed by the incumbent regime (Ninsin, 1993). Finally, for both authoritarian incumbents and their democratic opponents, a plurality formula in single-member districts helped to accommodate localized conflicts stemming from subregional and subethnic cleavages and personal factions, especially when a small number of large parties relied heavily on regionally concentrated voters, as, for example, in Malawi. Pattern 2 This pattern reflected the political legacy of the corporatist organization of interests favoured by authoritarian regimes in some Francophone countries. Specified social and economic interests (e.g. producers’ associations, labour, students, women, teachers, civil servants) were formally organized into state-sponsored peak associations, approximating representational monopolies usually tied to ruling parties. Elections, when held, were held in large multi-member districts by plurality formula, with electoral opportunities structured through single national party lists that were tightly controlled by central party elites. These opportunities discouraged ambitious politicians from establishing strong constituency linkages, deflecting their efforts to seek favour with central party elites who controlled valuable institutional and financial resources (e.g. media, candidate selection, patronage). In the third wave of democratization, the various corporate groups allied to single-party regimes were able to use the very corporatist structures through which they were organized to reduce collective action costs and spearhead the popular movements that forced reluctant autocrats to democratize. However, in the ensuing negotiations over electoral systems, the pro-democracy groups were either not as united as the pro-incumbent coalitions or were differentiated by organizational strengths and associated capacities to shape institutional outcomes.

424

Shaheen Mozaffar

Thus, in countries such as Côte d’Ivoire, Gabon and Togo, among others, the more united pro-incumbent coalitions were able to secure their preferred choice of plurality or two-round majoritarian systems that favoured their supporters with political bases in small single-member districts or in small multi-member districts where the use of party lists helped to accommodate intra-group and intra-party conflicts. In other countries, a small number of well-organized political parties, or groups that subsequently developed into political parties, emerged during the transition to dominate the pro-democracy coalitions and shaped the choice of electoral system. In Mali, for example, only three of the 47 political parties that emerged during that country’s chaotic democratic transition had a national political base. A second tier of parties were regionally based, but with the potential for becoming national parties. Below these groups were numerous smaller parties with predominantly local bases and thus with no prospect of entering the national government without coalescing with the larger parties. The resulting debate over the design of the new electoral system produced a two-round majority formula for National Assembly elections at the insistence of the larger parties interested in preserving their political hegemony. These parties, however, agreed to a proportional representation formula for local elections to accommodate the demands of the smaller parties (Vengroff, 1993, 1994). Pattern 3 This pattern reflected a different political legacy of the corporatist organization of interests in some Francophone and Lusophone countries. Pro-democracy and pro-incumbent coalitions in this pattern were almost equally balanced in their ability to shape institutional outcomes. The result was a strategic convergence on proportional representation (PR) as the preferred formula for the new electoral systems. The inclusion of all major groups in choosing new democratic institutions meant that no group was individually capable of imposing its preferred institutional outcome, hence no single group could be confident of securing decisive legislative majorities in competitive elections. For pro-democracy groups, political unity forged by common opposition to authoritarian regimes strengthened their bargaining position in choosing new electoral systems, but that unity could not be guaranteed in subsequent electoral competition. For the groups allied to discredited authoritarian regimes, the loss of incumbency removed a key institutional advantage and underscored their uncertain electoral prospects in unfamiliar electoral competition. Each group thus had the incentive to support proportional representation formulas to secure some legislative representation for itself. Pattern 4 In this pattern, the choice of electoral systems formed part of wider political negotiations to end violent civil conflict (South Africa) and protracted civil wars (Angola, Liberia, Mozambique, Namibia, Sierra Leone and Zimbabwe in 1980). In all seven countries, these negotiations produced proportional representation systems. In Angola, the d’Hondt formula was chosen for allocating 90 out of 220 seats in 18 five-seat regional constituencies, with the remaining 130 seats allocated in a single

Africa: Electoral Systems in Emerging Democracies

425

nationwide constituency (with three seats reserved for Angolans residing abroad). In Mozambique, the d’Hondt formula with a 5 per cent national threshold was chosen for allocating 150 seats in 11 multi-member constituencies (district magnitude range from 13 to 50). In Namibia, the Hare quota was chosen to allocate 72 seats in a single nationwide constituency. In South Africa, the Droop quota was chosen to allocate half the 400 seats in nine multi-member constituencies (district magnitude range from 4 to 43) corresponding to the country’s nine administrative regions and the other half in a single nationwide constituency, with regional votes counting toward the allocation of the national seats. In Zimbabwe, the rules for the 1980 elections provided, among others things, a 100-seat House of Assembly, of which 80 seats were allocated by Hare quota on a Common Roll (including black and white voters) in eight multi-member constituencies (district magnitude range from 6 to 16) corresponding to the country’s eight administrative districts), and 20 seats were allocated in singlemember districts by single transferable vote (STV) of white voters registered on a White Roll. These provisions were negotiated as interim measures, pending the drawing of new single-member constituencies, which was accomplished for the 1985 elections. In Sierra Leone, the d’Hondt formula with a 5 per cent national threshold was chosen for allocating 68 seats in a single nationwide constituency, while in Liberia the Hare quota with no threshold was chosen for allocating 64 seats in a single nationwide constituency. In both countries, proportional representation formulas were chosen as pragmatic attempts to legitimize competitive elections as mechanisms of conflict management in the wake of endemic violence and civil wars. Both countries had previously used plurality formulas in single-member districts, so their choice of proportional representation formulas indicates that the choice of electoral rules is neither historically determined nor culturally prescribed, but represents a strategically rational response to changing political contexts and contingencies. Pattern 5 In this pattern, the authoritarian incumbents’ goal of retaining power while accommodating demands for democratization motivated the choice of mixed electoral systems. In all four countries represented in this pattern, the incumbency advantage gave ruling parties the political leverage vis-à-vis the fragmented opposition parties to design the electoral systems. However, the specific institutional designs of these systems reflected the different configuration of political forces in individual countries. In the islands of the Seychelles, for example, the long-time ruling party and its major rival party dominated the constitutional process, agreeing first on the proscription of independent candidates altogether and then weakening the electoral prospects of the growing number of smaller parties by allocating 22 out of 33 seats by plurality formula in single-member districts and the remaining 11 seats by the Hare quota to parties meeting an excessively high threshold of 10 per cent of the national vote. In Cameroon, with a history of restricted electoral competition in single-party regimes, the fragmentation of opposition parties enabled the ruling party to design a complex electoral system that combined a plurality formula in single-member

426

Shaheen Mozaffar

districts and absolute majority with closed party list in multi-member constituencies. However, if no party list wins an absolute majority, half of the seats are allocated to the list with a relative majority and the other half are distributed by the Hare quota to the list winning at least 5 per cent of the vote in the district. In Guinea and Senegal, the design of mixed electoral systems emanated from the triple goals of new authoritarian incumbents who succeeded the long-time authoritarian rulers and founders of the two countries: (a) to consolidate their powers against the entrenched supporters of the old regime; (b) to create electoral opportunities for the growing number of opposition parties; and (c) to establish their democratic credentials. The ruling parties in both countries were characteristically only loose coalitions of divergent groups protected by authoritarian restrictions on political competition in the old regime. Democratization would thus not only subject these groups to the uncertainties of electoral competition, but would also enable the new rulers to replace them with their own supporters and simultaneously create opportunities for the representation of opposition groups. These calculations led to the choice of mixed parallel systems in both countries. The use of a plurality formula in single-member districts in Guinea and the use of a plurality formula with party block voting in multi-member districts in Senegal enabled the ruling parties to consolidate their traditional local strongholds, especially in the rural areas, and retain their numerical superiority in the national legislatures. And the creation of single nationwide electoral districts with allocation by proportional representation in both countries enabled the weakly organized opposition parties to secure legislative representation, but without threatening the power of the ruling parties due to the use of the parallel ballot that denied them compensatory seats.

The politics of electoral system reform For the most part, electoral systems chosen during third-wave democratic transitions have remained stable in Africa. However, a number of countries have reformed their electoral systems. Most of these reforms have involved changes in the size of the national legislature and in district magnitudes. In Lesotho and Madagascar, however, they have also involved changes in the electoral formula. Generally reflecting a learning process whereby political actors seek to adjust to the political consequences of their initial institutional choices, these reforms have typically involved attempts to (a) accommodate the inherent tension between party discipline and the responsiveness of legislators to their constituents and (b) enhance the legitimacy of the electoral process by expanding opportunities for the representation of smaller parties. Benin and Madagascar exemplify the first goal. In Benin, the size of the lower house of the national assembly was increased from the initial choice of 64 seats allocated in 6 regional constituencies to 83 seats allocated in 18 regional constituencies, reducing the average district magnitude from 11 to 5 seats. For the 1999 elections, the number of constituencies was further increased to 24 with a corresponding decrease in the average district magnitude to 3 seats (Magnusson, 1997, n.d.). In Madagascar as well, the initial choice of 137 seats for the national assembly was increased to 150 seats for the 1998 elections. And, for the

Africa: Electoral Systems in Emerging Democracies

427

first time among Africa’s third-wave democracies, the electoral formula was also changed from the initial choice of the Hare quota in 57 multi-member constituencies to a mixed system in which 82 seats are now allocated by plurality rule in singlemember districts and 68 seats by the Hare quota in 34 two-seat constituencies. In both Benin and Madagascar, the principal reasons articulated in the debates surrounding these reforms were the goals of bringing the members of parliament closer, and thus making them more responsive, to their constituents and the expectation that the reforms would help to reduce the high levels of party system fragmentation in both countries. Burkina Faso and Lesotho exemplify electoral system reforms aimed at strengthening the legitimacy of democratic elections by ameliorating the weakness of smaller parties in securing legislative representation. In Burkina Faso, the initial choice of 107 seats allocated in 30 regional constituencies by the d’Hondt formula for the 1992 transitional elections produced an excessively high vote–seat disproportionality index of 18.93, as measured by Gallagher’s Least-Squares (LSQ) Index, which corresponded to the inverse relationship between the high number of electoral parties, as measured by the effective number of electoral parties (ENEP) index of 3.63, and the low number of legislative parties, as measured by the effective number of legislative parties (ENLP) index of 1.82 (Gallagher, 1991; Laakso and Taagepera, 1979). For the 1997 elections, the size of the national assembly was increased to 111 seats allocated by the d’Hondt formula in 45 regional constituencies, but the changes failed to ameliorate the distortions (LSQ = 17.10, ENEP = 2.03, ENLP = 1.20). In response to the growing discontent of smaller parties, which had failed to win seats commensurate with their relatively substantial vote shares in the previous two elections, a series of political negotiations between the government and opposition parties, facilitated by international assistance, led to major systemic reforms: (a) 90 out of the 111 seats were now allocated in 13 regional constituencies (replacing the previous 45 constituencies corresponding to the country’s administrative districts); (b) the average district magnitude was now increased to 7 seats (ranging from 2 to 10), as compared to 4 seats in 1992 and 3 seats in 1997; (c) 21 seats were now allocated in a single nationwide constituency; (d) a single ballot was introduced whereby regional vote totals for each party became the basis for allocating national seats; and (e) the more proportional Hare quota replaced the more restrictive d’Hondt formula for allocating both regional and national seats. The results of the 2002 elections clearly reflected the positive effects of the reform: the LSQ index was 3.57, the ENEP index was 3.63 and the ENLP index was 3.30 (Santiso and Loada, 2003). In Lesotho, in the 1993 transitional elections, the initial choice of singlemember districts and plurality formula produced the distortions between votes and seats and the concentrated party systems typically associated with such systems, with the ruling party winning all the seats with 75 per cent of the votes (LSQ = 23.85, ENEP = 1.62, ENLP = 1.00). The situation did not change even after the number of seats was increased from 65 to 80 for the 1998 elections, with the ruling party winning 79 seats (99 per cent) with 61 per cent of the votes (LSQ = 32.11, ENELP = 2.22, ENLP = 1.03). In the wake of the political violence and

428

Shaheen Mozaffar

crisis precipitated by the lopsided results of the 1998 elections, the Southern African Development Community (SADC), a regional intergovernmental body, intervened to restore order and installed a multi-party Interim Political Authority (IPA). With the help of foreign assistance, including the contributions of a Danish political scientist, intense multi-party negotiations in the IPA produced Africa’s first mixedmember proportional (MMP) system. The new system features 120 seats, 80 of which are allocated by plurality formula in single-member districts with 40 compensatory seats allocated by Hare quota in a single nationwide constituency. The results of the 2002 elections realized the expected effects of the new system, as the ruling party won 77 out of 78 constituencies (two constituencies remain undecided), the leading opposition party won no constituency seats but won 21 out of 40 (53 per cent) of the compensatory seats, with eight smaller parties sharing the remaining 19 compensatory seats. These results produced an LSQ index of 6.50, the ENEP index was 2.54 and ENLP was 2.16. Finally, in Senegal a series of incremental electoral system reforms between 1988 and 1998 showed how authoritarian incumbents used the logic of embedded institutional reform to retain power while accommodating opposition demands for increased participation and representation. The logic of embedded institutional reform emphasizes the importance of intricate but often hidden relationships between elections and the rules governing them at multiple levels – presidential, legislative and local. Success at one level depends on performance at all levels. To retain control of the presidency, authoritarian incumbents opened the elections to the national legislature and local governments, which possess less institutional resources, to competition by the opposition parties. The opposition parties accepted these reforms because of the low entry costs of competing for national legislative and local government seats. In the short run, these multiple-level electoral reforms preserved the ruling party in power while expanding opportunities for the opposition parties, but also fragmenting them. In the long run, they encouraged splits within the ruling party and helped the opposition develop the increased ability to coalesce around a single opposition candidate, resulting in the defeat of the authoritarian incumbent and a democratic transfer of power through competitive elections in 1998 (Mozaffar and Vengroff, 2002). The new government introduced additional electoral system reforms in 2002 to consolidate its power and weaken the long-standing ruling party it had defeated in 1998. Retaining the mixed parallel system, it reduced the size of the national assembly from 140 to 120 seats, with 65 seats allocated in 30 multi-member constituencies by plurality formula and party block voting, and 55 seats allocated in a single nationwide constituency by the Hare quota. This last round of changes in the electoral system in Senegal reaffirms that electoral system reform can be used to expand democratic opportunities just as readily as it can be used to restrict them.

The political consequences of electoral systems Table 24.2 reports data on the mechanical effects of two dimensions of electoral systems, formula and district magnitude, on the index of disproportionality

Africa: Electoral Systems in Emerging Democracies

Table 24.2

429

Africa: political consequences of electoral systems

Electoral system variables

No. of cases Disproportionality Effective no. parties Relative (LSQ) reduction Votes Seats (ENEP) (ENLP)

Electoral formula Plurality Majority runoff Mixed parallel Proportional rep.

50 18 5 6 21

15.9 14.9 12.5 7.1

2.3 5.6 2.6 3.5

1.7 2.8 1.9 2.7

0.28 0.38 0.28 0.19

District magnitudes 1.00–1.50 2.10–5.20 7.86–13.00 22.70–400

46 14 18 6 8

14.7 12.3 8.3 3.7

3.3 4.0 3.6 2.7

2.0 2.5 3.2 2.3

0.28 0.31 0.15 0.12

between votes and seats (LSQ), on the indices measuring the effective number of electoral (ENEP) and legislative (ENLP) parties, and on the index of relative reduction in the effective number of parties from votes to seats (R). In the aggregate, both dimensions produce the expected mechanical effects. The disproportionality between votes and seats in countries using plurality and two-round majority runoff formulas is, on average, twice as high as countries using proportional representation formulas. Countries utilizing mixed-parallel formulas also manifest high average vote–seat disproportionality, even though five out of six of these countries combine a plurality formula with the Hare quota. In the sixth country, Cameroon, the Hare quota is used only in multi-member constituencies to allocate seats left over after half the seats in the constituency have been allocated to the party list winning a plurality of the votes. The discriminatory effects of the mixed-parallel system are also evident in the low average values of the indices measuring the structure of the party systems, as well as in the 24 per cent difference between the number of electoral (2.6) and legislative (1.9) parties. Both the plurality and the proportional representation formulas produce their expected mechanical effects on the structure of the party systems, with the former restricting the number of parties and the latter expanding them. However, party systems in countries utilizing the two-round majority runoff systems tend to be the most fragmented. While a number of factors generic to African countries in part account for this result, as discussed below, two-round systems typically create strategic opportunities for smaller parties to enter the first-round competition and establish their bargaining leverage with the larger parties seeking their support in the second. The average values of R reinforce the expected effects of the electoral formula. The two-round majority run-off formula produces a 38 per cent reduction in the effective number of electoral parties from votes to seats. The plurality and the

430

Shaheen Mozaffar

mixed-parallel systems each produce a 28 per cent reduction. The proportional representation formulas produce the least reductive effect. Consistent with theoretical expectations, district magnitude and vote–seat disproportionality are inversely correlated. In African countries, as elsewhere, singlemember districts tend to be the most disproportional, with a reductive effect of 28 per cent in the number of parties as votes are converted into seats. As district magnitude increases, the disproportionality between votes and seats decreases, with a corresponding decrease in the relative reduction values. The relationship between district magnitude and the structure of party systems does not meet theoretical expectations, however. District magnitude has a curvilinear effect on both the effective number of electoral and the effective number of legislative parties. Single-member districts correlate with an average ENEP index of 3.3 parties and an average ENLP index of 2.0 parties, with a high relative reductive effect of 28 per cent. Both indices, however, increase as district magnitude increases, but then decrease as district magnitude continues to increase. Note, however, that even as the party system becomes restricted with increasing district magnitude, the relative reductive effect of this inverse relationship declines as well. Thus the relative reduction in the effective number of electoral and legislative parties is 31 per cent when district magnitude ranges from 2.10 to 5.20 seats, which is twice as high as the 15 per cent reduction at the next higher range of district magnitude at 7.86–13.00 seats. And the relative reduction declines to its lowest value of 12 per cent at the highest range of district magnitude at 22.70–400 seats. These counterintuitive results, moreover, are also reflected in the relatively low ENEP and ENLP values across all electoral formulas and district magnitudes. In other words, the observed expected effects of both dimensions of electoral systems are only the aggregate mechanical effects of converting votes into seats. But these aggregate mechanical effects obscure the complex mediating influences of a combination of contextual and institutional variables. Clarifying these influences requires more detailed disaggregated analysis that cannot be presented here because of space limitations. Only a brief account, therefore, is offered. In Africa’s emerging democracies, political actors lack information about electoral institutions as sources of strategic coordination, and, in the wake of protracted authoritarian rule, also about the extent of their electoral support. They compensate for this information deficit by relying on alternative sources of information and strategic coordination. In contemporary Africa, for a variety of historical reasons, socially and politically constructed ethno-political groups and cleavages serve as these alternative sources. Ethno-political cleavages facilitate strategic coordination because they are strategic resources that are contingently (not reflexively) activated to define group interests and reduce the cost of collective political action in response to institutional incentives. Two key dimensions – group fragmentation and group concentration – facilitate the strategic coordination role of ethno-political cleavages. Analysts typically posit an isomorphic relationship between ethno-political cleavages and party systems on the assumption that each ethno-political group involved in a cleavage is totally separate from others, sufficiently large and internally

Africa: Electoral Systems in Emerging Democracies

431

cohesive to support a party by itself. Thus, ceteris paribus, large numbers of cleavages (high fragmentation) increase and small numbers of cleavages (low fragmentation) reduce the number of parties. However, the nature of constructed ethno-political groups in Africa (see Scarritt and Mozaffar, 1999, for explication) and the resulting cleavages manifest a complex group morphology that seriously militates against such a reflexive relationship between ethno-political cleavages and party system structure. Specifically, African ethno-political demography features politically salient differences within as well as among groups. The result is a high degree of ethnopolitical fragmentation that, ceteris paribus, either produces such a high degree of vote dispersion among large numbers of small parties that most are unlikely to secure enough votes to win seats, or produces small numbers of large multi-ethnic parties by encouraging them to campaign for votes across both inter-group and intra-group cleavages. Either way, high ethno-political fragmentation reduces the number of parties. While ethno-political group fragmentation increases the transaction cost of voter mobilization due to the combined presence of inter-ethnic and intra-ethnic cleavages, spatial concentration of ethno-political groups reduces it. Spatial concentration by itself, however, is unlikely to overcome the reductive effect of high fragmentation due to the presence of large numbers of small ethno-political groups. African countries with low fragmentation, moreover, feature a small number of large ethno-political groups that are also spatially dispersed, and therefore do not need concentrated voters to sustain a small number of political parties. Ethno-political group concentration and fragmentation, therefore, interact to foster but a moderate number of political parties. Ethno-political cleavages also interact with electoral institutions governing presidential elections to counteract the reductive effect of the proximity of presidential and legislative elections on the number of electoral and legislative parties. With the exception of Lesotho and South Africa, all of Africa’s third-wave democracies have chosen presidential governments, and most of them have adopted the two-round majority runoff formula for presidential elections. By influencing the effective number of presidential candidates, especially in the first round, ethnopolitical cleavages moderate the reductive effects of the proximity of presidential and legislative elections on the number of parties, encouraging the formation of electoral alliances between major presidential contenders seeking electoral victory and minor ones controlling critical blocs of votes that can secure it. The net result is again a moderate increase in the number of political parties. Finally, ethno-political cleavages interact with district magnitude in complex ways to shape party system structure. Large district magnitudes reduce the number of parties if ethno-political fragmentation is high and ethno-political concentration is low, as exemplified by South Africa, but they tend to increase the number of parties if both fragmentation and concentration are high, as exemplified by Benin. In South Africa, ethno-political groups are highly fragmented due to substantial cleavages among the nine groups that comprise the majority African population as well as among the English-speakers and the Afrikaners that comprise the white population. They are also spatially dispersed. However, the continued strategic

432

Shaheen Mozaffar

importance of race as a cost-effective basis of electoral mobilization diminishes the political significance of intra-group cleavages among African voters, while white voters typically tend to divide their votes among several smaller parties. As a result, the average effective number of both electoral and legislative parties in South Africa over two elections is 2.2, even though the average district magnitude is 40 seats and the electoral formula is the highly proportional Droop quota. In Benin, ethno-political groups are only slightly less fragmented than in South Africa, but they are more geographically concentrated, principally in the administrative provinces that form the electoral districts. Combined with an average district magnitude of 11 seats and with the highly proportional Hare quota, this concentration has resulted in one of the most fragmented party systems in Africa, with the average ENEP value of 13.0 and an average ENLP value of 7.0. Finally, spatial concentration helps to offset the constraining effects of plurality rule and singlemember districts, as, for example, in Malawi, where it has resulted in average ENEP and ENLP values of 2.8 and 2.8 respectively, with a correspondingly low LSQ index of 3.5. Due to these complex relationships between electoral institutions and ethnopolitical cleavages, Africa’s emerging democracies feature electoral and legislative party systems that are characterized by relatively low levels of fragmentation. For example, a recent analysis (Mozaffar, Scarritt and Galaich, 2003) of 62 elections in 34 countries revealed that for all elections the mean ENEP and ENLP values are 3.2 (standard deviation = 2.7) and 2.3 (standard deviation = 1.5) parties, respectively; the corresponding median values are 2.4 and 2.0 parties, respectively. For 47 multiple elections in 19 countries, the mean ENEP and ENLP values are 3.1 (standard deviation = 2.4) and 2.3 (standard deviation = 1.5) parties, respectively; the corresponding median values are 2.4 and 2.0 respectively. And for 15 single elections in 15 countries, the mean ENEP and ENLP values are 3.7 (standard deviation = 3.4) and 2.5 (standard deviation = 1.5) parties, respectively; the corresponding median values are 2.5 and 2.2 respectively. This remarkable convergence of Africa’s emerging democracies around some semblance of party system equilibrium obtains even in the face of high levels of underlying electoral and legislative volatility, as measured by Pederson’s (1983) index of volatility. For example, for the 47 multiple elections (ranging from 2 to 5 elections), the mean electoral volatility index (based on votes) is 27.87 and the mean legislative volatility index (based on seats) is 22.96. These indices are comparatively lower than those reported for new democracies in post-Second World War Western Europe and in Southern Europe in the 1970s, but are roughly comparable to those reported for the new democracies in Latin America in the 1980s and in the former communist states in Eastern Europe in the 1990s. In all these new democracies, however, high electoral volatility coincides with high party system fragmentation (Bielasiak, 2002). There are three reasons for the paradoxical coincidence of high electoral volatility with low party system fragmentation in Africa. First, because of the manifest programmatic and organizational weakness of African political parties, candidates rely principally on ethno-political group support to win. But, because of inter-group and intra-group cleavages, few African ethno-political groups possess sufficient

Africa: Electoral Systems in Emerging Democracies

433

numerical strength and internal cohesion to propel a candidate to victory on their own. Candidates, therefore, forge ethno-political coalitions to secure victory in each election. However, because of the very morphology that encourages them in the first place, these election-specific coalitions tend to shift from election to election, increasing electoral volatility. Second, there are characteristic information deficits in elections in postauthoritarian democracies concerning the incentive structure of new electoral institutions, the aggregate patterns of vote distribution, and the extent of electoral support for individual candidates and parties. This leads to excessively large numbers of candidates and parties competing in democratic elections in Africa’s emerging democracies. The typical result is a highly skewed vote distribution with a small number of candidates and parties winning by substantial majorities, high levels of electoral volatility, the emergence of a small number of electoral and legislative parties, and hence relatively concentrated party systems. Finally, these two reasons suggest a curvilinear relationship between ethnopolitical fragmentation and the structure of party systems in Africa’s emerging democracies. Low fragmentation correlates in a straightforward way with a small number of parties and low party system fragmentation. High fragmentation encourages the formation of inter-group coalitions that also help to reduce the number of parties and party system fragmentation. But moderate levels of fragmentation, combined with the countervailing effect of concentration, increase the number of parties and party system fragmentation, with district magnitude exerting an additional moderating influence to diminish the prospect of excessive increases in both. This complex interaction of electoral institutions and ethno-political cleavages thus structures the conversion of electoral volatility into party systems that are only moderately fragmented, offering crucial insights into, as well as reasons for cautious optimism about, the prospect of democratic stability in contemporary Africa.

Note 1. The data and the research presented in this chapter are drawn from a larger project that was supported by a grant from the National Science Foundation (SBER-9515439). I thank the Boston University African Studies Center for continued research support and Kathleen O’Doherty for invaluable research assistance. I am solely responsible for the chapter.

References Bielasiak, Jack (2002) ‘The Institutionalization of Electoral and Party Systems in Postcommunist States’, Comparative Politics, 34: 189–210. Chazan, Naomi (1979) ‘African Voters at the Polls: a Re-examination of the Role of Elections in African Politics’, Journal of Commonwealth and Comparative Political Studies, 17, 1: 136–58. Collier, Ruth Berins (1982) Regimes in Tropical Africa: Changing Forms of Supremacy, 1945–1975. Berkeley: University of California Press. Cox, Gary (1997) Making Votes: Strategic Coordination in the World’s Electoral Systems. Cambridge and New York: Cambridge University Press.

434

Shaheen Mozaffar

Elklit, Jørgen (2002) ‘Lesotho 2002: Africa’s First MMP Elections’, Journal of African Elections, 1, 2: 1–10. Hayward, Fred M. (1987) Elections in Independent Africa. Boulder, CO: Westview. Horowitz, Donald (1991) A Democratic South Africa? Constitutional Engineering in a Divided Society. Berkeley: University of California Press. Khadiagala, Gilbert and Donald Rothchild (eds) (forthcoming) The Impact of Colonial Bargaining on Inter-Group Relations in Africa. Lijphart, Arend (1990) ‘Electoral Systems, Party Systems, and Conflict Management in Segmented Societies’, in Robert A. Schrire (ed.), Critical Choices for South Africa. Cape Town: Oxford University Press, pp. 2–13. Lootvoet, Benôit (1996) ‘Guinèe, les tentations du passé: elements d’analyse de la scène politique’, in Afrique Politiques 1996: Democratisation: Arrèt Sur Images. Paris: Karthala, pp. 85–108. Mackenzie, William J. M. (1960) Five Elections in Africa. Oxford: Clarendon Press. Magnusson, Bruce (1997) The Politics of Democratic Regime Legitimation in Benin: Institutions, Social Policy, and Security. PhD thesis, University of Wisconsin-Madison. Magnusson, Bruce (n.d.) ‘After All Else Failed: Building Democratic Legitimacy in Benin’. Typescript. Mathur, Hansraj (1991) The Parliament in Mauritius. Rose Hill, Mauritius: L’Océan Indièn. Mozaffar, Shaheen (1997) ‘Electoral Systems and Their Political Effects in Africa: a Preliminary Analysis’, Representation, 34: 148–56. Mozaffar, Shaheen (1998) ‘Electoral Systems and Conflict Management in Africa: a TwentyEight State Comparison’, in Timothy Sisk and Andrew Reynolds (eds), Elections and Conflict Management in Africa. Washington, DC: United States Institute of Peace Press, pp. 81–98. Mozaffar, Shaheen (1999) ‘Cascading Interdependence, Democratization, and Ethnic Conflict Management in Africa, 47–59’, in John F. Stack (ed.), Ethnic Entanglements and World Politics. Greenwood, CT: Praeger. Mozaffar, Shaheen (2001) Institutions, Context and Outcomes. Paper presented to the American Political Science Association, San Francisco, 30 August – 1 September. Mozaffar, Shaheen (forthcoming) ‘Negotiating Independence in Mauritius’, in Gilbert Khadiagala and Donald Rothchild (eds), The Impact of Colonial Bargaining on Inter-Group Relations in Africa. Mozaffar, Shaheen and Richard Vengroff (2002) ‘A “Whole System” Approach to the Choice of Electoral Rules in Democratizing Countries: Senegal in Comparative Perspective’, Electoral Studies, 21: 601–16. Mozaffar, Shaheen, James R. Scarritt and Glen Galaich (2003) ‘Electoral Institutions, Ethnopolitical Cleavages and Party Systems in Africa’s Emerging Democracies’, American Political Science Review, 97, 3: 379–90. Ninsin, Kwame (1993) Political Parties and Democracy in Ghana’s Fourth Republic. Accra: Woeli Publishing Services. Nohlen, Dieter, Michael Krennerich and Bernhard Thibaut (eds) (1999) Elections in Africa: A Data Handbook. New York: Oxford University Press. Pedersen, Mogens N. (1983) ‘Changing Patterns of Electoral Volatility in European Party Systems, 1948–1977: Explorations in Explanations’, in Hans Daalder and Peter Mair (eds), Western European Party Systems: Continuity and Change. Beverly Hills, CA: Sage, pp. 29–66. Przeworski, Adam (1991) Democracy and the Market: Political and Economic Reforms in Eastern Europe and Latin America. New York: Cambridge University Press. Santiso, Carlos and Augustin Loada (2003) ‘Explaining the Unexpected: Electoral Reform and Democratic Governance in Burkina Faso’, Journal of Modern African Studies, 41, 3: 395–420. Scarritt, James R. and Shaheen Mozaffar (1999) ‘The Specification of Ethnic Cleavages and Ethno-political Groups for the Analysis of Democratic Competition in Contemporary Africa’, Nationalism and Ethnic Politics, 5: 82–117.

Africa: Electoral Systems in Emerging Democracies

435

Vengroff, Richard (1993) ‘Governance and the Transition to Democracy: Political Parties and the Third Republic System in Mali’, Journal of Modern Africa Studies, 31, 4: 541–62. Vengroff, Richard (1994) ‘The Impact of the Electoral System on the Transition to Democracy in Africa: the Case of Mali’, Electoral Studies, 13, 1: 29–37.

Further reading • Chazan, Naomi (1979) ‘Africa at the Polls: a Re-examination of the Role of Elections in Africa Politics’, Journal of Commonwealth and Comparative Politics, 17, 1: 136–58.

• Collier, Ruth Berins (1982) Regimes in Tropical Africa: Changing Forms of Supremacy, 1945–1975. Berkeley: University of California Press.

• Hayward, Fred M. (1987) Elections in Independent Africa. Boulder, CO: Westview. • Mozaffar, Shaheen (1997) ‘Electoral Systems and their Political Effects in Africa: a Preliminary Analysis’, Representation, 34: 148–56.

• Mozaffar, Shaheen and Richard Vengroff (2002) ‘A “Whole System” Approach to the Choice • • • •

of Electoral Rules in Democratizing Countries: Senegal in Comparative Perspective’, Electoral Studies, 21: 601–16. Mozaffar, Shaheen, James R. Scarritt and Glen Galaich (2003) ‘Electoral Institutions, Ethno-political Cleavages and Party Systems in Africa’s Emerging Democracies’, American Political Science Review, 97, 3: 379–90. Scarritt, James R. and Shaheen Mozaffar (1999) ‘The Specification of Ethnic Cleavages and Ethno-political Groups for the Analysis of Democratic Competition in Contemporary Africa’, Nationalism and Ethnic Politics, 5: 82–117. Schrire, Robert A. (ed.) (1996) Critical Choices for South Africa. Cape Town: Oxford University Press. Sisk, Timothy and Andrew Reynolds (eds) (1998) Elections and Conflict Management in Africa. Washington, DC: United States Institute of Peace Press.

436

Summary Table 5A

Africa: Assembly (lower or single chamber)

Country Law year Election years

Term

Districts

Ballot

Rule/Formula

Seats

No.

Magn.

4

64–83

6–24

3–14

Closed list

Proportional – Hare

5

31–40 (+7)

31–40

1

Single

Plurality

Cape Verde 1992: 1995, 2001

5

72

16

2–13

Closed list

Proportional – d’Hondt

Central African Republic 1993: 1993, 1998

5

109

109

1

Single

Majority/2nd round plurality

Ghana 1956: 1956, 1965 // 1969 // 1979 // 1992, 1996, 2000

5 4

104–200

104–200

1

Single

Plurality

Kenya 1963: 1963 // 1992, 1997, 2002

5

117–210 (+14)

117–210

1

Single

Plurality

Benin 1991: 1991, 1995, 1999 Botswana 1965: 1965, 1969, 1974, 1979, 1984, 1989, 1994, 1999

Threshold

Madagascar 1992: 1993

4

138

57

2–8

Closed list

Proportional – Hare

1998: 1998, 2002

4

160 68–82 68–92

68–82 34–46

1 2

Single Closed list

Coexistence Plurality Proportional – Hare

Malawi 1994: 1994, 1999

5

177–192

177–192

1

Single

Plurality

Mali 1992: 1992, 1997, 1997, 2002

5

116–147 (+13)

55

1–7

Multiple

Majority/2nd round runoff

Mozambique 1993: 1994, 1999

5

250

11

13–50

Closed list

Proportional – d’Hondt

Namibia 1990: 1994, 1999

5

72 (+6)

1 nat.

72

Closed list

Proportional – Hare

4

312–590

312–590

1

Single

Plurality

Nigeria 1959: 1959, 1964 // 1979, 1983 // 1992 // 1999

5% nat.

437

438

Summary Table 5A Country Law year Election years Senegal 1992: 1993, 1998, 2001

South Africa 1994: 1994, 1999

(Continued) Term

5

5

Districts Seats

No.

Magn.

120–140 65–70 55–70

1 30

65–70 1–5

400

1

400

Source: See Appendix: Notes and Sources for Summary Tables at the end of the book.

Ballot

Rule/Formula

Single

Two tiers Proportional – Hare Plurality

Closed list

Proportional (2 tiers) – Droop

Threshold

Africa: Electoral Systems in Emerging Democracies

Summary Table 5B Country Law year

439

Africa: Presidency Election years

Term

Rule

Benin 1991:

1991, 1996, 2001

5+5

Majority/2nd round runoff

Cape Verde 1992:

1996, 2001

5+5

Majority/2nd round runoff

Central African Republic 1993: 1993, 1999

6+6

Majority/2nd round runoff

Ghana 1979:

1979 // 1992, 1996, 2000

4+4

Majority/2nd round runoff

Kenya 1963:

1963 // 1992, 1997, 2002

5+5+5

Qualified plurality 25% in 5--8provinces/2nd round runoff

Madagascar 1992:

1992, 1996, 2001

5+5

Majority/2nd round runoff

Malawi 1994:

1994, 1999

5+5

Plurality

Mali 1992:

1992, 1997, 2002

5+5

Majority/2nd round runoff

Mozambique 1994:

1994, 1999

5+5

Majority/2nd round runoff

Namibia 1990:

1994,1999

5+5

Majority/further rounds

Nigeria 1979:

1979, 1983 //

4+4

Qualified plurality 25% in 2--3 states/college

1993:

1993

4+4

Qualified plurality 33% in 2--3 states

1999:

1999

4+4

Qualified plurality 25% in 2--3 states

Source: See Appendix: Notes and Sources for Summary Tables at the end of the book.

25 South Africa: Proportional Representation in the Puzzle to Stabilize Democracy Andrew Reynolds

The institutions of the Union of South Africa (established in 1910) were largely inherited from the British colonial institutions which had been first imported and then adapted by the British/Dutch/French settler regime which had controlled South African administration since the eighteenth century. Thus a classically British system of single-member districts and plurality rule was used for all elections to the effectively whites-only House of Assembly which sat until it was swept away in 1994. Between 1910 and 1948 governments were dominated by coalitions led by English South African parties. However, in 1948 D. F. Malan’s Afrikaner National Party won 52 per cent of the seats in parliament (with 42 per cent of the popular vote) to Jan Smut’s United Party’s 42 per cent of the seats (based on 52 per cent of the popular vote). Subsequently every government between 1948 and 1994 was based on a National Party majority. The ending of apartheid and the collapse of the white regime in South Africa led to a switch to list proportional representation to elect members of the National Assembly in 1994 and beyond. The system currently in use in South Africa is a large-district, closed-list proportional representation system using the Droop quota and largest remainders for seat allocation. Voters have a single vote but seats are allocated in a 200-member national district and nine provincial multi-member districts varying in size based on population. Provincial districts function for seat allocation purposes only and party seat shares in the national assembly are entirely based on national vote share and a quota based on a single national district of 400 members without an imposed threshold for representation.

Choosing an electoral system for multi-party democracy in 1994 Surprisingly, one of the least contentious issues throughout the entire negotiation process that transformed South Africa from racial authoritarianism to multi-party democracy was the agreement of almost all key players on the use of a proportional representation system to elect the Constituent Assembly in 1994. The whites-only parliament had inherited the British single-member district plurality system and it was long thought that the African National Congress (ANC) would seek to 440

South Africa

441

maintain the system, as they not only had electoral advantage in doing so, but because single-member district plurality fitted in with their desire for a majority rule constitutional dispensation. Nevertheless, the ANC ultimately agreed to the use of proportional representation and argued strongly that it would encourage participation by all groups with significant followings, avoid wasted votes and hence be an inducement to vote in areas where one party was dominant, leading to a more exact reflection of the popularity of political parties. They also maintained that it would avoid the time, expense and accusations of bias in the process of delimiting constituencies. There were also some pragmatic and self-interested concerns which lay behind the ANC’s enlightened stance. First, they realized that ‘giving in on the proportional representation issue’ would be perceived both domestically and internationally as a major concession to minority concerns. It would prove to be a strong bargaining chip with the Nationalists in the negotiations and a clear sign to the international community that the ANC were not about to try and seize power and turn the country into a racially exclusive one-party state. Second, the ANC estimated that the Group Areas Act had created ethnic bastions which would return minority members no matter what and thus would disenfranchise ANC supporters in non-hegemonic ANC areas. Despite the 1980 census showing whites to be in a majority in only five magisterial districts the ANC’s reasoning appears to have had some basis in fact as evidence suggests that under a plurality single-member district system the ANC would have been shut out of representation in Northern KwaZulu and much of the Western Cape. The ANC’s preferences when it came to the electoral system were heavily influenced by one of their leading constitutional specialists, Kader Asmal, and a paper he had published in October 1990, ‘Electoral Systems: A Critical Survey’. Asmal, the effective author of the electoral system found in the interim constitution, appreciated the way in which proportional representation could facilitate an inclusive polity which would convert potentially anti-system minority parties into pro-system parties with incentives to play their democratic rolls as constituent segments of a ‘loyal opposition’. As Sisk notes, the ANC chose it [proportional representation] after a well-considered assessment of not only what was desirable but also of what was possible, given the preferences of others. It evolved out of the process of strategic interaction (Sisk, 1995: 285). The National Party of F. W. de Klerk fairly early in the negotiation process (1990) demonstrated awareness of the importance of electoral system design to their role and position in any new multi-party dispensation. Under the new electoral realities they realized they would lose out through the maintenance of any aspect of the Westminster system and thus espoused power-sharing and anti-majoritarianism with the zeal of born-again converts. They sought a bicameral legislature, with a lower house elected by PR and an upper house with parties being given equal weight by their mere presence in provincial legislatures. That the National Party (NP) desired proportional representation as a component of any new constitutional dispensation was never in doubt as proportional representation was integral to any power-sharing arrangement which shied away from

442

Andrew Reynolds

separate voters’ rolls. Roelf Meyer, the NP’s lead negotiator, quickly realized that the existing single-member district plurality system had the potential to devastate their seat-winning abilities, even if they polled a significant minority of the vote. As it turned out their fears may have been somewhat unfounded as the concentration of their vote in the Cape and Transvaal in 1994 would have had seat-winning potential under a single-member district system, but in 1989 no one could have confidently predicted that the NP would do so well among the coloured community of the Western Cape or whether their votes would be geographically concentrated enough to win single-member districts in the face of an overwhelmingly dominant opponent. The smaller Democratic Party (DP) had perhaps the most sophisticated understanding of the electoral system debate and thus it was unsurprising that they came up with the most complicated proposal, most suited to their peculiar basis of electoral support and self-perceived role in South Africa. In their constitutional proposals of 1991 the DP proposed a proportional representation system with small multi-member constituencies. In their plan there would be 100 three-member districts with 100 additional (and compensatory) seats being drawn from a separate national list. In the eyes of the DP this arrangement fulfilled the requirements of proportionality while protecting the role of geographical accountability in the choice of representative. The DP also perceived the plan to be in their best interests as historically their support was highly concentrated in the white liberal suburbs of Cape Town, Johannesburg and Durban. As DP Members of Parliament in the old whites-only Assembly often had one or two colleagues representing adjacent constituencies, the proposed three-member districts seemed optimal. Among the other parties the Pan-Africanist Congress (PAC) and the Azanian People’s Organization (AZAPO) both accepted proportional representation, but this sat uneasily with the PAC’s insistence on majoritarianism and their opposition to the Nationalists’ power-sharing proposals. Nevertheless, Barney Desai, the PAC publicity secretary, argued that list proportional representation would avoid wasted votes, encourage participation and test exactly who would have a mandate in the Constituent Assembly. The Inkatha Freedom Party also supported proportional representation as they had done in the KwaZulu Indaba discussions of the 1980s. On the white right there was little consideration given to electoral system design. Those parties boycotting negotiations (the Conservative Party and AWB) were more preoccupied with partition and Volkstaats than institutional design, and those elements of the white right which did play some part in the negotiating process (the Afrikaner Volks Front and later the Freedom Front) saw proportional representation as the only option if they were to compete in a non-racial election. An area which received less debate than it deserved was the issue of the level of threshold which should be imposed for representation under the proportional representation system. The ANC and PAC originally had argued for 5 per cent but that demand was dropped in the midst of hard negotiations with the National Party government in 1993.

South Africa

443

The 1994 election When the South African people as a whole spoke for the first time on 26, 27 and 28 April 1994 in overwhelming numbers they said ‘Mandela’. By any objective standard the African National Congress (ANC) won a resounding electoral victory in the first non-racial democratic parliamentary elections in South Africa’s history. The 62.6 per cent of the national vote they obtained was a full 42 per cent ahead of their nearest rivals, the National Party (NP), and left them with a majority of 104 over all others combined in the new National Assembly. Such a high vote could only be achieved as a result of the ANC winning a huge proportion of the votes of the black South African community which in itself was overwhelmingly dominant among the divergent ethnic groups which constitute the people of South Africa. The simple mathematical fact that only 15 per cent of South Africa’s electorate were white and that De Klerk’s National Party managed to win over 20 per cent of the popular vote immediately tells us that many people who had never voted Nationalist before, because they had never been allowed to vote before, did so in spite of years of repressive and blatantly racist apartheid legislation directed at them. De Klerk won the Indian and coloured communities in South Africa by as large a margin as he won the white community and it was only the lack of inroads into the votes of black South Africans which failed to push his party’s vote up even higher than its already considerable base. Chief Mangosuthu Buthelezi’s Inkatha Freedom Party (IFP) could also feel well pleased with their efforts after rejoining the electoral process at such a late stage that they only had five days to ‘officially’ campaign. The IFP’s half share of the KwaZulu-Natal region gave Buthelezi not only a provincial base to rule but also (along with a scattering of votes from outside of KwaZulu) over 10 per cent of the national parliamentary seats, enough to catapult this mercurial political personality back onto the national stage. Even General Constant Viljoen’s Freedom Front could take some comfort in their fourth place, despite that fact that Viljoen’s 400,000 plus (white) votes only translated into 2 per cent (and 9 seats) of the national total. For in the most important respect Viljoen achieved all he set out to do. He split the ‘white right’ down the middle thus mitigating the danger of Afrikaner anti-system violence, bred from exclusion, which certainly was a more pressing possibility before he brought his ‘moderates’ into the democratic process. The only clear losers in this watershed election were the minor parties. The Democratic Party (DP) failed to break out of their suburban white base and so, because of the new realities of the franchise, were resoundingly crushed in the electoral fray. The Pan Africanist Congress (PAC) had far more potential but far less competence. Their election campaign was a textbook case in how to squander an increasingly favourable position through leadership ineptness, detachment from the community and organizational disarray. In the supposed PAC heartland of the Western Cape they failed to outpoll even the tiny African Christian Democratic Party (ACDP) and in the Xhosa-dominated Eastern Cape they came in behind the Democratic Party.

444

Andrew Reynolds

Election 1999 Five years into South Africa’s bold democratic experiment the key election time issues were fourfold. 1. Would the ANC be able to build on their dominance of the electoral arena by winning a two-thirds majority of votes cast, thus having the ability to rewrite much of the constitution without reference to other parties? 2. Would the New National Party be able to retain their position as official opposition without F. W. de Klerk at the helm? 3. Would the NNP be able to hang on to their one provincial government in the Western Cape and make inroads into the Northern Cape? 4. Could the IFP cling to their majority in KwaZulu-Natal? Ironically, Mbeki won a bigger share of the vote than the icon Mandela but whichever way one analyses the result it is unavoidable to note a sweeping ANC victory which positioned that party to be the dominant party of government (much as the LDP was in Japan for nearly fifty years). The ANC had just missed out on the magic two-thirds of the seats but it acheived this goal the following day with a formal alliance with the MF’s single MP. The ANC’s vote share was a little higher than 1994 on a poll of 3.5 million voters less. The Xhosa and Tswana votes they lost were more than made up for by winning increased numbers of coloured votes. The fact that the ANC held on to such a huge chunk of the vote was testament to two factors. First, there was no populist black opposition party which was able to challenge the ANC. Potential challengers such as the PAC and AZAPO were all but wiped off the map. New pretenders such as the UDM and UCDP proved to only have regional appeal and the imagined breakaway of a populist labour party never materialised. Second, the collapse of the NNP facilitated the ANC’s new-found success in attracting the coloured voters of the Western and Northern Cape. The DP also did much better than had been predicted, overwhelming both the New National Party (NNP) and FF for the votes and minds of white South Africans. Their ‘fight back’ campaign was painted as antagonistic and racist by some but it touched a nerve among the more privileged sections of society. Their national vote increased sixfold and their number of MPs shot up from 7 to 38. The DP’s campaign attracted conservative whites to the party like never before and quite remarkably they succeeded in crushing the NNP in the race for white votes taking between 55 and 60 per cent of the white vote compared to roughly 20 per cent for the NNP. The DP’s success was the NNP’s collapse. The party of apartheid, desperately trying to reinvent itself, was decimated in the race for white votes and had to rely upon its stalwart working-class coloured support in the Cape for much of their 7 per cent of the national vote. The NNP’s performance in the 1999 elections under its post-De Klerk leader Marthinus van Schalkwyk was disastrous in most respects. They lost over half of their national vote, 54 MPs, the province they had controlled

South Africa

445

outright, two-thirds of their white vote and a third of their coloured vote. Last, of the main players, Chief Mangosuthu Buthelezi once again demonstrated his endurance by easing the IFP to a level of support not much lower than in 1994 (8.5 per cent) and retained their three cabinet posts and their position as the third largest party in the Assembly.

Consequences of closed list proportional representation The list proportional representation electoral system clearly contributed to South Africa’s transition in three important respects. First, the inclusion of minority parties (especially in 1994) helped forestall anti-system violence from the extremes. Second, proportional representation was an integral part of the much broader framework of consensual arrangements which facilitated the transition from minority rule. Third, the system allowed for ethnic diversity in the voting bases of major parties and their resulting parliamentary caucuses. However, on the negative side it has been argued that the system did little to preclude ‘racial block voting’ (see Davis, 2003), encouraged campaigns based on hostile notions of communal identity (see Jung and Shapiro, 1995) and failed to provide the key components of geographical and personal accountability crucial to any representative system of governance (see Reynolds, 1999a). Partisan inclusiveness Using a plurality single-member district system in 1994, or imposing any threshold for representation within the proportional representation electoral system, the representation of smaller parties in the Assembly would have been dramatically curtailed. As it turned out, in 1994, even with the ANC’s broad electoral hegemony, the National Party and Inkatha Freedom Party would have been likely to pick up a number of seats under the plurality single-member district system because of their vote concentration in the Western Cape and white suburbs, and northern KwaZulu respectively. But all other parties would have been shut out of representation in the Constitutional Assembly. As Table 25.1 demonstrates, the minor change to a low 3 per cent threshold would have excluded the Afrikaner Freedom Front, the Democratic Party, the Pan Africanist Congress and the African Christian Democratic Party from the first Assembly. The inclusion ‘within the tent’ of the Freedom Front and PAC was particularly crucial and was correlated with a sharp decline in the activities of their militaristic anti-system wings. Furthermore, a 5 per cent threshold would have awarded the ANC the two-thirds majority to write the new constitution without reference to other parties. In 1999 a plurality single-member district system would have probably given the ANC an even greater seat bonus (than it might have done in 1994) with the NNP and DP’s votes being more geographically diffused and thus less effectively concentrated to win single-member districts. This time a 5 per cent threshold for representation would have reduced the number of parties in the National Assembly from 12 to 4 and given the ANC over 73 per cent of the seats based on their 66 per cent of the popular vote.

446

Andrew Reynolds

Table 25.1

South Africa: seat allocations under hypothetical PR thresholds Proportional systems threshold None (actual)

ANC DP IFP N/NP UDM ACDP FF UCDP PAC Others Total

1%

5%

1999

1994

1999

1994

1999

1994

266/66 38/9 34/8 28/7 14/3 6/1 3/1 3/1 3/1 5/1

252/63 7/2 43/11 82/20 – 2/1 9/2 – 5/1 –

276/69 40/10 36/9 28/7 14/3 6/1 – – – –

254/64 – 43/11 82/20 – – 9/2 – 5/3 –

291/73 41/10 38/9 30/7 – – – – – –

268/67 – 45/11 87/22 – – – – – –

400/100

400/100

400/100

400/100

400/100

400/100

Note: No. of seats/percentage. See party names in Table 25.3.

Ethnic inclusiveness Much of the analysis of the 1994 and 1999 elections revolved around the question of whether the results demonstrated a mere ‘ethnic census’ of the electorate – that is, ‘black parties’ representing black voters and ‘white parties’ representing whites – and whether or not the electoral system encouraged such an outcome. While there was a high degree of correlation between ethnicity and voting behaviour in 1994 and 1999 this did not imply an unthinking or ‘irrational vote’. Indeed, both on pocket book issues and social issues, it was highly rational for non-white South Africans to vote for the ANC. Indeed it can plausibly be argued that the large-district list proportional representation system, at the very least, left open enough space to encourage the major parties to attempt to appeal to a broad multi-ethnic voting base. The most interesting aspect of the 1999 election was that such multi-ethnic voting bases were slowly beginning to become visible. Three of the main parties became more multi-ethnic and cross-cutting in their support, even if their policy platforms were not. The ANC, NNP and DP (representing 84 per cent of the vote) became more inclusive and diverse in their voting bases in 1999. It is true that the smaller UDM, UCDP, FF, MF and AEB emerged as, or remained, ethnically rooted parties, and the IFP became somewhat more Zulu-based than in 1994, but these parties represented less than 15 per cent of the vote and the trend was clearly in the other direction. In 1994 the ANC won an estimated 80 + per cent of the black South African vote, less than 30 per cent of the coloured and Indian community’s votes along with a handful of white votes. In 1999 the ANC won a little less of the black South African vote, but not enough to injure their dominance, which was more than replaced by their new found success at appealing to middle-class and educated

South Africa

447

coloured voters in the Western Cape who had voted for the NP in 1994. Despite being elected with the support of mainly black South Africans the ANC’s vote was increasingly diverse in its community, language and ethnic base. If registered voters turned out in roughly equal numbers across language groups (as seemed to have been the case) then the ANC’s vote was approximately: 22 per cent Xhosa, 16 per cent Zulu, 13 per cent Pedi, 11 per cent Tswana, 11 per cent Sotho, 7.5 per cent coloured, 7 per cent Tsonga, 4 per cent Swazi, 3.5 per cent Venda, 2.5 per cent Ndebele, 1.5 per cent Indian, 1 per cent white. One of the ironies of the second democratic election was that, despite the DP campaigning quite explicitly for white votes and the ANC’s portrayal of them as the new reservoir for alienated conservative outdated white voters, their dramatically increased vote meant that the DP was somewhat more ethnically mixed than their tiny white liberal vote had been in 1994. Furthermore, the NNP, the party which had designed and operated apartheid, overwhelmingly became a nonwhite party in their electoral base (if not their leadership). The party which dominated South African whites-only elections between 1948 and 1994 struggled to win a fifth of the white vote in 1999. In 1994 it was estimated that the NP vote was based, by a slim majority, on the votes of coloured, black and Indian South Africans. Remarkably they were taking between 60 and 70 per cent of the coloured and Indian vote and a small percentage of the rural black vote. However, in 1999, with a greater collapse of the NNP’s white vote than its coloured vote, the NNP draw approximately 70 per cent of their national vote from those South African’s who they discriminated against less than a decade previously and excluded from the franchise from 1948 to 1994. Such ethnic diversity in voting base was reflected in the diversity of party caucuses in the 1994 and 1999 South African Assemblies (see Table 25.2). The first democratically elected South African Assembly was one of the most diverse parliaments ever assembled. Zulu, Xhosa, Venda, Sotho, Pedi, Tswana, Ndebele and Shangaan speakers sat alongside English- and Afrikaans-speaking whites, coloureds and Indian South Africans. This diversity, to some degree, cut across traditional party lines. In 1994 41 per cent of ANC MPs were Zulu, white, coloured or Indian, while 34 per cent of the NP MPs were not Afrikaners. Only half of Inkatha’s 43 MPs were Zulus (Reynolds, 1999b: 245). Such diversity was reflected in Nelson Mandela’s first cabinet (see Reynolds, 1994). The degree of ethnic diversity declined in the parliament elected in 1999 but whites and Indians are still significantly overrepresented in the Assembly and multi-ethnic party caucuses continue to be the Table 25.2 South Africa: ethnic diversity of the National Assembly members

1994 1999 Population

Black

White

Colored

Indian

208/52 232/58

130/32 104/26

29/7 42/10

33/8 22/5

75%

13%

9%

3%

Note: No. of seats/percentage.

448

Andrew Reynolds

norm. In both 1994 and 1999 the number of women MPs was comparatively high (88/400 in 1994, 120/400 in 1999), placing South Africa in the worldwide top ten for numbers of women in parliament.

Debates over electoral system reform post 2004 As noted earlier the large-district closed-list proportional representation electoral system has increasingly been critiqued and reform proponents have sought to press the government to change the system for elections in 2004 and beyond. Chiefly, unease with the existing system revolves around the belief that elected members of parliament (MPs) are detached from their constituents (having no real ties to a geographical constituency) and that the closed list means that MPs are beholden to party bosses who compile the candidate lists and not the electorate. There are also arguments that the system is so party orientated that it promotes ascriptive block voting along the lines of class and ethnicity. Since 1993 a number of electoral system reforms have been offered – mainly by academics. Because the constitution explicitly prescribes ‘proportionality’ in the national electoral system the main options revolve around either smaller districts with list proportional representation, or personalized proportional representation (mixed member proportional – MMP) systems along the lines of New Zealand and Germany. Related issues include: • Are the lists at the national or provincial level? • Should there be just one ballot counted for both the district and list races or are there two separate ballots? • Are the lists open or closed when it comes to the ability to choose between candidates as well as parties? In 1993 Reynolds proposed 300 small multi-member districts (5–12 members in size), a free list and 100 nationally compensatory seats, all using a single vote. This idea was re-elaborated and re-presented by Elklit and Reynolds in an Electoral Institute of Southern Africa workshop in Cape Town in 1999. Hennie Kotze, the author of the 1992 President’s Council Report on ‘A Proportional Polling System for South Africa in a New Constitutional Dispensation’ similarly argued for 300 districts (3–5 in size) with 100 compensatory seats and he leant toward open lists (a proposal he re-outlined in 1996). In 2002 President Mbeki set up a ‘South African Electoral Task Team’ to investigate possible reform options for National Assembly elections. Chaired by Frederik van zyl Slabbert the commission included academics, ANC appointees and Electoral Commission representatives and considered both smaller-district list proportional representation and MMP. Ultimately, by a 7–5 margin, they recommended a move to a system very similar to the Elklit-Reynolds and Kotze proposals but as the five minority members (who preferred the status quo) were the five ANC representatives, prospects for the government following through with the commission’s recommendation appear weak.

South Africa

Table 25.3

449

South Africa: National Assembly electoral results, 1994–99 1994

Party

Votes / %v

ANC NP/NNP IFP FF DP PAC ACDP UDM UCDP Others

12,237,655/63 3,983,690/20 2,058,294/10 424,555/2 338,426/2 243,478/1 88,104/1

160,175/1

Total

19,533,498/100

1999 Seats / %s 252/63 82/21 43/11 9/2 7/2 5/1 2/1

Votes / %v

Seats / %s

0/0

10,601,330/66 1,098,215/7 1,371,477/9 127,217/1 1,527,337/10 113,125/1 228,975/1 546,790/3 125,280/1 237,396/1

266/66 28/7 34/9 3/1 38/9 3/1 6/2 14/3 3/1 5/1

400/100

15,977,142/100

400/100

Key: ANC: African National Congress (Mandela/Mbeki); NP: National Party (De Klerk); NNP: New National Party (van Schalkwyk); IFP: Inkatha Freedom Party (Buthelezi); FF: Freedom Front (Viljoen); DP: Democratic Party (De Beer/Leon); PAC: Pan-Africanist Congress of Azania (Makwetu/Mogoba); ACDP: African Christian Democratic Party (Meshoe); UDM: United Democratic Movement (Holomisa-Meyer); UCDP: United Christian Democratic Party (Mangope).

References Asmal, Kader (1990) Electoral Systems: A Critical Survey. Bellville, UWC: CDS. Davis, Gavin (2003) Encouraging Exclusivity: The Electoral System and Campaigning in the 1999 South African Election. MA dissertation, Cape Town. Jung, Courtney and Ian Shapiro (1995) ‘South Africa’s Negotiated Transition’, Politics and Society, 23: 269–308. Kotze, Hennie (1996) ‘Proportional Representation in Multi-Member Constituencies’, in Jacques de Ville, Jacques Steytler and Nico Steytler (eds), Voting in 1999: Choosing an Electoral System. Cape Town: Butterworths. Reynolds, Andrew (1993) Voting for a New South Africa. Cape Town: Maskew Miller Longman. Reynolds, Andrew (1994) Election ’94: South Africa: An Analysis of the Results, Campaigns and Future Prospects. New York: St. Martin’s Press – now Palgrave Macmillan; Cape Town: David Philip; London: James Currey. Reynolds, Andrew (1999a) Election ’99: South Africa: From Mandela to Mbeki. New York: St. Martin’s Press – now Palgrave Macmillan; Cape Town: David Philip; London: James Currey. Reynolds, Andrew (1999b) Electoral Systems and Democratization in Southern Africa. New York and Oxford: Oxford University Press. Sisk, Timothy D. (1995) Democratization in South Africa. Princeton, NJ: Princeton University Press.

Further reading • Davis, Gavin (2003) Encouraging Exclusivity: The Electoral System and Campaigning in the 1999 South African Election. MA dissertation, Cape Town.

• Jung, Courtney and Ian Shapiro (1995) ‘South Africa’s Negotiated Transition’, Politics and Society, 23: 269–308.

450

Andrew Reynolds

• Reynolds, Andrew (1993) Voting for a New South Africa. Cape Town: Maskew Miller Longman.

• Reynolds, Andrew (1999) Electoral Systems and Democratization in Southern Africa. New York and Oxford: Oxford University Press.

• Sisk, Timothy D. (1995) Democratization in South Africa. Princeton, NJ: Princeton University Press.

Part 6 Asia and the Pacific

This page intentionally left blank

26 Asia and the Pacific: General Overview Allen Hicken

Asia is home to a wide variety of electoral and political systems. Unlike the Americas, no one system of government dominates in Asia. Parliamentary systems are the most common but presidential and semi-presidential systems are each well represented in the region. Many individual countries have experimented with more than one system over the course of their democratic history. There is less pluralism in terms of the electoral systems. What is striking about Asia vis-à-vis other regions of the world is the dearth of pure proportional representation systems. Out of the 26 Asian states examined in this study only five have used proportional representation to fill all of the seats in the lower or single legislative chamber (Cambodia, Indonesia, Israel, Sri Lanka and Turkey). The rest of the region has relied on various majoritarian rules (e.g. simple plurality or majority rules with a second-round runoff ) or semi-proportional rules such as the single non-transferable ballot. In recent years eight countries in the region have adopted mixed systems that combine majoritarian and proportional rules to elect different portions of the legislature. As will be shown in the following overview, the electoral arrangements of much of Asia were strongly shaped by countries’ colonial experiences. The political and electoral systems adopted at independence tended to reflect the systems used by a country’s colonial power. A few countries have left these original institutions in place throughout their democratic history but most have experimented with alternative ways of electing representatives. Nevertheless, even this experimentation bears the marks of path dependence. Rarely have states made the shift from majoritarian to proportional electoral systems or vice versa. Instead, electoral reforms have generally been confined to changes to the electoral rule, district magnitude or ballot structure within a given proportional or majoritarian framework.

Legacies of colonialism: regime type By far the largest colonizer in Asia was Great Britain – 10 of the 26 countries were under British colonial control (Australia, Bangladesh, Burma, India, Malaysia, Nepal, New Zealand, Pakistan, Singapore and Sri Lanka). Even as these states wrested or negotiated independence from Britain they initially adopted (or continued) 453

454

Allen Hicken

British-style political and electoral institutions. All of the former British colonies began independence as parliamentary democracies. Upon achieving their independence these states uniformly adopted British-style parliamentary institutions where the executive was selected by the legislature and dependent on the legislature’s confidence. Most have remained parliamentary systems, though a few (Bangladesh, Singapore, Sri Lanka) adopted presidential or semi-presidential systems in the years after independence. Parliamentary systems were also the norm in states with a French or Australian colonial heritage (Papua New Guinea, Cambodia, Laos and Lebanon). Asian states within the American sphere of influence were most likely to adopt presidential forms of government closely modelled on the US. Examples include the Philippines, South Vietnam and South Korea. Countries that were part of or strongly influenced by the Soviet Union have also tended to adopt presidential (Georgia) or hybrid systems (Armenia and Mongolia). Indonesia chose a presidential form of government during its independence struggle with the Dutch but after achieving independence in 1949 opted for parliamentarism. Five countries – Israel, Thailand, Turkey, Taiwan and Japan – avoided long-term colonial influence or control (although Japan’s post-Second World War democratic institutions were adopted under US military occupation). When making the transition to elected government, the political elite in Israel, Japan, Thailand and Turkey installed parliamentary government. On the other hand, the leaders of Taiwan’s ruling party, the nationalist KMT, opted for presidential and later semipresidential institutions. In the region’s presidential and semi-presidential systems, voters can either elect the president directly or choose representatives who then elect the president. Among the presidential and semi-presidential systems discussed above, five provide for the direct election of the president via the plurality formula: Bangladesh, Philippines, Singapore, South Korea and Taiwan. Armenia, Georgia and Mongolia, on the other hand, use majority rule with second-round runoff. Sri Lanka also directly elects its president but uses the supplementary vote, a form of preference voting. Each Sri Lankan voter ranks up to three candidates for president in order of preference. If no candidate gets an absolute majority of first preferences all but the top two candidates are declared defeated. The votes of the defeated candidates’ supporters are then transferred to whichever of the remaining candidates they have marked as a second preference. If still no candidate has an absolute majority, third preferences are distributed to the remaining two candidates and the candidate with the most votes is then declared the winner. The supplementary vote was designed to encourage candidates to actively campaign for the preference votes of minority groups. The idea for this unique electoral system had its roots in the French runoff system for electing presidents of the Fifth Republic. However, wary of the cost and security concerns connected with a runoff round of voting, Sri Lanka’s constitutional drafters devised the supplementary vote as a way to combine two rounds of voting in a single election (Reilly, 2001: 117). A few countries indirectly elect their presidents. From 1948 to 1952 the South Korean president was chosen by a two-thirds majority of the National Assembly.

Asia and the Pacific: General Overview

455

In Indonesia the president was chosen by the MPR, an electoral body consisting of the 462 members of the House of Representatives (the DPR) and 103 appointed members. A 2002 constitutional amendment will make the Indonesian president directly elected beginning in 2004. South Korea (1972–87) used a 5,000-member electoral college to elect the president by majority rule. Finally, voters in pre-1996 Taiwan elected an electoral body (National Assembly) which then elected the president on a majority basis.

Legacies of colonialism: the legislature Asian electoral systems also reflect colonial legacies. Without exception each of the British and US colonies in Asia initially adopted majoritarian electoral systems for the lower or single legislative chamber. By far the most common form was the plurality rule in single-member districts. British colonies adopting this rule in a parliamentary context included Australia, Burma, Malaysia, Nepal, New Zealand, Pakistan, Singapore, Sri Lanka and Bangladesh. The only exception is India. In its first two post-independence elections India mixed single-member districts with a smaller number of two- and three-seat districts. For the third election (1962) all districts became single-member districts. Presidential systems adopting singlemember district plurality were the Philippines (first as a commonwealth in 1935 and later as a fully independent state in 1946) and US-backed South Korea and South Vietnam. Papua New Guinea, under Australian Trusteeship, used Australia’s alternative vote until independence. After independence in 1975 the country switched to a unique form of single-member district plurality. The legislature is divided into two tiers – a local tier and a provincial tier. Each tier uses singlemember districts and each voter casts two votes, one for a local candidate and one for a provincial candidate. France’s former colonies also used majoritarian systems to elect their legislature. Cambodia’s first legislature was elected via single-member district plurality while Laos used the plurality rule in a combination of single-member and multi-member districts. Lebanon’s system of majority rule with a second-round runoff was modelled on the French system with an interesting twist. Given the deep divisions within Lebanese society a majoritarian electoral system ran the risk of over-representing some groups while leaving many under-represented. To compensate for the dangers of disproportionality under majority rule the electoral system was designed to ensure that all relevant groups were represented in parliament in a more or less proportional manner. Seat quotas were allocated to the country’s different religious sects with each multi-member district having a fixed number of seats for each sect. Candidates within these districts competed only against candidates from the same religious sect. Voters could cast as many votes as there were seats in a district. In other words, voters were not restricted to voting only for candidates from his/her sect. Within each sect any candidate receiving a majority of the vote (40 per cent in 1951) was awarded a seat. If no candidate received the required majority (or qualified plurality), a second round of elections was held that included any candidate that gained at least 15 per cent of the votes in the first round.

456

Allen Hicken

Other Asian democracies that also relied on majoritarian electoral rules include the former communist states of Armenia and Mongolia along with Thailand, Japan and Turkey. Armenia and Mongolia both held democratic elections in 1990 and both employed the same electoral system – majority rule with second-round runoff and single-member districts. For most of its elections Thailand used the plurality rule in districts with magnitudes ranging from one to as many as 15 (with three-seat districts being the norm). Voters were given as many votes as there were seats in the district and seats were awarded to the candidates with the most votes. Turkey relied on a similar system for electing its legislators from 1946 to 1957, with district magnitudes as small as one and as large as 27. Finally, the first Japanese election in 1946 also used multi-member district plurality but with a limited ballot. Unlike their counterparts in Thailand and Turkey, who were allowed to cast as many votes as there were seats in a district, Japanese voters were given only one, two or three votes in districts with magnitudes ranging from 4 to 14. In stark contrast to the prevalence of majoritarian rules in Asia is the scarcity of electoral systems using pure proportional representation. Only two countries used proportional representation in founding elections – Indonesia and Israel. Indonesia was and is an extremely heterogeneous society. In the first few years after independence Indonesia’s leaders struggled to forge a single country out of a society with pronounced ethnic, religious, linguistic, geographic and socio-economic divisions. Drafters of Indonesia’s electoral law favoured proportional representation with open lists as a means of ensuring representation for most of Indonesia’s diverse population. The country was carved up into 15 districts with an average district magnitude of approximately 17.1 The first and only election under this system was held in 1955. Not surprisingly given Indonesia’s heterogeneity and permissive electoral system, 172 parties stood for election, with 28 winning seats in the new assembly. The largest party controlled only 22 per cent of the seats. The resulting coalition, paralysed by partisan and intra-party conflict, was unable to tackle Indonesia’s growing economic and security problems. Finally, President Sukarno declared martial law in 1957, sweeping away the country’s democratic institutions. Democratic elections did not return to Indonesia until the fall of Sukarno’s successor, President Suharto, in 1998. For Indonesia’s first democratic election in over 40 years drafters once again selected proportional representation. Each of Indonesia’s 27 provinces became an electoral district with district magnitudes from 4 to 82. In addition to much larger districts, the biggest change to the 1999 version of proportional representation is the substitution of a closed list ballot for the open list ballot used in 1955. Forty-eight parties competed in the election with 19 winning seats. The PDI-P emerged as the largest party in the legislature with 30 per cent of the seats. Israel’s version of closed list proportional representation uses a single nationwide district to fill all 120 seats in the legislature. Israel is one of only a few countries that elects all its legislators from one national district. The origins of this unusual electoral system lie in the system of proportional representation used by the Jewish community (Yishuv) during the period in which Palestine was under a League of Nations mandate. The Jewish community was deeply divided. Religious and secular

Asia and the Pacific: General Overview

457

groups clashed over their vision of a Jewish state, there was animosity between Jewish immigrants coming from different countries, and there were stark ideological differences between the political left and right. With each party or group zealously guarding its political independence, nationwide proportional representation was viewed as a way to allow for the utmost representation of a diverse set of groups and interests. An additional attraction was the ability of the electoral system to rapidly and proportionally respond to changes in the electorate. This was considered crucial during a period when the make-up of the electorate was changing rapidly due to immigration. When Israel became a state, the cleavages within the Jewish electorate still existed. Statehood added a new cleavage – Jewish Israelis versus Arab Israelis. Again, national proportional representation was a way to ensure representation for a maximal number of competing groups.

Electoral system reform Unlike in Latin America, majoritarian systems in Asia have not, with a few exceptions discussed below, given way to systems with pure proportional representation. However, many Asian majoritarian states have undertaken significant electoral reform short of a move to proportional representation. After its first election in 1946 Japan switched from multi-member district plurality with a limited ballot to the much criticized single non-transferable vote. Under this system districts consisted of 2 to 6 seats and voters had a single vote for an individual candidate (in the 1947 election there was one single-member district). The candidates with the most votes won the available seats. Modest changes to electoral rules include India’s elimination of multiple-seat districts in 1961 and multiple changes of district magnitude in semi-democratic Thailand. More substantial have been the changes to the type of majoritarian formula used, as occurred in Australia, Lebanon, Mongolia and Singapore.2 Lebanon and Mongolia each switched from majority rule with a second-round runoff to simple plurality systems. In Lebanon this change was part of the strategy of then President Chamoun to curtail the power of the country’s traditional strongmen. These powerful individuals had been able to stifle electoral competition by using their money and influence to compile multi-member electoral lists that proved nearly invincible (Scheffler, 2001: 175). To undercut these strongmen Chamoun combined the move to plurality with a drastic reduction in district magnitude, from an average of 7 seats per district to 22 single-member districts and 11 two-member districts. His government was also the first to extend suffrage to Lebanese women and allowed women to stand as candidates. The move to plurality and women’s suffrage were lasting reforms. The dramatic reduction of district magnitude, however, was gradually reversed over time and by 1960 average district magnitude was near four. In 1990 Mongolia held its first democratic election. The ruling party MPRP maintained its hold on power with 61 per cent of the vote and 83 per cent of the seats. Despite this victory the party’s leadership was worried. Opposition parties had managed to capture nearly 40 per cent of the votes, despite the government’s

458

Allen Hicken

success at preventing many of the anti-communist leaders from standing for election by raising the age requirement for candidates. In a bid to shore up its electoral position prior to the 1992 election the MPRP replaced the existing majority run-off system with a multi-member district plurality system. District magnitude in the new system ranged from 2 to 4 seats and voters had as many votes as there were seats in a district. The new system gave the large and well-organized MPRP a decisive advantage over the smaller opposition parties and this was reflected in the electoral results. While the party’s vote share declined, from 61 to 57 per cent, its seat share increased from 83 to 92 per cent. The backlash to this outcome was swift and strong with critics arguing that the result was potentially dangerous to the country’s fledgling democracy. Bowing to pressure, the MPRP reformed the electoral system yet again, replacing multi-member district plurality with a singlemember district qualified plurality system. This system awards a seat to the plurality winner provided he or she manages to win at least 25 per cent of the vote. If no candidate wins at least 25 per cent then a runoff is held between the top two vote-getters. In 1918 Australia rewrote its electoral law, replacing its single-member district plurality system with the alternative vote system (known as preferential voting in Australia). Like single-member district plurality the alternative voting takes place in single-member districts. Voters must rank all the candidates in order of their preference. A candidate that receives a majority of first preference votes is deemed elected. If no candidate receives more than half of the first-preference votes then the candidate that received the fewest first-preference votes is eliminated and his or her votes are then transferred to the second-choice candidates as marked on the ballots. This process continues until one candidate has achieved an overall majority. Because of its ability to produce majority candidates in every district the alternative vote had been given serious consideration as the country’s first postindependence electoral system. However, because the alternative vote had never been put to the test in an actual election, the drafters of the first electoral law were wary and eventually opted for the better known and hence less risky option of single-member district plurality (Reilly, 2001: 35). Three events prompted Australia’s political elite to reconsider adopting the alternative vote in 1918. First, the alternative vote had been used in state elections for Australian state governments, specifically in Western Australia since 1907 and Victoria since 1911. The success of the alternative vote in these states convinced many that the system could work at the national level. Second, during the first two decades of the 1900s support for the Labor Party steadily grew until the party posed a serious threat to Australia’s established parties. Finally, the rise of a new conservative party after the First World War threatened to split the vote for conservative parties, to Labor’s even greater advantage. The alternative vote, then, was a way to counter the rise of the Labor Party and compensate for the effects of vote splitting between conservative parties. Its adoption encouraged anti-Labour parties to enter into a lasting coalitional arrangement but also allowed minor parties to survive and thrive. Finally, the alternative is generally viewed as a centripetal force in Australian elections pushing parties away from the political extremes and towards the median voter.

Asia and the Pacific: General Overview

459

Like their Australian neighbours New Zealand’s early twentieth-century politicians attempted to manipulate the electoral system in a bid to keep third parties (namely the Labour Party) out of the legislature. In 1908 the single-member district plurality system was replaced with single-member district majority runoff. This system was only used for two elections. Major parties did not do measurably better under the new system and in a few cases Labour candidates who lost the first round were able to win in the runoff election. As a result, single-member district plurality was reintroduced in time for the 1914 election. Singapore’s PAP party, continuously in power since Singapore’s independence, has also been adept at changing the electoral rules to its advantage. Prior to 1988 Singapore was a typical Westminster-style parliamentary system, complete with a single-member district plurality electoral system. In the first four post-independence elections the PAP easily captured every legislative seat. Consequently, the loss of two seats to opposition candidates in 1984 came as a profound shock to the PAP. The PAP reacted swiftly on multiple fronts. It began a public relations campaign, arguing that opposition parties and politicians were unnecessary and unhelpful in Singapore. It appointed members of parliament from its own party to be designated government critics. The party also worked to improve its voter outreach and consultation programmes. Finally, it reworked the electoral system. Gone was the simple single-member district plurality system. Instead, the government took about half of the single-member district seats and converted them into Group Representation Constituencies (GRCs) with three seats in each constituency or district. GRCs can only be contested by teams of candidates from the same party (or allied independents) and at least one member of each team must be from the Malay, Indian or another minority community. Voters cast a single vote for a team and the team with a plurality of the votes wins all the seats in that GRC. The official reason behind the move to GRCs was to improve the representation of minority groups in the legislature. However, many regarded the change as a strategy for curtailing the electoral chances of the opposition. Only the PAP was in a position to attract a large number of minority candidates and field a full team of strong contenders. The fact that GRCs were first adopted in marginal districts also aroused suspicion. As a result of the reform the number of legislative seats controlled by opposition parties fell to 1 in the 1988 election, while the number of minority representatives in parliament remained exactly the same (16) (Tremewan, 1994: 168). In subsequent elections the government has raised both the number of GRCs and the number of seats in each GRC. By the 2001 election only 9 singlemember districts remained versus 14 GRCs with district magnitudes of five or six seats. In total, GRCs now fill 75 of Singapore’s 84 elected seats. Electoral reform within Asia’s proportional representation systems has also been common. When Indonesia readopted proportional representation in 1999 it came with a much higher district magnitude and a switch from open to closed lists. Israel has also changed its electoral formula several times, alternating between the d’Hondt and the Hare quotas. In a bid to prevent extremely small parties from winning seats in parliament Israel introduced a 1 per cent electoral threshold in 1951 and then increased that threshold to 1.5 per cent in 1992. In an attempt to

460

Allen Hicken

strengthen the position of the chief executive the Israeli legislature also experimented with a directly elected prime minister in 1996, 1999 and 2001. The leader of any party contesting the election was eligible to run, provided he or she was nominated by a party or group of parties with at least ten seats in the outgoing legislature. Prime ministers were elected using majority runoff. Three countries in Asia have bucked the trend and completely overhauled their electoral systems – replacing majoritarian rules with proportional representation. In 1991 the Paris Peace Agreement brought an end to two decades of civil war in Cambodia. Included in the agreement were provisions for democratic elections. All of Cambodia’s major political parties preferred some sort of a majoritarian system like Cambodia had used in its past (Hartmann, 2001: 59). Yet the electoral system put in place was a closed list proportional representation system. The reason for the discrepancy between the preferences of Cambodia’s political elite and the electoral rules adopted lies in the strong influence of the United Nations on Cambodian affairs during this crucial transition period. The United Nations took a lead role in pushing the peace process forward and in organizing the postsettlement transition to elected government. The UN Transition Authority in Cambodia, known as UNTAC, was the biggest operation ever undertaken by the UN up until that time. Fortunately, the UN came armed with knowledge and experience gleaned from its mission in Namibia in the late 1980s – a mission very similar to UNTAC but on a smaller scale. Among the things the officials brought with them from Namibia was the UN-designed Namibian electoral system which was adapted for use in Cambodia. The biggest change to the Turkish electoral system occurred in 1961 when Turkey replaced its multi-member district plurality system with proportional representation. The impetus for this change was a military coup in 1960. Over the course of the 1950s Turkey’s majoritarian electoral system enabled a single party, the Democrat Party, to win and keep a majority in the legislature. As the Democrat Party consolidated its single party rule it came under increasing criticism from opposition parties, educated urban elite and segments of the urban and rural poor. The government responded to these criticisms by cracking down on its opponents and aggressively curbing press freedom. In the wake of increased outbreaks of violence in 1960 the prime minister declared martial law, suspending all political activity and outlawing protests against the government. When students in Istanbul defied the martial law and turned out in the streets to demonstrate against the government they were fired on by police and several were killed. The Turkish armed forces, which had historically considered themselves the defenders of the constitution, began to consider intervening in the growing crisis. Finally, the military leadership concluded that the government had strayed from the tenets of the constitution and that the future of the Turkish republic was in jeopardy, and so launched a coup against the government in May of 1960. Once in power the military outlawed the Democrat Party, arrested the prime minister (and other leaders) and tried and executed him for abrogating the constitution and instituting a dictatorship. The military leadership then undertook to rewrite Turkey’s constitution in a way that would clearly separate it from the previous constitution. One means to do this

Asia and the Pacific: General Overview

461

was to overhaul the electoral system. Gone were the highly disproportional majoritarian rules that produced manufactured majorities for the Democrat Party, allowing it to get and hold power as a single party. In their place was a proportional electoral system, designed in part to prevent any one party from ever seizing the reigns of power again. Under the new rules legislators were elected from districts with 1 to 44 seats. A form of open list balloting was used in the 1961 election. Voters were allowed to cast preference votes within a single list but if they opted to cast preference votes (rather than a single vote for a party list), they had to indicate their preference for at least half as many candidates as there were seats in the district. In other words, in the largest district, voters opting to cast preference votes were required to mark preferences for at least 22 candidates. In order to win seats in a district, party lists had to first exceed a district-level threshold equivalent to the Hare quota. The introduction of this new electoral system led to a dramatic fall in disproportionality. It also opened the door to small new ideological or issue-oriented political parties leading to a rise in the number of political parties. Since 1961 no party has been able to win a majority of seats in the legislature, just as the drafters intended. On the other hand, the switch to proportional representation was not a panacea. Along with multiple parties has come the necessity for coalition governments. In Turkey coalition members have often been unable to work together and as a result coalitions have been unstable and short-lived. Since its initial switch to proportional representation Turkey has made many changes to its electoral system. Open lists were used in 1961 and again in the 1973, 1977 and 1991 elections. But in the majority of elections closed lists were the rule. The d’Hondt rule has been the norm in all elections except for one – the 1965 elections in which the Hare quota was used. District magnitude and electoral thresholds (both district and national) have also been the objects of reform since 1982. A 10 per cent national threshold was established in 1983. This was in addition to district level thresholds which were set at the Hare quota in 1983, then raised to 20 per cent for some districts in 1987 and finally lifted to 20–25 per cent for all districts in 1991. All district thresholds were abolished in 1995 but the 10 per cent national threshold remained in place. District magnitude was also reduced dramatically in 1983, from an average magnitude of near 7 to 4.8 in 1983, 4.3 in 1987 and 4.2 in 1991. Magnitude increased again after 1991 with the average magnitude climbing to 6.6 in 1995. Another interesting twist to Turkey’s electoral system was introduced in 1987 – the contingency mandate. In districts with more than three seats one of the seats was allocated using plurality – a kind of bonus seat for the largest party – with the remaining seats allocated using proportional representation. Contingency mandates were used in the 1987 and 1991 elections and then eliminated prior to the 1995 election. Sri Lanka is a third example from Asia of a country that has made the jump from majoritarian rules to proportional representation. Recall that the electoral system in place in Sri Lanka immediately after independence was a single-member district plurality system. 3 For thirty years Sri Lanka held regular elections using this

462

Allen Hicken

system and a relatively stable two-plus party system emerged centred around two Sinhalese parties, the UNP and the SLFP. Over time, however, the demand for electoral reform grew. Frustration over the single-member district plurality system centred on two features common to many majoritarian systems: high disproportionality and seat volatility among the largest parties and, relatedly, chronic under-representation of third parties. Highly disproportional electoral outcomes were a regular feature of Sri Lankan elections with large disparities between parties’ vote and seat shares (manufactured majorities occurred in 1952, 1956 and 1970). The most egregious example occurred in 1970 when the second-place SLPF with 37 per cent of the votes was able to capture more than 60 per cent of the seats. In the following election the electoral system turned a 51 per cent vote share for the UNP into an 83 per cent seat share. Small changes in a party’s vote share often translated to enormous gains or losses in terms of seats. Between the 1970 and 1977 election the SLPF’s vote share declined by only 7 percentage points but its seat share fell from 51 per cent to 5 per cent. Sri Lanka also suffered from another frequent characteristic of single-member district plurality systems – under-representation of minor parties. Under-representation for Tamil parties was seen as especially problematic in the wake of greater demands for autonomy on the part of many Tamils and the growing enmity between Tamils and Sinhalese. Proportional representation was put forward as the best way to guarantee the representation of minority Tamil parties while at the same time doing away with the disproportionality and seat volatility associated with singlemember district plurality. In 1978 the Sri Lankan government adopted a new constitution that included proportional representation for the legislature.4 The version of proportional representation Sri Lanka adopted is unique. Sri Lanka uses a modified form of proportional representation with two tiers of seats. The first tier consists of 196 seats filled using open list preferential voting in districts with 2 to 20 seats. Each voter chooses a specific party list (or independent group list) and up to three of that list’s candidates. The party or independent group that receives the highest number of votes in the district receives one ‘bonus’ seat – awarded to its candidate with the highest vote total. The remaining seats in the district are then awarded to parties (including the largest one) in proportion to the votes received. Parties distribute the seats to the candidates from the party’s list with the highest vote totals. To be eligible for any seats in a given district a party list must garner at least 5 per cent of the vote in that district. The second tier consists of 29 seats that are allocated in a proportional manner based on national party vote shares.

The move to mixed systems The most discernable trend in Asian democracies over the last two decades is the move to mixed systems. Nine Asian countries currently use mixed systems: Armenia, Georgia, Japan, the Philippines, South Korea, Taiwan and Thailand. The defining feature of mixed systems is that one tier of seats is allocated nominally while the

Asia and the Pacific: General Overview

463

other tier is allocated on the basis of party lists (Shugart and Wattenberg, 2001). In the nominal tier members are elected solely based on votes cast for candidates by name. Usually the nominal tier relies on the plurality formula and singlemember districts. However, SNTV with multi-member districts has been used in South Korea and Taiwan. Armenia and Georgia have used qualified plurality with a potential second-round runoff. Mixed systems also contain a tier of seats elected from party lists that overlay the nominal tier. In all but one Asian case this ‘list tier’ is elected using the proportional representation formula. The exception is South Korea, which has used both plurality and a mix of plurality and proportional representation to elect its list tier (discussed below). Mixed systems vary in terms of how they combine the two tiers. At the one extreme, the allocation of seats in each tier can be independent of seat allocation in the other tier, as occurs in parallel systems with two votes per voter in Armenia, Georgia, Japan, the Philippines and Thailand, as well as in two-tier allocation of voters’ single vote, as in South Korea and Taiwan. In each of these cases candidates are awarded seats from single-member districts while parties are awarded seats in the list tier in proportion to their nationwide vote share. No consideration is given to how well or poorly the parties faired in the nominal tier when allocating the list tier seats. South Korea’s form of seat linkage has varied over time. From 1962 to 1972 the party that received 50 per cent or more of the votes automatically received two-thirds of the list tier seats. If no party won 50 per cent of the seats, the party with the largest number of votes automatically received a half of the list tier seats. The second-place party received one-third of the list tier seats as long as its vote share was more than double that of the third-place party. Failing this, the secondplace party received two-thirds of the remaining seats. Finally, any leftover seats were distributed to the other parties via proportional representation. From 1980 to 1987 the party that won the largest number of nominal tier seats automatically received two-thirds of the list tier seats. The remaining one-third of the seats were divided among the other parties on a proportional basis. In effect the systems used for allocating list tier seats in those past electoral systems in Korea made it very likely that the largest party would control a majority of the seats. However, the 1987–96 system formally assured a majoritarian outcome. The party that won the most seats in the nominal tier received enough seats from the list tier to ensure that it had a majority in the assembly. The remaining seats in the list tier were then divided proportionally among the other parties. In 1996 a new system was adopted under which the two tiers are not linked. Parties receive seats from the national tier in proportion to their vote total, with no majority-assuring provisions (Morriss, 1996). What are the circumstances under which states come to adopt mixed systems? In two countries, Armenia and Georgia, mixed systems were established as attempts at democratization. However, in all the other Asia cases mixed systems were the result of reforms of existing electoral arrangements. Why in these cases were mixed systems, with all of the uncertainties associated with a new electoral system, considered superior to the existing electoral system? It is possible to make some

464

Allen Hicken

generalizations. Often electoral reform was linked to a larger political crisis (e.g. Thailand and Japan). But regardless of whether or not there was a full-blown political crisis, in virtually every case the perception that the existing set of electoral institutions was failing was widespread among both the political elite and the masses. The nature of this failure differs from country to country but we can identify two broad categories: highly disproportional and hyper-personalized electoral systems (Shugart, 2001). A different system is the so-called additional-member system, a form of proportional representation used in New Zealand. Voters have two votes, but the party list vote is used to determine the total number of legislative seats to which the party is entitled. Seats from the list tier are allocated so that total allocation of seats from both tiers is as proportional as possible with reference to the party list vote share. New Zealand and to a lesser extent South Korea were highly disproportional electoral systems before the most recent reforms. In both cases parties receiving less than a majority of the votes won a majority of the seats in most elections. These manufactured majorities occurred in 60 per cent of pre-reform Korean elections and 88 per cent of New Zealand’s elections between 1949 and 1993. In addition, New Zealand experienced two cases (1978 and 1981) where the second largest party in terms of vote share actually received a majority of the seats. In short, elections in South Korea and New Zealand frequently produced governments representing far less than a majority of the electorate with serious implications for policy-making and social efficiency. In New Zealand governments often seemed to disregard the will of the majority, confident in their ability to return to office under existing electoral rules. For example, the 1984 Labour government and 1990 National government both forced major policy reforms on an unsuspecting electorate, in some cases directly counter to promises they had made to voters. In short, the high degree of disproportionality in Korea and New Zealand was a powerful argument for electoral system reform in both countries (Nohlen et al., 2001: 30). A proportional representation or a mixed system promised to improve proportionality by encouraging coalition governments and ensuring the representation of smaller parties. It is not surprising then that New Zealand opted for a compensatory system designed to maximize proportionality. Indeed, the new system seems to have worked as advertised. Proportionality has improved in the two post-reform elections. In South Korea the rhetoric for better proportionality ran up against the need for South Korea’s military leaders to secure their position electorally. Thus the mixed systems employed until 1996 were designed to first ensure that the largest party had a majority after which any remaining list tier seats were distributed on a proportional basis. As a result, Korea has not seen the large decline in disproportionality that New Zealand has. Another set of countries adopted mixed systems in response to hyper-personalistic electoral systems. In pre-reform Thailand, Japan and the Philippines, candidates’ personal ties with constituents were the most important factor in an electoral strategy. Candidates were not elected on the basis of their association with a particular party label, but rather based on the particularistic goods and services they were able

Asia and the Pacific: General Overview

465

to provide (promise) their constituents and supporters. In Japan and Thailand (and to a lesser extent the Philippines) the drive to cultivate a personal vote was a reflection of the intra-party competition mandated by the electoral system. In all three countries the electoral systems and the resulting hyper-personalism were associated with corruption, the lack of policy-oriented parties and failures to supply needed collective goods and policies. For these countries a mixed system promised to eliminate intra-party competition via a move to single-member districts which still allowed nominal tier legislators to develop and maintain close ties with their constituents. At the same time a proportional list tier would encourage some legislators to focus on the long-term health and reputation of the party and would also provide incentives for some individuals and parties to develop and pursue a broad, national policy agenda. Disproportionality was not a major concern in any of these three cases and so a compensatory system was not on the reform agenda. Japan was the first to adopt a parallel mixed system in 1994, followed by the Philippines in 1995 and Thailand in 1997. As we survey the region three patterns stand out. First, the nature of postindependence electoral systems was initially shaped by the countries’ colonial experience. Most Asian states initially adopted an electoral system modelled closely on that used by the colonial power. Second, it is striking the extent to which proportional representation remains an exception in Asia. Majoritarian systems dominate the region and, contrary to the experience in other regions, very few countries have shifted from a majoritarian electoral system to proportional representation. Most electoral reforms remain within a given class of electoral arrangements. Finally, to the extent that countries are moving away from majoritarianism, this is being accomplished via the adoption of mixed systems. In other regions of the world majoritarian systems have been gradually replaced with pure proportional representation. In Asia, however, reforms have thus far stopped short of proportional representation, and instead reformers have opted for mixed systems. The move to mixed systems in Asia conforms to the pattern discussed elsewhere in this volume – mixed systems have always been preceded by majoritarian (or authoritarian) electoral systems. In other words, mixed systems appear to have been adopted as an antidote to some of the failings of majoritarian systems in much the same way as proportional representation was adopted in parts of Western Europe and Latin America.

Notes 1. West Irian is listed as a district in Indonesia’s electoral law (for a total of 16 districts), but is not listed in any of the election results as reported by Feith (1957). 2. South Vietnam also changed its electoral formula. For elections during the 1st Republic (1959 and 1963) single-member district plurality was used. For the two elections of the 2nd Republic (1967 and 1991) South Vietnam used the plurality rule in an electoral system made up of a combination of single-member and multi-member districts. 3. A few two- and three-seat districts were occasionally used in an effort to improve the electoral prospects of minority candidates (Wilson, 1980: 80; Reilly, 2001: 115). Voters in multi-member districts could cast as many votes as there were seats and votes could be

466

Allen Hicken

distributed in any manner, including casting all the votes for a single candidate (cumulative voting). 4. The first election under proportional representation was not held until 1989 due to political unrest.

References Feith, Herbert (1957) The Indonesian Elections of 1955. Ithaca, NY: Cornell University Press. Hartmann, Christof (2001) ‘Cambodia’, in Dieter Nohlen et al. (eds), Elections in Asia and the Pacific: A Data Handbook, Vol. 2. Oxford: Oxford University Press. Morriss, Peter (1996) ‘Electoral Politics in South Korea’, Electoral Studies, 15, 4: 550–62. Nohlen, Dieter et al. (eds) (2001) Elections in Asia and the Pacific: A Data Handbook, 2 vols. Oxford: Oxford University Press. Reilly, Benjamin (2001) Democracy in Divided Societies. Cambridge: Cambridge University Press. Scheffler, Thomas (2001) ‘Lebanon’, in Dieter Nohlen et al. (eds), Elections in Asia and the Pacific: A Data Handbook, Vol. 1. Oxford: Oxford University Press. Shugart, Matthew Soberg (2001) ‘Electoral “Efficiency” and the Move to Mixed-Member Systems’, Electoral Studies, 20: 173–93. Shugart, Matthew Soberg and Martin P. Wattenberg (eds) (2000) Mixed-Member Electoral Systems: The Best of Both Worlds? Oxford: Oxford University Press. Tremewan, Christopher (1994) The Political Economy of Social Control in Singapore. New York: St. Martin’s Press – now Palgrave Macmillan

Further reading • Baxter, Craig, Yogendra K. Malik, Charles H. Kennedy and Robert C. Oberst (1991) Government and Politics in Southeast Asia. Boulder, CO: Westview Press.

• Beer, Lawrence W. (ed.) (1992) Constitutional Systems in Late Twentieth Century Asia. Seattle: University of Washington Press.

• Grofman, Bernard, Sung-Chull Lee, Edwin A. Winckler and Brian Woodall (eds) (1999) • • • • • • •

Elections in Japan, Korea, and Taiwan under the Single non-Transferable Vote. Ann Arbor: University of Michigan Press. Hassall, Graham and Cheryl Saunders (eds) (1997) The People’s Representatives: Electoral Systems in the Asia Region. Sydney: Allen & Unwin. Hicken, Allen and Yuko Kasuya (2003) ‘A Guide to the Constitutional Structures and Electoral Systems of East, South, and Southeast Asia’, Electoral Studies, 22: 121–51. Kahin, George McTurnan (ed.) (1964) Governments and Politics of Southeast Asia, 2nd edn. Ithaca, NY: Cornell University Press. Nohlen, Deiter et al. (eds) (2001) Elections in Asia and the Pacific: A Data Handbook, 2 vols. Oxford: Oxford University Press. Reilly, Benjamin (2001) Democracy in Divided Societies. Cambridge: Cambridge University Press. Sachsenroder, Wolfgang and Ulrike E. Frings (eds) (1998) Political Party Systems and Democratic Development in East and Southeast Asia, 2 vols. Aldershot: Ashgate. Sison, Carmelo V. (ed.) (1992) Constitutional and Legal Systems of ASEAN Countries. Manila: Academy of ASEAN Law and Jurisprudence.

Summary Table 6A

Asia and the Pacific: Assembly (lower or single chamber)

Country Law year Election years Armenia 1990: 1990 1995: 1995

1999: 1999

Australia 1902: 1903, 1906, 1910, 1913, 1914, 1917 1918: 1919, 1922, 1925, 1928, 1929, 1931, 1934, 1937, 1940, 1943, 1946, 1949,1951, 1954, 1955, 1958, 1961, 1963,1966, 1969, 1972, 1974, 1975, 1977,1980, 1983, 1984, 1987, 1990, 1993,1996,1998, 2001

Term

Districts

Ballot

Rule/Formula

Single

Majority/2nd round runoff

Threshold

Seats

No.

4

259

259

4

190 150 40

150 1 nat.

1 40

Double Single Closed list

Parallel: Q. plurality (25%)/2nd round runoff Proportional – Hare 5% nat.

131 75 56

75 1 nat.

1 56

Double Single Closed list

Parallel: Plurality Proportional – Hare

3

75

75

1

Single

Plurality

3

75–148

75–148

1

Alternative (single transferable)

Majority/transferable vote

4

Magn.

1

5% nat.

467

(Continued)

Country Law year Election years

468

Summary Table 6A

Term

Districts Seats

No.

Ballot

Rule/Formula

Single

Plurality

Threshold

Magn.

Bangladesh 1971: 1973 // 1979 // 1991, 1996, 2001

5

300

300

1

Georgia 1995: 1995, 1999

4

235 150 85

1–10 85

5–24 1

Double: Closed list Single

Parallel: Proportional – Hare 5–7% nat. Q. plurality (33%)/2nd round runoff

India 1950: 1952, 1957

5

484–494

312–314 87–91

1 2–3

Single Multiple

Plurality

5

520–543

520–543

1

Single

Plurality

Indonesia 1953: 1955 //

4

275 (+15)

15

2–63

Open list

Proportional – Hare

1999: 1999, 2002

5

462 (+38)

27

4–82

Closed list

Proportional – Hare

2003: 2004

5

550

c.80

c.3–12

Open list

Proportional – Hare

Israel 1949: 1949

2

120

1 nat.

120

Closed list

Proportional – d’Hondt

1961: 1962, 1967, 1971 // 1977, 1980, 1984, 1989, 1991, 1996, 1998, 1999

1951: 1951, 1955, 1959, 1961, 1965, 1969

4

120

1 nat.

120

Closed list

Proportional – Hare

1% nat.

1973: 1973, 1977, 1981, 1984, 1988

4

120

1 nat.

120

Closed list

Proportional – d’Hondt

1% nat.

1992: 1992, 1996, 1999, 2002, 2003

4

120

1 nat.

120

Closed list

Proportional – Hare

1.5% nat.

4

464

53

4–14

Limited (1–3)

Plurality

4

466–512

117–130

1–6

Single non-transferable

Plurality

4

480–500 300 180–200

300 11

1 6–33

Double: Single Closed list

Parallel: Plurality Proportional – d’Hondt

Single

Plurality

Single

Two tiers: Plurality Proportional – Hare

5% nat.

Single

Two tiers: Plurality Proportional – Hare

3% nat.

Japan 1946: 1946 1947: 1947, 1949, 1952, 1953, 1955, 1958 1960, 1963, 1967, 1969, 1972, 1976, 1979, 1980, 1983, 1986, 1990, 1993 1994: 1996, 2000

Korea (South) 1950: 1950 //

4

210

210

1

1962: 1963, 1967, 1971 //

4

175–204 153 22–51

153 1 nat.

1 22–51

273–299 227–253 46

227–253 1 nat.

1 46

1987: 1988, 1992, 1996, 2000

4

469

470

Summary Table 6A

(Continued)

Country Law year Election years Lebanon 1943: 1943, 1947, 1951 1953: 1953, 1957, 1960, 1964, 1968, 1972 // 1992, 1996, 2000 Malaysia 1957: 1959, 1964, 1969 // 1974, 1978, 1982, 1986, 1990, 1995, 1999

Term

Districts

Ballot

Rule/Formula

Seats

No.

Magn.

4

30–77

6–9

*7

Multiple

Majority/2nd round runoff

4

44–128

5–26

*4

Multiple

Plurality

5

104–193

104–193

1

Single

Plurality

1

Single

Majority/2nd round runoff

Threshold

Mongolia 1989: 1990 1992: 1992

5

430

430

4

76

26

2–4

Multiple

Plurality

1996: 1996, 2000

4

76

76

1

Single

Q. plurality 25%/2nd round runoff

Nepal 1959: 1959 // 1981, 1986, 1991, 1994, 1999

5

1

Single

Plurality

104–205

104–205

New Zealand 1890: 1890, 1893, 1896, 1899, 1902, 1905

1908: 1908, 1911 1914: 1914, 1919, 1922, 1925, 1928, 1931, 1935, 1938, 1943, 1946, 1949, 1951, 1954, 1957, 1960, 1963, 1966, 1969, 1972, 1975, 1978, 1981, 1984, 1987, 1990, 1993 1993: 1996, 1999, 2002

Pakistan 1970: 1970, 1977 // 1988, 1990, 1993, 1997 // 2002 Papua New Guinea 1975: 1977, 1982, 1987, 1992, 1997, 2002

70–76 (+4)

Plurality 58–64 4

1 3

Single Multiple

3

76(+4)

76

1

Single

Majority/2nd round runoff

3

76–95 (+4)

76–95

1

Single

Plurality

3

120 62–65(+7) 51–55

1 62–65 1 nat.

120 1 51–55

Double: Single Closed list

P-proportional – St-Laguë

4

200–300

200–300

1

Single

Plurality

5

109 1 1

Double: Single Single

Plurality

89 20

≥1

Single/bloc

Plurality

98–100

5% nat.

471

Philippines 1935: 1946, 1949

3

472

Summary Table 6A

(Continued)

Country Law year Election years

Term

Districts

Ballot

Rule/Formula

Threshold

Seats

No.

Magn.

102–200

102–200

1

Single

Plurality

220 206 14

206 206 1 nat.

1 1 14

Double: Single Closed list

Parallel: Plurality Proportional – Hare

5

89–168

89–168

1

Single

Plurality

1978: 1989, 1995, 2000

6

225

22

4–20

Open list

Proportional (2 tiers) – Hare

5% nat.

Taiwan (Yuan) 1991: 1992, 1995, 1998, 2001

3

161–225 125–176 36–49

27–28 2

1–27 6–41

Two tiers: Plurality Proportional – Hare

5% nat.

301–393

155–156

1–3

1953: 1953, 1957, 1961, 1965, 1969 // 1987: 1987, 1992, 1995 1995: 1998, 2001

Sri Lanka 1946: 1947, 1952, 1956, 1960, 1960, 1965, 1970 // 1977 //

Thailand 1978: 1979, 1983, 1986, 1988, 1992, 1992, 1995, 1996

4

3 3

4

Single non-transferable

Single/multiple

Plurality

2% nat.

1997: 2001

4

500 400 100

400 1 nat.

1 100

Double: Single Closed list

Parallel: Plurality Proportional – Hare

5% nat.

Turkey 1946: 1946, 1950, 1954, 1957 //

4

465–610

65–67

1–27

Single/multiple (approval)

Plurality

1961: 1961

4

450

67

1–44

Closed list

Proportional – d’Hondt

1965: 1965

4

450

67

1–44

Closed list

Proportional – Hare

1969: 1969

4

450

67

1–44

Closed list

Proportional – d’Hondt

1973: 1973, 1977 //

4

450

67

1–44

Open list

Proportional – d’Hondt

1982: 1983, 1987

5

400–450

67

2–7

Closed list

Proportional – d’Hondt

10% nat.

5

450

74

2–6

Open list

Proportional – d’Hondt

20–25% dis.

5

450

79–84

2–24

Closed list

Proportional – d’Hondt

10% nat.

1991: 1991 1995: 1995, 1999, 2002

1 quota

Source: See Appendix: Notes and Sources for Summary Tables at the end of the book.

473

474

Allen Hicken

Summary Table 6B

Asia and the Pacific: Presidency

Country Law year

Election years

Term

Rule

Armenia 1991:

1991, 1996, 1998

5+5

Majority/2nd round runoff

Bangladesh 1977:

1978, 1981, 1986

5+5

Plurality

Georgia 1995:

1995, 2000

5+5

Majority/2nd round runoff

1996, 1999, 2001 (Prime minister)

4

Majority/2nd round runoff

4+4+4

Plurality

1987:

1952, 1956, 1960 // 1963, 1967, 1971 // 1987, 1992, 1997, 2002

Mongolia 1993:

1993, 1997, 2001

4+4

Majority/2nd round runoff

4+4

Plurality

1987:

1946, 1949, 1953, 1957, 1961, 1969 // 1992, 1998

Sri Lanka 1978:

1982, 1988, 1994, 1999

6+6

Majority/supplementary votes

Taiwan 1996:

1996, 2000

4+4

Plurality

Israel 1996: Korea, South 1950:

Philippines 1935:

5

6

Source: See Appendix: Notes and Sources for Summary Tables at the end of the book.

27 Australia: Replacing Plurality Rule with Majority-Preferential Voting Marian Sawer

In 1918 farmers’ groups successfully pressured a conservative federal government into adopting a majority-preferential (alternative vote) electoral system for the federal House of Representatives. They threatened that otherwise they would run against conservative candidates, allowing, under the plurality rule, the Australian Labor Party to capture seats. The adoption of the majority-preferential system fostered an effective three-party system which survived for the rest of the twentieth century for the House of Representatives. It enabled the separate political representation of rural Australia, without threatening the major conservative party with which the rural party was usually in coalition. The exchange of preferences between conservative parties disadvantaged the Australian Labor Party, which did not itself become a major beneficiary of preference flows until the rise of post-materialist parties towards the end of the century, particularly in 1990.

The initial rules Formerly comprising six British colonies, in the 1890s Australia invented itself as a federated nation-state, although one that still had a strong British identity. The process included the popular election of delegates to a Constitutional Convention and the ratifying of the resulting Constitution through popular referenda. The former colonies became States of the new Commonwealth of Australia inaugurated in 1901. The new Constitution specified little about the electoral system, except that there were to be no plural votes and both houses of parliament were to be directly elected by the people. In 1902 there was considerable debate over the nature of the franchise and of the electoral systems to be adopted for the new Commonwealth. While the franchise was extended to all women in the interests of uniformity (two colonies had already introduced it) the same uniformity was not granted to Aboriginal Australians, despite the efforts of the government leader in the Senate. In terms of the electoral system, the colonies had already been experimenting with forms of preferential voting, particularly the single transferable vote form of proportional representation in Tasmania and the contingent vote in Queensland (Reilly, 2001). Under single-transferable vote, votes surplus to the quota required 475

476

Marian Sawer

for a seat were distributed to the next preferences indicated, as were the votes of candidates who did not reach a quota after the initial preference distribution. Contingent voting was an early version of preferential voting, in which the preferences for the other candidates were distributed between the top two if no candidate had an outright majority. It differed from the majority-preferential system later generally adopted in Australia in so far as there was no possibility of a candidate ranked third in terms of primary votes winning the seat as a result of favourable preference flows. Although rare, this can happen under majority-preferential voting. Under this system when no candidate has a majority, candidates are eliminated from the bottom and their preferences distributed upward until one candidate does have a majority. Any candidate still in the count can be the beneficiary of preference distribution. Electoral reformers were influential at the time of the design of the new federal political institutions (Uhr, 1999). The Tasmanian Attorney-General Andrew Inglis Clark (one of the fathers of the federation) had given his name to the ‘Hare-Clark’ single-transferable system introduced in Tasmania. Catherine Helen Spence, a leading South Australian reformer and correspondent of John Stuart Mill, had also helped popularize the single-transferable vote and founded ‘effective voting’ leagues in the 1890s. She believed that legislatures would work best if they mirrored the mind of the nation and ensured both that majorities ruled and minorities were heard. The University of Melbourne professor of mathematics, E. J. Nanson, was also an ardent advocate (both in Australia and internationally) of majority-preferential voting when single-member electorates were in place and of the single-transferable vote for multi-member systems. There was a strong push from electoral reformers to adopt these ‘best-practice’ electoral systems for the new Commonwealth parliament and the government included majority-preferential voting for the House of Representatives and singletransferable vote for the Senate in its 1902 Electoral Bill. Professor Nanson’s influence is evident in the drafting of these bills and, in particular, the use of the Droop quota in the single-transferable vote system. Opponents were successful, however, in deleting these provisions in favour of the traditional British plurality system for both houses. Some were concerned that the electoral system should strengthen the operation of a responsible two-party system rather than encourage fragmentation. Plurality rules were seen as most likely to do this, while majority-preferential was seen as taking the risk out of voting for third parties and the single-transferable vote as even more likely to lead to ‘kaleidoscopic’ representation.

Factors favouring new rules There was, however, a continued interest in majority-preferential systems in the light of three-cornered contests, with a protectionist prime minister suggesting its introduction to the Labor leader in 1906, in order to isolate the conservative free traders. Subsequently his government introduced a preferential voting bill, but the proposal lapsed with the 1906 election. The rising Labor Party refused to respond to the overtures of the progressive liberals. As a consequence the latter

Australia

477

were increasingly squeezed under the continuing plurality system and were forced to merge with the conservative free traders in 1909 in a new Liberal Party. The disparate interests gathered under the umbrella of this party (and its variously named successors) included free traders and protectionists, liberals and conservatives. This diversity required flexibility rather than discipline to hold it together (Graham, 1962) and multiple candidacies were common. The names of the main conservative party varied during the twentieth century; at the federal level it became the National or Nationalist Party in 1917 (both names were used), the United Australia Party in 1931 and the Liberal Party again in 1944. The Australian Labor Party, on the other hand, was benefiting from the new party discipline it had introduced – including the ‘pledge’. This did not permit a member of the party to run against an endorsed Labor candidate, unlike the non-Labour side of politics. Where the conservative vote was split between two or more candidates, the plurality electoral system enabled Labor victories. To deal with this situation and to try to frustrate the rise of Labor, there was a continuing interest in the introduction of preferential voting at both state and federal levels. Western Australia and Victoria introduced majority-preferential voting, in 1907 and 1911 respectively, while New South Wales introduced a majority second-round system in 1910. A Royal Commission on Commonwealth electoral law and administration established in 1914 once again recommended majority-preferential voting for the House of Representatives and single-transferable vote for the Senate. It justified the recommendation of preferential voting in the following terms: Under the prevailing party system, electors must either vote for the party nominee or refrain from voting. Political thoughts should not be confined in perpetuity to too narrow channels. There must necessarily be many shades of political opinion, which, in a democratic country, should be given expression to in the freest possible manner. In order that public opinion may be portrayed in distinct broad tones of thought, we strongly urge the adoption of preferential voting for the House of Representatives. (Royal Commission upon the Commonwealth Electoral Law and Administration, 1914: 7) The Prime Minister made a commitment to implement its recommendations in his election policy but was defeated at the polls. The new Labor government had little interest in introducing a change likely to benefit the Opposition. So far the story is that the Australian colonies had a record of electoral innovation and there were many reformers committed to preferential systems of voting. Not only was preferential voting discussed at length in parliamentary forums but there were partisan reasons for introducing it and precedents at the state level. Another important development in this period was the political organization of farming interests. Country-based parties made up largely of wheat farmers were already beginning to emerge before the war, but it was the experience of wartime marketing and price controls that galvanised the development of the Country Party of Australia at the national level. Wartime experience reinforced the belief

478

Marian Sawer

of farmers that they needed direct political representation to protect their interests and to ensure farmer representation on marketing boards; city-based parties could not be relied upon for this purpose. By 1914 there was an eight-member Country Party sitting in the Western Australian Legislative Assembly and by September 1916 various state farmers’ groups, led by the Victorian Farmers’ Union, formed the Australian Farmers’ Federal Organization. By the end of the war they were determined to stand their own candidates in federal seats. Their success in doing so in 1919 gave rise to the federal Country Party. The farmers were interested from the beginning in the introduction of an electoral system that would maximize their own bargaining power as a third party without automatically favouring ‘socialist’ candidates. Most of the state organizations saw themselves as natural allies of the main non-Labor party, but the Victorians always had the more radical position of ‘support in return for concessions’, regardless of the political complexion of the government. From the beginning the farming bodies were a highly effective political force, borrowing the organizational techniques of the Labor Party, if not its trade union ties or class-based political appeal. By 1918 the Victorian Farmers’ Union was bringing pressure on the Prime Minister, who was leading a conservative nationalist government, to introduce proportional representation for the Senate and majority-preferential voting for the House of Representatives. The General Secretary, J. J. Hall, wrote to the Prime Minister reminding him of his ‘promise to the people’ to ‘give at least preferential voting’. In fact the Prime Minister had not mentioned preferential voting in his 1917 election speech, but had been forced into voicing support in response to a planted question. His warning, reinforced by the federal body (the Australian Farmers’ Federal Organization), was as follows: It was with difficulty that organisations such as ours were restrained from running candidates at the last Federal election, when, with three candidates in the field, your cause would inevitably have suffered with the split vote. We cannot guarantee, and are not likely to give further immunity; in fact, our organisation at the last Federal election decided to run candidates in all future elections. This information is conveyed in good spirit, and not as a threat, so that it cannot be said at the last moment that, should you fail to pass preferential voting, we are doing an injustice to you or any other party. (Farmers’ Advocate, 1 February 1918)

The decision to introduce majority-preferential voting As we have seen, there were a number of factors favouring the introduction of preferential voting for the federal parliament. First there was a tradition in Australia of experimentation with electoral systems and already considerable experience with different forms of preferential voting. There was a pool of expertise unmatched in other countries and strong support for the system from electoral reformers.

Australia

479

Second, the rise of a disciplined Labor Party disadvantaged the non-Labor side of politics whose more individualistic philosophy and diverse components encouraged multiple candidacies and the splitting of the conservative vote. Third, non-Labor was in government in 1918 and controlled both houses of parliament, hence it was in a position to introduce electoral reform in its own interest. Fourth, there was the organization of farming interests as a nascent ‘third party’ intent on separate political representation and committed to preferential voting. Fifth, there was the immediate catalyst of rural grievances over price-fixing and the political opportunities presented by an above average number of federal by-elections in 1918. Sixth, the cause of preferential voting had also been taken up by other influential new groups such as the Returned Soldiers’ Political Federation. There were also factors militating against the introduction of preferential voting. One was Westminster ‘loyalism’, particularly on the conservative side of politics. Plurality voting systems were seen as intrinsic to the Westminster tradition of two-party responsible government. The Nationalist Prime Minister, W. M. Hughes, was also suspicious of the farmers’ movement and unsympathetic to the introduction of a voting system that might attract first preferences away from his own party (Souter, 1988: 165). Labor was generally not in favour of preferential voting, as under the plurality rule system it stood to benefit from the relative lack of discipline on the conservative side of politics. At the beginning of 1918 the Victorian Farmers’ Union was heartened by success in standing its own candidates in the Victorian State election, conducted under a preferential voting system. Fortuitously a resignation from the House of Representatives meant a federal by-election was needed in the rural Victorian seat of Flinders. The Nationalists had secured a blue-ribbon candidate, S. M. Bruce, seen as a future conservative leader. This was an ideal opportunity to bring pressure on the government, particularly as the Prime Minister, who had proved particularly intransigent on preferential voting, was overseas. The General Secretary of the Farmers Union, J. J. Hall, was promptly nominated for the seat. The Acting Prime Minister tried to get Hall to withdraw, making a personal offer to introduce a preferential voting bill during the next session of parliament. The Farmers’ Union held out and did not withdraw their candidacy until receiving a formal guarantee from the National Party that the Acting Prime Minister’s offer would be honoured (Graham, 1962: 210; 1966: 127). By then it was two days before the election and so Hall’s name remained on the ballot paper although he was no longer contesting the seat and the seat was duly won by Bruce. Paradoxically, having almost stymied his political career, the Country Party was subsequently responsible for Bruce’s ascent as Prime Minister in 1923. Despite the commitment on preferential voting extracted by the Victorian Farmers’ Union in May, the government was slow to progress its Electoral Bill. The Bill was a complex new piece of legislation, which still provides the basis of Commonwealth electoral administration today. It covered a number of matters recommended by the 1914 Royal Commission, including the restoration of postal voting in a more secure form. One unresolved issue was the Commission’s recommendation that proportional representation be introduced for the Senate.

480

Marian Sawer

In the meantime the political grievances of farmers were exacerbated by the issuing of federal regulations under the War Precautions Act which fixed the price of meat in metropolitan markets. This was seen as yet another example of the selling out of farmers by city-based parties catering to the interests of city voters. A series of protest meetings culminated in a mass demonstration in June 1918 in Melbourne, at that time the seat of the federal government. The death of Lord Forrest, Western Australian explorer and statesman, meant another by-election, this time for the federal seat of Swan. The Western Australia Farmers’ and Settlers’ Association nominated a strong candidate who refused to withdraw, despite a frantic telegram from the Acting Prime Minister and the fact that the Electoral Bill had now been introduced into parliament. As a consequence of the splitting of the conservative vote, the Labor candidate won the Swan by-election with only 34 per cent of the votes. While the conservative press was still lamenting ‘Labor’s cheap victory’, yet another by-election was called, the third of the year (Graham, 1962: 211). The federal government responded to the shock of the by-election loss in Swan and the possibility of another by quickly pushing its Electoral Bill through parliament, using for the first time the ‘guillotine’. This was a new standing order allowing for strict time limits to be imposed on parliamentary debate on ‘urgent’ Bills (Souter, 1988: 165). Labor stonewalled in the Senate in protest, with its Senate Leader making an epic speech lasting 12 hours and 40 minutes. Despite such efforts, the Bill passed, including last-minute amendments to enable the relevant parts to be proclaimed early and take effect in time for the by-election in Corangamite. This first use of compulsory preferential voting for the federal parliament resulted in the Victorian Farmers’ Union candidate winning Corangamite after the distribution of Nationalist preferences and those of two other conservative candidates. Under the plurality system Labor, with 42 per cent of the primary vote, would have won. Instead the Farmers’ Union candidate with only 26 per cent of the primary vote became the first Country Party member in the federal parliament. The rural-based Country Party has been represented in the federal parliament ever since, although in an effort to attract metropolitan support it changed its name to the National Party of Australia in 1984. The Labor candidate had the highest number of primary votes and would have won the seat if the election had still been conducted under the plurality system. The way in which the preferential system operates is for party workers or candidate supporters to distribute ‘how-to-vote’ cards outside polling booths, showing voters how to allocate their preferences in accordance with a party ‘ticket’. Of course voters do not necessarily follow the recommendations of their favoured party, but the percentage that do is sufficient to make the ordering of preferences an important matter of bargaining between political parties. Once preferential voting had demonstrated its benefits for the government, it was also introduced for the multi-member districts of the Senate. The way it worked was the same as for single-member districts in the first instance, with lowest ranking candidates being eliminated and their preferences distributed until one candidate had a majority. However, once the first Senator was elected, their votes were used

Australia

481

again, with all their preferences being distributed. Because the same votes were used to elect the second and subsequent candidates there was a built-in tendency to elect members from the same party. So preferential voting did not address the ‘windscreen wiper’ effect experienced under the previous system (plurality rule in multi-member districts). In 1910 and 1917 all Senators were elected from the winning party (respectively Labor and Nationalists) and, after the introduction of preferential voting, this situation was repeated in 1925 (Nationalists), 1934 (United Australia Party) and 1943 (Labor). While the desirability of proportional representation for the Senate had been widely recognized and had many parliamentary supporters, there were a number of obstacles. There was a Constitutional nexus between the number of seats in the Senate and the House of Representatives (with the latter having approximately twice as many seats as the former). The staggered six-year terms of the Senate meant that only three Senators were elected for each state division at each election, an unsatisfactory number for the purpose of proportional representation. An increase in the size of the parliament through electoral reform in 1948, including an increase to five of the number of senators normally elected in each state, facilitated the introduction of the single transferable vote.

Political consequences of majority-preferential voting The introduction of preferential voting for the House of Representatives in 1918 led to its general adoption in Australia, whether in the form of single-transferable vote for multi-member districts or compulsory or optional preferential voting for single-member districts. The introduction of preferential voting was justified by the Minister of Home Affairs and Territories in his second reading speech in terms of widening electoral choice by allowing voters to choose between candidates of the same party, rather than having this decision made for them by party pre-selection (Glynn, 1918: 6678). His argument was directed against control of pre-selection by the party organization (as in the Labor Party) and the consequent power of the party organization over members of parliament. He suggested that voters should have the right to choose between different candidates of the party they wished to support without this resulting in an opponent’s return. This was put even more bluntly by one of the government senators when preferential voting was being extended to the Senate the following year. He explained that the purpose of the Bill was simply an attempt, by a more scientific method, to ascertain the majority vote, and it will have the advantage of enabling rival candidates of the same party to face the electors without necessarily jeopardising the interests of their party, as has been the case in the past, when owing to a split vote, opposing candidates have been elected on a minority vote. (Russell, 15 October 1919, cited in Reid and Forrest, 1989: 118)

482

Marian Sawer

Labor critics, by contrast, stressed the importance of pre-selection in ensuring clear-cut electoral choice based on party policy. They suggested that preferential voting would weaken the two-party system with its clear policy mandates and disciplined two-party contests. They derided the likelihood of an increased number of political parties or a ‘parliament of faddists’. They also denounced the partisan intent of the legislation. A subsidiary argument in favour of preferential voting was its democratic credentials in enabling elections to be decided by the majority of electors rather than a simple plurality. This was regarded as congruent with Australia’s other democratic innovations such as the secret ballot, universal suffrage, a popularly elected upper house, provision and funding of electoral machinery by the state and ‘registration facilities in keeping with the scope of the franchise’ – i.e. continuous enrolment (Glynn, 1918: 6670). The distribution of preferences ensured that the outcome reflected majority choices or second choices. While majority-preferential voting may have the advantage over plurality rule systems in terms of reflecting majority choices in individual electorates, it does not ensure a match between votes and seats. For most of the twentieth century the Labor Party suffered from the concentration of its vote in industrial seats, meaning it had huge and wasteful majorities in these seats rather than having its vote spread more evenly across the country. This is a problem endemic to singlemember electorate systems rather than being specific to any one form of them. The introduction of preferential voting certainly enabled the non-Labor parties to field multiple candidates without political costs, and they continued to do so until the introduction of greater discipline by the new Liberal Party after 1944. The main consequence of the introduction of preferential voting has, however, not been to support multiple party candidacy but to foster the emergence of a three-party system. Both Nationalist and Labor voters at first gave Country Party candidates their second preferences in triangular contests, enabling Country Party victories sometimes from primary votes of less than 33 per cent (Graham, 1962: 215). After the 1919 general election 11 Country Party members took their seats on the cross benches and held the balance of power in the 75-member House of Representatives. They were determined to influence the Nationalist government in every way possible in favour of their own policies on economy in public expenditure, tariffs, the removal of price controls and the restoration of the prerogatives of parliament (Page, 1963: 62–75). After the next election they insisted on a change in conservative leadership as the price of their participation in a coalition government. Not only did preferential voting facilitate the rapid establishment of the Country Party as a significant presence in Australian parliaments, it also enabled its survival as a rural-based or ‘agrarian’ party long after the disappearance of such parties elsewhere. This is despite the century-long shift in population from rural areas – today farmers make up less than 5 per cent of the Australian workforce. Preferential voting gave the Country Party the leverage to insist on another electoral benefit – the weighting of electorates so that rural electorates had far fewer voters than metropolitan ones. At the federal level this was gradually rectified in the

Australia

483

1970s and 1980s. In 1974 the tolerance for variation in federal electorates was reduced from 20 to 10 per cent and in 1983 a new redistribution principle was introduced requiring redistributions to aim at equal enrolment halfway through the seven-year redistribution cycle. This meant anticipating population movements, so that rural seats now began the seven-year cycle with more rather than fewer electors. Majority-preferential voting favoured the political representation of a geographically concentrated party such as the Country Party and enabled supporters to vote for it without incurring the risk that such votes would result in ‘cheap victories’ for their least-preferred electoral choice. Because, except in Victoria, the Country Party was generally aligned with the main conservative party, the preferential system generally favoured conservative interests. As has been said by others, preferential voting rewards collaboration or coalition arrangements between parties, enabling the accumulation of votes between distinct but aligned parties (Reilly, 2001). The conservatives also benefited when the Labor Party split in the 1950s and the preferences of the anti-communist ‘Democratic Labor Party’ were channelled to the non-labour side of politics. Traditional labour voters, who would have found it difficult to vote directly for the ‘class enemy’, instead voted for the Democratic Labor Party which in turn directed their preferences to the non-Labor coalition. Conservative governments stayed in power after the 1961 and 1969 federal elections thanks to such preference flows. Of the 100 cases between 1969 and 1996 where the distribution of preferences changed outcomes, 81 favoured the conservative coalition (Hughes, 1997: 166). Because of these political effects, the Labor Party has in the past been suspicious of preferential voting and favoured optional rather than compulsory preferential systems. It was responsible for introducing optional preferential voting in New South Wales in 1981 and Queensland in 1992. Optional preferential forms operate more like plurality rule, leaving the marking of preferences up to the voter. Compulsory preferential, by contrast, requires the rank-ordering of all candidates for the vote to be formal. This is to preclude the possibility of preferences being exhausted before one candidate has a majority. Optional preferential voting minimizes the informal vote by those with low levels of literacy and this may favour the Labor Party. On the other hand, if a Labor government has become unpopular and protest votes are cast for other candidates, a compulsory system may push such voters to return to Labor through the full expression of preferences.2 The rise of post-materialist political parties (the Australian Democrats and the Greens) in the last two decades of the twentieth century also enabled the Labor Party to benefit more than the conservative parties from preference flows. The Labor Party does not currently have any policy to change the compulsory preferential system used for the House of Representatives. In the 1990 federal election minor party support was at a record level, and the Labor government pursued a ‘second preference’ strategy to woo minor party supporters on environmental and social justice issues. They were successful in obtaining two-thirds of the preferences of Australian Democrat and Green voters,

484

Marian Sawer

enough to determine the outcome of the election. For example, the National Party leader was defeated by a large flow of preferences from an anti-nuclear campaigner to Labor, which won the seat with only 27 per cent of the primary vote (Reilly, 2001: 88). The Australian Democrats and the Greens were new-style participatory parties, expressing the philosophies of the new social movements. Their electoral support was not geographically concentrated like that of the Country (National) Party, and so it was almost impossible for them to achieve direct political representation under majority-preferential systems. Increasingly, however, Australia’s bicameral parliaments (at both federal and state levels) have one chamber elected by single transferable vote, and post-materialist parties have obtained direct representation and often a balance of power position in these chambers. This has increased their profile and increased their political leverage. They have allocated their preferences, which may be crucial in determining lower house seats, in exchange for policy commitments as well as preference allocations to benefit their upper house candidates. Under majority-preferential voting it is the preferences of the minor parties that are more likely to be distributed than those of the major parties, whereas under single transferable vote it is major party preferences that often determine the final seat or seats. From the mid-1990s Australia has seen the appearance of populist parties such as ‘One Nation’ to the ‘right’ of the major parties, as well as the Australian Democrats and Greens to the ‘left’. Preference flows from the latter continue to favour Labor, while those from the former benefit non-Labor. One should, however, be careful not to exaggerate the partisan effects of preference distribution. The proportion of seats where the outcome is changed by preference distribution fluctuates, but is generally under 10 per cent (see Table 27.1). Some argue that preferential systems have a systemic effect in driving politics towards the middle ground because of the need for pre-election bargaining to obtain preference flows from other parties (Reilly, 2001). The best-known example of this dates from the decision of both the major parties to place the populist One Nation party, rather than their main opponent, last on their ‘howto-vote’ cards in the 1998 federal election. This resulted in the One Nation leader, Pauline Hanson, being defeated by a major party candidate who had come third in terms of the primary vote. In general, however, the major parties have been Table 27.1

Australia: House of Representatives, 1990–2001

Election year

Preferences distributed % seats

Outcome changed % seats

60 42 39 66 58

6 8 5 5 4

1990 1993 1996 1998 2001 Source: Hughes (1997: 167) (updated).

Australia

485

Table 27.2 Australia: transitional electoral results, 1914–25 Year

Labor Country National-Liberal Independents v/s v/s v/s v/s

1914 1917

51/56 44/29

– –

47/43 54/71

2/1 2/0

1919 1922 1925

43/35 42/39 45/31

9/15 13/19 11/19

45/49 40/41 42/49

3/1 5/1 2/1

Key: v/s: percentage of votes/percentage of seats.

moving toward the right, under the influence of populist discourse, rather than being drawn to the middle by the need to negotiate for the preferences of minor post-materialist parties. Nonetheless, majority-preferential voting is today firmly entrenched in Australian political culture. It is seen as more democratic than plurality systems, in that outcomes are determined by the wishes of the majority. It is seen as less effective than single transferable vote in achieving proportional and minority representation, but more like plurality systems in achieving a clear-cut outcome for one of the major parties. For this reason Australia generally uses majority-preferential voting for the chamber of parliament in which government is formed and single transferable vote for the chamber that is intended to act as a ‘house of legislative review’. This appears to be an optimal institutional design for Australian purposes. This is particularly the case where single-transferable vote results in neither government nor opposition controlling the upper house, but rather minor parties and/or independents with a vested interest in increasing the power of the legislature vis-à-vis executive government. In the Australian case the long-term political and institutional consequences of majority-preferential voting can only be viewed in terms of interaction with other voting systems used for the same parliament, not in isolation.

Notes 1. My thanks to Gillian Evans for her invaluable research assistance and to Peter McCarthy for his scrupulous editorial work. 2. Personal communication from Senator John Faulkner, Labor Leader in the Senate and Shadow Minister for Public Administration and Home Affairs, 22 January 2003.

References Glynn, P. (1918) Second Reading Speech on the Electoral Bill, Commonwealth of Australia Parliamentary Debates, House of Representatives, 4 October: 6669–82. Graham, B. D. (1962) ‘ The Choice of Voting Methods in Federal Politics, 1902–1918’, Australian Journal of Politics and History, 82: 164–81. Graham, B. D. (1966) The Formation of the Australian Country Parties. Canberra: Australian National University Press.

486

Marian Sawer

Hughes, Colin A. (1997) ‘Individual Electoral Districts’, in C. Bean et al. (eds), The Politics of Retribution: The 1996 Federal Election. Sydney: Allen & Unwin. Page, E. (1963) Truant Surgeon. Sydney: Angus & Robertson. Reid, G. S. and M. Forrest (1989) Australia’s Commonwealth Parliament 1901–88. Melbourne: Melbourne University Press. Reilly, Ben (2001) ‘Preferential Voting and Its Political Consequences’, in Marian Sawer (ed.), Elections: Full, Free and Fair. Sydney: Federation Press. Royal Commission upon the Commonwealth Electoral Law and Administration (1914) Report. Melbourne: Government Printer for the State of Victoria on Behalf of the Commonwealth. Souter, G. (1988) Acts of Parliament. Melbourne: Melbourne University Press. Uhr, J. (1999) ‘Why We Chose Proportional Representation’, in Marian Sawer and Sarah Miskin (eds), Representation and Institutional Change. Canberra: Department of the Senate, Parliament House.

Further reading • Hughes, Colin A. (2000) ‘STV in Australia’, in Shaun Bowler and Bernard Grofman (eds), • • • • • • •

Elections in Australia, Ireland, and Malta under the Single Transferable Vote. Ann Arbor: University of Michigan Press. Hughes, Colin A. and B. D. Graham (1968) A Handbook of Australian Government and Politics 1890–1964. Canberra: Australian National University Press. Sawer, Marian (ed.) (2001) Elections: Full, Free and Fair. Sydney: Federation Press. Sawer, Marian and Sarah Miskin (eds) (1999) Representation and Institutional Change. 50 Years of Proportional Representation in the Senate. Canberra: Department of the Senate, Parliament House. Sawer, Marian and Gianni Zappala (2001) Speaking for the People: Representation in Australian Politics. Carlton South, Vic.: Melbourne University Press. Wright, Jack F. H. (1980) Mirror of the Nation’s Mind: Australia’s Electoral Experiments. Sydney: Hale & Iremoner. Wright, Jack F. H. (1984) ‘An Electoral Basis for Responsible Government: the Australian Experience’, in Arend Lijphart and Bernard Grofman (eds), Choosing an Electoral System. Issues and Alternatives. New York: Praeger, pp. 127–34. Wright, Jack F. H. (1986) ‘Australian Experience with Majority-Preferential and QuotaPreferential Systems’, in Bernard Grofman and Arend Lijphart (eds), Electoral Laws and Their Political Consequences. New York: Agathon, pp. 124–38.

28 India: Majoritarian Democracy from Above Subrata K. Mitra

India belongs to a minority of post-colonial states that have made a successful transition from colonial to a stable democracy. But the path that India has taken to reach this coveted status is radically different from her western counterparts. Western democracies are the products of popular movements from below, resulting from industrial revolution, leading to the slow but inexorable rise of working men and subsequently women into political visibility, their violent struggle for participation, and finally the achievement of universal adult franchise. In contrast, there was no full-scale industrialization in India before independence. Large-scale changes in society and the economy lay in the Indian case not behind but ahead, as a beacon and a worthy goal for the nascent post-colonial state. Hence the puzzle: how could a democracy from ‘above’, based on alien values of individualism and institutions of foreign origin such as political parties and the rule of law, strike root on Indian soil, strengthening the foundations of the form of parliamentary democracy adopted at independence? Or, even more pointedly, how could the plurality rule first-past-the-post electoral system based on singlemember districts survive intact in spite of the apprehensions that this particular form of vote counting creates among minorities? A detailed answer to these complex questions is beyond the remit of this chapter. Instead, it concentrates on the catalytic role of the electoral system in the transition from colonial rule to popular democracy.

The odds against An electoral system based on plurality rule in single-member districts has a tendency to favour large organized groups competing against smaller groups. The logic of the system had already alarmed India’s Muslims and Christian minorities about the potential for an electoral tyranny of the majority Hindu community when the British rulers initiated the incremental devolution of power through the introduction of elections based on limited franchise during the early decades of the twentieth century. In contemporary India, with a powerful Hindu nationalist party in the fray, minority groups and parties can be reasonably expected to invoke Christ’s prophecy: ‘For whosoever hath, to him shall be given, and he shall have more 487

488

Subrata K. Mitra

abundance: but whosoever hath not, from him shall be taken away even that he hath’ (Rae, 1971: 134). In the event, however, the functioning of India’s electoral system has set at rest the worst of these apprehensions, causing Arendt Lijphart to comment: ‘A deeply divided society with, supposedly, a mainly majoritarian type of democracy, India nevertheless has been able to maintain its democratic system’ (Lijphart, 1996: 258). The sophisticated system of power-sharing that accounts for the efficacy and legitimacy of the electoral process will be taken up later in this chapter. First we shall consider some of India’s achievements, significant by any standards but particularly by the standards of developing countries (Malhotra, 1998; Enskat, Mitra and Singh, 2001). Regular elections have been held for the office of the President and to legislatures at national and regional levels from 1952 onwards (see Table 28.1). Parties in power have changed at the centre and in the regional states frequently as a result of elections. India’s elections have been generally free and fair though not devoid of violence or allegations of irregularity, which are routinely investigated by an independent Election Commission (Basu, 1985: 355) This overall picture of a functioning electoral system, however, has regional variations, for the functioning of a free and fair election system is hampered by the prevalence of ethnic conflict and mass insurgency in some parts of India. The constitution provides for direct rule from the centre through the imposition of President’s rule (Article 356), subject to parliamentary approval, and can normally last for a maximum of three years. The imposition of a State of Emergency on the whole country is also possible. The most well known of such incidents was the Emergency of 1975–77 when Emergency Rule was imposed and elections to the parliament postponed by a year at the behest of Prime Minister Indira Gandhi. On the whole, however, the abuse of executive privileges with regard to the electoral process has been kept under check due to a combination of factors such as judicial review, popular vigilance through the media, civil liberties groups and protest movements, and the countervailing forces of the political system resulting from the separation of powers and the federal division of powers. Attitudinal evidence of the successful functioning of India’s electoral system based on survey data suggests that the main hypothesis that accounts for the successful functioning of an electoral system is that political contestants will be bound by rules in the long run if and only if they are persuaded that the rules are fair (in the sense of being non-discriminatory), efficient in promoting their preferences, transparent and give free access to information about rules as well as the sanctions against rule-breaking (Mitra, 1999). Before we can analyse the evaluation of the electoral process by the electorate, we need to have a glimpse of its institutional arrangement and the opportunity structure it creates for the political actors. The Constitution of India describes the state as a socialist, secular and democratic republic, committed to holding regular, free and fair elections based on universal adult franchise. The elections determine the composition of the government at the centre, the constituent states of the federation, municipalities and the three tiers below the regional level, going down all the way to the village. The Constitution vests the responsibility for the

Table 28.1 Party:

1952 1957 1962 1967 1971 1977 1980 1984 1989 1991 1996 1998 1999

India: Lok Sabha electoral results, 1952–71, 1977–98

Communists

Socialists

CPM v/s

CPI v/s

v/s

– – – 4/19 5/25 4/22 6/36 6/22 6/33 6/35 6/32 5/32 5/33

3/26 9/29 10/29 5/23 5/23 3/7 3/11 3/6 3/12 2/14 2/12 2/9 1/4

16/21 10/19 9/18 8/36 3/5 – – – – – – – –

Congress

INC-2 v/s

INC-1 v/s

– – – – 10/16 2/3 5/13 2/5 0.3/1 0.4/1 2/4 – –

45/364 48/371 45/361 41/283 44/352 34/154 43/353 48/415 40/197 37/244 29/140 26/141 28/114

Nationalists

BKD/LD/SJP v/s – – – – 2/1 – 9/41 6/3 – 3/5 3/17 0.3/1 –

Swat. v/s

JP/JD v/s

BJS/BJP v/s

– – 8/18 9/44 3/8 – – – – – – – –

– – – – – 41/295 19/31 7/10 18/142 11/59 8/46 3/6 4/22

3/3 6/4 6/14 9/35 7/22 – – 7/2 11/86 20/120 20/161 26/179 24/182

Regional.

Indep.

Others

v/s

v/s

v/s

16/38 19/42 11/20 14/35 8/14 5/9 6/9 8/5 5/12 4/1 6/9 2/6 3/6

2/32 1/9 2/14 1/13 4/11 1/3 1/1 1/1 6/19 2/4 3/4 15/65 9/43

14/14 6/20 9/20 9/32 8/41 9/49 8/34 13/73 10/27 13/51 21/118 19/100 21/125*

Total seats

498 494 494 520 518 542 529 542 529 534 543 539 529

Key: Communists: CPI – Communist Party of India; CPM – Communist Party of India/Marxist. Socialists: Socialist Party; the Kisan Mazdoor Party; the Praja Socialist Party; the Samyukta Socialist Party. Congress: INC-1: Indian National Congress to 1967, Congress/Requisitionist to 1971, Congress/Indira to 1980; INC-2: Congress/ Organization, Congress/Urs 1980, Congress/Socialist 1984–. Nationalists: BKD – Bharatiya Kranti Dal; LD – Lok Dal; SJP – Samajwadi Janata Party: Swat. – Swatantra; JP – Janata Party; JD – Janata Dal; BJS – Bharatiya Jana Sangh; BJP – Bharatiya Janata Party. v /s: percentage of votes/percentage of seats. * This figure does not include the Bahujan Samaj Party, with 4% vote and 14 seats in 1999, which was classified as a regional party for all the other years.

489

490

Subrata K. Mitra

‘superintendence, direction and control of the entire process for conduct of elections’ in the Election Commission of India. All elections are conducted according to the provisions laid out by the law, most important of which is the Representation of the People Act 1950 which deals with the preparation and revision of electoral rolls, and the Representation of the People Act 1951 which provides for the detailed administrative arrangements related to the conduct of elections and the procedures related to post-election disputes. These rules constitute the basic legal structure for the functioning of elections. Elections to 543 seats in the Lok Sabha, the lower house of the national parliament, take place on the basis of plurality rule in single-member districts, usually called ‘first-past-the-post’. Two seats, set aside for Anglo-Indians, are filled in by nomination by the President of India. The members of the Rajya Sabha, the upper house of the parliament, are elected indirectly by members of state legislatures, using the single-transferable vote system. The President of India is elected by an electoral college consisting of elected members of the state assemblies and both houses of the parliament. The formula used for calculating the number of votes that members of the electoral college can cast seeks to provide equal weight to state assemblies and the national parliament on the one hand, and balances the states against one another in view of the great discrepancies in their size. The combination of plurality rule in elections to the lower house of the national parliament and state legislatures with single transferable votes to the presidency and the upper house gives a sense of proportionality to counterbalance the fears of majoritarian dictatorship and to dispel the notion that there are permanent winners or losers. This is of great importance to the prospects for electoral democracy in postcolonial countries where the founding fathers often feel tempted to substitute the popular will with their personal vision. The context of independence from British rule gives an insight into why India escaped this tragic fate. Partly under the impact of utilitarianism but mostly as a matter of expediency, the British had started experimenting with limited self-rule by the 1880s. This formed part of the British strategy of ruling India with the help of Indian intermediaries selected by a very restricted electorate of urban, rich and loyal subjects. The Indian National Congress was set up in 1885 by Sir Alan Octavian Hume, a retired British civil servant, in order to present Indian interests to the British Crown in a systematic and organised manner. It soon became the leading voice of the Indian middle classes, constantly clamouring for more jobs in native hands and for greater political participation. Thanks to the British strategy of incremental devolution of power as a method of effective governance with the minimum use of force, democratic institutions spread in India during the sixty years before independence. Six sets of national and provincial elections that took place between 1920 and 1944 acquainted Indians with the rules of the electoral game, gave people a taste of electioneering as an effective method of pursuing power and prestige, and built up an institutional knowledge of elections. The principle of elected representation was first recognized at the level of municipal councils through the 1882 Resolution on

India

491

Local Self-Government, confined to a small elite which gradually expanded. The 1892 Indian Councils Act was a step towards representative self-government at the national and provincial level though filled mostly by nomination. The 1909 Morley-Minto reforms extended the number of seats in provincial councils and in the Indian Legislative Council, but still with a strictly limited franchise based on wealth. The elections were indirect, and based on separate electorates for Hindus and Muslims, which guaranteed some proportional representation. The 1919 Government of India Act for the first time introduced the principle of direct elections, though based on limited franchise to about 3 per cent of the population. While the amount of real power that devolved to the colonized population was not significant, the separate electorates had the effect of mobilizing Indians on religious lines. There were provincial elections in 1920, 1923 and 1926. The last provincial elections under diarchy were held in 1929–30. The 1935 Government of India Act led to the 1937 provincial elections which led to a massive victory for the Congress party in eight out of eleven provinces. No elections were held at the federal level. The election held different meanings for the Congress and its nemesis, the Muslim League. The former, thanks to its success, saw elections and parliamentary democracy as the way forward, whereas the latter came to the conclusion that its only chance lay in the partition of India along religious lines which alone would produce guaranteed power for Muslims in a state created solely for them. In consequence, following independence and the partition of India, the leaders of the Muslim League migrated to Pakistan and the Congress, its links to the constituencies intact, went on to establish what Kothari has described as the Congress System. The ‘untouchable’ castes, considered polluted on account of their functions such as scavenging by upper-caste Hindus, also had their reservations against the simple majority system and, under the leadership of B. R. Ambedkar, a gifted lawyer, pleaded for separate representation. But a compromise was arrived at between their representatives and the Congress Party as a result of the Poona Pact 1932 which guaranteed a minimum representation for the members of these communities under a scheme of reserved seats in legislatures and public services. This method managed to gain their support for a plurality rule electoral system in India. The process was not as effortlessly incremental nor linear as it may sound. Periods of extensions to the franchise and cooperation between colonial rulers and elected representatives were interspersed with ruthless suppression and imprisonment of Indian leaders. The Congress Party itself was often divided in its opinion between collaboration with colonial rule with the hope of greater concessions and radical resistance to it on the agenda of independence. Gandhi brought these two strands together in his strategy of non-violent non-cooperation and built a powerful mass movement that united the peasantry and the national bourgeoisie under the banner of the Congress Party. By the 1930s, however, the national movement was split once again, this time on the issue of religion. The majority of Muslims, under the leadership of the Muslim League, had started agitating for an independent homeland for the Muslims of the subcontinent.

492

Subrata K. Mitra

The partition of India paved the way for the modernists under the leadership of Jawaharlal Nehru, the first Prime Minister of post-independence India, who saw the chance to mould the destiny of the nation in parliamentary democracy and majority rule, without being hemmed in by the constant objections and bickering of the Muslim League. On the other hand, the fact that the organization of the Congress Party survived the partition of the country almost intact ensured that the majoritarian idealism of the parliamentary leadership would be balanced by the realism of surreptitious proportionality espoused by the party organization. Well trained in the art of consensus, accommodation and short-term alliances, the Congress Party machine neatly reflected the geographic and social diversity of India. The immediate post-independence consequence of the introduction of universal adult franchise was the evolution of what we would call the ‘first party system’ with the Indian National Congress as the dominant party, to distinguish it from the ‘second party system’ when India firmly moved in the direction of a multi-party democracy.

Elections and the party system, 1952–99 The Indian National Congress formed the government at the centre and in all Indian provinces in the first general election held after Independence. Though parties of the left and right routinely took part in elections the fragmented character of the opposition and the combination of the ‘first past the post’ system in single-member constituencies systematically resulted in a Congress majority in the legislature. This hegemony of the Congress has caused this period to be described as the period of one-party dominance, and the party system as the ‘Congress System’. The opposition parties were present as active players in the parliament and in national politics but their role was confined to influencing policy from the sidelines of the institutional process, rather than making policy and alternating with the Congress in holding ministerial office. The power of the Congress declined sharply in the fourth general election of 1967 when the first coalitions of opposition parties took place at the regional level, leading to the breakdown of the dominance of the Congress Party in several states. At the national level, the Congress Party won, but with a reduced majority. The situation changed radically after the split of the Congress Party in 1969 into the Congress (Requisionist) and the Congress (Organization). The faction led by Indira Gandhi, referred to as Congress (R), brought about radical changes in the programme of the centrist Congress Party. These policies brought the party great electoral success in 1971 but the populist style adopted by its leader led to the corrosion of its organizational links with the electorate. In retrospect, the period 1967–77 can be thought of as a period of transition from one party dominance to multi-party democracy. The setback suffered by the Congress Party in the election of 1967 demonstrated the vulnerability of the centrist Congress to broad electoral coalitions of the left and the right. After its initial setback, however, Congress, under Indira Gandhi’s forceful leadership, turned its new policy of radical, populist leadership into its main asset. Its initial

India

493

success in the 1971 election was further reinforced in the assembly elections of 1972 when Indira Gandhi used India’s successful intervention in East Pakistan, leading to the birth of Bangladesh, in the electoral platform of the Congress Party. However, the radical rhetoric rebounded on the party when a number of interest groups including industrial workers, railway employees and students started political agitation. The culmination of this period of unrest was the authoritarian interlude of 1975–77. Elections were announced in 1977 when the regime got the impression that because of a significant increase in law and order, food supply and general prosperity, the election would lead to political gains. However, the elections reduced the Congress to a minority in the Lok Sabha for the first time, paving the way for the transition to a multi-party sytem and the politics of coalitions. The general elections of 1977 ushered in a new period of multi-party coalitions at the centre and stable two-, three- or four-party systems at the regional level. The transformation of the vote–seat ratio from one that favours the largest party to one that has a more proportional relationship between votes cast and seats obtained is the main lesson of the electoral results. Whereas during the first party system there was a threshold around 20 per cent of votes obtained that yielded no more than 10 per cent of seats and a cut-off point around 40 per cent of votes that brought in 55 per cent of seats or more, in the second period the relationship is more linear, with the distribution of votes and seats both more spread out and more ‘equitable’. The lesson is not lost on India’s smaller parties who have devised a way out of the one dominant party system by forming broad-based coalitions that make more effective use of their combined strength through a sophisticated system of seat-sharing prior to the election. In brief, the self-correcting mechanism of Indian democracy has found a way of delivering electoral justice without necessitating the change of electoral laws from majoritarian to proportional. The legitimacy of electoral laws is negatively affected in a context where politically conscious groups of voters feel trapped into an electoral cul-de-sac with no effective representation in the winning party or coalition. India’s multi-party system exhibits the effects of multiple cleavages which help to override some of the potential inequities of the majoritarian voting rules. The Congress Party cuts into all social cleavages. Parties of the left, such as the communists and the socialdemocrats, as well as the National Front, tend to get more support from the lower social classes, whereas parties of the right, such as the Bharatiya Janata Party, get more support from the upper social groups. However, no social segment is entirely colonized by a particular political party, nor does any particular social group have monopoly control over a special political party. In the promiscuous environment of India’s electoral culture, there are no ‘untouchables’, which gives the skilful and the entrepreneurial ample opportunities to build short-term, crosscleavage alliances, under the watchful eyes of the Election Commission which makes sure that political gamesmanship does not spill over into outright corruption or criminality. Electoral systems and parties can function effectively if voters perceive them as efficient instruments for the articulation of their interests. Survey data from 1996

494

Subrata K. Mitra

provide adequate evidence that such a sense of efficacy is indeed to be noticed in large sections of the Indian electorate. The question ‘Do you think your vote has an effect on how things are run in this country, or do you think your vote makes no difference?’ was asked to measure voters’ sense of efficacy. A majority of the electorate believed that their vote has an effect on the political state of affairs in the country. It is important to note here that even among the illiterate, the sense of efficacy is still respectable (47 per cent compared to 59 per cent for the average) while the Scheduled Castes and the Muslims as a group are actually above the national average in terms of the sense of political efficacy. If the existence of a sense of efficacy at the micro-level is a necessary condition for the effectiveness of a democratic political system then the perception of parties and elections as legitimate constitutes a sufficient condition for the functioning of an electoral system. The survey asked: ‘Suppose there were no parties or assemblies and elections were not held – do you think that the government in this country can be run better?’ The responses show how much significance people attach to the system which provides the basis for their direct participation in it. When asked to conjecture on a situation where no elections are available, asked deliberately in a manner that highlights the absence of elections rather than their presence, Indians overwhelmingly reject a future without parties and elections. An impressively large 69 per cent of the sample disagrees with the proposition that the country could be run better without elections. Significantly, this percentage has gone up from the relative low of 43 per cent in 1971 to 69 per cent in 1996. The young (71 per cent), the educated (74 per cent) and the urban people constituting the group called ‘opinion-makers’ have come out overwhelmingly in favour of sustaining the system. More importantly, they are also joined by Muslims (72.4 per cent) and Christians (72.5 per cent). The positive evaluation of the political system based on parties and elections by the better informed and the minority Muslims reinforces the picture of steady empowerment of the electorate through participation in electoral politics as structured by the first-past-the-post system. Electoral violence in India is widespread. The larger issue of the legitimacy of elections arises from the fact that Indira Gandhi, the unrepentant author of the 1975 Emergency, was actually brought back with popular acclaim in the parliamentary elections of 1980 when the Janata Party, who were the heroes of the antiEmergency movement, failed to deliver stable, efficient and honest governance. At the heart of the electoral system in India is a paradox: while people have faith in elections in general, people who run the institutions – bureaucrats, policemen and politicians – are seen as flawed characters, worthy at best of limited trust. When asked: ‘How much trust/confidence do you have in the different institutions of India?’ the results are highly positive for institutions such as the Election Commission and the Supreme Court but much less so for those who are responsible for running those institutions such as politicians, civil servants and policemen (Mitra and Singh, 1999). This finding of institutional trust but distrust of actors provides a glimpse into the institutional environment of India where abstract trust in institutions remains high but executives or political leaders who occupy them are not always thought of as worthy of the distinction. As a result, people

India

495

resort to ‘direct action’ which is often a euphemism for lawless behaviour when officials refuse to concede to their demands. The larger implications of this phenomenon are beyond this chapter but the problem for the long-term stability of multi-party democracy remains serious.

Conclusion The description of India’s transition from a colonized to a self-governing society as ‘democracy from above’ does not do full justice to the riots, uprisings and protest movements that have given it momentum from below. The fact remains, however, that the major steps in this process, such as the passage of the main Representation of the People Act, lowering the age of voting, the extension of direct democracy to the village level or legislation to ensure a minimum representation for former untouchables and subsequently women, came about not in response to popular pressure but elite initiative. The electoral process rather than popular protest acted as the engine of progress. Typically, as marginal social groups discovered the negotiable value of the vote, they became avid players in the political arena at the local and regional level. Established Jajmani systems – reciprocal social bonds based on the exchange of service and occupational specialization – broke down to create new groupings. Rather than being units in a hierarchical chain of oppression and domination, castes themselves were transformed into ‘caste associations’, serving as links between parties and the society, creating useful room to ‘manoeuvre in the middle’ for national, regional and local elites. Three specific attributes of the Indian context facilitate the full play of the electoral forces. The Gandhian legacy which combined institutional politics such as voting and lobbying with protest movements in a two-track strategy has became an integral part of the political culture of post-independence India. The ability of the national and regional elites to transfer knowledge from one arena to another, creating in the process new electoral arenas and mobilizing and subsequently institutionalizing new social forces, has reinforced the countervailing forces of the political system. Finally, the availability of the ‘Congress party system’ and the elite consensus that defined it during the first party system paved the way for the self-reinforcing character of the electoral system beyond the decline of the Congress hegemony. The Indian story will remain incomplete without a brief invocation of its path dependency, for there is an element of serendipity with regard to the successful career of the electoral system in India. The clean slate that the leaders of the post-colonial state were offered on which to try their legislative talents and political networking resulted in no small measure from the partition of India which sent most of the nay-sayers to the electoral system based on the first-past-the-post system in the Muslim League across the border to Pakistan. India still has a very large minority of Muslims spread across the nation, marxist state governments and a coalition dominated by Hindu nationalists in power at the federal level. However, in spite of their public bickering and ideological grandstanding, behind the scenes it is politics as usual which binds them in political transactions based on the fine calculation of self-interest. In the final analysis, despite the dark prophecies and

496

Subrata K. Mitra

strident recommendations for a changeover into proportional representation at the outset, India’s electoral laws have remained stable since independence, for they work in a manner that most people find satisfactory most of the time, which makes it possible for the losers to take the bad times in their stride and stay on with the game. More than the idiosyncrasy of Indian culture, effective institutional arrangement is the ultimate explanation of the successful career of the electoral system in India.

References Basu, Durga Das (1985) Introduction to the Constitution of India. New Delhi: Prentice-Hall. Enskat, Mike, Subrata K. Mitra and V. B. Singh (2001) ‘India’, in Dieter Nohlen, Florian Grotz and Christof Hartmann (eds), Elections in Asia and the Pacific: A Data Handbook. Oxford: Oxford University Press, pp. 559–84. Lijphart, Arend (1996) ‘The Puzzle of Indian Democracy: a Consociational Interpretation’, American Political Science Review, 90: 2. Malhotra, S. (1998) Manual of Election Laws: A Compilation of the Statutory Provisions Governing Elections to Parliament, State Legislatures, Panchayats and Urban Local Bodies, along with Instructions of the Election Commission of India. Delhi: Bahri Brothers. Mitra, Subrata K. (1999) Culture and Rationalily: The Politics of Social Change in Post-colonial India. Delhi: Sage. Mitra, Subrata K. and V. B. Singh (1999) Democracy and Social Change in India: A Cross-section Analysis of the Indian Electorate. Delhi: Sage. Rae, Douglas (1971) The Political Consequence of Electoral Laws. New Haven, CT and London: Yale University Press.

Further reading • Austin, Granville (1966) The Indian Constitution: Cornerstone of a Nation. Bombay: Oxford University Press.

• Basu, Durga Das (1985) Introduction to the Constitution of India. New Delhi: Prentice-Hall. • Lijphart, Arend (1996) ‘The Puzzle of Indian Democracy: a Consociational Interpretation’, American Political Science Review, 90: 2.

• Mitra, Subrata K. (1992) Power, Protest, Participation: Local Elites and the Politics of Development in India. London: Routledge.

• Mitra, Subrata K. (1994) ‘Caste, Democracy and the Politics of Community Formation in India’, in Mary Searle-Chatterjee and Ursula Sharma (eds), Contextualising Caste: PostDumontian Approaches. Oxford: Blackwell, pp. 49–71. • Mitra, Subrata K. (2000) ‘India’, in Gabriel Almond, G. Bingham Powell, Kaare Strom and Russell J. Dalton (eds), Comparative Politics Today: A World View. New York: Longman.

29 Indonesia: Transition and Change but Electoral System Continuity Andrew Ellis

Electoral politics in Indonesia has a limited, but complex, history. The independence of Indonesia from the Netherlands was proclaimed on 17 August 1945, and a constitution promulgated the following day. This constitution was based on the principle of the integralistic state, vesting overall power in a single highest state institution, the MPR (Majelis Permusyawaratan Rakyat). The legislative function was to be undertaken, however, by another chamber, the DPR (Dewan Perwakilan Rakyat), and the executive function placed in the hands of the President, who would be chosen by the MPR to serve a five-year term following each general election. While the new Republic of Indonesia was resisting efforts by the Netherlands to re-establish colonial control between 1945 and 1949, the 1945 Constitution could not be fully implemented. In practice, the republican government functioned as a parliamentary system during this period. It was then replaced firstly by the federal constitution which accompanied the independence agreement with the Netherlands and subsequently by the temporary 1950 Constitution which established institutions of a more parliamentary character. An account of the political institutions of this period is found in Budiardjo (1998).

The 1955 elections General elections, however, did not take place until 1955. The government proposal for a proportional representation system using 15 regional disticts (plus West Irian, which was still under Dutch control) was adopted: this was not publicly challenged. With the large number of small party representatives in the transitional Assembly, any other system would have been unlikely to gain agreement. Parties or organizations could submit lists of nominations, and individual candidates could submit nominations. A total of 257 elected seats were distributed between the regions in accordance with population, although a small extra allocation of seats was made to Outer Island areas of low population (East Kalimantan and Maluku). In addition, there were 15 appointed members, 3 in respect of West Irian and 12 from the Arab, Chinese and European communities. The Hare quota with largest remainders was adopted. Voters could vote either for a list or by writing 497

498

Andrew Ellis

the name of one candidate. Elections were conducted both for the DPR and for the Constituent Assembly, and were accepted as in line with democratic standards and principles (Feith, 1957, 1962). The resulting DPR contained representatives of 27 parties and lists, plus one individual member. The largest single party, President Sukarno’s Indonesian Nationalist Party (PNI), polled only 22.3 per cent. Three other parties also polled significantly: the modern Islamic Masyumi, the traditionalist syncretic Islamic Party in East and Central Java (NU) and the Indonesian Communist Party (PKI). The even split of party support meant that not only no single party, but no two parties were able to command a majority in the DPR. The results for the Constituent Assembly, which had 514 seats, were very similar. The Constituent Assembly was unable to reach agreement on the text for a permanent Constitution. Inability to reach agreement on the relationship between Islam and the state was one major contributory factor. The construction of governments was difficult and their ability to retain confidence in the legislature limited (Nasution, 1995). General loss of confidence in the political institutions and challenges to the unitary state through armed rebellion in some regions led President Sukarno, with military support, to issue a Decree on 5 July 1959 which reinstituted the 1945 Constitution. Sukarno had never been happy with the restricted role given to the President under the 1950 Constitution. He used this Constitution as the basis of his authoritarian ‘Guided Democracy’ regime, which lasted until he was gradually removed from power between 1965 and 1968 and replaced by the ‘New Order’ of President Suharto.

Elections under the ‘New Order’ The New Order retained and extolled the rather vague 1945 Constitution, using the concept of the integralistic state to create a system of virtually complete executive dominance. Some supporters of the New Order, notably the military, proposed a system of single-member districts by plurality rule as a result of the experiences of 1955, while nationalist and traditionalist Islamic politicians opposed this. Suharto resolved the dispute in favour of a proportional representation system and safeguarded his control through the addition of appointed members. In the MPR, the highest state institution, the elected and appointed legislators were joined by a further substantial number of appointed members. The number of seats for each province was based on population, but was qualified by a provision giving a minimum number of representatives for each province equal to the number of second-tier regions in the province. This was introduced specifically to balance representation between Java and the rest of Indonesia. Java received only just over half of the elected seats as Suharto sought to assuage fears of Javanese political domination. Java had over 70 per cent of the registered electorate in 1955, and even though this figure has fallen, it was still 58 per cent in 1999. Elections took place on a five-year cycle, in 1971 and from 1977 to 1997. Campaigns were heavily restricted, many candidates disqualified and rules applied disproportionately against opponents of the non-party movement originally

Indonesia

499

established under Sukarno and used as the instrument of the New Order, the Functional Groups or Golkar. Suharto forcibly reduced the number of electoral contestants to three in 1973 (although the law implementing this decision was not passed until 1975): Golkar, and two parties formed by the amalgamation of all Nationalist and Christian parties, the Indonesian Democracy Party (PDI), and of all Islamic parties, the Development Unity Party (PPP). The proportional representation system of the New Order was based on Indonesia’s then 27 provinces and used the Hare quota with largest remainders. However, complete central party control over the choice of candidates meant that the 1955 provision for individual candidates was not retained and the closed list system adopted. While there is little detailed literature about New Order elections, Gaffar (1992) produced a valuable study of these elections at the local level in villages near Yogyakarta. Suharto introduced the principle of elections to assemblies at the provincial and second-tier administrative area level in 1971, synchronous with the national-level elections. The electoral system was the same as the national system, and the pattern of executive dominance replicated at local level.

Transition to democracy The Suharto regime fell in May 1998. He was replaced by his vice-president, B. J. Habibie, and the process of developing a new democratic framework began. The lack of legitimacy of the Suharto era legislature brought agreement that new elections were urgently required, and these were fixed by the MPR to take place in June 1999. Three significant groups of actors dominated the preparation process. First, the Habibie government appointed a technical team, called ‘the Team of Seven’, responsible for drafting the package of new electoral legislation. Much of this teams’ drafting took place before international support or involvement was available: its only source of external technical advice was a resident Australian who volunteered help. Second, the parties within the existing legislature transformed quickly: the official Golkar rapidly began its transition to a party of the new era, while the Islamic PPP realized early that its only hope of survival was to present itself as the voice of reform within the old institutions. Third, new parties formed outside the legislature endeavouring to be voices of the reform movement. The most significant were: the Indonesian Democracy Party – Struggle (PDI-P), following former President Sukarno’s nationalist precepts and led by his daughter Megawati Sukarnoputri; the National Awakening Party (PKB), in which longstanding religious leader Abdurrahman Wahid was the most prominent figure, primarily representing the traditionalist Islam of East and Central Java; and the National Mandate Party (PAN), led by prominent reform campaigner Amien Rais, which attempted to appeal both to the modernist Islamic tradition and to an urban, liberal vote. The Team of Seven’s draft legislation proposed a personalized proportional representation system, under which about 80 per cent of the members would be elected from single-member districts by plurality rule. These would be joined by members elected through ‘national top-up seats’ allocated to candidates taken

500

Andrew Ellis

from lists nominated by each party at national level. The legislature was to have 462 elected seats, with about 380 single-member district seats and about 80 top-up seats proposed (the remaining 38 seats still being filled by military and police members appointed nationally by the forces command). This proposal of personalized proportional representation was designed to respond to reformist feeling that members of the legislature should not as previously be very heavily Jakarta-based and under central party direction. However, political support for this position came not from the reformists but primarily from Golkar, which rapidly moved to redefine itself as a party in the new era. Golkar believed that they could poll 30 per cent nationwide and would also benefit from their heavier support away from Java where (maintaining the principles of the New Order electoral legislation) single-member districts would have smaller electorates. Golkar also wanted to maintain the tradition of strong central party input into candidate nomination. The parties outside the legislature failed until very late in the debate to address electoral system issues at all because they had not recognized the political implications of what appeared to them to be technical matters: they finally insisted upon a proportional representation system. PPP saw the political necessity to position itself as the champion of reform within the legislature to ensure its survival. It reconciled this by espousing the use of the proportional representation system for the 1999 elections and the desirability of moving to single-member districts at a later date. The new electoral legislation was finalized in late-night negotiations against a time deadline in late January 1999. The electoral system – officially described as a ‘proportional system with district characteristics’ – was the clear product of incremental political negotiation. Each voter would cast a single ballot for a political party. The 27 provinces were retained as electoral districts, ranging in magnitude from 4 to 82. The number of seats won by each party in each province would be determined using the principles of proportional representation, and each candidate on each party list would be linked to one of the local districts within the province. The same system was established using smaller scale units at the provincial level. Given the positions and the power bases of the parties both inside and outside the negotiations, there was no way in practice that the system finally agreed could have been substantially different. Less controversial but also of great significance was the provision in the political party law which required all parties contesting the elections to be organized in at least nine provinces. It was not possible to register a specifically regional party. And while parties were required to submit nominations at provincial level, there was a single national ballot paper for the whole country. This rule encouraged parties to run candidates as widely as possible, and the stronger ones succeeded in doing so in all provinces. Equally, however, it was possible for votes to be cast for parties which had not nominated candidates in a particular province. This did not, however, happen on a wide enough scale in practice to enable a party without a candidate to qualify for a seat. There was more political scope for innovation on issues of election technicalities. However, the universal desire for a fast election left little option but to base the

Indonesia

501

election administration on the previously existing machinery. Any attempt to set up a new administration would have entered the political sphere, because of Golkar’s close connection with the civil service – indeed ending the ability of civil servants to campaign was a key issue in the negotiations. Although the legislation established a system of proportional representation in multi-member districts, it was not clear on essential detail. How votes cast would be converted into seats gained was not specified – there was no decision between highest average or largest remainder principles and no definition of the process to be followed. Nor were there any rules for identifying which particular candidates from a party’s list would take the seats gained by that party. These issues were resolved only at a very late stage. Several members of the Team of Seven, by then serving as government representatives on the General Election Commission (KPU) followed their earlier emphasis on links between members and specific geographical areas and sought local accountability of successful candidates. However, many of the 48 party representatives on the KPU wanted to ensure their central control of their parties and opposed such moves. The KPU’s final regulations for calculating seats gained used, as at previous elections, the Hare quota with largest remainders. The allocation of candidates to the seats won by each party was much more complex, and followed a concept used by the New Order for the 1971 elections only. Parties were required when submitting nominations to attach a list of candidates to each specific local district. Each party’s entitlement to seats in a province was established from the votes cast. The successful candidates of the party would then be determined according to the largest votes obtained by the party at local level, with the definition of ‘largest vote’ to be decided by each party and notified to the KPU before polling day. In the regulations published before polling day, in any local district where the party had polled one or more full quota of votes, the top candidate(s) of the party attached to that district would be declared elected for seat(s) linked to it. In addition, the plurality victor would be declared elected with only one linked seat, as long as his/her party had polled a full quota province wide. Each party leadership could then choose any candidates nominated by the party within the province to fill the remaining seats which had been won by the party in the province, with inter-party agreement on the linking of these members to local districts that did not yet have a full set of linked members. The desire of parties to guarantee the election of top national figures in Jakarta and specific local political factors in East Timor led to exceptions in these two provinces, where elected candidates were taken purely from the top of the provincial lists submitted by the parties. The formula was made even more complex by the late introduction by the KPU of ‘apparentements’, called ‘stembus accord’, by which groups of parties could reach agreements to pool their remainders with the intention of gaining more seats. This principle had been included in the 1971 regulations but had become irrelevant with the reduction of electoral participants to three. The attempt at reintroducing pre-election vote pooling was, however, unsuccessful. Some parties discovered as the results came in that given the actual votes cast, their groupings (or their parties within their grouping) would win more seats without the apparentements

502

Andrew Ellis

than with it. The result was the retrospective abolition of all apparentements – a further regulation to this effect being issued by the KPU considerably after polling day! Strictly applied, these rules could (and would in practice) have led to situations where parties had more candidates elected in some provinces than their seat entitlement in those provinces. It was thought that this problem might be resolved by restricting the total of elected candidates to the party’s total entitlement in the province and using the party’s notified ‘largest vote’ formula to define a ranking order. In practice, central party leaderships continued their moves to ensure that power did not pass to voters and local parties: few parties even complied with the requirement to notify a ‘largest vote’ formula. The regulations allocating candidates to seats won were changed twice after polling day as the leaderships sought to undermine the element of the system outside their control. The provision requiring a threshold of a ‘full quota’ disappeared and the party leaderships were given specific powers to notify the list of their elected candidates. As a result, at least eight candidates for the national legislature who should have been elected for their districts under the previous full quota provisions were replaced by others on the instruction of party leaderships. Two elected candidates for the DPR even switched province – a clear breach of the regulations – in order to get the leaders of small parties into the legislature. At the provincial and local level, where the same electoral system existed, there were wider violations of the regulations: there are examples where the political composition of the local elected body as constituted after the election was not that which would result from the election results by use of the election regulations. However, the political desire to move on, the general weakness of law enforcement in Indonesia and the specific weakness of the complaints and appeals provisions in the 1999 election legislation combined to allow such breaches of the regulations to stand. The 1999 elections were judged both domestically and internationally to have been acceptable overall and to reflect the will of the people, despite a number of specific or localised concerns (NDI, 1999). As with many transitional elections, the results inevitably contained a major element of a ‘referendum on change’. Golkar was seen as the party of the old system: others were against the old system. This cleavage will become less relevant over time – as has largely already happened, for example, across Eastern Europe. The 1999 results confirmed the identity of the five big parties, although there was the commonly found disappointment at a first election – especially within the National Mandate Party (PAN) and the National Awakening Party (PKB) – arising from unrealistic expectations. The Indonesian Democratic Party – Struggle (PDI-P) emerged as the largest single party, polling most strongly in Java, Bali and some other areas with significant Christian populations. Golkar was perhaps relieved to have held on to a position as the second strongest party, retaining its strength in many of the Outer Islands – especially Habibie’s home island of Sulawesi. They were beginning to position themselves as a party lying between the secular nationalism of PDI-P and the different Islamic positions of the PPP, the PKB and in some areas the PAN. The PPP gained many of its best results in the strong

Indonesia

503

Islamic areas in northern and western Sumatra, while the PKB totally failed to break out of its East and Central Java heartland. The relative strengths of the parties thus varied widely in the different parts of Indonesia: only the big five parties gained more than 2 per cent of the vote. Sixteen other parties gained representation. However, the 1999 electoral law includes a threshold of either 2 per cent of the national legislative seats or 3 per cent of the seats on regional authorities, which parties must exceed in order to qualify to participate in the election following. Only the big five and the conservative Islamic Crescent Moon and Star Party (PBB) had enough support to pass this threshold – and there was a rearguard action by many of the small-party representatives against the final validation of the election results. The results demonstrated a strong continuity between support for secular nationalist parties or parties representing the two major streams of Indonesian Islam in the 1955 elections and support for their identifiable successor parties in 1999 (King, 2000). The underlying electoral patterns and cleavages which appeared in 1955 re-emerged in 1999 despite the long intervening authoritarian period – with, for example, areas which had backed the PNI in 1955 found likely also to have supported the PDI-P in 1999. Gaffar’s (1992) findings showing correlation between religious stream identification and voting behaviour in his survey villages are consistent with this pattern. The passage of wide-ranging and deep devolution and regional autonomy legislation in 1999 was not reflected in the results of the elections for provincial and local authorities. In the vast majority of cases, the vote shares achieved by a party for the different level authorities in the same area vary by less than 1 per cent. Under the 1999 system, the effective number of parties in the legislature was 4.72 – comparable with its value in established multi-party proportional representation systems such as in Finland, Italy and the Netherlands. Given the diverse nature of Indonesian society, a system which produced a noticeably lower figure would appear likely to be a system which left significant groups unrepresented or badly under-represented. This is borne out by the sensitivity of the results to the electoral system chosen shown in calculations by Evans (2000), who used the 1999 votes cast to project notional seats won under a number of different systems. No account has been taken of the fact that parties are likely to run different types of campaigns under different electoral systems, which means that the corresponding estimations should be taken as indicative rather than precise. (Further useful material can be found in Evans (2003).) The actual 1999 results give a figure of 3.3 per cent for the least-squares index of disproportionality. This is comparable to average values found elsewhere in the world under proportional representation systems. In contrast, disproportionality for the projected systems shows values of 12.2 per cent for proportional representation based on local lists, of at least 19.8 per cent for plurality rule in single-member districts, and of 20.1 per cent for the alternative vote (which may be compared with an average value of 9.8 for 78 elections under plurality rule in single-member district systems in established democracies analysed by Lijphart, 1994). Indeed,

504

Andrew Ellis

estimated disproportionality for the plurality rule and alternative vote systems in Indonesia is greater than that for any of Lijphart’s cases with the single exception of the system used in India in 1952 and 1957. A system of single-member districts by plurality rule would likely lead to more disproportional results in Indonesia than almost anywhere else.

Major constitutional change The 1999 legislation envisaged a review of the conduct of the elections within three years at most, and the weaknesses perceived during the conduct of the elections created some momentum for this to start. Major dissatisfaction with the performance of the party representatives on the KPU – including substantial allegations of financial corruption – meant that its composition was addressed immediately. It was amended in 2000, the mixture of government and political party representatives being replaced by a commission with 11 independent members. However, the debate on the 1999 elections was rapidly overshadowed by the start of a full review of the 1945 Constitution, which the MPR initiated in 1999 following a key demand of the 1998 reform movement. This review was completed in 2002 after some heated debate and resulted in fundamental changes in the political institutions of Indonesia. Almost all participants in the debate were united in their support for the unitary state. Similar unity prevailed around ‘the presidential system’, although this terminology was used to describe both the unique integralistic system laid down in the original text of the 1945 Constitution which is in conventional terms neither fully presidential nor fully parliamentary, and the ‘conventional’ presidential system exemplified by the Philippines, various Latin American nations and the USA. This debate was thus confusing to outsiders. The constitutional review led to the abandonment of the integralistic state, in which the legislature, executive and judiciary were all subject to the MPR as the single highest state institution. The constitutional amendments introduced the separation of powers and the principle of checks and balances. A direct election for the presidency and vice-presidency by majority rule with a second round has been introduced. The newly established second regionally based elected chamber, the DPD (Dewan Petwakilan Daerah), will have limited powers regarding legislation and will also possess oversight powers. The participants in elections for the lower chamber DPR would be political parties; for the new region-based DPD, they would be individual candidates. The result is a set of state institutions which are recognizably within the mainstream of the family of presidential democracies. The first elections to these new institutions are to take place in 2004. The constitutional review process did not, however, entirely prevent work on the electoral legislation. In 2001, a joint team of Ministry of Home Affairs civil servants and academics produced three draft proposals for the revised election law. This contained three options for the electoral system, including both alternative vote and majority runoff. There were major technical and administrative problems with these drafts, and they were subsequently reconsidered internally by the Ministry alone.

Indonesia

505

Presidential electoral system With the constitutional review complete, consideration of the new political and electoral legislation by the legislature finally started in September 2002. There are four elements to this legislation: the election law itself, the law on political parties, the law on the structure and composition of state representative institutions and a new law on presidential elections. In addition, the vesting of the power to adjudicate disputes relating to election results in a new Constitutional Court requires the passage of implementing legislation to enable the establishment of this court before August 2003. The direct presidential election proposal had been controversial. Megawati Sukarnoputri of PDI-P, who succeeded as President in August 2001 after the removal from office of Abdurrahman Wahid, appeared initially lukewarm to the idea, but moved to support it. This perhaps followed a full appreciation of the difference in presidential tenure and accountability between the previous integralistic state system – under which the process to remove Wahid had largely been followed on policy grounds – and the new conventional presidential system (King, n.d.). As a result of the political negotiation which attended this constitutional change, a significant amount of detail on this issue was included in the amended constitution. The constitution now provides that president and vice-president will be elected on a single ticket. Tickets will be nominated by parties or coalitions participating in the general election. A ticket will be elected if it polls an absolute majority overall, and in addition polls at least 20 per cent in at least half of the provinces. This requirement is likely to be met in practice by an absolute majority winning ticket, but could conceivably force a second round where the leading ticket’s support is very heavily concentrated in Java (which contains only 6 of the current 30 provinces). Other issues, however, remained open in the constitution. For example, it is not specified whether the legislative elections and the presidential elections should be on the same day. If not, it would be possible to impose criteria for presidential nominations based on the performance of parties or coalitions in the legislative elections. This and other issues will need to be determined as implementing legislation is produced. Legislature electoral system The presidential election legislation, however, was set to follow the completion of the debate on the new legislation submitted by the government for the election of the legislative chamber (DPR), the region-based chamber (DPD) and the representative bodies of the regions. The new constitutional requirements limit the range of available options for electoral systems for the two chambers. The constitutional requirement that participants in DPR elections are political parties reflects previous practice and militates against the adoption of any kind of single-member district system. For the new DPD, election participants are to be individual candidates. Some understanding existed during the debate on the DPD that related changes

506

Andrew Ellis

might be necessary to the electoral system for the DPR – for example, the possible abolition of the bias against Java. However, in the course of subsequent debate, while the PDI-P – which has strength in Java – continued to argue the case for one person, one vote, one value (OPOVOV), Golkar – whose primary strength lies outside Java – argued for a system which retains the representational bias in favour of the other islands. The final compromise was a complex formula fixing the number of seats for each province based on a minimum of 325,000 inhabitants per seat in small provinces and a maximum of 425,000 inhabitants per seat in large ones, subject to a minimum of three seats per province (an echo of the 1995 rules). During 2000 and 2001, there was considerable advocacy, mostly driven by academic, media and NGO circles, in favour of single-member districts by plurality rule as a method of increasing member accountability – an attribute widely perceived outside the DPR to be lacking in the current legislature. Although 51 per cent of elected DPR members said in a survey (CETRO, 2000) that they regard district links as ‘very important’, and that given conflict 49 per cent would regard the district interest as decisive as against 31 per cent for personal beliefs and 19 per cent for the party group position, these figures appear to overstate the effectiveness of the 1999 district-linking system in practice. Even if the 1999 Indonesian electoral system is viewed not as a political deal but as a brave attempt to marry the principles of list proportional representation with the creation of an element of district accountability, it did not achieve this. No practical constituency/member link has really been created. In addition, steadily worsening relations between the elected legislature and many academic, media and civil society actors meant that the support for singlemember districts in the latter quarters became steadily less persuasive. Several factors have contributed to its dropping off the list of likely new electoral systems. They include the realization of the potential failure of a plurality rule system to adequately or at all reflect the diversity of Indonesia, the time and difficulty that would be involved in an acceptable districting process before 2004, and an appreciation of the negative effects of plurality systems on the election of women. In addition, viewed from outside the domestic political arena, it is perhaps wise to consider Curtice and Shively’s (forthcoming) conclusion: We are simply forced to conclude that there is nothing to choose between singleand multi-member districts. This may go against much intuition and the tenaciously held beliefs of the advocates of both forms of representation, but our evidence suggests that neither side has much of a case. This particular debate about electoral systems at least simply appears to be a blind alley and it certainly provides no grounds for preferring one electoral system over another. The lack of public confidence in the electoral machinery in Indonesia in relation to counting and tabulation restricts in practice possible electoral systems to those which can be counted in the polling station. The count at the polling station is a public event, at which not only party agents and domestic observers but many of the local community are present – and noisily involved as the vote on each

Indonesia

507

ballot paper is read out. Party agents and domestic observers are therefore able to compare lower level result protocols with higher level tabulations and ensure that no fraudulent changes have been made – maintaining confidence in the outcome of the tabulation process. The worldwide trend of increasing involvement of civil society in election processes through domestic observation and ‘quick count’ operations played an important role in Indonesia in 1999 and is likely to do so again in 2004. Single-member district systems (including by majority runoff) and list proportional representation systems are both amenable to polling station counting, but preferential systems, for example the alternative vote or the single transferable vote, are much less readily compatible. (Result protocols which record first preferences would go some way towards resolving this problem but do not do so completely.) Preferential voting as a mechanism for improving member accountability in Indonesia has thus largely been discounted. The preferential voting process is also perceived as too difficult, although this is less easy to sustain objectively. The initial draft of the election law submitted to the legislature by the government included a proportional representation system using multi-member districts. The pressure for accountability was reflected in two major proposals for change from the 1999 system: the use of open lists instead of closed lists, and the division of larger provinces so that most electoral districts would have a magnitude of between 5 and 9 members. This would result in a total of about 80 electoral districts in a proposed expanded legislature of 550 members. The difficulty of defining boundaries even on this more limited basis, coupled with the impracticality of a full open list system if provinces (some of which would qualify for 80 or so members) are used as the electoral districts, meant that this proposal was not technically viable in the form submitted. However, its basic form was finally adopted, with multi-member districts of magnitudes between 3 and 12 to be drawn by the KPU. An optional open list system was agreed, requiring voters to vote for one party and, if they wish, one candidate for that party. However, the open list votes will only result in the election of a particular candidate out of list order if the candidate gains more than a full Hare quota of individual votes. The positions taken by the parties in the legislature in the debate on these proposals reflected for the most part their perceptions of political advantage. Open or closed list, OPOVOV, member/second-tier region linkage, and Java/rest-ofIndonesia balance were all issues negotiated between the parties when the final deal was struck. Similar calculations of advantage are manifest in the debate over the electoral participation requirements for political parties. The larger parties support proposals to toughen the requirements that need to be met by a party which wishes to participate in the elections. Central party control over legislators is also an issue: proposals to restore the recall provisions over individual members as existed under Suharto remain on the table of the debate on the law on the structure and composition of representative institutions. The debate over the 2004 electoral law system was conducted in a context where the strength of party identity reflected in the 1999 results has continued. The major parties in the new Indonesian political system have remained in

508

Andrew Ellis

substantially the same form through the three years following the election. Indeed the country’s diversity and established traditions make it difficult to envisage any major reduction in the number of these stronger parties while retaining a configuration of parties coherently reflecting the cleavages of Indonesian society. Any such reduction would be likely to leave significant elements of society perceiving themselves as unrepresented or under-represented. While there have been serious internal differences within parties which have led to major tensions in Golkar and the PDI-P and have gone far enough to lead to the foundation of alternative versions of the PKB and PPP, indications are that the identity of the ‘mainstream’ versions of the existing parties will continue not to be seriously challenged in 2004. The central party leaderships show little inclination to relax their hold on their parties or to see significant change or realignment in the party system. There is no immediate likelihood of such change or realignment – even more so as there is no call to relax the effective prohibition of regional parties. The requirement that all parties are national in character remains, even in relation to provincial and local elections. The system for elections at the provincial and local level will remain the same as for the national level. Political debate is indeed conducted entirely in the context of the national level: the importance of regional authorities as decision-making bodies as the implementation of devolution and regional autonomy proceeds is not yet reflected in the perceptions of those debating electoral systems. The extent of the independence of the election administration was also controversial in the debate of the new law, with many of the political parties supporting a model where the KPU appoints its own staff while the government seeks a model in which KPU staff are civil service appointees. The final version gave the KPU power to choose its Secretary General from three civil servants nominated by the government and to define its own administrative structure within the terms of valid legislation. Most of the debate surrounding the electoral system for the legislative chamber (DPR) has thus been conducted around concepts which are already known and established in Indonesian electoral history. The region-based chamber (DPD), however, presents new territory for consideration: how to elect four members per province when the candidates are to be individuals, not parties. Candidates are required not to have been in an executive body of a political party in the four years before nomination. Government proposals requiring the signatures of 1 per cent of registered electors would have meant that only candidates with institutional support – in effect parties, religious organizations or the military – would find it practical to assemble a valid nomination. The DPR, however, reached agreement on a less onerous nomination requirement of the support of between 1,000 and 5,000 electors depending on province size. The government draft law proposed one obvious solution for the DPD electoral system: bloc vote in each province. This system appeared to benefit strong parties or organizations which promote or support slates of candidates. It might therefore have advantaged parties with support outside Java, where 24 of the 33 provinces are found. Unsurprisingly, the PDI-P put forward single non-transferable vote as an alternative and finally gained acceptance for it.

Indonesia

509

The 2004 legislation The 1999 Indonesian electoral system agreement had to be acceptable both to the parties of the New Order holding the levers of power and to the new parties outside on the street. The completion of the constitutional review in 2002 required a deal which could command support across the political spectrum. The legislative elections of 2004 are the subject of such a deal, which similarly needed the agreement of most of the political leadership. There were a limited number of solutions which could be reached in practice given the inherited traditions, the political background and positions of the actors, and the importance and sensitivity of national sovereignty and ownership of the electoral or constitutional process. The same will be true as agreement is reached on the details of the presidential election provisions. The key decisions on constitutional arrangements and on electoral legislation and systems are almost always made under time pressure in Indonesia. Even where more time appears to be available – as was at first sight the case in preparation for 2004 – the quest for political advantage drove finalization of the electoral and political arrangements, if not to the wire, certainly considerably later than election administrators would ideally have liked. At the same time, the extra time taken by the negotiating process clearly enabled legislators to identify and eliminate technical ‘rough edges’ in the draft legislation. The result was a system similar in principle to 1999, but with important differences of detail. The most important change in the Indonesian polity following 2004, however, will be that which results from the establishment of a conventional presidential system and the election of a president and vice-president on a single ticket. The second important change will be the establishment and election of the second, regional chamber. Likely changes in the electoral system for the legislature will have less impact in themselves.

Table 29.1

PNI Masyumi NU PKI PSII Parkindo PKatolik PSI Other

Indonesia: elections, 1955 Legislative DPR % v/s

Constituent Assembly % v/s

22/57 21/57 18/45 16/39 3/8 3/8 2/6 2/5 13/32

24/119 21/112 18/91 16/80 3/16 3/16 2/10 2/10 11/60

Key: PNI: Indonesian Nationalist Party; Masyumi: Modern Islamic; NU: Traditionalist Islamic; PKI: Indonesian Communist Party. Source: Komisi Pemilihan Umum (General Election Commission).

510

Andrew Ellis

Table 29.2

Indonesia: electoral results and simulations, 1999

Party: National vote % Actual system: seats List PR, provincial districts Simulations: seats – List PR, one national district – List PR, provincial districts, no bias against Java – List PR, local districts – Team of Seven’s personalized PR – Single-member, plurality rule – Single-member, alternative vote – Single member, plurality rule, no bias against Java

PDI-P Golkar PKB PPP PAN PBB 42 others Seats No. parties 34

22

13

11

7

2

11

153

120

51

58

34

13

33

462

21

156

104

58

50

33

9

52

462

35

159

112

55

58

37

12

29

462

19

198

165

44

33

21

1

0

462

6

240

134

33

28

12

3

12

462

18

270

136

30

20

5

1

0

462

6

277

56

51

48

28

1

1

462

7

291

113

34

20

3

1

0

462

6

Key: PDI-P: Indonesian Democracy Party – Struggle; Golkar: Functional Groups; PKB: National Awakening Party; PPP: Development Unity Party; PAN: National Mandate Party; PBB: Crescent Moon and Star Party. Source: Evans (2000), with minor corrections reported in Ellis (2000).

References Budiardjo, Miriam (1998) Menggapai Kedaulatan untuk Rakjat. Bandung: Mizan Press. Budiardjo, Miriam (2000) Dasar-dasar Ilmu Politik. Jakarta: PT Gramedia Pustaka Utama. Centre for Electoral Reform (2000) Survey of 252 Legislators. Jakarta: CENTRO. Curtice, John and Phil Shively (forthcoming) Who Represents Us Best? One Member or Many? Ellis, Andrew (2000) ‘The Politics of Electoral Systems in Transition: the 1999 Elections in Indonesia and Beyond’, Representation, 37, 3/4: 241–8. Evans, Kevin (2000) UNDP Jakarta, unpublished calculations. Evans, Kevin (2003) The History of Political Parties and General Elections in Indonesia. Jakarta: Arise Consultancies. Feith, Herbert (1957) Pemilihan Umum 1955 di Indonesia. Ithaca, NY: Cornell University Press. Feith, Herbert (1962) The Decline of Constitutional Democracy in Indonesia. Ithaca, NY: Cornell University Press. Gaffar, Afan (1992) Javanese Voters. Yogyakarta: Gajah Mada University Press. King, Blair (n.d.) ‘Negotiating Presidential Power: the Politics of Constitutional Reform in Indonesia, 1998–2002’. Draft doctoral dissertation, Ohio State University. King, Dwight (2000) The Elections of 1995 and 1999, Similarities and Continuities. Jakarta: Habibie Centre, May. Komisi Pemilihan Umum (General Election Commission), Jakarta. Available at <www.kpu.go.id>.

Indonesia

511

Lijphart, Arend (1994) Electoral Systems and Party Systems. New York and Oxford: Oxford University Press. Nasution, Adnan Buyung (1995) Aspirasi Pemerintahan Konstitusional di Indonesia. Jakarta: Pustaka Utama Grafiti. National Democratic Institute for International Affairs (NDI) (1999) Washington DC and Jakarta, Reports of 23 February, 28 May and 28 November 1999 and statements of the NDI/Carter Center Election Observation Mission from 5 June 1999 to 26 August 1999.

Further reading • Ellis, Andrew (2000) ‘The Politics of Electoral Systems in Transition: the 1999 Elections in Indonesia and Beyond’, Representation, 37, 3/4: 241–8.

• Ellis, Andrew (2002) ‘The Indonesian Constitutional Transition: Conservatism or Fundamental Change?’, Singapore Journal of International and Comparative Law, 6, 1.

• Feith, Herbert (1962) The Decline of Constitutional Democracy in Indonesia. Ithaca, NY: Cornell University Press.

• King, Dwight (2000) The Elections of 1955 and 1999, Similarities and Continuities. Jakarta: Habibie Centre.

30 Japan: Manipulating Multi-Member Districts – from SNTV to a Mixed System Junichiro Wada

After the Recruit scandal in 1988, there was a six-year political drama about electoral reform in Japan. In 1993, the Liberal Democratic Party (LDP) lost its power for the first time since its foundation in 1955. The new election rule was decided during that time. But the new mixed system is considered better for the LDP than previous systems with multi-member districts with single non-transferable ballot. However, although the electoral system favours the LDP, coalition governments are the norm after the reform. On 6 November 1987, when Noboru Takeshita formed his cabinet, it looked like a solid administration. The LDP had a stable majority in both the lower and upper houses and his faction was overwhelmingly strong in the party. But the Recruit scandal in 1988 muddied Japanese politics. As in many democracies, it was not unusual in Japan for politicians to receive bribes. But the Recruit scandal was bigger than usual. Recruit started as a company publishing job information magazines and was a firm dramatically growing in size. The company founder, Hiromasa Ezoe, wanted to have connections with politicians so he handed round pre-listed shares. These were bribes, but since they were not cash, many politicians carelessly accepted them. In this political scandal, politicians could not say that corruption was a personal problem. As an alternative, they found another argument: that the Japanese political (electoral) system made such scandals possible.

Single non-transferable ballot and malapportionment The two most distinguishing characteristics of the old Japanese election system were unequal apportionment and multi-member districts with single non-transferable ballot. Both of them were criticized at the time. The apportionment of representatives in Japan was unfair by any standard. Prefectures with small populations had more representatives than those with bigger populations. Many economists showed that this caused a political distortion in the Japanese budget (Gotoda, 1988; Wada, 1996). Even the Supreme Court, which had 512

Japan

513

been formed under the long LDP administrations, started to rule that the elections were unconstitutional. The thirty-sixth election of the lower house on 22 June 1980, when the ratio of the voters per representative between the over-represented and the under-represented districts was 1 : 3.95, and the thirty-seventh election on 18 December 1983, when the ratio was 1 : 4.41, were considered unconstitutional, though the Supreme Court did not call for another election. In 1986 the Diet increased the number of representatives in eight districts and reduced it in seven, bringing the ratio down from 1 : 5.12 to 1 : 2.99. This reapportionment was effected because in the judicial decision of the Supreme Court some of the judges ruled that a ratio of more than 1 : 3 was unconstitutional, regardless of any other social conditions. Everyone, including the politicians themselves, thought this reapportionment constituted a crisis because the ratio was too close to ‘the limit’ and went against tradition in making four two-member districts and a six-member district. Both the LDP and the Japan Socialist Party (JSP), which was the biggest opposition party, are widely based on rural over-represented areas. It was to be expected that the reapportionments would never be fair, but a reapportionment after the 1990 census was on the cards. The LDP politicians associated the political corruption with another characteristic: multi-member districts with single non-transferable ballot. Their logic was as follows. To get a majority in the Diet and to make a cabinet, any party must have plural candidates in many districts though voters could write only one name on the ballot. Politicians from a big party such as the LDP were forced to compete with each other. Since they would follow the same party’s policies, ideological differences became irrelevant; they sought money to compete with each other (see, for example, Ozawa, 1993). This argument was supported not only by politicians but also by the mass media and even some people in academia. It prevailed in politics.

‘Political reform’ – first stage: the Kaifu LDP cabinet The LDP was the biggest party and almost always had the majority in both houses of parliament. It is natural that their preference was for single-member districts with plurality rule. But in March 1988, LDP politician Masaharu Gotoda published a book suggesting an electoral system with both single-member districts and proportional representation with one ballot. He was considering a compromise with opposition parties for electoral reform. On 18 January 1989, he was chosen as chair of the political reform committee in the LDP and on 22 May 1989, the LDP proposed the following reforms: • investigating how to complement the single-member districts with proportional representation; • reducing the number of house members to the original 471; • improving the ratio of population per representative in the over-represented and under-represented prefectures to better than 1 : 2; • equalizing the ratio of population per representative between districts as much as possible.

514

Junichiro Wada

Though Takeshita stepped down and Sosuke Uno formed a cabinet on 3 June 1989, the LDP government kept the policy of political (electoral) reform. On 28 June, an Election System Council was set up. This council had been established by law but had not been formed. There were two members from business elites, one from the labour front, four from the academic world, six from bureaucracy, three from legal establishments and eleven from the mass media. According to the law, politicians could become special members but were not chosen. On 23 July 1989, the LDP was defeated at the 15th upper house election and lost its majority. The Recruit scandal, the adoption of a consumption tax, the liberalization of agriculture and so on may all have reduced the votes of the LDP. It was also good for electoral cooperation among opposition parties and bad for the LDP that Rengo, the Japanese Trade Union Confederation, ran in the election with its own candidates. Rengo succeeded in getting 12 seats. The personal style of Prime Minister Uno may well have appealed to the female candidates of the JSP led by Takako Doi, but a further reason for the LDP defeat was that its advocacy of keeping a liberal economic policy no longer appealed to the public. New voters, even in rural areas, did not hesitate to vote for the Socialist Party. Since the Cold War had ended and there was absolutely no possibility of socialism, they used the JSP to show their anger. The story after this election showed that their choice was not the LDP but it was not socialism either. After the election Uno resigned. The LDP selected Toshiki Kaifu as a new party leader and he formed a cabinet on 10 August 1989. Toshiki Kaifu was the deputy leader of a small faction in the LDP and so his image was cleaner than that of other LDP leaders. Although according to the Japanese Constitution it was not necessary, it was important for both politicians and the people that the upper house should also appoint Takako Doi as prime minister. The LDP had stayed in power from its establishment in 1955, but on that occasion they realized the possibility of a non-LDP government. On 26 April 1990, the Election System Council gave its conclusions. The most important proposals were as follows. • A mixed system should be established. • Voters should have two ballots: one for a single-member district and the other for party list proportional representation. • The total number of the members should be ‘about 500’. • Seats would be chosen 60 per cent from single-member districts and 40 per cent from proportional representation. • For apportioning single-member district seats in the prefectures, the method of simple quota and largest remainders would be used. • Eleven blocks would be organized as districts for the proportional representation part. • For allocating proportional representation seats to blocks, the method of simple quota and largest remainders would be used.

Japan

515

• The d’Hondt formula would be used to allocate seats to the parties. • Dual candidacy in single-member districts and block lists would be permitted. The total number of members was decided at ‘about 500’ because if they had settled on exactly 500, the total number for single-member districts would have been 300. With the census data of 1985 and the apportionment method of largest remainders, the least populous prefecture would then have received just one seat. Since this would have been a politically difficult problem, the Council slipped in a special apportionment rule: if, by giving two seats to a prefecture whose initial apportionment was one, the ratio of population per representative between prefectures became smaller, the prefecture would be given two seats. Dual candidacy for a single-member district and a proportional representation block was permitted. Candidates can be on the same rank of the proportional representation list. Candidates who are elected in the single-member districts are erased from the list and priority among the rest of candidates on the same rank is decided by the number of his or her votes divided by the votes of his or her singlemember district’s winner. This rule was in all parties’ agendas of the mixed rule until the final law, though it was diversely evaluated. The conclusions of the Council were exactly the same as those adopted as the final rule by the Diet, except for the apportionment method of single-member district seats for prefectures. The Council achieved a good compromise, but four years of political drama were needed before a new rule was adopted, and the apportionment became unfair during that drama. Opposition parties rejected the mixed system for the weight of single-member districts. It would be very natural for small parties to dislike single-member district plurality rule. They advocated that the reapportionment should be carried out first. The bases of support for the Communist JCP, the Buddhist-supported Clean Government Party (Komei) and the Democratic Socialist Party (DSP) are mainly in the cities, where the value of the votes is very low. Prime Minister Toshiki Kaifu was willing to push through the electoral reform. After the crossfire in the LDP, on 6 June 1991, the government established the principles of the reform and asked the Election System Council to make a districting plan. The principles were as follows. • There would be a mixed system. • Voters would have two ballots: one for single-member districts and the other for party list proportional representation. • The total number of members should be 471. • Three hundred seats should be chosen in single-member districts and 171 by proportional representation. • For apportioning single-member district seats in the prefectures, each prefecture should receive one seat at first and the rest of the seats would be allocated by the method of simple quota and largest remainders. • For the proportional representation part there should be one national block.

516

Junichiro Wada

• The d’Hondt formula should be used to allocate seats to the parties. • Dual candidacy in single-member districts and block lists would be permitted. Making the total number of members 471 and reducing the proportional representation part from 200 to 171 meant that the LDP intentionally introduced negotiable elements from the beginning (471 was the number written in the original electoral law; the number became bigger during reapportionments on several occasions). The apportionment rule for prefectures gave tremendous over-representation to less populous prefectures; the patent malapportionment was strongly in favour of the LDP whose main base is in rural areas. Although it was impossible to make equal sized districts with this malapportionment, the Council accepted the principle. (Actually it was impossible for them to make a districting plan with the ratio less than 1 : 2.) After receiving the districting plan of the Electoral System Council on 25 June 1991, the Kaifu cabinet submitted the bill on 5 August 1991. Prime Minister Toshiki Kaifu was willing to push through the electoral reform. Supporters included Masaharu Ito, Masaharu Gottoda and Masayoshi Takemura, but the opposition group was strong even in the LDP. On 30 September 1991, the chairman of the lower house Special Committee for Political Reform, Hikosaburo Okonogi, declared the deliberation incomplete and abandoned the bill. This action was considered to have been planned by former Prime Minister Noboru Takeshita, LDP Party Secretary Keizo Obuchi and the chairman of the Diet Affairs Committee Seiroku Kajiyama, who all belonged to the Takeshita faction. Prime Minister Toshiki Kaifu was caught by surprise and threatened to dissolve the lower house. The prime minister has a right to do this but without a strong base in the LDP he could not and had no other choice but to resign. Why did the LDP step back from electoral reform? The reason would be clear at the results of the thirty-ninth lower house election on 18 February 1990. The LDP succeeded in getting a stable majority. Many LDP politicians thought that the anger of the voters had blown over and did not want to have a tough time with electoral reform. Electoral reform was not accomplished in this period, but its key players reappeared. A Utopia study group, which was the matrix of the new Harbinger Party (Sakigake) led by Masayoshi Takemura, had been organized within the LDP on 2 September 1988. They had publicly revealed their receipts and payments for political activities on 3 March 1989, showing that Japanese political activities needed too much money. This brave action definitely advanced the political reform movement. On the opposition parties’ side, the emergence of Rengo, as mentioned, also helped to gather support.

‘Political reform’ – second stage: the Miyazawa LDP cabinet On 5 November 1991, the Kiichi Miyazawa cabinet was formed. It was to be expected that the LDP leaders should turn away from electoral reform. On 26 July 1992, the LDP

Japan

517

recovered its strength in the upper house election, though it could not get a majority because just half of the members were elected each time. At that time, many politicians would have put political reform behind them, but a new scandal involving the carrying company Sagawa Kyubin was emerging, a scandal made worse because it included crime syndicates. On 27 August, the deputy leader of the LDP, vice governor Shin Kanemaru, confessed he had received five billion yen (about four million dollars at that time) from Sagawa Kyubin and resigned. The LDP had no choice but to commit itself again to political reform and to reopen discussions for revising the existing electoral system. As a result of Kaifu’s failure to introduce a mixed system, a simple singlemember district system became popular in the LDP. On 22 September, Masaharu Gotoda mentioned that he was an advocate of simple single-member districts and had chosen the mixed system for negotiation with the opposition. On 17 October, Keizo Obuchi recommended simple single-member districts, reasoning that, with this system, no representative needed to change his or her power base. On 22 December, the LDP decided to propose single-member districts for political reform. While the LDP was preparing the agenda according to its own partisan interests, other actors began to appear on the stage. On 22 May 1992, Morihiro Hosokawa, former LDP member of the upper house and the governor of Kumamoto prefecture, suddenly announced the formation of a new party. Surprisingly, the Japan New Party (JNP) succeeded in getting four seats in the upper house election on 26 July, just two months after the announcement. This was understood as an omen for new party movements in the following year. A distrust of politics definitely existed, and people were looking to a non-established party. Another important actor was Minkan Seiji Rincho. This was a private conference for promoting political reform, its members including top business elites, leaders of the labour front, professors of prestigious universities and opinion leaders in the mass media. The preparatory assembly was held on 20 December 1991, and the founding assembly was held on 26 April 1992. A more peculiar but powerful actor was born from the Takeshita faction in the LDP. Shin Kanemaru was the president of the faction. After he resigned from the Diet on 14 October 1992, Ichiro Ozawa intended to make Tsutomu Hata the president of the faction. But on 28 October 1992, Keizo Obuchi was chosen president instead. Ichiro Ozawa and Tsutomu Hata left the faction and formed a new group on 18 December 1992. Thirty-five lower house members and nine upper house members joined the new Hata faction. They claimed to be reformers, although, given the origin of the group, this sounded unlikely. On 8 April 1993, in response to the LDP’s bill, the JSP and Komei presented their bill. The important points were as follows. • A German-style personalized proportional representation system should be set up. • Voters would have two ballots: one for single-member districts and the other for party list proportional representation. • Total number of members should be 500 (there may be extra seats).

518

Junichiro Wada

• There would be 200 single-member districts and 300 seats by proportional representation. • For apportioning single-member district seats in the prefectures, the method of simple quota and largest remainders should be used. • Twelve blocks would be organized as districts for the proportional representation part. • For allocating proportional representation seats to blocks, the method of simple quota and largest remainders should be used. • The d’Hondt formula should be used to allocate seats to the parties. The opposition parties were much smaller than the LDP. It was to be expected that they would prefer a proportional representation system and a proportional apportionment rule for the single-member districts, too. This was particularly favourable for Komei whose basis was in urban areas. The JCP, whose basis was also urban, insisted reapportionment came first. The LDP carried out a reapportionment with the support of Komei and DSP in December 1992. But the ratio of population per representative between over-represented and under-represented prefectures was still 1 : 2.77. The DSP independently insisted on a personalized proportional representation system using prefectures as blocks. It looked like stalemate. On 17 April 1993, Minkan Seiji Rincho, the private conference for promoting political reform, disclosed a further agenda. The important points were as follows. • An additional-member system would be set up similar to the German personalized proportional representation system but with an uneven distribution of seats by plurality and by proportional representation. • Voters would have two ballots: one for single-member districts and the other for party list proportional representation. In single-member districts seats would be allocated by plurality rule. In proportional representation blocks, the d’Hondt formula would be used, but the divisors series would start from N + 1, where N is the number of elected persons in the single-member district. • Total number of members should be 500. • There should be 300 single-member districts and 200 seats by proportional representation. • Apportionment and districting should be done by a third party. • Proportional representation would be based on prefectures as blocks or districts. • The ratio between districts in terms of population per representative should be less than 1 : 2 in principle. This agenda included 300 single-member districts, but the total proportionality should be much higher than with a mixed system, though it could not achieve as good proportionality as in Germany. It should have been an appropriate compromise agenda. On 2 July 1993, the JSP, Komei and DSP chose this agenda after establishing the number of single-member districts at 275 and the number of seats

Japan

519

by proportional representation at 225, though a compromise might have been possible. On 31 May, Prime Minister Kiichi Miyazawa committed to introduce political reform during that Diet session, promising he would never tell a lie. But on 14 June the LDP deputy leader, Secretary General Seiroku Kajiyama, told reporters that the LDP would not compromise and that the reform would be carried out not in the current session but after the upper house election scheduled two years hence. On 18 June, opposition parties presented a no-confidence motion against the Miyazawa cabinet which was passed by a vote of 255 to 220. Many members of the Hata/Ozawa faction voted yes or stayed away. On 18 June, the Takemura group left the LDP, though they had voted against the motion. The ten members founded a new party, Sakigake, on 21 June. The Hata/ Ozawa group had no other option than also to leave the LDP, so 36 lower house members and eight upper house members left the LDP on 22 June and founded the Japan Renewal Party (JRP) on 23 June. On 24 June the JSP, DSP, Social-Democratic Federation (SDF), Komei and JRP arrived at an agreement to aim for a non-LDP (and non-JCP) coalition government. Sakigake and JNP had a commitment to form a parliamentary group by themselves and did not attend the meeting.

‘Political reform’ – final stage: the Hosokawa coalition cabinet In the fortieth general election on 18 July 1993, the LDP retained its pre-election number of seats, though it was not a majority of the lower house. The JSP lost half of its seats and the five-party group could not acquire a majority either. Sakigake-JNP got the pivotal votes. On 20 July, in a TV show, Masayoshi Takemura, the leader of Sakigake, presented a request to other parties to come to a decision on a mixed electoral system with 250 single-member districts and 250 seats by proportional representation. This call was the idea of Takemura himself as Morihiro Hosokawa, the leader of the JNP, had no set policy with regard to the electoral system. Takemura had been the managing executive of the LDP committee for electoral reform in Kaifu’s time in office and conjectured that a mixed system would achieve a consensus. On 26 July, the DSP, SDF, Komei and JRP formed a pre-coalition, and on 27 July the JSP and LDP did so too. Sakigake-JNP joined the five-party group and thus, on 29 July, seven parties (JSP, DSP, SDF, Komei, Sakigake, JNP, JRP) and an upper house parliamentary group (Rengo) produced an accord for a coalition government. On 30 July the LDP chose Yohei Kohno as party leader, but he was destined not to become prime minister, only a governor. On 9 August, the Hosokawa coalition cabinet was formed and on 17 September the final debate started. There were two bills for electoral reform: one from the LDP and the other from the Hosokawa cabinet. The LDP’s bill was as follows: • There should be a mixed system. • Voters should have one ballot.

520

Junichiro Wada

• The total number of members should be 471. • There should be 300 single-member districts and 171 seats by proportional representation. • For apportioning single-member district seats in the prefectures, each prefecture should receive one seat at first and the rest of the seats should be allocated by the method of simple quota and largest remainders. • For the proportional representation part the prefectures should be used as blocks or districts. • For apportioning proportional representation seats, each block (prefecture) should receive one seat at first and the rest of seats would be allocated by the method of simple quota and largest remainders. • The d’Hondt formula should be used to allocate seats to the parties. • Dual candidacy in single-member districts and block lists would be permitted. The Hosokawa cabinet’s bill was as follows: • There should be a mixed system. • Voters should have two ballots: one for single-member districts and the other for party list proportional representation. • The total number of the members should be 500. • There should be 250 single-member districts and 250 seats by proportional representation. • For apportioning single-member district seats in the prefectures, each prefecture should receive one seat at first and the rest of the seats should be allocated by the method of simple quota and largest remainders. • The proportional representation part should use one national block. • For apportioning proportional representation seats, the method of simple quota and largest remainders should be used. • The d’Hondt formula should be used to allocate seats to the parties. • Dual candidacy in single-member districts and block lists would be permitted. The two bills proposing a mixed system with dual candidacy permitted looked very similar, but the LDP’s implied a large amount of malapportionment and an uneven, very high proportion of single-member districts. Since its proportional part was for just 171 seats and it apportioned one seat to each prefecture (total 47) at first, the seats given to prefectures proportionally were very few. The ratios of population per representative in the under-represented proportional representation blocks (prefectures) and in the over-represented blocks was 1 : 2.97. It was almost at ‘the limit’ as the Supreme Court had said that a ratio greater than 1 : 3 was unconstitutional regardless of any other social conditions. Since this bill proposed to adopt 47 prefectures as proportional representation blocks or districts, there would be 21 two-member and 13 three-member districts. They would not be able to give fair proportionality. A single ballot system would take away the freedom to choose and reinforce the importance of the single-member districts.

Japan

521

For apportioning single-member district seats in prefectures, the Hosokawa coalition cabinet adopted the apportionment rule giving one seat to each prefecture at first. If they had used the simple quota and largest remainders rule for 250 seats, the Tottori prefecture would have received a single seat and have become a political difficulty. That may be an irony of history since, if there had been 300 seats for single-member districts, using 1990 census data the simple quota with largest remainders rule could have given two seats to the Tottori prefecture. A long debate in the committee of the lower house was deadlocked. On 15 November Prime Minister Hosokawa and the LDP Governor Kohno held a summit meeting. Hosokawa proposed a compromise to increase the number of singlemember districts to 274, but Kohno did not accept. However, the coalition government adopted Hosokawa’s compromise proposal and it passed the committee on 16 November and the plenary assembly on 18 November. In the plenary assembly, 274 votes were in favour and 226 were against. According to the Asahi Shinbun newspaper, five JSP members voted against, 13 LDP members voted for while one JSP and seven LDP members were absent. Such a rebellion was rare in Japan where the parliamentary group system was strong and traditionally every member upheld the party decision. Another drama was waiting at the upper house, where the new parties, the JRP, Sakigake and JNP, were weak and the JSP was strong following the 1989 election. On 26 November the bill was sent to the upper house. On 21 January 1994, the upper house threw out the bill with 118 votes in favour to 130 against. Five LDP members voted in favour, but 17 JSP members voted against and three JSP members were absent. The JSP rejected the high proportion of single-member districts in the bill, but, by voting mostly against, they probably made a grave miscalculation (if anything had been calculated at all). In accordance with the Constitution, the Conference Committee of both Houses was held on 26 January 1994. The upper house committee was formed by opposition members, the LDP, the JCP and a minor party. The lower house committee was formed by coalition government parties. Since the overwhelming majority in the upper house committee came from the LDP, it proposed the LDP bill – which, as mentioned, consisted of 300 single-member districts and 171 seats by proportional representation in prefecture blocks. The lower house proposed a compromise with 280 single-member districts and 220 seats by proportional representation in seven blocks. The prime minister of the coalition government, Morihiro Hosokawa, had committed himself to accomplish political reform. The LDP governor Yohei Kohno was afraid that many LDP members would leave the party in the next vote established by the Constitution. They held a second meeting on 28 January and achieved the compromise plan of 300 single-member districts and 200 seats by proportional representation in 11 blocks. The final date of the Diet session was 29 January and there was no time to remake the law. The lower house bill was set up with no date to come into force. The next session started on 31 January 1994. The bill was passed on 4 March and published on 11 March.

522

Junichiro Wada

On 11 April 1994, the districting committee started discussions and submitted its report on 11 August. They could not make districts keeping the ratio under 1 : 2, but the proposal was adopted on 21 November. The law was enforced from 25 December 1994 and from that date the lower house electoral bill became the new electoral system.

The new rules The main points of the new electoral system were as follows: • There was to be a mixed system. • Voters would have two ballots: one for single-member districts and the other for party list proportional representation. • The total number of members would be 500. • There would be 300 single-member districts and 200 seats by proportional representation. • For apportioning single-member district seats, each prefecture was given one seat at first and the rest of the seats were distributed by the method of simple quota and largest remainders. • The proportional representation seats would be elected in eleven blocks or districts. • For apportioning proportional representation seats, the method of simple quota and largest remainders would be used. • Dual candidacy in single-member districts and party lists was allowed.

Political consequences – or are they still on the way? What did the reformers of the electoral system expect? At least, the leaders of the coalitional government, Ozawa of the JRP and Takemura of Sakigake, had different intentions. Ozawa wanted a two-party system and Takemura wanted a moderate multi-party system. Both had left the LDP at the same time, but their ideologies were different. Ozawa represented the right-wing of the LDP while Takemura represented the left. From the beginning of the coalition government, Ozawa and the JSP could not act in harmony. When Prime Minister Hosokawa resigned on 25 April 1994, they broke up absolutely. Sakigake and the JSP voted for Prime Minister Hata of the JRP, but they did not send members to his cabinet. Ozawa attempted to build a two-party system from the cabinet and formed a new parliamentary group, Kaishin, with his own party, the JRP, and the DSP, the JNP, etc. just after the resignation of the prime minister. The JSP was upset because Ozawa had not advised them in advance. On 29 June 1994 there was a surprising development. Prime Minister Hata resigned, and Tomiichi Murayama of the JSP was apppointed prime minister with the support of the LDP and Sakigake. Kaishin and Komei voted for former Prime Minister Toshiki Kaifu of the LDP. Everything descended into chaos.

Japan

523

Ozawa was still rooting for a two-party system. He launched a new party, the New Frontier Party, with Komei on 10 December 1994, but lost the forty-first lower House election on 20 October 1996 under the new electoral rule. Nevertheless, Ozawa created a new Liberal Party (LP) with his faithful followers. On 12 July 1998, the LDP lost the eighteenth upper house election making it more difficult to run the Diet. On 14 January 1999, the LDP and the LP formed a coalition government with an agreement to reduce the number of seats to be elected by proportional representation by 50. This would have been of benefit for both the biggest party and the party led by the leader who wanted a two-party system. But for the majority of the upper house, two parties were not enough. The LDP needed Komei, which has a strong organization supported by a Buddhist group. Komei’s support base was strong but not wide enough to win in single-member districts, and the party did not want a high proportion of single-member districts. For getting Komei to the cabinet, their agreement was to reduce the proportional representation part by 20 and the single-member district part by 30. In February 2000, the bill reducing the proportional representation part by 20 was pushed through without other parties attending the Diet. Ichiro Ozawa was not satisfied with the final version of the electoral law and decided to dissolve the governmental coalition on 1 April 2000, a decision which caused the breakup of his party. Some of its members formed the New Conservative Party (NCP) and stayed in the coalition while the LP lost power. Despite his efforts, why did Ozawa fail to achieve a two-party system? We can present at least three factors. The first is his ideological position on the left–right spectrum. Ozawa and his followers were on the right of the LDP, while the other non-LDP parties were on the opposite side. From this position it was very difficult to form an alternative party to the LDP. The second factor is the existence of the proportional representation part and the dual candidacy system, which makes coordination of candidates in the two parts quite difficult (Wada, 1996). The third factor, which is not as often taken into consideration, is the existence of upper house and local elections. Until 1998, 49 of the 126 members of the upper house were elected in multi-member districts with single non-transferable ballot and 50 members were elected by party list proportional representation. The proportions changed slightly in 2001, with 46 of the 121 members in multi-member districts elected by single non-transferable ballot and 47 by party list personalized proportional representation. All local assembly elections are held in multi-member districts elected by single non-transferable ballot and often the district is at-large. District magnitudes are huge, however, and by using the single non-transferable ballot, each voter can vote for only a single candidate. These elections have a negative impact on the prospects for a two-party system. How has the new electoral system been evaluated by the people? The turnout ratios may tell us: 59.65 per cent for 1996 and 62.49 per cent for 2000 were the worst and second worst on record in the lower house elections. Some politicians said people would soon adjust to the system. But during this period the turnout ratios for the upper house were disastrous, too. As can be seen in Tables 30.1 and 30.2, many politicians moved between parties and many parties appeared and disappeared

524

Table 30.1

Japan: lower house and cabinets, 1987–2001

Parties:

Communist Socialist

Harbinger

SDF

DSP JRP Clean-Gov. (Komei)

Others Vac. Total

LDP NCP LP

Prime minister

JCP

1987 Takeshita 1989 Uno

27

87

29

57

302

5

5

512

Election Upper 1989 1989 Kaifu

27

85

28

55

295

7

15

512

Election Lower 1990 1990 Kaifu (2)

16

139

5

14

46

285

7

0

512

16 16

141 134

10

4

13 13

36

46 45

274 222

7 17

15 15

512 512

15

77

52

19 Regeneration

60

52

228

8

0

511

15 15 15

74 74 74

15 22 21

130 126 126

52 52 52

206 206 200

17 14 21

2 2 2

511 511 511

15 15 15

72 72 70

21 21 21

187 178 176

201 200 208

13 22 17

2 3 4

511 511 511

15 15 15

63 63 37

22 23 14

170 170 163

209 207 206

18 19 18

14 14 18

511 511 511

Election Upper 1992 1992 Miyazawa Election Lower 1993 1993 Hosokawa 1994 Hata 1994 Murayama

JSP/SDP DPJ (Sakigake)

Conservatives

Reform

Election Upper 1995 1996 Hashimoto

40

Election Lower 1996 1996 Hashimoto (2)

Election Upper 1998 1998 Obuchi

Election Lower 2000 2000 Mori (2) 2001 Koizumi

26 26 26

15 15 15

52 52 98

2 2 2

26 26 26 26

14 13 14 14

92 94 94 96

2 2

20

19

127

152 129 46

238 250 259

47 52 48 48

263 265 269 267

31

233

42

15 26 11

0 0 1

500 500 500

20

40 39 39 19

14 7 10 9

2 2 0 1

500 500 500 500

7

22

21

0

480

Key: Parties: Communist: JCP – Japan Communist Party. Socialist: JSP – Japan Socialist Party; since 1996 SDP – Social Democratic Party. DPJ – Democratic Party of Japan. Harbinger (Sakigake) including JNP – Japan New Party; SDF – Social Democratic Federation; DSP – Democratic Socialist Party; JRP – Japan Renewal Party. Clean Government Party (Komei). Conservatives: LDP – Liberal Democratic Party; NCP – New Conservative Party; LP – Liberal Party. Coalitions: Regeneration (Kaishin), including DSP, SDF, JNP, enlarged to Reform (Kaikaku) or NFP, including DSP, SDF, JNP, Komei. Seat number key: prime minister’s party; cabinet’s party; supporting party. Source: Author’s elaboration of data from Asahi Shinbun (newspaper).

525

Japan: upper house and cabinets, 1987–2001

Parties:

Comm. Socialist DPJ

526

Table 30.2

Harbinger Rengo DSP JRP

Prime minister

JCP

JSP/SDP

1987 Takeshita 1989 Uno

16

42

Election Upper 1989 1989 Kaifu

14

72

Election Lower 1990 1990 Kaifu (2)

14

73

Election Upper 1992 1992 Miyazawa

11

73

Election Lower 1993 1993 Hosokawa

11

73

11 11 11 11

68 68 68 68

38 38 39 16

11

66

15

14 14

39 36

14 14 14

33 21 21

1994 Hata 1994 Murayama

Election Upper 1995 1996 Hashimoto Election Lower 1996 1996 Hashimoto (2)

(Sakigake)

5 23 41

Clean-Gov. NFP

Komei

Conservatives LDP NCP

Others Vac. Total

LP

12

24

144

14

0

252

12

10

21

110

13

0

252

12

10

21

109

12

1

252

4

11

11

24

106

12

0

252

4

11 Breeze

11

24

99

11

0

252

24 24 24 12

94 95 95 95

17 16 15 15

0 0 0 0

252 252 252 252

95

18

0

252

8

35 Heisei 47

3 3

68 68

111 111

17 20

0 0

252 252

3 3 3

68 59

110 112 119

18 20 17

1 0 0

252 252 252

25

–

12

Election Upper 1998 1998 Obuchi

2000 Mori Election Lower 2000 2000 Mori (2) 2001 Koizumi

23 23 23 23

14 14 13 13

55 56 57 57

23

12

55

3

24 24 24 24

106 104 107 113

– – –

12 12 12 5

15 19 16 16

0 0 0 1

252 252 252 252

23

106

6

5

21

1

252

Key: Parties: Communist: JCP – Japan Communist Party. Socialist: JSP – Japan Socialist Party; since 1996 SDP – Social Democratic Party. DPJ – Democratic Party of Japan. Harbinger (Sakigake) including JNP – Japan New Party; Rengo: Japanese Trade Union Confederation; DSP – Democratic Socialist Party; JRP – Japan Renewal Party. Clean Government Party (Komei). Conservatives: LDP – Liberal Democratic Party; NCP – New Conservative Party; LP – Liberal Party. Coalitions: Breeze (Shinryokufukai) including JNP, Sakigake, Rengo, DSP, JRP. Heisei, including NFP and Komei. Seat number key: prime minister’s party; cabinet’s party; supporting party. Source: Author’s elaboration of data from Asahi Shinbun (newspaper).

527

528

Junichiro Wada

Table 30.3

Japan: winners in single-member districts of mixed system

1996 election: Districts

Population

50 over-repres. 51–100 101–150 151–200 201–250 50 under-repres.

255,273–333,120 334,644–379,045 379,711–412,987 414,434–451,051 453,894–488,355 488,677–545,542

Total 2000 election: Districts 50 over-repres. 51–100 101–150 151–200 201–250 50 under-repres.

Population

255,273–333,120 334,644–379,045 379,711–412,987 414,434–451,051 453,894–488,355 488,677–545,542

Total

JCP SDP DPJ Sakigake NFP LDP Other

Total

1 – – 1 – –

1 1 – 1 – 1

3 1 4 4 – 5

– – 1 – – 1

7 13 14 22 19 1

35 34 30 21 28 21

3 1 1 1 3 1

50 50 50 50 50 50

2

4

17

2

96

169

10

300

SDP DPJ Komei LDP NCP LP Other

Total

1 – 1 1 – 1

8 7 11 17 18 19

– – 2 – 1 4

39 33 28 28 26 23

– 2 2 1 1 1

1 2 – – – 1

1 6 6 3 4 1

50 50 50 50 50 50

4

80

7

177

7

4

21

300

Key: JCP – Japan Communist Party. SDP – Social Democratic Party. DPJ – Democratic Party of Japan. Sakigake – Harbinger. NFP – New Frontier Party. Komei – Clean Government Party. LDP – Liberal Democratic Party. NCP – New Conservative Party. LP – Liberal Party.

between elections. After the electoral reform, people’s distrust of politicians may have increased. In the most recent period, in elections for prefecture governors or city mayors, candidates supported by almost all parties (‘ainori’ candidates) often lost against independents. One more factor to be considered is malapportionment and gerrymandering. Malapportionment and gerrymandering strongly favour the LDP (see Table 30.3). The rule of giving one seat to each prefecture, which was the origin of the problem, was reused for the reapportionment in 2002. This malapportionment gives strong political power to rural areas and, among other consequences, makes it difficult to introduce the economic reforms Japan needs to climb out of its ten-year depression.

References Ozawa, Ichiro (1993) Nihon Kaizo Keikaku. Tokyo: Kodansha. Gotoda, Masaharu (1988) Seiji toha Nanika. Tokyo: Kodansha. Wada, Junichiro (1996) Japanese Election System. London and New York: Routledge.

Newspapers and databases: Asahi Shinbun Yomiuri Shinbun

Japan

529

Mainichi Shinbun

Further reading • Curtis, Gerald (1999) The Logic of Japanese Politics. New York: Columbia University Press. • Gotoda, Masaharu (1988) Seiji toha Nanika. Tokyo: Kodansha. • Grofman, Bernard, Sung-Chull Lee, Edwin A. Winckler and Brian Woodall (eds) (1999) • •

• • • • • • • • • •

Elections in Japan, Korea, and Taiwan under the Single Non-transferable Vote. Ann Arbor: University of Michigan Press. Kohno, Masaru (1997) Japan’s Postwar Party Politics. Princeton, NJ: Princeton University Press. Lijphart, Arend, Rafael Lopez Pintor and Yasunori Sone (1986) ‘The Limited Vote and the Single Nontransferable Vote: Lessons from the Japanese and Spanish Examples’, in Bernard Grofman and Arend Lijphart (eds), Electoral Laws and Their Political Consequences. New York: Agathon, pp. 154–69. Reed, Steven (1997) ‘The 1996 Japanese General Election’, Electoral Studies, 16: 121–5. Reed, Steven (2002) ‘Evaluating Political Reform in Japan: a Midterm Report’, Japanese Journal of Political Science, 3: 243–63. Reed, Steven (2002) ‘Elections: Still Demanding a Change: Elections in Japan in 2002’, Japanese Journal of Political Science, 3: 281–3. Reed, Steven and Michael Thies (2001a) ‘The Causes of Electoral Reform in Japan’, in Matthew S. Shugart and Martin Wattenberg (eds), Mixed-Member Electoral Systems: The Best of Both Worlds? New York and Oxford: Oxford University Press. Reed, Steven and Michael Thies (2001b) ‘The Consequences of Electoral Reform in Japan’, in Matthew Shugart and Martin Wattenberg (eds), Mixed-Member Electoral Systems: The Best of Both Worlds? New York and Oxford: Oxford University Press. Sasaki, Takeshi (1999) Seijikaikaku 1800 Nichino Shinjitsu. Tokyo: Kodansha. Tanaka, Munetaka (1997) Seijikaikaku Rokunen no Dotei. Tokyo: Gyousei. Thies, Michael (2002) ‘The General Election in Japan, June 2000’, Electoral Studies, 22: 147–54. Wada, Junichiro (1996) Japanese Election System. London and New York: Routledge. Yutopia Seiji Kenkyukai (1989) Nagatacho Kakyubushitachi no Kekki. Tokyo: Kodansha.

31 New Zealand: Reform by (Nearly) Immaculate Design Jack H. Nagel 1

In 1993, a nationwide referendum replaced New Zealand’s long-established singlemember district plurality (first-past-the-post) elections with German-style personalized proportional representation, there innovatively called the mixed-member proportional system (MMP from here on). The history of New Zealand’s switch to proportional representation is remarkable for the absence of political influence or bargaining in the design and adoption of MMP. The non-political Royal Commission on the Electoral System that proposed MMP sought to devise and recommend what its members considered the best possible system for New Zealand with little regard for political feasibility. Through an unexpected series of events, their proposal was adopted almost unaltered despite opposition from most leaders in both major parties. New Zealand’s experience suggests that models of electoral choice based on political bargaining are not always applicable, and that reformers who shape their proposals to meet the interests of dominant political actors may ultimately have less influence than those who appear more quixotic. The system adopted in New Zealand did, however, depart from the Royal Commission’s recommendation in one major respect, by retaining a distinctive dual-constituency system guaranteeing representation to the Maori minority. Attractive characteristics of the combination suggest that serendipity as well as conscious design can play a valuable role in the development of electoral institutions.

Seeds of change in a textbook plurality system Continuously from 1914 to 1993 (and predominantly for half a century before 1914), New Zealand elected members of parliament using a single-member district, single vote plurality rule system. After two decades of Liberal Party hegemony (1891–1912), a period of three-party competition ensued, in which the Liberals (later renamed United) lost dominance to the mainly agrarian Reform Party, while Labour developed a mostly urban base. When the Depression swept Labour to power in 1935, Reform and United dissolved to form the National Party. For the next sixty years, the Duvergerian workings of single-member district plurality elections produced a two-party duopoly, with continuous one-party majority governments controlled by either Labour or National. The workings of the system 530

New Zealand

531

were so clear-cut that leading scholars of comparative politics used New Zealand as the textbook example for plurality elections and the purest case of the ‘majoritarian’ or Westminster type of democracy, in which single-member district plurality elections, two-partyism and single-party governments are among the defining elements. New Zealand’s plurality electoral system had one atypical feature. From 1867 onwards, members of parliament were elected from a dual-constituency system – two sets of racially separate districts – one for Maori (New Zealand’s indigenous people) and the other for Europeans. Just four seats were reserved for Maori representatives, whereas the number of European (later called ‘general’) seats varied over time from 70 to 95. The dual-constituency system can be visualized as a map with two overlays, one dividing New Zealand into numerous European districts and the other apportioning the same territory into four, geographically much larger Maori electorates. From the beginning, members representing both types of constituencies served together in Parliament with equal rights and privileges. Despite a shared interest in maintaining their duopoly, Labour and National frequently clashed over secondary aspects of the electoral system, e.g. the ‘country quota’, which until 1945 inflated the representation of rural areas, thus aiding National; the number of Maori electorates, which Labour, winner of all Maori seats from 1943 to 1990, desired to increase; and aspects of the quinquennial redistricting process. To prevent partisan ‘tinkering’ with the system, the Electoral Act 1956 entrenched six key provisions of the electoral law by stipulating that they could be repealed or amended only by a 75 per cent vote of all members of parliament, or by a majority vote in a nationwide referendum. The supermajority provision effectively meant that parliament itself could not enact any major electoral reform unless both Labour and National agreed to it. The referendum option, however, established an alternative route to reform that subsequently made possible the adoption of MMP. Beneath the surface of duopolistic stability, political developments created awareness of deficiencies in the electoral system and fostered constituencies for reform. Chief among these were: (1) the rise of minor parties; and (2) increased unease about representation of the Maori. 1. From 1954 to 1990, minor parties won at least 2 per cent of the popular vote a total of 18 times, resulting in an average vote for minor parties of 13.5 per cent. They were rewarded with a grand total of seven parliamentary seats, only 0.6 per cent of the seats available. The glaring discrepancy convinced many minor-party adherents of the need for proportional representation. As a corollary of minor-party votes, governing parties won smaller vote shares, never attaining an absolute majority during this period and twice falling below 40 per cent. Contrary to New Zealanders’ majoritarian ideology, it thus became apparent that governments’ electoral bases were, at best, pluralitarian, and that the electoral system manufactured parliamentary majorities. 2. By the 1970s, in the context of growing worldwide sensitivity to racial discrimination, New Zealand’s provisions for separate Maori representation became

532

Jack H. Nagel

a source of embarrassment to liberal whites and a focus of discontent among increasingly militant Maori. The fixed allotment of four seats became inadequate, as both the relative size of the Maori population and the number of non-Maori seats in parliament grew. Reforms in 1967 and 1975 gave Maori candidates the right to contest general constituencies and Maori voters the option of choosing to vote on either Maori or general rolls, but these well-intentioned measures did little to solve the problem of Maori under-representation. Only a few Maori candidates won general seats, and many Maori electors opted to remain on the separate roll. Meanwhile, a partisan impasse continued, as the National party, in power from 1975 to 1984, believed that all new Maori constituencies, like the old, would be safe for Labour, and so prevented any increase in their number. Although Maori and minor-party supporters were thus potentially receptive to proportional representation, both groups were minorities with no power to enact a new electoral system, and qualms of conscience over unfair representation could not outweigh partisan interests in the calculations of the major party leaders who controlled legislation. Instead, the impetus that started the process of electoral reform came from a third defect of the old system, one that adversely affected a major party. In both the 1978 and 1981 elections, Labour received a plurality of the popular vote, but National won a majority of seats in parliament and thus retained control of the government. The possibility of such an anomaly is inherent in any electoral system that aggregates votes by districts in order to choose a government, but its occurrence in consecutive elections suggested to Labour that it faced a systemic disadvantage due to the excessive concentration of its vote in urban and Maori constituencies. Consequently, in its 1981 and 1984 manifestos Labour included pledges, drafted by Geoffrey Palmer, member of parliament, to establish a Royal Commission to conduct ‘an authoritative and exhaustive reappraisal of electoral law’. When Labour finally came to power late in 1984, Palmer, now Minister of Justice and therefore responsible for electoral matters, moved quickly to establish a five-member Royal Commission on the Electoral System.

The Royal Commission Chosen through a process dominated by Palmer, the commissioners were distinguished and non-political. With John Wallace, a judge of the High Court as chair, the other members were a former government statistician, a research officer who was also Maori and two academics – one a constitutional lawyer and the other a political theorist. The Commission approached its task in a thorough and systematic manner. It employed two research officers and several consultants as well as support staff, received 804 written submissions, held public hearings, including five on Maori marae (gathering places), and travelled to Australia, Canada, Britain, Ireland and Germany. It presented its recommendations and analysis in a report of more than 500 pages, Towards a Better Democracy, an impressive work of applied political theory. In that volume, the Commission used ten criteria to

New Zealand

533

evaluate New Zealand’s existing single-member district plurality system and 13 alternatives. Three options were compared in depth – single-member district plurality, the single-transferable vote and mixed-member proportional (a label coined by the Commission). The Commission judged MMP best on seven of the criteria, especially fairness to political parties, representation of minority and special interest groups, Maori representation and effectiveness of parliament. The five commissioners therefore unanimously recommended MMP and proposed a national referendum to decide between it and the existing system. Although the Commission noted that MMP would preserve some valued elements of New Zealand’s political tradition, such as strong parties and constituency representation, its plan was, on the whole, remarkably unexpected and radical. As the first list-based proportional representation system in an Englishspeaking country, MMP would in all probability produce multi-party parliaments and coalition governments, thus ending two central characteristics of New Zealand’s Westminster system – two-party dominance and single-party government. How could such a recommendation have been produced by a commission appointed under the auspices of a party that for half a century had been part of a two-party parliamentary duopoly and at the time enjoyed a monopoly of governmental power? At no time from 1984 through to the referendum of 1993 did a majority of Labour’s parliamentary caucus ever support MMP, and, with the exception of Palmer, opposition was especially strong among cabinet members. At most Labour might have compromised on a parallel mixed system, in which all the existing single-member district plurality seats would be retained with the addition of 20–30 seats assigned on a proportional, but non-compensatory basis. Such a plan (known in New Zealand as a ‘supplementary member’ system) was proposed in the Party’s submission to the Royal Commission and, in 1988, by a Labour-dominated parliamentary select committee that inquired into the Commission’s report. The proximate answer to the puzzle lies in the commissioners’ own decision to seek what they judged to be the best electoral system for New Zealand, without concern for politicians’ preferences or political feasibility. Such a runaway commission became possible, though not preordained, because its members were highly capable and politically independent, and because its terms of reference were expansive. The appointments and charge in turn were due largely to the influence of Geoffrey Palmer. A professor of constitutional law before entering parliament, Palmer was an unusual politician, ultimately more dedicated to constitutional reform than to his political career or the fortunes of his party. Despite subsequent speculation that he was committed in advance to proportional representation, such a conclusion is not warranted by the evidence of his own previous writings and the commissioners’ consistent testimony that they were neither pre-committed nor pressured. Nevertheless, Palmer certainly did admit proportional representation as a possibility, both explicitly in the Commission’s terms of reference and by appointing prestigious, open-minded individuals known for independent judgment. After the Commission reported, Palmer became

534

Jack H. Nagel

a proponent of MMP, but – even when prime minister in 1989–90 – could not obtain support for the Commission’s recommendations from the Labour cabinet and caucus. Why then did the party fail to block off the possibility of radical reform when the Commission was established in 1985? Reasons include his colleagues’ deference to Palmer – who was deputy prime minister as well as the cabinet specialist on constitutional matters – and preoccupation with other matters as the party, returning to power after nine years in opposition, initiated major changes in economic, social and foreign policies. In addition, Labour MPs may have expected the Commission to focus on other aspects of its wide-ranging charge about which the party had long-standing concerns, such as redistricting, the number of Maori seats and campaign finance. As for the electoral system itself, few expected any proposal more sweeping than a very limited parallel mixed system, which may have been seen (mistakenly) as a remedy for the ‘wrong winner’ results of 1978 and 1981.

The surprising triumph of reform After the Royal Commission reported, horrified politicians of both major parties attempted to put the genie of reform back in the bottle. They succeeded in preventing action for six years, but then surrendered control of the issue by agreeing to referendums that produced a result they did not desire. In schematic summary, three main factors account for the politicians’ failure. First, the prestige of the Royal Commission and the persuasiveness of its report legitimized radical reform and unified reformers around MMP as the focal alternative to the status quo. Second, the Commission’s report inspired a poorly funded but vigorous and persistent grassroots movement, the Electoral Reform Coalition, that never let the press, public or politicians forget the Commission’s proposal for a referendum on MMP. Also converted to the cause were minorities within both National and Labour caucuses, as well as in their extra-parliamentary parties, which made it difficult for the parties to ignore or suppress the issue. In particular, the threat of a private members’ bill in 1989 helped force the parties to address the issue in their 1990 campaigns. Third, competitive dynamics, combined with error and miscalculation, helped put electoral reform back on the agenda. In a televised leaders’ debate during the 1987 election campaign, Labour Prime Minister David Lange, in response to a question from the Electoral Reform Coalition leader, promised to hold a binding referendum on electoral reform. Lange himself opposed MMP, and his party had adopted no such policy. Although later described as a gaffe, his action seems to have been an unreflective attempt to seize an advantage in the immediate context of the debate. After Labour won the election, its caucus refused to honour Lange’s rash and unauthorized promise. The National opposition, though even more opposed to reform than Labour, could not resist using the broken promise to embarrass the government. As a result, in their 1990 manifestos, both parties promised referendums. Because polls at the time showed public opinion evenly divided on the issue, politicians probably believed they had a good chance of

New Zealand

535

maintaining the status quo, especially as each hoped to manage the referendum in a favourable way. Labour promised only an indicative (non-binding) vote on MMP, and National was vague about the options it would offer to the public. After foot-dragging and ineffectual attempts to divert the reform impulse toward other objectives (e.g. re-instituting a second legislative chamber), the victorious National Party ultimately kept its promise, but instead of scheduling an early, one-on-one vote between mixed-member proportional (MMP) and first past the post (FPP, as single-member district by plurality rule was generally called in New Zealand), the government first held a complex indicative referendum in September 1992. In it, voters were asked whether first past the post should be retained and, if it were to replaced, which of four alternatives they preferred – STV, the alternative vote, a parallel mixed system or MMP. In an astonishing outcome, 84.7 per cent voted to reject first past the post, and in the vote among the reform alternatives, MMP won 70.5 per cent. This flood of opposition to the existing system resulted from the confluence of two previously independent political streams – the movement for electoral reform and the reaction to an extraordinary process of economic liberalization that had convulsed New Zealand since 1984. The 1992 vote coincided with a trough of economic discontent, as economic restructuring to that point had failed to fulfil politicians’ promises of improved welfare, delivering instead high unemployment, declining real wages and near-zero growth. Nevertheless, it would be a mistake to interpret the vote as merely a backlash against economic policies or conditions. The reaction of many New Zealanders was fundamentally political. In their eyes, the spectre of ‘elective dictatorship’ had become a reality as zealous factions exploited plurality elections, implicit logrolling, party discipline and one-party rule to impose radically disruptive policies without first winning broad-based popular consent. The Royal Commission’s primary objectives – representational fairness for parties and minorities – were thus augmented by a new appreciation for multi-party politics and coalition governments. ‘Consensus’ became the byword that summarized New Zealanders’ aspirations for policies based on widespread agreement and a more cooperative style of politics. Cowed by their overwhelming repudiation in the 1992 referendum, most leaders of both major parties decided that high-profile efforts on their part to resist MMP would be counter-productive. The National Party government prepared for a binding referendum between mixed-member proportional and first past the post, to be held at the same time as the 1993 parliamentary elections. The outcome seemed a foregone conclusion until a group of business leaders stepped into the breach left by demoralized politicians. Their pro-first past the post Campaign for Better Government launched a lavishly funded, sophisticated advertising blitz. Turning anti-politician sentiment against mixed-member proportional (MMP), the Campaign emphasized two unpopular secondary aspects of the proposal – closed lists and a 20 per cent increase in the size of parliament. The Campaign for Better Government advertisements depicted list members of parliament (MPs) as faceless ciphers and equated MMP with ‘More MPs’. Aided by a recovering economy and substantially higher turnout than in 1992, the Campaign effort succeeded in greatly

536

Jack H. Nagel

narrowing the gap but fell short of victory. The voters chose mixed-member proportional over first past the post by a 54–46 per cent margin. In addition to influences already discussed, the triumph of electoral reform in New Zealand was promoted by three strong norms in that country’s political culture: (1) popular sovereignty; (2) the importance of manifesto promises; and (3) fairness. 1. It was very difficult for politicians to resist appeals to ‘let the people decide’ given New Zealand’s strongly populist tradition, its long history of using referendums (albeit mostly to decide liquor laws) and specific provisions in the Electoral Act for reform by referendum. 2. For most of the twentieth century, New Zealanders’ trust in government was underpinned by a tradition that bound otherwise unfettered leaders through an ex ante contract with voters – a government should enact a policy if and only if it was included in its pre-election party manifesto. Violations of that norm in 1984–93 contributed greatly to disillusionment with politicians and the political system, but the tradition remained strong enough that reformers could exploit it, keeping both parties moving toward a fate neither wanted. In particular, National Prime Minister Jim Bolger insisted on honouring his party’s commitment to hold a referendum despite the reluctance of his colleagues and his own personal opposition to MMP. 3. In an egalitarian culture that values giving everyone a ‘fair go’, the reformers’ victory may have been saved by a last-minute backlash against the big-business interests who bankrolled the Campaign for Better Government. New Zealand’s tradition of fairness was also reflected in another influence that helped offset the huge funding disparity between the Campaign and the Electoral Reform Coalition. This was a well-financed government-sponsored information programme. Led by the chief ombudsman, the Electoral Reform Panel disseminated highly professional videos, advertisements and brochures. Although strictly neutral, this educational effort, combined with a similar programme before the 1992 referendum, improved public understanding of electoral alternatives.

Serendipity and the Maori seats In every respect but one, New Zealand’s post-1993 electoral system simply transferred to its smaller, non-federal setting the basic principles of MMP developed in West Germany after the Second World War. The distinctive and original element of MMP in New Zealand is its adaptation of the dual-constituency system for representing the Maori minority – a feature contrary to the Royal Commission on the Electoral System report which recommended abolition of the Maori seats. Indeed, a major reason for the Commission’s choice of a list proportional representation system was its perceived ability to promote fair and effective representation of Maori without the embarrassment of an illiberal racial distinction. After the 1992 referendum, the government set about framing the precise terms of the MMP system that would be submitted to a binding referendum in 1993.

New Zealand

537

The final legislation incorporated just two significant changes into the Royal Commission plan. One was an increase in the threshold from 4 to 5 per cent. The other concerned continuation of separate Maori constituencies. After a consultative process that included numerous hearings on marae, the parliament’s Electoral Law Committee concluded that there was a clear consensus among Maori in favour of keeping the dual constituency system. Maori were sceptical about the ability of MMP without the dual-constituency element to ensure their representation, but a second reason for their position appears to have been determination, in the face of immigration from Asia and the Pacific, to preserve their distinctive constitutional status – symbolized by the guaranteed seats – as tangata whenua (original people of the land) and as one of two partners in the founding of New Zealand by the 1840 Treaty of Waitangi between Maori chiefs and the British. In deference to this opinion, the Committee recommended that under MMP the single-member constituencies should continue to be of two types, general and Maori. They also recommended retention of the separate Maori electoral roll and the Maori option, which allows individuals of Maori descent to choose once every five years (before the scheduled redistricting) whether to vote on the Maori or general roll. In a major departure from past practice, the final bill also provided that the number of Maori seats would no longer be fixed at four, but instead would fluctuate up or down depending on the number of people on the Maori roll, using the same population quotas as would be applied to general constituencies. This principle, advocated by Maori leaders and the Labour party, had previously been blocked by National, which feared a partisan disadvantage under single-member districts by plurality because Maori seats had been monopolized by Labour. Ironically, and serendipitously, MMP permitted a resolution of that impasse, because it allocates seats to parties according to their party votes rather than the number of constituencies they win. Thus National no longer stood to lose by allowing the number of Maori seats to increase. Due to campaigns by Maori organizations encouraging voters to opt for the Maori roll, the number of Maori seats rose steadily in the early years of MMP – to five in 1996, six in 1999 and seven in 2002. As a percentage of single-member constituencies, the increase was dramatic – from 4.0 per cent under single-member districts by plurality in 1993 to 10.1 per cent in 2002. In addition, as anticipated by the Royal Commission, parties competing for party votes of Maori electors placed Maori candidates in winnable positions on their lists, resulting in the election of nine Maori list members of Parliament in 1996, seven in 1999 and nine again in 2002. Combined with the victory of several Maori candidates in general constituencies, the percentage of Maori members of Parliament in 2002 reached 15.8 per cent, approximately proportional to the number of New Zealanders of Maori descent.

Consequences of the new system The first three elections using MMP took place in 1996, 1999 and 2002. In a number of obvious ways, the new system fulfilled important expectations of its proponents (see Table 31.1). The Royal Commission’s judgements about fairness to parties

538

Jack H. Nagel

Table 31.1

New Zealand: electoral systems and political consequences Elections by plurality 1984

1987

1990

Elections by MMP 1993

1996

1999

2002

Index of electoral 16.9 14.0 17.2 13.2 18.0 14.5 26.8 volatility Index of 19.0 11.8 21.3 26.2 7.7 6.2 4.9 disproportionality No. of parties in 3 2 3 4 6 7 7 parliament Effective no. 2.0 1.9 1.8 2.2 3.8 3.4 3.8 parties in seats Type of government 1-party 1-party 1-party 1-party Majority Minority Majority formed majority majority majority majority coalition coalition coalition % women MPs 12.6 14.4 16.5 21.2 29.2 30.8 28.3 % minority MPs* 6.3 5.2 5.1 8.1 16.7 16.7 20.0 Note: 1993 was a transitional election – conducted in single-member districts by plurality rule, but with the adoption of MMP widely expected. * Maori, Pacific Islander and Asian.

and previously under-represented groups were quickly and dramatically realized. The index of disproportionality between parties’ votes and seats declined from an average of 19.6 in the last four first-past-the post elections to 6.3 in the first three MMP elections. Comparing the same periods, the proportion of MPs from ethnic minorities (not only Maori, but also Asians and Pacific Islanders) almost tripled, from 6.2 to 17.8 per cent. Women, who had already been advancing steadily under first past the post, increased their share of parliamentary seats by almost 50 per cent in the first MMP election and thereafter held steady at about 30 per cent of members of parliament. Indeed, from late 1997 to this writing, New Zealand has been continuously led by female prime ministers – first, Jenny Shipley of the National party and, since late 1999, Helen Clark of Labour. Also as expected, MMP fostered multi-party parliaments. With votes for minor parties now translated proportionally into seats (except for parties failing to qualify for list seats), each MMP election awarded seats to six or seven parties, and the effective number of parties roughly doubled compared with parliaments elected in single-member districts by plurality. However, the fortunes and sometimes the identities of parties in Parliament varied strikingly The index of electoral volatility was, on average, even greater under MMP than in the turbulent last years of FPP, and it reached its highest level in 2002. National and Labour survived the transition to remain the largest parties, occupying the centre-right and centre-left respectively, but National’s vote declined steadily from a peak of 47.8 per cent in 1990 to a mere 20.9 per cent in 2002. Labour, in disarray during the late 1980s, by the late 1990s had adapted successfully to MMP. On the left, the Alliance, which won 18.2 per cent of the vote (but only two seats) in the last first-past-the-post election,

New Zealand

539

effectively disintegrated by 2002, although one of its former components, the Greens, surpassed the threshold in both 1999 and 2002. On the economic right, the free-market ACT (originally the Association of Consumers and Taxpayers) maintained a stable vote but never a share of governmental power. The greatest instability occurred in the centre, where several competitors sought to establish a pivotal position between Labour and National. Centrism alone proved insufficient, so successful claimants combined moderation on left–right socio-economic policies with emotionally stronger appeals based on cross-cutting issues as well as the personalities of their leaders. The New Zealand First (NZF) party exploited an anti-immigration stand and the half-Maori heritage of its leader, Winston Peters, to attract votes from an odd-bedfellows combination of Maori and socially conservative, often elderly whites. NZF won 17 seats in 1996, fell to four in 1999 and then rebounded to 13 in 2002. In 2002, Peter Dunne, previously his party’s sole member of parliament (able to win his electorate because National did not run a candidate against him) led the United Future party to eight seats by performing well in televised debates and allying with previously unsuccessful Christian candidates. The major parties, no longer able to manufacture parliamentary majorities from a plurality (if that) of popular votes, were forced to govern in cooperation with smaller parties. Although coalition government fulfilled two major objectives of reform – providing a broader electoral base for legislation and a check on the power of the leading party – the manner in which the first coalition formed in 1996 gave MMP a very rocky start. After that election, the 17 votes held by New Zealand First were necessary to any feasible legislative majority. Most voters expected NZF to join with Labour and the Alliance, because during the campaign Winston Peters had concentrated his fire on the incumbent National government. Instead, Peters and his NZF colleagues engaged in nine weeks of simultaneous negotiations with both Labour and National. In the end, they drove a hard bargain with National, winning major policy concessions and a disproportionate share of cabinet posts. This process and its outcome discredited the concept of coalition government, confirmed opponents’ predictions that MMP would give minor parties excessive power, and disillusioned Alliance and Labour voters who had expected MMP would enable them to expel the unpopular National government from office. Public support for the new electoral system rapidly plummeted. It remained low through 1999, as the National-NZF coalition proved fractious and unstable, giving way to a National-led minority government in August 1998 when the NZF caucus disintegrated. If New Zealanders had voted again on MMP in the late 1990s, they might have replaced it with a single-member district plurality or a parallel system. The 1993 legislation, however, called for a trial of two elections followed by a parliamentary review before June 2002. The 1999 election and its aftermath restored support for MMP. Before that contest, Labour and the Alliance agreed on arrangements under which they would govern in coalition if, as expected, the voters turned away from the National-led government. Although the partners came out of the election two seats short of a majority, they quickly formed a minority government with support on crucial votes from the Greens. Labour Prime Minister Helen Clark provided

540

Jack H. Nagel

effective leadership and the standing of both her party and MMP rose markedly in opinion surveys. In 2001, a parliamentary select committee conducted the mandated review of MMP. By that time, Clark, previously a critic of MMP, realized her party could thrive under the proportional system. The committee was directed by parliament to reach its conclusions on the basis of unanimity or near unanimity, ‘having regard to the numbers in the House represented by each of the members of the committee’. This charge practically guaranteed that the main features of the status quo would not be disturbed. Minor parties, which held 32 of the 120 seats, had a strong interest in maintaining MMP, as did the 25 Labour and National MPs elected from party lists. Of the 290 submissions the committee received from individuals and groups, a substantial plurality favoured keeping MMP. Defying a 1999 citizens-initiated indicative referendum that overwhelmingly called for a reduction in the size of the House to 99 members, committee members from the Labour, Alliance and Green parties insisted on keeping the number at 120 – a decision important to maintaining proportionality, because any decrease would surely have been at the expense of list seats. Again defying public opinion, the same parties, joined by the ACT, rejected holding a new referendum on whether to keep MMP. On the issue of retaining separate Maori constituencies, only the ACT sought abolition. The committee also supported two controversial features of the status quo – closed party lists and the one-district alternative threshold for list seats. With the continuation of proportional representation assured, it appeared for a time, paradoxically, that New Zealand might return to single-party rule, because Labour’s support in some opinion polls during 2001–02 exceeded 50 per cent. Meanwhile, backing for its coalition partner, the Alliance, dwindled as that party had difficulty maintaining a distinctive position while also responsibly sharing power. Perhaps because some New Zealanders did not desire an unchecked government, in the election itself Labour won only 41.3 per cent of the vote and 52 seats. Joined by the two members of Parliament from Jim Anderton’s Progressive Coalition (a remnant of the Alliance), Labour formed a minority coalition government, with support on confidence votes guaranteed by United Future. To win legislative majorities, it was able to turn to any of three minor parties – the Greens, United Future and New Zealand First. Thus, in the early years of the twenty-first century, New Zealand under proportional representation moved toward a Scandinavian pattern, in which a dominant party on the centre-left led minority governments while parties constituting a fragmented opposition were unable or unwilling to coalesce against it.

Wider implications of the electoral reform in New Zealand Among the lessons that can be drawn from the adoption of MMP in New Zealand, three may be of particular value for readers interested in reform possibilities and processes elsewhere:

Table 31.2

New Zealand: electoral results, 1984–2002

Values/Greens Soc. Credit/ NewLabour/Alliance/ Labour United NZ/Future NZFirst National NZP/ACT Democrats Progressive NZ/United Future v/s v/s v/s v/s v/s v/s v/s v/s 1984 1987 1990 1993 1996 1999 2002

0.2/0 – 7/0 – – 5/7 7/9

8/2 6/0 2/0 – – – –

– – 5/1 18/2 10/13 8/10 2/2

39/56 48/58 35/29 35/45 28/37 39/49 41/52

– – – – 1/1 2/1 7/8

– – – 8/2 13/17 4/5 10/13

36/37 44/39 48/67 35/50 34/44 30/39 21/27

12/0 0.3/0 – – 6/8 7/9 7/9

Other

Total seats

v/s 5/0 2/0 3/0 4/0 8/0 5/0 5/0

95 97 97 99 120 120 120

Key: v: percentage of votes, aggregated across districts through 1993 and measured by the party vote under MMP from 1996 on; s: number of seats. Figures in bold type are for the governing party or parties in the initial post-election period. Greens were part of the Alliance in 1993 and 1996. Democrats were part of the Alliance in 1993, 1996 and 1999, and part of the Progressive Coalition in 2002. A rump Alliance party contested in 2002; its vote is included in ‘Other’. United NZ and Future NZ were separate parties in 1999. Vote total shown is the sum of their party votes. United NZ won a district seat in both 1996 and 1999.

541

542

Jack H. Nagel

1. Contrary to theories that portray existing electoral systems as ‘locked in’ because politicians in power will not change the rules under which they have succeeded (except perhaps to their further advantage), New Zealand’s experience demonstrates that fundamental reform can be accomplished in an established democracy, though it will remain a rare event. Generalizing from the process in New Zealand, elements that combine to make reform possible include the following: (a) clashing interests and competitive dynamics among the parties with a stake in the status quo; (b) the possibility that at least some powerful leaders will be concerned more with constitutional principles than with individual or partisan advantage; (c) provisions in law or practice allowing for influence or decision by disinterested actors; (d) the beliefs, commitment and skill of citizen activists; and (e) accident, miscalculation or historical conjunction. 2. When an advisory commission on electoral reform is established, its members may have more ultimate influence if they disregard short-term political feasibility. On this point, the contrast between New Zealand’s Royal Commission and the Jenkins Commission in Britain is instructive. Because the Royal Commission on the Electoral System used fundamental principles to evaluate a wide range of options, its recommendation had great persuasive force and moral authority. The Jenkins Commission never inspired a popular movement, in part because of the perception that its ingenious plan was contrived to satisfy the partisan requirements of the government in power. 3. If a reform is enacted, even a country with a long democratic tradition such as New Zealand may experience an unsatisfactory period of adjustment to the new rules. Delaying a reappraisal until at least two or three elections have been held offers a fairer trial – as well as the likelihood that a sufficient number of politicians will have developed a vested interest in the new system.

Note 1. The author wishes to thank Neill Atkinson, Paul Harris, Keith Jackson, Alan McRobie and Jack Vowles for information and helpful comments.

Further reading • Barker, Fiona, Jonathan Boston, Stephen Levine, Elizabeth McLeay and Nigel S. Roberts (2001) ‘An Initial Assessment of the Consequences of MMP in New Zealand’, in Matthew Soberg Shugart and Martin P. Wattenberg (eds), Mixed-Member Electoral Systems: The Best of Both Worlds. New York and Oxford: Oxford University Press. • Denemark, David (2001) ‘Choosing MMP in New Zealand: Explaining the 1993 Electoral Reform’, in Matthew Soberg Shugart and Martin P. Wattenberg (eds), Mixed-Member Electoral Systems: The Best of Both Worlds? New York and Oxford: Oxford University Press. • Department of Justice (1986) ‘The Electoral Law of New Zealand: a Brief History’, in Royal Commission on the Electoral System, Report of the Royal Commission on the Electoral System: Towards a Better Democracy. Wellington: Government Printer, Appendix A. • Electoral Commission (2002) The New Zealand Electoral Compendium, 3rd edn. Wellington: Electoral Commission.

New Zealand

543

• Jackson, Keith and Alan McRobie (1998) New Zealand Adopts Proportional Representation: Accident? Design? Evolution? Aldershot: Ashgate.

• Nagel, Jack H. (1996) ‘Constitutional Reform and Social Difference in New Zealand’, Cardozo Journal of International and Comparative Law, 4, 2: 373–94.

• Nagel, Jack H. (1999) ‘The Defects of Its Virtues: New Zealand’s Experience with MMP’, in • • • • • • •

Henry Milner (ed.), Making Every Vote Count: Reassessing Canada’s Electoral System. Peterborough, Ont.: Broadview Press. New Zealand House of Representatives (2001) Report of the MMP Review Committee. Available at: . Royal Commission on the Electoral System (RCES) (1986) Report of the Royal Commission on the Electoral System: Towards a Better Democracy. Wellington: V. R. Ward, Government Printer. Vowles, Jack (1995) ‘The Politics of Electoral Reform in New Zealand’, International Political Science Review, 16, 1: 95–116. Vowles, Jack, Peter Aimer, Susan Banducci and Jeffrey Karp (eds) (1998) Voters’ Victory: New Zealand’s First Election under Proportional Representation. Auckland: Auckland University Press. Vowles, Jack, Peter Aimer, Helena Catt, Jim Lamare and Raymond Miller (1995) Towards Consensus? The 1993 Election in New Zealand and the Transition to Proportional Representation. Auckland: Auckland University Press. Vowles, Jack, Peter Aimer, Jeffrey Karp, Susan Banducci, Raymond Miller and Ann Sullivan (2002) Proportional Representation on Trial: The 1999 New Zealand General Election and the Fate of MMP. Auckland: Auckland University Press. Wallace, John (2002) ‘Reflections on Constitutional and Other Issues Concerning Our Electoral System: the Past and the Future’, Political Science, 54, 1: 47–65.

Appendix: Notes and Sources for Summary Tables

Notes Electoral systems are distinguished for having changed at least one of these elements: single- or multi-member districts, ballot, rule or formula. In contrast, changes are recorded within a single system for term, number of seats, number of districts, magnitudes and thresholds, which are given as ranges within each period (or as an average*). Public ballot and indirect elections are indicated when they exist; otherwise assume secret ballot and direct elections. The following abbreviations have been used: Maj. = Majority; Plu. = Plurality; Prop. = Proportional. ‘Universal MS’ indicates the introduction of universal male suffrage. The sign // indicates a period without elections or with authoritarian fake elections. See Glossary for definitions.

Sources General Blais, André and Louis Massicotte (1997) ‘Electoral Formulas: A Macroscopic Perspective’, European Journal of Political Research, 32: 107–29. Cox, Gary W. (1997) Making Votes Count. Strategic Coordination in the World’s Electoral Systems. New York: Cambridge University Press. Freedom House (R. D. Gastil and Adrian Karatnycky eds) (1972–2002) Freedom in the World. The Annual Survey of Political Rights and Civil Liberties. New Brunswick: Transaction <www.freedomhouse.org>. Katz, Richard S. (1997) Democracy and Elections. New York and Oxford: Oxford University Press. Lijphart, Arend (1994) Electoral Systems and Party Systems. A Study of Twenty-Seven Democracies, 1945–1990. New York and Oxford: Oxford University Press. Mackie, Thomas T. and Richard Rose (1991) The International Almanac of Electoral History, 3rd edn. London and Washington, DC: Macmillan-Congressional Quarterly. Massicotte, Louis, and André Blais (1999) ‘Mixed Electoral Systems: A Conceptual and Empirical Survey’, Electoral Studies, 18, 3: 341–66. Nohlen, Dieter (1996) Elections and Electoral Systems. Delhi: Macmillan India. Reynolds, Andrew and Ben Reilly (eds) (1997) The International IDEA Handbook of Electoral System Design. Stockholm: International Institute for Democracy and Electoral Assistance. Rose, Richard (ed.) (2000) International Encyclopedia of Elections. London and Washington, DC: Palgrave-Congressional Quarterly.

Websites <www.dodgson.ucsd/edu/lij> (Arend Lijphart Archive, University of California, San Diego) <www.electionworld.org> (Elections Around the World) <www.essex.ac.uk/elections> (University of Essex) <www.idea.int> (International Institute for Democracy and Electoral Assistance) <www.ifes.org> (International Foundation for Electoral Systems) <www.parline.org> (International Parliamentary Union) 544

Appendix 545

Americas Basadre, Jorge (1980) Elecciones y centralismo en Perú. Apuntes para un esquema histórico. Lima: Universidad del Pacifico. Botana, Natalio ([1977] 1998) El orden conservador. La política argentina entre 1880 y 1916. Buenos Aires: Sudamericana. Bushnell, David (1993) The Making of Modern Colombia. A Nation in Spite of Itself. Berkeley: University of California Press. Cantón, Darío (1973) Elecciones y partidos politicos en la Argentina. Historia, interpretación y balance. Buenos Aires: Siglo XXI. Delgado, Oscar (2002) ‘Los sistemas electorales para el congreso en Colombia (1821–2002)’, Estudios Socio-Jurídicos, 4, 2: 67–129. Duarte Oropesa, José A. (1975) Historiología cubana, 3 vols. Miami, FL: Universal. González Casanova, Pablo (ed.) (1985) Las elecciones en México. Mexico: Siglo XXI. Graham, Richard (1990) Patronage and Politics in Nineteenth-Century Brazil. Stanford, CA: Stanford University Press. Jones, Mark P. (1995) ‘A Guide to the Electoral Systems of the Americas’, Electoral Studies, 14, pp. 15–21. Jones, Mark P. (1997) ‘A Guide to the Electoral Systems of the Americas: An Update’, Electoral Studies, 16, pp. 13–15. Nohlen, Dieter (ed.) (1993) Enciclopedia electoral latinoamericana y del Caribe. San José, Costa Rica: Instituto Interamericano de Derechos Humanos. Porto, Walter Costa (1989) O voto no Brasil. Da colônia à 5 a Republica. Brasilia: Senado Federal. Rodríguez Arechavaleta, Carlos M. (2003) La quiebra de la democracia cubana, 1940–1952. Doctoral dissertation, Flacso, Mexico.

Website <www.georgetown.edu/pdba> (Georgetown University, Center for Latin American Studies)

Western Europe Bartolini, Stefano (1984) ‘Sistema partitico ed elezione diretta del capo dello stato in Europa’, Rivista Italiana di Scienza Politica, 14, 2: 223–43. Cole, Alistair and Peter Campbell (1989) French Electoral Systems and Elections since 1789. Aldershot: Gower. Caramani, Daniele (2000) Elections in Western Europe since 1815. London: Macmillan. Carstairs, Andrew McLaren (1980) A Short History of Electoral Systems in Western Europe. London: George Allen & Unwin. Fernández Domínguez, Arturo (1992) Leyes electorales españolas de diputados a Cortes en el siglo XIX. Estudio histórico y jurídico-político. Madrid: Civitas. Georgel, Jacques, Geoffrey-J. Hand, Christophe Sasse et al. (1977) Les régimes électoraux dans la Communauté Européenne. Paris: Cujas. Grofman, Bernard and Arend Lijphart (eds) (2002) The Evolution of Electoral and Party Systems in the Nordic Countries. New York: Agathon. Rueda, José-Carlos (ed.) (1998) Legislación electoral española (1808–1977). Barcelona: Ariel. Tavares de Almeida, Pedro (1997) ‘Reformas electorales y dinámica política en el Portugal liberal (1851–1910)’, in Salvador Forner (ed.), Democracia, elecciones y modernización en Europa. Siglos XIX y XX. Madrid: Catedra, pp. 73–108.

Eastern Europe Birch, Sarah (2000) ‘Elections and Representation in Post-Communist Eastern Europe’, in Hans-Dieter Klingemann and Kenneth Newton (eds), Elections in Central and Eastern Europe: The First Wave. Berlin: Sigma, pp. 13–35.

546

Handbook of Electoral System Choice

Filippov, Mikhail G. and Olga V. Shvetsova (1995) Political Institutions and Party Systems in New Democracies of Eastern Europe. Paper presented at the American Political Science Association annual meeting, Chicago. Flores Juberías, Carlos (1995) ‘The Transformation of Electoral Systems in Eastern Europe and Its Political Consequences’, Journal of Constitutional Law in Eastern and Central Europe, 2, 1: 1–59. Ishiyama, John T. (1997) ‘Transitional Electoral Systems in Post-Communist Eastern Europe’, Political Science Quarterly, 112, 1: 95–115. Moser, Robert (1999) ‘Electoral Systems and the Number of Parties in Post-Communist States’, World Politics, 51: 359–84. Shvetsova, Olga V. (1999) ‘A Survey of Post-Communist Electoral Institutions: 1990–1998’, Electoral Studies, 18: 197–409.

Africa Mozaffar, Shaheen (1997) ‘Electoral Systems and Their Political Effects in Africa: A Preliminary Analysis’, Representation, 34: 148–56. Nohlen, Dieter, Michael Krenerich and Bernhard Thibaut (eds) (1999) Elections in Africa, A Data Handbook. New York and Oxford: Oxford University Press. Reynolds, Andrew (1998) Electoral Systems and Democratization in Southern Africa. New York and Oxford: Oxford University Press.

Asia-Pacific Croissant, Aurel (2002) ‘Electoral Politics in Southeast and East Asia: A Comparative Perspective, manuscript, University of Heidelberg, Institute for Political Science. Fuh-sheng Hsieh, John and David Newman (eds) (2002) How Asia Votes. New York and London: Chatham House. Hicken, Allen and Yuko Kasuya (2003) ‘A Guide to the Constitutional Structures and Electoral Systems of East, South and Southeast Asia’, Electoral Studies. Nohlen, Dieter, Florian Grotz and Christof Hartmann (eds) (2001) Elections in Asia and the Pacific. A Data Handbook, 2 vols. Oxford and New York: Oxford University Press.

Glossary and Index

Note K: number of candidates; M: magnitude or number of seats; V: number of votes per voter. Absolute majority rule. The winner is the alternative with more than half the votes, whether by simple majority or by qualified-majority. Additional-member system. See personalized proportional representation. Alternative vote. In a single-member district by majority rule, each voter can order all the candidates as preferred; seats are allocated by majority first preferences, or by elimination of least voted candidates and voters’ second and further preferences (K = V > M = 1). Also called majority-preferential (Australia) or instant-runoff (United States) or optional-preferential voting (Papua New Guinea). See: Australia 1918– Ireland president 1937–

467, 475 208

Apparentement (French). In proportional representation systems, parties present separate lists of candidates but declare themselves linked (parented) for joint vote counting and seat allocation. See: France 1951–56 Italy 1953

195–7, 209 200–1, 237

Apportionment. Allocation of seats among districts on the basis of population. Also called redistribution (Britain). See exact quota, sufficient quota. Approval ballot or voting. In a single-member district, usually by plurality rule, each voter can vote for as many candidates he or she wishes, whether more, less or in equal number to seats in the district (K > V ≥ M = 1). At-large. A single multi-member district to elect all the seats in the assembly. Ballottage (French). See runoff. Bloc ballot. In a multi-member district by plurality rule, each voter can vote for a closed list of candidates (K ≥ M = V > 1). See: Canada 1874–1968 97 United States 1789–1968 103, 108–9, 155 Categoric ballot. Each voter can vote for only one candidate or party list. See closed list, single ballot, single non-transferable vote. Choice voting. See single transferable vote. Closed list. In a multi-member district by proportional representation, each voter has to choose a list of candidates as given (K ≥ M = V > 1). Also called liste pré-ordonnée bloquée (French), lista cerrada y bloqueada (Spanish). Coexistence. Mixed system in which different rules are used in different districts to allocate different portions of seats in the assembly. See: Costa Rica 1913–48 Greece 1956 Madagascar 1998– Panama 1984– Portugal 1896, 1915

98–9, 133 198–9 437 102 203–4 547

548

Handbook of Electoral System Choice

College. Chamber of electors, usually elected in multi-member districts, who choose the president or some other officer. Early experiences include the elections of the pope and several central European kings since the eleventh century, as well as the German emperor. See for presidential elections: Argentina 1853–1946 1955–63,1983–89 Brazil 1834–38 (for the election of Regent) Colombia 1821–53, 1863, 1904 Costa Rica 1889–1909

105, 110 105, 121 106 106, 133

Cuba 1901–48 Mexico 1824–1911 Peru 1828–76 United States since 1789 Venezuela 1830–77

106 107 108 108–9, 155, 164 109

Compensatory seat allocation. See personalized proportional representation. Constituency. See district. Cumulative ballot or vote. In multi-member districts by plurality rule, each voter can cast more than one vote for a single candidate (K ≥ V ≤ M > 1). See: Brazil 1904–30 Chile 1879–1924 Peru 1861–1919 Switzerland 1918– United States: some local and state electiosns

96–7, 121 97 102 206–7, 279 155, 164

Deviation from proportionality. See disproportionality. D’Hondt formula. Proportional representation formula invented by civil law professor Victor d’Hondt (Ghent, Belgium, 1841–1901). See sufficient quota. Disproportionality. Measure of the aggregate deviation (D) between the proportion of votes obtained and the proportion of seats received by each party. It can be calculated 1 Σ|v − s |, or by least squares: D = Σ v 2 – s 2 ½, where v is the either by deviation: D = -( i i i i i ) 2 proportion of votes and si is the proportion of seats for each party i, or by some other formulas. District. Area for counting votes and allocating seats. See also magnitude, multi-member district, single-member district. Also called tier, constituency or division (Britain), electorate (Australia and New Zealand), circonscription (French), circunscripción (Spanish). Districting. The drawing of district boundaries. Also called redistribution (Britain). Double-round runoff. See runoff. Double-vote system. See personalized proportional representation. Droop formula. Proportional representation quota invented by lawyer Henry R. Droop (London, UK, 1831–84). Divisor between the total number of votes and the total number of seats plus one in a district, 1/M + 1 (in order to prevent the allocation of more seats than available, this value is increased by one unit or rounded up to the next integer). It requires a supplementary criterion or tier to allocate the remaining seats: see largest remainders, multiple tiers. See: Czechoslovakia 1920–35, 1990 Czech Republic 1992–98 Ireland 1922– Slovakia 1992– South Africa 1994–

326 326, 347 200 330 438, 440

Effective number of parties. Measure of the number of parties (N) weighed by their size, whether in votes or in seats. N = 1/Σ pi2, where pi is the proportion of votes or seats for each party i. Exact or simple quota. Divisor between the total number of votes (to allocate seats) or inhabitants (for apportionment) and the total number of seats in a district, 1/M. It

Glossary and Index 549 requires a supplementary criterion or tier to allocate the remaining seats: see largest remainders, multiple tiers. Also called Hamilton, Hare or Hare-Niemeyer formula. See: Albania 1992– Austria 1971– Belgium 1919– Benin 1991– Brazil 1932–45 Colombia 1931– Costa Rica 1953– Cuba 1940–50 Denmark 1915– Ecuador 1947–96 El Salvador 1963– Estonia 1989– France 1986 Germany 1920–33, 1985– Honduras 1981– Indonesia1953–55, 1991–

325 193 193 436 96–7, 121 98 98–9, 133 99 194 99–100 100 326–7, 332 196–7, 209 197–8, 222 100 468, 497

Greece 1926, 1932,1936–51, 1958– Guatemala 1944–53 Israel 1951–69,1992– Italy 1924, 1946–48, 1956–92 Lithuania 1920–26, 1992– Madagascar 1993 Namibia 1990– The Netherlands 1918– Nicaragua 1987– Poland 1991 Romania 1993– Slovenia 1990– Sri Lanka 1978– Turkey 1965 Venezuela 1989–

198–9 100 468–9 200–1, 237 327–8, 332 437 437 201–2 101 329, 369 329 330 472 473 104

First-past-the-post. See plurality rule. Fixed quota. Number of votes to allocate seats. Proposed separately by Gergonne and Gilpin. See: Germany 1919

197–8

Flexible list. See open list. Free list. See apparentement, open ballot, open list. Gergonne formula. Proportional representation formula invented by mathematician Joseph Diaz Gergonne (Nancy, France, 1771 – Montpellier, France, 1859). See fixed quota. Gerrymandering. Manipulative districting or apportionment. Gilpin formula. Proportional representation formula invented by manufacturer Thomas Gilpin (Chester C., Pennsylvania, USA, 1728 – Winchester, Virginia, USA, 1778). See fixed quota. Hagenbach-Bischoff formula. Proportional representation procedure invented by physics professor Eduard Hagenbach-Bischoff (Basel, Switzerland, 1833–1910) to calculate the ‘sufficient’ or d’Hondt quota by starting tentatively with the quota 1/M + 1 and decreasing it down until fitting. See sufficient quota. Hamilton formula. Proportional representation quota invented by politician Alexander Hamilton (Newis, West Indies, 1755 – Weehawken, New Jersey, USA, 1804). See exact quota. Hare formula. Proportional representation quota invented by lawyer Thomas Hare (London, UK, 1806–1901). See exact quota. Hare-Niemeyer formula. See exact quota. Highest average. See sufficient quota. Imperiali formula. Proportional representation quota based on the divisor between the total number of votes and the total number of seats plus two in a district, 1/M + 2. It can be a supplementary criterion to allocate remaining seats after using some other quota: see largest remainders, multiple tiers. See: Italy 1956–92

200–1, 237

Instant-runoff. See alternative vote. Jefferson formula. Proportional representation quota invented by politician and United States President Thomas Jefferson (Shadwell, Virginia, USA, 1743 – Monticello, Virginia, USA, 1826). See sufficient quota. Largest remainders. Frequent supplementary formula to allocate the remaining seats after using the exact (Hare), Droop or other quotas.

550

Handbook of Electoral System Choice

Lemas. Each voter can vote for a single candidate; seats are allocated on the basis of all a party’s candidates’ votes, and filled by those candidates with higher numbers of votes within each party. See apparentement, open list. See: Uruguay 1919–99

104

Limited vote. In a multi-member district by plurality rule, each voter can vote for fewer candidates than seats in the district (K ≥ M > V > 1). Also called lista incompleta (Latin America). See single non-transferable vote. See: Argentina 1912–30 Bolivia 1924 Brazil 1875–78, 1892–1930 Colombia 1910–29 Italy 1882–90

95–110 95–6 96–7, 121 98 200–1

Japan 1946 Portugal 1884–95,1901–11 Spain 1878–1936 United Kingdom 1868–80

469, 512 203–4 204–5 207

Lottery. The winner(s) is selected by some random device producing impartiality and rotation in office. Magnitude. Number of seats to be elected in a district. Majority rule. The winner is the alternative with more than half the votes. Also called simple majority rule. See alternative vote, runoff. See also plurality rule, qualified majority rule, qualified plurality rule. Majority-preferential. See alternative vote. Minimum barrier. See threshold. Mixed system. Combination of different rules and procedures for the election of a single assembly, usually including plurality or majority rules and proportional representation (K ≥ M = V > 1 / K ≥ M = V = 1). See coexistence, multiple tiers, parallel system. Mixed-member majoritarian. See parallel system. Mixed-member proportional. See personalized proportional representation. Multi-member district. A district electing more than one seat. Multiple ballot. In a multi-member district by plurality or majority rules, each voter can vote for as many candidates as seats in the district (K ≥ M = V > 1). See also approval vote, cumulative vote, limited vote. Multiple tiers. In proportional representation systems, each voter can vote only for a list, but two or more overlapping districts (tiers) are used to allocate different portions of seats from the same votes; it is used frequently with exact (Hare), Droop or other quotas requiring a supplementary procedure to allocate the remaining seats. See: Austria 1918– Belgium 1919– Croatia 2000– Czechoslovakia 1920–35 Denmark 1920– Dominican Republic 1966– Ecuador 1978–96 El Salvador 1984– Estonia 1992– Germany 1920–33

193 193 326 326 194 99 99–100 100 326–7, 332 197–8, 222

Greece 1926, 1932, 1936–51, 1958– Guatemala 1985– Italy 1946–48, 1956–92 Netherlands 1933– Norway 1989– Poland 1991–97 Sri Lanka 1978– Sweden 1970– Venezuela 1989–98

198–9 100 200–1, 237 201–2 202 329, 369 472 205–6, 265 104

In mixed systems, each voter has only one vote, either for a single candidate or for a list, but two or more overlapping districts (tiers) are used to allocate different portions of seats by different rules, including plurality rule and proportional representation. See: Croatia 1995 Korea S. 1963–71, 1987– Mexico 1963–76, 1986– Senegal 1992–

326 469 101,145 438

Glossary and Index 551 Taiwan 1991–

472

Ofen-Bader formula. See sufficient quota. Open ballot. In a multi-member district by proportional representation, each voter can vote for as many candidates as seats in the district from any party or list. Also called panachage (French). See: Switzerland 1918–

206–7, 279

Open list. In a multi-member district by proportional representation, each voter may select one or more individual candidates from a single list. Also called preferential vote, flexible list, free list, liste pré-ordonnée non bloquée (French), lista cerrada y no-bloqueada (Spanish); when the selection of an individual candidate is compulsory, quasi-list system. Optional-preferential voting. See alternative vote, open list. Ordinal ballot. Each voter can order all or a certain number of candidates as preferred. See alternative vote, single transferable vote, supplementary vote. Panachage (French). See open ballot. Parallel system. Mixed system in which each voter has two votes, one for a single candidate and one for a list, and different portions of seats are allocated on the basis of each vote by two different rules, including plurality or majority rule and proportional representation. Also called mixed-member majoritarian. See: Albania 1992– Armenia 1995– Brazil 1933 Bulgaria 1990– Georgia 1995– Hungary 1922–35, 1990– Italy 1993– Japan 1994–

325 467 96–7, 121 325–6 468 327, 359 200–1, 237 469, 512

Lithuania 1992– Macedonia 1998– Mexico 1977–85 Philippines 1995– Russia 1993– Thailand 1997– Ukraine 1993–

327–8, 332 328 101, 145 471–2 329, 382 472–3 330

Personalized proportional representation. In a system of proportional representation, each voter can vote either for a single candidate (which counts for the party) or for a single candidate in a single-member district and a party list in a multi-member district; seats are allocated on the basis of party votes; individual candidates are selected to fill part of the seats by plurality rule in single-member districts and the rest from party lists. Also called additional-member system, compensatory allocation, double-vote system or mixed member proportional. See: Bolivia 1994– Denmark 1918 Germany 1949– New Zealand1993– United Kingdom regional elections Venezuela 1989–98

95–6 194 197–8, 222 471, 530 294 104

Plurality rule. The winner is the alternative(s) with highest number(s) of votes, not necessarily achieving a majority or any other quota. Also called relative majority rule or first past the post (Britain and Commonwealth), winner-takes-all (USA), pluralité (French), pluralidad (Spanish). See for assembly elections: Argentina 1826–1962 Australia 1902–17 Bangladesh 1971– Botswana 1965– Brazil 1890–1930 Canada 1867–

95, 110 46, 475 468 436 96–7, 121 97

Malawi 1994– Malaysia 1957– Mongolia 1992 Nepal 1959– New Zealand 1890–1905, 1914–93

437 470 470 470 471, 530

552

Handbook of Electoral System Choice

Argentina 1826–1962 Chile 1833–1924 Colombia 1857–1929 Denmark 1848–1915 Ecuador 1895–1928 El Salvador 1886 France 1871–72 Ghana 1956– Greece 1864–1923, 1928, 1933, 1952 India 1950– Italy 1882–90 Jamaica 1962– Kenya 1963– Korea S. 1950 Lebanon 1953–

95, 110 97 98 194 99–100 100 195–7 436 198–9 468, 487 200–1, 237 101 436 439 470

Malawi 1994– Nicaragua 1984–90 Nigeria 1959– Pakistan 1970– Papua New Guinea 1975– Peru 1828–1919 Philippines 1935–95 Slovenia 1991– Spain 1868–1923 Sri Lanka 1946–1977 Taiwan 1996– Thailand 1978–96 Trinidad-Tobago 1962– Turkey 1946–57 United Kingdom 1832– United States 1789–

437 101 437 471 471 102 471–2 330 204–5, 265 472 472 472–3 103 473 207, 294 103, 155, 164

See for presidential elections: Bangladesh 1977– Bolivia 1956–64 Bosnia-Hercegovina 1998– Brazil 1945–60 Colombia 1853–60, 1910–90 Dominican Republic 1962–94 Honduras 1981– Korea S.1950– Malawi 1994–

474 105 331 105, 121 106 106 107 474 439

Mexico 1917– Panama 1946– Paraguay 1989– Peru 1896–1915 Philippines 1935– Slovenia 1991– Uruguay 1899–1994 Venezuela 1946–

107, 145 108 108 108 474 331 109 109

Preference voting. See alternative vote, single transferable vote, supplementary vote. Preferential vote. See open list. Proportional representation. In a multi-member district, seats are allocated to different parties or lists in proportion to the votes received. See exact quota, fixed quota, sufficient quota, personalized proportional representation, as well as d’Hondt, Droop, Gergonne, Gilpin, Hagenbach-Bischoff, Hamilton, Hare, Imperiali, Jefferson, Sainte-Laguë, Webster formula. Qualified majority rule. The winner is the alternative with certain proportion of votes fixed between majority and unanimity, such as 3/5, 2/3, 3/4, etc. Also called supermajority rule. Qualified plurality rule. The winner is the alternative with higher number of votes and certain fixed proportion lower than half, such as 1/4, 1/3, 2/5, etc. See runoff. Quasi-list system. See open list. Quota. Proportion of votes for winning a seat. See exact quota, fixed quota, simple quota, sufficient quota. Redistribution. See apportionment and districting. Relative majority rule. See plurality rule. Runoff. Second round of voting, either by majority rule between the two most voted candidates at the first round (K ≥ M = V = 1 / K = 2 > M = V = 1) or by plurality rule among a higher number of candidates (K ≥ M = V = 1 / K > M = V = 1). Also called double-round runoff, second round, second ballot, two-round system, ballottage (French). See for assembly elections: Albania 1991 Armenia 1990 Belarus 1996– Belgium 1831–98 Benin 1991– Brazil 1881–89

325 467 325 193 436 96–7, 121

Lebanon 1943–51 Lithuania 1919, 1978–90 Mali 1992– Moldova 1990– Mongolia 1990– The Netherlands 1848–1913

470 327–8, 332 437 328–9 470 201–2

Glossary and Index 553

Albania 1991 Central African Republic 1993– France 1817–71, 1875–1936, 1958–81, 1988– Germany 1871–1912 Greece 1844–62 Italy 1870–80, 1892–1913 Latvia 1979–90

325 436 195–7, 209 197–8, 222 198–9 200–1 327, 332

Lebanon 1943–51 New Zealand 1908–11 Norway 1906–18 Poland 1989 Portugal 1859 Spain 1836, 1837–67 Ukraine 1990 Uruguay 1996–

470 471, 530 202 329, 369 203–4 204–5, 253 330 104

After first round by qualified plurality rules: Mongolia 1996– Portugal 1852–56

470 203–4

See for presidential elections after first round by absolute majority rule: Argentina 1972 Armenia 1991– Austria 1945– Belarus 1996– Benin 1991– Bolivia 1899–1951, 1980– Brazil 1892–1930, 1986– Bulgaria 1991– Cape Verde 1992– Central African Republic 1993– Chile 1925– Colombia 1991– Costa Rica 1913–32 Croatia 1990– Dominican Republic 1995– Ecuador 1895–1996 Estonia 1992 El Salvador 1886– Finland 1988– France 1848, 1962– Georgia 1995–

105, 110 474 208 331 439 105 105, 121 331 439 439 105 106 106, 133 331 106 107 331–2 107 208 208 474

Germany 1918–30 Ghana 1979– Guatemala 1944– Honduras 1879–1957 Israel (for prime minister) 1996–2001 Lithuania 1993– Macedonia 1994– Madagascar 1992– Mali 1992– Moldova 1991– Mongolia 1993– Mozambique 1994– Namibia 1990– Nicaragua 1911–32, Peru 1978– Poland 1990– Portugal 1976– Romania 1990– Russia 1990– Ukraine 1994–

208 439 107 107 474 331–2 331 439 439 331 474 439 439 107–8 108 331, 369 208 331 331, 382 331

After first round by qualified plurality rules: Argentina 1994– Costa Rica 1936– Ecuador 1998– Kenya 1963–

105, 110 106, 133 107 439

Nicaragua 1995– Nigeria 1979– Peru 1931–63

107–8 439 108

Sainte-Laguë formula. A proportional representation formula invented by mathematician André Sainte-Laguë (Paris, France, 1882–1950) which is based on a quota higher than the sufficient quota, which is used to allocate seats together with half quotas. Procedures include the divisor series of odd numbers. Modified versions increase the first divisor to 1.4 or other value. Also called Webster formula. See: Bosnia-Hercegovina 1996– Denmark 1953– Latvia 1922–31, 1992– New Zealand 1993–

325 194 327, 332 471, 530

Norway 1953– Poland 1991, 2001– Sweden 1952–

202 329, 369 205–6, 265

554

Handbook of Electoral System Choice

Second ballot. See runoff. Second round. See runoff. Simple majority rule. See majority rule. Simple quota. See exact quota. Single ballot. In single-member districts, each voter can vote for a single candidate (K ≥ M = V = 1). Single non-transferable vote. In a multi-member district by plurality rule, each voter can vote for only one candidate (K ≥ M > V = 1). See limited vote. See: Japan 1947–93 Spain 1865–76

469, 512 204–5

Single transferable vote. In a multi-member district by some formula of proportional representation (usually Droop quota), each voter can order all the candidates as preferred; seats are allocated on the basis of quotas, elimination of least voted candidates and voters’ second and further preferences (K = V > M > 1). Also called choice voting, or preference voting (United States). See: Australia Senate 1949– Denmark-Schlesswig 1855 Estonia 1990 Ireland 1922–

467 194 326–7, 331 200

Also some local elections in Australia, Canada, South Africa, United Kingdom, United States. Single-member district. A district electing a single seat. Sounder and greater part rule. The winner is the alternative favoured by voters with special qualifications or seniority. Sufficient quota. Quota of votes (or inhabitants) able to allocate all seats in a district. Procedures to calculate the quota include the divisor series of natural numbers, using tentatively the 1/M + 1 quota, and others. Also called highest average, cifra repartidora (Latin America) as well as d’Hondt, Hagenbach-Bischoff, Jefferson or Ofen-Bader formulas. See: Argentina 1963– Austria 1918– Belgium 1899– Bolivia 1997– Brazil 1950–, Bulgaria 1911–31, 1991– Cape Verde 1992– Chile 1925– Croatia 2000– Czech Republic 2002– Denmark 1915–50 Dominican Republic 1966– Estonia 1919–32, 1992– Finland 1906– Germany 1949–83 Greece 1926, 1932,1936–51, 1958–85, 1990– Guatemala 1984–

95, 110 193 193 95–6 96–7, 121 325–6 436 97 326 326, 347 194 99 326–7, 332 194–5 197–8, 222 198–9 100

Israel 1949, 1973–88 Italy 1919–21 Moldova 1993– Mozambique 1993– The Netherlands 1933– Norway 1919–49 Paraguay 1993– Peru 1933–, Poland 1918–30, 1993–97 Portugal 1975– Spain 1977– Sweden 1909–48 Switzerland 1918– Turkey 1961, 1969– Uruguay 1918– Venezuela 1946–98 Yugoslavia 1992–

468–9 200–1 328–9 437 201–2 202 102 102 329, 369 203–4 204–5, 253 205–6, 265 206–7, 279 473 104 104 330

Supermajority rule. See qualified majority rule. Supplementary vote. In a single-member district by majority rule, each voter can order two or some other low number of candidates as preferred; seats are allocated

Glossary and Index 555 by majority first preferences, or by majority first-plus-second preferences (K ≥ M = 1 < V = 2). See: Sri Lanka president 1978– United Kingdom London local election

474 294

Threshold. In a system of proportional representation, minimum number or proportion of votes required to obtain representation. Also called minimum barrier. Threshold of exclusion. The maximum proportion of votes obtained by a candidacy not receiving a seat, or the minimum proportion of votes guaranteeing a seat. For a district, it may be estimated as T = 100% / M + 1, where M is the district magnitude (except for multimember districts with bloc vote by plurality rule, where M should be replaced with 1), or as the legal threshold, if higher. Threshold of inclusion. The minimum proportion of votes obtained by a candidacy receiving a seat. It is different for different electoral rules, but lower the higher the number of candidacies competing. Tier. See district, multiple tiers. Two rounds. See runoff. Two tiers. See multiple tiers. Unanimity rule. The winner is the alternative with either all votes, or selected by acclamation, passive acquiescence or assent to the result of some preliminary voting. Webster formula. Proportional representation quota invented by United States senator Daniel Webster (Salisbury, New Hampshire, USA, 1782–1852). See Sainte-Laguë formula. Winner-takes-all. See plurality rule, single-member district.