Referendum Democracy: Citizens, Elites and Deliberation in Referendum Campaigns

Referendum Democracy Citizens, Elites and Deliberation in Referendum Campaigns Edited by Matthew Mendelsohn and Andrew ...

Author: Matthew Mendelsohn | Andrew Parkin

44 downloads 1127 Views 2MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD

Referendum Democracy Citizens, Elites and Deliberation in Referendum Campaigns

Edited by Matthew Mendelsohn and Andrew Parkin

For Jackie

Contents

List of Figures and Tables

ix

Preface

xi

Acknowledgments Notes on the Contributors Introduction: Referendum Democracy Matthew Mendelsohn and Andrew Parkin

xiii xv

1

Part I The Rising Use of Referendums 1 Public Opinion and Support for Direct Democracy: a Grassroots Perspective Stephen C. Craig, Amie Kreppel, and James G. Kane

25

2 The Rise of Government-Initiated Referendums in Consolidated Democracies Laurence Morel

47

Part II The Impact of Referendums on LiberalDemocratic Institutions 3 Political Parties in Direct Democracy Ian Budge

67

4 Legislative Response to Direct Legislation Elisabeth R. Gerber and Simon Hug

88

5 Amending Constitutions through the Referendum Device Brian Galligan

109

6 Popular Control of Referendum Agendas: Implications for Democratic Outcomes and Minority Rights Shaun Bowler and Todd Donovan

125

7 The Medium is the Message: How Referendums Lead us to Understand Equality Avigail Eisenberg

147

viii

Contents

Part III The Impact of Referendums on Deliberation and Decision Making 8

9

10 11

Voters’ Decisions in the Nordic EU Referendums of 1994: the Importance of Party Cues Anders Todal Jenssen and Ola Listhaug

169

Are Voters to Blame? Voter Competence and Elite Maneuvers in Referendums Arthur Lupia and Richard Johnston

191

The News Media and Referendums Richard Jenkins and Matthew Mendelsohn

211

Constitutional Referendums and Democratic Deliberation Simone Chambers

231

Bibliography

256

Index

273

List of Figures and Tables Figures 3.1 Variations in party interventions in popular policy voting between various systems of direct democracy 3.2 Referendums and initiatives contrasted in terms of government control 3.3 Frequency of popular consultations and strength of parties: postwar democracies and selected territories 3.4 Distribution of direct democracy practices in terms of extensiveness of regulation and effects on political parties 8.1 Level of conceptualization and factual knowledge by strength and direction of party identification (pooled data from Finland, Sweden, and Norway) 9.1 Percentage of each group voting for the named initiative

70 73 82 85

178 198

Tables 1.1 Popular support for direct democracy in Florida, 1999 1.2 The cognitive mobilization and political disaffection hypothesis 1.3 The Gamson hypothesis 2.1 National referendums in 18 consolidated democracies (1900–98) 2.2 National referendums by country, initiator, and period (1940–98) 2.3 National government-initiated referendums by theme 2.4 Typology of governmental objectives during 22 of 27 referendums 3.1 Incidence of popular consultations and position of parties in 19 democracies and 51 selected territories: 1945–99 3.2 Existence of popular consultation related to strength of parties for 71 political systems 5.1 Constitutional referendums 1901 to 1988 6.1 Variation in elite control over procedures for direct democracy ix

34 38 40 50 52 53 54 81 82 117 128

x

List of Figures and Tables

6.2 Popular approval of anti-minority initiatives and referendums in Switzerland and the USA 6.3 Minority opinions towards direct versus representative democracy 8.1 Proportion of members of ad hoc groups who are also party members 8.2 Factual knowledge and assessments of personal knowledge on the EU issue % 8.3 Level of conceptualization before the Nordic EU referendums of 1994 8.4 Determinants of Yes vote: parties and cleavages 8.5 Change of party and change of EU position during the campaign 8.6 Influence of party and ad hoc movements on the vote: percentage saying group was ‘important’ or ‘very important’ 9.1 Figure 9.1 voters in ‘low info, no short cut’ group 10.1 Substantive differences between elections and referendums relevant to the media 10.2 Distribution of intervenors as major focus of news stories 10.3 Distribution of intervenors as minor focus of news stories 10.4 Focus of stories 10.5 News frames

134 139 173 176 177 183 187

189 199 215 223 224 227 228

Preface This volume is the product of a research project on referendums in liberal-democratic societies initiated and financed by the Donner Canadian Foundation. The volume brings together an eclectic mix of scholars from a range of countries with a variety of backgrounds. Although all have an interest in referendums, each comes at the question from the perspective of their particular area of the discipline of political studies. Taken together, the papers draw upon evidence from Western Europe, North America, and Australia and offer a consideration of referendums that incorporates insights from the fields of democratic theory, elections, public opinion research, the study of political institutions, and political communication. The result is a collection that advances the study of referendums by examining the various effects that the growing use of referendums is having on the contemporary practice of liberal democracy. The majority of the papers collected in this volume were originally presented at the conference on direct democracy, titled ‘Maturity or Malaise? The Growing Use of Referendums In Liberal-Democratic Societies,’ that was hosted by the Centre for the Study of Democracy at Queen’s University on May 14–16, 1999. The editors would like to thank the Donner Canadian Foundation for the financial support which made the conference, its associated activities, and the preparation of this volume possible. We would also like to thank George Perlin, the Centre’s Director, for his leadership in making the conference possible. We are grateful to Cameron McCormick for his assistance in the preparation of the manuscript for publication, and to the faculty and staff of the Centre for the Study of Democracy and the Department of Political Studies at Queen’s University who helped to make the conference a success, including Bob Burge, Fred Cutler, Evelyn McCaugherty, Shirley Fraser, Frances Shepherd, Celia Hatton, and Sarah Millar. All of the papers presented in this volume benefited greatly from the high quality of conversation that took place during the conference. In addition to those whose papers are collected in this volume, we would like to thank all of the conference participants, including those who acted as discussants, made presentations, or offered thoughtful commentary or intellectual support: Rick Arnold, Earl Bender, André Blais, Patrick Boyer, Catherine Conaghan, Pierre-F. Coté, Peter Cowan, Drude Dahlerup, xi

xii

Preface

James Davidson, Dorothy Ehrlich, Frank Graves, Andreas Gross, Janet Hiebert, Will Kymlicka, Lawrence LeDuc, Henry Milner, Will Robinson, Peter Russell, Peter Schrag, Hugh Segal, Leslie Seidle, Jürg Steiner, Alan Whitehorn, and Robin Wilson. This present volume is much improved because of their contributions. CBC-Newsworld broadcast a Forum on direct democracy that involved a number of the conference participants, and we would like to thank them for their involvement with the project and for helping to bring the issue to the attention of the Canadian public. We would also like to thank the authors themselves, not only for their contributions to the conference but also for their subsequent efforts in preparing their papers for publication. Matthew Mendelsohn acknowledges the Social Science and Humanities Research Council Post-doctoral Fellowship Grant # 75-930409, which first afforded him an opportunity to think about the issue of referendums. We also thank Jackie Hesketh and Eleanor Allen for their intellectual and moral support in the preparation of this volume. M ATTHEW M ENDELSOHN ANDREW PARKIN Kingston, March 2000

Acknowledgments Chapter 6 Popular Control of Referendum Agendas: Implications for Democratic Outcomes and Minority Rights The equal co-authors thank Liz Gerber, Simon Hug, and the editors for helpful comments on earlier drafts of this paper. We also thank participants in the Queen’s University Centre for the Study of Democracy Conference on the Growing Use of Referendums for their comments. Chapter 7 The Medium is the Message: How Referendums Lead Us to Understand Equality I am grateful to David Laycock, Stéphane Lévesque, Colin Macleod, and the editors for reading and commenting on this paper. I also thank David Hughes for his excellent research assistance. Chapter 8 Voters’ Decisions in the Nordic EU Referendums of 1994: the Importance of Party Cues Note: The data which we analyze in this paper are available from the Norwegian Social Science Data Services. We thank Kristen Ringdal for assistance with data analysis. Chapter 9 Are Voters to Blame? Voter Competence and Elite Maneuvers in Referendums The authors would like to thank the editors, James Druckman, Elisabeth Gerber, Simon Hug and participants in the international conference on ‘The Growing Use of Referendums in Liberal-Democratic Societies’ for helpful comments. Professor Lupia acknowledges the support of the Center for Advanced Study in the Behavioral Sciences. Chapter 10

The News Media and Referendums

The authors would like to thank Catherine Côté for conducting the content analysis and her suggestions and input throughout the preparation of the text. We would also like to thank Andrew Parkin for his insightful comments. Special thanks to the video archives in the Department of Political Science at l’Université de Montréal and Denis Monière for making the television news broadcasts available to us.

xiii

xiv

Acknowledgments

Chapter 11

Constitutional Referendums and Democratic Deliberation

For comments on earlier drafts the author would like to thank Rachel Barney, Amy Gutmann, Mark Lichbach, Jeffrey Kopstein, David Mapel, Melissa Williams, the editors, and the entire 99/00 Fellows Seminar at the University Center for Human Values at Princeton.

Notes on the Contributors

Shaun Bowler is Professor of Political Science at the University of California, Riverside. His major research interests concern comparative voting behavior. His work examines the interaction between institutional arrangements and voter choice in a variety of settings ranging from the Republic of Ireland to California’s direct democracy elections. He is co-editor of Citizens as Legislators: Direct Democracy in the United States (Ohio State University Press, 1998) and co-author of Demanding Choices: Opinion, Voting and Direct Democracy (Michigan University Press, 1998). Ian Budge is Professor of Political Science at the University of Essex. He is author of several books on political parties and British politics. His most recent book is The New Challenge of Direct Democracy (Polity, 1996). Simone Chambers is Associate Professor of Political Science at the University of Colorado. Professor Chambers is writing a book on the role of popular deliberation and democratic participation in constitutional change. Her primary areas of scholarship include political philosophy, ethics, and constitutional theory. She has published articles in the Journal of Political Philosophy, and Politics and Society. Her book, Reasonable Democracy: Jürgen Habermas and the Politics of Discourse, won the American Political Science Association, 1997, Best First Book in Political Theory Prize. In 2001 she will be at Princeton as a Laurance S. Rockefeller Fellow at the University Center for Human Values. Stephen C. Craig is Professor of Political Science at the University of Florida, as well as director of the university’s Graduate Program in Political Campaigning and of the Florida Institute for Research on Elections. He is author of The Malevolent Leaders: Popular Discontent in America (1993), editor of Broken Contract? Changing Relationships between Citizens and Their Government in the United States (1996), co-editor of After the Boom: The Politics of Generation X (1997), and has published numerous articles in professional journals. Dr. Craig has worked extensively with both academic and political surveys and continues to do polling and focus-group research for clients in Florida and elsewhere. xv

xvi


Todd Donovan is Associate Professor of Political Science at Western Washington University. He has published widely on ballot measures, voting behavior, minority representation, and third parties in the United States. He is co-editor of Citizens as Legislators: Direct Democracy in the United States (Ohio State University Press, 1998) and co-author of Demanding Choices: Opinion, Voting and Direct Democracy (Michigan University Press, 1998). Avigail Eisenberg is at the Department of Political Science of the University of Victoria, British Columbia. She is also on the board of directors of the British Columbia Civil Liberties Association. She teaches and writes in the areas of Canadian politics and democratic theory and is particularly interested in politics involving gender and cultural differences. She is the author of Reconstructing Political Pluralism (SUNY, 1995) and several articles on Canadian democratic politics and jurisprudence. Brian Galligan is Professor of Political Science at the University of Melbourne and Director of its Public Policy Centre. Before taking up his current position in 1995, he was a Professor of Political Science in the Research School of Social Sciences at the Australian National University and Director of its Federalism Research Centre. His recent publications include Citizens Without Rights: Aborigines and Australian Citizenship (Cambridge University Press, 1995), jointly authored with John Chesterman, and Defining Australian Citizenship: Selected Documents (Melbourne University Press, 1999), jointly edited with John Chesterman. He is also currently a joint editor of the Australian Journal of Political Science. Elisabeth R. Gerber is Associate Professor of Political Science at the University of California, San Diego. She received her Ph.D. from the University of Michigan. Her research is concerned with the policy consequences of electoral laws and other political institutions. She has published widely on the use of referendums in California and other states, and has completed two books on the subject, The Populist Paradox (Princeton University Press, 1999), and Stealing the Initiatives with A. Lupia, M. McCubbins, and D. L. Kiewiet (Prentice-Hall, 2000). She is currently involved in a major study of electoral laws in the American states and their effects on election outcomes and representation. Simon Hug is an Assistant Professor in the Department of Political Science at the University of Texas, Austin. His research interests include the formation of new political parties, the effect of institutions, and more particularly referendums and federalism, on decision-making and


xvii

conflict resolution. He has published on party formation, green politics, referendums and research methodology. He is the author of Altering Party Systems (Ann Arbor: University of Michigan Press, forthcoming) and co-author with Stefano Bartolini and Daniele Caramani of Political Parties and Party Systems. A Bibliographic Guide to the Literature on Parties and Party Systems in Europe Since 1945 (Sage, 1998). Richard Jenkins is currently a Social Science and Humanities Research Council of Canada postdoctoral fellow in the Political Studies Department at Queen’s University. He has published previously on election campaigns and the media. Anders Todal Jenssen is Associate Professor of Political Science at the Norweigan University of Science and Technology, Trondheim. His main fields of research are in political psychology and political behavior. He has published several books, most recently To Join or Not to Join, the major study of the EU referendums in the Nordic countries in 1994 (co-edited with Pertti Pesonen and Mikael Gilljam). Richard Johnston is Professor of Political Science at the University of British Columbia and co-author of Letting the People Decide: Dynamics of a Canadian Election and The Challenge of Direct Democracy: the 1992 Canadian Referendum. He was Principal Investigator of the 1988 and 1992–3 Canadian Election Studies, a consultant on the 1996 New Zealand Election Study, a member of the planning committee for the 1998 U.S. NES Pilot Study, and is co-Principal Investigator (with Kathleen Hall Jamieson) of the Annenberg School-University of Pennsylvania Year 2000 Study. His interests range from electoral history to media and informational dynamics in election and referendum campaigns. James G. Kane is editor of Florida Voter, a monthly political journal based on periodic statewide polls of registered voters. Kane also is an adjunct professor with the Graduate Program in Political Campaigning at the University of Florida, teaching courses in campaign management, lobbying, and survey research. Working with other faculty members at UF, he has co-authored several articles, book chapters, and conference papers dealing with various aspects of public opinion and voting behavior in the United States. Amie Kreppel is Assistant Professor of Political Science at the University of Florida. She is a specialist in comparative politics with an emphasis

xviii


on the institutions of the European Union, Italy, political parties, party systems and the legislature of Western Europe and the American Congress. Ola Listhaug is Professor of Political Science at the Norwegian University of Science and Technology, Trondheim. His main fields of research are in political behavior and comparative politics. He has published a large number of articles in various journals. Arthur Lupia is a Professor of Political Science at the University of California, San Diego. He has a Ph.D. from the California Institute of Technology. He studies the initiative process, the dynamics of voter decision making, and elections generally. His book, The Democratic Dilemma: Can Citizens Learn What They Need to Know? (co-authored with Mathew D. McCubbins and published by Cambridge University Press, 2000), combines insights from the disciplines of political science, economics, and cognitive science to explain how citizens gather and use information when making political decisions. In 1998, he was the first social scientist to receive the National Academy of Sciences’ Award for Initiatives in Research. Matthew Mendelsohn is Associate Professor in the Political Studies Department at Queen’s University. He has published on the media, public opinion, and public participation in decision-making in such journals as the British Journal of Political Science, the Canadian Journal of Political Science, the Journal of Politics, and Electoral Studies. He is currently writing about public opinion on constitutional issues in Quebec and the rest of Canada. He spent 1996–98 as Senior Advisor at the Privy Council Office, Government of Canada. Laurence Morel has published widely on referendum voting and why governments choose to use the referendum device in Europe. She teaches at Institute d’Études Politiques de Lille, France. Andrew Parkin is Assistant Director of the Centre for Research and Information on Canada. He previously held the position of Senior Research Fellow at the Centre for the Study of Democracy at Queen’s University. His teaching and publications have addressed issues in the fields of democratic theory, political participation, social movements, and Canadian politics.

Introduction: Referendum Democracy Matthew Mendelsohn and Andrew Parkin

Referendums – whether government-sponsored or citizen-initiated – are being used more frequently than ever before in liberal-democratic societies. In jurisdictions with a long history of direct democracy, notably many of the western US states, more questions are being placed on the ballot than in previous decades. In societies where the referendum has been used only sparingly in the past (such as the United Kingdom, Denmark, and Canada), it has now become more common, and pressure from both the public and political leaders for its increased use continues to grow. Still other societies, such as Italy, have joined the US and Switzerland in employing the initiative as a regular mechanism of decision making. It is no longer the case that Switzerland and a number of US states stand as the only exceptions to the traditional model of representative democratic government. The Swiss and American cases remain anomalous, yet other systems have integrated the referendum into their decision-making apparatus to the point that our understanding of what we mean by representative democracy needs to be re-examined. 1 We argue that the growing use of referendums is having a significant impact on the current practice of liberal democracy and that contemporary scholarship on liberal-democratic government has yet to come to terms with this development. We hope this collection, in a small way, changes some of the terms of discussion regarding the referendum. In particular, our intention is to open up a path whereby research on the referendum is no longer structured around a debate between the comparative advantages and disadvantages of two opposing systems of democracy: representative and direct.2 Such a debate has become less and less relevant as the referendum has become embedded within the heart of representative democratic systems. We propose to think of not two but one system, that of referendum democracy – a system in which 1

2

Matthew Mendelsohn and Andrew Parkin

the use, possible use, and threatened use of the referendum are fully integrated into the decision-making apparatus of representative, liberal democracy. Fortunately, recent years have seen the emergence of some vibrant scholarship which takes up this challenge by examining the ways in which the use of referendums is altering key aspects of liberaldemocratic politics. 3 The papers in this volume differ from much existing literature on referendums in that their approach is explicitly comparative and thematic. The standard approach has been to focus on the experience of a single country (Cronin, 1989; Kobach, 1993; Bogdanor, 1981) or even a single issue (Johnston et al., 1996; Jenssen et al., 1998). Even the important edited volumes on referendums in different liberal-democratic societies – the volumes edited by Butler and Ranney in 1978 and 1994, and by Gallagher and Uleri in 1996 – follow this pattern by offering chapters which, for the most part, give narrative accounts of the events in single countries. The papers in this volume break with this tendency and address particular themes – such as the impact of the use of referendums on parties, minority rights, or political discourse – so that they may examine and offer conclusions about the effect of the referendum in general. Even the quantitative analyses employed by several of the authors, which are necessarily limited to the consideration of data from only one or two jurisdictions, nonetheless are framed so as to bring their evidence to bear on an issue of general relevance. This thematic approach is useful in illustrating the major thesis of the editors: that the referendum is having a profound effect on liberaldemocratic government. As the referendum becomes a more regular component of decision making, it leaves few if any of the institutions, processes, and values of liberal, representative democracy untouched. Legislatures, courts, political parties, interest groups, and citizens each respond in different ways to the opportunities and challenges offered by the use or potential use of the referendum. The end result is a different kind of representative democracy than existed before. Another objective of this collection is to underline the disjunction between direct democracy in theory and in practice: in theory, referendums are designed to permit citizens to make decisions about public policy, but in practice referendums are used strategically by governments or groups seeking not to turn decision making over to citizens, but to persuade voters to ratify the position favored by the referendum’s sponsor. The power of elites to set the agenda and to define the choices facing citizens must therefore be given prominence in any discussion of the referendum (Setälä, 1999b). Certainly the public has a say – sometimes

Introduction: Referendum Democracy

3

the final say – in a campaign and it often rebuffs the referendum’s proponent. But one must remember that the decision to hold a referendum typically stems from a desire on the part of elites to achieve their preferred outcome, not from a normative commitment to greater public participation in decision making. In the remainder of this essay, we will highlight the contributions made by the papers collected in this volume while at the same time sketching out the essential features of what we have chosen to call ‘referendum democracy.’ The need for a new conceptual approach to the study of the use of referendums within representative democracies stems from the fact that the theories of neither representative nor direct democracy adequately describe the referendum’s contemporary role and significance. Theories of representative democracy do not account for how representative institutions have responded to the increasing likelihood that citizens may insist that legislation be ratified by a popular vote, or (in a number of countries) may launch legislative initiatives of their own. Theories of direct democracy tend to approach the issue of referendums from the point of view of voters exercising popular sovereignty, and so do not sufficiently account for how the referendum may be deployed strategically by elected representatives in order to accomplish a wide range of objectives, many of which – such as the goal of maintaining party unity in the face of a divisive issue, or of sapping one’s opponents’ limited financial resources – have little to do with a genuine desire to turn decision-making powers over to the public. It is the integration of the mechanisms of direct democracy into the everyday politics of representative democracy that requires investigation. Only then can we fully explain why the hopes of the advocates and the fears of the adversaries of the use of referendum – namely, that voters will replace elites at the center of democratic decision making – do not come to pass. This collection is organized around three themes. Following an introductory section which examines the growing use of the referendum device, the second part examines the consequences of the use of referendums for the key institutions, processes, and values of liberal democracy. The third part examines the question of the kind of decision making and deliberation encouraged by the referendum.

Referendum democracy Discussions of democratic government typically distinguish between two ideal types: direct democracy, in which policy decisions are made

4


by citizens, and representative democracy, in which citizens delegate decision making to elected legislators. In our discussion of the current use of the mechanisms of direct democracy, however, we prefer to make use of the term ‘referendum democracy’ because we find the usual dichotomy to be misleading. As has been pointed out elsewhere (Budge, this volume), direct democracy is not an alternative to representative democracy, since referendums have been successfully integrated into representative systems. But more importantly, we argue that the notion of direct democracy is misleading because it implies that voters have a direct impact on legislation, when in fact their actions are significantly mediated by elites, institutions, and electoral law. It is a serious oversimplification to say that, in cases where decisions are made by referendum, ‘the law of the land is made by citizens firsthand’ (Samuel C. Patterson, in Bowler, Donovan, and Tolbert, 1998, p. vii). Instead, it must be acknowledged that the referendum is a tool that competing elites (including elected representatives) can use to further their own agendas. For these reasons, we prefer to approach the study of referendums from a perspective that from the very start sees them as being intricately intertwined with the institutions and agents of representative democracy. And we believe as a consequence that the term ‘referendum democracy’ is a more apt description of the type of democracy that has resulted from the growing use of referendums and initiatives than is the label ‘direct democracy.’ There are currently many referendum democracies, which we define as liberal, representative democracies, in which the use or prospective use of referendums has become embedded within the political decision-making apparatus; there are no ‘direct democracies’ in which citizens enact their preferences at will. Furthermore, by using the term ‘referendum democracy’ rather than ‘direct democracy,’ it is our intention to direct attention towards the institutions and processes which structure referendums, and on the role of elites who more often that not determine which issues are put to a vote and under what circumstances. It is our contention that the integration of the referendum into systems of representative democracy has significantly affected the political dynamics of these systems as well as the institutions of liberal democracy. It has influenced the ways in which decisions are made, the possibilities for deliberation and coalition building, and even popular understandings of fundamental principles such as equality and democracy. This contention is strongly supported by the evidence offered in the papers which follow. In practice, referendum democracy takes on many forms. One must distinguish, first and foremost, between those societies that allow for


5

citizen-sponsored initiatives from those which provide only for government-sponsored referendums. This distinction is of fundamental importance because the two have such different theoretical underpinnings, objectives, and effects. By using the term ‘referendum democracy’ we by no means suggest that jurisdictions which widely use the initiative are the same type of system as those which make allowance for government-sponsored referendums. In fact, we wish to highlight exactly the opposite: the precise effect of the increased use of the referendum on liberal-democratic systems depends very much on who has the ability to determine when a vote will be held, and on what issues (see Smith, 1976). The extent to which the use of referendums affects the nature of democratic decision making will also depend very much on the type of issues put to a vote. In some cases, referendums are held only on constitutional amendments or on foundational issues such as the secession of a territory from a larger country or the accession of a country to a larger multinational union. In other cases, voters are called upon to pass judgment on divisive moral issues (such as abortion or divorce) or to approve legislation on routine matters of public policy. In determining the effect of referendums one must also consider the frequency with which referendums are held. Despite the many similarities between the two countries, referendum democracy takes a much different form in Canada, where only two national referendums have been held this century, to that in Australia, where constitutional referendums are common. And both these countries differ markedly from the situation in Switzerland and Italy where voters regularly are called upon to pronounce on a wide range of issues. The rules under which referendums are held must also be taken into account. Referendum democracy in the US states – where restrictions on campaign spending are prohibited and voter registration laws (and other aspects of American political culture) have substantially suppressed voter turnout among low-income voters – means something quite different from referendum democracy in countries where the state intervenes to maximize citizen participation and to ensure that as far as possible opposing sides operate on a level playing field. To a certain extent, then, each case is unique. Even the examples of Switzerland and the western US states, with their frequent initiatives, differ so significantly from each other that they can hardly be grouped together in any meaningful typology. Previous work, which examines these country-specific differences, is an important reference (Butler and Ranney, 1994; Gallagher and Uleri, 1996). Yet, at the same time, we suggest that the rise of the referendum, in all its forms, is changing the

6


nature of liberal-democratic government; by considering the impact of all referendums together, we hope to highlight the proposition that different societies integrate the use of the referendum in different ways with different effects on the political system. Although liberal-democratic societies increasingly share certain common features which can be attributed to the growing use of (and demand for) the referendum, this does not mean that the differences between these societies are becoming any less important. Rather, each society is having to come to terms with many of the same issues which stem from the fact that the referendum is being more widely used. To fail to compare how different societies respond to the same challenge – that of integrating the referendum into the decision-making apparatus of representative democracy – would be to miss an opportunity to highlight how liberal democracy has been evolving in the face of the requirement that the public be directly consulted more frequently and on a wider range of issues. There are many possible explanations for the rise of referendum democracy. The cause might be a confluence of unique events such as the regime changes sparked by the end of the Cold War, the signing of the Maastricht Treaty, and the expansion of the European Union (see Brady and Kaplan, 1994; White and Hill, 1996). It is more likely, however, that a more lasting shift in political attitudes has taken place, the effect of which has been to make citizens either more confident in their own ability to make key policy decisions or less confident in the ability of their elected representatives to do so. Supporting evidence has been found for each of these two contrasting explanations – one of which emphasizes the importance of cognitive mobilization, and the other which emphasizes political disaffection. Stephen Craig, Amie Kreppel and James G. Kane submit each of them to further scrutiny in their paper. Other, less well-known explanations also merit consideration. Craig and his colleagues raise the question of whether the catalyst driving the use of referendums is not the mood of the public per se, but the actions of ‘political entrepreneurs’ who are able to tap into the public’s diffuse support for direct democracy, mobilizing it for their own partisan advantage. This is an important argument to consider, for its implication is that the growing use of the referendum is not necessarily the product of the public’s demand for a more direct role in decision making but arises from the desire of elites to find new ways of advancing their own agendas. Through an examination of government-initiated referendums in consolidated democracies, Laurence Morel offers compelling arguments regarding the increased use of the referendum. First, the apparent


7

increase in the use of the referendum is largely explained by the growth in citizen-initiated referendums in Italy and obligatory constitutional referendums in Ireland, not a general across-the-board increase. Secondly, the growth in government-initiated referendums stems largely from the emergence of new issues, most notably European integration. And thirdly, the motivations for governments in using referendums tend to be instrumental rather than genuinely consultative. Governments’ use of the referendum is often due to the fact that they are interested in resolving tensions within their own parties, furthering their legislative agenda, or increasing their own power and electoral fortunes. As well, there are now de facto obligatory referendums – referendums which are not legally required but which in practice must be held because of a precedent that has previously been established. Once voters have been consulted on an issue they will expect to be similarly consulted should the same issue (or a related one) arise again. Governments which refuse to follow the precedent which they themselves have established risk undermining their own legitimacy. The implication is that the frequency of referendums is unlikely to diminish, since each type of issue that is put to a vote – be it electoral reform, constitutional change, or treaty ratification – establishes a precedent which makes subsequent referendums on similar questions more difficult to avoid.

The impact of referendums on liberal-democratic societies In this part, we will highlight the most prominent features of referendum democracy and outline the possible effects of the increased use of referendums on liberal-democratic societies. Political parties It is often argued – by both advocates and critics of direct democracy – that one effect of the use of referendums is to weaken political parties. But as Ian Budge demonstrates, this argument is overstated: parties play a key role in initiating many referendums and in organizing the debate in referendum campaigns – either by leading the yes and no sides or by taking positions which serve to guide voters. Ad hoc movements generally do not replace parties during referendum campaigns. As Jenssen and Listhaug show (this volume), even when ad hoc organizations play an important role in the campaign, these organizations are often led by and tend to be populated by current or past party members. And when one tries to remove parties from the referendum process, as occurred in Australia with the creation of the people’s convention on the future of

8


the monarchy in 1998, partisan machinations intervene (see Galligan, this volume). In short, referendums may curtail the ability of political parties to unilaterally set the legislative agenda, but they do not create a process which is somehow beyond the reach of parties or insulated from the effects of partisanship. Even the citizens’ initiative, which is cast as the biggest enemy of parties, nonetheless can be used to their advantage. By sponsoring or supporting initiatives, parties – including those outside of government – can succeed in recruiting new supporters and in shifting political debate to a ground more favorable to their own electoral success. Referendums can strengthen parties in other ways as well. For instance, they can be used by political leaders to manage conflict within their own parties or governing coalitions. 4 Instead of taking a stand on a key issue which divides party supporters, party leaders can adopt a more neutral stance and choose to have the issue decided directly by the voters. Similarly, the referendum allows parties to remove a divisive issue from the realm of electoral politics. In Quebec, for example, the Parti Québécois was able to win power in 1976 only after promising that it would not pursue its goal of secession from Canada without first gaining support for the proposal in a referendum. This strategy effectively decoupled the issue of secession from other election issues and allowed the party to become a stronger electoral force. Referendums also provide opportunities for small parties: by taking a position in a referendum campaign, they can potentially attract more publicity and gain credibility. This is especially likely to occur if the major parties are aligned on one side of the referendum question. When major parties unite, minor parties – as well as ad hoc organizations, social movements, and interest groups – come to play a prominent role in the campaign, and by doing so they may be able to carve out a more lasting place for themselves on the political landscape. Minor parties can also publicize their positions and enhance their support by sponsoring initiatives in those jurisdictions which allow for them – and by so doing they may transform the normal electoral agenda. One sees this most obviously in Italy, where the referendum abrogativo is used to overturn legislation. The procedure has been used effectively by the Radical Party to attract attention to its platform. Thus, referendums that may pose challenges for larger or established parties (to either maintain unity or ensure that division during the campaign does not permanently undermine unity) create opportunities for smaller or newer parties (to gain publicity and credibility, recruit members, and influence the political agenda).


9

Integration and constitutional change The widespread use of the referendum favors majoritarian policy outcomes by making ‘logrolling’ – the bundling of diverse measures together in a single package in order to broaden their basis of support – and concessions to minority concerns more difficult. The normal legislative process often involves deal-making between legislators, parties, or within cabinet in order to provide various interests with favorable outcomes, even if individual items could not secure majority support on their own merit. Moreover, the drawn out process of drafting and approving legislation also permits amendments that offer concessions to concerned minority opinion. During the process of legislative debate, committee meetings, and public hearings, proposed legislation is almost always amended to incorporate the views of those other than the sponsoring majority. This is often done in ways that do not jeopardize the majority’s project, but that likewise impose less harm on other groups than might have occurred had the initial version of the bill been adopted. When legislation is passed through referendums, the iterative and conciliatory nature of these normal legislative processes may be lost. The integrative qualities of many countries’ institutional arrangements, particularly those that use consociational processes and proportionaltype electoral systems, have been important in maintaining successful democratic government. These integrative qualities appear to be undermined by the widespread use of referendums, since referendums generally facilitate majoritarian politics. In some referendum democracies, however, institutional arrangements have been developed in order to offset this. The best known is the Swiss experience with the process of Vernehmlassung – a process through which elaborate consultations are undertaken before legislation is passed so as to integrate the concerns of as many communities as possible in order to forestall potential referendum challenges by making the measure ‘referendum-proof’ (Linder, 1994). In this way, the integrative qualities of Swiss liberal democracy are preserved. It may be argued that the more parties opt to finesse divisive issues by shifting the responsibility of decision making to the voters, the less adept they will become at reconciling opposing viewpoints and negotiating compromises among party factions or coalition partners. The need to forge consensus is less pressing if party leaders can appeal over the heads of dissident party or coalition members to the electorate at large. This may have important implications for how conflicts are managed in liberal-democratic societies. When party leaders broker deals among rival factions, compromises are made. By contrast, when the

10


matter is put to a referendum, voters typically have the option of only saying yes or no to an unamendable proposal. Parties’ long-term ability to build bridges between factions – factions which may correspond to important regional or cultural divisions within the country – may be strained. Brian Galligan underlines that in societies which use governmentsponsored referendums for the purposes of constitutional change, there may be an inertia towards the constitutional status quo. Yet he does not see this as evidence that the public is too short-sighted to make needed constitutional changes and does not see the referendum device itself as detrimental to the process of constitutional renewal. He points out that constitutional amendment should be difficult precisely because one is changing a higher law which will serve to subsequently structure and regulate future political conflict. In the case of Australia, the very low success rate of referendums provides strong evidence, not that voters are unable to recognize their long-term interests, but that political leaders failed to learn that one of their cherished projects – namely, the centralization of powers to the national government – was not supported by the public. It is worth noting that ‘megaconstitutional’ change – extensive changes to a country’s constitution bundled together simultaneously – can be especially difficult to pass by referendum because voters are rarely privy to or supportive of the interparty elite bargains that underpin these kinds of constitutional arrangements. As the Canadian government discovered in 1992, the use of the referendum as a device to ratify the outcomes of such elite-level bargaining sessions can be highly problematic. And as both the Canadian and Australian examples highlight, ordinary partisan politics invariably intervene in the referendum process. Using the referendum as a means of ratification provides greater legitimacy for constitutional change than alternative procedures and at the same time allows a people to symbolically affirm its status as a sovereign political community. It is, however, clear that there is a need to agree on the basic constitutional rules of the game before referendums can be easily used for constitutional amendment. Failing to agree beforehand on such questions as who has a veto and what constitutes a sufficient majority can lead to the process of constitutional change through a referendum being itself contested. In these circumstances, referendum democracy becomes very risky. Again, the effect of the referendum is heavily influenced by the referendum’s institutional design. For example, when the public can both draft and ratify constitutional amendments, rather than simply ratify amendments drafted


11

by governments, constitutional change may be facilitated, something which has occurred in many American states. In such circumstances, the risk is that constitutional change becomes too easy for groups to engineer. Policy outcomes, legislatures, and interest groups Policy outcomes under conditions of referendum democracy may be biased toward the status quo (see Immergut, 1992). This is the case both with initiatives and with government-sponsored constitutional change. There is a good deal of evidence that uncertain voters are risk averse and tend to vote ‘no’ to proposed changes. Furthermore, once a measure has been passed through a referendum it is often politically unacceptable (or legally prohibited) to change it without another referendum. This makes it more difficult to bring about evolutionary change by means of periodic legislative amendments. Even the ordinary process of making small amendments to legislation to reverse unintended consequences or adapt to new circumstances is prohibited in many jurisdictions that use the initiative: once legislation is passed by the public, it can only be changed by the public. Indeed, one of the qualities inherent in referendum democracy in California is that legislation passed through referendums often imposes very strict rules on future decisions. As the number of these referendums passed grows, restrictions multiply and legislatures’ ability to react to new circumstances becomes seriously constrained (Schrag, 1999). Referendums are increasingly used to regulate the functioning of political institutions themselves. Examples include initiatives used to change the way representatives are selected in Italy or to impose termlimits on representatives in the United States. One can also note that the attempt to confer greater powers on elected officials was rejected in Australia by a public which refused to confer on parliament the right to select the head of state. These somewhat different examples have in common the self-evident feature that they constrain the governing elite; but more importantly, they highlight the fact that referendum democracy forces elected officials to abandon their virtual monopoly on the passage of laws that govern their own behavior and tenure (see also Tolbert and Hero, 1998). As power is distributed more widely through the political system, the role of interest groups changes. Few would deny the important and constructive role of interest groups in shaping public policy through normal legislative processes. In referendum democracies, however, interest groups become both more and less powerful. Privileged access

12


to decision-makers is no longer sufficient to guarantee power because other groups may choose to use the referendum to pass legislation regardless of the preferences of the legislature. Yet at the same time, interest groups have an added weapon, the initiative, with which to prod reluctant legislatures (Gerber, 1999). In Gerber’s formulation, policy outputs represent a spatial game between the sponsor of the initiative (often interest groups), the legislature, and the ‘median voter.’ Interest groups increasingly must change the nature of their argumentation and efforts so that they are appropriate for victory on the public stage during a referendum contest, rather than appropriate for private lobbying and persuasion. Importantly, some groups may prove to be more adept at this than others. Elisabeth Gerber and Simon Hug summarize a wide body of literature on the question of whether the referendum advantages voters as a whole or narrow, well-financed groups. They conclude that where policies differ in those jurisdictions with direct legislation as compared to those without, the difference is in the direction of the ‘median voter,’ wherever that voter lies on the ideological spectrum (liberal or conservative). The evidence is quite strong that policies are more likely to reflect standing public opinion in direct legislation jurisdictions than in places where there is no provision for direct legislation. This confirms the initial suspicion of Butler and Ranney that neither ‘conservative’ nor ‘liberal’ causes are systematically advantaged by the use of the referendum (1978, p. 224), and suggests that narrow interest groups that do not have the backing of majority opinion are not strongly privileged in a system of referendum democracy. Gerber and Hug also argue that even if questions do not get to the ballot, policies are more likely to reflect majority opinion because legislatures anticipate challenges from outside groups and enact legislation that is responsive to those concerns. The presence of the initiative thus has an ‘indirect effect’ and forces legislatures to change their behavior. These ‘indirect effects’ are largest for those groups with resources or who can mobilize majority opinion. Institutions and regulatory regimes will of course mediate these conclusions: a country with strict spending limits and free access to media may produce different outputs than one with no limits and no unpaid access to the media. Yet the observation that policies move towards the median voter under situations of referendum democracy remains an important one. However, in situations of low voter turnout, such as in California, this means that the referendum may amplify the opinion of those most likely to actually vote: white, middle-class, suburban voters (Schrag, 1999).


13

Minority rights Shaun Bowler and Todd Donovan address one of the most important objections to the use of referendums, namely that they provide a means through which the majority may tyrannize minorities. Some express the concern that the rising use of referendums might led to the curtailment or reversal of legislative gains won by minority groups. Bowler and Donovan question this. Their study of the outcomes of Swiss and US state initiatives shows that voters are not as intolerant as the critics of direct democracy often claim them to be; at the same time, their US survey data show that minorities are not uniformly opposed to the initiative as a means of deciding policy. They also make two important observations: first, it is not yet known whether legislatures in jurisdictions without the initiative have passed more or less anti-minority measures than those which allow for the initiative; and second, that cases where anti-minority measures have been passed by referendums must be weighed against cases where referendums have produced gains for minority groups. There may, of course, be more subtle ways for the majority to target minorities than to use the initiative to roll back anti-discrimination legislation or place restrictions on immigration. Referendums that mandate lower taxes or balanced budgets, for instance, may de facto harm minority groups if the end result is the curtailment of government services that are disproportionately used by these groups (Schrag, 1999). But again, the key question is whether the outcomes of referendum votes on these matters are different from the outcomes of more traditional policy making. Over the past 20 years, conservative administrations in countries with and without the referendum have cut back on services that are disproportionately relied on by minorities. Seen in this light, the important question to ask is not ‘do referendums allow for the tyranny of the majority or the restriction of services used by minorities?’ but rather ‘under what conditions do referendums allow for outcomes which are more majoritarian and more restrictive than the outcomes which would otherwise result from legislative action?’ The effect on citizens and their conceptions of liberal democracy Avigail Eisenberg examines the conditions under which referendum outcomes are likely to be worse for minorities by considering the effect of the referendum on political discourse and our conceptions of equality. Eisenberg’s unique contribution to the debate about the impact of referendums on minorities is to establish that the referendum is not

14


a ‘neutral’ decision-making tool. The very attractiveness of the referendum, she argues, is that it evokes a particular conception of equality – namely one which sanctions a view of the community as composed of undifferentiated citizens. Its use serves to erase distinctions among groups of citizens by positioning them as members of a single voting community within which each voter is equal to one another in terms of their political power, and from which emerges a ‘general will’ in the sense that the majority view counts as the will of the whole community. This has implications for minority rights. She suggests, for example, that a referendum campaign can more easily sustain a claim made by minorities to the same rights as those of other citizens than it can a claim to different status (e.g. to aboriginal self-government, affirmative action, recognition as a national community, or special status). Because the logic of the referendum itself is one which denies differences between citizens, arguments in favor of differentiated equality must be advanced through a process which implicitly denies their validity. Thus the long-term effect of referendum democracy may be to provide minority claims for equal treatment with a rhetorical advantage, while hobbling claims for different treatment. The implication is that referendum democracy has the effect of shaping the terms of liberal-democratic political discourse by favoring certain kinds of arguments about the meaning of equality over others. How referendums change the nature of other aspects of political discourse and alter our understanding of important democratic concepts deserves increased attention. Whether referendums are detrimental to claims for differentiated equality may further be influenced by considerations such as whether voters are only allowed to vote on a ‘take it or leave it’ final agreement or whether instead they are able to use successive referendums to shape an evolving text, and whether the referendum is preceded and accompanied by mechanisms which draw citizens together in a deliberative process. Consideration of these alternatives serves to underline the importance of the institutional design of the referendum, and ultimately to shine the spotlight back on the role played by elites. As Bowler and Donovan stress, elites do not drop out of the picture because a matter is being put to a popular vote. Rather, they structure the referendum process and provide the cues which influence voting behavior. Measures of all kinds that restrict minority rights have a very difficult time passing when they are opposed by a united front of elites. Thus, whether or not referendums produce anti-minority outcomes depends at least as much on the maneuvering of elites as on the majoritarian sentiments of voters.


15

Many have held out the hope that referendum democracy would politicize citizens and make them more engaged political actors (Key and Crouch, 1939). Thus far, the jury is still out on the question. Although some find that the regular use of citizen initiatives produces little effect (Citrin, 1996, pp. 286–7), others find that those in states which use direct democracy are more efficacious and knowledgeable about government than those in states without (Bowler and Donovan, forthcoming). It is sometimes suggested that the use of the referendum in Switzerland fosters a variety of civic virtues (Frey and Stutzer, 2000); yet in other jurisdictions the occasional use of government-sponsored referendums on issues of national importance produces little impact on efficacy and politicization more generally (Mendelsohn and Cutler, 2000). Overall, we suspect that, contrary to the expectations of its advocates, the use of the referendum has not accomplished the task of reducing the ‘democratic deficit.’ The irony, according to Craig and his colleagues, is that as politics becomes more ‘democratic’ in the US, people become more disaffected. The dynamics of referendum democracy as practiced in the US in part creates a permanently disenchanted electorate. This suggests that if one has a normative commitment to increased citizen participation in democratic politics, referendums are not a stand-alone solution; they need to be made part of a much broader process of citizen politicization and more open, porous, decision making.

Deliberation and voting in referendums Some have argued that voters are incapable of making wise choices in referendum campaigns (Mueller, 1989). Whether this is true depends on how we define ‘wise choices,’ and, as we shall see, the papers collected here approach this question in different ways. There is general agreement about how voters make decisions during referendum campaigns; it is the implication of this process of decision making that is contested. Those who believe that voters make reasonable decisions are interested in whether or not their votes ‘make sense’ based on the voters’ interests and values; those who do not suggest that voters must engage in deliberation which is public and dialogic so that they can gain a better appreciation of their opponents’ views. Anders Todal Jenssen and Ola Listhaug examine the role of parties in structuring the vote during the 1994 referendums in the Nordic countries on membership of the EU. They conclude that parties played an important role in providing information to voters, and that the referendum campaigns were a continuation of normal electoral politics.

16


Beyond the point that partisanship acts as an important cue for voters, Jenssen and Listhaug advance several general propositions. Their data suggest that when parties are polarized on the referendum question (as occurred in Norway), there will be more information available to voters and more cue-taking by voters. In these cases, voting behavior will be very similar to what we would expect in an election. They also find that when mainstream parties are united on a referendum question, this will lead to the formation of ad hoc movements that will attract a great deal of attention. In this second case, partisanship may be less important than during elections. Partisanship will also be less important if parties are internally divided. Their work underlines that traditional factors, such as ideology (Pierce et al., 1983), social status (Hahn and Kamieniecki, 1987), core values (Siune, Svensson, and Tonsgaard, 1994), and reference groups all play a role in voters’ reasoning in referendum campaigns (Svensson, 1984). Arthur Lupia and Richard Johnston make an impressive case that most voters are capable of ‘getting it right’ using a variety of cues. They suggest that ‘a voter’s choice is competent if it is the same choice that she would make given the most accurate available information about its consequence. Would she make the same decision if fully informed about the consequences of her actions? If yes, then her choice is competent’ (p. 194). They contend that the argument disparaging referendums because voters lack information is flawed because it conflates information and competence. They reject the assertion that a voter is only competent when they know where the parties stand on, say, 100 different issues and also know their own positions on this same breadth of subjects. Few voters meet this standard, but by taking cues from prominent intervenors most voters vote in a way consistent with those who do meet this standard; that is, they vote competently, consistent with their underlying values and interests. Simone Chambers sets a different standard: knowing where parties stand on 100 issues does not guarantee competent decision making unless one has arrived at one’s own positions through deliberation and public discussion. Instead of focusing on voting, Chambers focuses on talking. Processes that enable voters to talk in public with others – i.e. to deliberate – can moderate self-interest, mediate difference, and lead to more reasonable policy. This deliberation should be dialogic and discursive; that is, one should talk to others in order to air and test one’s own views rather than engage in strategic conversation. The dialogic model can be integrative because through deliberation citizens build trust and understanding. She argues that referendums, particularly those


17

dealing with constitutional issues, can undermine deliberation because ‘they introduce an extreme form of majoritarianism which is inappropriate for deep constitutional questions that divide a nation [and] second, they often present the voter with the image of irreversibility which serves as a disincentive to adopt the attitudes of impartiality and respect’ (p. 231). The question Chambers raises, then, is not so much whether voters can make the ‘right’ decision on a given question, but rather what type of decision-making processes are likely to be facilitated by the use of referendums. Richard Jenkins and Matthew Mendelsohn examine the ways in which the news media covered the 1995 referendum on Quebec sovereignty, and advance two general propositions. First, because of its continued focus on traditional parties and their leaders, the news media does not always provide a wide range of cues about the positions of intervenors during government-sponsored referendums. This facilitates votes based on partisanship. Second, they conclude that media coverage of referendum campaigns looks very much like media coverage of elections: it provides narration of the day’s activities, interpreted through a strategic frame, and provides little that would allow for decision making based on anything but minimal cues concerning some intervenors’ positions. Thus they argue that the way in which the media go about covering referendums does not facilitate deliberation or collective will formation. Referendum democracy appears to be a democracy that shuns the deliberation inherent in the legislative process in favor of a majoritarian process of the aggregation of individual opinions. Lupia and Johnston argue persuasively that referendum outcomes are usually pretty good at measuring pre-existing public opinion. Chambers makes a good case that referendums fail to facilitate a discursive public conversation – a failure which, according to Jenkins and Mendelsohn, is abetted by the kind of strategic conversations encouraged by the news media. As liberal-democratic societies use more referendums, it appears that accurate and efficient measurement of standing opinion becomes the most favored value.

A comparative research agenda on referendum democracy Our portrait of referendum democracy in this introductory chapter has focused on general patterns. Of course, the political culture, regulations, and institutions particular to each country produce substantial variation. Our overview has not accounted for the important differences in law

18


and conventions in some countries that attenuate the general version of referendum democracy sketched above. For example, we mentioned only briefly how in some cases the majoritarian nature of referendum democracy is tempered by appropriate institutional arrangements. By highlighting the general tendency of referendum democracy towards majoritarianism, it is not our intention to suggest that these types of outcomes are inevitable, but rather that they are likely, unless the tendency is challenged by counteractive mechanisms. The referendum, elites, and liberal-democratic systems The widespread adoption of the referendum device changes the way government works, creates new rules and institutions, and, in the process, restructures political conflict. Referendum democracy facilitates certain kinds of conflict and certain kinds of outcomes, though who wins and what gets organized into politics in each case is partially contingent on the particular regime rules in force. To identify clear winners and losers is difficult; while those outside the realm of established political elites stand to be advantaged, it is often only those who can credibly launch an initiative campaign (or make a compelling case for a referendum) who prevail in practice. But regardless of who wins and loses, in all cases the political terrain is affected. In Canada, for example, the 1992 referendum permanently weakened the ability of elites to broker constitutional deals, producing a different kind of democracy than existed previously in the country. In Italy, the institutionalization of the initiative in 1970 permanently threatened established parties’ control of government. Referendum democracy may allow power to be more widely distributed through the system, particularly in regard to the political agenda. A greater number of entry points to the political system are created, the pool of political actors is enlarged, and the monopolization of power is diminished. In Italy, the referendum abrogativo has played an important role in agenda-setting on such moral issues as abortion and issues relating to electoral reform (Uleri, p. 109). It has also opened up the political process to groups emerging from civil society rather than from the established party system, most notably the Radical Party and a variety of Catholic groups on the right. In Switzerland, while groups such as pacifists and environmentalists failed to win passage of referendums abolishing the army or prohibiting research on genetically modified organisms, the strength of their vote alerted the government to important opinions in society, increasing the likelihood that subsequent decisions may reflect some of these groups’ concerns.


19

Even though the pool of actors is enlarged, the power of elites remains formidable. In Italy, governments have dissolved parliament more than once in order to avoid a referendum, and on other occasions parliament has used its discretion to pass laws that responded to the spirit of the referendum abrogativo, but did not reflect exactly what proponents of the referendum would have liked. The use or threatened use of the referendum, particularly in the US states, has become integrated into the strategic thinking of political actors. For example, once an initiative has qualified for the ballot, established party elites may choose to propose a similar but not quite so radical departure from the status quo with the hope that voters will approve their less extreme measure. Alternatively, groups threatened by the original measure may attempt to qualify counter-initiatives. This is done with the hope that voters will become so confused by the conflicting messages and TV spots that they will vote ‘no’ to all of them (Banducci, 1998, pp. 119–28). A focus on elites rather than voters therefore is crucial to a proper understanding of referendum democracy. The referendum is one device among many used by parties, politicians, and interest groups to shape the political agenda, influence political outcomes, resolve internal disputes, and alter the political dynamics affecting their interests. Those with resources – whether these are established parties looking to insulate themselves from divisive internal conflict or powerful civil society forces looking to place issues on the agenda – will inevitably focus their attention on the most important sites of decision making, whether these are legislatures or referendums. Variations of referendum democracy, innovation and deliberation The contested nature of referendums is in part about the contested nature of democracy itself: is democracy about the promotion of selfinterest and the aggregation of individual opinion or is it a process of deliberative and collective will formation? Discussions of referendums often tap into both streams of thought by talking simultaneously about the public having their say at the expense of elites (where the assumption is that there is a standing public opinion waiting to be measured that has for too long been thwarted by elite machinations), and also about societies coming together to debate their collective future. Those who are drawn to the more deliberative model of democracy look to fold referendums into a variety of more robust, discursive democratic institutions (Barber, 1984). The debate over the introduction of

20


referendums is therefore embedded in the evolving debate in liberal democracies between advocates of an efficient, rationalized democratic space, and a deliberative one founded on non-commercialized public spheres. Chambers argues that ‘Referendums, without supplementary institutions and forums that promote deliberation, invest so much in numbers rather than arguments that it is hard for the losers not to read the outcome as “might makes right”’ (p. 243). The use of referendums without such supplementary institutions produces the referendum democracy we have described above. We argue that it is possible nonetheless to use referendums as part of a creative process that can promote deliberation. Tully (1994) has suggested that constitutional agreements that need to be submitted to a referendum should be treated as rolling drafts rather than final accords, so that amendments could be accommodated and suggestions solicited. Under such circumstances, the chance of passage during a referendum would be significantly augmented. Mendelsohn (2000) labels such a process ‘public brokerage.’ Such an iterative model, with its focus on an ongoing process of consultation rather than a once and for all ratification, is contrasted with our description of referendum democracy as currently practised throughout most liberal-democratic societies. This underlines the importance of institutional design when structuring referendums in democratic politics. There are many further variations to the model of referendum democracy sketched above. For example, the ability of one-third of the representatives in the Danish parliament to ask for a referendum provides a protection for minority opinion and an incentive to engage in cooperative bargaining before legislation is passed. Both Svensson (1996) and Bogdanor (1994) suggest that this has encouraged parties to seek consensus. Likewise, the possible requirement that a measure be passed by concurrent majorities in different constituent communities also establishes a process which goes beyond simple majority rule. The use of the indirect as opposed to the direct initiative in some US states has allowed legislatures to be more fully involved in the process, permitting them to offer amendments to propositions before they become law, promoting more dialogue between groups and institutions. This is not unlike the way the Italian system responds to the referendum abrogativo. Unlike the contention of Bogdanor (1994) that Italy’s form of referendum is no more than a veto on government policy, it de facto allows outsiders to initiate policy by placing issues on the agenda and encouraging parliament to respond in ways consistent with the spirit of the initiative. The dialogue between the different actors – the public, the groups sponsoring


21

the referendums, the established political parties, and the courts which determine whether or not a referendum can be put to a popular vote – has become a feature of Italian democracy that has changed the distribution of power and political resources. The danger with both the indirect initiative and the referendum abrogativo, however, is that in providing a forum for ongoing debate and bargaining – a forum that may moderate some of the referendum’s most vulgar extremes – referendums may lose a great deal of their appeal: providing a clear decision rule which produces decisive outcomes that established elites are forced to obey. The use of a people’s convention to set the parameters for debate before the referendum is itself called can promote the integration and deliberation that referendums sometimes lack. The funding of a variety of civil society organizations leading up to a vote, as occurred in the Danish referendums on the European Union in 1992 and 1993, helped promote deeper deliberation. This avoided some of the pitfalls of overly structured contests where umbrella committees are set up and are the sole organizations responsible for the two campaigns – a format which can constrain a variety of actors who may favor or oppose a proposal for reasons entirely different from those of the established leaders of the umbrella committees. More innovative uses of the referendum, such as those which allow for multiple options to be placed on the ballot or a second round of voting, both of which occurred during New Zealand’s referendums on electoral reform in 1992–3, can likewise produce a more constructive process. Providing institutional incentives so that proponents of initiatives are encouraged to form broad-based coalitions when establishing committees to draft initiative proposals, so that some level of negotiation occurs prior to the question being placed on the ballot, can both attenuate the majoritarian qualities of referendums and promote more citizen rather than professional involvement in the process. Future research The papers in this collection highlight the diversity of scholarship being conducted on the referendum and lay out a comparative research agenda built around two related themes. First, how does the rising use of the referendum affect parties, constitutions, legislatures, rights, courts, political discourse, policy outputs, and citizens themselves, and how do political actors adapt themselves to the presence of the referendum device? And second, how do different contexts, institutions, communication networks, and statutory regimes affect citizens’ ability to deliberate and make political choices? This collection begins to sketch out a way to answer these questions.

22


Notes 1 One must begin with a definitional note. We will use the term ‘referendum’ in its most generic sense to refer to all forms of popular consultation. We will use more specific terms, such as ‘citizen-sponsored initiative’ (or just ‘initiative’) and ‘government-sponsored referendum’ when we refer to particular variants of the referendum device. 2 Even those who recognize that the contrast between representative and direct democracy is a false one still see the utility in contrasting these ‘two systems’ (Bobbio, 1987, p. 53). 3 It is striking that in Butler and Ranney’s classic study (1978), they promise a discussion of the impact of referendums on liberal-democratic institutions (p. 1), but say hardly a word about this question. 4 For instance, referendums in Italy on divorce and Sweden on nuclear power emerged as responses to tensions within the governing coalitions (Bjørklund, 1982).

Part I The Rising Use of Referendums

1 Public Opinion and Support for Direct Democracy: a Grassroots Perspective Stephen C. Craig, Amie Kreppel, and James G. Kane

The use of referendums and initiatives has increased sharply over the past quarter century, both in the United States (Cronin, 1989; Waters, 1999) and in other democratic countries around the world (Butler and Ranney, 1994; Dalton, 1996; Uleri, 1996; Setälä, 1999a). However, though it is clear that this phenomenon has occurred, there is little understanding or agreement as to the reasons why. A review of the literature suggests numerous possible explanations, most of which can be grouped into one of three broad and relatively distinct (but not necessarily mutually exclusive) categories. One perspective emphasizes the critical importance of cognitive mobilization (Dalton, 1996; Inglehart, 1977, 1990), a dynamic process that allegedly has produced heightened popular demand for access to governmental decision making; that is, as a consequence of rising levels of political skills and resources among contemporary electorates, there is an emerging belief that ‘the people’ can – indeed, should (Craig, 1993) – be trusted to make complex political decisions and otherwise to participate actively in the democratic governing of society. A second explanation focuses on growing cynicism or disaffection toward government, and the loss of public confidence in traditional democratic structures. In this view, initiatives and referendums are utilized more frequently than in the past as a means of protecting ‘the people’ (Us) from ‘the politicians’ (Them),1 thereby ensuring that citizens’ rights are safeguarded and their wishes attended to by political leaders (Sears and Citrin, 1982; Schmidt, 1989; Craig, 1993). 2 Finally, there are those who argue that the trends described here do not really reflect changes at the grass roots; rather, direct democracy at the dawn of the twenty-first century is said 25

26

Stephen C. Craig, Amie Kreppel and James G. Kane

to represent, for the most part, a kind of faux populism in which special interest groups and political entrepreneurs have discovered a new means of political access and, accordingly, have developed the tools and strategies needed to bend the use of popular political action to their own narrow purposes (Smith, 1998). All three of these explanations possess some degree of plausibility, and examples certainly can be found that would seem to lend empirical support to each. Yet the evidence is mixed and, as a result, the underlying questions (how do citizens feel about the expanding opportunities for people like themselves to engage in political decision-making, and why do they feel that way?) remain largely unanswered. In the analysis presented below, public opinion data from the United States are used to test, more or less directly, the cognitive mobilization, disaffection, and faux populism hypotheses; in doing so we hope to provide insights that will assist other researchers as they explore the popular roots of contemporary direct democracy.

Overview Trends By all accounts, the instruments of direct democracy are employed more frequently today than in the past. In the USA’s federal system, 49 of the 50 states have a top-down (or legislative) referendum, 3 but only 27 permit some form of statewide initiative or popular (bottom-up) referendum – and 23 of the 27 adopted the practice prior to 1960. 4 While this number could rise as politicians elsewhere call for direct democracy to be implemented in their own states (Waters, 1999, p. 128), for now our conclusion must be that the formal (legal) opportunities for direct citizen input have not changed much over the past forty years. On the other hand, more statewide initiatives were actually placed on the ballot in the 1990s than in any previous decade: a total of 323 from 1991 to 1998, compared with the previous high of 290 during the height of the progressive movement in 1911–20 (Waters, 1999, p. 128).5 Even as some states try to make it more difficult for citizeninitiated ballot measures to qualify,6 political considerations (as well as the ‘professionalization of the initiative industry’ (Citrin, 1996, p. 282)) probably preclude anything close to a full-fledged retreat from current practice. The United States does not, of course, permit citizens to participate directly in the making of national policy. In fact, bottom-up mechanisms can be found at the national level in only a few countries – most

Public Opinion and Support for Direct Democracy

27

notably Switzerland, where both the initiative and referendum have existed for purposes of constitutional revision since 1848 (following similar procedures introduced in the cantons as far back as the thirteenth century (Kobach, 1994)), and Italy, where citizens can under certain circumstances utilize the ‘abrogative’ referendum to reject laws already passed by the Parliament (Bogdanor, 1994). Legislative referendums are much more common, though the rules governing their use (for example, national or subnational, binding or advisory, constitutional or statutory) vary from place to place. Setting aside these variations, the evidence once again seems to suggest that the frequency of citizen lawmaking (including the bottom-up variety, where it is available) has increased in recent years (Butler and Ranney, 1994; Setälä, 1999a). Can we assume that this increase is itself a reflection of the public mood? Put another way, is the apparent rise of direct democracy truly a democratic phenomenon? And if it is, does heightened popular demand reflect a positive response to the perceived competence and judgment of electorates, or a negative assessment of the fitness or integrity of elected leaders? Explanations In US politics, public support for direct democracy is often seen as an extension of that nation’s ‘populist’ heritage. As described by Daniel A. Smith (Smith, 1998, p. 45), ‘populism entails a grass-roots, bottom-up form of protest and participation by the masses.’ It is a process that pits ‘Us against Them’ indicating ‘a mass outcry of a “common people” aimed at an established elite, their norms, and their practices.’ The underlying cause of any such uprising is more likely to be negative (reflecting frustration with unresponsive leaders) than positive (involving a desire to achieve some policy goal or procedural reform independent of the belief that the public’s will is being thwarted).7 Yet either way, it presumably involves high levels of support among citizens for both the general principles and the specific instruments of direct democracy. The faux populism hypothesis, in contrast, holds that the increased use of initiatives in recent years has relatively little to do with the political mobilization of ordinary citizens. Speaking of the United States, Smith (1998, p. 48), contends that what we have witnessed is the rise of faux populism, which depends first of all on the existence of ‘an unequivocal (though usually latent) public mood’ on some matter, that is, ‘the notion that a rather large number of people out in the country are thinking along certain common lines.’8 This public mood, Smith says, ‘is real, but is poorly articulated; it is fragmented and ill-defined.’

28


Consequently, for a faux populist movement to be successful, there also must be ‘a populist entrepreneur’ with ‘sufficient charisma and organizational resources . . . to channel and fashion the public mood into a coherent, popular message’; further, the entrepreneur must be able to get that message on the ballot via the initiative process, and to persuade voters that ‘it is the solution to the widely perceived public problem’ (Smith, 1998, p. 48). Missing from faux populism is ‘the active laboring and protest of the masses.’ However genuine the public mood being tapped may be, the movement itself tends to be ‘fairly sterile’ because the ‘common people’ are not the ones organizing themselves to promote change (Smith, 1998, pp. 48–9). With the ‘professionalization of the initiative industry’ noted earlier, there are many more faux populist entrepreneurs than there used to be. A growing number of political consulting firms ‘now draft and circulate petitions, collect signatures, raise funds through direct mail, prepare campaign advertising, conduct polls, and get out the vote. A mass organization is no longer needed to go over the heads of elected officials and take one’s case to the people. In California today, all it takes to qualify an initiative for the ballot is money, money, and more money’ (Citrin, 1996, p. 282). What is true for California (McCuan et al., 1998) is increasingly true in other states that permit citizen-initiated measures to appear on the ballot. 9 And while the opportunities for bottom-up activity are not currently widespread outside the United States, we anticipate that the diffusion of American-style campaign techniques may ultimately contribute (if it has not already done so) to a similar trend in countries such as Switzerland and Italy, and to heightened pressures elsewhere as citizens exert pressure on elites to employ referendums with greater frequency, and on a wider range of issues, than in the past.10 There are no existing data of which we are aware that can be used to determine whether, as stated above, the increased use of initiatives and referendums worldwide is truly a ‘democratic’ phenomenon. However, polls in the United States have consistently revealed what appears to be a high level of popular support for direct democracy. Most citizens, for example, would like to see the establishment of a binding nationwide initiative (Cronin, 1989), while those living in states that already have initiatives tend to regard them in a positive light. 11 Yet there are signs that the bloom may be starting to come off the rose for some voters who have been exposed to direct democracy on a regular basis – in California, for example, where the proportion of residents saying that ‘statewide ballot-proposition elections are a good thing’ dropped from


29

83 percent to 66 percent between 1979 and 1990 (perhaps in reaction to the professionalization of the process (Citrin, 1996, pp. 272–3)). Rather than a deeply felt desire for more citizen legislation, survey questions on the topic may be ‘tapping a somewhat ritualistic commitment to the principle of popular consent while at the same time permitting respondents to express contempt for the professional politician, an enduring negative symbol in American political culture’ (Citrin, 1996, p. 273). We examine the faux populism hypothesis by testing the other two competing explanations for rising levels of popular support for direct democracy. Recall that the cognitive mobilization hypothesis stresses changes that have occurred in the ability of citizens to be effective political actors. ‘Cognitive mobilization involves two separate developments. First, there has been a decrease in the cost of acquiring information about politics [for example, due to expansion of the mass media, especially television]. Second, there has been an increase in the public’s ability to process political information [due mainly to higher levels of education]. Cognitive mobilization thus means that more citizens now have the political resources and skills necessary to deal with the complexities of politics and make their own political decisions’ (Dalton, 1996, p. 21). It also means, we would imagine, that more people possess confidence in their own ability (that is, they have a strong sense of internal efficacy (Converse, 1972; Craig et al., 1990)), and in the ability of the mass public in general, to make important decisions that were once left almost entirely to elected leaders. Whether or not this confidence is warranted remains a matter of some dispute (Cronin, 1989; Budge, 1996; Citrin, 1996; Bowler and Donovan, 1998). Nevertheless, citizens who believe themselves capable of effective political action are likely to expect that the system will provide them (and others like them) with meaningful opportunities to participate; in cultures where democratic norms are strong, they also may become frustrated and cynical if it is perceived that such opportunities are unavailable (Craig, 1993; Abramson and Inglehart, 1995). For our analysis, the cognitive mobilization hypothesis leads us to predict that support for direct democracy will be associated with such variables as education, internal efficacy, political sophistication and attentiveness, and strength of partisanship.12 The disaffection hypothesis takes a rather different view of the motivational forces that are driving the recent surge in direct democracy: ‘Whereas the populist impulse (with its built-in skepticism toward politics and politicians) has always existed in the United States,’

30


observed Citrin (1996, p. 269), ‘since the late 1970s there has been a sharp rise in anti-political sentiments and a resurgence of interest in increased popular control of the government’s activities.’ This loss of public trust in governmental leaders and institutions is a global phenomenon. It is evident, for example, in Canada, Britain, Italy, Spain, Belgium, the Netherlands, Norway, Sweden, Ireland, and Japan (Nye, 1997, pp. 1–2; cf. Listhaug and Wiberg, 1995; Kaase and Newton, 1995; Nevitte, 1996; Inglehart, 1997; Mishler and Rose, 1997), and there are those who believe it to be the main reason why more citizens are demanding that they be permitted to play a larger role in public policy making. Moreover, even though individuals may give these developments a positive spin – for example, saying that they like voting on policy because it gives them ‘a chance to express their opinion’ (Citrin, 1996, p. 272) – the truth is that the content of many ballot measures conveys the impression of a disgruntled electorate attempting to rein in elected officials who could not otherwise be counted on to behave honorably. 13 We therefore anticipate that our analysis will show support for direct democracy to be greatest among respondents who score low on measures of political trust and external efficacy (the latter referring to beliefs about whether government is responsive to the wishes of ordinary citizens (Converse, 1972; Craig et al., 1990)), and who give low job performance ratings to elected leaders and representative institutions.14 Finally, we need to consider the joint effects of cognitive mobilization and disaffection on popular attitudes toward the use of referendums and initiatives. Some years ago, William Gamson (Gamson, 1968, p. 48) proposed that ‘a combination of high sense of political efficacy and low political trust is the optimum combination for mobilization – a belief that influence is both possible and necessary.’ Although the Gamson hypothesis has not fared particularly well in empirical tests (Craig, 1993, pp. 169–70), one reason for this may be that its terms are frequently conceptualized and operationalized improperly. As reformulated by Craig (1980 and 1993), the argument would seem to be that among the internally efficacious (an indicator of cognitive mobilization), feelings of political mistrust and external inefficacy (especially the latter, since it explicitly taps dissatisfaction with existing participatory structures) will increase the potential for mobilization. But what, exactly, constitutes ‘mobilization’? Prior research has generally examined various kinds of ‘aggressive’ or ‘unconventional’ behavior (civil disobedience, potential for political violence) – activities that we expect will appeal to relatively few citizens in established democratic


31

societies, unless perhaps they are deeply alienated (which most are not). Alternatively, Craig (1993, p. 171) has suggested that the mix of orientations described by Gamson might do a better job of accounting for involvement in ‘elite-challenging’ (as opposed to ‘regime-challenging’ (Inglehart, 1977, pp. 299–300)) modes of participation. This would presumably include support for direct democracy. Our own test of the Gamson hypotheses will consider each of these possibilities.

Data and methods The present study is based on a statewide telephone poll of 608 registered Florida voters (weighted N = 696), conducted in the spring of 1999 by the Florida Voter survey organization.15 Respondents’ attitudes toward the role of citizens in making public policy were tapped by a series of questions spaced throughout the interview schedule. The content of these questions, which constitute the dependent variables in our tests of the cognitive mobilization, disaffection, and Gamson hypotheses, will be discussed more fully in the next section. For now, let us simply point out that while Floridians do not vote on ballot propositions as often as the residents of some other states (for example, Oregon and California), they do so often enough that the questions we posed have some experiential basis for most people – especially given that our sample consists of registered voters, who over the past decade have been asked (either by elites or by entrepreneurs operating through the initiative process) to render judgment on issues relating to taxes, term limits, capital punishment, the environment, governmental restructuring, public education, and a variety of others. Florida voters should, then, be reasonably familiar with referendums and initiatives, and their views should therefore be about as well grounded as those of citizens anywhere else in the US. Our measures of cognitive mobilization include the following: (1) education (number of years formal schooling); (2) internal efficacy (an index built with two alternative-choice questions that tapped respondents’ beliefs about their own competence to participate effectively in politics);16 (3) political knowledge (number correct out of four questions relating to different aspects of national and state politics);17 (4) political interest (a single item asking whether respondents followed what goes on in government and public affairs most of the time, some of the time, only now and then, or hardly ever); (5) exposure (though not necessarily attentiveness (Price and Zaller, 1993)) to news media (how many days in the past week respondents

32


read a daily newspaper, and how many days they watched local news on TV); and (6) partisan strength (whether respondents considered themselves to be pure independents, independent leaners, weak identifiers, or strong identifiers). Political disaffection is operationalized in terms of (1) political trust (an index using some fairly standard forced-choice questions); 18 (2) job approval ratings for President Bill Clinton, the US Congress, Florida Governor Jeb Bush, and the Florida state legislature (as well as beliefs about whether, generally speaking, ‘things here in Florida are moving in the right direction’ vs. ‘off track and moving in the wrong direction’); (3) governmental responsiveness (two items drawn from the American National Election Study series); 19 (4) incumbent-based external efficacy; and (5) regime-based external efficacy. 20 In a society such as the United States with a political culture where the norm of governmental responsiveness is very strong, it is difficult to distinguish empirically between political trust and incumbent-based efficacy (because citizens will seldom place much trust in leaders who are thought to be inattentive to the public) (Craig et al., 1990; Craig, 1993). Although our survey included one question that tapped incumbent-based external efficacy, we suspect that the regime-based measure (an index reflecting respondents’ beliefs about the responsiveness of the system as a whole) will have a greater impact. Before going further, we want to acknowledge that the research presented here is limited in three ways. First, our data are from just one US state and therefore may not be representative of the patterns and relationships that exist elsewhere.21 Second, the data are cross-sectional rather than longitudinal in nature, and therefore cannot tell us everything that we would like to know about whether cognitive mobilization (by definition, a process that takes place over time) or the loss of public trust in government are responsible for increased popular support for direct democracy. Third, the data are at the individual rather than the aggregate level, and therefore cannot tell us much about the linkage (if any) between changing citizen demands for access to public policy making and the actual spread of direct democracy in many jurisdictions around the world. Strictly speaking, the questions posed earlier – are initiatives and referendums used more frequently today because of an increase in the number of people who want to play a role in making policy? and, if so, why? – are not ones that we can answer with the evidence at hand. What we can do is take an in-depth look at how one group of voters feels about citizen legislation, and explore the reasons why they feel that way. Our findings should provide useful clues for


33

those who want to investigate the popular roots of direct democracy elsewhere.

Results Public support for direct democracy Predictions by populist and progressive reformers in the United States that the establishment of direct democracy would stimulate voter interest and increase turnout have not really been borne out (Magleby, 1984; Cronin, 1989; Citrin, 1996; but cf. Smith, 1999). Although some issues do capture the public’s attention, others are largely ignored; turnout rates, on average, tend to be lower – and sometimes much lower – in proposition elections than in candidate elections. 22 Shall we conclude, then, that the faux populism hypothesis is correct – that citizen support for direct democracy is not as widespread or as intense as its advocates would have us believe? In the Florida survey (see Table 1.1), respondents appeared at first glance to be overwhelmingly in favor of both the referendum (84 percent approve, 59 percent strongly) and initiative (80 and 54 percent, respectively). Among voters who did approve, less than half wanted to see more of these elections in the future (42 percent referendum, 41 percent initiative), 23 but very few endorsed a cutback in either area. By a margin of better than two to one, most Floridians also rejected the notion that people today don’t have enough time to become wellinformed about complex issues; to the contrary, 63 percent believed that it would be ‘a good thing for the average citizen to become more directly involved in deciding what policies are adopted by our government’ (versus 27 percent who said it would be ‘not such a good thing’). On the other hand, there was less certainty about the idea of having referendum and initiative elections at the national level: just 54 percent approved (31 percent strongly) versus 35 percent disapproved. The breakdown was almost the same for one of the alternative-choice measures described earlier: 55 percent believed that ‘in our political system, all citizens should have the chance to participate directly in making government policy,’ while 36 percent felt that ‘what’s important is whether government policies get results, even if only a few people actually participate in making those policies.’ On a second alternative-choice item, just 46 percent claimed that ‘elected leaders should do what the voters want them to do even if they think it is a bad idea,’ compared with 43 percent who took the Burkean position that ‘elected leaders sometimes pay too much attention to what the public opinion polls tell them their

34


constituents want.’ Some degree of concern about the dangers of direct democracy also can been seen in the fact that only one voter in four (26 percent) expressed ‘a great deal’ of confidence in ‘the wisdom and good judgment of the American people’ versus 55 percent who had ‘some’ confidence and 18 percent who had ‘not very much’ (or ‘none’, a volunteered response). Table 1.1

Popular support for direct democracy in Florida, 1999

Item 1: As you may be aware, Florida sometimes has referendum elections, in which voters may either accept or reject laws that have been passed – or at least proposed – by the state legislature. When it comes to letting voters have a direct say in making government policy through referendum elections, would you say that you approve strongly, approve not so strongly, disapprove not so strongly, or disapprove strongly? Approve strongly Approve not strongly Disapprove not strongly Disapprove strongly Other/Don’t know

59.4% 24.4% 5.7% 4.1% 6.4%

Item 2: (If approve) In the future, would you like to see more referendum elections in Florida, fewer of them, or just about the same number as we’ve had in recent years? More About the same Fewer Other/Don’t know

41.8% 51.0% 6.8% 0.3%

Item 3: What about initiative elections, in which citizens are able to propose a new law directly – by gathering signatures on a petition – and then voters either accept or reject that proposed law at the ballot box on election day? When it comes to letting voters influence government policy through initiative elections, would you say that you approve strongly, approve not so strongly, disapprove not so strongly, or disapprove strongly? Approve strongly Approve not so strongly Disapprove not so strongly Disapprove strongly Other/Don’t know

53.7% 25.9% 6.5% 7.4% 6.5%

Item 4: (If approve) In the future, would you like to see more initiative elections in Florida, fewer of them, or just about the same number as we’ve had in recent years? More About the same

40.7% 46.9%

Public Opinion and Support for Direct Democracy Fewer Other/Don’t know

35

7.7% 4.7%

Item 5: Some people believe the country would be better off if more citizens became directly involved in political decision making. Others feel that as complicated as most issues are today, the average person does not really have enough time to play an active role in trying to solve them. What about you . . . Do you think it would be a good thing or not such a good thing for the average citizen to become more directly involved in deciding what policies are adopted by our government? A good thing Mixed/In between Not such a good thing Other/Don’t know

62.6% 5.8% 26.8% 4.8%

Item 6: What about having referendum and initiative elections at the national level – that is, letting voters make decisions directly about laws that would affect the entire country? Is this something you would approve of strongly, approve of not so strongly, disapprove of not so strongly, or disapprove of strongly? Approve strongly Approve not so strongly Disapprove not so strongly Disapprove strongly Other/Don’t know

31.0% 22.6% 13.6% 21.3% 11.4%

Item 7: Statement one – In our political system, all citizens should have the chance to participate directly in making government policy. Statement two – What’s important is whether government policies get results, even if only a few people actually participate in making those policies. (Do you feel strongly or not so strongly about that?) All citizens participate directly – strongly All citizens participate directly – not strongly Mixed/In between Whether policies get results – not strongly Whether policies get results – strongly Other/Don’t know

46.0% 9.2% 3.8% 12.9% 22.8% 5.3%

Item 8: Statement one – Elected leaders should do what the voters want them to do even if they think it is a bad idea. Statement two – Elected leaders sometimes pay too much attention to what the public opinion polls tell them their constituents want. (Do you feel strongly or not so strongly about that?) Do what voters want – strongly Do what voters want – not strongly Mixed/In between Pay too much attention to polls – not strongly Pay too much attention to polls – strongly Other/Don’t know

31.8% 14.1% 6.2% 14.3% 29.1% 4.5%

36


Table 1.1

(continued)

Item 9: In general, how much confidence do you have in the wisdom and good judgment of the American people when it comes to making political decisions – would you say that you have a great deal of confidence, some confidence, or not very much confidence? A great deal Some Not very much/None Other/Don’t know

25.5% 54.5% 18.0% 1.6%

Item 10: (If approve of initiative elections – see above) Is there any one issue that you would especially like to see placed on the ballot for people to vote on? (And which issue is that?) Taxes/Government spending Social issues (e.g., abortion, school prayers, gay rights, affirmative action, gender equality, welfare reform) Environment Other issues (e.g., gambling, campaign reform, term limits, tort reform, immigration, transportation) No one issue/Don’t know

5.2% 22.6%

3.6% 17.4% 51.2%

Source: Florida Voter statewide survey of registered voters, March–April, 1999 (weighted N = 696).

Two additional observations. First, respondents who approved of initiatives were asked to name one issue they ‘would especially like to see placed on the ballot for people to vote on’; even among this group (that is, setting aside those who disapproved of initiatives to begin with) more than half (51 percent) were unable to do so. Second, if it is true that an absence of consistency or ‘constraint’ is a sign that citizens either do not possess ‘true’ attitudes about a subject (Converse, 1964) or, alternatively, that the attitudes they possess are ambivalent (Zaller, 1992), then we may safely conclude that many Floridians are not altogether certain how they feel about the prospect of giving ‘the people’ more say in making public policy. To be sure, there is a tendency for individuals who express support on one of the questions in Table 1.1 to express support on others as well – but the tendency is not very strong, and there are numerous exceptions. For example, neither item 7 (citizens participate vs. getting results) nor item 9 (confidence in people’s judgment) is significantly correlated with item 1 (referendum approval); and item 9 along with item 8 (do what people want vs. too much attention to polls) are not correlated with item 6 (support for national referendum/initiative).


37

Thus, consistent with the faux populism hypothesis, the movement toward more citizen legislation appears to be a ‘movement’ without a reliable and committed mass base. Even though ‘the people’ (or most of them) are willing to go along for the ride, our data suggest that the success of contemporary direct democracy may indeed have more to do with the skills and resources of entrepreneurs who have learned how to recognize some ‘latent public mood’ (Smith, 1998) and turn it into viable electoral activity. This is not to say that Americans are opposed to direct democracy. Our claim is simply that when we look beneath the surface, there is less genuine enthusiasm for the process than the initial poll numbers would lead us to believe. Cognitive mobilization and political disaffection To the extent that we can take expressed attitudes about direct democracy at face value, what can the Florida data tell us about the factors helping to shape those attitudes? Despite the modest correlations noted above, we were able to build an index of (barely) acceptable reliability 24 using the four core questions presented in Table 1.1: item 1 (referendum approval), item 3 (initiative approval), item 5 (should citizens become more directly involved in policy making), and item 6 (national referendum/initiative). This additive Index of Support for Direct Democracy will be the principal dependent variable for our tests of the cognitive mobilization, disaffection, and Gamson hypotheses. As is evident from the top portion of Table 1.2, the cognitive mobilization hypothesis fares quite poorly. First, none of the indicators shown here – Education, Internal efficacy, Political knowledge, Political interest, Exposure to news media (separate questions for newspapers and television news), and Partisan strength – had a significant effect on support for direct democracy. Second, the outcome was essentially the same when we examined the three action components of our Index (Referendum approval, Initiative approval, National direct legislation) separately: only 5 of 21 coefficients achieved the 0.05 probability threshold (2-tailed test), and some of the relationships that did were in the opposite direction from what was expected. Frequent newspaper readers were less rather than more inclined to endorse initiatives (r = 0.10), national-level direct legislation appealed most to the uninformed (0.15) and those not interested in politics (–0.11) – and the internally efficacious were, as predicted, slightly more likely than the inefficacious to approve of both referendums (0.10) and initiatives (0.09).25 Yet these were minor exceptions to the overall pattern. Generally speaking, our data did not leave the impression that cognitive mobiliza-

–0.16* –0.09* –0.10* –0.12*

–0.06 –0.01 –0.10* –0.04

Clinton approval

0.06 0.10* 0.09* –0.03

Internal efficacy –0.01 0.05 –0.01 –0.11*

Interest in politics

–0.07 0.03 –0.11* –0.01

Congress approval –0.11* –0.01 –0.10* –0.10*

Gov’t responsiveness

B. Political Disaffection

0.05 –0.00 0.04 0.15*

Political knowledge

0.04 0.14* 0.04 0.03

Right/wrong direction

0.05 0.05 0.10* 0.08

Read newspaper

–0.04 –0.00 –0.03 –0.02

External efficacy (I)

0.01 0.07 0.02 –0.05

Watch TV news

–0.04 0.02 –0.05 –0.06

External efficacy (R)

–0.03 0.00 –0.06 –0.02

Partisan strength

Source: Florida Voter statewide survey of registered voters, March–April 1999 (weighted N = 696).

Note: Table entries are Pearson’s r; correlations with an asterisk (*) are significant at p ≤ .05 (2-tailed test). Low scores indicate support for direct democracy and (1) for cognitive mobilization variables: low education, high internal efficacy, low political knowledge, high interest in politics, low attentiveness to newspapers and television news, strong partisanship; and (2) for political disaffection variables: high trust, high approval of Clinton and the Congress, high perceptions of responsiveness, Florida going in the right direction, and high incumbent-based (I) and regime-based (R) external efficacy.

Direct democracy index Referendum approval Initiative approval National direct legislation

Political trust

0.03 0.01 –0.01 0.08

Education

A. Cognitive mobilization

The cognitive mobilization and political disaffection hypotheses

Direct democracy index Referendum approval Initiative approval National direct legislation

Table 1.2

38


39

tion factors were driving popular opinion toward the use of direct democracy. Nor, for the most part, was political disaffection (see the bottom half of Table 1.2). As measured by the Index, pro-direct democracy orientations were somewhat more common among the mistrustful (r = –0.16) and respondents who felt the government was unresponsive (–0.11). 26 In addition, individuals who disapproved of the job being done by President Clinton and by the US Congress were more likely to favor initiatives (–0.10 and –0.11, respectively). On the other hand, those who said that things in Florida were going in ‘the right direction’ expressed a more positive view of referendums (0.14). 27 Overall, the relationships were inconsistent and (when they existed) of modest magnitude. Our preliminary conclusion must therefore be that political disaffection has only a limited impact on citizens’ support for direct democracy. 28 What about the Gamson hypothesis? Perhaps it is not either cognitive mobilization or political disaffection alone, but rather the combination of the two, that makes some people less willing to leave policy-making power solely in the hands of their elected representatives. We tested this possibility by looking at the interaction between internal efficacy and three of our disaffection measures from Table 1.2: trust, responsiveness, and regime-based external efficacy (the latter being the dimension that seemed likely, on theoretical grounds, to have the greatest effect (Craig, 1980)). In fact, only two of the multiplicative terms in Part A of Table 1.3 were related to any dependent variable – a lackluster r = 0.09 with Referendum approval for both Internal*Responsiveness and Internal*External. In Part B, where the relationship between disaffection and attitudes toward direct democracy is observed for three levels (high, medium, low) of Internal efficacy, we find little indication that the strength of this relationship was enhanced by a strong sense of confidence in one’s skills as a political actor. If anything, the correlations appear to be marginally stronger among those at an intermediate level of internal efficacy, providing little evidence for the Gamson hypothesis. We originally planned to close this paper with a multivariate analysis that incorporated not only the most important of the variables discussed earlier, but also partisanship, ideological orientation, and additional demographic factors. Under the circumstances, such an exercise would surely be overkill (see note 26) – especially given that the bivariate correlations for party, ideology, and demographics are every bit as unimpressive as those for cognitive mobilization and political

40


Table 1.3

The Gamson hypothesis

A. All respondents

Internal x Trust

Internal x Responsiveness

Internal x External (R)

Direct democracy Referendum approval Initiative approval National direct legislation weighted N = 696

–0.03

–0.00

0.05 0.03 –0.04

0.09* 0.04 –0.06

0.02 0.09* 0.03 –0.05

B. By internal efficacy

Political trust

Govt responsiveness

External efficacy (R)

High efficacy: Direct democracy Referendum approval Initiative approval National direct legislation weighted N = 161

–0.12 –0.06 –0.10 –0.13

–0.11 –0.10 –0.12 –0.04

–0.04 –0.01 –0.09

Medium efficacy: Direct democracy Referendum approval Initiative approval National direct legislation weighted N = 321

–0.17* –0.11 –0.13* –0.15*

–0.11 0.00 –0.23* –0.13*

–0.07 –0.01 –0.07 –0.12

–0.17 –0.05

–0.08

–0.02

0.00 0.15 –0.10

–0.05

Low efficacy: Direct democracy Referendum approval Initiative approval National direct legislation weighted N = 158

0.02 –0.01

0.01

0.05 0.01

Note: Table entries are Pearson’s r (see Table 2 for direction of coding); correlations with an asterisk (*) are significant at p ≤ 0.05 (2-tailed test). External efficacy (R) is the regime-based measure. Source: Florida Voter statewide survey of registered voters, March–April 1999.

disaffection. The data simply do not allow for firm conclusions to be drawn concerning the origins of popular support for direct democracy.

Discussion The data from our Florida survey suggest that, at the individual level, neither cognitive mobilization nor political disaffection measures – taken singly or in combination – are able to explain much of the variation in citizens’ attitudes toward direct democracy. Nonetheless, favorable views


41

toward direct democracy are somewhat more pronounced among those who lack confidence in governmental leaders and institutions, and so it is likely that the loss of public trust in the US and elsewhere over the past few decades may have contributed to electorates increasingly seeing direct democracy as a viable means for holding elites accountable. It is also possible that one or two of the cognitive mobilization variables (internal efficacy is probably the best candidate) have had a similar effect. If our findings can be generalized to other settings, however, this is likely to be true (if at all) only at the margins. Where, exactly, does this leave us? Having worked from an empirical literature that provided little theoretical guidance, it is possible that we have omitted critical variables from our analysis – though we are presently at something of a loss to imagine what those variables might be. Alternatively, perhaps we have simply done an inadequate job of operationalizing popular attitudes toward direct democracy. It is extremely difficult to write survey questions that convey (1) enough information about referendums and initiatives that respondents understand clearly what they are being asked to evaluate but, at the same time, (2) not so much information that one runs more than the usual risk of ‘creating’ attitudes (or ‘non-attitudes’) where none previously existed. For the most part, we are satisfied that the principal reason for our meager results has less to do with the questions per se than with the fact that many Floridians – in all likelihood, many Americans in general – do not have well-developed attitudes about this particular topic. We commented earlier on the lack of internal consistency among our measures of direct democracy, and in the previous section it became apparent that these measures were at best weakly correlated with a wide range of theoretically relevant outside variables. Whether what we are seeing here is evidence of non-attitudes or ambivalence (or a mix of the two), the bottom line is the same: regardless of poll results to the contrary – and not all poll results are to the contrary, e.g. look again at the response distributions for items 6 through 9 in Table 1.1 – we suggest that a sizeable number of voters are just not sure how they feel about the prospect of letting ordinary citizens like themselves participate directly in making public policy. When the issue is framed primarily in symbolic terms (including ‘Us versus Them’), there is little contest: almost everyone agrees that empowering the electorate is a good thing. But when asked about specific reforms (more citizen legislation? a national referendum?) or about the implications of extending the present system (will people make wise choices? is participation more important than getting results?), many respondents appear to have

42


second thoughts. Accordingly, those who subscribe to the faux populism model and who believe that the recent trend toward direct democracy stems more directly from the strategies of political entrepreneurs than from the deep-seated attitudes and values of the general public would seem to have a valid point. If this is true in the United States, where ‘democracy’ is a (maybe the) core element of national identity, should we expect that it will also be true in other settings? Not necessarily, since the pattern in any nation should reflect, to a significant degree, the specific history and political development of that nation; in other words, Americans (or at least those living in states that have referendums and initiatives) may give these instruments of direct democracy little thought precisely because they have become such a familiar part of the electoral landscape. 29 One wonders whether the same thing would be true today in the emerging democracies of Eastern Europe – or, for that matter, in the United States itself back in the 1890s. Alas, without new breakthroughs in time travel, the latter question must remain unanswered. We do, however, hope that our findings are of some use to scholars who wish to understand better the levels and bases of public support for direct democracy that exist among contemporary electorates around the world.

Notes 1 The language of Us versus Them is borrowed from Citrin (1996). 2 Depending upon electoral majorities to protect the rights of all citizens may, of course, be problematic. See Gamble (1997; but cf. Citrin, 1996, pp. 279–81; Cronin, 1989, pp. 90–9; Donovan and Bowler, 1998c, pp. 264–70; Bowler and Donovan, this volume; Frey and Goette, 1998) for an illustration pegged specifically to the recent history of direct democracy in the United States. 3 Most such proposals originate in the legislature, but there are exceptions. In Florida, for example, a commission is appointed every twenty years to recommend changes in the state’s constitution. Thirteen amendments were developed in this fashion and placed on the 1998 general election ballot; following a low-key campaign that was largely devoid of controversy, 12 of the 13 passed – a success ratio that exceeds the historical average, even for top-down referendums (which, in turn, have a better track record than initiatives and popular referendums; see Waters, 1999). 4 South Dakota, in 1898, became the first state to enact the initiative and popular referendum. Most of the others that chose to follow suit – primarily in the West, where populist sentiment was (and probably continues to be) strongest – did so prior to 1920. 5 The decade of the 1980s ranks third overall (276 statewide), while the fewest initiatives were considered by voters during the 1940s (133), 1950s (108), and 1960s (79). Interestingly, the success rate of initiatives in the 1990s (46 percent)


6

7

8

9

10

11

12

43

and 1980s (44 percent) also is higher than the 100-year average of 40 percent (Waters, 1999, pp. 124, 128). Their ability to do so was limited in important respects by a 1999 ruling (Buckley v. American Constitutional Law Foundation) in which the US Supreme Court overturned major portions of a Colorado statute which required that signature collectors ‘wear name badges, be registered to vote, [and] report their addresses and the fees they were paid’ (Bender and Tenner, 1999, p. 61). Nevertheless, according to Waters (1999, p. 123), one reason why fewer initiatives were on the ballot in 1998 – 55 overall, a decline of 35 percent from two years earlier – is that ‘state legislatures (not just Colorado) had made the process more difficult to use.’ Or there may be an element of both. As Citrin (1996, p. 290) has observed, ‘populism clearly has a punitive impulse, but the popularity of direct democracy also expresses a positive yearning for voice – for the chance to be heard and to participate.’ The latter quote is actually from John Kingdon, cited in Smith (1998, p. 48). The specific public mood with which Smith was concerned – and around which he built his argument – had to do with the anti-tax sentiment that first became evident with California’s notorious Proposition 13 in 1978 (see also Sears and Citrin, 1982), and subsequently spread to many other states where initiatives are permitted. This is becoming all the more true as the opponents of certain initiatives turn to counter-propositions – rapidly qualified initiatives ‘designed to deflect attention’ from the original measure. ‘Rapid qualification of these counterproposals would not be possible,’ according to Donovan and Bowler (1998a, p. 10), ‘without petition management firms and the maturation of a campaign industry that has made it easier for “grassroots” and industry alike to get on ballots in populous states.’ Any such trend will depend to some extent on country-by-country variations in laws governing campaign finance and spending; for example, see Gallagher and Uleri (1996). Also, in the American National Election Study surveys of 1972 and 1974, almost 70 percent of the American public indicated that they would like (versus just under 20 percent who said they would not like) more issues to ‘be decided by the voters at the polls.’ Outside the United States, support for direct democracy can perhaps be inferred from the rise in post-materialist value priorities – one component of which is the belief that citizens should have ‘more say’ in political decision making – that is said to have occurred in many liberal democracies since the end of World War II (Inglehart, 1977 and 1990). Diminished party loyalties (along with higher levels of issue voting and participation in some of the ‘more demanding forms of political activity’) are among the other consequences of cognitive mobilization identified by Dalton (1996, pp. 26–7). Also see Bogdanor (1994, p. 91), who has attributed the increased use of referendums in Western Europe to an ‘unfreezing of the political alignments and institutions formed after the achievement of universal suffrage in the first two decades of the twentieth century’ (cf. Lipset and Rokkan, 1967). Central to this process is a growing ‘detachment’ from the traditional parties (Setälä, 1999, p. 9) combined with greater support for

44

13

14

15

16

Stephen C. Craig, Amie Kreppel and James G. Kane the ‘weapons of direct democracy, which . . . (enable) the people to control their leaders’ (Bogdanor, 1994, p. 94). Hence, we expect that non-identifiers in our own sample – along with those who believe that government does not pay enough attention to citizens’ views (that is, the externally inefficacious, as measured by the questions outlined in note 20 (cf. Setälä, 1999a)) – will be more likely than strong partisans to sanction the use of referendums and initiatives. This seems to be the case both in the US (for example, constraints on taxing and spending, term limits on elected officials – see Magleby, 1994; Bowler et al., 1998) and elsewhere (for example, Italy, where direct democracy has been used to rebuke elites – see Uleri, 1996; Bogdanor, 1994). A variant of the disaffection hypothesis might focus specifically on citizens’ economic dissatisfactions (Weatherford, 1987); that is, individuals should be more inclined to endorse direct democracy when they believe that elected leaders are doing a poor job of managing the nation’s economy. While our survey does not test this directly, the governmental performance ratings and right/wrong direction item discussed in the next section (and portrayed in Table 1.2) probably capture respondents’ economic assessments fairly well. An alternative explanation that combines elements of the disaffection and cognitive mobilization hypotheses points to the long-term decline of deference to established authority observed in the United States, Canada, and Western Europe (Nevitte, 1996). This perspective is represented, again indirectly, by several of the independent variables in our analysis, for example political trust, internal efficacy, and exposure to newspapers and television news (the latter are sometimes seen as having contributed to the public’s loss of confidence in governmental institutions and leaders – see Cappella and Jamieson, 1997). Florida Voter publishes a monthly, non-partisan political journal based upon periodic surveys of registered voters within the state. The findings reported here are from a survey done between 25 March and 8 April 1999. Respondents were randomly selected from a list of registered voters, with up to three callbacks being made for each working number in an effort to obtain a completed interview. Additional information concerning this poll can be obtained from Florida Voter directly (800–899–7655), or from the Florida Institute for Research on Elections in the Political Science Department at the University of Florida. Because of the serious problems of response set that plague traditional agree–disagree questions (Converse and Presser, 1986; Krosnick, 1999), we have chosen to use the alternative-choice format instead. Respondents were read two contrasting statements and asked to indicate which one came closest to their own opinion; having done so, they were then asked whether they preferred that statement ‘strongly’ or ‘not so strongly’. The two statement pairs used to measure internal efficacy were (1) ‘I feel that I have a pretty good understanding of the important political issues facing our country vs. Sometimes politics and government seem so complicated that a person like me can’t really understand what’s going on’; and (2) ‘I often don’t feel sure of myself when talking with other people about politics and government vs. I think I am better informed about politics and government than most people’ (Craig et al., 1990).


45

17 Respondents were asked (1) what political office is currently held by Janet Reno (US Attorney General); (2) which party has the most members in the House of Representatives in Washington and (3) in the state legislature (Republican in both cases); and (4) which party, if either, is more conservative than the other at the national level (again, Republican). 18 Respondents were asked (1) ‘How much of the time do you think you can trust the people who run our government to do what is right – just about always, most of the time, only some of the time, or almost never?’ (2) ‘When government leaders make statements to the American people on television or in the newspapers, how often do you think they are telling the truth – just about always, most of the time, only some of the time, or almost never?’ and (3) ‘Would you say the government is pretty much run by a few big interests looking out for themselves, or that it is run for the benefit of all the people?’ (Craig et al., 1990). 19 These questions asked respondents to indicate how much attention ‘the government pays to what the people think when it decides what to do’ and how much ‘having elections makes the government pay attention to what the people think’ – a good deal, some, or not much. 20 All external efficacy questions were in the alternative-choice format. Incumbent-based efficacy was measured with the following statement pair: ‘Generally speaking, those we elect to public office lose touch with the people pretty quickly vs. Most public officials are truly interested in what the people think.’ The regime-based efficacy index was constructed from two items: (1) ‘There are many legal ways for the average citizen to influence what the government does vs. The average citizen doesn’t have much say about what the government does’; and (2) ‘In the end it is the people who have the final say about how our country is run vs. If public officials are not interested in hearing what the people think, there is really no way to make them listen’ (Craig et al., 1990). 21 In fact, we consider it highly likely that the forces giving rise to an increase in direct democracy are at least marginally different from country to country (Setälä, 1999), and that some of those forces involve elements other than those examined here (see note 14). 22 This is generally true outside the United States as well (Butler and Ranney, 1994), though it is possible that turnout rates will vary from country to country – or in a single country over time – due to such factors as referendum frequency (more elections producing voter fatigue), subject matter (some topics are more salient than others), and level of satisfaction with national conditions and the performance of governmental institutions (a positive assessment making some citizens more willing to let others decide). 23 These are not always the same people. Some of those who want more referendums say they would prefer fewer (5 percent) or about the same number (27 percent) of initiatives as the state has now. A similar proportion of those who favor more initiatives would like there to be fewer (2 percent) or about the same number (30 percent) of referendums. 24 Standardized item alpha = 0.619, with a mean inter-item correlation of r = 0.29. 25 This last finding is odd given that there is a fairly strong positive relationship between Internal efficacy and Political knowledge (r = –0.41, with the negative

46

26

27

28

29

Stephen C. Craig, Amie Kreppel and James G. Kane sign indicating simply that the two variables are coded in opposite directions), and we would therefore expect each to be correlated in a similar fashion with our dependent variables. The relationships in Table 1.2 are so weak, however, that we probably should not attach too much substantive meaning to the apparent inconsistency. Both relationships weakened when the other variable was held constant, though the correlation between trust and democracy (r = –0.12) remained statistically significant. In this context, it should be noted that the results obtained from multivariate analysis (OLS regression) did not differ substantially from those reported in Table 1.2. Although ‘right direction’ voters had a more favorable attitude toward referendums in general, it was their ‘wrong direction’ counterparts who (by a small margin) proposed having more of these votes in the future. Also, support for both referendums (r = 0.14) and initiatives (0.09) was stronger among those who approved of Jeb Bush’s performance as governor; job ratings for the Florida state legislature were unrelated to any of our dependent variables. At the aggregate level (where the relationship of interest is not mistrust support for direct democracy but direct democracy trust), Citrin (1996, pp. 286–7) found levels of political trust to be about the same in states with direct democracy as in states without; that is, the experience of ‘going over the heads of elected officials’ did not appear either to enhance or to undermine ‘the legitimacy of representative democracy.’ It is not clear, however, that direct democracy in the states should be expected to influence citizens’ attitudes toward the national government (which is what Citrin happened to be measuring in this instance). By the same token, support for the idea of citizen involvement is hardly an American phenomenon. In the 1990 World Values Surveys, for example, 91 percent of Canadians and 86 percent of Europeans – compared with 83 percent of Americans – agreed with the statement that ‘our government should be more open to the public’ (Nevitte, 1996, p. 100).

2 The Rise of Government-Initiated Referendums in Consolidated Democracies Laurence Morel

The referendum has been an important decision-making device since the mid-twentieth century in very few countries. Since the beginning of the 1960s, however, there has been a striking increase in the use of referendums to settle an ever growing range of questions. This is particularly true in the consolidated democracies, though similar trends can be found in other political contexts. Two questions are raised by the increase in the frequency of referendums in the consolidated democracies. First, what can explain it? Is it circumstantial factors, such as the emergence of new cleavages, to which parties will soon respond and, by so doing, end the apparent ‘crisis of representation’ that is seemingly provoking popular consultations? Or is it part of a more structural trend, an ineluctable process whereby democracies are becoming, by means of referendums as well as other devices (such as opinion polls and televoting), more and more direct and less and less mediated? Second, is this growing use of referendums necessarily a positive development, given that ‘more referendums’ does not necessarily mean ‘more democracy’? Such a phenomenon could, in fact, merely mask the use of the referendum by governments not for genuine consultative purposes, such as giving the public a greater say in shaping legislation, but instead for instrumental and strategic purposes, such as the consolidation of governments’ own power. This paper examines in detail the reasons why governments have initiated referendums in the last fifty years in consolidated democracies. Through this study, I will be able to offer an explanation for the increase in government-initiated referendums at the end of the twentieth century. In addition, by examining the motivations of governments, 47

48

Laurence Morel

I will be able to offer some insight into whether the increase in referendums initiated by governments is fueled by motivations that are genuinely democratic or primarily instrumental. I find that governments’ motivations for initiating referendums can be grouped into four categories: to consolidate their own power (‘plebiscitary motivations’), to resolve internal divisions within the governing party or coalition (‘tension resolving referendums’), to pass legislation that otherwise would be impossible to pass (‘legislative motivations’), or because the referendum was essentially required by the fact that passing the policy without referring it to the people would seriously undermine the legitimacy of this policy (‘ de facto obligatory legitimation referendums’).

The study I focus on government-initiated referendums. I have made this choice because this variable – who initiates the referendum? – has the greatest effect on the practice and dynamics of referendums. Other criteria, such as the nature of the issue or whether or not the referendum is binding, have far less impact. Whether it is the government or another group that initiates the referendum has the greatest effect on the referendum itself. It is for this reason that the majority of referendum typologies grant a central place to the variable of who initiates the referendum. Of these, the one developed by Gordon Smith deserves special mention (Smith, 1976). Smith classifies referendums according to two criteria, the first of which relates to whether or not the decision to initiate a referendum is controlled by the government and those that are not. In his view, a government-controlled referendum is usually intended to produce results that favor the government’s authority. On the other hand, referendums that are not government controlled (the typical case being that of the popular initiative) generally are launched to bring about changes that the government resists. The actual effect of the referendum is the second criterion identified by Smith. The effect of the referendum is generally pro-hegemonic when controlled by the government and generally anti-hegemonic when controlled by others. However, the correspondence is not perfect: it is possible to have referendums initiated by non-governmental groups which end up in reinforcing the government, and it is possible that the government could initiate a referendum that in fact ends up damaging the government itself. Therefore, Smith’s typology distinguishes between four types of referendums: (1) controlled/ pro-hegemonic; (2) controlled/anti-hegemonic; (3) uncontrolled/ pro-hegemonic; and (4) uncontrolled/anti-hegemonic – with the first and fourth

The Rise of Government-Initiated Referendums 49

types being far more common (there is also a fifth residual category, which Smith calls ‘non-functional’). Employing Smith’s typology, we can say that government-controlled referendums could be less democratic than those not controlled by government. Because of the degree to which they can be manipulated by the government for its own partisan or strategic ends, they are liable to be biased in practice, both because the government’s aims may be instrumental to begin with, but also because the rules could be biased in favor of the government and hence violate the norms of fair process. This does not mean that referendums by popular initiative or by minority initiative (e.g. those that can be initiated in Denmark by a minority of members of parliament or in Switzerland and Italy by a small number of cantons or regions) are free from strategic manipulation for instrumental ends. Anyone who initiates a referendum is by definition a political actor with a variety of motivations, including the maximization of their own power and influence. However, the tools at the disposal of the government to structure the referendum process are far greater than those that can be employed by other actors. Who initiates the referendum therefore constrains or facilitates the exercise of political power. Government-initiated referendums are thus a very specific category of referendums, both in terms of why they are held and how well they are likely to conform to a democratic standard. The study of the motivations that lead governments to use referendums will allow us to be more precise concerning these two points. In particular, if we find that there has been a substantial increase in the number of referendums undertaken for instrumental or strategic purposes, or an increase in the number of referendums undertaken to reinforce the political regime as a whole, this will not be consistent with the image that an increased use of referendums is leading to ‘more democratic’ governing styles with more control for popular movements outside of government to participate in legislation. The study has been limited in three respects. First, it focuses only on national referendums. This is done for practical reasons: the large number of regional and local referendums that have been conducted would make a comprehensive study virtually impossible. Table 2.1 provides a list of these referendums in 20 consolidated democracies since 1900. Second, the study focuses on the period between 1940 and 1999. And third, the study will focus on referendums which are formally initiated by governments. Referendums formally initiated by governments fall into three categories: referendums initiated by the executive (in practice, this means the president of a republic), referendums initiated by the legislature (a majority in either chamber), and referendums initiated by

11 14 4

14 1 1

5

2

(1) 1 1 (1) 1

(3)

3 3

32

1 5

1

2 3

23

5

(1)

(4)

2

(1)

1 1

(1)

1930–39

1

1920–29

17 1 5 6

7

(1)

1

(1) 2

4

1940–49

1 2

45

2 7

1

1

1 2

1950–59

2 6

26

12

(1)

2

(1)

4

6

1960–69

11 3

86

18

2

1

5 3

1 1

1

3

1

1970–79

6 3

62

1 1 20

4 12

1

1

1980–89

98 1 2 9

1 52

1 2

10 32

3 1 1

1

1990–98

414 2 47 40

2 1 18 2 12 2 1 1 4 23 48 3 6 2 3 5 133

Total

Source: Butler and Ranney (1994); Centre d’études et de documentation sur la démocratie directe (Université de Genève; Faculté de droit).

Note: Numbers between brackets indicate referendums held in authoritarian contexts. They are not included in the totals. Large countries which have not had any referendums since democracy are the Netherlands, the United States, Israel, Japan, and India.

Other Countries Switzerland Canada Australia New Zealand

2 1

1

1

1910–19

2

1900–09

National referendums in 20 consolidated democracies (1900–98)

Western Europe except Switzerland Austria Belgium Denmark Finland France Germany Great Britain Greece Iceland Ireland Italy Luxembourg Norway Portugal Spain Sweden Total

Table 2.1 50


both the executive and legislature. This third criterion is somewhat different from Smith’s, who based his typology on the real initiator as opposed to the formal initiator, as indicated by the fact that some obligatory referendums are put in the category of government-controlled referendums (in particular the Irish referendum on EEC membership) and some referendums initiated by the government in the category of uncontrolled referendums (in particular the British referendums on EEC membership). However, selecting referendums according to their real initiative cannot constitute the starting point for an empirical study; identifying the real initiator of a referendum can only be the result of a study, a study which takes the formal initiative as its starting point. These limitations leave us with 38 referendums. Of the 508 referendums in the 20 countries appearing in Table 2.2, I am clearly dealing only with a small minority. Even if we exclude the 334 referendums held in Switzerland, these 38 referendums still constitute a very small number. Nevertheless it must be noted that the government-initiated referendum has been used in 15 of the 18 countries which have had national referendums since 1940, and that it was the most frequently used type of referendum in 11 of the 18 countries that have had referendums during the period. Therefore, putting aside certain exceptional cases such as Switzerland, Australia, and Italy, the government-initiated referendum is by far the dominant form of referendum in most countries.

The rising use of government-initiated referendums Table 2.2 indicates that the government-initiated referendum accounts for a very small portion of the explosion of referendums beginning in the 1970s. To a large extent, the increase in number had more to do with the rise in popular initiatives and, to a lesser extent, obligatory referendums. The increased use of these two types of referendums is also concentrated in a small number of countries: Switzerland and Italy for the popular initiative, and Switzerland, Australia, and Ireland for the obligatory referendum. However, despite having contributed very little to the overall increase in the number of referendums, the use of government-initiated referendums did rise in half of the countries in Western Europe by about one third, with the 13 held between 1940 and 1969 growing to 17 between 1970 and 1999. Table 2.3 shows that this increase in government-initiated referendums stems largely from the holding of numerous referendums on the various stages of European integration.

0

4

1

6

47 8 10

35

2

1

17

4

139

46

9

Non-government initiated*

2 13 4

Counterproposal

2 2 2 2 17

1 1

1 1 3

1

Government initiated

46

3

2

4

Obligatory

1

0

4

Non-government initiated*

19 9

90

27

1

19

6

1

Obligatory

334 2 27 29

2 1 15 1 12 0 1 1 2 22 48 0 2 2 3 4 116

Total

* Non-government initiated referendums in Denmark are initiated by one-third of the members of the legislature; in Italy and Switzerland, they have been initiated by popular petition.

Other countries Switzerland Canada Australia New Zealand

Counterproposal

1970–99

1

2

7

1

Government initiated

1940–69

National referendums by country, initiator, and period (1940–98)

Western Europe except Switzerland Austria Belgium Denmark Finland France Germany Great Britain Greece Iceland Ireland Italy Luxembourg Norway Portugal Spain Sweden Total

Table 2.2 52

The Rise of Government-Initiated Referendums 53 Table 2.3

National government-initiated referendums by theme 1940–69

1970–98

Total

8 3 0 2 3 1 1

6 1 9 1 1 2 0

14 4 9 3 4 3 1

18

20

38

Constitutional Territorial European integration NATO/conscription Moral issues Nuclear/driving on right Pensions Total

Why do governments initiate referendums? I undertook an examination of the objectives of governments in initiating referendums. Motivations are often multifaceted, and include strategic and partisan objectives, or the recognition that it will be easier to pass the legislation through the referendum than through the legislative process. Such motivations often overlap: a party may choose to use a referendum to resolve tensions within its own party, but if this decision also divides the opposition, so much the better. While recognizing that the objectives of a government can be multifaceted and overlapping, in many referendums it is possible to identify one primary objective. These objectives can be grouped in four general categories, summarized in Table 2.4 for 22 referendums. The categories include: referendums designed to resolve divisions within the governing party or coalition; those that are necessary in order to advance the legislative agenda of the governing party or coalition; those that serve a power consolidation and electoral function; and those that were de facto (though not de jure) obligatory. The divisions-solving referendum Since 1970, referendums designed to resolve divisions within the governing party or coalition have been particularly frequent. In fact, part of the general increase in government-initiated referendums since 1970 is accounted for by governments turning to the referendum device to resolve internal tensions. These kinds of referendum have dealt with issues that include the European union (4 referendums), civil nuclear power (2) and membership of NATO (1). These referendums were organized by the governments of the day to cope with two situations: a split in the governing party (Norway, 1972; Great Britain, 1975; Austria, 1978; Sweden, 1980 and 1994; Spain, 1986), which was sometimes magnified by a split in the governing coalition (Sweden, 1955 and Finland,

54

Laurence Morel

Table 2.4 Typology of governmental objectives during 22 government-initiated referendums Division-resolving referendums

Austria Finland Norway Spain Sweden Sweden Sweden UK

1978 1994 1972 1986 1955 1980 1994 1995

Nuclear energy EU membership EEC adhesion Remaining in NATO Driving on the right Nuclear energy EU membership Remaining in EEC

Legislative referendums

Belgium Canada Denmark France France France France Sweden

1950 1986 1986 1961 1962 1962 1988 1957

Return King Leopold III Charlottetown Accord EC Act Self-determination in Algeria Independence of Algeria Direct election of president Agreements on New Caledonia Pension schemes

Plebiscitary referendums

France France

1958 1969

France France

1972 1992

Constitution Senate reform and creation of regions Enlargement of EEC Maastricht treaty

Canada Norway

1942 1994

Military conscription EEC adhesion

‘Politically obligatory’ legitimation referendums

1994). The decisions to hold these referendums were motivated by two somewhat different kinds of goals: to use the referendum as a mediation device between competing factions, and to make use of the referendum as a palliative to avoid the electoral repercussions of a divisive issue (Bjørklund, 1982). The mediation function first characterized the Norwegian, English, and Spanish referendums. In these three cases the decision to organize the referendum was taken by the prime minister with the aim of restoring the unity of the party, or avoiding a split. Thus, in Norway, the decision in favor of holding a referendum was taken by the leadership of the Labour Party as early as 1962, at the beginning of the negotiations for that country’s entry into the European Economic Community (EEC). The aim, obviously, was to defuse the bomb which the European question represented for the party. Similarly, in Great Britain, Harold Wilson, the Labour Prime Minister, saw the referendum as the only


means of avoiding the rebellion within his own party that would have been caused by Parliament’s adoption of renegotiated membership of the EEC (a decision that would have been taken regardless of dissenting Labour votes by virtue of the support of members of the opposition Conservatives and Liberals). The Labour Party was dominated by those opposed to membership, so the call for a referendum by the ‘Europhobes’ permitted Wilson to use the device to mediate tensions within his own party. During the 1982 election campaign in Spain, the Socialist Party had promised to hold a referendum on NATO. The leader and eventual prime minister, Felipe Gonzalez, would probably not have kept his promise if not for relentless pressure from the left of his party. In office, Gonzalez had rallied support for Spain’s continued membership of NATO, but pressure from forces within his own party – youth and unions, and allies on the more extreme left such as environmentalists and pacifists – opposed to membership of NATO forced him to keep his word to hold a referendum. This referendum represented a means of resolving a conflict within the Spanish left, both inside the party in power and between the various components of the left generally. With regard to the Swedish referendum on ‘driving on the right’ and the Finnish referendum on membership of the European Union, the situation was complicated by the fact that the divisions on these questions simultaneously affected the individual parties of the coalition and the coalition itself. In Sweden, it was the Agrarian Party, coalition partner of the Social Democrat Party, that convinced the latter to organize a referendum. For the majority in the former, the referendum was a two-pronged weapon: contrary to both the Social Democrats and to a minority within its own ranks, it wanted to maintain driving on the left. The referendum was thus a device for gaining leverage in the coalition and a means of managing its own internal divisions on the subject. For the Social Democrat prime minister, supportive of the proposed changes, the referendum had to be used to prevent this issue from harming the cohesion of the coalition, and, like the Agrarian Party, it had also to neutralize the internal divisions within its own party. Likewise, in Finland, it appears that the decision to hold a referendum on the EU was made by the centrist prime minister with the twofold aim of ensuring the survival of his government and finding a decision-making procedure acceptable to a divided party. The use of the referendum as a palliative for electoral risks constituted the principal motivation in the Austrian and Swedish referendums on nuclear power, as well as of the Swedish referendum on membership of the European Union in 1994. In these three cases, it was decided that

56

Laurence Morel

the referendum was to be held because of forthcoming elections, with the aim of removing the divisive issue from the electoral campaign and preventing it from doing any harm. As early as 1991, it was a consideration of this kind which prevailed in the decision of the Swedish Social Democrat prime minister to hold a referendum on the entry of his country into the EU. The deciding factor seems to have been the pressures of other parties which were threatening to organize a minority parliamentary initiative referendum on the revision of the Constitution made necessary by the Maastricht Treaty. Those referendums initiated by a parliamentary minority must be held at the same time as the elections, a possibility which frightened the Social Democrat prime minister who was not at all certain about the unity of his party on the question. He was thus forced to act pre-emptively, to ensure that the issue of the EU did not divide his party during the subsequent elections. Similarly, in the long discussions between the parties on the date of the referendum, the PM did everything to prevent the referendum from being held on the same day as the elections. The Swedish Social Democrats had had a similar experience with nuclear power at the end of the 1970s. In the 1976 election, they had done poorly in part because they allowed divisions within the party over nuclear power to spill over to the electoral arena. Their decision in 1980 to initiate a referendum on the issue, to be held after the forthcoming election, was motivated by their wish to avoid repeating this mistake, especially since the internal opposition within the party had intensified after the Harrisburg accident in the US. The three parties of the right also agreed easily on holding this referendum, as they were also willing to avoid the nuclear issue during the election campaign, with the aim of not wasting their chances to form a coalition after the vote (the coalition that they had formed in September 1976 had broken down in October 1978 owing to the Centrists’ opposition to the national nuclear energy program). The Austrian Socialists may have likewise learned from the Swedish experience. The head of Austria’s socialist government of the day, Bruno Kreisky, was cognizant of the unhappy experience of the Swedish Social Democrats in 1976 and thus chose to use a referendum in 1978 to address the question of nuclear power and manage the divisions within his own party. This effectively decoupled this controversial issue from the upcoming elections of 1979. The legislative referendum In the 22 referendums studied, the legislative goals behind governmentinitiated referendums dominated the field before 1970 (5 out of 9), but


declined in importance thereafter (3 out of 13). And whereas prior to 1970 the questions raised by these referendums were varied, afterwards they concentrated solely on territorial and sovereignty questions. But in all such cases it was a head of government (both ministerial or presidential) who resorted to the referendum to promote a specific piece of legislation which would not have passed through the normal parliamentary process. Three different situations have explained the need for such a referendum: the existence of a minority government (France, 1961, April 1962, October 1962 and 1988; Denmark, 1986); the existence of a majority government divided over a major question, with the prime minister being unable to pass the preferred policy through normal parliamentary channels (Belgium, 1950; Sweden, 1957); and third, a procedure for adopting legislation which required unanimity in a federal state (Canada, 1992). These explanations should not be confused with the ‘division resolving referendums’ discussed earlier, where the principal motivation for recourse to the referendum was the strategic decision to manage a potentially difficult electoral issue that risked doing damage to the party in power. In the cases under discussion in this section, the principal motivation for turning to a referendum was to get a particular piece of legislation passed that the government could not pass – due to its minority situation, divisions between the executive and the legislature, or the requirement for unanimity – without the use of the referendum. Although frequently forgotten, three of the five referendums organized in France by General de Gaulle were instances in which the referendum was used as a weapon by a minority government. Conventional wisdom has defined ‘Gaullist referendums’ as those whose motive is to attain a vote of confidence in favor of the president of the republic. But in the three cases of interest here, this plebiscitary function was only secondary. Instead, one must recognize that until 1962 there was no presidential majority in parliament, making the pursuit of legislative actions by the president very difficult. It is in fact the referendum of October 1962 that puts an end to this situation by acting as the catalyst for the direct election of the president of the republic. And indeed, the subsequent election produced a bipolarization in French political life, securing a solid majority for the elected president of the republic, and reducing the number of conflicts between parliament and the president. Later, in 1988, one would again be faced with a minority situation, and, in this case, Prime Minister Michel Rocard initiated a referendum on the future of New Caledonia to facilitate the approval of the Matignon Agreements. Although the right was not diametrically opposed to these agreements and would have likely eventually voted in their favor in

58

Laurence Morel

parliament, this would have slowed down the peace process. The referendum was therefore again used to advance the legislative objectives of a government that could not count on majority support in the legislature. This referendum, however, underlines how the identification of one principal objective for the calling of a referendum does not deny secondary reasons. In 1988, Prime Minister Rocard also called the referendum to add legitimacy to the agreement and to divide his opponents and cajole some parties of the centre to rejoin the minority government. Much like the referendum of 1961 on Algeria, domestic political forces and the creation of legitimacy were contributing factors to the decision to hold a referendum. In Denmark, Poul Schluter’s minority government used the referendum in 1986 to attempt to secure passage of the Single European Act after it was rejected by parliament. Although some have seen the motivation for this referendum as tactical, designed to weaken the Social Democrats who have been traditionally divided over European integration, such a reading confuses one possible outcome – the strategic use of the referendum – with the referendum’s primary purpose: to secure passage of legislation that could not otherwise be passed. In both Belgium in 1950 and Sweden in 1957, the governing majority was reduced to minority status on controversial issues. The governments chose to use the referendum to rectify these situations. In Belgium, the return of King Leopold III became a source of political struggle during parliamentary debate in the months preceding the election of February 1946. Parliamentary deadlock on the question persisted for five years. The Social-Christian Party (PCS) had begun calling for a referendum on the question as early as 1945. The opponents to the return of the king – the Liberals, the Socialists, and the Communists – were then in the majority in parliament and it appeared unlikely that the results of an election could succeed in reversing this alignment. Nonetheless, following the election of June 1949, when the PCS fell but two votes short of a parliamentary majority, it succeeded in convincing the Liberals to approve a bill establishing a referendum vote for March 1950, thus turning to the referendum to attempt to achieve its legislative ends. The principal motivation for the governing Social Democratic Party in Sweden in 1957 to hold a referendum on its pension proposal was to advance its legislative agenda. As in the 1955 referendum over ‘driving on the right,’ the coalition between the Social Democrat and Agrarian parties was divided, and the conflict on the pension scheme had become one of the central political questions of the day. The referendum was first called for by two opposition parties (the Conservative Party


and the People’s Party) and the proposal was rejected by the Social Democrats. It was only in mid-1957 that the Social Democrat prime minister, Tagus Erlander, suddenly changed his mind and decided to organize a referendum. There were several reasons for this, including a desire to manage conflict within the coalition and a desire to prevent the establishment of ‘minority initiated referendums.’ However, the most important reason was that the Social Democrats could not secure passage of their preferred legislation through ordinary parliamentary processes because they were in the minority on the question. By turning to a referendum, the Social Democrats increased the chances for passage of their proposal by permitting them more time to make their case to the public. In Canada, the final decision to organize a national referendum on the Charlottetown Accord (a constitutional package designed to meet the demands for greater autonomy for the French-speaking province of Quebec and for the aboriginal populations) was taken at the beginning of September 1992 by federal Prime Minister Brian Mulroney. Parliament had already adopted in June an act allowing for the possibility of a referendum. This decision was related to the necessity for unanimous approval of the Charlottetown Accord by the federal parliament and the legislatures of the ten provinces in order for it to become law, and referendums which had already been scheduled in Quebec and other provinces. Within this context, the only way the federal government could pass its package of constitutional changes was to turn to a national referendum. The plebiscitary referendum This plebiscitary function of the referendum has characterized only the French referendums initiated by the president. This is not to say that such referendums are an exclusively French phenomenon. A vote of confidence was clearly exploited in other countries by certain referendum initiators in the hope of using the anticipated positive results as a springboard to imminent electoral success. Obviously this was very much a part of Bruno Kreisky’s aspirations in 1978 and those of Felipe Gonzalez in 1986. But in these cases, these motivations were secondary. In France, on the other hand, this has often been the primary motivation, in part facilitated by institutional factors. For example, the president can practically start a referendum alone, in so far as he needs only the agreement of the prime minister, and the latter, appointed by the president, is generally accountable to him (except for periods of cohabitation). As well, the direct election of the president, coupled with the

60

Laurence Morel

length of the mandate – seven years – together often compel the president to seek some kind of endorsement at mid-mandate to reinforce his own position. Accordingly, three of the four French referendums whose principal function was plebiscitary were held mid-mandate (1969, 1972, and 1992). The only exception was the referendum of September 1958, the obvious function of which was to provide some kind of popular legitimation for de Gaulle’s assumption of power. Moreover, in the French case, cultural and historical factors have also played a part in encouraging the plebiscitary motivations of the president, above all the existence of a plebiscitary Bonapartist tradition resurrected by de Gaulle. De Gaulle’s long stay in power concluded in April 1969 as it began: with a vote of confidence by referendum, which in this case he lost. This referendum, held a year after the crisis of May 1968, was sometimes interpreted as political suicide, with the overwhelming victory of the right during the June 1968 elections having already shown the confidence the nation had in the President. However, de Gaulle worried that this apparent confidence may have been directed more toward his prime minister, as the leader of the majority, than himself, and thus looked to a referendum victory to reinforce his position. The plebiscitary function of the referendum, which was explicit with de Gaulle, became an implicit aim of his successors. Pompidou and Mitterrand carefully avoided calling for direct votes of confidence, seeking rather to strengthen their position indirectly through popular approval of changes that they had proposed. Yet there is no doubt that both used the referendum to improve their personal position. A secondary function of the referendums which they sponsored was to divide the opposition: the emerging alliance between Communists and Socialists in 1972, which led to their Common Programme of government, and that between the UDF and the RPR in 1992. The politically obligatory legitimation referendum Governments often claim that the reason why they organize a referendum is the necessity to grant a special legitimacy to a decision, but this is in fact rarely the case. When one examines more carefully the reasons behind the holding of these referendums, one finds other motivations at the fore. This does not mean that referendums may not have a strong legitimating impact on some policies. In fact they often do, and sometimes even help to resolve deep conflicts by rendering a decision more acceptable to the minority (like in the case of the French referendums on Algeria and New Caledonia). But the ‘official’ legitimation function


often has little to do with the true motivation, which is generally of a more political kind. In fact, only a small number of referendums fall into the category of legitimation referendums, and it should be added that these referendums have all been ‘politically obligatory,’ in other words de facto obligatory. That means that governments could not have afforded not to consult the people on the matters put to the referendums. Their decisions would have been viewed as lacking legitimacy. With reference to our previous discussion regarding the formal versus real initiative, we have clearly here a case of governments which can be said to have been only the formal initiators of the referendums. In Canada, the politically obligatory nature of the 1942 referendum on military conscription resulted from the electoral pledge made by the leader of the Liberal government that he would not resort to conscription for overseas service in case of re-election. As the war continued and the need for more soldiers became dramatic, the government thought that a way to escape the deadlock would be to hold a referendum asking the Canadian people to release the government from ‘any obligation arising out of any past commitments restricting the methods of raising men for military service.’ In Norway, it was the existence of a precedent, namely the 1972 referendum, which supplied the need for the 1994 referendum on the Maastricht Agreement. There were of course other motivations as well. The Labour Party wanted this referendum because of its strong internal divisions on the question of membership, and also because the referendum was the only hope of creating the threequarters parliamentary majority necessary to ratify the agreement. However, the precedent had been established and a conventional understanding had developed that to enter the EU would require another referendum. In order to better illustrate the politically obligatory referendum based on precedents, one could invoke a number of substate referendums held in Western Europe. In France, the organization of a referendum on independence in New Caledonia before the end of 1998 was one of the provisions contained in the Matignon Agreements, approved by the referendum held in the whole of France earlier in the year. The referendum of November 1998 was therefore made obligatory by the result of the earlier referendum, although certain room for maneuver evidently existed (as testified by replacing the question of independence with the proposal to postpone the independence referendum). Nonetheless, avoiding a referendum would have violated the Matignon Agreements, and the reason why a referendum was made part of these agreements was clearly related to the precedents of the earlier referendums in New

62

Laurence Morel

Caledonia in 1958 and 1987 on the maintenance of this territory as a part of France. Similarly, the referendums organized in parts of the United Kingdom in 1997 and 1998 had as their principal explanation the existence of precedents. 1 With regard to the 1997 devolution referendums in Scotland and Wales, it is clear that it was impossible, after having consulted the population on the same issue in 1979, not to consult it a second time – a fortiori since this would have implied defying the will expressed at previous referendums. As well, the 1998 referendum on the Good Friday peace agreement in Northern Ireland can be seen as necessary due to the precedent set by the so-called Border Poll in the province in 1973. The fact that the authorities at the time planned to repeat the referendum every ten years, coupled with the constitutionally obligatory character of the referendum in the Irish republic, made a popular vote in 1998 inevitable. There were of course other motivations in both cases: Tony Blair’s search for a plebiscitary electoral boost and the nationalist and unionist parties’ hope that the referendum could be a way out of internal strife. Nonetheless, the precedents had been set and could not be avoided, even if one had wanted to.

Conclusion The first goal of the study was to explain the increased use of referendums in consolidated democracies. First, it is important to remember that the increased frequency of referendums has come in large part due to the increase in citizen-sponsored initiatives in Italy and obligatory referendums in Ireland and, to a lesser extent, Denmark. As well, the number of referendums in Switzerland increased substantially in the 1970–98 period when compared to the 1940–69 period. But my primary focus was on government-initiated referendums and the explanation for the increase in their frequency since the beginning of the 1970s. The evidence suggests that there has been an eruption of new political issues which cut across traditional cleavages, such as nuclear energy and, especially, European integration. These questions divide parties and their electorates, and have given rise to an increase of referendums initiated by governments with the goal of neutralizing such divisions. By turning over the question to the electorate, governing parties have hoped to manage their own internal party divisions and, at times, exploit divisions within the opposition. At other times, parties have chosen to hold referendums due to their need to have their legislative agenda passed, despite being in the minority on the key issues.


It is not necessarily clear that the growth in the use of governmentinitiated referendums will continue. Although the number of cross-cutting issues will probably continue to increase for various reasons – such as the decline of ideology, the increased complexity of the social structure, and the rise of the more independent voter – referendums have not fulfilled with great efficiency the objectives set out for them by the initiating parties. They have often put governing parties in awkward positions: on some referendums they have to side with their opponents, on others the division within their own party is not resolved by turning the issue over to the electorate, and, perhaps most fundamentally, the outcome of the referendum is increasingly difficult to predict or control, as evidenced by the referendums on the Maastricht Treaty. Moreover, the more frequent use of referendums could further convince voters that parties are not necessary and therefore further weaken the ties between voters and parties. It therefore seems unlikely that the use of government-initiated referendums will grow because it does not appear to be in governments’ interest to do so: simply because they initiate a referendum does not mean that they can control its outcome or effects, something which would make any government wary of pursuing that path. The only way one could see a significant increase in the number of government-initiated referendums is with a growth in the number of de facto obligatory referendums created by precedents. However, even here one should not overstate the case: although it is unlikely that Denmark will accept significant changes to its status in the EU without a referendum, France, Sweden, and Finland all avoided referendums on the Amsterdam Treaty despite referendums on Maastricht. Therefore, although some precedents may have been established in some countries on some issues, governments can still usually avoid referendums when they so choose. We have concluded that there is not an ineluctable trend towards an increased use of the government-initiated referendum. Should we also conclude that this is a good thing for democracy, because when governments choose to initiate a referendum, it tends to be for instrumental and strategic purposes rather than a genuine interest in consultation? Our study shows that things cannot be put in such stark terms, as governments can have various motivations for calling referendums, not all of which produce biased practices. In particular, referendums used to promote legislation or because they appear to be ‘politically obligatory,’ which constitute almost half of the referendums studied, are conducive to good democratic practice. Even the referendums intended to resolve the internal divisions within parties or coalitions are not inherently

64

Laurence Morel

anti-democratic. So, while it remains true that governments have far more control over the referendum than other actors, their motivations for calling referendums do not unequivocally indicate that they must necessarily manipulate them. Rather, what the analysis of these motivations clearly shows is that the question of who initiates a referendum – though a good predictor of its practice – cannot on its own be used to develop a typology that predicts whether the referendum will follow good democratic practices.

Note 1 The function of the earlier devolution referendums, held in 1979, can be described as legislative. Indeed, the devolution proposals of the Labour minority government probably would not have been approved by the House of Commons if the promise of the referendum had not existed.

Part II The Impact of Referendums on Liberal-Democratic Institutions

3 Political Parties in Direct Democracy Ian Budge

Democracy today is essentially party democracy; that is, it is constituted by the competition between parties to gain votes and form governments. Though parties run candidates for office, the latter get elected less on their individual appeal than on the party endorsement. At the same time parties put forward more or less integrated sets of policy priorities as the program which, if elected, their nominees will carry through. This not only provides a necessary degree of policy coordination in government, it also enables parties to be held responsible for program implementation at the next election. Because electors presumably vote for the policies they most prefer, parties serve as crucial intermediaries in the (rough) translation of majority wishes into government policies – a central claim of democracy. By bringing candidates for legislative and executive office into the same organization and disciplining them to ensure their united action and common responsibility, modern parties have subverted traditional theories of individual representation and constitutional doctrines of the separation of powers. Nevertheless, they still operate within the forms and institutions of representative democracy, even if its conventions have changed. The question we address here is what difference will direct democracy, with its focus on deciding policy through direct votes of the people, make to parties which have traditionally played an autonomous role in defining and brokering such policies? Our concern to answer this question is sharpened by the hostility shown by many proponents of direct democracy towards political parties. Rousseau’s (1762/1973) distrust of ‘partial interests’ corrupting the General Will is well known. Similar suspicions motivated the Progressives who instituted popular referendums and initiatives in many American states during the early years of the twentieth century. They 67

68

Ian Budge

intended their reforms to undermine the hold of corrupt political machines, and discouraged or explicitly prohibited party participation. Reformers were not unhappy to transfer approval of legislation from elected, party-dominated legislative assemblies to the sovereign people. Modern-day libertarians active in promoting popular initiatives against ‘big government’ also see parties as part of the problem. Against such a background, direct democracy can indeed seem inimical to the traditional deliberative institutions of representative democracy – especially to political parties and parliaments. Ironically, many of the accusations levelled by progressive reformers against the early twentieth-century parties – that they stifled free debate and discussion, hindered due deliberation of issues, and left the door wide open for selfish interests to covertly dominate policy – are now levelled by defenders of (party-dominated) representative democracy against referendums and initiatives! The fact that these arguments have been reversed is of course no reason for not taking them seriously. This is particularly so because they have been taken up by such respected democratic theorists as Sartori (1987) and analysts of referendums and initiatives themselves (Magleby, 1984; Kobach, 1994).

Perceived threats from popular involvement The main ways in which direct policy voting is seen to undermine parties are as follows: (a) It takes control of the political agenda out of their hands. In particular, decisions are forced on issues neither raised nor ‘framed’ by them. In a related way, the accountability of parties may decline. Parties’ strength is based in part on the fact that they can be held accountable for policy successes and failures; if government policies are not under their direct control, parties may become less accountable stewards of public policy and hence less relevant. As a result, popular votes reduce the consistency of the policy program which is implemented; it may not be adopted as a complete package with a coherent philosophy, since individual items can be voted on separately at the behest of interest groups. (b) Single issue groups – which unlike parties do not have to seek accommodations among the competing interests of their own supporters – are more extreme and less compromising than political parties, so popular consultations encourage the taking of extreme

Political Parties in Direct Democracy

69

positions and reduce the possibility of negotiation and agreement (Magleby, 1984, pp. 188–90; Cronin, 1989, p. 248). Popular policy voting thus undermines the deliberative processes that have been so much stressed in recent democratic theory, almost as a defining characteristic of the democratic system itself (Elster, 1989; Bohman and Rehg, 1997). (c) Party leaders may take opposing sides in popular consultations and electors may ignore or even defy the party lead where it is given. Direct democracy thus weakens parties by promoting internal splits and schisms (Kobach, 1994, p. 132). (d) As a result of being less focused on the party as a unique policy vehicle, activists and political entrepreneurs shift their attention to interest groups outside parties on which they base personal organizations. This tendency contributes to a ‘hollowing-out’ of parties into ‘shell organizations’ within which candidate-centered factions may retain a common label but share little else (Pomper, 1980; Wattenberg, 1990). This is a common observation made about contemporary US parties, seen at its most extreme in the western US states and symbolized most famously by California. Of course, to the Californian example we can contrast the European cases of Switzerland and Italy. In Switzerland popular policy voting has coexisted happily with parties and legislatures over the past 150 years. Heavy popular involvement may be said to have existed for as long as representative institutions themselves. From the institution of the modern Federation around 1870, mandatory referendums and popular initiatives were prominent features of the constitution – not only at the federal but also at cantonal and local levels. They were used by excluded parties – first Catholics and then Socialists – to bargain for a share of political power. This process culminated in the establishment of the present semi-permanent coalition of the four leading parties at the federal level, with cabinet seats shared between them by the ‘magic formula’ of 2–2–2–1, where the three largest parties receive two cabinet seats each, and the fourth party receives one. This is a situation which looks extremely odd to Anglo-Saxon eyes. And the fact that all important legislation and treaties have to be approved – or can be challenged – by popular vote does take much of the heat out of legislative and partisan debate. But, however idiosyncratic the arrangements, nobody could deny that Switzerland is a well-governed country, in spite (or because) of being the closest thing we have to a fully fledged direct democracy. A point that is often overlooked is that it is also a party

70

Ian Budge Idealized Actual Western US France Italy Athenian Athenian states Quebec democracy democracy (California) Switzerland Australia UK Totally

Totally

unmediated by parties

mediated by parties

Figure 3.1 Variations in party interventions in popular policy voting between various systems of direct democracy. Notes: Idealized Athenian democracy is contrasted with actual practice, since the political clubs to which many (male) Athenians belonged seem to have served as proto-parties (Bonner, 1967, pp. 45, 61). California is the most famous modern example of popular voting taking place without party intervention. This is in contrast to Switzerland, Italy and particularly democracies with occasional referendums, often prompted and controlled by parties.

democracy where parties clearly dominate government and constitute the central participants in political debate. The other counter-example to the presumed negative effects of direct popular involvement is Italy, where popular interventions from the 1970s onward seem if anything to have strengthened parties rather than undermined them. Here the peculiar form of the abrogative referendum has been used since the 1970s to pass proposals which would otherwise have been blocked in parliament. These initiatives have been promoted by the parties excluded from government up to 1992 – primarily the Communists and Radicals. Unusually, parties split in such referendums less than they did in parliament. The general judgment on these initiatives has been favorable: they updated the system, notably with regard to social matters during the 1970s, alleviating popular frustrations and enabling the main parties to stagger on until they collapsed under the weight of their own internal corruption in the early 1990s, which had nothing to do with the initiative. When they did collapse, popular voting allowed reasonable electoral reforms to be instituted with a fair level of general agreement. Italy thus constitutes a strong case for claiming that under certain circumstances popular involvement actually strengthens parties and legislatures rather than weakens them. The examples of Switzerland and Italy serve to focus our attention on a general theoretical point. There are clearly different forms of direct democracy on offer. One is the populist, unmediated form to which California seems the closest approximation in practice. The other is the more organized and mediated form seen particularly in Switzerland but also in most countries where popular consultations take place. Here parties have a full role to play. Figure 3.1 illustrates the wide range of


71

forms direct democracy can take, depending on the institutional arrangements and rules made for it. It shows that most existing forms incline to the mediated and brokered Swiss arrangements. Populist forms of direct democracy exist mostly in critics’ minds – a point underlined by the contrast between the idealized and actual Athenian democracy shown in the figure. However, critics concerned about the negative effects on parties tend to focus on the unmediated forms, as that suits their argument best. A fair assessment, however, must also take the mediated forms into account.

Defining terms of discussion To examine the extent to which the criticisms made in Part I regarding referendums do in fact act on parties, we need to take them one by one and consider arguments for and against them. In order to conduct a reasoned review, however, we need first to consider what kinds of evidence are available and relevant to assessing the criticisms. This is because objections to direct democracy are commonly formulated at two levels. One is what actually occurs when referendums or initiatives are held on policy matters (point (c), for example, where Kobach (1994) argues that exactly this happens in Switzerland – parties do often split during popular consultations). But other points are more abstract in nature: for example, point (a) concludes that having separate votes on each policy must of necessity make the government program less coherent. Different kinds of evidence are required to consider these different levels of argument. Where criticism is based on a purely factual assertion one can cite factual evidence without any problem. For example, Kobach’s description of the internal dissension within parties at the leadership and electoral level produced by popular consultation is contradicted by other authors (Kriesi, 1994, pp. 71–3) who see Swiss electors following a clear party lead when this is consistent with basic party ideology. Theoretical and logical arguments about the effects of direct democracy are relatively impervious to simple factual observations (it may be very difficult to conclude with rigorous evidence that party platforms have become ‘less coherent’); they assert that certain consequences for parties must follow any extension of direct democracy, given its nature. (For example, voting on individual policies must necessarily disrupt overall government programs in a Westminster system, as it either inserts new policies into them or raises the possibility that integral parts of the program will be defeated.)

72

Ian Budge

These more abstract criticisms do, however, rest on certain assumptions, sometimes of a definitional nature. Crucial ones here are that government programs are fairly tight and cohesive in the absence of direct policy voting, and that parties form passive registers of popular decisions rather than being active shapers of them. Both assumptions are questionable. To see why we need to consider the nature of popular consultations and political parties within a comparative context. Who sponsors the vote? Policy consultations in the contemporary world are commonly distinguished by the terms ‘referendum’ and ‘initiative.’ The difference between them is that referendums are commonly called by governments to decide some matter by a popular vote. Initiatives, on the other hand, can be triggered by a popular petition with a requisite number of signatures. Initiatives of this kind, which exist in 27 of the US States, Switzerland (at local, cantonal and federal level), one Canadian province, and in Italy,1 are usually taken as the purer form of popular policy voting, as they are free to take place on any issue, not simply those dictated by party governments. So the question of who initiates a popular consultation is quite crucial in assessing its effects on parties. Government-sponsored referendums, which are often held to bridge over internal disagreements, may help parties stay united in the long term, while initiatives have more potential to split them open. We should note qualifications on this simple distinction. Many referendums are constitutionally mandated on certain issues (particularly for changes in the constitution itself). Hence, they cannot simply be called at the behest of the government but must be held regardless of whether it is politically convenient to do so (though, of course, the government may well have raised the issue calling for a referendum in the first place). The important distinction between popular consultations may therefore be whether they are ‘controlled’ in some sense by governments or whether they take place at a time not chosen by them (Smith, 1976). Some authors regard most referendums as being strongly influenced if not wholly ‘controlled’ by governments (Lijphart, 1984, p. 203; Butler and Ranney, 1978, p. 222). Like many other assertions in this field such a view has been controverted (Qvortrup, 1999, p. 180). However, it does seem plausible to say that on a continuum of control, referendums called at will by a government are diametrically opposed to initiatives wholly beyond their control, with other kinds of referendum falling in between. Initiatives also differ depending on whether parties actively participate or not. The distinctions are shown schemat-

Referendums and initiatives contrasted in terms of government control.

Figure 3.2

(e.g. Italian referendum on divorce 1974)

(e.g. Irish referendum on single European Act 1987)

(e.g. French referendum on direct election of President 1962)

(e.g. UK referendum on membership of EU 1975)

CONTROL

HIGH

(e.g. Californian Proposition 13)

Initiative promoted Initiative prompted by citizens’ petition by citizen petition, stimulated by some minimum control by parties parties

Referendum strictly mandated by constitution, requirement interpreted by courts

Referendum required by constitution but subject to government interpretation

Referendum called when and how government wants

CONTROL

MINIMUM

73

74

Ian Budge

ically in Figure 3.2. The consequences for parties of each type of popular vote therefore may well differ. Rules and regulations of the vote The consequences for parties will also differ depending on the rules under which campaigning and voting on a policy proposal take place. The highly regulated framework for referendums in Quebec, for example, differs very much from the minimally regulated initiatives in California. In Quebec groups for and against a proposal have not only to declare themselves and accept limits on the money they contribute but have also to register themselves as members of an umbrella committee for and against a proposal (Coté, 1999). In California there is no obligation to declare oneself, no limit on financing – and far from registering an interest, new groupings can be invented to confuse the issue (‘California Democrats for Life,’ for example, when the actual Democratic Party is ‘pro-choice’ on abortion) (Schrag, 1999). Political culture The variety of regulations, all with clear consequences for the way parties can operate, also points to the importance of the countries whose experience we draw upon in coming to conclusions about the effects of direct democracy. As our contrast between California and Switzerland has shown, we cannot take the experience of one place at one time and use it to draw definitive generalizations. On the contrary we should draw on as wide a range of cases as possible. One must keep in mind that the nature of parties is affected by the overall political context of the country. Certain qualities may be used as evidence for party decline due to the presence of referendums, when such characteristics are taken as normal or even praised within another system. Take factionalism, for example. It is normal in Italy or Japan for parties to have several organized factions, often with a secretariat and recognized leader. This does not make the party a ‘shell,’ as all the factions subscribe to the common ideology albeit with different emphases, and manage to hammer out an agreed program at elections. Having different factions appeals to a wider range of voters. We should bear this in mind when assessing criticism that the US parties are being ‘hollowed out’ by factional tendencies. Furthermore, even if they are being ‘hollowed out,’ this may well be due to other features of American politics than popular voting, which, after all, takes place in only half the states.


75

Political and social change more generally In so doing, we should try to clarify whether changes in parties can simply be attributed to the impact of direct democracy as such, or to a wider set of factors among which popular consultation is only one, and not necessarily the most important. Long-standing features of American politics such as the separation of powers, as well as the rise of the media, may have much more to do with the ways parties operate – particularly if the same tendencies are visible in states with and without the initiative (Budge, 1996, p. 122). Whether negotiated compromises between two legislative houses and an independent executive favor programmatic coherence any more than popular voting is debatable. The effect on the party system as opposed to large parties alone Which parties will take part is also an important question. Effects on parties are often discussed with exclusive regard to established ‘governmental’ parties – it is their legislative negotiations and compromises that are seen as being undermined by populist initiatives. There is a lot of evidence from many countries in Europe that small parties, excluded from legislative politics, have promoted popular votes and campaigned in them precisely so as to draw attention to themselves and get established. The Republican comeback in many American states was also associated with such behavior. In assessing the effects of direct democracy on parties let us therefore think of the party system as a whole rather than just large parties. Their stranglehold on power may be weakened but this could contribute to a strengthening of the system as a whole in terms of policy competition.

From criticisms to facts: does direct democracy weaken political parties? Having cleared away some of the underbrush we can now better assess the main criticisms made of the impact of direct democracy on political parties. (a)

The first was the extent to which agenda control is transferred from political parties to irresponsible interest groups, making governments less accountable. Interest groups cannot be held accountable for the consequences of the decisions they campaign for and ‘front’ organizations often fade away after the campaign is over. The considerations put forward above, however, made it clear that such a criticism is not directed at policy consultations as such but at relatively unregulated ones. These do not even extend to all

76

Ian Budge

the initiatives called by popular petition. For in Switzerland and Italy political parties – generally out of government – play a great part in organizing and promoting such petitions. Only in the western states of the US would the description of non-party irruptions into the policy process hold true. Yet even there the question of whether Democrats or Republicans take the governorship or capture the legislature is important for policy making, and these parties actively use the initiative to advance their policy agenda (Gerber and Hug, this volume). If we turn to government-sponsored referendums rather than initiatives it is even more accurate to say that parties are heavily involved. They either decide to call the referendum where this is optional or raise the issue which requires a referendum in the first place. Once we put the question of the agenda in comparative perspective therefore it becomes clear that parties retain considerable influence. (b) Does voting on issues one by one really undermine the consistency of party and government programs? As pointed out above, the force of this argument depends heavily on the idea that such programs are very coherent and detailed in the first place. Actual analyses of their content tend rather to conclude that they are loose groupings of priorities favored by the parties, put together contingently rather than from logical necessity (Robertson, 1976; Budge, Robertson, and Hearl, 1987). They are thus relatively easily adapted to the outcomes of popular voting. Thus, if a tax-cutting or limiting measure is carried, a left-wing government can always safeguard welfare programs by borrowing money or cutting military expenditure. (c) What about the accusation that any popular policy debate inevitably simplifies issues and by doing so leaves less room for party compromise and deliberation? It is certainly true that more people get the opportunity to deliberate than if the issue is confined to the legislative chamber. But popular discussion as we have seen can also be guided by the parties. Anybody claiming that the quality of popular debate is necessarily less ‘deliberative’ than legislative debate has only, however, to look at the records of parliamentary assemblies. Particularly where a one-party government and opposition confront each other the quality of debate is low and deliberation and compromise nonexistent. Where compromises operate more powerfully, inside coalition governments or under the separation of powers, they have the same consequences for the coherence of the government program for which popular consultations were criticized under the


77

previous point – they undermine any coherence they have. Rather than being convinced by argument, most parliamentarians vote under strict party instructions whether convinced by them or not – deliberation does not enter in. The criticism may, however, relate less to external debate than to deliberation within parties, and to their ability to forge compromises among their own membership. These may be placed under strain by single-issue votes which give members no incentive to ‘back down’ from their positions. A first point here is that referendums are often called by parties to bridge over matters on which they cannot agree internally. In this context a single-issue vote is a strategy for creating a compromise which the party cannot achieve by itself. A second point is that some parties in some countries (Italy being the obvious case) are more united in the popular campaign than they are in parliament. The rare cases of direct party involvement in state initiatives in America have been where they felt their central policies were being affected and they were fairly united (Democratic opposition to tax limitation in Michigan for example). Of course, their nonparticipation on many other initiatives may indicate an inability to agree on a compromise. Examples could be multiplied. But it is probably inadvisable at this point to make an omnibus statement about the effects of single-issue voting on deliberation within parties. What we can say is that it depends a lot on context, the position of the party, and the nature of the issue. Clearly these deserve systematic comparative investigation before we can be definitive about them. (d) Does party discipline not break down generally, however, where popular consultations are called? This is a point particularly stressed by Kobach (1994, p. 132) with regard to Swiss parties. In response, we note that Swiss parties are not permanently split by these controversies. On the contrary they have survived such situations for over a hundred years, with their ideological and social identities as intact as parties in other countries. It is unrealistic to set up a universal ideal of a cohesive, disciplined party, in which any dissent is a sign of weakness and decline. On the contrary, some of the most successful parties in the world are highly factionalized – one need think only of the Italian Christian Democrats from 1948 to 1992, the Japanese Liberal Democrats, or the American Democrats and Republicans. Furthermore, referendums and initiatives have given opportunities for smaller parties to publicize and establish themselves, in

78

Ian Budge

Switzerland, for example. It is established parties which divide internally, while small opposition ones remain united. This happened with some of the current government parties (Catholics and Socialists in particular) but also with ecological and nationalist movements in the last quarter-century. This last point has a general relevance. Much of the literature on party decline focuses on large, established parties. The loss of their members and voters is usually taken as indicating a general party decline – overlooking the fact that votes and activists are generally lost to small rising parties. We must distinguish between a weakening of incumbent parties, and the decline of all parties. Because incumbent parties usually lose popularity to opposition parties, we can hardly speak of universal party decline. (e) Party factionalism and the creation of personal organizations inside a party ‘shell’ are certainly more likely to occur when internal splits emerge. Where opposition parties seize on popular voting to provoke splits in established parties, factionalism inside the latter may ensue. Yet this is hardly an automatic consequence of policy consultations as such. Quite as often a referendum is called by the ruling party because it is already split over the issue. By holding a consultation it can ‘agree to disagree’ in this limited area while preserving its overall unity (Butler and Ranney, 1994, pp. 259–60). Again, therefore, the effects of popular consultations may depend on the conditions and rules under which they are held, rather than on their inherent nature. Factionalism itself may be a perfectly viable way for a party (or party system) to adapt to political circumstances and is not necessarily a sign of decay.

Referendums and party decline: a causal relationship? On closer examination therefore most of the adverse consequences anticipated for parties from the extension of popular policy voting are subject to strong counter-arguments. Can we go beyond this to a systematic survey of the observable effects that direct democracy has on parties? Beyond Switzerland and the US states, answers are difficult to provide because of the infrequency of consultations and because it may be impossible to disentangle the influence of popular interventions from concurrent developments such as the rise of media-based politics. Still more difficult is tracing the effects on political parties – not because of a lack of research in this case but because of their extreme complexity as social and political entities. We can distinguish at least seven levels of party activity.


79

1. The top party leadership, whose concern is with strategy and immediate political problems, since they may form the government. Leaders may be reasonably united or permanently factionalized – both situations are normal and can continue for long periods. Even at the best of times there will be strong personal rivalries among individuals for leadership positions. 2. A parliamentary group linked to leaders, and itself possibly united or prone to split into factions. 3. A bureaucratic organization devoted to research, finances, campaign management, and so on. 4. Party activists and workers, usually formally enrolled in the party and organized on a territorial basis. 5. Specialist groups, women, workers, farmers, youth, and so on, who may have strong semi-autonomous organizations within the party. 6. Party voters, themselves falling into the two groups of regular supporters and those who vote more casually for the party. 7. Ideology set down in books, pamphlets, constitutions and other documents, which is to a greater or lesser extent shared by all the foregoing groupings.

In keeping with the general tendency among commentators to (over-) react to dramatic events, change at one of these levels is often identified with decline, while developments at the others are ignored. Thus, the perceived abandonment of ideology in the 1960s was often linked to a perceived decline of the party per se. The re-emergence of ideological divisions in the 1980s has often been interpreted as a victory of the right, while socialist parties are now seen as having abandoned their true inheritance. (For comparative analyses of programmes showing that, on the contrary, parties of the left have stayed just where they always were, see Klingemann, Hofferbert, Budge et al., 1994, passim.) Similarly, voting changes and, in particular, greater volatility between parties are interpreted as evidence of declining electoral loyalty to parties (Franklin, Mackie, and Valen, 1992), ignoring the fact that when voters switch they vote for other parties. Again, we see here the general tendency to identify the fate of parties with that of established parties, which is not necessarily the same thing. As von Beyme (1985, p. 305) remarks, much depends on the capacity of new parties to absorb the voters who move and on the capacity of the system to accommodate new parties. So far they have shown a marked ability to do this. These considerations underline the point that parties are complex multi-level

80

Ian Budge

organizations. This indeed is what makes them so unique and valuable, as they are the only institutions that tie governments into large popular groupings. Thus it is particularly dangerous to concentrate on only one aspect or level of their activities. Change often constitutes regrouping and necessary adaptation rather than irreversible decline. In considering parties therefore we cannot simply concentrate on any one level or any one indicator of vitality or health. In a judicious review, Mair (1992) sees much more evidence of change and adaptation than of decline. Indeed, as he remarks, we hardly have the data about all aspects of parties which are relevant to spot decline if it were actually occurring. The most obvious phenomenon is the continued dominance of parties over governments and over electoral competition, where the incidence of popular consultations hardly makes a dent in any democracy in the world (von Beyme, 1985, p. 372). Established parties may occasionally go down, but rivals old or new succeed them, often injecting party competition with new vigor. The best prognosis is that this will continue, even if the mass membership party is succeeded (but only in part) by organizations much more centered on the use of the media – of which the now looser-knit but more ideologically focused American parties may be the forerunners. Any attempt therefore to link referendums and initiatives to the strengthening or weakening of parties in general is fraught with difficulties. Still, the question is an interesting one and a general comparison serves at least to put it in context, even if the detailed results have to be viewed with some suspicion. Table 3.1 accordingly takes the major established post-war democracies, enumerates the popular consultations held in them from 1945 to 1999, and notes the standing of their parties and party systems in the latter part of the period. To the established national democracies it adds those internal entities where the effects of popular policy voting have often been discussed – the US states, plus Quebec. Table 3.1 is innovative in showing the position of parties not just where there is popular consultation but where there is none. If we want to see how consultation impacts on parties we need to consider what happens where it does not occur. While the number of referendums is a hard datum, judgments of party strength are extremely ‘soft’ for the reasons given above. Where authoritative writers have contended that parties are weaker, even if such a judgment is contested by others (including by the author) parties have been recorded as ‘possibly weaker.’ These judgments relate largely to established parties and to indicators such as voting support and loyalty, and leadership cohesion.


81

Table 3.1 Incidence of popular consultations and position of parties in 20 democracies and 51 selected territories: 1945–99 State

Number of referendums

Australia New Zealand Canada United States Japan Israel India Austria Belgium Denmark Finland France (Fifth Republic) Germany1 Ireland Italy 2 Netherlands Norway Sweden Switzerland United Kingdom1 Quebec American states with consultations (27) American states without consultations (23) 1 2

24 9 1 0 0 0 0 2 1 13 1 11 1 17 40 0 2 4 343 5 2 1103 0

Change in party strength Unchanged Unchanged Unchanged Possibly weaker Unchanged Possibly weaker Possibly weaker Unchanged Possibly weaker Unchanged Unchanged Unchanged Unchanged Unchanged Possibly stronger Unchanged Unchanged Unchanged Possibly weaker Unchanged Possibly stronger Possibly weaker Possibly weaker

Important regional referendums counted as national ones. Most existing Italian parties collapsed under accusations of corruption in 1992–93. They were succeeded, however, by bigger and arguably stronger parties under the second Republic from 1994 onwards.

Sources:

Butler and Ranney (1994); Qvortrup (1999), p. 180.

The overwhelming impression, based on these judgments, is that parties have remained as strong as ever, regardless of whether popular consultations are held or not. It is, however, true that areas with many consultations – Switzerland and some of the American states – have been judged by some commentators to have seen a weakening of parties. The counterbalancing point is that in several countries without direct policy voting, in particular in US states without it, parties do not seem to be doing too well either. There may be marginal differences between the strength of parties in the Eastern US (less popular consultations) and the West (much more). But commentators like Pomper

82

Ian Budge

Table 3.2 Existence of popular consultation related to strength of parties for 71 political systems With consultation Parties unchanged/stronger Parties possibly weaker

Without consultation

14 (13) 29 (2)

2 (2) 26 (3)

Note: Table 3.2 is based on the list in Table 3.1 – bracketed entries are for countries only.

Strength of parties

(1980) and Wattenberg (1990) have seen a general decline of Democrats and Republicans, not state-by-state variation. These tendencies emerge more clearly from Table 3.2, which summarizes the lists in Table 3.1 in two-by-two form. At the country level the most common situation is for there to be some form of popular consultation, with parties whose position is unchanged. Cases where parties have possibly become weaker are actually more prevalent, relatively speaking, among countries without any form of consultation at all. Adding the American states to the table changes the balance to one where parties are possibly weaker in the majority of all cases. But still, party weakening is relatively more evident among systems with no popular consultation at all. One might object that the important distinction is really between systems which either have no or very few consultations on the one hand, and relatively frequent ones on the other. Figure 3.3 roughly plots this relationship graphically, and shows a curvilinear relationship,

0

10

20

30

40

50

350

1100

Number of popular consultations

Figure 3.3 Frequency of popular consultations and strength of parties: postwar democracies and selected territories


83

with the non-consultative jurisdictions, together with countries like Belgium, pulling it down at one end and Switzerland and the other American states doing so at the other. Were the information not so tenuous, this might lead to the conclusion that parties are strengthened by a modicum of direct democracy, but not too much direct democracy; such a finding could cause concern for parties if states continue to increase their use of referendums. A more cautious and probably more correct inference is that the relative strength of the party system is not closely related to direct democracy at all but to the precise arrangements made for popular consultations and to other features of the political system.

How do variations in the forms of popular consultation affect parties? A more profitable question to pursue therefore is how variations in the forms of direct democracy affect parties. We have already seen (Figure 3.2) that different rules radically affect the extent to which governments (and hence established parties) can influence policy consultations. We can now suggest three further generalizations. First, where parties, by constitutional rule or convention, do not habitually take part in popular campaigning for or against policy proposals, they lay themselves open to the negative developments highlighted in the first section of this chapter on perceived threats from popular involvement. That is, control of the agenda slips out of their hands, they get landed with policies disruptive of their programs, and they cease to become the primary policy vehicle for candidates and activists, who may well therefore form factions or ally with groups outside the party to create their own campaign organizations. There is of course an exception to this general rule in that parties may actually strengthen themselves and prevent internal splits by ‘agreeing to disagree’ on a particular issue and ‘letting the people decide.’ But they cannot make this a general reaction. In the modern world parties either take stands on issues and programs or lose relevance. If they allow themselves to be passively affected by policy decisions which they have not tried to influence, they lose the cue-giving function which gives them their electoral standing and their dominance of government, which is what seems to have happened in some of the western American states. Yet parties rarely stand aloof, as the experience of most countries suggests. They both initiate and influence referendums, keeping parties at the center of policy making, where they need to remain.

84

Ian Budge

Second, it is tempting to conclude from this that referendums called by governments are the best kind of (controlled) direct democracy from the viewpoint of maintaining party viability. This still leaves parties in control of the agenda. Yet we must be careful to distinguish between different kinds of parties. Certainly, controlled referendums will favor the parties in government, but they may disfavor the opposition and possibly keep smaller parties impotent, particularly if, as in the referendums in Quebec on sovereignty, the leaders of the two major parties are legally designated as leaders of the two opposing committees for the referendum campaign. If we are concerned about party viability in general we have to recognize that constitutionally mandated referendums and initiatives called by popular petition will often give opposition parties the opportunity to rally round a cause of their own choosing and reinvigorate themselves. Different kinds of direct democracy have different consequences for different parties – considerations which a ‘constitutional engineer’ might want to balance in designing a political system. Third, it is likely that if popular consultations are heavily regulated, large established parties are less likely to be undermined. The rules may give them a privileged position and they certainly give those with more resources the opportunity to anticipate what is going to happen and prepare. Conversely, with fewer rules, small parties and non-party actors will find it easier to prompt consultations and campaign in them in a relatively uninhibited fashion. Even so, such a process does not represent a general threat to party viability and can often represent a renewal of the party system. Though a heavily regulated environment may work against the strengthening of new parties, a minimum of regulation seems necessary for small parties to flourish. A complete absence of a regulatory framework can provide well-financed interest groups with great power to control the agenda and the debate. Unless such groups are regulated in terms of spending limits and the declaration of their sources of funds, parties – both large and small – are likely to suffer and give way to interest groups. In most cases, such regulations exist, providing additional evidence for the contention that popular consultations do not systematically undermine parties. Figure 3.4 sums up most of the points made above by setting up a continuum between maximum and minimum regulations and placing both parties and countries on it. It demonstrates: (a) how practices of direct democracy vary between political systems, and (b) how they benefit different kinds of political parties or harm them. It accordingly underpins our earlier conclusion that there is no one effect of direct


MAXIMUM

Benefits Benefits Benefits government established small new parties opposition parties parties

REGULATION UK, Quebec

Italy

Harms all parties

85

MINIMUM

Switzerland California REGULATION

Figure 3.4 Distribution of direct democracy practices in terms of extensiveness of regulation and effects on political parties.

democracy on political parties: it depends on the nature of the system, the referendum regime, and the parties.

Functions of parties in direct democracy Concluding that direct democracy as such does not harm political parties is important because direct democracy, no more than any other type of modern democracy, cannot function without them. Even in a system which went further down the path of direct policy-voting than Switzerland, parties would be needed to run the government – even if matters of principle were decided by initiatives. As we know from the experience of representative democracy, the best way to select a government and to hold it responsible is by electoral choice between freely competing groups. Whether these had policy programs subject to later popular review and recall or whether they were selected for purely administrative or executive functions would depend on institutional arrangements within the particular democracy. But it is inconceivable that any could do wholly without party government. In terms of popular voting, party intervention in direct policy campaigns seems the best way to improve the quality of electoral choice. Party endorsements provide an easy way for voters to relate a particular issue to their broad stance over a range of political issues (Lupia, 1994; Lupia and Johnston, this volume; Jenssen and Listhaug, this volume). Although it simplifies a much more complex underlying reality, the left–right ideological continuum offers an orientation to a range of central policy questions, and parties – even American ones – position themselves consistently on this continuum (Klingemann, Hofferbert, Budge, et al., 1994, p. 40). Where the party position is known, electors can use it to relate the proposal to their own preferences. In Italy, for example, even though interest groups were present in initiative campaigns, parties successfully mobilized their supporters and publicized their position, hence neutralizing interest groups’ ability to get electors to vote for proposals which they ordinarily would reject (Uleri, 1994).

86

Ian Budge

One appealing way for parties to function in a full-blown direct democracy would be to have them run candidates in general elections, as happens now, to form a government. The parties would produce a general program whose plausibility and appeal would obviously be a factor in being elected. Policy-voting by citizens would take place between general elections – including on key elements of the program adopted by the governing party – but possibly under varying institutional safeguards, given that electing a government also involves considering its general program. Given the loose-knit nature of most party programs, the government could live with defeats on some of its measures. With too many it might have to threaten an election and seek a popular vote of confidence, as it does now with legislatures. There could be a moratorium on raising issues certified as forming part of the government program for a certain period of time after the general election. Or government positions could be protected by requiring absolute or qualified majorities for their rejection. Whatever institutional safeguards (or none) for the general program were adopted, the government would clearly be bound to give a lead to electors on the various measures up for consideration through referendum. By providing guidance of this kind, parties could sketch out an overall position consistent with their general program for government. They would thus improve the overall policy coherence of single-issue balloting by tying it in with the overall governmental program supported at the previous election. Electors are clearly predisposed to follow party cues, as the strong discrepancies between the success rates of initiatives with and without government sponsorship shows in both the US and Switzerland (Magleby, 1994, p. 93; Kriesi, 1994, p. 70; Linder, 1994, pp. 99–100). This guards against possible effects, such as a new dissatisfied popular majority voting down proposals previously passed by an earlier majority. There is no evidence of this actually happening anywhere, but party loyalties would be an additional safeguard against it. Parties can reveal the interests and groups involved in promoting initiatives. This is not simply a function of the more extended and informed debate that parties are likely to stimulate. Their participation means that groups and single-issue movements have to line up with parties, thus revealing their ideological preferences. Again, parties provide easily understood information to electors about the general implications of issues. Furthermore, where parties get engaged, their capacity to mobilize support and tap into their pre-existing, multi-purpose organization helps deprive well-financed interests of most of their advantages.


87

Thus there is no doubt about the institutional practicability of having party government, and party leads on single issues, under direct democracy. Nor, given the analyses of existing direct legislation, is there much doubt of the desirability of parties playing a role, and providing the benefits listed above. The only question is whether parties might not be so seriously weakened by having to cope with direct legislation that they simply could not perform an organizing and guiding role nor survive as political actors. This is the possibility we have considered and rejected here. With sensible institutional safeguards of the kind discussed above parties should flourish under direct democracy just as they do under other kinds of democracy. Obviously given the patchy and opaque nature of much of the evidence, and the importance of the topic, this conclusion requires more research. It would be particularly useful to have a database on all popular policy consultations held over the postwar period, with concurrent information on the rules under which they were held and information on how they impacted on parties. In this way we could look more systematically at the relationships examined crudely in Tables 3.1 and 3.2 and Figure 3.3 above. A further desperate need is to define and measure systematically what we mean by party decline. There has been much diffuse discussion of this question and some systematic measurement of membership and support base in the shape of party identifiers. But these form only two of the levels of party activity. Is the continued ability of parties to dominate governments and parliaments not a better indicator of their general health? Whether or not it is, we need an agreed set of measures of party viability before we can make sweeping pronouncements about the effect of direct democracy on parties. From the available evidence there is no reason to believe that parties cannot organize voting and government in a properly regulated direct democracy just as efficiently as they now organize them under representative democracy.

Note 1 The abrogative referendum in Italy is usually classified as an initiative rather than a referendum as it can be requisitioned by popular petition, subject to review by the courts. Although formally it can only aim at abolishing existing legislation, in practice the desired alternative is spelled out during the campaign and is legislated by parliament if there is a majority. In British Columbia, the one Canadian province with the initiative at the province-wide level, no question has ever been placed on the ballot since the adoption of the legislation in 1994.

4 Legislative Response to Direct Legislation Elisabeth R. Gerber and Simon Hug

Current scholarly knowledge suggests that direct legislation affects policy making in various ways.1 One way is directly, when voters pass laws by initiative or referendum. A second way is indirectly, when legislators or other policy-makers respond to the threat of, or act in anticipation of, initiatives or referendums. Many recent studies have advanced our understanding of direct effects by analyzing what types of policies voters pass by initiative and referendum. For example, Tolbert (1998) notes that many laws passed by direct legislation involve governance policies that regulate elected officials. Schrag (1998) claims that middle-class white voters use direct legislation to pass conservative fiscal and social legislation in the US states. Gerber (1999) shows that broad-based citizen groups use direct legislation to pass majoritarian policies. Gamble (1997) argues that citizens use direct legislation to take away legislative protections for minorities, although Donovan and Bowler (1998b) and Frey and Goette (1998) claim that the effects on minorities are much less detrimental than Gamble asserts. A number of recent studies have also tried to study indirect effects. Compared to the research on direct effects, however, this body of research has produced a much murkier picture. Some studies claim to show that indirect effects are substantial and important (Gerber, 1996 and 1999; Matsusaka, 1995; Hug, 1999); others argue that they are minimal (Lascher et al., 1996, Camobreco, 1998). Some argue that indirect effects benefit majoritarian interests (Gerber, 1996); others claim that they may, under certain circumstances, benefit narrow special interests (Matsusaka and McCarty, 1998; Hug 1999). In this chapter, we argue that the debate over indirect effects results largely from two methodological problems in most studies of legislative response. First, any study of legislative response must rely on strong 88

Legislative Response to Direct Legislation

89

counterfactual arguments – that is, claims that policy outcomes would have been different if direct legislation were not possible. But few scholars studying legislative response to initiatives and referendums address this methodological challenge head on. As a result, their studies rest upon problematic inferences. Second, studies of legislative response must be sensitive to issues of model specification. Theory suggests that legislative response to the threat of direct legislation is dependent on voter preferences. This means that policy outcomes should not only differ between polities with and without institutions for direct legislation, but that they should differ in predictable ways as a function of voter preferences. These theoretical advances have not been incorporated into some of the empirical models in the recent literature, suggesting that these models may be inappropriate for capturing the effects of direct legislation on legislative behavior. In our review of the literature, we argue that failure to make counterfactual claims explicit plagues virtually all one-country (or one-state) case study analyses of direct legislation. Comparative case studies and large-scale statistical analyses are on somewhat firmer ground. While none of these research designs can completely circumvent the problem of making hypothetical counterfactual claims to demonstrate indirect effects (which we describe in much greater detail below), the latter two typically make such reasoning more explicit. As such, we find the results of the comparative case studies and statistical analyses more compelling. However, while the comparative case studies and statistical analyses more directly address the first methodological problem, few address the second – that is, their models fail to capture important relationships illuminated in recent theoretical models. Those studies that do capture these relationships find that legislative response to direct legislation empowers citizens and interest groups whose preferences reflect the polity’s median voter. In this chapter, we provide a critical overview of the insights gained from the literature on legislative response to direct legislation. We begin by describing the nature of direct and indirect effects. We then summarize a body of case study based research that attempts to identify indirect effects. We use this discussion of the literature to illuminate the counterfactual claims that underlie all empirical studies of indirect effects (and indeed, all analyses of the effects of institutions on policy). We next describe the recent theoretical research on legislative response as the basis for statistical modeling. We argue that the greatest methodological challenge in studies of legislative response is to construct a research design

90

Elisabeth R. Gerber and Simon Hug

that is fundamentally comparative, and that incorporates theoretical insights into the relationships between voters, legislators, institutions, and policy. We then assess the literature on legislative response in terms of how each study deals with these methodological challenges. We conclude by emphasizing the need for further systematic studies to gain a better understanding of legislative responses to the instruments of direct legislation.

Legislative responses and indirect effects Legislative responses to direct legislation are best conceived as one component of the policy consequences of allowing citizens to vote directly on policy. Provisions for direct legislation affect policy outcomes in multiple ways. 2 One type of effect occurs when citizens pass laws by initiative or referendum. These effects are called ‘direct’ since they refer to changes in policy that are the immediate consequences of popular votes. A second type of effect occurs when citizens reject legislative proposals or adopt proposals opposed by the legislature. These effects are also ‘direct’ in the sense that they come about when citizens take action at the polls, but they differ from the first type in that they are taken in response to actions of the legislature. Finally, direct legislation also has ‘indirect’ effects. For instance, legislatures may react to the presence of the direct legislation option by adopting proposals to fend off an initiative or referendum threat (for example, Gerber, 1996 and 1998). Similarly, legislatures may adopt policies to avoid having their agenda-setting powers wrested away from them.3 Following these examples, it is tempting to consider as ‘direct effects’ those things that happen at the ballot box, and to consider as ‘indirect effects’ those that transpire as the result of votes in the legislature. However, the following additional examples reveal that such distinctions are overly simplistic. Consider the claim made by some authors that direct legislation institutions allow legislatures to shirk and avoid making tough decisions (for example, Magleby, 1984, pp. 186 and 192; Cronin, 1989, p. 229). Consequently, the outcome of a particular ballot vote might be incorrectly attributed to direct effects, despite the fact that the same legislature in a purely representative democracy would have adopted the same policy. Similarly, a legislative proposal being adopted or defeated in a popular vote might have looked differently in a purely representative democracy. In both examples, the direct and indirect effects of direct legislation interact in complex and subtle ways. They may even cancel each other out.


91

For this reason, it is important to differentiate clearly between direct and indirect effects. We do so as follows. Direct effects of initiatives and referendums appear if a policy is adopted at the ballot box, but would not have found the necessary support in the legislature. Indirect effects of direct legislation refer to differences in policies that do not depend on the success of a popular vote on a given proposal. Differentiating them in this way highlights the difficulty in identifying, in a precise way, the effects of direct legislation.

Extant literature It is not surprising that much of the empirical research on legislative response to direct legislation focuses on polities where citizens frequently vote on policy matters. In countries like Switzerland (with an average of six votes per year) or in some American states like California (with an average of three votes per year), direct legislation seems to loom large over every aspect of policy making.4 Thus, scholars interested in these places would be missing a large part of the political landscape if they ignored the effects – direct and indirect – of direct legislation. Some of the reasons the literature focuses on Switzerland and the American states also appears to be instrumental. In countries where legislatures frequently deal with the prospect of initiatives and referendums, we would expect to see greater adaptation of their practices to this political reality. In other words, when direct legislation is common, legislative response is likely. Observers of Swiss democracy stress that direct legislation not only allows citizens to vote on certain issues, but that it also affects other aspects of policy making (for example, Rappard, 1912). Perhaps the clearest early statement of this belief is offered by Neidhart (1970). Neidhart shows how the federal government developed procedures to consult interest groups about new legislation to ensure that policy proposals would not face referendum challenges. Specifically, in most policy domains, legislative proposals are first discussed and amended by expert groups composed of representatives of interest groups and concerned organizations. After this vetting, the proposals are formally submitted to the political parties, interest groups and the cantonal governments for official reaction in a process called Vernehmlassung. Only after this lengthy procedure is a legislative proposal considered by parliament. Consequently, Neidhart’s (1970) main conclusion is that referendums have transformed Swiss democracy into a bargaining democracy. He also suggests that this transformation has led to differences in

92


policy outcomes: policy outcomes in federal legislation much more strongly reflect the preferences of interest groups that can credibly threaten to launch a referendum. Immergut (1992) extends this reasoning when comparing the health policies in three countries, including Switzerland. While both France and Sweden were able to overcome opposition to comprehensive health insurance plans (France only under the Fifth Republic), Switzerland failed to do so. Immergut argues that after a policy proposal failed due to a referendum in 1911, politicians became prudent and only minor reforms were possible. The federal government and the legislature remained hostage to powerful interest groups that could threaten with a referendum challenge (Immergut, 1992; see also Neidhart, 1970, pp. 150–71). These studies emphasize how popular referendums empower defenders of the status quo. Immergut (1992) relates this effect to three elements. First, referendums inherently have a status quo bias, since uninformed voters tend to reject new proposals. 5 Second, given the uncertainty over the outcome of a referendum vote, the legislature is likely to prefer heeding the demands of groups capable of launching a referendum campaign. Third, legislatures might be impatient to adopt a new policy, and a referendum challenge inevitably drags out the process. This might then lead legislatures to compromise with powerful interest groups, thus avoiding a referendum and accelerating the legislating process, while at the same time inhibiting ambitious policy departures. Whereas the popular referendum seeks to nullify an existing legislative decision, the popular initiative can focus on any issue citizens choose. In contrast to the popular referendum, the popular initiative is often portrayed as a counterbalance against this ‘conservative bias.’ At first glance, this claim might come as a surprise, since the passage rate of popular initiatives is extremely low in Switzerland – less than 10 percent of them succeed at the ballot box (Trechsel and Kriesi, 1996). However, authors such as Delley (1978), Werder (1978), Kriesi et al. (1981), and Hofer (1987) argue that despite this low success rate, the indirect effects of the popular initiative are considerable. Kriesi et al. (1981, p. 566), for instance, estimate that in the postwar period up to 1978, about a third of all initiatives had indirect effects on policy outcomes, even though only a single initiative was accepted at the ballot box during that period (Trechsel and Kriesi, 1996, p. 191). The indirect nature of the Swiss popular initiative renders studies of their indirect effects quite straightforward. In contrast to the direct initiative in California, for example, an initiative proposal duly submitted


93

with the required signatures (100 000 at the time of writing) must first be examined by the Swiss federal government and both chambers of parliament. The legislature may submit a counterproposal to the initiative, which appears on the same ballot.6 As Hofer (1987) shows, many initiatives are withdrawn once a counterproposal is adopted in parliament. Given that no vote takes place in such cases, their effects on policy are clearly indirect. But even initiatives that are unsuccessful at the ballot box may subsequently have policy consequences. An often-cited example refers to a series of anti-foreigner initiatives that appeared on the ballots in the 1970s. Despite all major political parties forming a unified front against these proposals, one measure aimed at limiting the share of foreign residents failed by only a very small margin. Several authors trace back Switzerland’s relatively restrictive immigration and asylum policies to these ‘nearly successful’ popular initiatives (for example, HoffmannNowotny, 1985; Trechsel and Kriesi, 1996, p. 203; Killias, p. 1997). Considerations about the indirect effects of direct legislation find their parallels in scholarly work on the American states. In an early study of the initiative and referendum in California, Key and Crouch (Key and Crouch, 1939, p. 499) present evidence for a 1923 case in which the governor refused to sign a bill giving him more control over the appointment of members of boards and commissions – known as the ‘Bromley Spoils System Bill’ – due to the threat of a popular referendum. Schmidt (1989, p. 26) also refers to a series of instances where such indirect effects are apparent. Frequently, just the filing of an Initiative petition will spur legislators into action. Laws passed under the pressure of an Initiative petition include Massachusetts’ laws to reduce the air pollution that causes acid rain (1985) and ban the experimental use of cats and dogs from pounds (1983); Arizona’s laws abolishing the state sales tax on food (1980), making funding available for Medicaid (1982), and restricting water pollution caused by toxic chemicals (1986); and Wyoming’s law to maintain stream flows adequate for fish and wildlife (1986). Accordingly, most scholars of direct legislation in the United States concur with the view that indirect effects may be important (for example, Magleby, 1984, p. 186; Cronin, 1989, p. 224). Often, however, they rely on only a few passing empirical examples to support their claims. Cronin (1989, p. 206), for instance, cites one example of a legislature reacting to failed initiatives before writing in his concluding chapter

94


(Cronin, 1989, p. 224) that ‘[I]ndeed, the mere circulation of petitions for an initiative, referendum, or recall sometimes “encourages” officials to reconsider what they are doing and how they are doing it.’ Banaszak’s (1991 and 1996) careful comparison of the introduction of women’s suffrage in the United States and Switzerland brings together the experiences of these two countries. The countries differ in size but share some common institutional features (for example federalism, bicameralism, and direct legislation at the subnational level). Despite these similarities, the introduction of women’s suffrage in the two countries followed very different paths. While some American states introduced the women’s right to vote in the early years of the twentieth century, most Swiss women had to wait until the 1970s, and some until the 1990s, to become full citizens. In the latter case, the referendum is often identified as the main culprit (for example, Papadopoulos, 1998, p. 183). But Banaszak’s careful comparison of the Swiss and American paths toward women’s suffrage casts some doubts on this argument. Several of the American states that introduced the women’s right to vote early in the twentieth century also had to overcome the referendum hurdle in order to change their state constitutions, and after several attempts, they succeeded. Banaszak (1996) argues that the American women’s suffrage movement used the referendum tool much more skillfully to pressure political elites and legislatures into extending the suffrage. Swiss advocates of women’s suffrage were much more reluctant to use the referendum tool. While this short overview suggests some common themes in the research on legislative response in the two countries that most often use direct legislation, similar conclusions appear in scholarly work on other direct legislation countries. In the case of Italy, Bogdanor (1994, p. 64) cites the example of five laws which were challenged by referendums between 1970 and 1987 but subsequently altered by parliament within the four-month grace period established for this purpose under the terms of the referendum abrogativo. Thus, under the very real threat of referendums, parliament reacted and incorporated some of the demands voiced by the groups launching the referendums. Uleri (1994, p. 425; 1996, p. 121) supports this view and discusses in more detail the abortion referendums of 1981. Two referendums – one aimed at legalizing abortions and one attempting to outlaw them – were on the ballot and both failed by large margins. Nevertheless, Uleri (1996, p. 121) argues that parliament responded by adopting changes in the abortion laws. Even more critical in this respect were the changes in the electoral laws forced upon the political parties by proposed referendums. In 1993, only one


95

referendum on electoral law was admitted by the constitutional court, namely the one changing the electoral rules for elections to the Senate. But following the passage of this referendum, parliament adopted changes for elections to the lower house as well (Uleri, 1996, p. 118). These changes were instrumental in the process leading to the demise of the former Italian party system. As such, the examples provide support to Bogdanor’s (1994, p. 65) claim that, ‘perhaps (as in Switzerland), the most important referendums in Italy have been those that have not been held, the threat of the referendum having persuaded governments to put forward, or to refrain from putting forward, particular bills.’ Another country that allows for direct legislation (in this case, referendums launched by actors other than the government) is Denmark. The abolition of the second legislative chamber in 1953 was accompanied by a provision for referendums initiated by a minority in the lower house. According to article 42 of the Danish Constitution, one-third of the MPs may launch a referendum on almost any legislation. While this procedure has only been used once since its introduction (Svensson, 1996, p. 44), authors disagree as to whether it has had any secondary consequences. Bogdanor (1994, p. 73) argues that article 42 forces the government to accommodate the interests of the opposition, especially if the latter are in accord with those of the citizens. Svensson (1996, p. 48) suggests that the observable consequences of direct democracy are somewhat limited. Without thorough case studies, however, it is difficult to decide to what extent there may have been invisible effects, by stimulating more consideration for the minority or greater willingness to compromise.

Methodological problems: counterfactual reasoning All empirical studies of the effects of institutions on policy rely on a crucial counterfactual argument, and studies on the effects of direct legislation institutions on legislative behavior and policy are no exception. Key and Crouch (1939, p. 564) provide an eloquent description of this problem: In examining direct legislation one may ascertain the procedures followed, the number and nature of measures considered, legal problems arising in connection with its operation, reported campaign expenditures, electoral participation, and other indisputable facts. Of greater significance, however, is the broad question of the effect upon legislation in general of the adoption of these devices of direct

96


legislation. To answer this in keeping with the rigorous demands of scientific method would require retrospective prophecy; in other words, a determination of what would have happened in the realm of legislation had the initiative and the referendum not been adopted as compared with what has occurred since the direct legislative procedures have been in use. Key and Crouch’s (1938, p. 564) ‘retrospective prophecy’ is simply one form of a counterfactual statement. Taking Immergut’s (1992) analysis as an example, we can easily illustrate the retrospective prophecy or counterfactual argument that underlies her conclusions. To conclude that the threat of referendums shaped Swiss health policy in the way she describes, we must believe that without the possibility of such threats, Swiss health policies would be radically different today. Contrary to most scholars writing on this topic, Immergut appears to be aware of this problem. Her comparative case study focuses on three countries that started at the turn of the century with similar health policies and similar visions about how these policies should be reformed. One major difference between the three countries, of course, was that only Switzerland allowed direct legislation. Thus, by selecting countries that were similar on some dimensions but differed in their institutional arrangements, Immergut’s research design allows her to rule out some alternative explanations for why the three countries ended up with such different health policies. Still, there are many other factors that are not ‘held constant’ across her three cases. For example, and especially in light of the negative outcome of the 1911 health policy referendum, it might well be the case that Swiss citizens were simply less supportive of broader health care regulations, independent of any institutional effects. Immergut (1992, pp. 175 ff.) implicitly acknowledges this counterfactual reasoning when she notes that: ‘[i]t is not at all clear that similar votes held in other countries would have yielded different results. For attempts to compare public attitudes to social programs do not show clear national trends.’ Conspicuously absent from the studies she refers to in support of her claim, however, are public opinion data on Swiss citizens’ opinion on health policies (Immergut, 1992, p. 176). Thus, these sources can hardly underpin the counterfactual claim she wishes to make. Similar caution must be applied to Neidhart’s (1970) thesis and the studies demonstrating how failed or withdrawn initiatives nonetheless led to policy changes due to legislative responses. In the case of withdrawn initiatives, we again are asked to believe the strong counterfactual claim


97

that policies in a given issue area would not have changed if initiatives had not been launched. For instance, in the field of Switzerland’s immigration policies, we must believe that policies would be considerably different if direct legislation was not possible. Some cautionary remarks to this effect are implied in Soysal’s (1994) careful comparative analysis of immigration regimes. Her findings suggest that even in policy areas where direct legislation does not directly apply, Switzerland has adopted very restrictive immigration and naturalization measures; hence it is difficult to claim that direct legislation caused the legislature to adopt certain policies when they also adopted similar policies in the absence of direct legislation. The illustrative cases marshaled in support of legislative responses in the American states also rely on counterfactual claims. Whether we consider Key and Crouch’s (1939, p. 499) example of a spoils bill or Cronin’s (1989, p. 206) example of the Michigan legislature adopting tax cuts out of fear of more radical policies forced on them through the initiative process, all rely on such claims. But counterfactuals also underlie the statistical analyses reported in Banaszak (1996, pp. 182–4), which show that states that had several referendums on women’s suffrage extended the right to vote earlier to women than other states. The relationship between counterfactual arguments in case studies and statistical analyses is described eloquently in Fearon (1991). Whether a researcher employs case studies, large-scale comparisons, or statistical analyses, some sort of counterfactual reasoning is always present. This becomes even more apparent in King, Keohane, and Verba’s (1994, pp. 76ff., 81ff.) definition of a causal effect. They define this notion on the basis of a counterfactual argument, which is very similar to Key and Crouch’s (1939, p. 564) notion of ‘retrospective prophecy’: ‘[T]he causal effect is the difference between the systematic component of observations made when the explanatory variable takes one value and the systematic component of comparable observations when the explanatory variable takes on another value.’ The crux of this definition obviously lies in the notion of ‘comparable observations,’ that is, in which all other factors except the explanatory variable of interest are identical, irrelevant, or controlled statistically. This definition of causal effects applies equally to qualitative and quantitative studies. Nevertheless, it suggests that comparisons are virtually inevitable in order to support causal claims. Immergut’s (1992) study is a prime example. Her claims gain in credibility because she can compare different institutional settings and thus attempt to rule out some of the competing explanations for the causal effects she is interested in.

98


Studies relying on single cases or unique institutional settings in general fail to offer this possibility of comparing across contexts.

Methodological problems: theoretical underpinnings These pitfalls of counterfactual reasoning become even more evident in light of recent theoretical models which demonstrate legislative responses to direct legislation. In the formal literature on direct legislation, Romer and Rosenthal’s (1978 and 1979a) pathbreaking work marks an important turning point. Literature predating Romer and Rosenthal largely suggested that in initiatives and referendums voters strictly get their way (see Mueller 1989, part II for a detailed survey of these earlier studies). Romer and Rosenthal’s agenda-setting model for school bond referendums showed how the proposer of a bond measure may take advantage of her strategic role. Since the agenda-setter may have control over the reversion point (for example, the policy that is implemented if the voters should reject the referendum proposal), policies may fail to correspond to the citizens’ most preferred policy. A series of scholars have extended this basic model to account for various agenda-setting situations, some of which apply directly to direct legislation. For instance, Lupia (1992 and 1994) extends the basic Romer and Rosenthal model to address the question of how citizens become informed in initiative campaigns. Steunenberg (1992) and Gerber (1996) extend the basic Romer and Rosenthal model to study legislative response. Their models rely on the assumption of a one-dimensional policy space and complete and perfect information. The key insight is that if direct legislation has an effect on policy outcomes, the final policy is closer to the voters’ preferred policy.7 For this change in policy to come about, however, votes on initiatives or referendums need not occur. Politicians may prefer to pass laws other than their most favored in order to pre-empt an initiative. The models show the conditions under which such legislative responses may occur. They also show the representational consequences of legislative responses to direct legislation, that is that the threat of initiatives should lead to policies more preferred by the majority of voters. This last theoretical result contrasts with the common view that direct legislation empowers narrow interest groups and results in policies favoring special interests. 8 More recently, several scholars have developed models that consider how this basic dynamic changes when governments and legislatures are uncertain about various aspects of the policy process. Moser (1996) proposes an incomplete information model that builds upon Steunenberg’s


99

(1992) model of popular referendums. Matsusaka and McCarty (1998) study a model extending Gerber’s (1996) model of the initiative. Finally, Hug (1999) proposes a model covering both popular referendums and initiatives, as well as constitutionally required and government-initiated referendums. These models come to very similar conclusions about the expected policy consequences, as do those models that rule out informational asymmetries. Under wide-ranging conditions, if policy outcomes differ in settings with direct legislation, the final policy outcomes are more reflective of citizen preferences (that is, a majority of voters). 9 While in these incomplete information models votes occur in equilibrium, policy outcomes may also be affected when no votes occur. Again, these changes reflect legislative responses to direct legislation. Given that these incomplete information models predict that initiatives and referendums will sometimes occur, as they do in the real world, they offer some insight into when legislatures will act to pre-empt initiatives (that is, when they have enough information to act effectively) and when they will not (that is, when they cannot anticipate a credible initiative threat). The incomplete information models also hint at the considerable difficulty in clearly sorting out the direct and indirect effects of direct legislation. They all highlight the crucial interplay between legislative proposals, which attempt to anticipate initiative and referendum challenges, and the strategic responses by interest groups in threatening to use direct legislation. These models, while limited to one-dimensional policy spaces and particular types of informational asymmetries, suggest that legislative responses to direct legislation will generally lead to policies preferred by the majority of voters.10 More precisely, if policy outcomes differ, they do so in the direction of the median voter’s preferred policy. Thus, the effect of direct legislation on policy outcomes crucially depends on the voters’ preferences. This insight clearly harks back to the counterfactual argument that Immergut (1992) implicitly wishes to make, namely that differences in preferences between Switzerland and other countries cannot alone explain differences in health policies. But all of these models suggest that if direct legislation does have an effect, it should be to mediate the impact of preferences on policy outcomes. These theoretical models have important implications for the appropriate research design to study legislative responses to direct legislation. First, they make it clear that it is extremely difficult to sort out direct and indirect effects when studying the policy consequences of direct legislation. Second, they show that voter preferences play an important mediating role in shaping indirect effects. Consequently, systematic

100


studies of these policy consequences that take voter preferences into consideration appear to be the most fruitful approach to studying legislative responses.

Systematic studies of policy consequences Despite Key and Crouch’s (1939, p. 564) plea for studies of legislative response, few scholars of their era followed their call. Only recently have scholars proposed systematic studies on the direct and indirect policy consequences of direct legislation. In the late 1960s and 1970s, most work on this topic was done by economists who examined whether economic performance in communities or states with direct legislation differed from performance in those without. Their underlying model was a demand function for public goods which included a number of economic variables. The economists then tested whether (and how) direct legislation altered economic outcomes. As we discuss below, these early models were highly reductionist, especially in their treatment of voter preferences. More recently, several scholars have improved upon these early economic models with more sophisticated model specifications and direct measures of voter preferences. Most of these latter studies focus on policies outside the realm of economic performance. Below, we first summarize studies focusing on economic outcomes before discussing some recent analyses of other policies. Economic outcomes Studies of the effect of direct legislation on economic outcomes are often informed by the argument that politicians in purely representative democracies pursue interests different from those of voters. These divergent interests are assumed to result in higher spending on wasteful projects, higher taxes, higher levels of debt, and other economic inefficiencies. Direct legislation, in this context, provides a useful check on the activities of elected officials and thus aligns spending and tax levels with citizens’ preferences. Early studies at the level of communities in the United States found little evidence to this effect.11 Pommerehne (1978), however, did find important differences in direct legislation and non-direct legislation communities in Switzerland. To study expenditure patterns in Swiss municipalities, Pommerehne built an econometric model based on a demand function of the median voter. He finds that his model performs better for communities with direct legislation than for those without. He takes this as evidence for a stronger influence of the median voter’s


101

preferences in municipalities with direct legislation. 12 Kirchgässner, Feld and Savioz (1999, chapter 4) extend Pommerehne’s study by including data from recent years and more municipalities. Instead of estimating different equations for each subset of municipalities, they pool their observations and introduce a dummy variable to capture the effect of direct legislation. They find a significant negative effect, thus demonstrating that expenditures are lower in municipalities with direct legislation. It is interesting to compare these two studies for their different approaches and interpretations. Pommerehne (1978) bases his argument on the explanatory power of his median voter model. As such, his work relates directly to the recent theoretical models, which suggest that the median voter’s preferences should be more strongly related to policy outcomes in political entities that allow for direct legislation. Kirchgässner et al. (1999), by contrast, base their inferences on the size, sign, and significance of the coefficient on their direct legislation dummy variable. The implication of Kirchgässner et al.’s estimated model and results is that expenditure levels are biased in a specific direction in municipalities with direct legislation. This bias may or may not be in the direction of the median voter. A study by Feld and Savioz (1997) also builds upon the argument that direct legislation provides a check against politicians’ wasteful spending habits. Feld and Savioz test this proposition by estimating the effect of direct legislation on GDP in the cantons of Switzerland. Using as their baseline a Cobb-Douglas production function, they estimate the marginal effect of provisions for direct legislation in the realm of fiscal policy. Their results suggest that in cantons with direct legislation, GDP is, on average, one-sixth larger than in all other cantons. Whether this effect is due to the increased influence of the median voter in cantons with direct legislation in the realm of fiscal policy is, however, hard to establish with this specification of the empirical model. Similar analytic approaches are taken by other scholars studying the effect of direct legislation in the United States and Switzerland. Matsusaka (1995, 2000) provides perhaps the most detailed accounts of how direct legislation affects spending, deficit, and tax patterns at the level of the American states. For the period 1960–90, his analysis suggests that states with popular initiatives had lower expenditures, taxes, and deficits. He reaches this conclusion by estimating a base model that attempts to predict state spending, taxing, and deficit patterns. His model controls for economic and demographic factors, and includes a dummy variable for initiative states. The coefficient on the direct

102


legislation dummy turns out to be negative, indicating that spending levels in initiative states are significantly lower. Extending his analysis backward in time, Matsusaka (2000) finds that for the first half of this century, the economic effects of initiatives are strikingly different. Estimating a similar model for the earlier period, he finds that spending was actually higher in initiative states. Analyzing support for a series of ballot proposals, Matsusaka argues that in the earlier period, voters preferred higher spending on social programs. Thus, spending patterns were more consistent with the median voter’s preferences in initiative states: ‘the initiative does not appear to be inherently an antispending device’ (Matsusaka, 2000, 641). It is interesting to compare these two conclusions in light of the insights produced by the theoretical models discussed above. As Matsusaka (2000) clearly states, voter preferences seem to have shifted over time, thus producing the observed sign reversal on the estimated effect of direct legislation. This suggests that the direct legislation dummy variable picks up not only the effect of the initiative process, but also the direction of the voters’ preferences. Obviously, if the voters’ preferences change over time, the coefficient on this dummy should also change in predictable ways. However, the statistical models used by Matsusaka and others capture these changes in preferences only indirectly, by estimating the model separately at different points in time. If changes in voter preferences take place over time, they might also occur across space, that is, across cantons, states, and countries. These changes, however, cannot be captured directly in the specification of the statistical models discussed above. Given that they only allow for tests of policy differences across all units considered, the estimates in these studies will be biased toward zero if voter preferences are in one direction in some units and in the other direction in other units. Consequently, the effects of direct legislation on economic outcomes, conditional on the voters’ preferences, are likely to be even stronger than suggested by these empirical results. Recent work by Camobreco (1998) challenges Matsusaka’s (1995) results. Studying similar fiscal policies across states, Camobreco attempts to control for citizen preferences by including in his analyses a measure of state ideology constructed by Erikson, Wright and McIver (1993). Camobreco fails to find support for the claim that policy is more closely related to voter preferences in initiative states than in non-initiative states. But Matsusaka (1999) clearly demonstrates that the results from Camobreco’s linear specification of the empirical model do not necessarily show that the median voter’s preferences are better or worse repre-


103

sented in direct legislation states. The reason for this ambiguity is that Camobreco’s linear specification can only show whether the effect of the median voter’s preferences on fiscal policies is stronger or weaker in direct legislation states. But theory suggests that what matters is not this absolute difference in policy, but rather the relationship between the median voter’s preference and the policy adopted. The spatial theories predict that this relative distance should be smaller in states with initiatives or referendums. Matsusaka’s (1999) critique thus corresponds in part to Romer and Rosenthal’s (1979b) critical review of empirical tests of the median voter result. Kiewiet and Szakaly (1996) address a slightly different question using similar methods. Their research question focuses on whether various measures adopted to render the issuing of bonds by state governments more difficult reduce state indebtedness. Controlling for a host of economic and political variables, among them a measure of voter preferences, they find that required referendums on bonds effectively reduce a state’s level of indebtedness. These economic effects of referendums also invoke responses by state legislatures. Kiewiet and Szakaly’s (1996) study provides the most clear-cut evidence in this regard. Given that required referendums make new debt more difficult to incur, Kiewiet and Szakaly show that legislatures are forced to adapt to the fiscal constraints imposed by direct legislation. These legislative responses to debt constraints are indicative of the types of responses we expect in other settings where direct legislation changes the economic environment. Policy consequences One of the major differences between the economists’ studies of the effects of direct legislation on economic outcomes, and recent work by political scientists on the effects in other policy areas, is how each body of work operationalizes voter preferences. Most economic studies estimate citizen or voter preferences with aggregate data measuring population characteristics. Most political science studies employ more direct measures of preferences based on survey results. Given the important mediating role of preferences highlighted in the theoretical models, it would seem that more direct measures of preferences would produce stronger, and more reliable, results. Gerber (1996 and 1999) uses such direct preference measures in her study of state laws requiring parental notification or consent for teenage abortions. Gerber estimates each state’s median voter’s preference for consent and notification policies from public opinion surveys

104


(Miller et al., 1993). The surveys asked random samples of adult citizens in each state whether they supported or opposed these (and other) policies. The surveys also asked whether respondents voted in recent elections. Gerber takes the mean response of self-reported voters from each state as the estimate of the median voter’s preference. She finds that voter preferences have a stronger impact on the likelihood of having these policies in states with initiatives than in those without. This finding is perhaps the strongest indication of legislative responses to initiatives in the literature. Given that no initiative or referendum had occurred on parental consent and notification laws prior to the study, policy differences across states, controlling for confounding factors, are due strictly to the indirect effects of initiatives. Using similar direct measures of voter preferences, Gerber (1999) reports nearly identical results for the presence and absence of the death penalty in the American states. Again, the preferences of voters for capital punishment are much more strongly reflected in criminal law in states with direct legislation than in those without. Given that in the course of the century several votes have occurred on this contentious issue, the effects of direct legislation in this policy area are most likely a mix of direct and indirect effects. In most policy areas, however, direct preference measures are unavailable. Thus, scholars must use proxy measures that are (presumably) related to underlying policy preferences. Hence, Lascher, Hagen and Rochlin (1996) and Camobreco (1998) estimate policy preferences using a composite ideology scale based on survey responses over a period of 15 years (Erikson et al., 1993). Analyzing eight policies, Lascher et al. (1996) fail to find evidence of a stronger effect of voter preferences in initiative states. The same non-results appear in Camobreco’s (1998) study of fiscal policies. We caution against drawing strong inferences from these null results, however. As we explained in the context of Camobreco’s study above, the specification of the empirical models proposed in these two studies is problematic. First, both studies rely on an ideological scale which may or may not be closely related to the policies under consideration. Second, they analyze policies that are (naturally) measured by interval variables (for example, expenditures). Given that these policies are measured continuously, it is crucial to assess the median voter’s ideal point relative to this continuous policy space.13 If policy and preferences are measured on different metrics, inferences from their specifications may be very misleading. In other words, since the theory predicts that direct legislation biases policies toward the preferences of the median voter, the simple linear specification employed


105

by Lascher et al. (1996) and Camobreco (1998) cannot provide an adequate test of the theoretical models. Matsusaka (1999) shows that differences in slope coefficients between direct legislation and nondirect legislation states, which both sets of authors use to justify their claims, do not tell us anything about whether policies reflect the voters’ preferences more closely or not. A recent study by Gerber and Hug (1999) on the effects of direct legislation on minority rights suggests that the use of a preference proxy may yield results consistent with theory, provided that the policy space under consideration is defined by the presence or absence of a particular policy and a non-linear model is employed. Studying various protective measures for minorities, they find clear evidence that direct legislation biases policy outcomes toward the majority position. More precisely, they find that in conservative states direct legislation makes protective measures for minorities less likely. In liberal states, it renders them more likely. In extremely conservative states, the effect of direct legislation largely disappears, since the preferences against protective measures transpire in both direct legislation and non-direct legislation states.

Conclusion Most scholars concur that legislative responses to direct legislation are a fact of life. These effects prove, however, to be very elusive in empirical research. Case studies of particular countries or states employing direct legislation often rely on strong and mostly implicit counterfactual claims to demonstrate how legislatures react to the threat of initiatives and referendums. Scholars of Swiss, American, Italian, and Danish direct legislation, among others, provide interesting insights into the legislative responses in these countries. Some of these scholars suggest that the legislatures in these polities are more responsive to the preferences of the voters; others argue that they become more responsive to powerful interest groups. Since these case studies fail to address the underlying counterfactual logic directly, some degree of skepticism is warranted when evaluating their findings. This skepticism is further heightened in light of recent theoretical work on the policy effects of direct legislation. Results obtained from formal models suggest a close intertwining of both direct and indirect effects of direct legislation on policy outcomes. Only under very rare circumstances can the indirect effects of direct legislation, which form the essence of legislative responses, be easily teased out.

106


Both the methodological problems inherent in case studies and the implications of recent theoretical work suggest a need to resort to systematic studies of the policy consequences of direct legislation. These policy consequences comprise both direct and indirect effects, and in some rare cases, the legislative response can be clearly assessed. Such systematic studies of policy consequences have mostly addressed economic outcomes at the subnational level in Switzerland and the United States. Recent studies at the state level in the United States have also addressed the effect of direct legislation on other policies. Among the studies of economic outcomes, some consensus appears that direct legislation biases policy outcomes in particular directions. For instance, Feld and Savioz (1997) find that Swiss cantons with direct legislation have, on average, a much higher GDP than those without. Matsusaka (1995, 2000) finds that expenditures in initiative states were lower in the second half of the twentieth century and higher in the first half. But it is not clear whether, as theory predicts, these policies vary in accord with voter preferences. With voter preferences estimated on the basis of aggregated economic data, establishing this claim directly is, at best, a difficult enterprise. Studies of non-economic policies attempt to make this link between preferences and expected changes in policy outcomes much more directly. However, problems of model specification often seriously hamper the conclusions drawn from empirical studies. Perhaps the strongest evidence of legislative responses to direct legislation appears in Gerber’s (1996 and 1999) work on laws requiring parental consent for teenage abortion and the death penalty. Relying on similar data sources and model specifications, Gerber and Hug (1999) and Hug (1999) find similar results indicating legislative responses as predicted by the recent theories. This body of research shows how legislatures are more responsive to voter preferences in states with direct legislation. The future challenge for studies of legislative responses to direct legislation will be twofold. First, further studies employing theoretically motivated model specifications must consider whether the results obtained by Gerber (1996 and 1999), Gerber and Hug (1999), and Hug (1999) equally apply to other policy areas. Second, they must address Kiewiet and Szakaly’s (1996, p. 64) cautious remark about whether results found at the state level may be transposed to the national level. So far cross-national comparisons have been seriously hampered by the limited variation in the institutions of direct legislation across countries. With the recent extension of direct legislation in several East European countries (for example, Brady and Kaplan, 1994; White and


107

Hill, 1996) future cross-national research will become much more feasible. Hopefully, these future studies will allow for much more firmly grounded insights into the legislative responses to direct legislation.

Notes 1 We employ the term ‘direct legislation’ to refer to any votes by citizens on policy. The two types of direct legislation we consider in this chapter are initiatives, where citizens draft, qualify, and vote on legislation, and referendums, where elected representatives draft and pass laws, and citizens vote to ratify or reject them. 2 For a contrary view, see Cronin (1989, p. 232), who argues that policies do not differ between states with and without direct legislation. 3 Legislative responses may also take the form of institutional innovations (for example, in Switzerland). We discuss some of these innovations below. While such responses may have important implications for many aspects of governance, we focus our attention on how they affect policy per se. 4 Figures for Switzerland are calculated for 1951–94 (Trechsel and Kriesi, 1996, p. 191). Figures for California are calculated for 1950–92 (Magleby, 1994, p. 226). 5 This claim, which appears widely in the literature, finds support in Bowler and Donovan’s (1998) recent systematic study of voting behavior. They attribute this tendency to a rational voting strategy in situations of informational scarcity. 6 Until 1987, voters were restricted to voting for either the initiative or the counterproposal but not both. In other words, casting two favorable votes was not valid, but casting two unfavorable votes was. Since 1987, voters can vote for or against the proposals in any combination. 7 The voters’ preferred policy outcome is represented as the position of the median voter in these models. Given the assumptions in these models regarding the dimensionality of the policy space and the preferences of the actors, Black’s (1958) median voter theorem applies. 8 The reasoning appearing in the substantive literature relies on notions of the government’s or legislature’s uncertainty and impatience (for example, Immergut, 1992). While models dealing with uncertainty exist, as we discuss below, to our knowledge none address the issue of impatience. 9 Only under very specific conditions do these models predict that voters are worse off under institutions of direct legislation. This occurs in Matsusaka and McCarty’s (1998) model when the legislature is a perfect agent of the median voter, but compromises with the interest group to avoid an initiative. Similar effects appear in Hug’s (1999) model of direct legislation. 10 Tsebelis (forthcoming), relying on his theory of veto players, presents results showing that a similar tendency towards a central location among the idealpoints of the voters also obtains in multidimensional spaces under complete and perfect information. This central location is, however, different from the notion of ‘median voter,’ since in multidimensional spaces Black’s (1958) median voter theorem generally fails to apply.

108


11 In what follows, we rely heavily on the excellent overviews in Kirchgässner, Feld, and Savioz (1999, chapters 4 and 5). 12 Romer and Rosenthal (1979b) criticize this and similar work, emphasizing that a relationship between, for instance, median income and government expenditures does not demonstrate that the median voter’s preference is realized. It might well be that the government expenditures are systematically lower or higher than what the median voter prefers. 13 By contrast, the policies in Gerber’s model are naturally dichotomous (the state either does or does not have the policy). This leads to the non-linear specification she employs.

5 Amending Constitutions through the Referendum Device Brian Galligan

This chapter is concerned with the use of the referendum device as a mechanism for constitutional change. Such an institutional mechanism is grounded in democratic constitutional theory and is a well-established part of some constitutional democracies. Current interest in direct democracy is mainly focused on popular initiatives and referendums as alternative mechanisms to ordinary decision making by representatives. Such ‘unmediated popular voting’ is supported by some as a purer form of democratic participation that cuts out the mediation of representatives, or what economists might call the transaction costs of indirectness (Budge, 1996, chapter 2) on issues such as immigration or taxation. Such referendums might be seen as being in conflict with, or narrowing the scope for, the normal decision making of representative democracy. Constitutional referendums are quite different. They need to be distinguished from policy referendums because they are concerned with the basic institutions of government. Such constitutional referendums determine the shape of the political process or the way institutions of government are structured. For example, citizens might amend their constitution to provide for citizen-initiated referendums on certain policy matters. That changes the rules on policy making to directly involve citizens in future decisions. Such referendums are not concerned with policy per se but with the institutions of government that determine the policy-making process. Constitutional referendums are also fundamentally significant as a legitimating mechanism for democratic government. Through collectively deciding upon the system of government under which they live, the people express their democratic sovereignty. In so doing they constitute themselves as the sovereign body that makes authoritative decisions about their system of government. 109

110

Brian Galligan

The first section of this chapter sets out the core theoretical justification for constitutional referendums. The second section of the chapter examines the politics of constitutional referendums, with particular reference to Australia that has a century of practice. Its finding is that, despite the difficulties of achieving special majorities in a complex federal system and the endemic problems of partisanship and point scoring that often mar the process, constitutional referendums are practically workable. The third substantive section examines innovative uses of deliberative mechanisms to buttress the referendum device. The chapter advances four propositions: (1) constitutional referendums are a most legitimate and appropriate means of constitutional change in a democracy; (2) despite a range of difficulties in staging popular referendums, they are practically feasible; (3) the Australian people have not, as some contend, been too conservative to will needed constitutional change, but instead have tended to show good sense coupled with a healthy dose of skepticism; and (4) it is not, as some contend, impossible to change constitutions through the referendum device, though it is difficult, which is how it should be with constitutional change.

Theoretical justification In a liberal-democratic constitutional system, the constitution is the higher law or controlling instrument that constitutes the system of government. The constitution determines the form of government, including its system of representative democracy, which, in federal systems, includes the allocation of powers between national and subnational governments. The constitution is the supreme law of the land and the institutional embodiment of the sovereign people’s will. Hence, adopting the constitution in the first place or changing its provisions from time to time requires a special mechanism. The constitutional referendum provides such a mechanism. Indeed, according to the arguments provided here, constitutional referendums provide the best mechanism for enabling the sovereign people to decide constitutional issues. In voting on the basic law of the constitution that determines how they are to be governed, the people act as the sovereign constitutive body. Hence constitutional referendums are of deeper significance as affirmations of the sovereign people than for their practical utility as instruments of constitutional change. Arend Lijphart comments that the popular referendum is an ‘instrument for moving representative democracy in the direction of direct

Amending Constitutions through the Referendum Device 111

democracy’ (Lijphart, 1984, p. 200). That is no doubt true of policy referendums that were discussed above. Constitutional referendums, while also being instruments of direct democracy, have as their fundamental purpose the originating and changing of the system of government itself. Constitutions that are established and changed by popular referendum are directly based upon consent or the will of the people. We might therefore say that the constitutional referendum is an appropriate instrument for establishing a constitutional system of government based upon popular consent, whether that system uses representative or participatory mechanisms for routine politics. Thus, from the viewpoint of constitutional and democratic theory, the referendum device is the most direct and appropriate way to institutionalize government based on the consent of the governed. According to the early champion of liberal politics, John Locke, only government that has the consent of the governed is legitimate. While Locke did not advocate popular referendums and was no democrat, his theory of legitimate government is based squarely on the consent of those governed, and in a constitutional democracy, the appropriate mechanism for consent, by way of extending Locke, is the constitutional referendum. More a champion of democracy, Rousseau preferred participatory over representative democracy, though he thought that such a regime was unlikely because it required special conditions of smallness in scale as well as high levels of virtue, simplicity of lifestyle and equality among members. But practicalities aside, Rousseau taught that government depended on the sovereign will of the people. And again, although Rousseau might not have considered it plausible, the constitutional referendum is an appropriate mechanism for institutionalizing popular will. Another way of justifying constitutional referendums is through the adoption of Bruce Ackerman’s notion of ‘dualist democracy’ where two quite different kinds of political decisions are made. One kind involves the people making decisions on their constitutional system of government; the other involves governments elected by the people making decisions about the ongoing business of government. As Ackerman says, we need to distinguish constitutional politics from normal politics: ‘the will of We the People from the act of We the Politician.’ The two are quite different, with the former requiring special conditions that Ackerman describes as follows: Decisions by the People occur rarely, and under special constitutional conditions. Before gaining the authority to make supreme law

112

Brian Galligan

in the name of the People, a movement’s political partisans must, first, convince an extraordinary number of their fellow citizens to take their proposed initiative with a seriousness that they do not normally accord to politics; second, they must allow their opponents a fair opportunity to organize their own forces; third, they must convince a majority of their fellow Americans to support their initiative as its merits are discussed, time and again, in the deliberative fora provided for ‘higher lawmaking.’ It is only then that a political movement earns the enhanced legitimacy the dualist Constitution accords to decisions made by the People. (1991, p. 6)

Popular sovereignty as the source of constitutional legitimacy was translated into institutional form by the American inventors of modern federal republicanism. As Samuel Beer pointed out in his book on American constitutionalism, the ‘principle that made possible the distinctively American form of constitutionalism and so of federalism was popular sovereignty.’ Beer concluded: ‘This fundamental principle of American republicanism was irreconcilably in conflict with the principle of hierarchy which, as embodied in the Old Whig constitution, necessarily implied parliamentary sovereignty over a unitary system’ (Beer, 1993, p. 137). Although the American founders quite deliberately constrained popular participation within their constitutional design, sovereignty of the people was the potent principle that grounded their republican system. Other more recent constitutions, such as the Australian, followed the American model but used the mechanism of constitutional referendums as a direct institutional embodiment of the principle of popular sovereignty. The American founders had used the popularly-elected constitutional convention which, as Beer points out, was ‘a body elected specially, and in some cases solely for the purpose of establishing or altering the fundamental law of one of the states, an innovation which R. R. Palmer has called “the most distinctive institutional contribution of the American revolutionaries to Western politics”’ (Beer, 1993, p. 309). According to Beer, such a body was necessary for establishing ‘properly authorised governments’ on the basis of ‘a fundamental law binding the legislature itself.’ Since the people were sovereign it was appropriate that the convention be elected by them. To entrust the process of drafting a new federal constitution to delegates appointed by the constituent states – the ‘compact’ model – might be appropriate for a confederal system where the member states are sovereign. But popular election of


delegates was more appropriate for a federal democracy where the people rule. The process of drafting a constitution is distinct, both in terms of the time required for drafting a constitution and the process used for its ratification. Ratification can be carried out by special conventions, either elected or nominated, as was the case in the United States. Alternatively, it can be achieved by way of endorsement by the people voting in a referendum. The latter means is a more direct and appropriate expression of popular sovereignty. It was the method used for the Australian constitution in 1900 when the people of the states (minus women in four of the six states who had not yet been enfranchised, and most Aboriginal people who were excluded either through state legislation or practice) voted affirmatively on the draft constitution before it was formally passed by Westminster. A comparable distinction between drafting and endorsement applies to referendums more generally. The two-phase process allows for different variations of government and popular control in each phase. For example, governments can keep tight control over the drafting of referendum proposals or there can be a more open system of citizen-initiated referendums. Where governments keep control over the drafting as well as the timing of a referendum, it is said to be ‘controlled,’ as distinct from ‘uncontrolled’ where popular initiative determines the calling, timing, and wording of a referendum (Lijphart, 1984). The Australian constitution, where the national or Commonwealth government controls constitutional referendums, is an example of the former, while the Swiss Constitution that allows popular initiative is an example of the latter. Typically, as Gordon Smith has pointed out, strong control is associated with referendums ‘that have foreseeable results in favour of the government’ because otherwise the government would hardly propose them (Smith, 1976, p. 6). Clearly, government control of referendums is significant in determining the sorts of proposals that are put to the people. We would expect controlled referendums to be supportive of the government’s interests, or ‘pro-hegemonic.’ In fact, as Lijphart points out, most referendums are both controlled and pro-hegemonic: ‘when governments control the referendum, they will tend to use it only when they expect to win’ (Lijphart, 1984, p. 203). But crucially, governments do not control the actual referendum process of popular vote so cannot determine outcomes. That might explain patterns of poor success in referendums, as in the Australian case. There, successive Commonwealth governments have been poor judges of what the people will support

114

Brian Galligan

and have persisted in putting forward self-serving and unpopular proposals. This brings us to the next phase of our discussion: the actual politics of constitutional referendums. Normative considerations aside, are constitutional referendums an appropriate mechanism for changing constitutions? It might be that the constitutional referendum is theoretically the most appropriate instrument for constitutional change, but in practice, the politics of referendums might be sufficiently unworkable to turn the tables against it, with too many nefarious effects on liberaldemocratic systems. I will argue that this is not the case, using the Australian experience to illustrate some of the main problems as well as opportunities that the referendum device offers for constitutional change. This will entail giving some detail of Australia’s constitutional politics and record, and critically assessing the more general lessons about constitutional referendums that might be learned.

Political experience It is all very well to justify constitutional referendums as the most appropriate means of constitutional change on theoretical and normative grounds. But are they practical? And do they work? How do they affect the political system? What are the problems likely to be encountered in referendum politics? And can they be resolved? What are the more promising variations of practice that might facilitate popular decision making in constitutional matters? These are important questions and raise complex matters of practical politics. In examining the role played by the referendum on the process of constitutional change, it is particularly difficult to sort out the relevant aspects of constitutional politics from the complexity of idiosyncratic national politics and ongoing political tussles in which constitutional politics are invariably embedded. For example, Canadian referendums on national issues are inseparable from Quebec nationalism; and Australian referendums in the past have been pushed as a means of achieving greater national power by centralizing Labor governments. Hence recent referendums in Canada and past referendums in Australia can only be properly understood in their embedded political context and history. This is clearly a very large area for study where both detailed examination of particular country experiences and comparative analysis can contribute. As a first step, detailed case studies are necessary. I shall focus on the Australian experience and argue that constitutional referendums work moderately well, but that there are


serious problems of partisanship and spoiling that blight the process. Some of this is probably inevitable in such an important sphere of politics, but partisanship is heightened by Australia’s party and parliamentary systems. In the rest of the chapter we examine the record of popular involvement in constitutional politics, and recent attempts to enhance the process by limiting the role of politicians in favor of a popular assembly. Australian constitutional politics offers strong evidence in favor of involving the people as final decision-makers. Popular participation was an integral part of the original process of constitution-making one hundred years ago and of constitutional change during the ensuing century. The people were part of the original process in two respects: choosing the delegates to the 1897–98 convention that drafted the constitution and endorsing the draft constitution before it was formally passed by the Westminster parliament. This process of popular choice and ratification required that the form and detail of the constitution accorded with popular taste. Popular decision making was built into the mechanism for constitutional change through section 128 of the constitution. That section requires a double majority of the electors overall and a majority in a majority of the states (that is, four out of the six) to effect constitutional change. This dual requirement reflects the federal character of the constitution that preserves the states as quasi-autonomous political entities. There is an additional requirement that key matters affecting a state require approval of the majority of electors of that state. The right to initiate referendums, and hence in practice their drafting, is vested exclusively in the Commonwealth Parliament with procedures to allow for disagreement between the Houses of Parliament. The disparity between the initiation and ratification parts of the Australian amending formula helps account for the historic pattern of referendum results – many proposals put to referendum by governments but few passed by the people. This again underlines the general principle highlighted earlier that the institutions and patterns of political competition within a society, mediated by the particular form of the referendum in use, can either facilitate or constrain the process of constitutional change. It is claimed by critics that Australia has a poor record of constitutional change: many have shared Geoffrey Sawer’s 1967 conclusion that, constitutionally speaking, Australia remained ‘the frozen continent’ (Sawer, 1967, p. 206). In particular, some suggest that the requirement to use a referendum for constitutional change has constrained Australia’s ability to adjust and evolve to new circumstances. They point to the fact that only 8 out of 44 referendum proposals put to the Australian people

116

Brian Galligan

have been passed, while more than one hundred bills to alter the constitution have been introduced into parliament. For the most part, such critics have favored progressive reformist or Labor centralist policies based on concentrating power in Canberra. They usually champion the streamlining of federal and bicameral arrangements in favor of more direct majoritarian ones. This view has drawn support from a selective interpretation of the referendum record, with critics often blaming negative outcomes on the apathy and ignorance of the people. Comparative analysis, however, suggests that Australia’s record of infrequent constitutional change is not exceptional. Countries that began the twentieth century with undemocratic constitutions, or that have experienced revolutions or regime changes following military defeat, or that have to accommodate distinct linguistic communities, have records of greater constitutional change. In addition, a good deal of constitutional adjustment and change is made through judicial review by supreme courts. That aside, stable constitutional democracies like the United States, Canada, and Australia typically have modest change. Australia is distinctive not in the modesty of its constitutional change, but in the remarkable persistence of its governments – particularly Labor ones – in putting doomed centralist proposals to a vote. This disparity between hegemonic proposals, to use Lijphart’s terminology, and popular choice is the key to understanding Australia’s slim record of referendum successes (Galligan, 1995, chapter 5). Commonwealth governments typically put questions that would enhance their powers; the people are naturally suspicious of giving more power to Canberra and usually say no. This explains virtually all the referendum defeats up to 1973 (Table 5.1). Of the 28 referendum questions put between 1901 and 1973, 24 were for increasing commonwealth power; of these 24, only two passed. Those two were for powers over social services in 1946 (the constitutional basis for much of the postwar welfare state) and to make laws with respect to Aboriginal people in 1967. Both commanded majorities in all states, the former with 54 percent support overall and the latter with an extraordinary 90 percent. The 22 failures included expansion of the commonwealth’s legislative powers broadly in 1911, 1926, and 1944, and over many more specific matters. In addition to these 24 referendums, there were four proposals for changing the machinery of government or restructuring financial arrangements, three of which passed. One of these was concerned with senate elections (1906) and two with financial arrangements and state debts (1910 and 1928). The one that failed was to break the Senate nexus and decrease the relative size of the Senate (1967).

Amending Constitutions through the Referendum Device 117 Table 5.1 Year

Constitutional referendums 1901 to 1988

Proposal

Government submitting

States approving

% of electorate approving

1906 Senate elections*

Protectionist 6

82.65

1910 Finance* State debts*

Fusion Fusion

3 (Qld, WA, Tas) 5 (all exc. NSW)

49.04 59.95

1911 Legislative powers Monopolies

Labor Labor

1 (WA) 1 (WA)

39.42 39.89

1913 Trade and commerce* Corporations* Industrial matters * Railway disputes * Trusts * Monopolies *

Labor Labor Labor Labor Labor Labor

3 3 3 3 3 3

(Qld, SA, WA) (Qld, SA, WA) (Qld, SA, WA) (Qld, SA, WA) (Qld, SA, WA) (Qld, SA, WA)

49.38 49.33 49.33 49.13 49.78 49.33

1919 Legislative powers* Monopolies *

Nationalist Nationalist

3 (Vic, Qld, WA) 3 (Vic, Qld, WA)

49.65 48.64

1926 Legislative powers Essential services

Nat. – CP Nat. – CP

2 (NSW, Qld) 2 (NSW, Qld)

43.50 42.79

1928 State debts*

Nat. – CP

6

74.30

1936 Aviation Marketing*

UAP UAP

2 (Vic, Qld) 0

53.56 36.26

1944 Postwar powers

Labor

2 (SA, WA)

45.99

1946 Social services* Labor Marketing* Labor Industrial employment * Labor

6 3 (NSW, Vic, WA) 3 (NSW, Vic, WA)

54.39 50.57 50.30

1948 Rents, prices

0

40.66

Labor

1951 Communists

Liberal/CP

3 (Qld, WA, Tas)

49.44

1967 Nexus Aborigines

Liberal/CP Liberal/CP

1 (NSW) 6

40.25 90.77

1973 Prices Incomes

Labor Labor

0 0

43.81 34.42

1974 Simultaneous elections * Amendment* Democratic elections* Local government*

Labor Labor Labor Labor

1 1 1 1

1977 Simultaneous elections Casual vacancies Territorial votes Retirement of judges

Liberal/NCP Liberal/NCP Liberal/NCP Liberal/NCP

3 (NSW, Vic, SA) 6 6 6

62.20 73.30 77.70 80.10

2 (NSW, Vic) 0

50.60 47.10

1984 Simultaneous elections * Labor Inter-change of Labor powers*

(NSW) (NSW) (NSW) (NSW)

48.32 48.02 47.23 46.87

118

Brian Galligan

Table 5.1 Year

(continued)

Proposal

Government submitting

States approving

% of electorate approving

1988 Parliamentary terms Fair elections Local government Rights and freedoms

Labor Labor Labor Labor

0 0 0 0

32.92 37.60 33.62 30.79

1999 Republic Preamble

Liberal/NCP Liberal/NCP

0 0

45.13 39.34

Notes: *Referendum held at same time as a federal election. Italicized subjects achieved sufficient majorities for alteration to the constitution. Source: Constitutional Commission (1986), p. 36, and figures from Australian Electoral Commission for 1988 and 1999 results.

By 1974 it had become clear to federal politicians that proposals to increase commonwealth economic powers were a waste of time, which is not necessarily evidence that referendums make constitutional change impossible, but only that the Australian population was generally wary of centralizing power. It would also be a mistake to argue that the referendum prevents societies from undertaking necessary evolution, for governments often find other avenues for pursuing their objectives when one route – in this case, the referendum route – is blocked. Australian politicians in Canberra recognized that there was no need to continue to pursue constitutional change through doomed referendums because the commonwealth could colonize most areas of state jurisdiction through its open-ended spending power or, increasingly in recent decades, its power over treaties and the subject matter of treaties. The commonwealth government had become jurisdictional aggressor and fiscal bully of Australian federalism, aided and abetted by decisions of the High Court. This is not unlike what has happened in Canada, where constitutional recognition of Quebec’s distinct status in confederation has been stalled by the referendum process but has become a de facto reality in most government decision making. After 1974, referendums shifted entirely to machinery of government issues. Of the 16 proposals put to the people, 13 failed and three passed. Those three concerned casual senate vacancies, territorial votes, and the retirement of judges, all in 1977. But 1977 was not a clean sweep: the proposal to have simultaneous elections for the Senate and the House of Representatives failed to gain support in a fourth state. That same issue had failed before in 1974 and would fail again in 1984 and 1988.


Again there is a pattern of consistency: proposals to interfere with the separate electoral cycle of the Senate are not accepted by the Australian people, and federal governments are rather slow to learn. Hence, the referendum record does not show that Australia is constitutionally a frozen continent, or that the people are incapable of prudent constitutional politics. Rather it shows that the string of failed referendums is due largely to central governments persisting with selfserving amendments that would increase their legislative powers and decrease the independence of the Senate that is an integral part of federal machinery and in recent decades a substantial check on governments based in the House of Representatives and not in control of the Senate. The presence of the referendum device makes it difficult for governments to engage in self-serving and expedient processes of constitutional change. While it is no doubt accurate to say that constitutional change in Australia has been difficult, in part due to the presence of the referendum, this is as it should be: governments carry the burden of proof before a country’s constitutive document is changed, and the public will be hesitant unless the government has made a persuasive case.

A people’s convention Reaching a positive conclusion about the overall record of constitutional referendums in Australia should not cloud the real problems that mar constitutional politics. There was something of a recent consensus that politics and partisanship were the main obstacles to constitutional change by referendum and that involving other elites and the broader public in conventions was the way forward. This was tried in limited ways in the latest two episodes of constitutional change (1988 and 1999). Both failed badly as Table 5.1 shows. The 1988 proposals were modest changes intended as precursors to more sweeping proposals, including an entrenched bill of rights, recommended by a Constitutional Commission of eminent persons. These proposals were all defeated by record margins of two-to-one against. In part this was due to the fact that the Labor government of the time took little role in defending the proposals to which it was not strongly committed, while the Liberal-National Party opposition mounted a crusade against them. Partisanship and normal politics then – ‘We the Politicians’ rather than ‘We the People’ – were key factors in those defeats. A people’s convention in the 1999 republican referendum was a more ambitious attempt at taking normal politics and partisanship out of

120

Brian Galligan

referendums. This novel approach to constitutional referendums reflected a widespread concern in Australia that partisanship and politicians were the main spoilers in the referendum process. The remedy was a ‘people’s convention,’ touted as an independent forum of popular voice. This Constitutional Convention – dubbed the Con-Con forum – met in Canberra in February 1998 and was ‘to consider whether Australia should become a republic, which republic model should be put to voters to consider against the present system of government, and in what timeframe any change might take place’ (Constitutional Alteration (Establishment of Republic) 1999, Explanatory Statement, p. 6). In other words, the Convention rather than the Commonwealth government was to test whether there was sufficient consensus for a referendum and devise the model for a new head of state (Galligan, 1998; Nathan, 1998). The model sought to remove the initiation process from the hands of governments and put it in the hands of a more popular body, though without moving to a model of citizen initiative. The convention’s selection was novel and, as a result, its composition quite different from parliament. Half the delegates, or 76, were appointed by the prime minister who used the procedure to ensure the representation of special groups and individuals as well as political leaders and parliamentarians. Forty of the appointed delegates were senior parliamentarians, including all the premiers and opposition leaders of states and territories, with political affiliations in aggregate reflecting roughly the party composition of the various parliaments and assemblies. The other 36 appointed delegates included representatives of special groups, such as indigenous people, immigrants and youth, as well as eminent public figures. The election for the other half of the delegates was also a departure from usual electoral practice. Electors voted by mail rather than personally at a polling booth, voting was not compulsory, and candidates were listed in a separate booklet and not on the ballot paper. With the established parties not contesting the election, a kaleidoscope of individuals and groups put themselves forward: there were 609 candidates in 80 groups and 176 ungrouped candidates. Perhaps because of these arrangements, but probably because of lack of enthusiasm for the Convention and republicanizing the head of state, only a minority of Australians, or 47.9 percent, bothered to vote. The majority of those elected supported relatively well-defined positions: the Australian Republican Movement delegates supported a precise ‘minimalist’ model for change, ‘real republicans’ were pledged to an elected head of state and broadening the debate to more substantive issues of reform, and constitutional monarchists


championed the status quo. There were also idiosyncratic delegates with no obvious affiliation to established groups (Australian Electoral Commission, Your Guide to the Constitutional Convention, 1997). Given the diversity of the convention delegates and the absence of party discipline, it is perhaps surprising that the convention worked as well as it did. That was probably because the Australian Republican Movement and its dominant leader, Malcolm Turnbull, had sufficient support and organizational skills to focus attention mainly upon their well-publicized model. Moreover, this ‘minimalist’ model for an appointed head of state had been in the public domain for several years and received broad support from elites. The main challenge came from the ‘real republicans’ who wanted direct election of the head of state and more extensive constitutional change, but finally achieved only the commitment to hold further conventions in the future. Others supported the ‘ultra-minimalist’ model of Richard McGarvie, previously Governor of Victoria, who attended as an appointed delegate (McGarvie, 1999). With the prime minister attending, and other senior federal ministers and premiers ‘coming out’ on the republican issue, the convention had extensive media attention. The outcome was more or less as expected. There was majority support (59 percent) for Australia becoming a republic and overwhelming support (88 percent) for putting the question to referendum. The snag, however, was that there was no clear majority for the preferred model of selecting the head of state, which would have mandated that the head of state was selected by the prime minister and endorsed by a twothirds majority of both houses of the Commonwealth Parliament. Moreover, the head of state office was to be left with the traditional unspecified powers of the governor-general, but subject to instant dismissal by the prime minister. This ‘minimalist’ model was finally supported by 48 percent of convention delegates, opposed by 37 percent with 15 percent abstaining. Thus, while most wanted a referendum and a republic, there was no consensus on the model to replace the monarchy. The model devised by the Convention, albeit with only minority support, was put to the Australian people on 6 November 1999 (Constitution Alteration (Establishment of Republic) 1999, Exposure Draft Bill; Uhr, 1999). It was defeated in all states and the Northern Territory, receiving only 45.13 percent support overall. The sole jurisdiction that endorsed the proposal was the Australian Capital Territory, and Victoria had the highest state support of 49.84 percent. Although this referendum was quite unique in its subject matter as well as its origins in the convention, the outcome was predictable (Galligan, October, 1999).

122

Brian Galligan

Opinion polls showed consistently that the majority of Australians favored a republic, and a larger majority favored electing the head of state. This is explained by the fact that some monarchists prized the monarchy for its independence and saw the best republican alternative as a head of state selected independently of the prime minister. The main reason for the failure of the republican referendum is quite simple: the people were presented with a model for a new head of state that the majority did not support (for a contrary view see Turnbull, 1999) and which, on balance, was not obviously better than the status quo. Although advocated by many prominent elites and the Labor Party, the model was savaged by others while the prime minister remained a staunch monarchist. The whole exercise of the Constitutional Convention and republican referendum was fraught with ironies and contradictions. The convention itself was not really a people’s convention in any directly democratic sense, but an alternative representative assembly composed of appointed and elected delegates. Although partisanship was largely banished from the process, amateurish groupings of selfappointed elites proved a poor substitute for professional politics. The republican model itself was curiously flawed: the new president was to be apolitical with practising politicians explicitly precluded, but through appointment and dismissal procedures would be the creature of politicians. Given all of this it was hardly surprising that the Australian people should reject the proposal. If anything, this is further evidence of their serious sense in constitutional politics. Again, Canadian and Australian experiences reflect one another. Much elite commentary in Canada following the defeat of the broad package of constitutional reforms proposed in 1992 (the Charlottetown Accord) suggested that the Canadian people could not gather the will or good sense to approve necessary constitutional change. In retrospect, an alternative perspective has taken root: the Charlottetown Accord was deeply flawed, it was the product of partisan politics, it responded to the exigencies and short-term considerations of political leaders of the day, and it did not reflect a broad popular consensus on a constitutive document. As with the Australian republican option, the Charlottetown Accord was not clearly better than the status quo and may in fact have been worse. Elites often become convinced that constitutional change is necessary, and once this assumption takes hold it is very difficult for them to restrain their constitutional ambitions or abandon their projects. The public tends not to get swept up in such processes, comes in at the end of the day, and often concludes that there is either no pressing need for change, or that the proposed change will not clearly address


the problem without creating additional ones. In this way, while it is certainly true that referendums can make the process of constitutional change difficult, they also make it more likely that constitutional amendments have broad popular support before they are passed.

Conclusion Australian experience with constitutional referendums for a century clearly supports the proposition that the people are responsible and capable of making prudent decisions. Proponents of various sorts of changes over the decades think otherwise and blame the pattern of failure on popular ignorance and unthinking conservatism. But examination of the record and the politics of referendums, as sketched above, proves the contrary. The attempt to banish political partisanship that is often a significant negative in referendum campaigns in favor of a ‘people’s convention’ was a qualified failure, producing a disorderly array of amateurish substitutes. There is no substitute for political and popular consensus if constitutional change is to succeed. Achieving such consensus is hard – elites acting alone might more easily achieve agreement – but that degree of difficulty is not inappropriate for changing the fundamental institutions of government in a constitutional democracy. Obviously such a conclusion depends on one’s accepting the normative theory of constitutional referendums outlined in the early part of the chapter. Changing the constitution requires a higher level of consensus than changing ordinary law, and under these circumstances, the difficulty experienced by Australia in successfully amending its constitution is not unreasonable. This chapter has highlighted that certain kinds of constitutional changes are difficult to pass through the referendum. In federal countries, centralizing measures appear to have a difficult time, while changes to the institutions of government may fare slightly better. More fundamentally, constitutional change via the referendum device usually requires broad consensus across political parties. This is not such a bad thing: when parties are divided, or feel that they can exploit existing divisions within the society for political advantage, perhaps the time is not ripe for changing a country’s constitutive document and basic law. This is of course the opposite from what has happened in many of the American states, where citizens have the power to initiate constitutional referendums and pass change with a bare majority. In the process they have transformed ordinary politics into constitutional politics by ratcheting up the number of citizen initiatives that in reality deal with normal

124

Brian Galligan

legislation but are passed as constitutional amendments, often with difficult conditions preventing their future amendment. Whether and to what extent Australia’s experience with popular participation in constitutional politics can be generalized depends on further study of particular country experiences and comparative analysis. All countries are special cases, but relevant characteristics of Australia’s liberal-democratic and federal constitutional political culture are not unique. Therefore, one should suppose that constitutional referendums are not only a good thing per se, but are workable in many countries with comparable conditions.

6 Popular Control of Referendum Agendas: Implications for Democratic Outcomes and Minority Rights Shaun Bowler and Todd Donovan

One of the major concerns voiced repeatedly in discussions of direct democracy is that it raises the possibility of abusive majority rule (Gamble, 1997; Bell, 1978; Butler and Ranney, 1994; Eule, 1990; Linde, 1993; Magleby, 1984). Three major assumptions underlie this concern. The first is that the process of direct democracy allows a majority relatively free rein over drafting and implementing laws. The second is that popular majorities are expected to be intolerant of minorities and civil rights, and especially intolerant when compared to elites. A third, albeit often implicit, assumption is that elites play little or no role in a process which is shaped by (intolerant) voters themselves. In this paper, we examine each of these assumptions. We demonstrate that nearly all forms of direct democracy require legislative elites to draft policies, and suggest that this constrains the potential for directly abusive effects of referendums on minority rights. The assumptions both of elite tolerance and also of elites as passive bystanders are thus brought into question. Because it is the initiative, as used in Switzerland and the US, that conforms most closely to the paradigmatic version of popular majority rule, we examine the initiative in detail. We find that citizens voting on these measures are not as intolerant as many commentators fear.

Minorities and majority rule At first blush the empirical track record of direct democracy in the United States and Switzerland might be seen as decidedly anti-minority. From 125

126

Shaun Bowler and Todd Donovan

1970 to 1996, there were at least 21 federal and canton level initiatives and popular referendums about foreigners and their rights in Switzerland (Frey and Goette, 1998; Linder, 1994, p. 97). Earlier in the century, antiSemitism was given voice in Switzerland’s initiative process in a proposal to prohibit the killing of animals by kosher rites (Steiner, 1995, p. 133). In the US, during the period 1978 to 1998, no fewer than 13 anti-gay initiatives appeared on ballots in seven different states – to take but one example of minority rights under threat. In part this was only interrupted by a United States Supreme Court ruling that reversed a voter-approved Colorado initiative (Romer v. Evans, 1996). The ruling, however, has failed to halt completely attempts to use the initiative process to roll back gay rights (Donovan et al., 1999). On the other hand, there have been several instances of attempts to use direct democracy to advance minority rights of varying kinds. Steiner (Steiner, 1995, p. 134) notes the Swiss example of attempts to extend the rights of conscientious objectors (an ‘opinion minority’). Similarly, in the US, gay rights advocates have sometimes tried to use the process to advance gay rights (Donovan et al., 1999). The Swiss and US experiences provide somewhat conflicting examples of the use of the initiative as a vehicle for advancing women’s rights (Banaszak, 1996). Yet these scattered examples are rare, and usually fail to pass. Attacks on rights are feared to be more frequent and more successful. To some extent, citizen’s initiatives attacking gay rights in the US represent a continuation of earlier cycles of political activity directed at other minorities (Cronin, 1989). Movements against various minorities via the statewide initiative process have waxed and waned throughout the twentieth century, often corresponding to more generalized social backlashes associated with political gains of specific minority groups (Donovan and Bowler, 1998b; Gamble, 1997). The legislative gains made by African-Americans in the 1960s as a result of the racial desegregation of schools and the integration of housing and accommodations, for example, spawned a number of state and local ballot measures whose intention was to reinstate the previous order (Bell, 1978; Wolfinger and Greenstein, 1968). Similarly, increased accommodation for language minorities led to a series of ‘Official English’ initiatives in the 1980s (Citrin et al., 1990). These have been accompanied by a more recent series of proposals targeting ethnic minorities by placing limits on the provision of government services for illegal or undocumented immigrants, on affirmative action and on bilingual education (Tolbert and Hero, 1998; Tolbert and Hero, 1999). In each of these areas, policies benefiting minorities were subjected to direct public evaluation after political activists

Popular Control of Referendum Agendas 127

succeeded in gathering enough signatures to force a public vote. And it was only subsequent intervention by the courts which halted or slowed down many of these attempts to roll back the gains minorities had made. Anecdotally, then, American and Swiss-style direct democracy has provided a vehicle for a series of majoritarian attempts to roll back minority rights – rights which were often introduced by more tolerant legislatures. Hence it is not surprising that many of the authors we noted at the outset voice concerns about the safety of rights, and especially minority rights, under direct democracy (see also Cain, 1992; Papadopoulos, 1995). These worries are based on the concern that mass views are markedly intolerant, with writers such as Stouffer (1955) suggesting that individuals’ responses to outgroups were almost universally intolerant. Toleration only evolves after individuals devoted time and effort to considering the broader implications of any proposed repression (Wenzel et al., 1998), something which may be absent during the initiative process. If one assumes that minorities receive more favorable treatment from elites than from the mass public, then the move toward the more direct forms of democracy – as depicted in Table 6.1 below – has struck some writers as an especially worrisome development for minority rights (see, for example, Putnam, 1976). Despite these long-standing concerns, it has not been empirically established that direct democracy necessarily produces outcomes that are decidedly more anti-minority than those produced by legislatures (Donovan and Bowler, 1998b; Cronin, 1989). Furthermore, it is even less clear that direct democracy operates without the active and important intervention of the same political elites who dominate representative institutions. In what follows we address each of these points. We argue first that, depending on the form of direct democracy, elite intervention is key in determining outcomes in the actual practice of direct democracy. It is therefore misleading to worry about majority tyranny without considering the role of elites in fostering tolerance. Second, in taking the paradigmatic case of the initiative process and minority rights – where elite influence is at its (relative) weakest – we show that mass opinion is not excessively intolerant of minority rights.

Elites in direct democracy Direct democracy as it is typically practiced in the USA and Switzerland is but one variation on how popular participation in legislation might be structured. Different provisions for direct democracy can affect how much the process might be used as a vehicle to abuse rights. Nations

128


differ in terms of subjects that may be put to a popular vote and the type of majority required for approving a ballot question. There is also great variation in rules about how much discretion elected representatives have in structuring ballot questions and whether decisions reached by the public are binding or not. Combined, these differences affect how much elites as opposed to citizens themselves control the agenda on questions of minority rights. A brief overview of some of this variation is provided in Table 6.1. Forms of direct democracy are listed in a rough rank-order from those where elites in government have total control over the referendum agenda to those where elites have relatively little control. We arrive at this ranking by considering if proposals originate inside or outside the legislature, and if popular approval of a measure is binding on the government. At one extreme of elite control are authoritarian plebiscites put forward by an anti-democratic government for the purposes of legitimating its own stranglehold on power (see Lupia and Johnston, this volume). At the other end of the continuum are direct popular initiatives. Here voters themselves set the agenda and exercise veto power. Other procedures are ranged in between these two endpoints and show varying degrees of elite agenda-setting power. Table 6.1 Variation in elite control over procedures for direct democracy (ranked in order of agenda-setting influence) Plebiscite:

Examples:

Advisory referendum:

Examples:

Legislative referendum:

Vote to endorse anti-democratic government and/ or approve broad (extra-constitutional) changes proposed by government/head of state. France (Napoleon I and III); Romania (Antonescu); Germany (Hitler); Serbia (Milosevic, 1990); Belarus (Lukashenko, 1994). Non-binding vote on policy issue drafted by legislature, placed on ballot by act of legislature or government. Government/legislature has discretion in responding to popular vote. Spanish vote on NATO, 1986; Sweden and nuclear power, 1980; Finland; used in 8 other European countries that also allow binding referendums; New Zealand and electoral system reform, 1992 (first vote). Binding vote (de jure or de facto) on policy issue drafted by legislature, placed on ballot by act of legislature. Electorate responsible for approving legislation.

Popular Control of Referendum Agendas 129 Examples:

Compulsory referendum:

Examples: Popular referendum:

US states; Switzerland; New Zealand; Australia (federal and state); France and Senate reform, 1969, Danish vote on Maastricht, 1992; Irish vote on divorce, 1986; Welsh and Scottish votes on devolution, 1997; Canadian vote on Charlottetown, 1992. Constitution forces public vote if legislature seeks to change certain spending or revenue policies; requirement to put proposed constitutional amendments to a vote. Swiss cantons, Australian constitutional changes, Alberta (Canada). Binding vote on policy issue previously drafted by legislature, placed on ballot by act of citizens. Provides a popular veto of existing policy.

Examples:

Switzerland; numerous US states and cities since 1898; Italy since 1970. .................................................................................................................................. Advisory initiative:

Examples: Indirect initiative:

Examples:

Direct initiative: Examples:

Non-binding vote on policy proposal initiated outside of legislature. Citizens place proposal on ballot, legislature has discretion in response. New Zealand, 1990s. A legislative vote on policy proposal is made to take place by decision of voters. Legislature may respond by approving, amending or sending proposal to voters. Maine (US), Massachusetts (US); Mississippi (US), Wyoming (US). Some US states use direct and indirect initiatives. Binding vote on policy initiated outside of legislature. Decision by voters alone. States in western US; Switzerland (constitutional only); Italy; British Columbia (never used).

Among these many variants of direct democracy, a major distinction is that between the legislative referendum, where votes are largely over policy issues drafted within a legislature, and the citizen’s initiative, where voters make decisions about policies drafted by groups outside of the legislature. It is this second variety of direct democracy (listed below the dotted line in Table 6.1) which allows the agenda to be set, in varying degree, by voters themselves. Even within this category, however, considerable variation exists over the degree of influence that voters

130


have. Some procedures may only allow for proposals which are advisory to the government. Others may simply require the legislature to vote on a proposal (but not necessarily to pass it). More obviously effective participation is given when citizens can either propose or negate laws directly. Some mixtures of these procedures may be present within a given polity. Taken together, the list in Table 6.1 illustrates that there is not one single institution of direct democracy but rather, a range of institutional arrangements that might provide different avenues for popular majorities to abuse the rights of minorities. In most variations of the referendum, politicians essentially control the agenda and place ‘take-it-or-leave-it’ offers before the electorate, where details of the proposal have been negotiated among parties within a legislature. Sometimes, it is true, as in the case of the 1969 French referendum on senate reform, politicians may guess wrong about the state of popular opinion. Legislative referendums might also fail when politicians overextend their request and roll so many items into a proposal that they inadvertently expand opposition – as in the case of the 1992 Charlottetown Accord vote in Canada (Johnston, 1993). Sometimes, politicians may lose simply because they are forced to hold a referendum when such a process was not their first choice (as in the Danish vote on the Maastricht Treaty). But by and large, ‘take it or leave it’ offers favor the actors making the offer. Legislative referendum proposals in much of the world pass more often than fail, and are particularly successful when there is elite and party consensus (Butler and Ranney, 1994, pp. 5–6). Under processes such as the initiative, however, citizens can make any proposal they want subject to constraints imposed by constitutions and the practical problems of collective action. By ‘citizens’ we mean any group, interest, or entrepreneurial individual who assumes the organizing costs of qualifying a measure for ballot status. In its most pure form, the direct citizen’s initiative allows non-legislative actors to ‘move first’ in the agenda-setting game by directly drafting the policy choices that voters will face. Under this process, the legislature is thoroughly excluded from framing the issue and, as with popular constitutional proposals in California, from subsequently amending any policy that voters approve (Ferejohn et al., 1995; Gerber, 1995). Unless the process is advisory, as in New Zealand, a citizen-drafted and voter-approved initiative becomes binding law without legislative input or action. Some of the processes described in Table 6.1 present more complex pictures of agenda control, as the sequence of moves made by legislators and citizens matter much more in terms of who ultimately shapes policy. Under the popular referendum politicians move first in passing


legislation and then citizens (with large organizing costs) seek a veto of the legislature’s policy at the election. Under the indirect initiative, citizens move first in drafting a proposal and then offer it to politicians for adoption. If the legislature does not wish to adopt the indirect initiative, it is then presented to voters, possibly with an alternative proposal drafted by the legislature (Magleby, 1984). It is reasonable to expect that variations in agenda control under different forms of direct democracy provide different degrees of opportunity for popular majorities to abuse the rights of minority groups within a polity. By and large, most of the procedures described in Table 6.1 leave governments and legislators with the upper hand in that it is they who move first and frame the decisions presented to voters. By contrast, the initiative process gives the greatest leeway to voters to be directly repressive of other sections of society. While all other variants of direct democracy provide for the intervention of elites to express, amplify, or muffle anti-minority aims, under the initiative process the participation of elites is not required. This is not to say that elites are irrelevant during the initiative process. Elites play a central role in shaping public opinion in the initiative process by endorsing or opposing initiatives (Bowler and Donovan, 1998; Karp, 1998). Even if not agenda setters, elites may act as opinion leaders. Elites may, for example, provide cues for voters who lack information or pre-existing views on a proposal (Lupia, 1994, 1992). The nature of divisions among elites can be critical, since visible cleavages between rival parties send voters clear signals that allow them to understand the referendum in partisan terms. Visible cleavages within rival parties, however, send a far muddier signal to voters. In such circumstances voters find it harder to orient their own opinions on a given issue by use of a standing party attachment (Franklin et al., 1994). This relationship between elite consensus and public opinion formation also follows when the issues at stake are ones of rights. We have suggested elsewhere that consensus among elected elites against anti-minority initiatives lead to the defeat of such measures, while tacit elite endorsements lead to greater popular support (Donovan and Bowler, 1997). Voters, no doubt, are motivated by more than just elite cues. We might anticipate that factors such as animus toward a targeted group, racism or ‘symbolic’ racism (Kuklinski et al., 1997; Kinder, 1986), and perceptions of minorities as a ‘threat’ (Hero, 1998; Giles and Buckner, 1993; Giles and Evans, 1986; Key, 1949) also structure decisions. This could cause race, religion, or ethnicity to be major fault-lines dividing supporters and opponents. Studies of support for anti-minority initiatives in the US have found that race and group status affect voting even when other forces

132


were held constant. Whites voted against affirmative action more than other groups (Alvarez and Butterfield, 1998), ‘born again’ Christians voted for restrictions on gay rights (Donovan and Bowler, 1997), and whites were more supportive than Asians and Hispanics of pro-English language policies (Citrin et al., 1990). Whites were also more supportive of limits on services to immigrants (Tolbert and Hero, 1998) and of ending minority hiring practices (Tolbert et al., 1999). Yet party identification was still a major factor in most of these cases, illustrating the potential for partisan exploitation of anti-minority referendums and initiatives. Elected officials do more than provide voters with a way to orient their own positions; they often use the initiative process themselves. In many places, including Switzerland, Italy, and the US, no formal rule prevents elected officials from promoting their own initiatives. In California, for example, the highest ranking elected Republican (Pete Wilson) personally championed initiatives (Smith, 1999) targeting welfare recipients (in 1992), service provision to immigrants (in 1994), and minority hiring preferences (in 1996). The first popular referendum in Italy (to repeal divorce laws) was facilitated in 1970 by the Christian Democrats after they became a minority in Parliament and could no longer block such legislation (Bogdanor, 1995, p. 65). The use of the initiative or referendum by small Swiss parties ‘often influences the large parties to anticipate the demands of small parties’ (Steiner, 1974, p. 19; see also Linder, 1994, p. 118) and major opposition parties use the process as an obstructionist technique (Kobach, 1994, p. 130). Minority parties have also used initiatives to incorporate themselves into government. In Switzerland, nongovernment parties (Conservatives, Socialists, and the Agrarian Party), ‘frequently and successfully made use of the referendum against legislation. As a result, they were gradually integrated into the government to render less uncertain the public’s reaction to proposals emanating from the state’ (Papadopoulos, 1995, p. 428; see also Linder, 1994, pp. 118–19). Substantial evidence thus points to the way in which, rather than being subject to the discipline of voter-inspired initiatives, politicians often use the initiative as a vehicle for their own ends. In summary, we have argued that direct democracy – including the citizen’s initiative – grants a key role to elites, one which is often ignored in discussions of the relationship between direct democracy and rights. In the next section, we consider the potential for initiatives to directly affect rights by examining in more detail the evidence on voter approval of anti-minority measures from the two places that use direct citizen’s initiatives frequently enough to provide a reasonably lengthy empirical record: Switzerland and the American states.


Voter approval of anti-minority initiatives In the US, Gamble’s study (Gamble, 1997) evaluated the electorate’s propensity to approve anti-minority initiatives. Summarizing three decades of initiatives in housing and accommodations desegregation, AIDS testing, gay rights, language policy, and school desegregation, she concluded that there is strong evidence that popular majorities use the initiative to deprive minorities of their civil rights at the local and state levels (Gamble, 1997, p. 246). We have suggested elsewhere (Donovan and Bowler, 1998a; Donovan and Bowler, 1998b) that these conclusions may be overstated, and that the likelihood of an anti-minority (specifically anti-gay) initiative passing depends in part on the homogeneity of a jurisdiction’s population and the relative proportion that is college educated. More importantly, it depends on the size of the population. Where communities are relatively homogeneous – yet still containing some small minority group – it seems reasonable to suppose that majority tyranny is more likely than in places where communities are diverse. A jurisdiction’s degree of homogeneity is likely to correlate with population size, since larger communities are often more diverse and include numerous minorities. In diverse communities where no single group forms a monolithic majority, it should be more difficult to piece together a big enough voting bloc to be overtly repressive toward any single group. This logic mirrors that employed by James Madison, who feared small, homogeneous publics when developing his constitutional scheme for controlling majority tyranny (Madison, [1787] 1961, p. 83). Specifically, US anti-minority initiatives pass with relatively high frequency only at the local level, particularly in smaller places. Gamble found 78 percent of the 74 state and local anti-minority initiatives she identified were approved between 1959 and 1993. The vast majority of these (48, or 65 percent) were local measures, including many anti-gay initiatives from small rural towns, and there is no evidence that the initiative results were different from those produced by municipal councils in similar places that have no provision for the use of the initiative. Since small communities are ones where it is, presumably, easier for deliberation and discussion to take place, this finding also has uncomfortable implications so far as many proponents of deliberative democracy are concerned. As Table 6.2 illustrates, 26 of the state-level initiative measures listed on US state ballots from 1959 to 1996 overtly target civil rights of minorities (Gamble, 1997). This represents about 4 percent of the

134


Table 6.2 Popular approval of anti-minority initiatives and referendums in Switzerland and the USA (USA 1959–95; Switzerland 1970–96) USA

Switzerland

n/a

20 6% 20%

State level/canton of Zurich Number of anti-minority measures % of all initiatives anti-minority % of these with anti-minority result % anti-gay approved

26 4% 46% 38%

13 4% 62% n/a

Local level/city of Zurich Number of anti-minority measures % of all initiatives anti-minority % of these with anti-minority result

< 48 unknown 78% (est.)

Federal level Number of anti-minority initiatives % of all initiatives ‘anti-minority’ % of these with anti-minority results

31 unknown 23%

Note: For the US, ‘anti-minority’ measures include those that limit desegregation of accommodations and schools, measures limiting rights of gays and lesbians, English-only laws, and restrictive AIDS measures. US local estimated from Gamble (Gamble, 1998) based on incomplete data. For Switzerland, measures include policy toward foreigners and immigrants, prohibitions on the actions of certain groups (Catholics, Freemasons), drug policies, religious liberties and public housing. Sources: USA: Gamble (1997); Donovan and Bowler (1998); Switzerland: Frey and Goette (1998). Additional calculations based on Magleby (1994).

approximately 670 initiatives on state ballots during this period (Magleby, 1994, p. 232). When we examined results in one state policy area – anti-gay rights initiatives – we found significantly more tolerance than illustrated in Gamble’s sample of small rural towns. US state-level voters are no more likely to pass anti-gay measures than other measures. Thirty-eight percent (5 of 13) were approved from 1978 to 1998 (Donovan et al., 1999), which is identical to the 38 percent approval rate for all state initiatives from 1898 to 1992 (Magleby, 1994). This is a lower approval rate than found for other ‘majoritarian’ issues that dealt with broad social and moral questions. In California, 58 percent of majoritarian measures were approved between 1986 and 1996 (Donovan et al., 1998). When anti-gay initiatives are considered with all statewide anti-minority (civil rights) initiatives identified from 1978 to 1998, just under half of the measures were approved by voters. Most of those that were approved have been overturned by courts or acts of the US Congress (Donovan and Bowler, 1998b, pp. 268–9).


These patterns should not be interpreted as casting a rosy hue on the American initiative process. Some observers in the US suggest that the mere presence of the citizen’s initiative can threaten (or encourage) legislatures to act in ways that more closely match voter preferences – in other words, legislatures in initiative states might be more likely to translate existing anti-minority preferences into policy, regardless of whether such measures reach the ballot (Gerber and Hug, this volume; Gerber, 1996; Matsusaka, 1995, 1998). Similar observations of indirect effects of initiatives have been made in Italy and Switzerland (Bogdanor, 1995, p. 65; Hoffmann-Nowotny, 1985, p. 214; Linder, 1994; Trechsel and Kriesi, 1996, p. 185–200). Furthermore, some anti-minority initiatives may have secondary effects where, even while failing to alter policy, they build hostility to targeted minorities (Wenzel et al., 1998). Moreover, initiatives that do not explicitly affect civil rights can nevertheless be seen as anti-minority. For example, measures for public services for minorities might receive more support than those expressly targeting civil rights. The conclusion we wish to draw from such a record is not so much that the initiative process is friendly towards minorities, but that it is not as hostile as its critics contend. Nor is it necessarily more hostile than elite opinion allows. Indeed, as we argued above, under some conditions it is as hostile as elites allow or encourage. It is worth remembering that many dramatic recent examples such as California’s Propositions 187 (immigration) and 209 (affirmative action) were embraced, backed, and in some instances largely authored by party elites. Insurgent outsiders and populist ‘grassroots’ activists – the ones we are supposed to suspect as agents of majority tyranny – have had less direct success with antiminority initiatives than figures within the major parties. In Switzerland, a systematic investigation of the pass rate of antiminority initiatives from 1866 to 1996 showed that, if anything, Swiss voters are even more tolerant of minorities than American voters. Historically, only 10 percent of Swiss initiatives have been approved (Kobach, 1994), as have 50 percent of constitutional referendums (Frey and Goette, 1998). A systematic investigation of all measures appearing on Swiss ballots and in Zurich (city and canton) indicates that anti-minority measures are rare (6 percent of all federal initiatives), and from 1970 to 1996 were rarely approved either at the federal level or in the city of Zurich (Frey and Goette, 1998). Frey and Goette find that of 20 federal votes on minority issues, four (20 percent) had anti-minority outcomes. In 11 cases from 1866 to 1996 where the public was asked to restrict minority rights, all measures failed. Conversely, 11 federal measures

136


from this period asked voters to broaden civil rights for minorities and eight (72 percent) were approved. At the local (city of Zurich) level, 31 civil rights initiatives and referendums appeared on ballots between 1970 and 1996. Twenty-eight of these measures proposed a change in the status quo that expanded civil rights. Zurich voters approved 23 of these 28 expansionist measures. By far the most common product of initiatives and referendums in Zurich then is the extension of civil rights. Only seven of the 31 popular votes (23 percent) had what Frey and Goette categorize as an antiminority result, and most of these were cases where the vote maintained a status quo condition. These results led Frey and Goette to conclude (1998, p. 1346) that, ‘While civil rights opponents might be given more possibilities of attacking civil rights, their record of success is poor. Rather, the proponents [of minority rights] seem to benefit from direct democratic institutions.’ When these Swiss results are compared to those from the US, it appears that minorities fare best at the polls in places where the voting population is relatively heterogeneous and diverse. In Switzerland as in the US, opponents of civil rights had the greatest success via direct democracy in relatively rural (canton of Zurich) rather than urban places (city of Zurich). This discussion of direct democracy and minorities suggests that although the initiative process has been used by insurgent popular majorities (or party elites) against minority rights, it is not necessarily used in that way often or successfully. Indeed, the fact that the overwhelming majority of initiatives concerns issues such as taxation rather than minority rights is circumstantial evidence that the repression of minorities is not a pressing concern or that the majority does not find the initiative a particularly useful instrument of repression. Furthermore, it should be noted that in most democratic settings, elites have many other institutional mechanisms available to them for abusing minorities if that was their goal – through laws, regulations and the courts. Hence discriminatory or repressive elites would not typically need to have their actions rubber-stamped by referendums. Direct democracy is only one part of the democratic puzzle, and it is a part that interacts with and depends on all the other pieces. Patterns of discrimination can be seen under a variety of institutional regimes, including those that do not embrace direct democracy. Still, given the number of times that the initiative process has been used against specific sections of society one might expect to see that minorities targeted by initiatives would be hostile to the process. Even if most state and national attempts to use the initiative process against


minorities failed, either at the polls or in the courts, it would still be readily understandable if minorities greatly disliked the initiative process itself. We examine below whether minorities dislike the process, but first note that this only provides part of the picture. That is, to look at direct democracy in isolation from other components of the democratic process is to engage in a partial analysis. Although this may be the only course of action given data limitations, ideally one would compare what minorities think of both representative and direct democracy in light of our earlier comments.

Public opinion towards the initiative process Survey evidence shows high and persistent support for the initiative as a process, at least in California and Washington. Roughly speaking, about one third of California voters and one fifth of Washington voters remain skeptical of the process.1 One straightforward way of looking to see how repressive minorities find the initiative process to be is to see if, in fact, minorities comprise the bulk of those people who dislike the process, especially as compared to legislative politics. Here we reach the definitional point which we have ducked so far: which minorities should be opposed to the process? Given that the initiative process has been used against a wide range of different groups we would ideally like indicators for several different kinds of minority status. A 1996 California Poll (1996–04) included a series of instruments that are especially relevant to this study. In addition to ethnicity, this poll also recorded the language of interview. Although we may expect non-Anglo voters to dislike the initiative by virtue of their minority status, we may expect Spanish speakers to be especially fearful of the process – and hence hostile towards it – in light of California’s ‘Official English’ initiative, its 1996 anti-affirmative action initiative (Proposition 209), and its 1994 initiative affecting services to immigrants (Proposition 187). The survey also asked voters their sexual orientation, thus allowing us to test directly whether gays dislike the process. Since 1978, five initiatives affecting gays appeared on California ballots, one in 1978 limiting the rights of gay teachers, and four others during the 1980s targeting AIDS victims. A number of other markers for minority status were also included in this survey. One of these was religious affiliation – since Christians make up a majority of the religious faithful we might expect non-Christians to be wary of the initiative process. Some Christian religious groups have used the initiative in US states to target gays and lesbians and to advance their moral agenda. As such, Christians might prefer it as

138


a means of government. The county of residence of respondents marks an important divide in the state. With the bulk of the population living in a handful of counties in the smoggy south, initiatives can really only pass with sizable support in Los Angeles and Orange counties. Voters in the rural north or the Central Valley are likely to have different views on land use, animal management, and water rights than the suburban voters of the south. Similarly, left-wing voters in the San Francisco Bay Area constitute a minority interest in the face of these more conservative counties. Hence it seems plausible to expect to see a regional division of opinion. We may therefore also speak of opinion minorities. Although the discussion so far has concentrated on standard demographic fault lines of minority politics, we should also note that opinion groups constitute a minority, too. It is possible for a demographically heterogeneous but ideologically coherent group to be in a permanent minority. In California politics membership in these groups is reflected, at least in part, by registration with one of the several minor parties in the state.

Some results of models of public opinion Table 6.3 presents results from some very straightforward statistical models of the expectations outlined above. Overall, these results are consistent with the view that some minorities are sensitive to the initiative process to the point of disliking it. However, not all minority groups are hostile towards the process, nor is the hostility to the process by some groups expressed across the board. By the same token, ‘majority groups’ are not overwhelmingly in favor of the process. If it is the case, then, that minorities are aware of who they are, the majority does not seem especially self-aware of who they are, nor that the initiative process advantages them as a group. In more detail, Table 6.3 presents results from seven logit equations. The first equation is a general assessment of the initiative process in terms of whether the respondent finds it a good or a bad thing. For this first column, respondents are coded 1 if they like the process, 0 if otherwise. The following columns are estimations from a series of questions where voters were asked to make comparisons between representative and direct democracy. Voters are asked which process can be expected to make the ‘right’ decision (column 2); which process is dominated by special interest groups (column 3); which process is best suited to deciding big government programs and projects (column 4); which process enacts more coherent policies (column 5); which is best suited to a thorough review of proposals (column 6); and which is best suited to deciding technical

Age

Spanish speaker

Education

Woman

Anglo

Gay

0.27 0.17 –0.46 0.41 –0.13 0.57 –0.06 0.14 –0.15** 0.03 –1.58** 0.30 –0.003 0.004

Do right thing?

Initiative good? –0.62** 0.32 0.46 0.65 0.17 0.33 –0.38 0.26 0.18** 0.05 1.25** 0.49 0.02** 0.007

2

1

–0.14 0.17 0.79** 0.42 –0.36** 0.16 –0.002 0.14 –0.16** 0.03 –0.39 0.29 –0.004 0.004

Influenced by interest?

3

–0.03 0.15 –1.05** 0.42 –0.09 0.16 –0.09 0.13 –0.09** 0.03 –0.99** 0.29 –0.0009 0.004

Decide on Big Problems?

4

Minority opinions towards direct versus representative democracy

Republican

Table 6.3

0.15 0.16 –0.45 0.43 –0.16 0.16 0.015 0.13 –0.18** 0.03 –0.17 0.28 0.002 0.004

Enact coherent policies?

5

0.11 0.16 –0.07 0.42 –0.43** 0.16 –0.05 0.14 –0.22** 0.03 –0.86** 0.30 0.003 0.004

Thorough review?

6

0.35** 0.16 –0.34 0.46 –0.41** 0.16 0.02 0.14 –0.20** 0.03 –0.27 0.29 –0.0003 0.004

Technical issues?

7

139

0.08 0.14 –0.48* 029 0.26 0.17 0.27 0.33 950 47.3** 69.6

Influenced by interest?

3

0.09 0.13 0.26 0.25 0.10 0.15 1.005 0.31 948 26.3** 62.03

Decide on big problems?

4

–0.09 0.13 0.36 0.25 –0.13 0.15 0.99** 0.31 933 41.1** 57.4

Enact coherent policies?

5

0.07 0.14 0.48** 0.25 0.38** 0.16 0.68** 0.31 930 70** 61.2

Thorough review?

6

7

–0.07 0.13 0.08 0.25 –0.0 0.15 0.84** 0.31 948 52.7** 62.2

Technical issues?

Note: Dependent variable = 1 if response is initiative good (column 1) or = 1 if response ‘voting public’ (columns 2–7). Coded 0 if otherwise. See appendix to chapter.

* = significant at p < 0.05 ** = significant at p < 0.01 (Wald test).

N X2 imp -2ll % correct

constant

Christian

0.10 0.14 0.03 0.26 0.02 0.16 1.9** 0.34 940 44.9** 69.5

Do right thing?

Initiative good? –0.33 0.25 –0.59 0.54 –0.11 0.29 –4.46** 0.64 1013 32.8** 93

2

1

(continued)

Minor party

South

Table 6.3

140


issues (column 7). For these estimations, a score of 1 indicates a response of ‘the voting public’, 0 if otherwise (other responses include ‘elected representatives,’ ‘both,’ ‘neither,’ and ‘no opinion’). The appendix gives full coding, question wording, and marginals for each of these variables. In column 1, the estimates of overall assessment of the initiative process make it clear that Spanish-speaking voters greatly dislike the initiative process. Older voters also dislike the process, while Republicans like it. Looking across the columns, one of the most systematic patterns we see is that Spanish-speaking Californians think the initiative process is not an especially appealing means of conducting government business. Gays think that the voting public is more influenced by ‘special’ interest groups than elected representatives, and gays are joined by Spanish speakers in being more likely than others to think that representatives (rather than voters) should make decisions about government programs. Spanish speakers are also less likely than others to think that the voting public can be trusted to ‘do the right thing’ on important issues, and are less likely to think that voters can give issues a thorough review. Minor party voters tend to be more supportive of the process, in large part, we assume, because they tend to be excluded from representation in the legislature. The more interesting findings in this table are not so much that minorities consistently resent the initiative process but that majorities do not persistently favor it. Members of the putative majority groups (whites and Christians) do not especially like the initiative process in any consistent manner. Christians are more likely to claim that voters rather than legislators can give proposals a thorough review, but are not otherwise distinguished in any other estimation here. Whites are only distinguished from other respondents in that when other factors are held constant, they are less sympathetic to voters in three of the seven estimations. Thus it appears that members of the majority do not seem to favor the initiative on the basis of a perception that it might be a major vehicle for imposing their own views. For example, members of the majority may believe that the initiative process is more likely to do the ‘right thing’ (column 2 of Table 6.3) since that could easily be perceived by them to mean what is right according to the majority’s point of view. Yet this is not the case. In fact, across the different responses there seems no ‘majority’ (white or Christian) point of view in support of the process. This might reflect that fact that the concept of ‘opinion majority’ is rather elusive. When we consider the large range of issues put to a vote in places like California, it is likely that no single voter supported all of the measures approved in a single election, for reasons related to the Condorcet Paradox (Brams, Kilgour,

142


and Zwicker, 1997, 1998). Thus, the most consistent pattern across all these responses is simply the differentiation of opinion according to education level. More highly educated voters consistently believe that the initiative process is worse than representative democracy across all categories. This in itself is interesting given the correlation between education, tolerance, and support for civil liberties (Nie et al., 1996). In short, the evidence of Table 6.3 shows that some minorities do dislike the initiative process, but others (gays, women, minor party loyalists) are not consistently opposed. We also see very little evidence that a consistent majority favors the process. Our attempt at identifying minorities who perceive negative effects of the initiative is complicated by the fact that some minorities such as gays and lesbians who might be targeted by initiatives in one place or time (or at one level of government) are nevertheless able to take advantage of initiatives elsewhere (Donovan et al., 1999). Respondents from some minority categories might thus perceive the initiative to be advantageous, muting any consistent effects for a group.

Discussion Of all the different variants of direct democracy the initiative process is the most likely to allow for the abuse of minorities by popular majority rule. Other variants of the process involve, and often require, the formal participation of elites in setting an agenda that is abusive of minority rights. In fact, under other systems elites may not need the rubber stamp of a referendum to repress minorities and may actively pursue repression through other means. Under the initiative process, however, popular majorities can set an anti-minority agenda regardless of elite behavior – and this is what causes concern. Even here, however, elite behavior nevertheless retains a major impact; we have argued that elite intervention is key in determining how the actual practice of direct democracy affects minority rights. Thus, in discussions of the impact of direct democracy upon rights, the tolerance level of elites as well as masses is an important concern. How the legislature would have behaved on the question at hand – absent the initiative – is often missing from discussions of the impact of direct democracy on civil rights. This does not mean that intolerance at the level of mass opinion is of no concern since, in principle, mass intolerance could be given voice through the initiative process even in the face of tolerant elites. In looking at the record of outcomes produced by the initiative process, however, we conclude that – although we can indeed find examples of abuse – direct democracy is not uniformly or even frequently tyrannical of


minorities. In the United States, at least, on occasions when direct democracy has produced policies that deny minorities their civil rights and liberties, the courts have intervened. Indeed, a study of the court’s role in invalidating US (state) initiatives found that 81 percent of antiminority measures approved by voters were challenged in court, with 63 percent of these invalidated in part or in whole by state and federal courts (Miller, 1999, p. 21). Two further points should be made. First, with the data we have access to, it is hard to find evidence of ‘the majority’ as a cohesive or self-aware group. The patterns in Table 6.3, for example, are not consistent with the hypothesis that a group of voters conceive of themselves as a majority and are aware of the power of majorities under direct democracy. Majorities that occasionally form in support of an issue (or set of issues) might be such elusive, transitory things, that even under the direct initiative, they have a hard time constituting a consistent force of abuse. Different issues on the same ballot will be passed by different coalitions of voters, and everyone, at some time, may be part of a minority on individual issues. Second, we can see that some demographic minorities (specifically gays and Latinos) are hostile to the process. The fact that these groups have been recently targeted (while other minorities may have escaped such targeting) may be an important part of the explanation for why some minorities are more opposed to the process. We should note, though, that some opinion minorities may see the initiative as holding out the chance to win the war of ideas. Minority party supporters (an opinion as opposed to demographic minority) may well make proposals through the initiative process to further their interests. There is, after all, a tension within the initiative process. Direct democracy is the weapon of ‘outsiders,’ but those outsiders are often likely to form a minority within the participating electorate and so find initiatives of limited use. But many outside groups may also form a minority in the legislature, and see the initiative as a means of influence. In principle, outsiders should be able to use the process to limit or discipline the excesses of elected politicians. Our own surveys of partisan elites show this to be the case. Minor party legislative candidates in western US states are far more sympathetic towards the initiative than major party candidates and incumbents (Bowler et al., 1999). Underlying these points is a definitional one: what constitutes a minority? We have distinguished here between opinion minorities and demographic minorities. To be sure, we can think of different demographic traits as identifying minorities. And these are the ones analysts

144


usually have in mind when discussing the potential for abuse under direct democracy. But once we begin to add up the different sorts of groups we may arrive at a fairly large number of demographic and opinion minorities within the population. This is especially likely to be the case where the political system includes large numbers of voters. In the end, then, we return to an older tradition in political science which speaks of the underlying cleavage structure in society that allows us to make some more specific claims. Where there exist a number of different, and possibly cross-cutting cleavages, direct democracy seems unlikely to pose a consistent threat to minorities. Where such cleavages reinforce rather than cross-cut, however, we are likely to see greater prospects for repression. Empirically such cases seem to appear more likely at the local level where residential segregation coupled with the partition of an area into discrete districts can produce populations with a relatively high degree of homogeneity. Repression may also be more likely where permanent cleavage lines are defined in terms of demographic groups, rather than majorities that form around individual issues. Opinion minorities not only shift from issue to issue, but losers stand to fight again someday over other issues. Demographic minorities (if initiatives are tyrannical and withstand court scrutiny) stand to lose their civil rights (or access to education, language, or essential services) that may affect their very ability to participate in the political process. More generally, institutions of direct democracy are embedded within a wider institutional and political context. They represent just one part of the democratic process which affects minority rights. The actual working of that process depends, in part, on the underlying distribution of preferences of citizens and elites. We have found that when elites are relatively united against a proposal to limit or undermine civil rights, it is very difficult for such a measure to pass. With these broader constraints in mind we draw the conclusion that direct democracy, even in the form of the binding citizen’s initiative, may not be as injurious to minority rights as may be feared.

Note 1 Data taken from our own poll of Washington voters (conducted by Applied Research Northwest) and from Field/California Polls. We focus on California Poll data, in part, because it provides variation in minority categories. Only 5 percent of Washington respondents classified themselves as an identifiable race/ethnicity other than white.


Appendix: Coding of variables for Table 6.3 Question wording: Column 1: Voting on propositions allows voters to make decisions on state government laws and policies. Another way to make state laws and policies is through our elected representatives in the state legislature. Dependent variables for columns 2–7 take the value of 1 if the answer to the following questions is 2, otherwise the variable is coded 0. Column 2: Which do you feel can be trusted more often to do what is right on important government issues – elected representatives in the legislature or the voting public through proposition measures? No. of cases

Value

198 655 50 45 75

1 2 3 4 8

1023

Value label Elected representatives Voting public Both Neither No opinion

Total

Column 3: Which do you feel is more influenced by special interest groups – elected representatives in the legislature or the voting public through proposition measures? No. of cases

Value

580 292 80 6 65

1 2 3 4 8

1023


Total

Column 4: Which do you feel is better suited to decide upon large scale government programs and projects – elected representatives in the legislature or the voting public through proposition measures? No. of cases

Value

312 562 63 19 67

1 2 3 4 8

1023

Total


146


Column 5: Which do you feel generally enacts more coherent and well thoughtout government policies – elected representatives in the legislature or the voting public through proposition measures? No. of cases

Value

384 451 63 44 81

1 2 3 4 8

1023


Total

Column 6: Which do you feel is better suited to decide upon highly technical or legal policy matters – elected representatives in the legislature or the voting public through proposition measures? No. of cases

Value

502 375 56 26 64

1 2 3 4 8

1023


Total

Column 7: Which do you feel gives a more thorough review to each particular aspect of a proposed law – elected representatives in the legislature or the voting public through proposition measures? No. of cases

Value

466 395 54 23 85

1 2 3 4 8

1023

Total


7 The Medium is the Message: How Referendums Lead us to Understand Equality Avigail Eisenberg

In 1998, the provincial government of British Columbia completed negotiations and signed a treaty with the aboriginal Nisga’a people that established limited self-government for the Nisga’a and recognized their claim to a portion of their traditional lands.1 The treaty generated an enormous amount of attention and controversy in the province. Some non-Nisga’a Aboriginal peoples opposed it because, they argued, it accepted the Nisga’a claim to land claimed by other Aboriginal groups and because it established an unacceptable template for future treaty negotiations. Some non-Aboriginal people opposed the treaty because it created ‘special rights’ for Nisga’a people, and because it limited the rights of non-Nisga’a people living on Nisga’a lands. Some of these opponents called for a referendum on the Nisga’a agreement. This call drew enthusiastic support mostly from those in the province who were already eager to hold referendums whenever possible. 2 One reason not to hold a referendum on the treaty is that referendums are viewed as favoring majorities not minorities. This view was expressed by the Premier of British Columbia at the time, Glen Clark, who argued that ‘[i]t is dangerous and improper to subject minority rights to the will of majority’ (Clark, 1998). Yet, interestingly, in the scholarship about referendums a stark disagreement exists about what effect referendums have on minority interests and rights. Some studies confirm that, indeed, referendums are poor ways to protect minority interests (Butler and Ranney, 1994). And in a survey study of the impact of referendums on minority civil rights in the United States, Barbara Gamble concludes that ‘the record shows that American voters readily repeal existing civil rights protections and enthusiastically enact laws 147

148

Avigail Eisenberg

that bar their elected representatives from passing new ones’ (Gamble, 1997, p. 262). Yet, other scholars are just as emphatic in arguing exactly the opposite, namely that, based on the record of successes and failures, initiatives are not employed to deprive minorities of their basic rights any more than are ordinary legislative processes (Zimmerman, 1986, p. 95). ‘[D]irect democracy devices can only rarely be faulted for impairing the rights of the powerless,’ especially when the record of referendums is compared to the protections afforded by representative government (Cronin, 1989, p. 92). More generally, other studies suggest that the frequent concern that the majority acts in tyrannous ways towards minorities is overblown (Magleby, 1984, p. 196; Boyer, 1992, pp. 96–112), and, at worst, the record of referendums is mixed when it comes to protecting minority rights (Donovan and Bowler, 1998; Bowler and Donovan, this volume). This disagreement among scholars about the impact that referendums have on minority interests can be explained in a variety of ways, but a distinction may be drawn between two general types of explanatory strategies which scholars employ: (1) those explanations that focus on general elements of political, cultural, or socioeconomic context that are external to the referendum being analyzed; and (2) analyses that focus on factors internal to the specific referendum being analyzed. External factors are factors such as the nature of the community, the historical context, or the economic conditions in which the referendum takes place. The impact of referendums on minority rights will differ when these external factors differ. For example, some communities might be better than others at directing voters to apply ‘other-regarding behaviour’ rather than to base their vote on narrow self-interest (Gillette, 1988, p. 954). Studies have suggested that economic hardship leads voters to be more grudging in regard to the interests of others (Donovan and Bowler, 1997, p. 119). Moreover, the size of the community is also thought to influence referendums because larger political communities ‘offer the heterogeneity of interests that make it difficult for a cohesive majority to tyrannize a minority’ (Donovan and Bowler, 1998b, p. 1021). With respect to factors internal to referendums, a growing body of literature suggests that referendum outcomes are dependent on factors particular to each referendum campaign. Three internal factors often cited are, first, the attitudes that elites take to the specific issue at stake in the referendum (Butler and Ranney, 1994, pp. 5–6; Donovan and Bowler, 1997, pp. 113–17). Voters tend to vote the way that people they trust vote, in particular, the way that elites they trust vote (Jenssen and

The Medium is the Message 149

Listhaug, this volume). Second, voters are influenced by cues used during the campaign to activate racial, religious, and other sentiments. And third, voters are influenced by the wording of the question, for example whether the question is confusing or not, whether it is openended or offers the possibility for revisiting the issue in the future (Chambers, this volume; Derriennic, 1995). The influences of internal and external factors are not entirely independent of each other. For example, the success of internal cues used during the campaign is likely to depend on external factors such as the nature and history of the community in which it is used. Nonetheless, by emphasizing different factors, each form of explanation provides a different avenue for analyzing voting outcomes and each approach can be helpful in explaining why a majority of voters might or might not endorse minority interests through a referendum. Here I put aside the debate about which of these factors are decisive, and focus on a factor which is neither internal nor external but which I argue influences the outcome of referendums. This factor is the referendum itself. A referendum is a decision-making mechanism that is based on a particular understanding of political equality, namely one in which all individuals enjoy precisely the same amount of political power. Referendums invoke this conception of equality for voters. That is, in employing a referendum as a means to resolve a pressing issue, governments or elites legitimize the referendum as the appropriate way to decide the issue and in so doing legitimize the particular understanding of political equality that underpins the democratic idea of referendums. This understanding requires that each voter be given the same amount of political power in making the decision. One of the main appeals of referendums is that they attempt to eliminate distinctions that might otherwise divide or differentiate members of a community. In this way, referendums suggest to voters that each voter is given the same power in making the decision at hand because each has a similar stake in the decision. Whether or not this appeal to what is here called undifferentiated equality reflects the best view of political equality is the subject of considerable debate in western societies today. This debate will not be reviewed here. Rather, the aim here is, first, to point out that the conception of political equality to which referendums appeal does not reflect the only view of political equality currently endorsed in liberal democracies, and second, that often the controversy surrounding minority rights claims is based on contending views of what political equality requires. In particular, claims put to a referendum that involve

150

Avigail Eisenberg

or appear to involve special treatment for minorities conflict with the view of political equality conveyed by referendums and by most mechanisms of direct democracy. In this respect, the very invocation of a referendum as a legitimate decision-making mechanism may influence the treatment of minority claims. Unsurprisingly, claims for minority rights differ along several dimensions. Often they differ in terms of what is needed to protect minority interests. The focus here is on but one distinction in the realm of minority rights claims, namely the distinction between claims for similar treatment and claims for different treatment. The first part of this paper distinguishes between these claims in terms of the conception of political equality each type of claim requires. Claims for similar treatment draw on an undifferentiated notion of equality while claims for different treatment require a differentiated conception of equality. The second part explains why referendums contain a logic that favors claims based on undifferentiated equality. Those who advocate referendums typically find them attractive not merely because they express the majority’s values but rather because referendums express minoritarian and communitarian sensibilities. These sensibilities, as will be shown below, are both based on an understanding of political equality that equates equality with similar treatment of and status among individuals. The third section examines the evidence that referendums tend to favor claims for undifferentiated equality by looking at the manner in which cues about equality are used in conjunction with the internal factors thought to influence the outcomes of campaigns, namely the position of elites, the background cues, and the wording of the question. The final section explores the limitations of using referendums in light of their bias towards undifferentiated equality.

Minority rights and claims to special status Today, two distinct understandings of equality are at the centre of most debates about minority rights in liberal democracies. First, advocates of undifferentiated equality argue that the best way for each citizen to enjoy similar benefits, opportunities and protections is for each individual to be treated approximately the same way by public institutions in the sense that each is granted similar political, social, and civil rights (for example, Nozick, 1974). Undifferentiated equality requires, foremost, that individuals be protected against discrimination in public and private institutions and, more specifically, that they not be denied opportunities or benefits because of their religion, race, ethnicity,


gender, or sexual orientation. In terms of citizenship, undifferentiated equality requires that the liberal state remain neutral between the cultural and religious agendas of different groups (see Glazer, 1978; Morton, 1985; Walzer, 1982). Once the state involves itself in protecting the language or culture of one group, it leaves itself open to minoritarian demands which cannot and should not be addressed by using state resources (see Glazer, 1983). Therefore, those who understand equality as requiring undifferentiated treatment argue that the business of the state does not extend to helping individuals to negotiate their membership in cultural or religious associations. To the contrary, the state’s main concern is to protect the rights of citizenship that transcend these other associational ties and unite all individuals into a single statebased community. Beyond this, undifferentiated equality requires that the state allow individuals to associate with whomever they choose and that, at most, it ensure that individuals are at liberty to exit the groups to which they belong (Kukanthas, 1992). Differentiated equality also requires that individuals not be denied opportunities or benefits because of their religion, race, ethnicity, gender, or sexual orientation. But its advocates argue that these protections against discrimination must sometimes be augmented by group rights or special status for groups that suffer unjust disadvantages which cannot otherwise be rectified (see Kymlicka, 1995; Tully, 1995; Young, 1990). The group rights claimed by national minorities and indigenous peoples offer ‘special status’ to these groups in the sense that the benefits they confer – for example, the right to use a particular piece of land, or the right to be educated in one’s own language – are available to group members only. In either case, these rights are different from the ones enjoyed in the broader community. Like individual rights to be protected from discrimination, rights to distinct status are defended on the basis that they secure equality. Sometimes the aim of differentiated equality is to compensate groups for disadvantages they suffer due to the systemic biases within institutions and practices. Sometimes the aim is actively to protect certain cultural minorities whose historical presence and whose cultural fragility justify special status. In other words, very different types of claims are defended using a differentiated notion of equality. Language rights, the right to group representation, affirmative action and regional autonomy are just a few possible examples. These are, admittedly, very different types of claims and they require different types of justification. 3 The objective here is not to provide such justification, but to focus on one element that they share, namely an understanding of equality that

152

Avigail Eisenberg

requires individuals and groups to be treated differently in order for equality to be enjoyed. In sum, two different understandings of equality can inform claims by different types of minority groups for protection: undifferentiated and differentiated equality. While these understandings are not necessarily antagonistic, often the politics surrounding disputes about minority claims to special status involve one side arguing that equality demands that all individuals enjoy the same rights and that the state ought to remain neutral towards all, and the other side arguing that equality demands different treatment in light of the disadvantage that a group suffers. Disputes over minority rights claims that involve these contending notions of equality are often the subject of referendum campaigns. But as the following two sections argue, referendums are themselves often justified on the basis of an undifferentiated notion of political equality. The question then arises whether referendums create a political atmosphere for thinking about minority rights claims that is biased against claims based on differentiated equality.

Referendums and equality: majoritarianism vs minoritarianism One might think that the only political atmosphere created by a referendum is one in which the majority rules and that this atmosphere disadvantages minority claims of all sorts. This position appears to be adopted by Gamble who surveys 74 US civil rights referendums on racial, ethnic, and language minorities, gay men and lesbians and people with AIDS and finds that, in 78 percent of the cases, direct democracy leads to ‘majority tyranny’ (Gamble, 1997, p. 253). One obvious question that arises from her study is why do majorities sometimes deny minorities rights? One can assume that the campaigns examined by Gamble differed in terms of many of the external and internal factors that are thought to influence referendums. But, if these factors are not always decisive in influencing the majority, what explains the majority’s tendency to deny minority rights in some cases and not in others? Or alternatively, what factors might induce majorities to support minority rights claims in 22 percent of the cases Gamble examined? One answer to this question appeals to what can be called a ‘minoritarian sensibility.’ A majority might support a minority claim if a majority of those outside the minority believe that, by endorsing the minority’s claim, they thereby advance their own interests. Relevant


examples include minority rights to privacy or to protection against employment discrimination. These rights are ones that individuals in the majority or in other minorities are likely to view as crucial to their own interests as well. Those outside the minority may have an interest in endorsing such rights in order to ensure that the rights remain vital and protected. This sort of motivation is ‘minoritarian’ because it draws on the idea, more fully developed in pluralist theories of democracy, that it is often in the majority’s interest to protect minority interests because the majority is, in many senses, simply a coalition of minorities. In pluralist societies, everyone is a member of a minority group and often belongs to more than one sort of minority group. Majorities are, in some senses, merely shifting coalitions of minorities that are formed to advance a particular set of interests. Based on this view, pluralist democracy is ‘government by minorities’ (Dahl, 1956 and 1961). Majorities, in pluralist societies, are not sufficiently stable to rule for very long. Nor are they stable enough to tyrannize minorities. Once they have accomplished their limited purpose, they break apart and reform through different coalitions for different purposes. What unites these groups, at least according to Dahl’s formulation, is a ‘democratic creed’ that includes, among other values, respect for and tolerance of minorities. Each minority adopts a minoritarian sensibility, partly as a means of ensuring that it will be tolerated in turn. In referendum politics, something like a minoritarian sensibility seems to work to bridge the gap between how voters feel about a minority group and what sort of rights they are willing to extend to that group. Voters may be inclined to endorse minority rights claims if they view these claims as endorsing values important to the majority as well. For instance, in their assessment of anti-gay and lesbian ballot initiatives, Donovan and Bowler explain that, despite the fact that ‘[a] clear majority of Americans disapprove of homosexuality,’ the mass public is firmly committed to protecting the civil rights of homosexuals (Donovan and Bowler, 1997). In terms of the minoritarian hypothesis, the majority is likely to be committed to protecting the rights of homosexuals because it embraces tolerance towards minorities. By extending tolerance towards others, we protect and even enhance the prospects that our own distinctiveness will be tolerated. Moreover, by protecting the rights of even unpopular minorities to freedom of religion and speech, the right to vote, and access to education and employment without being subjected to discrimination, we extend to others the same rights and privileges we want to enjoy ourselves.

154

Avigail Eisenberg

The power of this minoritarian sensibility is best illustrated by seeing what happens when it is absent. Pluralist theory was not designed to accommodate all types of minorities and, therefore, not all types of minority claims may benefit from the minoritarian sensibility. In particular, claims made for differentiated equality are unlikely to benefit because the reciprocity that otherwise encourages those outside the minority to protect those inside the minority is no longer present. For example, the 1992 Canadian referendum on the Charlottetown Accord asked Canadians whether they approved of a constitutional package that included a provision to recognize Quebec as a ‘distinct society.’ The ‘distinct society’ clause was crucial to the package in two ways. First, it was a way to reconcile the strained relations between Quebec and the rest of Canada. Second, opposition to this clause was partially responsible for the Accord’s rejection in the referendum. The problem was that the clause was very difficult to sell in the rest of Canada (see Johnston et al., 1996), in part because it claimed for Quebec a form of recognition that was not open to other groups. Canadians outside Quebec were being asked to recognize a status for Quebec that they could not claim for themselves. Therefore, in order to endorse the provision, voters had to either approve of the clause for itself, or care more about securing the whole package of amendments than about defeating one part of it ( Johnston et al., 1996, p. 95). But the option of relying on their minoritarian sensibilities – that is, extending to others the rights and privileges that they claim for themselves – was decisively closed because the Accord seemed to claim for one group a distinction that was not open to others. As this example makes clear, minoritarianism can only work for claims that reflect or are made to reflect undifferentiated equality.4 The minoritarian sensibility offers one answer to the question posed at the beginning of this section, namely why majorities might endorse minority rights. A majority (or a coalition of minorities) supports the claim of one particular minority as a means of advancing interests that the majority considers important for itself. Referendums are more likely to be ‘minority-friendly’ when minorities advance claims that rest on undifferentiated equality. This is not to say that the majority will always support claims that appeal to minoritarianism or always defeat claims that require special treatment for one group. Instead, minoritarianism is best thought of as a resource whose availability partly depends on external factors, such as the self-conscious pluralism of a society, and partly on the issue involved, that is, whether the issue rests on differentiated or undifferentiated equality. When issues require that we


extend to others rights that we ourselves enjoy or want to enjoy, then referendum campaigns can draw on the minoritarian sensibility to motivate majorities to protect minority rights.

Referendums and equality: majoritarianism vs communitarianism In addition to minoritarianism, a second sensibility that referendums invoke is a communitarian sensibility. Communitarianism focuses on the values inherent in community life, including sharing values, beliefs, identity, and membership (for example, see MacIntyre, 1984; Sandel, 1984). Communitarianism refers to community values that cannot be calculated by merely aggregating the private interests of a majority. Nor are communitarian values reduced to specific interests that people in the same neighbourhood or state happen to share. Rather, communitarian values refer to the value that individuals attach to their common membership in a community and to projects and interests they share as members of a community, whatever those projects and interests might be. The emphasis of communitarianism is on what is shared, what unites, and what members have in common. Democratic theorists have repeatedly cited the value of community as one of the primary justifications for direct democracy. For instance, Rousseau described the general will as not merely the will of the majority: ‘[W]hat makes the will general is less the number of voters than the common interest uniting them’ (Rousseau, Book 1, chapter 4, emphasis added). Other theorists make a similar point: direct democracy does not seek to empower the formal will of a mathematical majority. Rather its aim is to capture the ‘spirit which does justice to the whole of the community’ (Barker, 1942, pp. 67–8). Benjamin Barber is emphatic on this point: ‘[m]ajoritarianism,’ he argues, ‘is a tribute to the failure of democracy; to our inability to create a politics of mutualism that can overcome private interest’ (Barber, 1984, p. 198, emphasis added). Some analysts deny that referendums give voice to any particular values other than those to be calculated through the simple aggregation of private, individual preferences. Those who advance this position are primarily concerned that referendums curtail the deliberative aspect of political decision making by sanctioning a procedure in which voters may decide on an issue based on their private preferences without engaging in public deliberation (Elster, 1986). While the concern that votes in a referendum may not be based on a sufficient amount of deliberation is not disputed here, the position that votes are merely cast on

156

Avigail Eisenberg

the basis of private preferences which are formed in isolation from the community largely misses the appeal of many referendums. In other words, the view that voters come to a decision based on purely private values fails to comprehend why referendums end up being attractive in some societies and circumstances and not in others. One of the key appeals of referendums is their leveling effect. That is to say that referendums erase differences in status and advance a notion of political equality in which each participant has precisely the same political power. This leveling effect is conveyed in at least three ways. First, referendums have often been proposed as a way of undermining the power of political elites who are the real target of calls to ‘give power to the people.’ The distrust of elites is a central explanation for the appeal of referendums today and, historically, for the appeal of direct democracy in general (see Craig et al., this volume). Second, referendums are sometimes regulated by strict spending limits to prevent the debate from being dominated by the most affluent members of society. 5 When referendums are not regulated in this way, the concern is that democracy has been thwarted by wealth and that the political equality which made referendums so attractive in the first place has been usurped. Third, the understanding of political equality that fuels the reaction against elites and against wealthy interests can give rise to a distrust of all that divides and differentiates the people. This distrust is usually conveyed today in terms of a disdain for ‘special interests.’ The call for referendums is often accompanied by the sense that divisions among the people are not genuine conflicts of interests at all but, rather, are manufactured by elites or special interests. When these three means of leveling distinctions are present, the main appeal of referendums as a tool of democracy becomes their ability to coalesce communitarian values by rendering all distinctions in the community irrelevant.6 They are a means to ‘cut through the obfuscation’ of differences and to operationalize an undifferentiated notion of political equality that treats all citizens the same by according to each the same political power. The 1992 Canadian referendum provides a particularly vivid illustration of this process. According to Johnston et al.’s analysis, the politics surrounding the referendum ‘cut through coalition-of-minorities obfuscation’ (Johnston et al., 1996, p. 5). It advanced an ideal of voters applying ‘common sense without the encumbrances of special interests’ and especially without the interference of elites who the voters believed were ‘susceptible to lobbying by special interests who pay the costs of electioneering with surplus extracted . . . from ordinary voters’ ( Johnston et al., 1996, p. 12). So, in


part, the target at which voters aimed in casting their vote against this complicated package of constitutional amendments was ‘elite arrogance’ and the ‘very logic of coalition-building.’ As Johnston et al. put it, ‘however complicated elites saw the situation to be, voters made the question simple for themselves’ (Johnston et al., 1996, p. 11) Given that referendums are often driven by anti-elite sentiments, and more generally that referendums heighten the communitarian sensibilities of a population by diminishing the significance of distinctions among members of the community, it is somewhat ironic that elites are so interested in using them, particularly to coalesce nationalist sentiments. In Yugoslavia, for instance, the parade of nearly 20 referendums between 1990 and 1993 became one of the key means by which voters gave voice to their distinct community membership (Brady and Kaplan, 1994, p. 207). To facilitate the expression of these communitarian sentiments, the different sets of elites ensured that each referendum was either held in a distinct area of the country – such as Slovenia or Croatia – where a particular ethnic community was known to hold a majority, or was open only to members of one ethnic group, such as Serbs in Croatia or Albanians in Macedonia. Indeed, only one attempt was made to hold a Yugoslavia-wide referendum, and it failed because elites could not agree on the question (Brady and Kaplan, 1994, p. 210). In Quebec, referendums have also been used by elites as a means to encourage the expression and development of a nationalist community. The first sovereigntist provincial leader, René Lévesque, solidified support for his party by campaigning on the promise that, if elected, he would hold a referendum on Quebec’s secession from Canada – which he held in 1980.7 What has been crucial about the 1980 referendum and the subsequent referendum held on secession in 1995 is that neither is accepted by elites in Quebec as decisively a vote against nationalist aspirations. 8 Despite a majority (though a slight majority)9 of people in Quebec voting against the sovereigntist option in both referendums, the prospect of holding another referendum continues to be attractive to some nationalist leaders. The attractiveness is due to the fact that referendums can coalesce communities and draw out communitarian values. As much as the referendums have alienated the non-Francophone and Aboriginal minorities in the province, they offer Quebec nationalists yet another opportunity to focus the attention of Francophones on what they share – namely their membership in a distinct national community that is centered around a shared language. By focusing the attention of voters on what members of a community share, referendums are a means of transcending differences and thereby

158

Avigail Eisenberg

emphasizing the common status of all members. Direct democracy helps to express the common interest of a community by diminishing the significance of diversity in the community whose will it seeks to express. This does not mean that referendums are good means to heal communities damaged by divisions based on religion, language, or ethnicity. To the contrary, referendums might exacerbate divisions based on such differences. Whether or not the communitarian sensibilities activated by referendums successfully transcend differences largely depends on whether the minority interests at stake require that similar or distinct treatment be extended to minority groups. Minorities seeking similar status and treatment might find a referendum campaign to be a good opportunity to find acceptance as equal members in the community on similar terms of membership enjoyed by other members of the community. However, minorities seeking distinct status are far more likely to find referendums an alienating event and to insist – as Aboriginal people did in Quebec, and as did various ethnic groups in Yugoslavia – that separate referendums be held in their community or that the vote of their community be counted and reported separately. Advocates of direct democracy often point out that referendums are a means of operationalizing an ideal of equality that, in giving exactly the same weight to each individual’s vote, requires that the differences among people be ignored in the course of performing this ultimate act of citizenship – namely voting in a referendum. In other words, just as participating in a referendum is a testament to one’s membership as an equal in a community, giving each person’s vote equal weight in constitutive decision-making processes is a testament to a particular understanding of political equality, namely an undifferentiated one. In the absence of countervailing attempts to structure referendums so that they bring out community differences (for example, requiring concurrent majorities from different constituencies within the larger community), referendums create an atmosphere in which sensibilities such as minoritarianism and communitarianism resonate. These sensibilities are consistent with an undifferentiated understanding of equality and are hostile to or inconsistent with differentiated equality. In many cases, the tension between these two understandings of equality is itself the subject of the controversy that the referendum is being used to resolve. So far, the argument has shown that one of the crucial determinants of whether a minority rights claim is likely to be endorsed in a referendum is whether the claim is based on undifferentiated equality and, therefore, requires similar treatment for the minority group, or whether it is based on differentiated equality and requires distinct status for the


group. Majorities often have access to a minoritarian sensibility which they draw on in deliberating about whether or not to endorse minority rights. The minoritarian sensibility is only useful in cases where minorities are advancing a claim for similar treatment. Where minorities are claiming distinct status, they are asking voters to endorse rights that a majority of the voters do not (and cannot) enjoy themselves. This request challenges the minoritarian ideal that we are all members of minorities with similar status. Second, referendums also give rise to a communitarian sensibility that motivates voters to focus on their shared membership and to diminish the significance of differences among those who are considered members of the community. This resource is, again, useful only to groups that seek undifferentiated equality. Claims for distinct status, a distinct voice, or distinct treatment conflict with the communitarian sensibility because these claims signal either that membership is not shared or that members do not share similar status. The presence of either or both of these sensibilities is no guarantee that minority claims for undifferentiated equality will be endorsed in referendums, nor that claims for differentiated equality will automatically be defeated. Rather, the claim is that these sensibilities, which are activated by the very logic of using a referendum to make political decisions with respect to minority rights, are resources upon which campaigns may draw. Moreover, these sensibilities may act as obstacles to a campaign’s success and, indeed, adversaries will often use these sensibilities to influence voters. One cannot suppose therefore that referendums are neutral means of deciding issues involving political equality. And they are likely to be particularly biased within societies in which debates about the sort of equality that ought to be guaranteed to minorities are ongoing.

The nature of equality as a factor in influencing the outcome of referendums As stated above, the outcomes of referendums are thought by scholars of direct democracy to be influenced by a variety of external and internal factors. The implication of these analyses is that the referendum itself is a neutral medium in which political decisions are made. Thus far, the argument advanced here points to the minoritarian and communitarian sensibilities invoked by referendums to argue that referendums are not a neutral medium for decision making. In addition to the appeal often made to these sensibilities by those defending

160

Avigail Eisenberg

referendums in theory and in practice, another way to detect the biases of referendums is to notice how the effects of the internal factors are often mediated by the relation of these factors to a particular understanding of political equality. The three internal factors cited above are the influence of elites, the cues used during the campaign to activate racial, religious, and other sentiments, and the wording of the question. These factors are not independent of the influence exerted on them by the referendum itself. Specifically, the relation of each factor to the notion of undifferentiated equality will partly determine how each influences voters during a campaign. In other words, part of what accounts for the way that elites or cues or the wording of the question might influence voters is the message that each factor conveys about the nature of political equality. First, the influence of elites will partly depend on whether they are viewed as allied with special interests, that is interests that enjoy special rights or protected status in society. Elites who are viewed as allied to causes for differentiated equality might have far less influence in a campaign than those who are viewed as fighting for the equal rights of all individuals. When elites strongly endorse claims for differentiated equality, they might, in effect, compound voter suspicion about these sorts of issues. In other words, the extent to which elites may help voters decide to protect minority rights will largely depend on whether the minority rights in question seek to establish differentiated or undifferentiated equality. Elites that can tap into the voters’ minoritarian sensibilities might be helpful in cases about extending to minorities similar treatment and might be fatal to causes involving distinct treatment. Perhaps the worse case scenario is one in which all elites endorse a minority claim for special status. This scenario might trigger, more strongly than any other, the communitarian sensibility that distrusts elite power at the same time as it seeks to diminish the significance of special interests in the community.10 Second, the cues used during the campaign can also be related to differentiated equality in so far as these cues offer contending views of community and therefore contending views of which shared values ought to be viewed as most important. Several examples have already been discussed above in which nationalist sentiments are activated by the campaign. Those who activate these sentiments are doing so in order to convince voters to vote primarily with the values of the nationalist community at heart. Those who oppose nationalist sentiments will employ cues that activate sentiments in favor of supranational or subnational communities. The reason why, for instance, communitarian


sensibilities failed to tie all residents of Quebec or of Canada together in the 1992 pan-Canadian referendum is that the more powerful cues during the campaign asked that voters situate themselves within separate nationalistic or provincial communities. By tapping into communitarian sensibilities about membership, these cues helped voters understand which rules of membership ought to be applied in deciding the issue. Religious cues are particularly powerful in this respect and have had a significant impact in the United States in referendums on gay and lesbian rights and on abortion. Like nationalist cues, they remind voters of the community, in these cases, the religious community, to which some voters belong. In doing so, they remind voters of which communitarian sensibilities ought to be referred to in assessing the issue at hand. By applying these communitarian sensibilities, voters work out who is a member and who is not and thus which shared project might be at stake in the decision at hand. For instance, in the case of gay rights, opponents may invoke religious attitudes, particularly fundamentalist ones, to cue voters to think about who belongs to God’s community, who deserves similar treatment in that community, and who does not. This appeal might not always be successful, but its purpose, to consolidate a particular notion of community, is, nonetheless, clear. Finally, the wording of the question is a particularly powerful means by which the debate between differentiated and undifferentiated equality may influence voters during referendum campaigns. Partisans will often attempt to word the question in a way that directly frames the issue in terms of differentiated or undifferentiated equality. For example, in the Soviet Union’s referendums in 1991, nationalists and unionists each tried to co-opt the voter’s sentiments by adding questions to the ballot that either reminded voters of their nationalistic community or offered them a way of reconciling their nationalistic pride with an undifferentiated equality for members outside their nationalist group. Kirghizia and Uzbekistan, which were both dominated by the conservative Communist party elites who supported the continuation of the Soviet Union, worded the question to emphasize the values of undifferentiated equality by asking voters whether they approve of ‘the preservation of the USSR as a renewed federation of equal sovereign republics.’ Kazakhstan’s question also appealed to undifferentiated equality and did the work of reconciling this notion with the nationalist sentiments that might otherwise induce voters to vote against the Union. In this case, the referendum asked whether voters wanted ‘to maintain the USSR as a union of sovereign states of equal rights’ (Brady and Kaplan, 1994, p. 194). In each case, voters were

162

Avigail Eisenberg

asked to view their distinctiveness as consistent with a notion of undifferentiated equality and, therefore, to endorse the union. Differentiated equality was made to look like, or at least to be consistent with, undifferentiated equality to help secure passage. Similarly, attempts are often made to make claims for undifferentiated equality appear to be ones for differentiated equality in order to help ensure their defeat. One example is the 1992 Colorado initiative that amended the state constitution in a way that barred state legislators from passing laws or regulations that prohibited discrimination on the basis of sexual orientation. Proponents realized that one way to convince voters to support an initiative that withheld undifferentiated equality from some members of the community was to frame the issue as though it was about extending to gays and lesbians special status. As Donovan and Bowler note, proponents conveyed the issue in terms of ‘no special job rights for gays’ (Donovan and Bowler, 1997, p. 120; also see Gamble, 1997, pp. 257–60). Framing the issue in this way was a good strategy to employ for the proponents because it denied opponents easy access to the minoritarian sensibilities that would otherwise convince voters that, despite the unpopularity of the group, similar treatment ought to be extended to its members. Again, the suggestion here is not that voters cannot resist claims that are presented as ones about undifferentiated equality or that they will consistently vote against ones for differentiated equality. While the influence of this factor – that is, what sort of equality is being sought – is not always decisive, it is significant enough to influence our understanding of when minority rights claims are more or less likely to be approved through referendums. Other factors are influential as well. The support of elites, the cues invoked in the course of the campaign, and the wording of the question are three factors likely to be influential. But rather than being discrete factors, their influence may be mediated by their perceived connection to two different understandings of equality. Elites might compound the difficulties in passing referendums that attempt to secure differentiated equality for one group if, in allying themselves with ‘special interests,’ they remind voters of the communitarian sensibility that referendums invoke in the first place. This sensibility distrusts elites at the same time as it suspects any interest that divides the community. The cues relevant in referendums are often those that remind voters of one particular community membership they hold and, therefore, of the particular membership and shared projects to which they are committed. And finally, the wording of the question is crucial because, irrespective of the substance of the issue,


it can be reframed to appear to be about either undifferentiated or differentiated equality.

Conclusion In the end, the key question that might seem to arise from this discussion is: ought referendums be used to decide whether or not to endorse minority claims to special status, such as the Nisga’a Treaty previously mentioned? The argument here has suggested that, all things being equal, claims for differentiated equality (or those that are made to appear as such) will be disadvantaged in the context of a referendum. This is partly because the two types of sensibilities to which referendums appeal – minoritarianism and communitarianism – both reinforce the ideal of undifferentiated equality and thus create a politically biased atmosphere in which debates involving contending understandings of political equality are resolved. The minoritarian sensibility reminds voters that we are all members of a minority and therefore ought to extend to minorities the same rights that we enjoy ourselves. The communitarian sensibility focuses the voter’s attention on the value of shared membership and common ends. One of the primary appeals of using referendums as a means of democratic decision making is to activate this sensibility and thereby to level distinctions that might otherwise divide the community. Again the notion of differentiated equality conflicts with this appeal. Moreover, the atmosphere created by referendums that privileges an understanding of equality as undifferentiated treatment and status for all members of the political community mediates the way in which elites, cues, and the wording of the question influence voters. The connection between elites and ‘special interests’ is often a means of justifying the use of referendums in the first place and thereby signaling the legitimacy of an equal distribution of political power. The cues that often resonate in referendums are those that attempt to remind voters of the type of community they belong to and thus to whom equal status must be extended. And finally, debates about how to word the question, like the cues and context, are often about reminding voters of the nature of community and about the similar status that all members ought to enjoy. It would be remiss to conclude without noting that referendums as instruments of decision making are not entirely independent variables in the politics of minority rights. That is, referendums are often structured in ways that attempt to mediate the effects of undifferentiated

164

Avigail Eisenberg

equality, for example by requiring that the votes of minorities or of geographical regions in which a minority lives are counted or reported separately from the votes of the broader community. If British Columbia, for instance, was compelled to hold a referendum on the Nisga’a Treaty, one way of ensuring that the undifferentiated understanding of political equality relayed by the atmosphere of a province-wide referendum did not go unchallenged would be for the government to insist on and advertise the reasons for counting and reporting the votes of the Nisga’a people separately from those of non-Nisga’a. Yet, the need to structure referendums in this and other compensatory ways only reinforces the argument being made here – that referendums do not provide an unbiased atmosphere for resolving issues that involve contending understandings of political equality. And finally, notwithstanding this bias, other considerations in favor of holding a referendum on a minority’s right to distinct rights or status might override the concerns outlined here. For instance, a community conflict might seem so debilitating and a referendum might appear so decisive a way to resolve it that holding a referendum, even on a matter that clearly involves minority rights to special treatment, might be, on balance, justified. The argument here does not suggest that holding a referendum on a minority claim to special status is entirely unjustified in all cases. However, it does hold that one cannot justify the use of a referendum on the basis that referendums are a politically neutral means of resolving the issue. Just as the nature of the question, the nature of the debate, and the views of political elites might bias the outcome of a referendum, the very use of referendums creates an atmosphere that biases the proceedings against claims for differentiated equality.

Notes 1 Canada, Nisga’a Final Agreement, December 1998. Also see BC Studies, 120 (Winter 1998/99) special edition on the Nisga’a Treaty. 2 The British Columbia Liberal Party launched a court challenge to the Nisga’a Agreement on the grounds that the Agreement constitutes an amendment to the Canadian Constitution and therefore, under British Columbia’s Constitutional Amendment Approval Act, requires that a provincial referendum be held. Constitutional experts have since argued that the Agreement does not constitute a constitutional amendment. See R. Vick Farley, Memorandum on Constitutional Amendment Approval Act and the Nisga’a Final Agreement, British Columbia Ministry of the Attorney-General, 24 July 1998; Peter Hogg, Letter to the Deputy Minister Attorney General, Re: Constitutional Amendment Approval Act and Nisga’a Final Agreement, 29 July 1998. Both legal opinions can be found at www.nisgaa.org.

The Medium is the Message 165 3 Kymlicka argues that there are three types of group rights: (1) self-government rights; (2) polyethnic rights; and (3) special representation rights. See Kymlicka (1996). 4 As discussed in the final section, a strategy often employed in referendum campaigns is to make claims that rest on undifferentiated equality appear as though they require differentiated equality. Therefore, proponents of distinct status or special rights might find it useful to present their claims as though what they are asking for is no different from what the majority already enjoys. Philosophically, this is precisely how cultural rights are defended, in that the right to cultural security is something that the majority in a democracy already enjoys by virtue of its numerical dominance. See Kymlicka (1989). 5 Legislation on spending limits during referendum campaigns varies significantly. In the United States, spending limits were ruled unconstitutional in Buckley v. Valeo 424 US 1 (1976). A recent court case in Quebec upheld strict spending limits in that province for the express purpose of ensuring that the debate was not dominated by wealthy interests. See Libman v. Quebec (Attorney-General) [1997] 3 SCR 569. 6 In addition, referendums in which each vote is given the same weight remove the distorting effects that electoral boundaries may have on undifferentiated political equality. Partly for this reason, the pro-sovereigntist Parti Québécois can repeatedly succeed in forming a majority government (with a first-past-the-post electoral system) but repeatedly fail to convince a majority of the Quebec electorate to vote in favour of sovereignty. I am grateful to Stéphane Lévesque for pointing out this dimension of the argument to me. 7 In the 1980 referendum, the Government of Quebec asked voters for ‘a mandate to negotiate sovereignty-association’ with the rest of Canada. 8 Interestingly, some Quebec Nationalists have explained their referendum defeats in terms of the success that Federalists have had in mobilizing differences based on ethnicity, income, age, and gender within the Quebec community. 9 In the 1980 Quebec referendum, nearly 60 percent of the population voted No; in 1995, 50.6 percent voted No. 10 As Bowler and Donovan argue, the empirical evidence suggests that minority rights are most likely to be upheld in referendums when elites are united in endorsing them (see Bowler and Donovan, this volume). The claim being made here is that the power of elite endorsement will vary depending on whether a referendum involves issues in which different ideals of political equality conflict.

Part III The Impact of Referendums on Deliberation and Decision Making

8 Voters’ Decisions in the Nordic EU Referendums of 1994: the Importance of Party Cues Anders Todal Jenssen and Ola Listhaug

Parties play a minor role in most textbook comparisons between representative democracy and direct democracy. Parties are not only absent from the traditional depiction, they are sometimes described as having given up their democratic responsibility in the case of referendums. They leave the responsibility for solving difficult problems to citizens and withdraw from public debate, failing to give citizens the political guidance they badly need. In this portrait, the political scene is left wide open to various political extremists, and referendums are frequently described as rule by the least informed, the inept, and the irresponsible. Ordinary voters are fooled, seduced, and frightened into making incompetent or even disastrous decisions. Even the ‘populist’ refutation of the above argument contends that voters are more competent than usually believed. With the exception of Budge (Budge, 1996), the assumption that parties withdraw from the politics of direct democracy is rarely questioned. In this paper we investigate to what extent parties play a central role in the process of attitude formation by voters. Our main thesis is that referendum campaigns, at least in the countries under investigation, have much more in common with regular election campaigns than anticipated by some scholars. Despite the fact that parties usually have no formal role in referendum campaigns, party cues play a central role in the opinion formation process. This does not mean that we believe the differences between regular elections and referendums are trivial. Our research strategy is simply to take a first step in the discussion of the unique aspects of attitude formation in referendums by identifying what elements referendums have in common with regular elections. 169

170

Anders Todal Jenssen and Ola Listhaug

In the first section of the paper, we highlight the control that parties, especially parties that are part of a majority in parliament, have over the referendum institution. We argue that in some cases this control can be exercised to help parties take a leading role in referendum campaigns and to influence public opinion. We examine to what extent parties can theoretically influence public opinion through their formal control over such things as the timing of referendums, the allocation of campaign resources, the role played by ad hoc organizations, and the implementation of the referendum results. In the second section of the paper we discuss the problem of attitude formation during referendum campaigns. If the issue submitted to a referendum is new to most voters, how do voters manage to form reasonably developed attitudes? Through an examination of levels of information, we find that part of the answer is that political parties play a crucial role in helping voters understand issues. In the third section, we examine whether parties actually influence the vote. Do parties persuade and mobilize their traditional followers, and, if so, does that mean that referendum voting can be explained through reference to those factors traditionally used to explain election voting? We proceed by undertaking an examination of comparable national surveys conducted in Finland, Sweden, and Norway in 1994 during the referendums on the European Union. Data from these surveys (Moen et al., 1996) are used for most of the analysis in the present study.1 Other sources of information, mainly from the Norwegian referendum study, are utilized in subsections of the discussion.

The institutional framework The decision to hold a referendum Referendums are rare in Finland, Sweden, and Norway. National referendums have been held twice in Finland, five times in Sweden, and six times in Norway. In all three countries the right to initiate referendums lies with parliament and the government (Suksi, 1996; Ruin, 1996; Wyller, 1996). No mechanism for any form of popular initiatives exists. This means that parties exercise a great deal of control over many aspects of the process, such as timing. For example, the sequencing of the three Nordic EU referendums was hardly a coincidence. All three governments favored EU membership, though opinion polls and historical circumstances indicated that support for membership would be strongest in Finland and weakest in Norway. The referendum in Finland was held on 16 October, in Sweden on 13 November, and in Norway on 28 November. It is commonly assumed (and never publicly denied) that

Voters’ Decisions in the Nordic EU Referendums 171

the government parties planned for a cross-border bandwagon effect (Jahn and Storsved, 1995). The objective was to be both psychological – the activation of a subconscious fear of isolation – and political, with the future of Nordic unity (whether outside or inside the EU) of concern to many voters. In Norway this presumed effect was even given a unique name – ‘svenskesuget’ (‘the Swedish suction’) – and one estimate suggests that the Yes side may have gained as much as four points from this effect (Aardal and Jenssen, 1995, p. 38). The dominant party or parties exercise great control over whether and when to hold a referendum, typically calling one when they fear internal divisions (Bjørklund, 1982 and 1996). In Norway and Sweden the social democratic parties in office feared electoral losses in the parliamentary elections to be held in the period leading up to the referendums due to internal disagreement. Both parties therefore tried to ‘isolate’ the EU issue from the general election. In both countries the referendum was held after the election and in both cases the social democrats bluntly refused to discuss the issue during the election campaigns. The Norwegian party also delayed the final party vote on the issue until a new social democratic government was in place. Although in principle a simple majority in parliament is the only group that can initiate a referendum, in practice parties are sensitive to shifting public moods. If significant groups of citizens (in terms of number, power, or authority) demand a referendum, the parties may choose to initiate one. One form of pressure may be the example set by other countries. Suksi (1996) argues that the decision to hold a referendum on Finnish entry into the EU was very much inspired by previous referendums on EU integration in other countries. In Norway there was virtually no debate over the procedure, with the decision to hold a referendum taken unanimously by the Storting (the Norwegian parliament). This was in part because EU membership was turned down in the 1972 referendum, and no politician dared suggest that the decision could be overruled through a simple parliamentary vote. Hence, in this particular case, parties were left with very little choice – indicating that some of the control over whether and when to hold a referendum has been wrested away from parties. This is consistent with Morel’s argument (Morel, this volume) that the more frequent use of referendums is largely explained by the fact that an increasing number of referendums have become de facto obligatory. Resource Allocation In the Nordic countries, political parties are heavily subsidized by the state in a variety of ways. Information on party finances is scarce, but it

172


is likely that the lion’s share of party resources originates from the state. In other words, parties in parliament hold the key to their own economic well-being. For example, Norwegian parties in parliament granted NOK 12 million to their own EU referendum campaigns, in addition to NOK 12 million for the years 1990–92. These figures represent considerable amounts by Norwegian standards, and provide the parties with a significant material advantage during the referendum campaign. Although parties may prefer to abstain from referendum campaigns, particularly if they are deeply divided over the issue and fear party splits (Bjørklund, 1982), most parties in Norway, Sweden, and Finland campaigned vigorously in 1994. In his study of the strategies of the Norwegian parties before the EU referendum, Hjelseth (1995) discovered that all parties emphasized an ambition to integrate the EU issue into their broader ideological platform. The party’s position on the EU issue was to be explained as compatible with or even enhancing the party’s core values. This strategy reduced the risk of estranging long-time supporters and at the same time it increased the chances of success in the referendum. The strategy made the parties more significant as campaign organizations as well. As we will see later, they helped voters make sense of a very complicated issue by offering understandable frames of reference in terms of traditional party values and a ‘friends and foes’ schema. The parties were not the only political actors taking part in the 1994 referendum campaigns. Ad hoc organizations emerged in all three countries, the strongest of these being the Norwegian group ‘Nei til EU’ (No to the EU) (Jahn et al., 1998; Jenssen et al., 1998). At its peak, the organization had 140 000 members, more than any Norwegian political party. The organization was highly visible during the campaign. The leading ad hoc organization of the yes side was ‘Europabevegelsen’ (the European Movement) with 35 000 members. Both ad hoc organizations applied for state money for their campaign efforts. Each were granted two million NOK by the Storting for the years 1993 and 1994, significantly less than the parties, though such grants nonetheless underline that parties are more likely to be compelled to share power during referendum than election campaigns. The ad hoc groups are sometimes described as substitutes for political parties during referendum campaigns, with ad hoc activists supposedly quite different from the citizens participating in ‘routine politics.’ This was definitely not the case in the Nordic EU referendums of 1994. The ad hoc organizations can be better described as supplements to the

Voters’ Decisions in the Nordic EU Referendums 173 Table 8.1 Proportion of members of ad hoc groups who are also party members (Norway only; percentages reported)

Presently a party member (any party) Previously a party member Never a party member Source:

‘No to the EU’

‘European Movement’

Population

40

48

10

21

23

–

41

29

–

Jenssen et al. (1994); Election survey (1993).

parties. It is hard to find any signs of ad hoc organizations trying to overshadow the political parties; on the contrary, the ties between the ad hoc groups and the parties seem to have been quite strong. Informal talks were held to coordinate the campaign effort between them, with the Norwegian Center Party going so far as to copy the slogans of the ‘No to the EU’ movement word for word. The links between the parties and the ad hoc movements are illustrated by the number of party members that joined the ad hoc groups (see Table 8.1). Although large groups within both ad hoc organizations were not involved in party politics, a majority within both reported membership in political parties. Forty percent of those in the ‘No to the EU’ and 48 percent of those in the ‘European Movement’ were party members compared to only 10 percent within the general population. An additional 21 to 23 percent of members with the ad hoc organizations were past party members. Särlvik (1959) has reported even higher percentages of party members in the ad hoc organizations of the Swedish 1957 referendum campaign on supplementary pension plans. In addition, with the exception of a substantial group of social democratic party members within the ranks of the Norwegian No movement, members of parties supporting the Yes side joined the ‘European Movement’ and members of No parties joined the ‘No to the EU’ movement. Furthermore, the leadership of the two major ad hoc organizations was dominated by party members. The notion that ad hoc organizations represent ‘grassroots politics’ or an elite-directing (as opposed to elite-directed) mode of politics (Dalton, 1988) is clearly not valid in the case of the Norwegian 1994 referendum. The ad hoc organizations were supplements to the parties, not substitutes. The pattern seems to have been the same in Sweden and Finland (Jahn et al., 1998), with one possible exception. The Swedish ‘People’s movement against EU’ deliberately chose ‘ordinary’ citizens to present their arguments in radio and TV discussions.

174


Party control of interpretation and implementation Although referendums in Norway, Sweden, and Finland are advisory, they have sometimes been interpreted as de facto binding (Bogdanor, 1994, p. 27). This is not the case. The Swedish Riksdag (parliament) decided to overrule the 1955 referendum on right-hand driving. In Norway, paragraph 93 of the Constitution states that a change in the constitution after a referendum requires a three-quarters majority in parliament and prior to the referendum more than one-quarter of Norwegian MPs said they were ready to block EU entry if the Yes side won only a narrow majority. This would probably have led to some sort of political confrontation (Jenssen et al., 1998) and while the minority is within its right to veto change, whether this would have been morally and politically defensible is an open question. That being said, to argue that any result would have been automatically binding is to go too far. Space for party maneuvers becomes larger if more than two alternatives are offered in a referendum. The Swedish referendums on supplementary pension plans (1957) and nuclear energy (1980) provide interesting examples. In both referendums the voters were faced with three different alternatives. Not surprisingly, no alternative received a majority and no second ballot was offered. In subsequent parliamentary debates, the governing party was able to adopt its preferred option. Even in countries with binding referendums, such as Denmark, parties administer the outcome of the votes. The two Danish referendums on the Maastricht Treaty provide an interesting example. When the Maastricht Treaty was turned down in the 1992 referendum, the parties in government renegotiated some elements of the treaty (labeled the ‘Edinburgh agreement’ in Denmark) and submitted it to a second referendum the following year. With the new version of the treaty, the governing parties were able to split the popular opposition and win a comfortable majority. Even binding referendums are perhaps not that binding after all. In the Nordic countries, it becomes apparent that the depiction of referendums as events beyond the realm of party politics is highly suspect.

Thinking about the unknown Few citizens are heavily involved in politics. Even during election campaigns, only a minority actively seeks out information. This has been explained both as a result of the costs involved in gathering information (in the rationalistic tradition) and as a result of limitations in human cognitive capacity (in the tradition of schema theory and


symbolic politics). In electoral politics, voters are offered several information shortcuts, of which the most important is the well-known historical records of the political parties. By anticipating that parties stand for what they used to stand for, the need for cognitive processing is minimized. The typical referendum issue – if such a thing can be said to exist – is not unknown to the voters, but it is new in the sense that many voters have no prior knowledge about or view on it. Hence, one can argue that referendums are more demanding for the average citizen than elections. On the other hand, referendums usually offer only two alternatives, whereas several parties with different positions on numerous issues compete in most democratic states. One could also argue that if the parties offer ‘cues’ to voters in referendums by taking positions on the questions at hand, which they did in the Nordic referendums of 1994, the task of making up one’s mind is not very different from regular elections. The literature on attitude formation describes a variety of ‘cognitive shortcuts’ and ‘heuristic tools’ that are supposed to ease the process of attitude formation. Nevertheless, citizens do not take positions on all issues. Some issues are simply deemed irrelevant by many. In their discussion of turnout in referendums and elections, Butler and Ranney (1994) find that average turnout is lower in referendums than in elections. A closer look at their data, however, shows that while average turnout is lower, standard deviation of turnout is much higher, with turnout varying greatly by the type of issue being submitted to a vote. The lowest turnouts in democratic states frequently occur on issues of low salience to many voters, like new adoption laws (Ireland, 1979: 28.0 percent), hunting laws (Italy, 1990: 43.3 percent) or driving on the right-hand side (Sweden, 1955: 53.2 percent). The EU issue does not fall into this category. The issue was submitted to referendum partly because of its extraordinary importance. Consequently, most voters might be expected to hold reasonably informed attitudes and be able to give arguments for their voting intention. It is impossible to establish absolute standards for what we might call ‘informed attitudes,’ yet we can begin by examining respondents’ ability to answer correctly basic factual questions related to the issue. To avoid a political bias in the questioning (the two sides tended to highlight different facts) the questions focus on indisputable but perhaps somewhat trivial matters. To help us evaluate the level of factual knowledge, we have also included assessments of personal knowledge. The results reported in Table 8.2 suggest that most voters were familiar with some basic facts about the EU, like the number of member states and

176


Table 8.2 issue %

Factual knowledge and assessments of personal knowledge on the EU

Feel ‘very well’ or ‘fairly well’ informed on EU issue Know the number of EU member states Know the location of the European Commission Know the number of votes for own country in the Council of Ministers

Norway

Sweden

Finland

53.5

–

37.1

67.3 97.0

62.0 93.7

61.1 93.3

50.5

51.2

–

the location of the European Commission. When confronted with more demanding questions (like the number of votes accorded to their own country in the Council of Ministers) only a bare majority was able to give the correct number. The level of factual knowledge seems to have been somewhat higher in Norway than in the other countries. This is not surprising given that the Norwegian campaign lasted longer than the campaigns in Sweden and Finland, and the fact that the question of EU membership had been submitted to referendum before. For the same reason more Norwegian voters than Finnish voters felt ‘very well’ or ‘fairly well’ informed on the issue. The ability to formulate arguments in one’s own words (‘level of conceptualization’) is another relevant indicator. In our study, voters were asked to give up to three arguments in their own words for their voting intention. They were also asked to give up to three arguments for the opposite position, if they believed that relevant counter-arguments existed. This second question thus serves as an indicator of the level of polarization as well as an indicator of cognitive involvement. Almost all voters with a voting intention were able to refer at least one argument as a motive for their voting intention (see Table 8.3). On average, they were able to mention about two arguments. Again the Norwegian voters seem to be somewhat more cognitively involved, but the differences are small. The level of conceptualization is somewhat higher among Yes voters, probably the result of a somewhat higher level of education in this group. Fewer voters were able and willing to refer to counter-arguments. Although many felt that no valid counterarguments existed, these results also reflect basic cognitive limitations. Most people find it hard to reproduce information inconsistent with their own views, though if we bear in mind the polarized nature of the

Voters’ Decisions in the Nordic EU Referendums 177 Table 8.3

Level of conceptualization before the Nordic EU referendums of 1994 Norway

Able to give at least one argument (%)a Average no. of arguments Mentioned at least one counter-argument (%)a Average no. of counter-arguments a

Sweden

Finland

Yes voters

No voters

Yes voters

No voters

Yes voters

No voters

95

93

96

86

94

95

2.19 61

2.17 46

1.82 58

2.05 49

1.93 48

1.92 35

1.81

1.60

1.66

1.49

1.48

1.48

The interviewers were instructed to record a maximum of three arguments.

Source:

Oskarson and Ringdal (1998), tables 7.4 and 7.5.

debate, a surprisingly large number of voters were able to refer to what they perceived as reasonable counter-arguments. During election campaigns parties are a primary source of information, mediated through the mass media. Studies of media content during the final stage of the Norwegian referendum campaign reveal that 34 percent of all individuals that appeared in the TV coverage (Nilsen, 1996) and 35 percent of those appearing in two leading newspapers (Ramberg, 1995) were active top-rank politicians. Yet the parties’ role in the news media may have been even more important than these figures suggest. When respondents in the Norwegian survey were asked if they trusted anybody in particular among all those publicly taking a stand on the EU issue, 89 percent of the persons mentioned were leading party figures. Very few mentioned the leaders of the ad hoc organizations. Some voters are more predisposed to accept party messages. According to Figure 8.1, we see that information and levels of conceptualization are higher among those who identify with the party sending the message, as long as the voter favors or at least leans towards the position advocated by the party. For those voters who experience cross-pressure (that is, identifying with a party whose position on the referendum is different than one’s own) exposure to the party’s campaign message may hinder attitude formation. The level of conceptualization is clearly higher among party identifiers supporting the party’s position than among voters without party identification and among voters experiencing cross-pressure.

178


2.5 Conceptualization among ‘No’ voters

Number of arguments/correct answers

Conceptualization among ‘Yes’ voters Knowledge among ‘No’ voters Knowledge among ‘Yes’ voters 2

1.5

1 Strong PID: ‘No’ party

Weak PID: ‘No’ party

No PID

Weak PID: ‘Yes’ party

Strong PID: ‘Yes’ party

Figure 8.1 Level of conceptualization and factual knowledge by strength and direction of party identification (pooled data from Finland, Sweden, and Norway).

Party influences on the vote The analysis in the section on the institutional framework above supports the view that parties play an important role in the Nordic EU referendums when it comes to setting the agenda, controlling the resources for the campaign, and recruiting the leadership and activists. The following section has shown how parties play an important role in providing voters with information and helping to conceptualize the issue for them. We now extend the analysis to voting by examining the influence parties have on the vote. Positioning the parties Parties did not avoid taking public positions on the EU issue, despite the fact that for many the issue created intra-party problems and controversy and despite the fact that the decision to hold a referendum was taken in part to minimize the problems for the parties by lifting the


issue out of the conventional political arena (Bjørklund, 1997). Of the 25 parties that we have studied, only the Green Party in Finland refrained from giving advice to its followers (for a more detailed analysis of some of the questions raised in this section, see Listhaug, Holmberg, and Sänkiaho, 1998). Parties were not evenly distributed on the two sides, with all major parties in all three countries favorable towards joining the European Union. In Finland, the tiny Rural Party, the small Christian Party, and the somewhat larger Left Party were the only parties to go against membership. The balance in favor of membership was equally skewed in Sweden where the Left Party and the Greens were against, and all the traditional governing parties were for membership. In Norway, the situation was somewhat more balanced with three centrist parties against – the Liberals, the agrarian Center Party, and the Christian People’s Party. The Socialist Left Party was also against membership. Labor, the Conservative Party, and the Progressive Party were in favor of membership. How the parties chose to position themselves on the EU question was strongly influenced by each country’s history and industrial structure, as well as by its contemporary social and economic problems (Jenssen et al., 1998, chapter 14). The short version of the comparative pattern includes the following elements. The breakdown of communism gave Finland an opportunity to become a full-fledged member of Western Europe, and leave behind its non-aligned status between East and West. Moreover, the economic crisis that hit the country in the first part of the decade, partly as a result of loss of markets in Russia and Eastern Europe, led to record levels of unemployment, peaking near the 20 percent mark. The economic problems strengthened the case for full access to new markets and to the political influence in the institutions that regulate these markets. Sweden was also hit by an economic downturn in the period leading up to the referendum, including unemployment reaching above 8 percent. Norway suffered an economic setback in the first part of the decade, but unemployment peaked at a lower level than in Sweden. The economic recovery in Norway came early and was well under way at the time of the referendum. Contributing to the impression that membership was less urgent for Norway was the widely held view that Norway commanded a wealth of resources, notably oil, that gave her independence and weakened economic arguments for membership. Within some parties, internal opposition and a considerable division of opinion among supporters created problems of party unity and strategy and put limits on how hard the party could push the issue and how

180


distinct and clear their position could be. Such a situation should undermine the ability of parties to mobilize their supporters through clear cues. A prime example is the Finnish Center Party, which was strongly divided both among the leadership and at the grass roots. The Swedish Center Party also had trouble taking a stand on the membership issue. The Center Party had tried to run on a platform emphasizing support for regions outside the large metropolitan areas and, at least partly, on green issues. Such an issue legacy was difficult to reconcile with the pro-EU stance taken by the party. Opposition to EU membership by many within the party, as well as a divided party elite, led the party to a run a low-profile campaign (Holmberg, 1996). The other party in this category in Finland is the Green Party, with its no-advice position. Most parties, however, were able to advocate their chosen position so clearly that voters perceived them to be fairly distinct on the issue. In the interviews, respondents were asked to place the parties on an 11-point scale where zero indicated a position strongly negative towards membership and 10 a position very positive. A reading of 5 signified the neutral point. We assume that parties that are located between 4 and 6 were giving very weak signals, and thus provide the voters with unreliable cues on how to decide in the referendum. In total, only four parties are located in this area (results reported here are available in tables from the authors). Two of the Finnish parties are in the neutral zone, the small Green Party and the important Center Party, with mean perceived positions at 4.38 and 4.82, respectively. In Finland the most extreme parties are the Rural Party (2.74) and the Conservatives (8.42). In Sweden the only party in the neutral zone is the Center Party at 5.30. In Sweden the most distinct No party is the Left Party at 1.09 and the most distinct Yes party is the Conservatives at 9.44. In Norway, the Christian People’s Party is in the neutral zone with a reading of 4.21. This is on the side that the party advocated, but is nevertheless unexpected on the basis of the relatively clear No cues given by the party. The Center Party is the most extreme No party with a mean perception of 0.84. On the opposite side, Labor is perceived to be favorable towards membership, recording a mean of 8.62. It might be somewhat of a surprise that Labor is slightly more pro-EU than the Conservative Party despite the fact that opposition to the official party line was stronger within Labor than within the Conservative Party. A possible explanation is that Labor was in government and the Conservatives were not. The leading Labor ministers, including the extremely popular Prime Minister Brundtland, were all strong advocates of membership, giving the party high visibility as a pro-EU force.


Party cues and the vote That parties take a stand and that they send their message are only the prerequisites for a real impact on voting behavior. In a detailed study in Norway, the effect of party identification on the vote has been measured in a multivariate model (Pettersen, Jenssen, and Listhaug, 1996). After controlling extensively for a variety of cleavages, the research demonstrated that party support had a significant impact on the referendum vote. Research for Finland and Sweden has been limited to bivariate analysis of the relationship between party support and referendum vote (Listhaug, Holmberg, and Sänkiaho, 1998) or to the impact of cleavages and the vote (Ringdal and Valen, 1998). In this section we will expand the analysis of Finland and Sweden in the same multivariate framework as we have developed for Norway. Party systems as well as voting behavior in Scandinavia can be explained by a cleavage model (Rokkan, 1967 and 1970; Rokkan and Valen, 1964). According to this model, lasting antagonisms were created at important historical crossroads in the nation-building process and by the industrialization of Western Europe. Although the historical cleavages have lost some of their political impact during the last decades, the cleavage pattern is still recognizable in electoral politics (Franklin et al., 1992) and can be central to referendum voting (Steiner, this volume). It is indeed a main conclusion of major studies of the two EU referendums in Norway that the battle over membership was fought along the lines of traditional cleavages (Valen, 1973; Jenssen and Valen, 1995). In fact, historical cleavages explain somewhat more of the variance in the Norwegian referendum of 1994 than of 1972 (Jenssen et al., 1995, table 7.7). Territorial, cultural, and economic cleavages can also be used to explain the vote during the referendums on EU membership. Other cleavages, such as gender and sector of employment, have also been linked to referendum voting, with women and those in the public sector more likely to vote no (Pettersen, Jenssen, and Listhaug 1996; Gilljam, 1996; Ringdal and Valen, 1998). There are many plausible ways to include party and cleavages in a single model. We have chosen to expand on the comparative cleavage model of Ringdal and Valen (1998). They include a large set of relevant cleavage variables that have equal comparative relevance in the three countries. The model includes important indicators of traditional class, aspects of region and urban structure, as well as new cleavages of gender and sector of employment. The comparative nature of the model means that we have dropped some cleavages that are specific to Norway, such as some aspects related to traditional culture (Pettersen, Jenssen, and Listhaug, 1996).

182


The results of the multivariate analysis are laid out in Table 8.4. Since the focus of this paper is the impact of parties, we will just briefly mention the most important aspects of the impact of the cleavage variables. We observe that most of the signs are in the same direction in the three countries: support for membership in the European Union tends to be greater among urban, middle-class, private sector, high education, and male voters (variations in sample size makes it easiest to obtain statistically significant results in Norway, and most difficult in Finland). A striking difference between the countries is that farmers are strongly opposed to membership in Finland and Norway while Swedish farmers are as likely to vote Yes. This has a structural explanation: Swedish farms are more competitive than farms in Finland and Norway and less dependent on government subsidies. We also observe that geographical patterns in the vote are somewhat stronger in Norway and Sweden than Finland, though the Yes vote is concentrated in the southern parts of all countries, especially in the counties containing the national capitals. In general, parties had a strong impact on the vote decision. In Finland three parties stand out as mobilizers of the Yes vote. These are the Conservative Party, the Social Democratic Party, and the Swedish People’s Party. The effect for parties opposed to membership is also clear, with strong effects recorded for the Christian Party (–0.50), the Rural Party (–0.21) and the Left (–0.17). The number of supporters for the Rural Party is, however, very small, and the coefficient is not statistically significant. A quite interesting aspect of the Finnish case is the negative impact of support for the Center Party on the vote. This is clearly a case where the core of party supporters went against the advice of the leadership, although the leadership was divided. The effect, however, was very weak. The Green Party, which, like the Center Party, was perceived in the neutral zone (and also took a neutral position), records the same coefficient as the Center Party (–0.11) – not significant, though it suggests that Green supporters were as divided as their leaders, but leaning in an anti-EU direction. In Sweden there are four party effects of some significance. Support for the Liberals and for the Conservative Party are strongly associated with a pro-EU vote while those who identified with the Greens and the Left Party followed cues and voted No. The Swedish case is remarkable in that three parties that urged their supporters to vote in favor of membership were unable to mobilize their voters in this direction. The Social Democrats, the Center Party, and the Christian Democrats record coefficients close to zero, with the Center Party and the Christian Party even having a wrong sign. The Swedish case demonstrates that even

Union member Income: High Region: Middle Capital area

–0.04 0.09 0.04 0.09

0.04 0.10

0.03 0.06 0.04 –0.11 0.12 0.03 –0.04

0.04 0.08 0.09 –0.22 0.17 0.05 –0.06 –0.04 0.10

0.05 0.06 0.07 0.08

Beta

0.05 0.02 0.01 0.07

B

Finland

0.28 0.05

0.24 0.00

0.38 0.12 0.18 0.00 0.00 0.42 0.19

0.09 0.12 0.10 0.02

p value

0.11 0.13

–0.05 0.05

0.07 0.07 0.05 0.01 0.10 –0.01 –0.08

0.09 0.04 0.02 0.08

B

0.11 0.12

–0.05 0.04

0.07 0.04 0.03 0.00 0.04 –0.00 –0.04

0.09 0.13 0.13 0.08

Beta

Sweden

Determinants of Yes vote: parties and cleavages (all three countries)

Male Age in 10 yrs Education in years Private sector Class: Low non-manual High non-manual Self-employed Farm Student No class Unemployed

Table 8.4

0.00 0.00

0.07 0.11

0.01 0.13 0.27 0.86 0.16 0.90 0.07

0.00 0.00 0.00 0.00

p value

0.13 0.15

–0.02 0.10

0.02 0.09 0.08 –0.12 0.11 –0.02 –0.02

0.07 0.01 0.01 0.05

B

0.12 0.14

–0.02 0.08

0.02 0.06 0.04 –0.05 0.06 –0.01 –0.01

0.07 0.02 0.07 0.05

Beta

Norway

0.00 0.00

0.27 0.00

0.35 0.00 0.05 0.01 0.01 0.65 0.68

0.00 0.35 0.00 0.03

p value

183

(continued)

SLEFT SSOC DEM SCENT SLIB SCONS SCHR SGREE

FCENT FSOC DEM FCONS FLEFT FGREE FSPP FRP FCHR

Pop. density: Middle density High density

Table 8.4

–0.11 0.20 0.25 –0.17 –0.11 0.17 –0.21 –0.50

0.03 0.06

B

–0.07 0.14 0.17 –0.06 –0.04 0.05 –0.04 –0.12

0.03 0.06

Beta

Finland

0.04 0.00 0.00 0.03 0.19 0.07 0.19 0.00

0.43 0.17

p value

–0.33 0.01 –0.05 0.23 0.27 –0.12 –0.33

0.14 0.11

B

–0.09 0.01 –0.02 0.08 0.16 –0.02 –0.06

0.14 0.09

Beta

Sweden

0.00 0.83 0.43 0.00 0.00 0.26 0.01

0.00 0.00

p value

0.05 0.12

B

0.04 0.12

Beta

Norway

0.07 0.00

p value

184

1069

N (weighted)

0.00

–0.05

1725

0.18

0.45

0.09

0.25 0.18 0.36 –0.28 –0.29 –0.26 –0.00

2366

0.28

0.21 0.04 0.24 –0.10 –0.14 –0.10 –0.00 0.07

0.00 0.03 0.00 0.00 0.00 0.00 0.99

Note: A more detailed description of the cleavage variables can be found in Ringdal and Valen (1998). The dependent variable is coded 1 (yes to membership in the EU) 0 (no to membership in the EU). Male is coded 1, female 0, age is coded in ten-year units with 20 as zero point, private sector is coded 1, public sector or unknown sector is coded 0, the occupational class variable is coded as dummy variables for each category with 1 if the person is member of the category, and 0 otherwise (no class is a category for persons that cannot be allocated to any of the categories, mostly because we lack information about their occupation), unemployed is coded 1 and 0 otherwise, union members are coded 1, all others are coded 0, high income is the highest quartile and is coded 1, others are coded 0, region is coded as two dummy variables, 1 for middle, or central part of the country, 0 for those who do not live in this region, 1 for the region of the capital, 0 for those who do not live here, population density is also coded as two dummies (1, 0) for the two categories middle and high density. The reference categories for some of the variables (in parentheses) are: gender (female), sector (not private), class (workers), unemployed (not unemployed), income (lowest quartiles), region (north), population density (low). The first letter in the party names is F for Finland, S for Sweden, and N for Norway. The Finnish parties are: CENT, Center Party, SOC DEM, Social Democratic Party, CONS, Conservative Party, LEFT, Left Party, GREE, Green Party, SPP, Swedish People’s Party, RP, Rural Party, CHR, Christian League. The Swedish parties are: LEFT, Left Party, SOC DEM, Social Democratic Party, CENT, Center Party, LIB, Liberal Party, CONS, Conservative Party, CHR, Christian Democratic Party, GREE, Greens. The Norwegian parties are: LAB, Labor Party, PROG, Progressive Party, CONS, Conservative Party, CHR, Christian People’s Party, CENT, Center Party, SOC LEFT, Socialist Left Party, LIB, Liberal Party.

0.21

0.27

R sq.

Constant

NLAB NPROG NCONS NCHR NCENT NSOC LEFT NLIB

185

186


parties that are accurately placed spatially (the mean perception for the Christian Democrats was 6.29 and for the Social Democrats 7.68) may fail to move their voters in the direction of the party. A few speculations on the Swedish results can be offered. First, all three parties had trouble in reconciling their pro-EU position with the traditional legacies of the parties. This is maybe most obvious for the Center Party which is a party favoring rural interests as well as environmental values, stances that ran into conflict with images of how European Union was seen by voters. Second, Sweden has seen a long-term decline in trust in government and politicians. This could have weakened parties that defended positions linked to the interests of the governing political elite. Membership in the European Union clearly was an issue position which was favored by the power elites in Sweden. Norway is a stand-out case for party effects. Only the Liberal Party is unable to move its supporters in the direction advocated by party leadership. It is remarkable that the Labor Party, which suffered from an intra-party opposition not much different from that of the Swedish Social Democrats, was able to move its party supporters much more clearly in the pro-EU direction (b = 0.25). Support for the Conservative Party has a coefficient of 0.36; for the Progressive Party it is 0.18 – barely of statistical significance. In spite of the fairly neutral image of the Christian People’s Party their supporters are as strongly opposed to membership as supporters of the other main No parties, the Center Party and the Socialist Left Party. The results in Table 8.4 are derived from cross-sectional data measured for one point in time, hardly satisfying if one wants to make statements about the causal mechanism in the dynamics between voter and party. A closer examination, however, shows that the link between party position and referendum vote gets somewhat stronger during the referendum campaign. This development is especially marked in Norway (Listhaug, Holmberg, and Sänkiaho, 1998). There are two main explanations for the strengthening of the link between party position of the voters and referendum vote over the period of the campaign. Voters can change position on the question to bring themselves in line with the position of their preferred party (that is, persuasion). Or they can change party affiliations to correspond to their position on the issue (that is, policy voting). The availability of panel data from the referendum campaign allows us to examine this question, though some problems remain in attempting to establish the direction of causality. Table 8.5 presents the main results. The table is a somewhat simplified version of the results of the earlier analysis for Sweden (Holmberg,

Voters’ Decisions in the Nordic EU Referendums 187 Table 8.5 Change of party and change of EU position during the campaign (% reported)

Stability – same party group and issue position Change – but no relationship between party group and issue position Change – a relationship between party group and EU position – Change of both positions – Change of EU position caused by party group – Change of party group caused by EU position Sum (N =)

Finland

Sweden

Norway

53

68

55

24

11

17

23

21

28

100 (1111)

4 11

2 12

5 15

7

6

7

100 (1533)

100 (2202)

1996) and the comparative analysis (Listhaug, Holmberg, and Sänkiaho, 1998). In all countries stability is by far the most prevalent pattern: 53 percent in Finland, 68 percent in Sweden, and 55 percent in Norway maintained their position on the issue and their attachment to a party group between the first and second surveys. By party group we mean parties that share the same position on the issue, Yes parties as contrasted to No parties. For a second group we observe panel change (either on partisanship or the EU issue), but without a systematic relationship between party group and EU position. The third group is the most interesting and includes those who changed position or party in a systematic pattern during the campaign related to the EU issue. The size of this varies from 21 percent in Sweden, 23 percent in Finland, and 28 percent in Norway. Within this group of switchers, the largest group in all countries is those who changed position on the issue in line with the position of the party group which they supported. It seems likely that these voters were persuaded by their parties. The size of this category is 11 percent in Finland, 12 percent in Sweden, and 15 percent in Norway. The group of policy voters is about half that in each country. Ray (1999) has studied the EU issue in Norway over a longer timeframe, including the parliamentary election of 1993, where the EU played a major role. His research shows that policy voting was more important than persuasion in the 1993 election, but that persuasion

188


was more important than policy voting in the referendum campaign. His study is based on balance theory: if voters disagree with their parties on an important issue, balance can be restored either by policy voting or by persuasion. The relative strength of the two processes will depend on the cognitive task at hand. Because the choice of party is the important task in a parliamentary election, one will find a greater number of voters voting for the party they agree with on an issue. During a referendum, the task is to decide on a particular issue, and one may use party advice to guide opinion. Parties played a role in influencing voters’ decisions in the referendums, but parties were not alone. In all countries ad hoc movements were created. The No movements were stronger than the Yes movements. The movement against membership was especially powerful in Norway, where it built on the legacy and experience from 1972. Much of the research on the Norwegian EU referendums has focused on the ad hoc movements, with one leading scholar arguing that parties stepped aside in favor of these movements (Bjørklund, 1997, p. 105). In the survey respondents were asked to list the importance of the relevant actors in influencing their vote. The following actors were listed: family, closest friends, colleagues at work/school, party, newspaper of choice, TV, and the major pro- and anti-EU movements. In all countries, TV was listed as the most important influence. In Finland, newspapers came second, followed by family and then parties (tied with close friends). In Sweden, parties were tied with family in second place. In Norway, parties came second, just behind TV. In all countries parties were rated ahead of the ad hoc movements. However, among those voting No, the impact of the No movement was considered more important than party, while Yes voters considered party more important than their ad hoc movements. The likely explanation for this is imbalance in the position taken by the major parties of both the right and the left in favor of the Yes outcome in all countries. It appears that only when no major parties assume a leadership role in opposition can an ad hoc movement emerge to fill the gap. The effect of party, though, is substantially greater for those who identify with a party. Despite some differences between the three countries, the data reported in Table 8.6 show that parties were especially important to those on the Yes side with a party identification, though even those Yes voters without a party identification relied heavily on parties. Although the No movement was credited by No supporters with slightly more importance than the party, it was only in Finland where those without a party identification were significantly more likely to

Voters’ Decisions in the Nordic EU Referendums 189 Table 8.6 Influence of party and ad hoc movements on the vote: percentage saying group was ‘important’ or ‘very important’ (N in parentheses) Vote in the referendum Yes Party identification Finland

Party Anti-EU movement Pro-EU movement

Sweden


Norway


No

PID

No PID

PID

No PID

27 (376) 11 (372) 16 (372)

12 (282) 10 (290) 17 (291)

23 (255) 28 (226) 13 (225)

8 (200) 22 (205) 16 (207)

45 (411) 9 (408) 14 (409)

24 (427) 8 (427) 13 (426)

22 (317) 25 (314) 11 (314)

18 (394) 22 (390) 12 (390)

53 (707) 20 (689) 20 (689)

32 (404) 17 (414) 13 (413)

40 (579) 45 (579) 12 (569)

27 (552) 35 (581) 13 (565)

rely on the No movement. It appears that for both those with and without a party identification, when major parties fail to mobilize opinion, ad hoc groups become more important.

Conclusion We began this paper with the observation that referendums and normal electoral politics are often thought to be two distinct varieties of political behavior. In parliamentary elections, parties are key actors; in referendums political parties are often sidelined and replaced by ad hoc political movements. We have demonstrated that, contrary to this view, parties took an active part in the politics of the referendum process. Parties had a key role in the institutional aspects of the activity by deciding on the timing of the referendums, recruiting leadership to the ad hoc organizations, and controlling the economic resources that were put into the campaigns. Most of the parties took fairly strong stands on the issue and they played a leading role in influencing opinion. The results of

190


our analysis have pointed to important areas of commonality between elections and referendums, in particular the role of party identification and enduring political cleavages in influencing the vote. One difference, however, is that during referendums voters are more likely to switch positions on an issue in order to follow a party’s cue, whereas during elections voters are more likely to shift parties to reflect their issue positions. We must keep in mind that the three EU referendums in Finland, Sweden, and Norway in 1994 give us a very limited basis for generalizations, which leads us to recommend that scholars follow the path of the recent study by Hug and Sciarini (2000) and work on strengthening the cumulative research on referendums across nations and over time. Under what circumstances will parties have their greatest effect and why are some parties more successful than others in mobilizing their supporters? A larger database on referendums would allow us to systematically compare the effects of small and large parties, of parties in power and parties in opposition, and of parties that are united or split on the issue.

Note 1 A full description of the three studies, including design, fieldwork, funding, etc. is given by Moen et al. (1996). In all three countries representative samples of the adult population were interviewed before and after the referendum. In the Norwegian study respondents were interviewed three times, twice in the Swedish and Finnish studies. To ensure identical wording, the group of researchers involved in the project (cf. Jenssen et al., 1998) closely coordinated the national studies. Data and full documentation is available from Norwegian Social Science Data Services, Hans Holmboesgate 22, N-5007 Bergen.

9 Are Voters to Blame? Voter Competence and Elite Maneuvers in Referendums Arthur Lupia and Richard Johnston

The history of referendums is one of triumph and tragedy. For some, referendums were the platform from which new and important political freedoms were launched. For others, referendums ended important freedoms or were critical in the dismantling of democracy itself. What is the proper lesson to draw from the history of referendums? Many opponents of direct democracy claim the lesson to be that voters cannot be trusted. Some supporters of the process claim the reverse, seeing referendums as necessary checks on elites. In reality, both arguments are problematic. To help clarify the implications of policy making by referendum, we present an essay that addresses two questions: 1. Are referendum voters as incompetent as commonly portrayed? 2. Are the real villains of the piece in history’s most pernicious plebiscites the elites who made the reference in the first place, chose the wording of the ballot question, and controlled the subsequent interpretation of the vote? Our answer to the first question is that referendum voters are not as incompetent as commonly portrayed. Our answer to the second question is that elite maneuvers are a critical part of the referendum process and that elites are far from blameless for its foibles. Together, these answers provide a different vision of the referendum process and imply a different set of remedies to some of the processes’ perceived imperfections. We begin with voter competence. Voters are not well informed as a rule, even about basic facts that seem relevant to connecting political 191

192

Arthur Lupia and Richard Johnston

means and ends. We take as fact the characterization of voters as commonly ignorant of political facts, for the burden of evidence is overwhelming. We then explore the relationship between information and competence. The relation proves to be remarkably loose. In the referendum context, voters are typically given only two choices: vote yes or vote no. This aspect of the process simplifies the voters’ decision significantly. It makes much of the information available to voters superfluous with respect to their ability to cast the same vote they would have cast if they were so well informed as to resemble walking political encyclopedias. As a result, we find that the conditions for competent performance – properly defined for referendums – are not as difficult to satisfy as much commentary suggests. Certainly competent performance does not require that voters be fully informed about all of the minute details of a ballot question. From a historical perspective, however, some referendum electorates clearly made catastrophic choices, to the point of undermining the very democratic values in whose name referendums are conducted. The plebiscitarian rises to power of Napoleons I and III come to mind, as does that of Adolph Hitler. To be certain, results such as these are often the result of accounting fraud, not to be held against voters themselves. Sometimes, however, anti-democratic accusations stick, in the sense that unsavory referendum results appear to be an authentic expression of popular will. Are voters to blame? Our research exposes voter incompetence, the traditional and easy answer to this question, as one for which evidence is sparse. Our search for a more plausible explanation brings us to elite actions. Political elites play an important role in the referendum process. For in the typical referendum, someone must choose to send a question to the people, someone must choose the wording of the question, and someone must choose the timing of the referendum. The actors in question are political elites who, in addition to taking the actions just listed, are in a privileged position to affect the plebiscite’s subsequent interpretation after the election (that is, they make claims about what the vote really meant). That elites take these actions is neither random nor accidental. Substantial and multiple elite maneuvers are a necessary component of the referendum process. To clarify the extent to which elite maneuvers affect voter liability in the referendum process, we explore the role of referendums in the fall of the French First and Second Republics – examples that commonly fuel antipathy to the institution. While we find that the extent to which voters are to blame for these low points is quite limited, our efforts do not suggest

Are Voters to Blame? 193

that the referendum process can survive any degree of voter ignorance. They do, however, suggest that when voters use referendums to maim basic democratic virtues, elites are usually close by, making victims and sharp instruments readily available. Together, the account of voter competence in the first part of the paper and the historical reconstruction of a referendum low point in the second part provide a picture of the referendum process that defies convention. It portrays referendum outcomes not as the rash acts of an incompetent populace, but as the product of elites and citizens seeking simplicity in the face of complex choices. Accordingly, we hope that our examination of referendums’ informational requirements and our recounting of the elite moves that precede referendums serve to replace two fruitless caricatures of the referendum process. The first caricature involves disparaging referendums because voters lack information. This caricature is flawed because it conflates information and competence. The second caricature is to support referendums as mechanisms guaranteeing control of elites. It is flawed because it assumes that referendums can somehow occur independently of elite initiative and influence.

The question of competence Do referendum voters know enough to make competent decisions? A common answer to these questions evokes an image of voters who cannot be trusted with important matters. Supporting this idea is the ever-expanding body of survey evidence on citizen ignorance, about which Somin (1999, p. 417) draws a representative judgment: Overall, close to a third of Americans can be categorized as “knownothings” who are almost completely ignorant of relevant political information, (Bennett, 1988) – which is not, by any means, to suggest that the other two-thirds are well informed. Three aspects of voter ignorance deserve particular attention. First, voters are not just ignorant about specific policy issues, but about the basic structure of government and how it operates (Neumann, 1986; Delli Carpini and Keeter, 1991 and 1996, chapter 2; Bennett, 1988). Majorities are ignorant of such basic aspects of the US political system as who has the power to declare war, the respective functions of the three branches of government, and who controls monetary policy (Delli Carpini and Keeter, 1996, pp. 707–71). This suggests that voters not only cannot choose between specific competing policy programs,

194


but also cannot accurately assign credit and blame for visible policy outcomes to the right office-holders. In addition to lacking information, voters commonly substitute simple, potentially misleading cues for the information they lack. This fact does not go unnoticed by the high-priced firms behind modern media campaigns, most of whom earn their keep by reducing debates to bumper sticker sized slogans and the slick imagery of 30-second television advertisements.1 But is information the same thing as competence? More to the point, does competence require that voters mobilize encyclopedic information about politics? To answer these questions we must be precise about what we mean by competence. For the referendum context, we define the term as follows. A voter’s choice is competent if it is the same choice that she would make given the most accurate available information about its consequence. Would she make the same decision if fully informed about the consequences of her actions? If yes, then her choice is competent. 2 Far from requiring that the voter actually be fully informed, however, the definition stipulates only that the voter chooses as if she were. It specifically leaves open the possibility that the information itself is not required. This definition stands at odds with prevailing definitions of competence in political science – definitions that demand levels of knowledge which are encyclopedic in character and focus on attentiveness to political detail (for example, Luskin’s (1987) political sophistication). For something like sophistication or knowledge of details to be a necessary condition for competence, the following must also be true: people can make reliable predictions about the consequences of their actions only if they know a particular, detailed set of facts about these actions. But this assumption is false. Citizens can and regularly do use limited amounts of information to make the same decisions they would have made if they knew the consequences. Suppose, for example, that knowledge of a particular set of facts is sufficient for a competent choice (for example, suppose that knowing Bill Clinton’s position on 100 political issues is sufficient for a competent vote in the 1996 US presidential election). Then, if a person does not know these facts, and cannot access any other facts that allow her to make the same choice, then she cannot choose competently. If, however, there exists another, perhaps simpler, set of facts that leads her to make the same choice (that is, Bill Clinton is endorsed by the Sierra Club), then knowing the initial set of facts is not a prerequisite for


competence. When a few, simple pieces of information can lead citizens to make the same choices as many, complex pieces of information, citizens can be competent without detailed knowledge. Competence and information are different. Competence is the ability to make accurate predictions; information is data. Competence requires information because accurate predictions require data – at a minimum you need some data to verbalize the prediction you are making. By contrast, you can know a long list of facts and fail to put them together in a way that allows you to make accurate predictions. Thus, while competence requires information, you can have information without being competent. What does competence – as we have defined it – mean in the referendum context? The answer to this question lies in an almost universal attribute of referendums – of the thousands of initiatives held around the world, nearly all ask voters to make a binary choice. The typical referendum allows voters to choose one of two alternatives – the piece of legislation described on the ballot or the pre-existing status quo. For the typical referendum and from our previous definition, a voter is competent when she chooses the same alternative that she would choose given the best available information about the consequences of doing so. Therefore, all that the voter needs to do, regardless of how complex the referendum, is to figure out whether the policy described on the ballot is better or worse than the policy entailed by the status quo. Even if both alternatives are very technical, it is sufficient just to know the difference. Therefore, questions pertaining to voter competence in referendums can be restated as: can voters choose the better of two alternatives?

Evidence Answering questions about competence is problematic because competence is not easy to observe. For example, it is often the case that a decision-maker who seems competent to one observer (say, a Conservative) seems foolish to another (for example, a Liberal). There are, however, ways to determine the extent to which individuals or groups can emulate the behaviors they would have engaged in if they were better informed. If we believe that better-informed people make accurate predictions, then the ability of lesser-informed people to emulate their behavior bespeaks their competence as well. 3 In what follows, we describe research that is relevant to the question of voter competence in the referendum context. This research clarifies

196


the extent to which referendum voters are able to adapt to the information they lack. It shows that many people who look stupid when asked standard political survey questions are nonetheless competent when acting as referendum voters. We know that referendum voters usually fail to read the fine print on the ballot paper. As a result, we should not be surprised when they cannot recount the details of important ballot measures. For instance, in the days immediately following Canada’s intense 1992 referendum on the Charlottetown Accord, respondents in the post-referendum survey reported in Johnston et al. (1996) struggled to recall so much as one element in the package. Does this lack of attention to detail imply incompetence? It does not. To see why, first note that each of us makes thousands of decisions each day but consciously works through the details of only a few. If asked detailed questions about many of our decisions, we tend to find ourselves at a loss for words. However, we come to regret relatively few of the decisions we make. How can this be? The secret is information short cuts. We look for quick and effective cues about the likely consequences of our actions. Traffic lights, brand names, and personal reputations are all types of information short cuts that allow us to make complex decisions quickly and effectively. Indeed, this mode of reasoning is not something that we turn on and off, but a fundamental characteristic of how we live (Churchland and Sejnowski, 1992; Pinker, 1998). A consequence of this cognitive truth is that a voter’s ignorance of a referendum’s details is sufficient for incompetence only if no short cut allows her to make the same choice she would have made given the details in question. Of course, short cuts are no panacea. If used incorrectly, reliance on short cuts can lead to grave errors. Since research into human cognition has settled that people will use devices such as short cuts in contexts such as referendums, the key to understanding referendum voters’ competence is to determine how they use the short cuts available to them. In what follows, we show how a particular information short cut allowed less-informed voters to emulate the behavior of voters who were like them in all other respects save for the fact that they could answer more detailed political survey questions. Lupia (1994) reports on the result of an exit poll whose purpose was to determine the extent to which relatively uninformed voters could use information short cuts to cast the same votes they would have cast if better informed. The exit poll was of 339 California voters who were confronted by five distinct and complex propositions regarding the


regulation of the insurance industry. Three of the propositions were placed on the ballot by the insurance industry, while one each was put on the ballot by trial lawyers and consumer activists. Lupia’s pollsters asked respondents how they voted on the propositions, some socioeconomic and insurance rate related questions, and a series of questions designed to elicit their knowledge of the propositions. These latter questions were designed to learn not only what respondents knew about the content of the insurance reform debate but also to gauge respondent beliefs about the preferences of persons or groups who took a public stand on the initiatives during the campaign. Figure 9.1 depicts the study’s central finding, identifying a class of voters who, while possessing relatively low levels of information about the content of insurance reform initiatives, used an information short cut that allowed them to emulate the behavior of well-informed voters. It shows that relatively uninformed voters who could correctly identify the insurance industry’s official position on a particular proposition were more likely to emulate the behavior of relatively well-informed voters on that proposition than were similarly uninformed voters who did not know the insurance industry’s position. Figure 9.1 divides voters into three categories. In the first category are voters who knew neither the answers to the detailed questions about the propositions nor the insurance industry’s preferences. The second category contains ‘model citizens’ – voters who consistently gave correct answers to detailed questions about the insurance initiatives and knew the insurance industry’s preferences. The third category of the table contains respondents who, unlike the model citizens, could not consistently answer questions about the propositions’ details but, like the model citizens, knew the insurance industry’s preference. The conclusion that knowledge of the insurance industry’s preferences provided an effective short cut for voters who lacked encyclopedic information comes from a comparison of Figure 9.1’s three voter categories. Voters in the second and third categories look very much like each other, while voters in the first category (who were ignorant of the endorsements) look very different from the other voters. Voters in the second category voted as if they were ‘model citizens.’ Why should the insurance endorsement work in this way? The answer to this question is quite clear. First, the insurance industry spent over $65 million attempting to pass the three initiatives sponsored by its members and to defeat the other two. Second, the campaign made the insurance industry’s position on each issue widely known. Consumer activist Ralph Nader, in particular, made frequent public appearances.

198


Prop. 106

Low info, no endorsement

Prop. 104 High info, knows endorsement

Prop. 103

Prop. 101 Low info, knows endorsement Prop. 100

0

50

Prop. 100 Low info, no endorsement High info, knows endorsement Low info, knows endorsement Figure 9.1

100

Prop. 104

Prop. 106

26

12.5

45

Prop. 101 Prop. 103

27

15

53

8

72.5

17

11

53

5

73

17

12.5

Percentage of each group voting for the named initiative.

His efforts generated substantial media coverage of the fact that the insurance industry was pouring unprecedented amounts of money into the campaign. Third, most voters had prior interactions with the insurance industry and believed that their future insurance rates were at stake. In particular, the initiatives sponsored by insurance companies (Propositions 101, 104, and 106) would allow insurance companies to raise premiums and lower benefits, while the initiative sponsored by activists (103) restrained the industry’s ability to make these moves. As

Are Voters to Blame? 199 Table 9.1

Figure 9.1 voters in ‘low info, no short cut’ group

Proposition

%

100 101 103 104 106

13.8 15.5 11.4 10.9 20.9

a result, when many less-informed people learned the insurance industry’s preferences, they voted against them. While we cannot be certain that this was an optimal move for the less-informed voters, the fact that the best-informed voters acted the same way is suggestive of the prior group’s competence. This finding provides support for the claim that low-information voters used an information short cut to emulate the voting behavior they would have exhibited if they were as informed as the best-informed persons in the survey. Subsequent and more rigorous analyses in Lupia (1994) demonstrate that this result is robust to the introduction of a wide variety of contextual and socioeconomic factors. Not everyone made use of short cuts. Table 9.1 shows the percentage of the respondent pool who fell into the ‘low information, no short cut’ category. The size of this group ranges from a low of 11.4 percent for Proposition 103 to 20.9 percent for Proposition 106. As Figure 9.1 indicates, these voters tended to vote in very different ways from the rest of the sample. Consistent with a broader set of findings in Bowler and Donovan (1998), uninformed voters tend to vote no. The direction of the differences in each case is also consistent with messages in insurance industry campaign advertisements. Indeed, these figures may be evidence that their massive media campaign influenced this portion of the electorate. It is important to note, however, that despite outspending consumer activists by a ratio of over 30 : 1, only the activists’ initiative was victorious: it received 51.1 percent of the vote. The industry’s three initiatives, by contrast, failed decisively – Propositions 101, 104, and 106 received 13.3 percent, 25.4 percent, and 46.9 percent of the vote respectively. And in each of these four cases, the outcome of the election was the same as it would have been had voting privileges been extended to only the more informed subset of respondents. Is this example part of a general pattern, a spurious correlation, or a unique phenomenon? It is quite possible that voters who rely on short

200


cuts are regularly or easily misled, despite appearances to the contrary. Related research on binary decision making suggests that the exit poll’s results are an example of more general phenomena. In large campaigns, many claims are made and many opinions are available. No voter can attend to all of them; each voter must choose to whom and to what they will listen. Moreover, competitive campaigns produce contradicting claims. As a result, voters must make choices about which to believe and which to dismiss. To supplement our argument about the competence of referendum voters, we describe research that addresses two questions: • How do people choose among bits of the information available to them? • Do they choose in ways that enhance their competence? Lupia and McCubbins (1998, chapter 7) ran a series of experiments on how people choose whom to believe. Their experiments placed subjects in an environment that replicates important aspects of referendums – they were faced with a binary choice, they were initially given little information about the consequences of their choice. The key experimental variations were on the availability of short cuts and the conditions under which the short cuts were available. The experiments reveal people who are systematic in how they choose whom to believe and who make these choices in ways that improve their competence in referendum contexts. In the experiments, each subject was randomly assigned one of two roles. Some subjects played the role of a voter, while others played the role of an information provider (that is, endorser). The voter’s job was to choose one of two alternatives, while the endorser’s job was to advise the voter. Specifically, Lupia and McCubbins asked each voter to predict whether a tossed coin, hidden from their view, landed on heads or tails. At the same time they asked the endorser to advise the voter about the best choice. Lupia and McCubbins paid all subjects for their actions. They paid the voter a fixed amount, usually $1, for a correct prediction of the coin toss. They tested hypotheses about how people use short cuts by varying the endorser’s information and incentives. In some trials, the endorser also earned $1 when the voter made a correct prediction. In other cases, the endorser earned $1 when the voter made an incorrect prediction. In still other cases, Lupia and McCubbins made the voter uncertain about how the endorser earned money and, in particular,


whether the endorser gained or lost from the voter errors. Lupia and McCubbins also varied whether or not the endorser actually witnessed the coin toss as well as the voter’s beliefs about what the endorser knew. The experiment’s central hypothesis concerned conditions for persuasion (Lupia and McCubbins, 1998, chapter 3). They proved that for an endorser to persuade a voter in the experimental environment, the following conditions are individually necessary and collectively sufficient: • Absent incentive-altering external forces, persuasion requires perceived common interests and perceived endorser knowledge. • In the presence of such forces, these requirements can be reduced. In other words, with respect to persuasion, external forces can be substitutes for common interests (and for each other). Lupia and McCubbins use a game theoretic model to prove that voters in situations exemplified by the experiment should base their vote on the endorser’s advice if the voter perceives the speaker to have the relevant knowledge and an incentive to reveal that knowledge truthfully. In a context where there are no outside pressures on their communication act, these requirements translate into the voter’s perception that the speaker is knowledgeable and shares common interests. Otherwise, Lupia and McCubbins identify a list of incentive-altering external forces that offer alternative bases for one person to trust what another person says. The forces they describe include penalties for lying (for example, a loss of reputation for saying false things, penalties for perjury) and the threat that claims will be verified. These forces are present in culture, markets, and, most importantly for our argument here, the political contexts in which many referendums are held. These forces affect competence by altering what people choose to say and what people choose to believe. 4 Because the point about external forces is important to the question of voter competence, we take a moment to elaborate on it. How external forces such as penalties for lying affect competence in situations where voters rely on endorsers for advice should be familiar to any member of an advanced industrial economy. Every day, millions of people buy goods from, and sell goods to, people about whom they know little or nothing. Each of these transactions requires some degree of trust (for example, that the currency offered as payment is legitimate and that a good has its advertised characteristics). Since buyers and sellers do not know each other well, they must have an alternative and effective means for evaluating credibility. One such means is the external forces

202


that substitute for unobservable personal characteristics. For example, laws and customs realign strangers’ incentives making them transparent to others and giving people a basis for trust in billions of situations where it would not otherwise exist. These external forces are the substitutes that make advanced economies possible. Lupia and McCubbins (1998, chapters 7–9) show how political institutions supply analogous substitutes that make it possible for people to learn from others in modern democracies. Lupia and McCubbins’ experiments provide a consistent picture of how experimental subjects chose whom to believe. In over 2500 cases, their experimental subjects consistently and systematically chose to follow the advice of people whom they perceived to be knowledgeable about the consequences of their actions and for whom they had a reason to trust what they heard. Indeed, when subjects placed in experimental conditions in which Lupia and McCubbins predict competent performance, subjects predicted the coin toss correctly over 86 percent (1443/1669) of the time. These subjects neither believed everything they heard nor did they choose whom to believe at random. By contrast, when placed in experimental conditions where the speaker’s knowledge or interests could not be assessed accurately, subjects performed as expected – operating only at the rate of chance (1039/ 2057). People employ information short cuts for nearly every conscious decision they make. We collapse complex phenomena into simple categories that we can later use and process quickly. We do not take these actions randomly, but when we observe aspects of the environment that have systematic and similar properties, we convert them into short cuts, some of which are better known as brand names, reputations, or party ideologies. Short cuts are such a fundamental aspect of our being that it is not surprising that we take them for granted. We should be careful not to forget this fact when assessing voter competence. Indeed, improving voter competence will be better served by focusing the debate on voter competence away from whether or not citizens can answer survey designers’ political trivia questions and towards whether or not voters have sufficient information about the quality of the short cuts they will inevitably use. In sum, the theory and evidence presented above reveal that common stereotypes about voter competence rely on shaky foundations. If there are people who are willing to provide short cuts to voters and sufficient competition for voters to learn the motives or reliability of the short


cuts they receive, then voters can approximate the decisions they would have made if better informed. Of course, it may be true that polities encounter issues for which experts are themselves not well informed. In this case there are two possible scenarios. In the first scenario, no one in the polity is sufficiently well informed – including the experts on whom professional legislators rely. As a result, the informational advantages of a professional legislature relative to a referendum are non-existent. In the second scenario, the professional legislature’s experts are knowledgeable, but the electorate’s experts are not. In a closed society, where the channels of communication are centrally controlled or in a society where effective channels of communication do not exist, the second scenario is imaginable. In such a situation, it is unlikely that uninformed voters could overcome their problems and it is likely that putting the referendum in their hands could do damage. European and North American states, however, are not closed societies. Each has access to modern forms of communication and competitive political environments in which referendums occur. If someone has the opportunity to expose the opposing side’s weaknesses, the competitive nature of politics gives them a strong incentive to go public. In such cases, it is possible, but unlikely, that competing elites will conspire to withhold important information from potential supporters in the electorate.

Elite footprints: Bonapartist examples The distinctive characteristic of referendums, as opposed to most other forms of policy making, is the electorate’s direct participation. When pundits and scholars discuss referendums, they often focus on the voters’ role while underplaying or ignoring the role of elites. However, every public referendum is preceded by a series of elite machinations. Political elites, and not voters, determine whether or not to hold a referendum, when to hold it, and the wording of its question. They are also in a privileged position to influence interpretations of the event’s meaning, ex post. When pundits and scholars criticize the use of referendums, their preoccupation with the role of voters often leads them to blame voter incompetence for the more disturbing episodes in the history of mass democracy. But are voters to blame? In this section, we shift our focus from the mechanics of voter competence to a search for elite footprints in some of history’s most notorious experiences with referendums.

204


For our case study we turn to France and the fall of the First and Second Republics. The French Revolution yielded in less than fifteen years to Bonapartism. A mere four years were required to repeat that tragedy, in the rapid transition from the Second Republic to Second Empire. Referendums played a key role in each episode and give the impression of the people using referendums in a rush back to their chains. In what follows, we examine the validity of this impression. We find some truth to it, but even more fiction. It is not clear that the referendums were examples of mass incompetence, if by incompetence we mean that French voters chose what they should have known would be oligarchic and anti-liberal institutions over a reasonably justifiable status quo. Indeed, the evidence suggests, first, that opinion was commonly divided and that the collective outcome was arguably the superior one under the circumstances. The cases Three referendums marked the death of the French Revolution, and each contributed to the consolidation of the regime of Napoleon Bonaparte. The rise of Louis Napoleon’s Second Empire from the ashes of the Second Republic involved two referendums. As a result, the referendums of that era are remembered as effective tools for those who desire consolidation of executive power and as the enemy of those whose conception of the proper republic emphasizes parliamentarism.5 But did the French rush to their chains in these imperial episodes? The record in each case is remarkably ambiguous. It seems undeniable that most voters preferred consolidation of the executive to the alternative, and that both times the popularity of Napoleon the person or Bonapartism the idea was a factor. There can also be no denying that authoritarian rule ensued, each time combining secret police at home with adventurism abroad. But what were the circumstances under which the plebiscites were held? What can we say about how elite decisions about the structure, wording, and timing of the votes affected what voters chose or why they chose what they did? The first relevant referendum was the vote to approve the constitution proposed in the aftermath of the coup, the ‘18th Brumaire,’ that ended rule by the Directory.6 The vote proclaimed on 7 February 1800 involved male citizens signing the local register, for or against. The new constitution replaced the five-person Directory with a three-person consulate. The politically central fact was that the First Consul was the heroic general and Brumaire conspirator, Bonaparte. In Cobban’s sardonic words, ‘the people performed what was to become its customary


role in dictatorships in exemplary fashion: 3 011 007 voted for and 1 562 against.’ (Cobban, 1961, p. 113). The problem with leaving the account there is that, even accepting the official figures, the total vote represents less than half the eligible electorate. 7 And it is now clear that the official figures were massively falsified, that turnout was about half that reported (Lyons, 1975, p. 235). This hardly seems like a democratic people rushing back to the chains of empire. The margin for the consular constitution was still wide, of course, but undoubtedly wider than true sentiment in the country, as the non-secret ballot left dissidents exposed to public pressure and perhaps worse. The next referendum, in 1802, made Napoleon consul for life – the key vote in consolidating his personal rule. Voting once again was by open register and this time registers were open for three weeks. Turnout this time was genuinely higher than in 1800, again though undoubtedly partly as a reflection of social pressures. The one-sidedness of the result is almost certainly a misleading gauge of opinion. The last in this first series of major referendums, held in 1804, endorsed Napoleon as Emperor. Once again, the result was nearly unanimous. However, even official returns suggest dramatic drops in turnout, and the record again suggests considerable correlated fraud in the direction and turnout figures. Enthusiasm for the final step to absolute rule was clearly less than for the earlier, less obviously anti-republican steps. But if the margin was inflated, few doubt that executive consolidation around the person of Bonaparte was what most French voters wanted. So the French may have sauntered rather than ran. Were they, in fact, embracing chains? This seems like an exaggeration, fueled by Bonaparte’s own propaganda together with our subsequent knowledge of his actual regime and eventual megalomania. The vote in 1800 was arguably for energy in the executive, something with which the Hero of the Nile was amply endowed. It was probably a rejection of the Directory, but not as much as Bonapartist propaganda painted it. The 1800 constitution is not hard to portray in a republican light, and the coup it ratified was arguably no worse than the ‘self-coups’ (to use a late twentieth-century term) that the incumbent Directors themselves had perpetrated against their legislature. The Directory itself seemed to inspire no more than indifference, dogged as it was by economic hard times and military embarrassments. And the unsatisfactory Directorial constitution was almost impossible to amend within its own terms. The 1800 vote was followed by a display of, first, military energy and, then, successful peacemaking, even as Napoleon nudged his fellow Consuls

206


aside. The constitution left consular succession an open question, but then no earlier constitution had come up with a workable mechanism. In that context, the 1802 proposal, ‘consul for life,’ seems less sinister than it does with hindsight. The 1804 imperial vote is more mystifying, but with its low turnout, it is also the least credible as an indication of true opinion. Besides, the First Empire left much of the institutional and social legacy of the Revolution intact. 8 The second imperial episode of the mid-nineteenth century also involved referendums. It was quicker and truly more farcical that the first episode, but no less ambiguous. After much debate, the Second Republic’s constitution provided for direct election of a President. Although advocacy of direct election was not simply a Bonapartist ruse, the runaway victor was Louis Napoleon. The Republic lurched from crisis to crisis, until on 2 December 1852, Napoleon declared himself President for life and secured popular ratification for this later that month. Two years and another referendum later, he was Napoleon III, Emperor of the French. In one sense, these results speak clearly. Although balloting was still open and fraud was rampant, turnout was clearly higher than in the first Napoleonic episodes (Butler and Ranney, 1994, appendix A) and each time the yes vote seems a reasonable gauge of French opinion. But opinion on what? In the 1852 instance, it was not clear who was closer to the side of liberalism and democracy – Louis Napoleon, or his opponents. The assembly, dominated by royalists, was preoccupied with narrowing the franchise. The 1852 referendum may have signaled authoritarian consolidation but it also saw restoration of universal manhood suffrage. Campbell (1978, p. 10) argues that a coup was widely expected almost from the moment Louis Napoleon took office, and that the real question was whether he was tough enough to see it through. Oddly, he was the first Emperor anywhere to have campaigned for the job, as he published widely before 1848, associating Bonapartism with Saint-Simonian schemes to promote industrial growth. 9 It seems fair to conclude that Bonapartist schemes enjoyed real support at their outset and that some, if not all, of the relevant referendums correctly indicate the direction of opinion, if not the exact balance. Does this record then indicate incompetence, in the sense that voters would have chosen differently had they known the authoritarian consequences? The answer may be yes, but only if you allow evidence from more than a decade after the event. And if we admit evidence from still later, the picture actually muddies. In a post-ancien régime period of


intellectual confusion about how to constitute executive power, Bonapartism was arguably as plausible a program as any other. Even Victor Hugo, who made a celebrated break with Napoleon III, admitted a certain fondness for the first emperor (Cobban, 1961, p. 149). The Second Empire itself is now coming to be seen not so much as a temporary betrayal of France’s republican destiny but as a vital period in the development of civil society and a regime which deepened the idea of citizenship (Hazareesingh, 1998). This paper is not the place to take positions on such issues. It suffices to observe that on key questions embedded in Bonapartist authoritarianism, French elites were no more clear-headed than French voters. This fact clouds the claim that incompetent voters, rather than many competent voters in non-secret ballot elections together with maneuvering elites and some incompetents in both groups, were responsible for the events of the nineteenth century in France. Through the haze of fraud and propaganda, then, the following seem to be true. First, among the referendums, the more obviously authoritarian the choice the less one-sided its genuine support appears to have been. Voters sensed the magnitude of departure from the status quo and responded accordingly, by sitting on their hands. Second, the status quo was not the unequivocally superior alternative to Bonapartism, not even with hindsight. And binary choice, of course, is what French voters were asked to make, not the relative merit of a regime whose final trajectory was tragic versus some idealized regime combining democracy with energy. If these episodes are supposed to constitute the most disturbing evidence for voters’ incapacity, then the case seems at best not proved. Most of the mud, we submit, sticks to elite actors who, together with historical circumstance, framed the choices.

Conclusion Referendums are a story of elites who seek simplicity by attempting to narrow fundamental political issues to a single question, and of voters who seek simplicity by attempting to determine what the question means. Risk and error are associated with both simplifying moves. If we want to remedy problems with the referendum process, then we should understand where that risk and error lie. To that end, we have argued the following: • The informational requirement for competence in referendum contexts is minimized by the fact that most referendums entail binary choices.

208


• When short cuts are available, competence requires some information, but less than is commonly claimed. • Elites – and not voters – determine the wording, timing, and often the subsequent interpretation of referendums. As a result, they must be held at least partially responsible for the consequences of referendum usage. • Many of the perceived problems with referendums, both past and present, have less to do with voter competence than is commonly assumed; even if electorates contained only competent voters, it is likely many of history’s most notorious plebiscite outcomes would have nevertheless occurred. This view of the referendum process implies different solutions to some of its perceived problems. Suppose, for example, there are concerns about voter competence on certain types of question. A common remedy suggested for this is to give voters as much detailed information as possible. On the one hand, there is nothing wrong with this suggestion as it is better than intentionally withholding information. On the other hand, if a polity wants to use scarce resources to boost voter competence, it is better served by attempting to improve the quality of information rather than the quantity. For we know that voters use short cuts and that when they are available they will attend to them, leaving many details aside. As a result, a polity that wants to enhance competence should direct its efforts towards making it easier for voters to choose effective short cuts. This effort should include the following steps: • If well-known interests take learned positions on referendums, make it easier for voters to access these endorsements. • For position-taking interests that are less well known, make sure that they can be held accountable for what they say, perhaps by ensuring that both sides in a referendum battle have access to public airwaves. • If there is money involved in the promotion of, or opposition to, a referendum, make it easier for others to follow the money, for money trails are great places for learning about others’ true interests. While remedies to voter competence problems are easy, remedies to the troubles caused by elite maneuvers are more difficult. They are difficult because bringing a question to the people requires that it be stated in some fashion. While one could attempt to restrict the dominant role of elites by requiring them to offer multiple alternatives to voters (for


example, give them many choices instead of only two), the logistics of such an option makes it impossible. Not only are the informational requirements for a multiple choice more demanding than that of a binary choice, but declaring a winner is problematic (for example, the choices include a series of run-off elections – which is very inefficient – or a plurality counting rule – where an alternative can be declared a winner even though far less than a majority of voters chose it). As a result, if we feel that referring questions to the voters is an important democratic virtue, then we must accept the fact that elites will play a large role in deciding what the questions are. Having said this, we notice a tendency in the study of referendums to hold voters, rather than voters and elites, responsible for the outcomes of the process. This is a mistake. In both the study of referendums and the practice of a referendum campaign, it can be constructive to raise questions about elite decisions. It is in these decisions that much of the history of referendums has been determined.

Notes 1 Not every scholar accepts this characterization of the electorate. Some findings of voter ignorance seem to be the product of error in measures (Achen, 1975). Sniderman et al. (1991) usefully point out that the debate touched off by Achen turns as much on ontological assumptions as on statistical demonstration. Yet no scholar has made findings about the lack of factual knowledge go away. 2 Our defense of this competence definition follows from a similar argument by Lupia and McCubbins (2000) in defense of the concept of a reasoned choice. This definition of competence also has parallels to definitions of competent performance in cognitive science (see, for example, Newell, 1990). 3 This approach is increasingly common in the study of voter competence. Bowler and Donovan (1998) is a particularly instructive and carefully crafted example of such research. 4 The conditions they identify augment earlier insights such as those in Aristotle’s Rhetoric, by social psychologists of the 1950s (for example, Hovland, Janis, and Kelley, 1953), the philosopher Jürgen Habermas (1984) and by economic signaling theorists of the 1980s (for example, Crawford and Sobel, 1982; Sobel, 1985). 5 See the discussion in Morel (1996). Parliamentary rule was reinstated in 1870 by referendum, but no other question was referred in the Third Republic’s 70 years. Four referendums were required to launch the Fourth Republic, but once that Republic was launched no other referendum was conducted until the 1958 vote, which ended it. 6 This referendum also represented a procedural departure from earlier ones, which was conducted in local assemblies, and so combined some deliberation with face-to-face social pressure.

210


7 See Appendix A in Butler and Ranney (1994). 8 Our sense of the times owes much to the synthesis in Crook (1998). 9 Ironically so, for Napoleon’s initial political base was in the peasantry, exactly as Marx described it. See Campbell (1978), chapter 1, and Cobban (1961), pp. 146–7.

10 The News Media and Referendums Richard Jenkins and Matthew Mendelsohn

The way the news media cover referendums is crucial to a proper understanding of what we can expect from these exercises in democratic decision making. Yet the question of how the news media cover referendums has not been systematically considered.1 Proponents of referendums who idealize their deliberative possibilities may forget that political discourse will be mediated through the news values of media organizations, which are often inimical to collective deliberation. Opponents of referendums who are concerned about voter competence during referendums may forget how much useful information citizens can in fact pick up by watching the evening news. We begin by providing a brief review of how the news media cover elections. We then outline the theoretical differences between elections and referendums, and underline how these should matter for news coverage. Nonetheless, we hypothesize that news values and institutional imperatives will supersede the differences between elections and referendums, creating coverage of referendums that closely mirrors that of elections. We then examine news coverage of the 1995 Quebec referendum on secession and find that our suspicions are borne out. In the end, we conclude that media coverage of referendums looks much like that of elections, despite the fundamental differences between the two kinds of consultations. We suggest that this has normative implications for evaluations of referendum democracy: the way the news media cover referendums may help citizens in voting in a manner consistent with their pre-existing opinions, but it does not facilitate a process of democratic deliberation in which citizens can constructively challenge one another’s beliefs in a search for understanding. The news media therefore help encourage a kind of referendum democracy which amounts to an elaborate means of measuring public opinion, but not of collective will formation. 211

212

Richard Jenkins and Matthew Mendelsohn

The news media and election campaigns Elections are regular occurrences in representative democracies and in spite of important country-specific differences the news media tend to go about their job of mediating the election for voters in much the same way. This allows us to make a number of generalizations about the nature of news coverage during election campaigns that will be relevant to assessing the differences that direct democracy campaigns may produce. Two things stand out from a cross-national comparison of news coverage. First, parties and their leaders structure most news coverage of election campaigns. Second, the coverage of the campaign tends to subordinate substantive issue discussion in favor of strategy, the horserace, and the narrative description of events. Parties, leaders, and the organization of coverage Party leaders are the primary focus of the news media, and parties have adapted their campaign techniques to maximize leader visibility. The obvious manifestation is the leaders’ tour which features the party leader traveling across the country with an entourage of media personnel. The media are, of course, willing participants in the leaders’ tour since it offers the promise of access and visuals for their stories, and leaders’ speeches provide easy choices regarding the selection of soundbites. In proportional representation systems, less emphasis may be placed on the person at the top of the party list, with some countries, such as Switzerland, having lower profile leaders; but even some European countries that have traditionally placed less emphasis on leaders than the Anglo-American democracies have noticed a ‘presidentialization’ (or ‘Americanization’) of their campaigns (Mughan, 1995). The consequence of media and party practices is that there is little scope for non-party sources in the news coverage. For illustrative purposes, in the 1997 Canadian election 87 percent of the television news stories focused on one of the large parties, and of those stories, 72 percent focused on the leader.2 One could imagine a news coverage that featured other party members, interest groups, minor parties, and expert perspectives on the issues of the campaign, but there appears to be little of this in the mainstream coverage of elections. The importance of parties extends to the key decisions about how much attention a party and its leader deserve in the news coverage. To ensure balance, the news media in many countries monitor the amount of coverage that is given to each party or are mandated by law to allocate coverage in a fair manner (Semetko et al., 1991). Even where monitoring


213

or legal requirements are not in place, we should expect balance to emerge because a party’s position in the race will define its importance for journalists. Empirically this has been demonstrated in the Canadian context as new parties’ share of television news coverage in the 1993 federal election was closely related to their share of vote intentions (as measured by surveys) as the campaign evolved (Jenkins, 1999a). There is little space for non-party voices in election coverage. Party leaders are the key foci for news personnel and the definition of newsworthiness privileges an emphasis on party activity. Election coverage thus reflects and reinforces a view of politics as a competition between party elites rather than a collective exercise in democratic judgment. The creation of news stories Patterson has observed that ‘an understanding of the news media’s version of reality begins with a recognition that reporters must have a story to tell’ (Patterson, 1993, p. 96). Reporting the news requires the selection of a narrative or interpretative framework. Since the outcome of an election campaign is the formation of a government, it is natural that most reporting of elections uses a metanarrative that places each day’s ‘news’ in the context of who is likely to be successful on election day. Thus one of the most common observations about, and criticisms of, election coverage is that it is dominated by stories about who is ahead and why rather than stories about the substantive stakes of the election (Robinson and Sheehan, 1983; Brady and Johnston, 1987). Considerable attention is devoted to the underlying strategies and motivations of the parties, and their comparative standing with the public. An additional feature of the coverage is an emphasis on leadership and character. This emphasis is consistent with the media’s tendency to attribute outcomes to individuals (Ansolabehere and Iyengar, 1991) and the decline of political parties (Wattenberg, 1990). Coverage of election campaigns tends be narrative and episodic; that is, there is an emphasis on describing the day’s activities and episodes as discrete events. The campaign is therefore framed as an event to be covered, rather than as a collective debate. Despite the focus on leaders, the length of soundbites, and therefore the length of time politicians are able to discuss the campaign in their own words, has declined. Hallin characterizes this evolution as a shift toward a ‘more mediated, journalistic-centered form of journalism’ (Hallin, 1992, p. 5). This change produces a press that is more active in interpreting the campaign and more adversarial toward politicians and the political process (Lichter and Noyes, 1995; Lichter and Smith, 1996).

214


Finally, citizens are notoriously absent from election coverage. Although they may provide a visual backdrop for a leader’s speech, or play the role of extras as a leader shakes hands in a public square, their voices tend not to be heard. Elections may be about democratic accountability and the sovereignty of the public, yet members of the public are not factored into the dominant narratives of election coverage. This is not to imply that the news media ignore election issues. Rather, we wish to highlight two theoretical points. First, the coverage of issues is framed in terms of leadership, tactical motivations, and the horserace. When attention is directed toward issues, the arguments of political actors tend to be interpreted through a strategic frame – the political tactics that may lie behind a given political direction rather than a programmatic one – the principles that lie behind a given program of action and the possible consequences of such a program. Second, the goal of journalists is to identify where parties stand on issues, not to engage in substantive debate about the merits of contending courses of action. The consequence of such reporting is that it may cue voters as to where parties stand on issues (Jenkins, 1999b) and enable voters to compare parties’ positions with their own, but will rarely cause them to reconsider their positions or persuade them to change their minds. Such a form of coverage may permit some issue voting and prime certain issues over others, but does not facilitate debate and deliberation. This second feature is in part a consequence of the objectivistic nature of news coverage. Although some states continue to have newspapers that are allied with particular parties, television news (and most newspapers in most states) is non-partisan. Accordingly, the news media have adopted practices that ensure the appearance of neutrality and objectivity (Hackett and Zhao, 1998). Reporting on what is said by political leaders is much easier to justify in terms of objectivistic coverage because it relates to indisputable facts. When offering more elaborate assessments of parties’ records, their ideological motivations for pursuing particular programs, the groups to which they are allied, and the viability of their promises, one leaves oneself much more open to accusations of bias. Furthermore, providing contextual information of this type, which voters need in order for them to situate the facts of the election, may be beyond the role of the news media which has a functional need to ‘tell stories’ within the available limits of time and space. The provision of contextual information may properly be the role of the family, the professional association, the peer group, the union, the interest group, and the church.


215

Table 10.1 Substantive differences between elections and referendums relevant to the media

Objective Key actor Voters’ background Democratic responsibility Required information Usual media practice

Elections

Referendums

Choosing governments Parties Established loyalties Horserace/information Cue giving Established guidelines

Making policy Committees/anybody None/unknown Information Cue giving/deliberative Open

Summary A highly stylized characterization of news coverage of elections is that parties structure the coverage, and that substantive issues are only important to the extent to which they assist in the construction of stories about the electoral contest. One could conclude that the reporter during an election fulfills four complementary roles: theater critic (offering evaluation of the performance of leaders), campaign strategist (identifying the tactics of parties and explaining why they may or may not be successful) ( Jamieson, 1992), sports reporter (reporting on the winners and losers of the day and the current standings), and war correspondent (narrating the key battles and activities in adversarial terms). How relevant are these observations about the coverage of elections to the question of how the media cover referendum campaigns? The answer to this question hinges on two things. First, are referendums and elections different in key respects that could matter theoretically for news coverage? And second, if they are different in important ways, are the news media capable of and interested in covering referendums in some way other than the way they cover elections? The following section looks at the first question.

Referendums, elections, and news coverage Both referendums and elections feature campaigns led by high-profile individuals and mass electorates casting ballots. Other than these qualities, referendums and elections have little in common substantively. These differences are summarized in Table 10.1. Referendums are about making policy rather than choosing governments and many of the differences follow from this overarching distinction. Unlike elections, referendums involve more and different actors that extend beyond party sources. Voters may have less familiarity with these actors and

216


with the issue facing them and may therefore require different kinds of information in order to make an informed choice. Finally, the news media have adopted a set of institutional norms associated with news coverage of elections, which change only modestly from election to election. Media practices during referendums, on the other hand, are less established due to the relative infrequency of referendum campaigns in most states. Even if citizens are essential participants in the election outcome, an election campaign is by definition a competition between parties for power and the focus of media coverage, almost inevitably, turns toward parties and their leaders. During a referendum campaign, the spotlight could more easily turn towards the public. As documented elsewhere (Lupia and Johnston, this volume; Jenssen and Listhaug, this volume), elites are central to referendums, yet the imagery surrounding referendums and the discourse used to discuss them privilege voters. One might expect the news media to internalize the discourse about popular control and pay more attention to voters than they would during an election campaign. At a minimum, this would mean an expansion of voices beyond party leaders to include the leaders of key groups and associations, such as business and labor. More expansive change is also possible, moving beyond well-known public personalities and leaders to include a wide range of civil society actors not normally heard in the media, and offering them an opportunity to participate in the collective conversation. Referendum campaigns can often be far less structured than elections. For most referendums in most countries there are legally recognized Yes and No committees which serve to structure both campaign spending and argument. These committees, however, lack the institutional foundation of political parties and do not prevent interventions by ad hoc movements, interest groups, and citizens. Such groups can often support or oppose the referendum issue for reasons very different than those advocated by the leaders of the committees; if a variety of different groups are offering a variety of different reasons for support or opposition, one could imagine a far less structured debate. Although such a discursive, open-ended debate could offer the media far more choice, it may provide the news media with a challenge beyond their capabilities. The news media is faced with more difficult choices during a referendum campaign. During an election, ‘balance’ is fairly easy to define: all major parties receive nearly equal coverage, with such a requirement stipulated by law in some countries (although such a definition does leave open the question of how much coverage to afford smaller parties).


217

During a referendum campaign, ‘balance’ becomes far trickier. One could offer coverage almost exclusively to the leaders of the two committees (if such organizations exist), but such an attempt at balance may fail to reflect the true nature of the debate because alternative voices who may be equally active or important would be ignored. The imposition by the media of one clear narrative – that is, one well articulated definition of what the vote is really about – and one easily understood definition of balance is particularly difficult when the campaign pits a large centrist coalition of established elites in opposition to a mixed bag alliance of those on the left and right. This describes quite well the coalitions in the French and Danish referendums on Maastricht in 1992. In the Australian referendum of 1999 on a new republican head of state, the debate featured a compromise middle position being attacked by both those who thought it went too far and those who felt it did not go far enough. ‘Balance’ may be difficult to operationalize in such less traditionally structured contests because one easily grasped narrative that identifies ‘what the campaign is about’ may be absent. Does one offer leaders of ad hoc organizations the same nightly access to the media as the prime minister? Does one offer both the small left and the small right parties the same amount of coverage as the large centrist parties? A variety of different plausible operationalizations of balance are hence possible during a referendum campaign. Since both referendum and election voting require citizens to make informed choices about the best alternative given their preferences and values, we might expect that the information needs of citizens would be the same during referendums and elections. And we know that citizens are able to make reasonable choices during elections, even though they fall short of the ideal of citizen competence (Sniderman et al., 1991) and even though the media do not contribute to informed debate (Patterson, 1993). Are there reasons to believe that referendums will be different, and that, despite the fact that voters are able to make do with sporadic attention, incomplete information, and strategically oriented media coverage during election campaigns, they will be unable to do so during referendum voting? Elections usually feature familiar parties, most voters have standing attitudes towards parties, and there are additional short-hand cues – such as the state of the economy – that many voters use to pass judgment on governments. Furthermore, theories of issue voting suggest that election campaigns are opportunities for voters to identify where parties stand on issues and vote for the party that comes closest to sharing their views, or, more likely, the party that proposes to move away from the

218


status quo in a direction consistent with their own positions, but are not occasions when voters listen to debate between parties and re-evaluate their own views. These factors imply that for many voters, exposure to extensive and substantive campaign information during elections is not a necessary precondition for decision making. The news media may be ignored or, for many voters, followed out of interest to see how one’s preferred party is performing and how the dynamics of the campaign are shaping up. Despite the common criticism that campaign news coverage is superficial, it may fulfill its democratic function by providing the minimal informational cues that voters require. Referendums ask voters to offer a verdict on a particular question, not choose between parties or select a government. Because of this, voters may in fact need more information during a referendum campaign than during an election, and may need information framed in a different way. The information needs will depend on the degree to which the issue has been the subject of sustained debate, whether voters have standing opinions on the issue, the complexity of the issue, and the nature of elite disagreement on the issue. For example, even if one has a strong standing opinion on a general issue – say, the liberalization of trade or the European Union – one is unlikely to have a standing opinion or be highly familiar with the contents of the particular proposal up for a vote (say, the Maastricht Treaty or Edinburgh Agreement). In general, there is a greater need in most referendum campaigns for more substantive information than required during most election campaigns. Furthermore, the provision of this information may need to be designed to facilitate decision making by voters in a deliberative context. One might therefore expect the news media to recognize this and feel compelled to fulfill what they perceive as their pedagogic responsibility in a liberal-democratic society by presenting information framed in terms of competing interests and values rather than strategies. In suggesting that information needs may be higher in referendum contests, there is a danger that we overestimate the needs of the average citizen. Lupia (1992 and 1994) has provided empirical support for the proposition that citizens are able to use cues about where key intervenors (groups and prominent individuals) are located to determine whether they prefer the status quo to the referendum proposal given their own beliefs about the issue. While this research confirms that voters do not need encyclopedic information about even complex ballot propositions, some qualifications should be made. First, propositions examined by Lupia (1994) involved considerable advertising expenditures which may have ensured a greater awareness of the group-based cues that are


219

essential for short-cut based reasoning. Lacking such advertising, it may be necessary for the news media to provide coverage of a wider range of intervenors in order to communicate this kind of information to voters. Second, while Lupia demonstrates that those voters who can identify where intervenors stand can mimic the behavior of the well-informed, a large percentage of voters fail even to be capable of identifying the positions of intervenors, and consequently vote in ways that ‘do not make sense.’ During election campaigns, such poorly informed voters may have access to cues like party identification or the state of the economy, whereas they may be at a complete loss during referendums. In close referendum votes, such voters could make the difference in the result. Most crucially, however, referendum campaigns can take many different forms. Bowler et al. (1998) propose an original and intuitively appealing typology of campaigns that allows one to distinguish among a variety of different referendums according to whether the proponent and opponent are narrow and organized or diffuse. In most cases, the referendums examined by Lupia (see Lupia and Johnston, this volume) feature narrow, well-financed groups taking on other narrow, wellfinanced groups (for example, the insurance industry vs. trial lawyers). Such contests often feature highly technical regulatory legislation, and public attitudes toward the referendum proponents and opponents may be particularly useful in such contests. Such intervenor information (‘do I trust the insurance industry?’) may be less useful during referendums where diffuse constituencies battle other diffuse constituencies, such as occurred during the Quebec referendum on secession. In such contests, one can often find representatives of business, unions, and other prominent groups on both sides of the issue. The competitive interest group referendums examined by Lupia are also those where deliberation may be less important than opinion measurement. According to Lupia, referendums succeed if citizens do not vote in ways that overtly violate their pre-existing beliefs. Such a standard may be appropriate for the referendums examined by Lupia, but may be an insufficient standard when examining referendums dealing with broad constitutional or economic change featuring heated debate and involving a complicated map of issue dimensions and a variety of cross-cutting cleavages. Summary In elections, the focus is invariably on parties, voters have well-established feelings towards the principal actors, the electorate must choose between parties, the media need only provide cues to electors to activate their feelings towards parties, and the media have a responsibility to tell

220


the story of the horserace. This type of coverage has developed over decades and now is part of a well-ensconced, institutional pattern of news coverage. None of these things necessarily apply to referendum coverage. During a referendum the media may have far more power to impose their own narratives and interpretations of the substantive issue. One could therefore expect qualitatively different coverage, including a focus on different intervenors (that is, a focus on non-party actors), the provision of more information, and the interpretation of this information through a more deliberative frame. Yet whether differences between referendums and elections produce a qualitatively different kind of coverage hinges on whether the media recognize these differences and, presuming that they do, whether such a recognition would be important enough to overcome the forces which work to produce the kind of electoral coverage that is apparent in most liberal democracies. In the next section, we show that the forces and conventions shaping coverage are more powerful than any perceived differences between the two kinds of popular consultations and, therefore, there is good reason to expect referendum coverage usually to look much like election coverage.

Empirical evidence concerning the news media and referendums The discussion of the possibilities for appropriate referendum coverage directs us to focus on two aspects of the news coverage. The first is the balance of voices that get represented in the news. Do elites dominate coverage, and if so, how diverse is the range of elites? The second is the way the stories are told. If the news is told from a strategic framework, voters are unlikely to be provided with the means to engage in a substantive discourse themselves. Here, a content analysis of the news coverage during the 1995 Quebec sovereignty referendum is used, along with findings from other studies, to address these questions. The 1995 referendum campaign was the second referendum on the question of the secession of Quebec from Canada. In both cases, the referendum question was the initiative of the governing Parti Québécois (PQ) and opposed by the Quebec Liberal party and federal politicians. Quebec’s referendum legislation provides for umbrella committees to run the campaign and have a virtual monopoly on campaign spending, with the leadership of these committees being in the hands of the two provincial party leaders (the Premier acting as leader of the Yes committee, and the Leader of the Opposition as leader of the No committee). Because of the highly structured nature of the umbrella committees,


221

there is good reason to believe that these politicians will receive most of the coverage, along with the leaders of other parties at both the provincial and federal level. Yet there was a diversity of groups and voices which intervened in the debate and had very different positions from those of the umbrella committees. Some on the sovereigntist side felt that the official PQ position was insufficiently sovereigntist since it called for a continued partnership with the rest of Canada. But the real disagreements came on the federalist side, where some groups – particularly English-speaking Quebeckers – argued that the leadership of the No committee was insufficiently committed to Canada and challenged the legitimacy of the umbrella committee system. Aboriginal peoples also urged a No vote, but for reasons very different from the official No committee, and towards the end of the campaign Canadians from outside Quebec attempted to participate in the debate without going through the formal committees. How the media frame a referendum in such circumstances – highly structured umbrella committees coupled with vibrant alternative voices – should tell us something about the kind of information available to voters during a referendum campaign. For each day of the official campaign period (1 to 30 October) the referendum coverage on Quebec’s most popular television news programs (Radio-Canada and TVA)3 and Quebec’s largest circulation broadsheet (La Presse) was analyzed. Each story was identified in terms of its placement, the intervenors who were referred to in the story, the major focus of the news item, and the frame through which the story was told. Television news devoted considerable attention to the referendum campaign: 190 stories were broadcast during the 30 days of the campaign, with half these stories appearing in the first part (first four items) of the newscast. Given that newspapers are only constrained by advertising space and news value rather than the length of the news broadcast, there were more news items (400) in La Presse. Eleven percent of these stories appeared on the front page, 12.5 percent in the remaining pages of the first section, and 76.5 percent in other sections of the paper, particularly the National News second section. There is no apparent evolution in the distribution of stories, either in terms of the number of items per day or the placement of stories in the newscast, as the campaign neared the finish. Balance of voices Comparative evidence suggests that operationalizing ‘balance’ between the two sides of a referendum question is problematic, especially when there is an imbalance of elite or public views on the question. In both

222


the 1975 (Common Market) and 1979 (Scottish devolution) British referendums, most of the partisan players were lined up on one side. The British broadcast news media, however, adopted a 50–50 rule for allocating coverage (Fowler, 1981). In contrast, the media in some countries covered the sides of the Common Market referendum in proportion to party standing, an approach which produced an imbalance (Smith, 1976). In the 1995 Quebec referendum, a 50–50 rule seems most likely to emerge given the distribution of opinion in Quebec before the campaign and ultimately the closeness of the race. Although there is a slight edge for the sovereigntist side in our content analysis, the data confirm that the federalist and sovereigntist politicians were equally likely to be the major focus of the news stories both on television and in newspapers (Table 10.2). While we lack data on the position of each intervenor, there is no reason to believe that one side was advantaged in the amount of coverage each side received. In a careful study, Monière and Guay (1996, chapter 4) found that the French-language media treated both sides in the 1995 referendum virtually equally. Perhaps more important than the balance of attention to the two sides is the evidence concerning who got covered for each side. In other campaigns, it has been argued that politicians are the primary focus for stories in referendum campaigns (Seymour-Ure, 1976; Fowler, 1981; Robinson, 1998; see also discussion in Jenssen and Listhaug, this volume). As one analyst observed concerning the 1979 referendum on Scottish devolution: ‘The politicians had once again held on to their monopolistic position as the sole authorised diviners of political issues’ (Fowler, 1981, p. 135). But in the Quebec case, a variety of non-party groups had a stake and position on the question, with business, labor, and social movements and interest groups of various kinds actively and intensely involved in the debate. Despite this, there is no question that politicians dominated news coverage. Tables 10.2 and 10.3 shows that politicians were the most frequent type of intervenor in news stories. Of those stories which had a person or group as a major focus, 70 percent of the intervenors in television news and 53 percent of intervenors in newspaper stories were politicians. Beyond these politicians officially affiliated with one of the two camps, only ‘personalities’ – which includes some retired politicians and politicians from other provinces – and business groups were frequently the major focus of a news story. Newspaper coverage was more diversified in its use of intervenors in its story constructions. Experts, citizens, community groups, Aboriginal peoples, and students are all more prevalent in newspaper stories compared with television.

The News Media and Referendums Table 10.2

Distribution of intervenors as major focus of news stories Television frequency

Active affiliated politicians Jacques Parizeau Jean Charest Jean Chrétien Lucien Bouchard Daniel Johnson Mario Dumont Other party politicians, federalist Other party politicians, sovereigntist

Other perspectives Personality Artist Experts Citizens

%

Newspaper frequency

%

18 6 9 19 17 3 15

12.6 4.2 6.3 13.3 11.9 2.1 10.5

28 11 12 33 21 6 35

8.6 3.4 3.7 10.1 6.4 1.8 10.7

13

9.1

27

8.3

(100) Interest organizations/ representatives Business Labor Agriculture Fédération des Femmes du Québec Alliance Québec Professional organization Community groups Aboriginal peoples Student

Total

223

(70)

(173)

10 1 – 1

7.0 0.7 – 0.7

1 – 1 2 1 (17)

0.7 – 0.7 1.4 0.7 (11.9)

– 1 11 5 3 (48)

– 0.3 3.4 1.5 0.9 (14.6)

16 4 3 3 (26)

11.2 2.8 2.1 2.1 (18.2)

51 10 28 16 (105)

15.6 3.1 8.6 4.9 (32.2)

143

100

20 5 1 2

(53)

326

6.1 1.5 0.3 0.6

100

a

Multiple foci were allowed. Not all stories had a major focus. Approximately 54 percent had a single major focus, while 20 percent had two. b Numbers in italic represent totals for the subcategories.

When citizens were present, they were more likely to be the minor focus of the story. But only 30 percent of the stories had a minor focus, indicating that not much attention was paid to the public in the news. More importantly, citizens appeared in stories which were primarily about campaign events rather than substantive issues. As Robinson

224


Table 10.3

Distribution of intervenors as minor focus of news stories Television frequency

Active affiliated politicians Jacques Parizeau Jean Charest Jean Chrétien Lucien Bouchard Daniel Johnson Mario Dumont Other party politicians, federalist Other party politicians, sovereigntist

Other perspectives Personality Artist Experts Citizens Total

Newspaper frequency

%

17 7 11 14 14 5 21

8.0 3.3 5.2 6.6 6.6 2.4 9.9

15 7 8 14 15 6 57

3.0 1.4 1.6 2.8 3.0 1.2 11.6

18

8.5

47

9.5

(169)

(34.1)

(107) Interest organization/ representatives Business Labor Agriculture Fédération des Femmes du Québec Société St-Jean-Baptiste Alliance Québec Community groups Aboriginal peoples Student

%

(50.5)

19 3 – 2

9.0 1.4 – 0.9

24 7 2 2

4.9 1.4 0.4 0.4

– – – – 3 (27)

– – – – 1.4 (12.7)

1 1 22 7 24 (90)

0.2 0.2 4.5 1.4 4.9 (18.3)

22 5 16 35 (78)

10.4 2.4 7.5 16.5 (36.8)

75 13 53 93 (234)

15.2 2.6 10.8 18.9 (47.5)

212

100

493

100

a

Multiple foci were allowed. Not all stories had a minor focus. For newspaper coverage only 30 percent of the stories had a minor focus and up to seven intervenors were noted in those stories with a minor focus. b Numbers in italic represent totals for the subcategories.

(1998) notes about the 1980 referendum experience, the way in which the media covered the campaign cast the public in a passive position. The low level of attention devoted to the business community, and the almost complete invisibility of labor, is striking. Although unions


225

tended to side with the sovereigntist side and the business community with the federalist side, they were by no means monolithic. This created opportunities for rich news coverage of the competing positions of various factions, but also made the traditional frame for news stories – for example, business vs. labor – difficult to use if one wanted to accurately reflect the complexity of the debate and the nuances of the various actors’ positions. Such nuances took a back seat to more traditional narratives dealing with the competition between several high-profile politicians. Community groups, interest groups, and social movements representing organizations of citizens brought together for collective action were relatively absent from news coverage, especially television. This absence extends to a number of groups that are significant for understanding Quebec politics. These include Alliance Québec, the Société St-JeanBaptiste, and the Fédération des Femmes du Québec. Some of these organizations’ positions on the question of sovereignty are well-known, one being an English-rights group and one being an important nationalist organization, but others, such as the women’s group, are not clearly associated in the public mind with one side or the other, and their absence from coverage may have deprived voters of important cues. Even Aboriginal peoples, who have an important stake in the outcome of the question, and who organized their own referendum in the northern part of Quebec to highlight the fact that their concerns were not adequately considered by the Quebec government, were a very minor focus of attention. Because low-information voters are generally the least attentive to the news media during referendum campaigns (Borg and Esaiasson, 1998), they are unlikely to get cues from the media concerning the positions of prominent intervenors (Johnston et al., 1996; Mendelsohn and Cutler, 2000). The nature of coverage reported above suggests that cues concerning groups’ positions may not have been picked up by those voters who needed them most during the 1995 referendum. Types of referendum stories It is generally acknowledged that journalists tend to construct campaign stories on the basis of a few common frames and that these frames can have political consequences because they influence what considerations the audience uses when evaluating political objects (Iyengar, 1991). Comparative evidence suggests that the news media narrate referendum campaign activities through a competitive, electoralist schema. Looking at the Scottish referendum on devolution in 1979, Brown found that

226


despite a great deal of attention devoted to the issue, newspapers did not cover devolution and the referendum ‘in their own right,’ but only when they could link them to campaign events, personalities, or conflicts (Brown, 1981, p. 111). Robinson’s (1998) study of the 1980 Quebec referendum lends support to the idea that the referendum campaign was treated like an election campaign. Not only did the parties and leaders dominate, as they would in an election, but the way the stories were told conformed to narrative styles drawn from election coverage. ‘Within the newscast, each subevent (for example, the poll, the leader interviews) acquired its meaning from the state of the referendum “game” as a whole’ (Robinson, 1998, p. 95). The state of the contest between the Yes and No sides was used to weave disparate events and interventions together into a single story. The key, of course, is that this story lacks the capacity to engage voters in a conversation about what is at stake. The election frame downplayed the significance of the referendum and turned it into normal politics rather than a debate about Quebec’s future (Robinson, 1998). We defined four main categories of story focus for analytical purposes: substantive issues, cues about intervenor support, leadership, and campaign stories (Table 10.4). Approximately 73 percent of television and 62 percent of newspaper stories were primarily about the campaign or leadership.4 It appears that the referendum was treated like an election. A small share of the coverage was primarily about substantive issues of which the economic cost of secession and post-sovereignty institutions were most important. In addition, there was, especially in the newspaper, a significant number of stories which conveyed information about where a variety of groups stood. This information could be used by voters to make vote decisions, but is less likely to assist with deliberation. In fact, voters do learn during a referendum campaign. The key is that information gains are likely to be the product of citizens paying more attention to the campaign, not due to the fact that the media provided a great deal of substantive information, and that such gains are likely to come in terms of the ability to identify intervenors’ positions (Mendelsohn and Cutler, 2000). For each story an attempt was made to identify the overarching frame through which it was told. In essence, identifying the frame is a way of identifying what was important to the media about the story. Before the coding began, six possible frames were conceptualized. The first possibility, substantive/programmatic frame, identifies those stories that provide analysis of an issue in terms of its substantive impact or in terms of principled debate or discussion. A second possibility, captured by the

The News Media and Referendums Table 10.4

Focus of stories TV

Substantive issue Constitution Economic cost of secession Post-secession institutions/ negotiations/‘partnership’ Current economic situation Cost of a No Quebec’s power as a minority/majority Language Costs of federalism Lack of recognition of Quebec’s identity Cues about group or individual support Leadership Gaffe Popularity Adversary critique Dissatisfaction/support/ conflict between leaders Character Performance Campaign Polls Rallies State of the campaign Campaign strategy Campaign events/activities Total

227

Newspaper

All

N

%

N

%

N

%

2 10 17

1.1 5.3 8.9

10 28 17

2.5 7.0 4.3

12 38 34

2.0 6.4 5.8

5 1 1

2.6 0.5 0.5

9 8 5

2.3 2.0 1.3

14 9 6

2.4 1.5 1.0

– – –

– – –

1 3 1

0.3 0.8 0.3

1 3 1

0.2 0.5 0.2

(36)

(18.9)

(82)

(20.5)

16 (16)

8.4 (8.4)

70 (70)

17.5 (17.5)

118 86 (86)

(20.0) 14.6 (14.6)

5 3 9 4

2.6 1.6 4.7 2.1

8 6 34 11

2.0 1.5 8.5 2.8

13 9 43 15

2.2 1.5 7.3 2.5

1 1 (23)

0.5 0.5 (12.1)

7 9 (75)

1.8 2.3 (18.8)

8 10 (98)

1.4 1.7 (16.6)

10 13 29 18 45 (115)

5.3 6.8 15.3 9.5 23.7 (60.5)

16 48 21 23 65 (173)

4.0 12.0 5.3 5.8 16.3 (43.3)

26 61 50 41 110 (288)

4.4 10.3 8.5 6.9 18.6 (48.8)

190

100

400

100

590

100.0

competing interests frame, identifies those stories which focus on how different groups share different interests. Stories of this kind would provide voters with the group cues that would allow them to identify who stands where. A third possibility, competing parties frame, identifies those stories framed in terms of partisan goals and interests. The last three

228


Table 10.5

News frames TV

Substantive/programmatic Competing interests Competing parties Strategies of actors in referendum Consequences of story for campaign Factual report of an event Total

Newspaper

All

N

%

N

%

N

%

5 13 18 45

2.6 6.8 9.5 23.7

6 76 33 107

1.5 19.0 8.3 26.8

11 89 51 152

2.2 18.1 10.4 31.0

11

5.8

22

5.5

33

6.7

98

51.0

156

39.0

255

43.1

190

100.0

400

100.0

491

100.0

frames – strategies, consequences on the campaign and campaign events – relate to those stories where the campaign as strategic contest and spectacle are what is considered important. Campaign-based frames clearly dominated the news. Over half of the stories on television and 39 percent of the stories in the newspapers were factual reports of campaign events (Table 10.5). This conforms to the news-as-events or news-as-drama requirements of television. Another quarter of the stories were framed in terms of the strategies of the referendum actors. Overall, about 80 percent of television stories were framed by one of the three campaign-related frames, with little attention to the more substantive frames. Newspaper stories were, however, more likely than television to be framed in terms of competing interests. This frame is the most likely to generate public understanding of where groups stand and why. This finding is consistent with the earlier conclusion that newspaper stories were more likely to cue voters about where different groups stood on the question. Still, about 70 percent of all newspaper items were framed by the three campaign categories. The news media approach referendums in the same manner that they approach elections. Together the focus of the stories and the frame of the stories demonstrate that the news is about the campaign process and strategy rather than the substantive issues that were at the heart of the referendum question. Although newspaper coverage was more substantive and more likely to convey some cues about the positions of groups, neither medium framed issues in a manner that would promote deliberation.


229

Conclusion Despite significant differences between elections and referendums (Table 10.1), it turns out that they produce quite similar kinds of news coverage. This does not mean that the news media are compelled to cover public events in this manner, though the fact that the news media end up covering referendums like elections does underline how immutable news conventions have become. For the most part, news was about strategy and campaign events. There was little opportunity for deliberation to emerge from or be stimulated by the news coverage. This does not mean that deliberation did not occur among many voters, only that the news media did not facilitate it. At most, the news media may provide cues to voters about where some prominent intervenors stand, though even on that score the media may fall somewhat short. Newspaper coverage provided a somewhat wider range of cues about intervenors’ positions, focusing on some civil society groups. There is no evidence that the media can go beyond simply cueing voters as to where groups stand and providing a forum and a context for collective debate. The imagery of referendums, rooted in popular sovereignty, is that of decision-making power turned over to citizens. The reality, however, is that referendums are launched by particular elites with particular agendas who have an interest in presenting material to reach a given end, not in sacrificing power and control to the public. The media continue to turn to elites for the provision of information and interpretation of the issue, which will almost inevitably frustrate voters who may already feel they have no control over the process. By adopting election coverage rules as the standard of news presentation for referendums, the news media undermine the deliberative potential of referendums. The analysis presented here reveals the contradiction between participation and control, which is implied in the discourse about referendums, and spectatorship, which is the role that citizens are cast in by media practices. In order for referendums to function as devices of popular sovereignty, the public must be active participants in the discourse. While the data presented here may be further evidence that the news media, which is designed to tell stories in compelling ways, is incapable of fulfilling its democratic function (Patterson, 1993), this may be too pessimistic. The news media does not have to undergo a radical transformation and attempt to fulfill an alien pedagogic function. At minimum the news media could expand the range of actors and perspectives to voters. It may also be appropriate to rethink the use of the election

230


framework – the treatment of the referendum as a competitive strategic struggle between elites – for referendum coverage. Our research has focused on one referendum, while other referendum research on the media has emerged largely from other Anglo-American democracies. Whether such findings can be generalized to other countries, particularly those of continental European countries with different political and media cultures, remains an open question. How media coverage constrains and shapes political discourse during referendums, and whether it consistently acts to undermine the possibility of democratic deliberation, should be studied in more depth.

Notes 1 Most treatments of the media are focused on attempts by the two sides to convince voters using advertising (see Zisk, 1987). 2 Content analysis data for the 1997 Canadian election was undertaken by the Canadian Election Study team of André Blais, Richard Nadeau, Elisabeth Gidengil and Neil Nevitte. 3 Most of the analysis was taken from Radio-Canada, but on those days when videotapes for Radio-Canada were unavailable, those of TVA were substituted. No systematic differences between the two could be identified. 4 Even if these stories were primarily about other things, about one-quarter of these campaign and leadership stories had some identifiable issue content to them.

11 Constitutional Referendums and Democratic Deliberation Simone Chambers

Democratic theory has taken a deliberative turn. More and more theorists are turning away from liberal, pluralist, or economistic understandings of democracy and toward a view anchored in conceptions of accountability and discussion. Voting-centric democratic theory is being replaced by talk-centric democratic theory.1 The voting-centric view sees democracy as the arena in which fixed preferences and interests compete via fair mechanisms of aggregation. In contrast, deliberative democracy focuses on the communicative processes of opinion and will-formation that precede voting. Theorists of deliberative democracy are interested in how deliberation shapes preferences, moderates self-interest, mediates difference, produces reasonable opinion and policy, and potentially leads to consensus. While nineteenth- and early twentieth-century democratization was concerned with expanding the vote to include everybody, today democratization often focuses on expanding the public sphere to give everyone a say. 2 Voice rather than votes is the new vehicle of empowerment. Voice, however, does not translate into authoritative decisions without votes. This raises a number of interesting questions for deliberative theory. Different voting procedures, for example, have different effects on deliberation by creating different incentive structures.3 One of the challenges facing deliberative theory, therefore, is to evaluate voting procedures from the point of view of whether they promote deliberation, are neutral with regard to deliberation, or actually sabotage the possibility of deliberation. In this paper I investigate this issue with regard to referendums. I argue that particularly in the case of constitutional policy, where we want citizens to try and achieve a high level of impartiality and mutual respect, referendums tend to undermine deliberation. They do this for two reasons. The first is that they introduce an extreme form of 231

232

Simone Chambers

majoritarianism that is inappropriate for deep constitutional questions. I develop this very common criticism of referendums in a slightly unusual direction by drawing out how the prospect of a majority win is likely to affect people’s willingness to deliberate in the first place. Second, referendums often present the voter with the image of inflexibility (debate cannot alter the framing of the question) and irreversibility (constitutional proposals are entrenched outside the ongoing iterative process of normal politics). These factors act as disincentives to adopt a deliberative point of view. In what follows I begin with a short outline of the main features of democratic deliberation. My aim here is to stress that even mildly productive deliberation contains stronger dispositional requirements than many theorists of deliberative democracy seem ready to admit. Citizens and elites must be willing to act discursively rather than strategically. I then discuss why deliberation, as I have outlined it, is particularly important for contemporary constitution-making. The main point is that constitution-making is becoming a more democratic process (whether we like it or not) and we need to find appropriate mechanisms to include citizens. Referendums are an obvious choice and have been used in many contexts. In the two sections that follow, I discuss majoritarianism first and then move to inflexibility and irreversibility. I focus on why the problems introduced by these characteristics are particularly acute in constitutional referendums. In the final section I make some preliminary proposals for mitigating the negative effects of referendums.

A discursive model of deliberation What does it mean when we say that we want citizens to deliberate about public policy and the political choices they must make? One of the ways that answers to this question differ is where they fall along a continuum between monological and dialogical. At the monological end, one finds a conception of deliberation that simply means that we want citizens, particularly when they are voting, to be thoughtful and informed. Here deliberation is understood in contrast to things like impulse, prejudice, or habit. Deliberation entails thinking through one’s choices or ‘weighing a thing in the mind.’4 This type of deliberation can be achieved while safely ensconced in an armchair. Furthermore, this conception often leaves open how citizens should be thinking about public policy, that is from what perspective. At the other end of the spectrum are dialogic conceptions of deliberation in which the conditions

Constitutional Referendums and Democratic Deliberation 233

of thoughtfulness are laid out in detail. 5 The argument here is that the best way to deliberate about issues, especially issues of public policy affecting large numbers of people, is in dialogue and debate with others affected by the policy. The dialogic conceptions of deliberation folds the monological requirements of ‘thoughtful and well-informed’ into a more demanding model. Take for, example, Lynn Baker’s excellent, although incomplete, definition of deliberation: First, deliberation requires that one be well-informed about proposed legislation. This includes being knowledgeable about competing – and minority – ideas, needs and perspectives. Second, it requires that this information be thoughtful and rationally considered rather than reacted to emotionally . . . Third, deliberation requires that one be able to exchange views on proposed legislation with other decision makers, if one chooses. Fourth, deliberation requires open-mindedness. One’s preferences must be revisable in light of discussion, debate, and new information. 6 Although only the third and fourth requirements are explicitly dialogical, a dialogical model of deliberation usually maintains that the first and the second requirements cannot adequately be met without some form of critical exchange. Why is the dialogic nature of deliberation thought to be so important? Dialogue can perform two functions that even the best of monological thinking cannot. The first is that it subjects individual and group preferences to a publicity test. 7 Having to defend one’s policy preferences in an open public forum requires, first of all, that citizens think through the reasons why they hold certain opinions or policy preferences. At a minimum, this forces participants to be thoughtful in a way that the monological model does not. Gut reactions, prejudice, and indefensible assumptions, as well as misinformation and ignorance, are often exposed in critical debate forcing participants to think through their reasoning more clearly. Furthermore, not all types of reasons and justifications are acceptable in public debate. For example, public policy preferences which could only be justified by an argument like ‘this benefits me but no one else’ are not likely to be persuasive or survive a publicity test. And, indeed, few people argue this way in public. Ideally, the activity of giving and listening to reasoned argument will initiate a learning process whereby participants abandon positions for which there are no generalizable reasons. Of course, this does not preclude the possibility of cynical

234

Simone Chambers

participants who tailor their arguments to what they can realistically get away with. Looking at the state of public debate today, it is difficult not to conclude that many so-called generalizable arguments really mask self-interest. Our public sphere falls very short of the ideals of deliberation and publicity. But it is not impossible to imagine a differently structured public sphere, perhaps with more equal access, where public debate does a better job at identifying the wolves in sheep’s clothing. Thus, present conditions not withstanding, the claim here is that public debate can serve as a ‘laundering’ mechanism in which preferences are developed and pursued in light of what other citizens are likely to say in response. 8 A dialogic model of deliberation contains a higher threshold of accountability than a non-dialogic model. As we shall see, this function is particularly important in constitutional debate where we do not want citizens to approach questions of fundamental law and rights from an exclusively private or selfish perspective.9 Publicity is under particular threat, however, when referendums are introduced as ratification mechanisms for constitutional politics. The reason is not that there is no public debate. The reason is that accountability, in the sense of sincerely responding to challenges and directing arguments to others, is often overshadowed by strategic campaign maneuvering. The second function that dialogue can fulfill that mere thoughtfulness cannot is integration. Deliberation on this model is not just about well-informed and thoughtful voters, it is also about citizens who, in deliberating with each other, build bridges of recognition, understanding, and sometimes consensus. Although critical, and sometimes contestatory not to mention acrimonious, deliberation is cooperative in the sense that ideally participants share the goal of arriving at jointly acceptable principles and policies. Deliberation stands against adversarial bargaining models. The object in deliberation is to be accountable to and understood by one’s interlocutor, not to vanquish her. The practice of accountability involves participants in a process of mutual understanding. This process does not create strong communal ties or common identities but, if successful, it can create recognition between groups such that individuals and groups feel that they count for something in the process. Again, this function is particularly important in constitutional politics. I acknowledge, however, that there are many political contexts in which more adversarial approaches might be appropriate. The discursive model of deliberation I defend here, while quite strong, is not intended to replace all forms of democratic participation. Preference ‘laundering’ and integration cannot be achieved by just any type of conversation and debate. A deliberative conversation must


meet certain conditions. The first set of conditions entails ensuring that the debate meets an equality requirement. What this means is that, ideally, anybody and everybody can participate, all participants have equal status as participants, no one may be silenced and all opinions heard. In the less than perfect world of real politics, we try to approximate this condition. It is relatively clear why the equality condition is necessary to fulfill the two functions. An integrative sense of counting for something is not likely to be achieved in a public sphere where advantages and privileges are severely skewed in favor of certain groups and individuals. And if certain groups have unfair advantage in the conversation then their interests and claims will never meet credible challenges and so, although public, will not have undergone a publicity test. The conditions of equality are minimally met through such things as free speech guarantees and citizenship rights. They are more substantively guaranteed through exposing the ways in which public spaces fail to be level playing fields despite formal equality. Sometimes these problems can be addressed through such things as voter registration and education campaigns, subsidized media access for the under-represented, or campaign finance and access regulation. Although failure to create a level playing field can have serious consequences for the caliber of deliberation in a referendum, and indeed it is this question which animates much of the debate in the United States over the desirability of initiative democracy, I will only discuss these problems tangentially. This is for two reasons. First, the United States is quite unique in its resistance to egalitarian regulation of the public sphere.10 Although probably possessing the most advanced and well-developed direct democracy movement, the United States is still very far from introducing a referendum at the national level although many states allow their constitutions to be amended via initiative. Ultimately, however, these initiatives are subject to judicial review that is highly insulated from direct democratic influence. Thus, the American context, where the problem of money is particularly acute, is not central to the cases that interest me. Second, I want to argue that even in a ‘well ordered’ public sphere, that is in a public sphere that meets basic requirements of justice and equality, referendums pose problems for deliberations. Although formal and some forms of substantive equality are necessary conditions for deliberation, they are not sufficient. Deliberation will not succeed in fulfilling the functions of publicity and integration unless participants possess certain dispositions and interests. Let us say that participants must be, for the most part, discursive as opposed to

236

Simone Chambers

strategic actors. These are ideal types. None of us is ever purely one or the other but always some combination. Nevertheless, this distinction is a helpful heuristic in identifying the defining features of democratic deliberation. Strategic actors, although willing to compromise, are essentially interested in maximizing their outcome in a world which they believe is populated by others with the same interest. By contrast, discursive actors are interested in persuading others and so finding solutions to public problems that are acceptable to all. Neither of these goals, that of maximizing one’s own interests or maximizing generally acceptability, are ever fully achieved. They do set in motion different interactive dynamics, however. One way to conceptualize the difference is to note that the strategic actor is rational and the discursive actor is reasonable. 11 The rational is usually understood instrumentally and refers to individual calculation of utility and efficiency. The reasonable incorporates mutuality and reciprocity. The reasonable person is not simply the person who has good reasons for her actions, it is also the person who is willing to listen to objections, who is open to suggestion, who will re-evaluate a position in light of new evidence, and so on. But above all, the reasonable person understands that the mutuality of common life creates an obligation to explain herself to others. A reasonable policy choice is one that makes sense given that one lives in a world with other people who also have policy preferences that are just as important to them as yours is to you. Reasonableness asks voters and deliberators to think through public issues from a cooperative perspective. This does not mean that citizens need to be altruistic or should refrain from pursuing their interests if these are controversial or not widely shared. It means that in pursuing our interests in the public realm we ought to be aware that we are not the only ones with claims, and in pressing these claims in the public sphere we ought to justify them in terms that sincerely speak to our fellow citizens’ concerns. Again I argue that constitutional questions demand a higher level of discursive action than most day-to-day politics. There are many occasions in the public life of a society when a strategic stance is appropriate. When discussing constitutional essentials that will set the fair rules of the game for the future, however, mutual acceptability should be our reference point more than maximization. Rights are not the sort of things we should be using as bargaining chips or subjecting to horsetrading. While almost all advocates of deliberative democracy agree that successful deliberation requires some form of reasonableness, they are not always clear on why we should expect people to be reasonable. In


particular, there is a failure to grapple with what can be called the Hobbesian predicament: I accept in principle that constitutional politics ought to be approached from a reasonable perspective, but I also know that if I am reasonable and everyone else is rational, I am likely to come out very badly in the end. Although the dilemma is Hobbesian the solution, thank goodness, is not. There are institutional mechanisms that can promote reasonableness short of a Leviathan. Referendums, however, are not one of them.

Are the people smart enough to speak for ‘We the People’? The epistemic objection I have outlined a strong conception of deliberation. It is important to stress that not all forms of ‘thoughtfulness,’ exchange, or debate are forms of deliberation in this sense. At the core of deliberation is an ideal of mutual accountability among citizens and groups, and between citizens and elites. Mutual accountability requires reasonableness on the part of participants. In particular this entails that one’s preferences must be revisable in light of discussion, new information, and especially the acknowledgement of other people’s reasonable claims. Although deliberation certainly has a role to play in day-to-day politics, I am primarily interested in the case of constitutional politics. Constitutional documents set out the basic principles, laws, and procedures by which a group of people, who are often very diverse, will resolve their future disputes and set the moral side constraints for their interaction. All modern or liberal democratic constitutions are either explicitly or implicitly prefaced by ‘We the People . . . ’ Although modern constitutions often appeal to natural right or even God in justifying first principles, their legitimacy rests in the assumption that they speak for the people. In 1787, a small American elite could begin a constitutional document with ‘We the People’ and (perhaps) believe that they spoke for the people, despite the fact that women, slaves, and Native Americans were excluded and most citizens were only minimally involved in the process.12 The pressures of egalitarian democracy combined with demands for the inclusion of different voices has made this type of constitution-making increasingly untenable. 13 Most of the constitutions ratified (or almost ratified) in the last decade of the twentieth century (for example, Canada, South Africa, Ethiopia, Uganda, Chile, Australia, some Central and Eastern European countries), while clearly influenced by the American model, went significantly beyond that model both with regard to the content of the

238

Simone Chambers

constitution and the democratic procedures used to frame and ratify the constitution. 14 The demand for democratic inclusion in the process of constitutionmaking has raised a dilemma. On the one hand, real citizen participation appears to allow us to come closer to the ideal of ‘We the People’ by actually canvassing the diverse interests, perspectives, and concerns in society. 15 On the other hand, by including citizens we risk making constitutions hostage to shallow and misinformed opinions, destructive ‘special interest’ politics, or divisive group politics. Thus, one of the major tasks facing modern constitution-making is to design procedures of popular consultation that allow the people real voice at the same time as mitigating the worst aspects of populism. Referendums are a prime example of this dilemma. On the one hand, referendums and plebiscites appear to bring the people into the closest possible relationship to the constitution. Here popular sovereignty is taken at its word. Canada, Australia, and New Zealand have all used referendums as a mechanism of inclusion and legitimacy. Bruce Ackerman has made a powerful argument that popular sovereignty would be better served if the American people were able to amend the constitution through a referendum. 16 On the other hand, the prospect of the people directly deciding constitutional matters raises the specter of majoritarian and populist abuse. Often cited as an example of what people will do if they had such power is Amendment 2 in Colorado in which a ballot initiative succeeded in amending the Colorado constitution, in effect, to curb the civil rights protections of gays and lesbians. This amendment was subsequently struck down by the US Supreme Court confirming many people’s intuitions that we want a thoughtful, well-informed, and rights-friendly elite deciding issues of fundamental rights, not the general public. 17 The arguments against referendums often rest on an assessment of citizens, existing levels of knowledge and commitment. So, for example, Philip Weiser is suspicious of Ackerman’s proposal because referendums lack ‘the filter necessary to prevent Americans from voting away essential protections of liberty, basic civil rights, and thoughtfully entrenched institutional arrangements.’ He adds that ‘the present low levels of duty to the polity, understanding of constitutional principles, [and] empathy for others’ makes the use of constitutional referendums particularly problematic.18 Although I too am wary of using referendums to decide constitutional matters, my argument does not rest on an evaluation of existing public opinion or how rights-friendly is the general population. There are a number of problems with this particular


criticism of referendums. First, the general levels of sympathy for and understanding of rights is in dispute. For example, the debate about whether direct democracy is likely to be bad or good for minority rights is quite evenly split. 19 So is the debate about how well or ill informed, sophisticated or unsophisticated are voters.20 Second, and more important, this approach to the problems and perhaps dangers of referendums takes a static view of the relationship between existing public opinion, elite interpretation of the constitution, and referendums. It investigates the desirability of referendums from the point of view of the discrepancy between existing public opinion and elite or ideally ‘enlightened’ public opinion. The public opinion approach, first of all, fails to see how the procedures of participation can affect opinion and enhance learning. Second, this criticism by itself fails to tackle the problem of legitimacy. If one is to argue that we should avoid direct democracy, particularly in fundamental questions of rights, because citizens are not capable of making these decisions thoughtfully, then one must reconcile this epistemic argument with the claims of popular sovereignty. Finally, arguments based on existing levels of knowledge and thoughtfulness fail to tackle the problem of stability. If it is the case that there is a serious and potentially dangerous gap between elite rights-discourse and how ordinary citizens understand rights then the solution cannot be and should not be to insulate elite discourse from the ‘unenlightened’ public even further. For in the end, without an underlying understanding and endorsement of rights on the part of the ‘unenlightened’ public, rights are at risk. The relationship between citizen participation in constitutional matters and the stability of constitutions over time is particularly evident in countries trying to establish a constitutional tradition. In South Africa, for example, citizens were exhorted to participate in the constitutional process with the slogan ‘it is your right to decide your rights.’21 From a strictly public opinion point of view this appears to be a very foolhardy claim. Many South Africans had little or no understanding of what rights were or even what a constitution was.22 But it was precisely because of this lack of knowledge that framers were insistent that the public be involved at every stage of the process. 23 The process of participation itself was an education in constitutional politics that strengthened legitimacy and stability. There is no reason why this might not also be the case in established democracies. My argument against some forms of direct democracy is not based on a negative evaluation of citizen competency, capability, or knowledge levels. Indeed, I would be more inclined to argue that we want more

240

Simone Chambers

rather than less citizen input in constitutional matters. At first glance referendums appear very attractive as mechanisms of constitutional renewal, ratification, and amendment. They can potentially serve as catalysts for a higher level of engagement and participation and this is a good thing. This is certainly Bruce Ackerman’s optimistic evaluation of constitutional referendums. He feels that in allowing citizens to exercise popular sovereignty at times of transformation, ‘apathy will give way to concern, ignorance to information, selfishness to serious reflection on the country’s future.’24 Ackerman argues that a national referendum can promote deliberation in the best and strongest sense by getting citizens involved and galvanizing them on important issues. This can sometimes be the case. By many accounts the Irish referendum of the peace accord was such an experience. However, what I want to argue next is that advocates of direct democracy often fail to see the way in which referendums, if not very carefully designed, undermine meaningful participation. I am making the opposite argument to the epistemic skeptics who worry about popular opinion contaminating fundamental law. I am saying that referendums do not do a good enough job representing the views of ‘We the People.’

‘We the people’ who won versus ‘We the People’ who lost: maintaining deliberative accountability Although the South African constitutional process was highly consultative, citizens were not asked to vote in a referendum. A referendum, or more precisely the prospect of a referendum, did play a role in the process, however. The ANC used the threat of a referendum to keep negotiating parties at the table. 25 Because the ANC could carry a clear majority of the voters, any constitutional proposal that they endorsed would more than likely pass a plebiscitary test. On a number of occasions, walk-outs by the IFP (Inkatha Freedom Party) were ended by the threat ‘either come and talk or take your chances at the polls.’ In this situation there was a clear and expressed tension between deliberation and referendum voting: initiating a referendum would signal the end of talk, not the beginning. This tension resides in the majoritarian nature of referendums. I want to suggest that the strong majoritarianism of referendums derails deliberation and in so doing undermines the legitimacy of outcomes. My argument here is not that referendums allow majorities to oppress minorities or that majorities are often just looking for the opportunity to oppress minorities. My argument is that the prospect of a majority win affects the type of reasons participants are


likely to appeal to in debate and this in turn affects the functions of publicity and integration. I have found jury deliberation studies helpful in highlighting some of the problems I want to discuss. Although the analogy between jury deliberation and democratic deliberation is by no means perfect, it can identify potential problems with public deliberation. 26 Studies that compare juries that have unanimity as a rule versus juries that have majority rules show significant differences in the internal dynamic of deliberation. Indeed, some have argued that where there is a simple majority rule, jurors do not really deliberate in the sense of formulating arguments that could persuade interlocutors. ‘When a jury is required to reach a unanimous verdict, the jurors must deliberate with each other. Any rule that permits a majority to prevail raises the possibility that the majority will not listen to a minority once sufficient votes exist to return a verdict.’ 27 The South African experience seems to confirm some of this. Once a referendum is called or even knowing that the conversation will end in a referendum diminishes the sense that what one has to say counts for something. Both majorities and minorities lose an important incentive to be reasonable. Even in situations where there is no clear majority, referendums invite participants to approach debate strategically rather than discursively, that is it creates the incentive to find arguments that will sway only the needed number of voters. Knowing the rules of the end game exercises a huge influence on how participants approach the process. This fact has been overlooked by much of the literature on deliberative democracy. The tendency here has been to introduce voting as an unproblematic decision procedure for deliberation. The argument often goes like this: if deliberation does not end in consensus or if there is not enough time to reach consensus then participants simply vote.28 This fails to acknowledge that (a) the prospect of voting might affect the quality and substance of deliberation and (b) there are many different ways to take a vote, often introducing different incentive structures and so having different effects on deliberation. I want to suggest that, if deliberation requires citizens to be reasonable in addition to being rational then we need to look at the way voting mechanisms promote or hinder that disposition. The problem pointed to in jury studies is that once unanimity is rejected as unfeasible, there is always the risk that majorities will not listen to minorities, that dominant groups will not try to understand the points of view of those without power, that participants will not engage each other in a cooperative search for a solution. For their part, minorities too may feel that there is no point in trying to formulate persuasive arguments as these will fall on deaf ears. One possible outcome of this

242

Simone Chambers

logic is that participants in referendum debates are more likely to resort to negative and aggressive tactics in the hope, not of persuading or exercising accountability, but simply in discrediting. This suggests that majoritarianism is antithetical to the spirit of deliberation, not simply because its outcomes may be morally unjustifiable to minorities, but because as a decision rule it does not give citizens strong reasons to engage in mutual accountability in the first place. The problem of majoritarianism is particularly acute when constitutional referendums are used. As long as there is no permanent minority, day-to-day politics and representative democracy always hold out the possibility that in the next election tables will turn and majorities will be whittled down to minorities and vice versa. This does not prevent the majoritarian deliberative dynamic from plaguing ordinary elections but the ongoing nature of day-to-day politics can sometimes create additional incentives to act discursively. Constitutional referendums which are often perceived as decisive and almost irreversible sometimes create the impression of entrenching a permanent minority status for those who lose the vote. This creates the divisive picture of a nation being built on the division between ‘We the People’ who won versus ‘We the People’ who lost. Let me try and offer a few details about this dynamic and ways we can perhaps avoid it. When fundamental or constitutive issues are raised, we want to get citizens to approach the question discursively which means we want citizens to listen to each other, respond to each other’s claims and grievances, argue sincerely not strategically, weigh alternatives dispassionately, approach the dispute from the point of view of possible agreement rather than from the point of view of maximizing one’s or the group’s utility, and finally, be swayed by reason. One of the most obvious ways of doing this is to introduce a specially constructed ratifying formula designed to encourage deliberation. At a minimum, the ratifying (and amending) formulae should reflect a generally accepted threshold of consent often, but not always, higher than required for ordinary legislation. In addition to this, the construction of ratifying (and amending) formulae also sometimes reflects concerns about the quality of deliberation. Article V of the American Constitution, for example, is designed to ensure not only that a legitimate threshold of consent is reached, but also that cool heads and dispassionate deliberation will prevail.29 In contrast, referendums and plebiscites by themselves are conceived exclusively from the point of view of the threshold of consent. This can be problematic. A referendum preceded by little or very acrimonious debate but resulting in 75 percent of the


vote supporting the proposition might hypothetically have less democratic legitimacy than a protracted deliberative process with many stages, much discussion, and lots of opportunity for challenge and criticism but with the winning side receiving only 60 percent of the vote. Why would deliberated outcomes that garner less of the popular vote sometimes be more legitimate than non-deliberated outcomes that gain a larger percentage of the vote? Deliberation (as I have described it) lends a stronger legitimacy to majority decisions because majorities must be accountable to minorities even while they can outvote minorities. It is not reasonable to say something like ‘this constitutional initiative is just because it benefits us and we are numerically stronger than you.’ Majorities need to find arguments that at least appear to appeal to generalizable interests. 30 And such appeals can be publicly challenged and questioned, forcing majorities to find more reasonable policy options or lose even their own supporters through a loss of credibility. As many theorists of deliberative democracy have pointed out, the argumentative dynamic of deliberative accountability has a logic of its own that can push even cynical and deceptive participants toward either persuasive arguments or exposure as cynical and deceptive.31 The more robust deliberation is, in the sense of maintaining conditions of equality and accessibility and encouraging participation, publicity, and cooperation, the more accountable majorities will be to minorities and the more decisions based on majority rule will reflect reasonable policy. Reasonable policy here means that the policy is backed by reasons that have been scrutinized and defended in light of minority objections. Coherent majoritarianism requires that it is not simply the most votes that signals a legitimate outcome, it is the most votes produced by the best arguments. 32 Dewey captured this point well when he said: Majority rule, just as majority rule, is as foolish as its critics charge it with being. But it is never merely majority rule. . . . The means by which a majority comes to be a majority is the important thing: antecedent debates, the modification of views to meet the opinions of minorities. . . . The essential need, in other words, is the improvement of the methods and conditions of debate, discussion, and persuasion. 33 Referendums are particularly prone to the problem of weak deliberative conditions underpinning a majority decision. Referendums, without supplementary institutions and forums that promote deliberation, invest so much in numbers rather than arguments that it is hard for the

244

Simone Chambers

losers not to read the outcome as ‘might makes right.’ This undermines the legitimacy and stability of outcomes but, more important for my argument, creates a chilling effect on the debate that precedes the vote. Thus, although it is correct, it is not enough to say we need to talk before we vote. We need to talk before we vote, but knowing that at the end of the day there will be winners and losers weakens the incentive to talk deliberately. The latest turn in the Quebec sovereignty debate is a case in point. The referendum on sovereignty in which the vast majority of Anglophones (who are a minority) supported the ‘no’ and a small majority of Francophones (who are in the majority) voted ‘yes’ was extremely divisive.34 The campaign, outcome, and recriminatory aftermath threaten to undo the trust and recognition that has been slowly growing between the English and French of that province over the course of the past twenty years.35 The referendum on sovereignty was accompanied with little or no effort to engage citizens in deliberation outside of the intense campaign period prior to the vote.36 In particular, there was no attempt to respond to or answer minority fears about what would happen on the ‘Day After.’ There was a great deal of talk prior to the decision but very little of it was deliberative. The sovereignty debate has taken a further turn away from deliberation. Rather than talking about the pros and cons of sovereignty, the debate is now about whether the federal government and/or the internal opponents of sovereignty will recognize as legitimate a 50 percent plus 1 type of win for the independentist option. Without a clear agreement about what could constitute a fair procedure in deciding such a fundamental issue, everything is in limbo. The dispute about fair procedures is irresolvable, however, if it remains at the level of numbers and does not include conditions of fair and reasonable deliberation. To argue solely in terms of the numerical threshold of consent adds up to a discussion about how large a majority has to be before it can simply do what it wants. Rather than discussing the threshold of consent, a more fruitful approach might be to discuss venues of popular consultation such that in the event of, say, separation, minorities feel confident that the majority is not only very much aware of their worries and fears but that these have been addressed in the separation proposal. Furthermore it is important that these discussions take place outside the parameters of a referendum campaign. Here is the paradox: majoritarianism requires deliberation in order to be legitimate but majority decision rules create disincentives to engage in deliberation, that is to participate as a


discursive actor. Participants must feel that what they say makes a difference. The stronger is the majoritarian dynamic, the less likely participants will feel that what they say will make a difference, thus the less likely they will be to talk reasonably rather than act strategically.

The lobster trap phenomenon: the problem of irreversibility The majoritarian dynamic outlined above points to a further problem with referendums. Voting disrupts talking. Voting signals closure which means that voting literally ends the talking. Voting means taking a decision that is more or less final. Think of the difference between conversations at a faculty retreat where no decision will be taken and then discussion in a faculty meeting where a tough decision must be made. 37 Not surprisingly, participants are often more flexible and open under the former conditions than the latter. It would appear that participants are more likely to be discursive actors where there is no decision at stake than when there is an impending one. The problem I am discussing first struck me as I was trying to interpret a strange and disturbing shift in the Canadian constitutional debate. In the summer of 1992, a number of conferences were set up across Canada to discuss possible solutions to the various constitutional impasses facing the nation. Conference participation represented a relatively broad cross section of Canadians and all five conferences were broadcast across Canada.38 By all accounts these were very constructive.39 Advocates even of quite extreme positions were willing to sign on to ‘consensual’ documents that reflected the general tenor of the weekend-long debates. The procedures adopted within the conferences conformed to ideals of deliberation in that there was a conscious attempt to guarantee dialogical equality so that everyone could speak and be heard and to exclude influences, such as money and power, that might distort the conversation. But even more interesting was the fact that no votes were taken. The mornings were taken up by workshops which came together in afternoon plenary sessions to work out position papers. These papers reflected a conference-wide ‘consensus’ in that they recorded what people were able to agree on as well as those issues upon which no agreement was reached. Many of the proposals worked out at these conferences made their way into a formal constitutional proposal that was put to the citizens of Canada in the form of a referendum in the fall of 1992. The hope of many was that the positive experience of the conferences could be recreated at the national level. Since the conferences themselves

246

Simone Chambers

had a high approval rating among citizens at large, this hope did not seem so far fetched. It did not, however, pan out. It is not simply that the proposal failed to be ratified, for, after all, the conferences could have got it wrong or failed to take into account a crucial interest or claim.40 Another type of disappointment was focused on the caliber and tenor of debate. Lacking was the openness and flexibility of the conference participants, that is the willingness to revise claims to make them fit with other, now perceived as equally legitimate, claims. Instead, the referendum campaign appeared to harden positions. Furthermore, leaders and spokespersons began to talk in zero sum terms, to the effect that any concession to other interests would be a loss for their side. Rather than reasonable argument, fear of being a loser in the deal was played up. All in all, the referendum debate had an effect opposite to what deliberation is supposed to have: it moved participants further apart, heightened distrust, exacerbated misunderstanding, and left Canadians in a worse place than when they started, certainly with regard to Quebec/ Canada relations.41 So we have two conversations about the same thing with roughly the same interests being represented, both of which did a fair job of maintaining conditions of formal equality and equal access. What can account for the very large difference? Obviously there are many differences between five face-to-face meetings, each containing about 250 people who are thrown together for an intense weekend, and a 56-day referendum campaign trying to sway millions of citizens. A full comparison would have to discuss and deal with all differences. At the moment I want to single out one difference that offers an interesting perspective on the respective tones of the debates. This difference has to do with the relationship between deliberation and authoritative closure. Not only did participants at the conferences not take any votes or make any internal decisions, but their role in the larger constitutional debate was left vague. They had no clear mandate. The outcomes of these conferences did not bind in any authoritative way. The end of the conferences did not represent any kind of closure or decision. It is not quite right to say that nothing was at stake, for it was not an academic conference discussing constitutional options for ancient Athens but rather citizen-participants who were aware of the seriousness and urgency of the issues and knew that they had the ear of the public as well as the political elite. Nevertheless, it is correct to say that a binding or authoritative decision was not at stake and this freed participants from the fear of premature or disadvantageous closure. Participants could ‘afford’ to be flexible, open, and cooperative. This in turn led to a real movement


of positions and convergence on issues. During the referendum campaign, however, a whole new procedural dynamic was set in motion that made it difficult to recreate that flexibility. Now participants were confronted with a deadline at which point the proposal, if passed, would be authoritative and very difficult if not impossible to revise.42 The fear of not being the loser appeared to overwhelm any principled desire to reach cooperative agreement. This quite natural fear, raised with any prospect of closure on a disputed issue, was played on and exacerbated by competing political elites.43 The negative effect of a vote that is perceived as closure on a particular issue can also be seen in the most recent vote on sovereignty in Quebec. This campaign, as I mentioned earlier, was very divisive and lacked many, if not most of the characteristics of a healthy and productive deliberation. In an unguarded moment, Quebec Premier Jacques Parizeau, leader of the ‘yes’ campaign, identified an unstated fear that lurked behind the debate. The metaphor he chose for a pro-sovereignty vote was the lobster trap: once Quebeckers had voted yes they would be like lobsters with the trap door closing behind them – unable to escape their fate. When closure is perceived to mean no escape, the motivation not to be the loser may be overwhelming. Under these circumstances, referendums may bear more resemblance to peaceful civil wars than democratic deliberations. Like civil wars they leave scars as reminders of what people will do and say when they perceive that they are involved in an all or nothing battle. Secessionist disputes are often the most susceptible to this dynamic because secession is usually all but irreversible. The possibility of reversing a decision somewhere down the line when more information is available or the debate takes a new turn is a very important component to stable democratic decisions. Constitutional questions also often fall within this logic because constitutional referendums are often extraordinary political events not regularly repeated. Indeed this raises a paradox for constitutional referendums. On the one hand, we want citizens to take constitutional change and ratification seriously and perceive it as more fundamental than ordinary legislation, but on the other hand, the more it is perceived as the last word in the matter the more likely it is the strategic logic of a zero sum game will kick in.44 One might think that this argument, rather than making a case against referendums, really makes a case against deliberation as I have conceived it. If the dispositions necessary for authentic deliberation can only be expected to be present in conversations which are, on the one hand, non-decisive but, on the other, deal with issues needing decisions, this appears to drastically limit the relevance of deliberation. Decisions need

248

Simone Chambers

to be taken. Deliberation needs to end in a vote. But if the very prospect of taking a vote makes cooperative deliberation unlikely we have a serious problem. One might want to draw a pessimistic conclusion: talking (understood as deliberation) and voting do not mix well. When voting is involved, particularly voting that entrenches, talking is likely to be strategic, competitive, and adversarial. But, given the complexity of modern political society, all talking must end in voting.

Prognosis: the art of institutional design The prognosis for deliberation need not be so gloomy. My argument leads to two conclusions that I will introduce again by appealing to jury research. Some juries are described as ‘evidence driven’ while others are ‘verdict driven.’45 In the verdict driven juries many votes tend to be taken while in evidence driven juries discussion is less focused on closure, and more on the spirit of airing all the facts while holding off a decision. Evidence driven juries try not to discuss conclusions. In both cases, of course, a vote is taken at the end of the deliberative process. What is interesting is that in interviews and polling after jury duty, participants in evidence driven juries had a stronger sense of satisfaction with the process, the verdict, and the reasonableness of their fellow jurors. This is not to say that they came up with better verdicts (although, interestingly, they tended to be more lenient) but it does point to the fact that the process was perceived as more legitimate and they had a stronger sense of efficacy, that is, that they as individuals had made an important contribution to the process and that they were actually listened to. Again it is risky to make huge generalizations from these studies to the political world. Nevertheless there do seem to be some implications. The foregoing discussion leads to two conclusions. The first is that some component citizen (and elite) participation in constitution-making should be insulated from majoritarian decision procedures. Second, when using popular referendums, we should try and find ways to avoid investing the outcome with too much finality. Let me take up the first conclusion. Referendums, as I have argued, need to be supplemented by deliberative institutions that can create deliberative accountability between citizens. It is almost impossible to do this during a referendum campaign. These campaigns are ‘verdict driven.’ What we want is to promote ‘evidence driven’ conversations. One reason we want this is that, like in jury deliberation, this type of conversation can give citizens a sense of confidence that they were listened to, that they had the


opportunity to voice their concerns, that all the information and issues were aired, and that the conclusion is fair and just. Perhaps in jury deliberation it is not that important how the jurors ‘feel’ about the process as long as they get the answer right. But in constitutional deliberation this is not the case. The right answer to questions like ‘What rights should we recognize and entrench?’ is not independent of who are the ‘we’ in question. We cannot divorce outcome from process in constitutional deliberation because the outcome is about process. Constitutional debate is about the rules that govern citizen interaction and debate in the future. Constitutional deliberation is like a jury deciding on what rules will govern the jury system. Thus it is important that the process itself reflect the principles that will become entrenched and this in turn means that it is important that participants ‘feel’ like they have been treated as equals. Unlike isolated jury deliberations, the stability and legitimacy of the outcome often depend on how citizens perceive the process. 46 This is especially true for newly democratized countries. A ‘good’ or ‘liberal’ constitution that is introduced without concern for citizen participation or the process of ratification will often lack stability.47 One way to begin creating a constitutional culture that can anchor a constitutional document in the hearts and minds of citizens is to use the processes of consultation and ratification as educative tools. But again it is important that participation is evidence and not verdict driven. How do we do this? Now, one way to ensure that citizens do not get distracted by the endgame is not to invite them to participate in that game, or not directly. So, for example, in South Africa there was a great deal of popular consultation by way of public meetings, sectoral hearings, workshops, a telephone hotline and Internet site. Two million submissions were received and maintained on an Internet database in addition to a newsletter that kept citizens informed of the progress of controversies. These initiatives were essential in maintaining the legitimacy of the process and by extension the document itself. Citizens were not asked to vote directly on the document, however.48 The South African experience is not necessarily less democratic than the Canadian one. Much depends on an evaluation of the connection between popular consultation and the deliberations of the Constitutional Assembly. If the Assembly was authentically accountable to the people even without the background threat of a vote, then we could say that the process was democratic despite the lack of direct voting. This is not the place to make that evaluation. But I do want to stress that, in principle, informal consultation can fulfill the role of authenticating a hypothetical ‘We the People’ preamble.

250

Simone Chambers

Although risky, I do not want to say, however, that referendums are always out of the question from the point of view of deliberation and democratic legitimacy. After all, voting at the end of the day does appear to introduce a more obvious mechanism of elite accountability than informal consultation without a vote. Furthermore, Ackerman is right to say that referendums contain a potential galvanizing impetus to get citizens concerned and involved. But if referendums are introduced then they should be very carefully designed so as to promote evidence driven debate. This means not simply that governments are obligated to develop deliberative forums and opportunities to participate. It also means that these must be properly insulated from majoritarian voting. Constitutional referendums need to be supplemented with unofficial and informal consultation. Citizen workshops, town hall meetings, and people’s conventions as well as creative use of the media and the Internet to engage citizens should take place well before referendum dates are set or wording is formalized. Developing, encouraging, and funding a variety of civil society organizations to help promote deliberation also can serve as a healthy counterbalance to the referendum dynamic which sees all opinion coalesce under two umbrella organizations advocating the ‘yes’ or ‘no’ side. This too has a tendency to weaken accountability as it aggregates differences in one of two camps where many voices with particular concerns may be silenced or not heard above the campaign din. Finally the campaign itself should be circumscribed within a set period, limiting campaign advertisements and canvassing to that period. In addition to insulating deliberation from the process of decision making, decision making itself needs to be structured in such a way as to minimize the lobster-trap phenomenon. One possible solution to the inhibiting effect of perceived irreversibility is to draw out the process of popular ratification such that votes do not carry the weight of entrenchment and yet still constitute meaningful consultation. James Tully, for example, has suggested that constitutional agreements that need to be submitted to a referendum should be treated as rolling drafts rather than final accords, so that amendments could be accommodated and suggestions solicited.49 Such an iterative model would focus on an ongoing process of consultation rather than a once-and-for-all ratification.50 Drawing out the process into many stages would also address another common complaint made against referendums. In many uses of initiatives and referendums, once the question is on the ballot the debate cannot change it. Thus, one reason why one might construct arguments that appeal to others is lacking. In contrast, representatives can be influenced by deliberation and shift their position in response to objections


heard during the election campaign. Even if one is in a hopeless minority, there is sometimes the chance that one’s arguments may constrain the future action of the winning majority’s representative. No such incentive to put forward persuasive arguments exists with regard to initiatives and referendums. If, however, the referendum process itself included citizens’ input in the framing, this could enhance the deliberative quality of debate. Indeed the less citizens themselves have to do with the framing of the question the more referendums are a strategic ‘card’ to play by elites. Mendelsohn and Parkin note in this volume’s Introduction that in theory referendums are designed to permit citizens to make decisions about public policy, but in practice referendums are used strategically by governments or groups seeking not to turn decision making over to citizens, but persuade voters to ratify a decision that has already been made. . . . the decision to hold a referendum typically stems from a desire on the part of elites to achieve a preferred outcome, and not from a normative commitment to greater public participation.51 It would be better if discrete democratic decisions were not presented as the last word on the matter. If referendums are used, we should consider making them part of an ongoing process of consultation rather than a once-and-for-all ratification. Another measure designed to weaken the ‘endgame’ dynamic is to disaggregate issues into separate votes. Rather than voting on a constitutional ‘package,’ voters should be asked their opinion of discrete clauses and components to the package. This too makes debate more manageable, weakens a bargaining and trade-off approach, and encourages substantive discussion on issues. Finally, introducing measures that guarantee future review or the opportunity to return to a particular thorny issue in the future often has the effect of inducing participants to be more flexible. This could include such things as transition clauses that say something to the effect – let’s try this for five years and then review the result. 52 There are many ways to use and structure referendums in democratic politics. We need to be creative in designing the spaces and procedures through which citizens participate in the democratic process and keenly aware of how democratic institutions often shape the content of participation. Constitutional politics are founding politics. As important as the content of our constitutions are the procedures that we use to bring them about. We should try to adduce as much evidence as possible about how various procedures, including referendums, shape the debate

252

Simone Chambers

and, by extension, the very content of our constitutions. I have tried in this paper to pose a few preliminary questions and offer a few provisional answers to the question of institutional design and constitutional change. My general conclusion is that constitutional referendums need not be out of the question but we should be very wary when we design them and try to build in deliberative conditions as much as possible.

Notes 1 Some of the main contributors to talk-centric democratic theory are Benhabib (1996), Bohman (1996), Fishkin (1991), Gutmann and Thompson (1996), Habermas (1996), Manin (1987), and Miller (1992). 2 There are, of course, many places around the world where the struggle is over the vote but even here we see a keen awareness of the relationship between meaningful voting and a deliberative public sphere. 3 Another challenge that has yet to be fully met by theorists of deliberative democracy is the rational choice challenge. The problem that rational choice points to is the following: even if opinion formation is deliberative and policy preferences are reasonable, many if not most voting procedures fail as accurate translation devices and in fact restructure opinion in all sorts of ways. Imagine two identical deliberations with the same reasonable opinions coming to the surface but one feeding into a PR (proportional representation) system while the other goes into an FPP (first past the post). We would get two very different outcomes despite starting with the same opinion input. See Van Mill (1996). 4 OED. A great deal of political science, particularly voting behavior research, uses this idea of deliberation. See Page and Shapiro (1992). 5 Habermas (1993) probably has the most demanding model. The more a theory of deliberative democracy is tied to strong epistemological and moral claims the more demanding will be the conditions. 6 Baker (1991). 7 I talk here only about subjecting citizens’ opinions to a publicity test. Publicity tests are also and perhaps more characteristically understood as tests to keep government honest (Habermas, 1989; Holmes, 1990). 8 For a discussion of how democratic deliberation may transform preferences, see Cohen (1997), Miller (1992), and Goodin (1989). 9 Sunstein (1984) argues that the American Constitution assumes that citizens will put aside ‘naked preferences’ when dealing with fundamental constitutional issues. 10 The Supreme Court took a strong stand in Buckley v. Valeo when it insisted that ‘the concept that government may restrict the speech of some elements in society in order to enhance the relative voice of others is wholly foreign to the First Amendment’ (Buckley v. Valeo, 1976). Most advanced industrialized countries have come to exactly the opposite conclusion. Referendum regulation, especially when referendums are used to decide fundamental constitutional issues, are even more highly regulated than ordinary elections. For example, Quebec upheld strict spending limits in that province for the


11 12

13 14

15 16

17 18 19 20 21 22

23 24 25 26

27 28 29

express purpose of ensuring that the debate would not be dominated by wealthy interests. See Libman v. Quebec, 1997. I borrow this formulation from Rawls (1980). Citizens had little or no input into the process of constitution-framing, not even being asked to choose delegates to Philadelphia. At the ratification stage low voter turnout for delegates to state ratifying conventions is one possible indicator that citizens were not particularly engaged in the debate (Ackerman, 1998). For a stronger critique of the claim to speak for the people see Beard (1986), Derrida (1986), and Tully (1994). Others have also noted this trend toward citizen inclusion. See, for example, Mendelsohn (2000). The main difference in content is that all these constitutions contain a stronger recognition of diversity claims and most of them contain a stronger acknowledgement of social and economic rights. This is, at any rate, the larger argument I make in the book. See Ackerman (1998) (he argues that such a referendum would be in keeping with a long-standing role for the People in higher law-making). See also Amar (1988). Romer v. Evans, 1996. Weiser (1993). See Gamble (1997) for an argument that referendums are bad for minority rights and Baker (1991) for an opposite view. Page and Shapiro (1992). Nical (1997). Many people did not even have a word for constitution in their language and had only a vague idea of what a right might be. At one of the first public meetings, for example, people asked for things like the right to pay lobola (bride price) in cattle only (Nical, 1997). Hyden and Venter (forthcoming), Nical (1997). Ackerman (1993). Nical (1997). Michael Walzer notes, for example, that jurors are expected to be impartial and are given common criteria upon which to base their deliberation while citizens legitimately pursue their interests and passions and base deliberation on very different criteria. This is often but not always true. Walzer fails to recognize that different types of political decisions will call on different types of citizen deliberation. While much of day-to-day politics is not like jury deliberation (voting for President is not like convicting a criminal), some issues of justice do require citizens to deliberate impartially using similar criteria (Walzer, 1999, p. 63). In any case, my appeal to jury deliberation does not rest on the assumption that jury and political deliberation require similar moral perspectives. It is an argument regarding the way decision rules affect strategy. Saltzburg (1993), p. 362. Jürgen Habermas and Joshua Cohen make this argument, among others (Habermas, 1996, p. 178–79; Cohen, 1997). Amendment requires the support of two-thirds of both Houses (or a successful constitutional convention called for by two-thirds of the states) and then ratification by three-fourths of state legislatures (or by conventions in threefourths of the states).

254 30 31 32 33 34 35

36 37

38 39 40 41 42

43 44

45 46

47 48

Simone Chambers Manin (1987). I have tried to trace this dynamic in Chambers (1996). See Gutmann (1999). Dewey (1981). The sovereignty option was narrowly defeated by a 50.6 to 49.4 margin. For example, although the vast majority of Quebec Anglophones are against Quebec independence, the last fifteen years has seen a growing acceptance, understanding, and even sympathy for national aspirations. The majority of Anglophones in Quebec have supported the inclusion of a ‘distinct society’ clause in the constitution and a majority also agree that Quebec has a legitimate interest in protecting and maintaining French culture. For a more in-depth analysis of the dialogue between Anglophones and Francophones in Quebec see Chambers (1996). It was, however, an extremely level playing field with regard to campaign financing and spending, media access, and advertisement. While most faculty decisions fall very short of being ‘constitutional’ or fundamental in a moral sense and so we need not be concerned that participants fail to be discursive, some faculty decisions are about fundamental questions of justice and equity within the institution and do call for participants to rise above strategic attitudes. For a more detailed discussion of the deliberative potential of these conferences see Chambers (1998), pp. 143–72. See, for example, Russell (1993), p. 177. Why Charlottetown failed is a complicated not to mention disputed question. See Cooke (1994) and McRoberts and Monahan (1993). See especially James Tully in Cooke (1994) and Errol Mendes in McRoberts and Monahan (1993). The referendum was in principle non-binding as the existing constitution does not make provision for plebiscitary amendment. If passed, however, there would have been little choice but to entrench it. Conversely, it was not possible for government officials sponsoring the accord to overturn or ignore the outcome. Studies show that elites have particularly powerful influence on voters in referendums. See Butler and Ranney (1994, pp. 5–6). Another effect is that the more fundamental and irreversible a decision, the more conservative are voters, that is the less likely they are to change the status quo. Hans and Vidmor (1986). This is not always true of jury deliberations. For example, later public statements made by jurors to the effect they were unhappy with the verdict or process can undermine the perceived legitimacy of the outcome. Hyden and Venter (forthcoming). The process: political negotiations in 1993 result in the agreement on 34 basic principles as the framework for a national constitution. Democratic elections in 1994 indicate popular support for the constitutional project. The Constitution is then framed by a Constitutional Assembly in consultation with the public but also using closed-door sessions of intense negotiation. The resulting document is reviewed by a Constitutional Court which also takes submissions from the public (the court sends it back for more work) and is passed into law by the National Assembly.

Constitutional Referendums and Democratic Deliberation 255 49 Tully (1995), p. 154. 50 Ackerman’s proposal is very drawn out. A President can propose an Amendment only in his second term. Then it must be approved by Congress in a super majority and then, rather than going to the states for final ratification, it is placed on the ballot at the next two Presidential elections (Ackerman, 1991, p. 410). 51 See Setälä (1999b). 52 Both Canada and South Africa have used this mechanism during constitutional negotiation, although Canada did not use it at the time of the 1992 referendum.

Bibliography Aardal, Bernt and Anders Todal Jenssen (1995) ‘Opinionsutviklingen 1972– 1994’, in A. T. Jenssen and H. Valen (eds), Brussel midt imot. Oslo: Ad Notam Gyldendal. Abramson, Paul R. and Ronald Inglehart (1995) Value Change in Global Perspective. Ann Arbor: University of Michigan Press. Achen, Christopher H. (1975) ‘Mass Political Attitudes and the Survey Response,’ American Political Science Review, 69: 1218–37. Ackerman, Bruce (1998) We the People: Transformation. Cambridge, MA: Harvard University Press. Ackerman, Bruce (1993) We the People: Foundations. Cambridge, MA: Harvard University Press. Alvarez, Michael and Tara Butterfield (1998) The Revolution against Affirmative Action in California: Politics, Economic and Proposition 209. Paper presented at the Western Political Science Association meeting. Los Angeles. Amar, Akhil R. (1988) ‘Philadelphia Revisited: Amending the Constitution Outside Article V,’ University of Chicago Law Review, 55: 1043. Ansolabehere, Stephen, Roy L. Behr, and Shanto Iyengar (1991) ‘Mass Media and Elections: An Overview,’ American Politics Quarterly, 19: 109–39. Archbold, Claire (1999) ‘The Incorporation of Civic and Social Rights in Domestic Law.’ Ms prepared for the UN University Project on The United Nations System in the 21st Century: Human Rights, Princeton University. Baker, Lynn A. (1991) ‘Direct Democracy and Discrimination: A Public Choice Perspective,’ 67 Chicago-Kent Law Review, 67: 707–35. Banaszak, Lee Ann (1991) ‘The Influence of the Initiative on the Swiss and American Women’s Suffrage Movements,’ Annuaire suisse de science politique, 31: 187–207. Banaszak, Leeann (1996) Why Movements Succeed or Fail: Opportunity, Culture and the Struggle for Women’s Suffrage. Princeton, NJ: Princeton University Press. Banducci, Susan (1998) ‘Direct Legislation: When Is It Used and When Does It Pass?’ in S. Bowler, T. Donovan, and C. Tolbert (eds), Citizens as Legislators: Direct Democracy in the United States. Columbus: Ohio State University Press. Barber, Benjamin (1984) Strong Democracy: Participatory Politics for a New Age. Berkeley: University of California Press. Barker, Ernest (1942) Reflections on Government. London: Oxford University Press. Beard, Charles (1986) An Economic Interpretation of the Constitution of the United States. New York: Free Press. Beer, Samuel (1993) To Make a Nation: The Rediscovery of American Federalism. Cambridge: Cambridge University Press. Bell, Daniel (1960) The End of Ideology. New York: Random House. Bell, Derrick A. Jr (1978) ‘The Referendum: Democracy’s Barrier to Racial Equality,’ Washington Law Review, 54: 1–29. Bender, Earl and Sean Tenner (1999) ‘Guide to a Petition Circulator’s Legal Rights,’ Campaigns and Elections, 20: 61–3. 256

Bibliography

257

Benhabib, Seyla (1996) ‘Toward a Deliberative Model of Democratic Legitimacy,’ in S. Benhabib (ed.), Democracy and Difference: Contesting the Boundaries of the Political. Princeton, NJ: Princeton University Press. Bennett, Stephen (1998) ‘“Know-Nothings” Revisited: The Meaning of Political Ignorance Today,’ Social Science Quarterly, 69: 476–90. Bjørklund, Tor (1982) ‘The Demand for Referendum: When Does It Arise and when Does It Succeed?’ Scandinavian Political Studies, 5: 237–59. Bjørklund, Tor (1997) Om folkeavstemninger i Norge og Norden 1905–1994 (On Referendums in Norway and the Nordic Countries 1905–1994). Oslo: Scandinavian University Press. Black, Duncan (1958) Theory of Committees and Elections. Cambridge: Cambridge University Press. Bobbio, Norberto (1987) The Future of Democracy: A Defence of the Rules of the Game. Cambridge: Polity Press. Bogdanor, Vernon (1981) The People and the Party System: The Referendum and Electoral Reform in British Politics. New York: Cambridge University Press. Bogdanor, Vernon (1994) ‘Western Europe,’ in D. Butler and A. Ranney (eds), Referendums Around the World: the Growing Use of Direct Democracy. Washington, DC: American Enterprise Institute. Bohman, James (1996) Public Deliberation: Pluralism, Complexity, and Democracy. Cambridge, MA: MIT Press. Bohman, J. and W. Rehg (eds) (1997) Deliberative Democracy. Boston: MIT Press. Bonner, R. J. (1967) Aspects of Athenian Democracy. New York: Russell & Russell. Borg, Sami and Peter Esaiasson (1998) ‘Exposure to the Campaigns,’ in A. T. Jenssen, P. Pesonsen and M. Gilljam (eds), To Join or Not to Join: Three Nordic Referendums on Membership in the European Union. Oslo: Scandinavian University Press. Bowler, Shaun and Todd Donovan (1998) Demanding Choices: Opinion and Voting in Direct Democracy. Ann Arbor: University of Michigan Press. Bowler, Shaun and Todd Donovan (forthcoming) ‘Direct Democracy and Attitudes about Citizen Influence on Government,’ British Journal of Political Science. Bowler, Shaun, Todd Donovan, and Caroline J. Tolbert (eds) (1998) Citizens as Legislators: Direct Democracy in the United States. Columbus: Ohio State University Press. Bowler, Shaun, Todd Donovan, Max Neiman, and Johnny Peel (1999) Elite Attitudes about Direct Democracy. Paper presented at the Annual Meeting of the Western Political Science Association, Seattle. Boyer, Patrick (1992) Direct Democracy in Canada: The History and Future of Referendums. Toronto: Dundurn Press. Brady, Henry E. and Richard Johnston (1987) ‘What’s the Primary Message: Horse Race or Issue Journalism?’ in G. Orren and N. Polsby (eds), Media and Momentum: The New Hampshire Primary and Nomination Politics. Chatham, NJ: Chatham House: 127–86. Brady, Henry E. and Cynthia S. Kaplan (1994) ‘Eastern Europe and the Former Soviet Union,’ in D. Butler and A. Ranney (eds), Referendums Around the World. Washington: AEI Press. Brams, Steven D., Marc Kilgour, and William Zwicker (1997) ‘Voting on Referenda: The Separability Problem and Possible Solutions,’ Social Choice and Welfare, 16(3): 359–77.

258

Bibliography

Brams, Steven D., Marc Kilgour, and William Zwicker (1998) ‘The Paradox of Multiple Elections,’ Social Choice and Welfare, 15(2): 211–36. Brown, Michael (1981) ‘The Scottish Morning Press,’ in J. Bochel, D. Denver, and A. Macartney (eds), The Referendum Experience. Aberdeen: Aberdeen University Press. Brown, Roger (1986) Social Psychology: The Second Edition. New York: Free Press. Budge, Ian (1994) ‘A New Spatial Theory of Party Competition,’ British Journal of Political Science, 24: 443–67. Budge, Ian (1996) The New Challenge of Direct Democracy. Cambridge: Polity Press. Budge Ian, David Robertson, and D. J. Hearl (eds) (1987) Ideology, Strategy and Party Change. Cambridge: Cambridge University Press. Butler, David and Austin Ranney (eds) (1978) Referendums: A Comparative Study of Practice and Theory. Washington, DC: American Enterprise Institute. Butler, David and Austin Ranney (eds) (1994) Referendums around the World: The Growing Use of Direct Democracy. Washington, DC: American Enterprise Institute. Caciagli, M. and P. V. Uleri (eds) (1994) Democrazie e referendum. Laterza: RomeBari. Cain, Bruce (1992) ‘Voting Rights and Democratic Theory: Toward a Color-Blind Society?’ in B. Grofman and C. Davidson (eds), Controversies in Minority Voting. Washington, DC: Brookings. Camobreco, John F. (1998) ‘Preferences, Fiscal Policies, and the Initiative Process,’ Journal of Politics, 60: 819–29. Campbell, Angus, Philip E. Converse, Warren E. Miller, and Donald E. Stokes (1960) The American Voter. New York: Wiley. Campbell, Stuart L. (1978) The Second Empire Revisited: A Study in French Historiography. New Brunswick, NJ: Rutgers University Press. Cappella, Joseph N. and Kathleen Hall Jamieson (1997) Spiral of Cynicism: The Press and the Public Good. New York: Oxford University Press. Chambers, Simone (1996) Reasonable Democracy: Jürgen Habermas and the Politics of Discourse. Ithaca, NY: Cornell University Press, pp. 212–27. Chambers, Simone (1998) ‘Contract or Conversation? Theoretical Lessons from the Canadian Constitutional Debate,’ Politics and Society, 26(1): 143–72. Churchland, Patricia S. and Terrence J. Sejnowski (1992) The Computational Brain. Cambridge, MA: MIT Press. Citrin, Jack (1996) ‘Who’s the Boss: Direct Democracy and Popular Control of Government,’ in S. C. Craig (ed.), Broken Contract? Changing Relationships Between Americans and Their Government. Boulder, CO: Westview. Citrin, Jack, Beth Reingold, and Evelyn Walters (1990) ‘The “Official English” Movement and the Symbolic Politics of Language in the United States,’ Western Political Quarterly, 43: 553–60. Clark, Glenn (1998) ‘On the Offensive: BC Premier Glen Clark Responds to His Critics,’ Macleans, 111(44): 30. Cobban, Alfred (1961) A History of Modern France, Vol. Two: From the First Empire to the Fourth Republic. Harmondsworth: Penguin. Cohen, Joshua (1997) ‘Deliberation and Democratic Legitimacy,’ in J. Bohman and W. Rehg (eds), Deliberative Democracy. Boston: MIT Press. Constitutional Alteration (Establishment of Republic) (1999) A Bill for an Act to alter the Constitution to establish the Commonwealth of Australia as a republic with a President chosen by a two-thirds majority of the members of the Commonwealth Parliament. Exposure Draft.

Bibliography

259

Constitutional Alteration (Establishment of Republic) (1999) Explanatory Statement. Constitutional Commission (1988) Final Report, 2 vols. Canberra: AGPS. Converse, Jean M. and Stanley Presser (1986) Survey Questions: Handcrafting the Standardized Questionnaire. Newbury Park, CA: Sage. Converse, Philip E. (1964) ‘The Nature of Belief Systems in Mass Publics,’ in D. E. Apter (ed.), Ideology and Discontent. New York: Free Press. Converse, Philip E. (1972) ‘Change in the American Electorate,’ in A. Campbell and P. E. Converse (eds), The Human Meaning of Social Change. New York: Russell Sage. Cooke, Curtis (ed.) (1994) Constitutional Predicament: Canada After the Referendum of 1992. Montreal: McGill-Queen’s Press. Corry, J. A. (1978) ‘The Uses of a Constitution,’ in The Constitution and the Future of Canada, Special Lectures of the Law Society of Upper Canada. Toronto: Richard De Boo: 9–10. Coté, P. F. (1999) ‘Québec’s Experiences with Referendums,’ for Conference on Maturity or Malaise? The Growing Use of Referendums in Liberal-Democratic Societies. Queen’s University, Canada. Craig, Stephen C. (1980) ‘The Mobilization of Political Discontent,’ Political Behavior, 2(2): 189–209. Craig, Stephen C. (1993) The Malevolent Leaders: Popular Discontent in America. Boulder, CO: Westview. Craig, Stephen C., Richard G. Niemi, and Glenn E. Silver (1990) ‘Political Efficacy and Trust: A Report on the NES Pilot Study Items,’ Political Behavior, 12: 289–314. Crawford, Vincent and Joel Sobel (1982) ‘Strategic Information Transmission,’ Econometrica, 50: 1431–51. Cronin, Thomas (1989) Direct Democracy: The Politics of Initiative, Referendum and Recall. Cambridge, MA: Harvard University Press. Crook, Malcolm (1998) Napoleon Comes to Power: Democracy and Dictatorship in Revolutionary France, 1795–1804. Cardiff: University of Wales Press. Dahl, Robert (1956) Preface to Democratic Theory. Chicago: University of Chicago Press. Dahl, Robert (1961) Who Governs? Democracy and Power in An American City. New Haven, CT: Yale University Press. Dalton, Russell (1988) Citizen Politics in Western Democracies: Public Opinion and Political Parties in the United States, Great Britain, West Germany and France. Chatham, NJ: Chatham House. Dalton, Russell (1996) Citizen Politics: Public Opinion and Political Parties in Advanced Democracies, 2nd edition. Chatham, NJ: Chatham House. Delley, Jean-Daniel (1978) L’initiative populaire en Suisse. Mythe et réalité de la démocratie directe. Lausanne: Réalités Sociales. Delli Carpini, Michael X. and Scott Keeter (1991) ‘Stability and Change in the United States Public’s Knowledge of Politics,’ Public Opinion Quarterly, 55: 583–612. Delli Carpini, Michael X. and Scott Keeter (1996) What Americans Know About Politics and Why It Matters. New Haven, CT: Yale University Press. Derrida, Jacques (1986) ‘Declaration and Independence,’ New Political Science, 15: 7–15. Derriennic, Jean-Pierre (1995) Nationalisme et Democratie. Montreal: Boreal. Dewey, John (1981) ‘The Public and Its Problems,’ in J. Boydston (ed.), The Later Works. Carbondale, Illinois: Southern Illinois University Press.

260

Bibliography

Donovan, Todd and Shaun Bowler (1997) ‘Direct Democracy and Minorities: Opinions on Anti-Gay Rights Initiatives,’ in S. Witt and S. McCorkle (eds), Anti-Gay Rights: Assessing Voter Initiatives. Westport, CT: Praeger. Donovan, Todd and Shaun Bowler (1998a) ‘An Overview of Direct Democracy in the American States,’ in S. Bowler, T. Donovan, and C. Tolbert (eds), Citizens as Legislators: Direct Democracy in the United States. Columbus: Ohio State University Press. Donovan, Todd and Shaun Bowler (1998b) ‘Direct Democracy and Minority Rights: An Extension,’ American Journal of Political Science, 42: 1020–5. Donovan, Todd and Shaun Bowler (1998c) ‘Responsive or Responsible Politics?’ in S. Bowler, T. Donovan, and C. Tolbert (eds), Citizens as Legislators: Direct Democracy in the United States. Columbus: Ohio State University Press. Donovan, Todd, Shaun Bowler, David McCuan, and Ken Fernandez (1998) ‘Contending Players and Strategies: Opposition Advantages in Initiative Elections,’ in S. Bowler, T. Donovan and C. Tolbert (eds), Citizens as Legislators: Direct Democracy in the United States. Columbus: Ohio State University Press. Donovan, Todd, Jim Wenzel, and Shaun Bowler (1999) ‘Direct Democracy and Gay Rights after Romer,’ in C. Rimmerman, K. Wald, and C. Wilcox (eds), The Politics of Gay Rights. Chicago: University of Chicago Press. Dryzek, John (1996) Democracy in Capitalist Times: Ideals, Limits, and Struggles. New York: Oxford University Press. Dubois, Philip and Floyd Feeney (1992) Improving the California Initiative Process. Berkeley: California Policy Seminar. Duverger, M. (1951) Les partis politiques. Paris: Colin. Elster, Jon (1989) ‘The Market and the Forum,’ in J. Elster and A. Hyland (eds), Foundations of Social Choice Theory. New York: Cambridge University Press. Erikson, Robert S., Gerald C. Wright, John P. McIver (1993) Statehouse Democracy: Public Opinion and Policy in the American States. New York: Cambridge University Press. Eule, Julian (1990) ‘Judicial Review of Direct Democracy,’ Yale Law Review, 99: 1504. Farley, R. Vick (1998) Memorandum on Constitutional Amendment Approval Act and the Nisga’a Final Agreement. Victoria: British Columbia Ministry of the AttorneyGeneral, 24 July. Fearon, James (1991) ‘Counterfactuals and Hypothesis Testing in Political Science,’ World Politics, 43(2): 169–95. Feld, Lars P. (1997) ‘Exit, Voice and Income Taxes: The Loyalty of Voters,’ European Journal of Political Economy, 13(3): 455–78. Feld, Lars P. and Marcel R. Savioz (1997) ‘Direct Democracy Matters for Economic Performance: An Empirical Investigation,’ Kyklos, 50(4): 507–38. Ferejohn, John et al. (1995) ‘Is It Too Easy to Amend the Constitution?’ in B. Cain and R. Noll (eds), Constitutional Reform in California. Berkeley, CA: IGS Press. Fishkin, James (1991) Deliberation and Democracy: New Directions for Democratic Reform. New Haven, CT: Yale University Press. Fowler, John (1981) ‘Broadcasting: Television,’ in J. Bochel, D. Denver, and A. Macartney (eds), The Referendum Experience. Aberdeen: Aberdeen University Press: 121–35. Franklin, Mark, Tom Mackie, and Henry Valen et al. (1992) Electoral Change: Responses to Evolving Social and Attitudinal Structures in Western Countries. Cambridge: Cambridge University Press.

Bibliography

261

Franklin, Mark, Michael Marsh, and Lauren McLaren (1994) ‘Uncorking the Bottle: Popular Opposition to European Unification in the Wake of Maastricht,’ Journal of Common Market Studies, 32(4): 455–72. Franklin, Nicola E. (1992) ‘Initiative and Referendum: Participatory Democracy of Rolling Back the State,’ in M. M. Clark (ed.), Citizen Participation in Government. Sydney: Hale & Iremonger. Frey, Bruno and Lorenz Goette (1998) ‘Does the Popular Vote Destroy Civil Rights?’ American Journal of Political Science, 42: 1343–8. Frey, Bruno and Alois Stutzer (2000) ‘Happiness, Economy and Institutions.’ Economic Journal, 110(466): 918–38. Gallagher, Michael and Pier Vincenzo Uleri (eds) (1996) The Referendum Experience in Europe. London: Macmillan Press (now Palgrave). Galligan, Brian (1995) A Federal Republic: Australia’s Constitutional System of Government. Cambridge: Cambridge University Press. Galligan, Brian (1998) ‘The Constitutional Convention I: The Republic Model,’ Quadrant, 42(4): 17–21. Galligan, Brian (1999) ‘The Republican Referendum: A Defence of Popular Sense,’ Quadrant, 43: 46–52. Gamble, Barbara (1997) ‘Putting Civil Rights to a Popular Vote,’ American Journal of Political Science, 41: 245–69. Gamson, William (1968) Power and Discontent. Homewood, IL: Dorsey. Gerber, Elizabeth (1995) ‘Reforming the California Initiative Process: A proposal to Increase Flexibility and Legislative Accountability,’ in B. Cain and R. Noll (eds), Constitutional Reform in California. Berkeley, CA: IGS Press. Gerber, Elizabeth (1996) ‘Legislative Responsiveness to the Threat of Popular Initiatives,’ American Journal of Political Science, 40: 99–128. Gerber, Elisabeth (1998) ‘Pressuring Legislatures Through the Use of Initiatives. Two Forms of Indirect Influence,’ in S. Bowler, T. Donovan, and C. Tolbert (eds), Citizens as Legislators: Direct Democracy in the United States. Columbus: Ohio State University Press, pp. 191–208. Gerber, Elisabeth (1999) The Populist Paradox: Interest Group Influence and the Promise of Direct Legislation. Princeton, NJ: Princeton University Press. Gerber, Elisabeth and Simon Hug (1999) Minority Rights and Direct Legislation. Theory, Methods, and Evidence. Manuscript. La Jolla: University of California. Germann, Raimund (1985) Experts et commissions de la confédération. Lausanne: Presses Polytechniques Romandes. Giles, Michael and Melanie Buckner (1993) ‘Percent Black and Racial Hostility: an Old Assumption Re-examined,’ Social Science Quarterly, 58: 848–65. Giles, Michael and Arthur Evans (1986) ‘The Power Approach to Intergroup Hostility,’ Journal of Conflict Resolution, 30: 469–86. Gillette, Clayton P. (1988) ‘Plebiscites, Participation, and Collective Action in Local Government Law,’ Michigan Law Review, 86: 930–88. Gilljam, Mikael (1996) ‘Det kluvna Sverige?’ (‘Sweden divided?’), in M. Gilljam and S. Holmberg (eds), Ett knapt ja till EU. (A Small Majority for EU). Stockholm: Norstedts Juridik. Gilljam, Mikael, Pertti Pesonen, and Ola Listhaug (1998) ‘The Referendum in Representative Democracies,’ in A. T. Jenssen, P. Pesonen, and M. Gilljam (eds), To Join or Not to Join: Three Nordic Referendums on Membership in the European Union. Oslo: Scandinavian University Press.

262

Bibliography

Glazer, Nathan (1978) ‘Individual Rights Against Group Rights,’ in A. Tay and E. Kamenka (eds), Human Rights. London: Edward Arnold, pp. 87–103. Glazer, Nathan (1983) Ethnic Dilemmas: 1964–1982. Cambridge, MA: Harvard University Press. Goodin, R. (1989) ‘Laundering Preferences,’ in Foundations of Social Choice Theory. Cambridge: Cambridge University Press, pp. 75–102. Greer, John (1996) From Tea Leaves to Opinion Polls: A Theory of Democratic Leadership. New York: Columbia University Press. Gutmann, Amy (1999) ‘How Not to Resolve Moral Conflicts in Politics,’ Ohio State Journal of Dispute Resolution, 15(1): 1–22. Gutmann, Amy and Dennis Thompson (1996) Democracy and Disagreement. Cambridge, MA: Harvard University Press. Habermas, Jürgen (1984) The Theory of Communicative Action. Boston: Beacon Press. Habermas, Jürgen (1989) Structural Transformation of The Public Sphere, trans. Thomas Berger. Cambridge, MA: MIT press. Habermas, Jürgen (1993) Justification and Application: Remarks on Discourse Ethics, trans. Ciaran P. Cronin. Cambridge, MA: MIT Press. Habermas, Jürgen (1996) Between Facts and Norms: Contributions to a Discourse Theory of Law and Democracy. Cambridge, MA: MIT Press. Hackett, Robert A. and Yuezhi Zhao (1998) Sustaining Democracy? Journalism and the Politics of Objectivity. Toronto: Garamond. Hahn, Harlan and Sheldon Kamieniecki (1987) Referendum Voting: Social Status and Policy Preferences. Westport, Conn: Greenwood Press. Hallin, Daniel C. (1992) ‘Sound Bite News: Television Coverage of Elections, 1968–1988,’ Journal of Communication, 42(2): 5–23. Hans, Valerie P. and Neil J. Vidmor (1986) Judging the Jury. New York: Plenum. Hastie, R., S. Penrod, and N. Pennington (1983) Inside the Jury. Cambridge, MA: Harvard University Press. Hazareesingh, Sudhir (1998) From Subject to Citizen: The Second Empire and the Emergence of Modern French Democracy. Princeton, NJ: Princeton University Press. Hero, Rodney (1998) Faces of Inequality: Social Diversity in American Politics. New York: Oxford University Press. Hjelseth, Arve (1995) ‘Partienes EU-strategi’, Working Paper No. 6. The Referendum Project. Department of Sociology and Political Science, Norwegian University of Science and Technology. Hofer, R. (1987) ‘Die Volksinitiative als Verhandlungspfand,’ Annuaire suisse de science politique, 27: 207–36. Hoffmann-Nowotny, Hans-Joachim (1985) ‘Switzerland,’ in Tomas Hammar (ed.), European Immigration Policy: A Comparative Study. New York: Cambridge University Press. Hogg, Peter (1998) Letter to the Deputy Minister Attorney-General, Re: Constitutional Amendment Approval Act and Nisga’a Final Agreement. Victoria: British Columbia Ministry of the Attorney General, 29 July. Holmberg, Sören (1996) ‘Partierna gjorde så gott de kunde’, in M. Gilljam and S. Holmberg (eds), Ett knapt ja till EU. Stockholm: Norstedts Juridik. Holmes, Stephen (1990) ‘Liberal Constraints on Private Power? Reflections on the Origins and Rational of Access Regulation,’ in J. Lichtenberg (ed.), Democracy and the Mass Media. Cambridge: Cambridge University Press: 26.

Bibliography

263

Hovland, Carl I., Irving L. Janis, and Harold H. Kelley (1953) Communication and Persuasion: Psychological Studies of Opinion Change. New Haven, CT: Yale University Press. Hug, Simon (1999) Occurrence and Policy Consequences of Referendums. A Theoretical Model and Empirical Evidence. La Jolla: University of California. Hug, Simon and Pascal Sciarini (2000) ‘Referendums on European Integration. Do Institutions Matter in the Voters’ Decision?’ Comparative Political Studies, 33: 3–36. Hughes, Colin (1983) ‘Commonwealth Constitution: Methods of Initiating Amendments,’ Appendix 3 to Constitutional Amendment Sub-Committee’s Report to Standing Committee, June 1984, Brisbane 1985, vol. 1. Hyden, Goran and Denis Venter (forthcoming) ‘Introduction,’ in G. Hyden and D. Venter (eds), Constitution-Making in Africa. Pretoria: Oxford University Press. Immergut, Ellen M. (1992) Health Politics: Interests and Institutions in Western Europe. Cambridge: Cambridge University Press. Inglehart, Ronald (1977) The Silent Revolution: Changing Values and Political Styles Among Western Publics. Princeton, NJ: Princeton University Press. Inglehart, Ronald (1990) Culture Shift in Advanced Industrial Society. Princeton, NJ: Princeton University Press. Inglehart, Ronald (1997) ‘Postmaterialist Values and the Erosion of Institutional Authority,’ in J. S. Nye, Jr, P. D. Zelikow, and D. C. King (eds), Why People Don’t Trust Government. Cambridge, MA: Harvard University Press. Iyengar, Shanto (1991) Is Anyone Responsible? How Television Frames Political Issues. Chicago: Chicago University Press. Jahn, Detlef, Pertti Pesonen, Tore Slaatta and Leif Åberg (1998) ‘The Actors and the Campaigns,’ in A. T. Jenssen, P. Pesonen, and M. Gilljam (eds), To Join or Not to Join: Three Nordic Referendums on Membership in the European Union. Oslo: Scandinavian University Press. Jahn, Detlef and Ann-Sofie Storsved (1995) ‘Legitimacy through Referendum? The Nearly Successful Domino-Strategy of the EU-Referendums in Austria, Finland, Sweden and Norway,’ West European Politics, 18(3): 18–37. Jamieson, K. H. (1992) Dirty Politics. New York: Oxford. Jenkins, Richard W. (1999a) ‘How Much Is Too Much? Media Attention and Popular Support for an Insurgent Party,’ Political Communication, 16(4): 429–45. Jenkins, Richard W. (1999b) How Campaigns Matter: Priming, Persuasion and Information in the 1993 Campaign. Paper presented at the Annual Meeting of the Canadian Political Science Association, Sherbrooke. Jenssen, Anders Todal (1995) ‘Ouverturen,’ in A. T. Jenssen and H. Valen (eds), Brussel midt imot. Oslo: Ad Notam Gyldendal. Jenssen, Anders Todal, Ola Listhaug, and Per Arnt Pettersen (1995) ‘Betydningen av gamle og nye skiller,’ in A. T. Jenssen and H. Valen (eds), Brussel midt imot (Opposing Brussels). Oslo: Ad Notam Gyldendal. Jenssen, Anders Todal, Kathrine L. Moen, Inge Ramberg, Kristen Ringdal, and Arve Østgård (1994) Medlemsundersøkelsen. Holdninger og sosial bakgrunn blant medlemmer i Nei til EU, Europabevegelsen og Sosialdemokrater mot EU. Dokumentasjon. Department of Sociology and Political Science, Norwegian University of Science and Technology. Jenssen, Anders Todal, Pertti Pesonen, and Mikael Gilljam (eds) (1998) To Join or Not to Join: Three Nordic Referendums on Membership in the European Union. Oslo: Scandinavian University Press.

264

Bibliography

Jenssen, Anders Todal and Henry Valen (eds) (1995) Brussel midt imot. Oslo: Ad Notam Gyldendal. Johnston, Richard (1993) ‘An Inverted Logroll: The Charlottetown Accord and the Referendum,’ PS: Political Science and Politics, 26: 43–8. Johnston, Richard, André Blais, Elisabeth Gidengil, and Neil Nevitte (1996) The Challenge of Direct Democracy: The 1992 Canadian Referendum. Montreal: McGillQueen’s University Press. Kaase, Max, and Kenneth Newton (1995) Beliefs in Government. Oxford: Oxford University Press. Karp, Jeffery (1998) ‘The Influences of Elite Endorsements in Initiative Campaigns,’ in S. Bowler, T. Donovan, and C. Tolbert (eds), Citizens as Legislators: Direct Democracy in the United States. Columbus: Ohio State University Press. Key, V. O. (1949) Southern Politics in State and Nation. New York: Knopf. Key, V. O. and Winston Crouch (1939) The Initiative and the Referendum in California. Berkeley: University of California Press. Kiewiet, D. Roderick and Kristin Szakaly (1996) ‘Constitutional Limitations on Borrowing: An Analysis of State Bonded Indebtedness,’ Journal of Law, Economics and Organization, 12(1): 62–97. Killias, Martin (1997) ‘Immigrants, Crime, and Criminal Justice in Switzerland,’ in M. Tonry (ed.), Ethnicity, Crime, and Immigration: Comparative and Cross-national Perspectives. Chicago: University of Chicago Press. Kinder, Donald (1986) ‘The Continuing American Dilemma: White Resistance to Racial Change 40 Years After Myrdal,’ Journal of Social Issues, 42: 151–71. King, Gary, Robert O. Keohane, and Sidney Verba (1994) Designing Social Inquiry. Princeton, NJ: Princeton University Press. Kircheimer, Otto (1966) ‘The Transformation of Western European Party Systems,’ in J. La Palombara and M. Weimer (eds), Political Parties and Political Development. Princeton, NJ: Princeton University Press. Kirchgässner, Gebhard, Lars P. Feld, and Marcel R. Savioz (1999) Die direkte Demokratie der Schweiz: Modern, erfolgreich, entwicklungs- und exportfähig. Basel: Helbing & Lichtenhahn. Klingemann, H. D., R. Hofferbert, Ian Budge et al. (1994) Parties, Policies and Democracy. Boulder, CO: Westview. Kobach, Kris (1993) ‘Switzerland,’ in D. Butler and A. Ranney (eds), Referendums Around the World. Washington, DC: American Enterprise Institute. Kornhauser, A. (1960) The Politics of Mass Society. London: Routledge & Kegan Paul. Kriesi, H. (1994) ‘La lunga e complessa vicenda della Confederazione Elvetica,’ in M. Caciagli and P. V. Uleri (eds), Democrazie e referendum. Laterza: Rome-Bari, pp. 63–78. Kriesi, Hanspeter, Rene Levy, Gilbert Ganguillet, and Heinz Zwicky (1981) Politische Aktivierung in der Schweiz, 1945–1978. Diessenhofen: Rüegger. Krosnick, Jon A. (1999) ‘Maximizing Questionnaire Quality,’ in J. P. Robinson, P. K. Shaver, and L. S. Wrightsman (eds), Measures of Political Attitudes. San Diego: Academic Press. Kukanthas, Chandran (1992) ‘Are There Any Cultural Rights?’ Political Theory, 20(1): 105–39.

Bibliography

265

Kuklinski, James H., Paul Sniderman, Kathleen Knight, Thomas Piazza, and others (1997) ‘Racial Prejudice and Attitudes Toward Affirmative Action,’ American Journal of Political Science, 41: 402–20. Kymlicka, Will (1989) Liberalism, Community, and Culture. Oxford: Oxford University Press. Kymlicka, Will (1995) Multicultural Citizenship. Oxford: Clarendon. Kymlicka, Will (1996) ‘Three Forms of Group-Differentiated Citizenship in Canada,’ in S. Benhabib (ed.), Democracy and Difference. Princeton, NJ: Princeton University Press, pp. 153–70. Lascher, Edward L., Michael Hagen, and Steven Rochlin (1996) ‘Gun Behind the Door – Ballot Initiatives, State Policies and Public Opinion,’ Journal of Politics, 58(3): 760–75. Lichter, S. R. and Richard Noyes (1995) Good Intentions Make Bad News: Why Americans Hate Campaign Journalism. Lanham, MD: Rowman & Littlefield. Lichter, S. R. and Ted Smith (1996) ‘Why Elections Are Bad News: Media and Candidate Discourse in the 1996 Presidential Primaries,’ Press/Politics, 1(4): 15–35. Lijphart, A. (1984) Democracies. New Haven, CT: Yale University Press. Linde, Hans (1993) ‘When Initiative Law Making is not Republican Government: The Campaign against Homosexuality,’ Oregon Law Review, 72: 20–39. Linder, Wolf (1994) Swiss Democracy: Possible Solutions to Conflict in Multicultural Societies. New York: St. Martin’s Press (now Palgrave). Lipset, Seymour Martin and Stein Rokkan (eds) (1967) Party Systems and Voter Alignments: Cross-National Perspectives. New York: Free Press. Listhaug, Ola, Sören Holmberg, and Risto Sänkiaho (1998) ‘Partisanship and EU Choice,’ in A. T. Jenssen, P. Pesonen, and M. Gilljam (eds), To Join or Not to Join: Three Nordic Referendums on Membership in the European Union. Oslo: Scandinavian University Press. Listhaug, Ola and Matti Wiberg (1995) ‘Confidence in Political and Private Institutions,’ in H. Klingemann and D. Fuchs (eds), Citizens and the State. Oxford: Oxford University Press. Lupia, Arthur (1992) ‘Busy Voters, Agenda Control, and the Power of Information,’ American Political Science Review, 86: 390–403. Lupia, Arthur (1994) ‘Shortcuts versus Encyclopedias: Information and Voting Behavior in California Insurance Reform Elections,’ American Political Science Review, 88(1): 63–76. Lupia, Arthur and Mathew D. McCubbins (1998) The Democratic Dilemma: Can Citizens Learn What They Need to Know? New York: Cambridge University Press. Lupia, Arthur and Mathew D. McCubbins (2000) ‘The Institutional Foundations of Political Competence: How Citizens Learn What They Need to Know,’ in A. Lupia, M. D. McCubbins, and S. L. Popkin (eds), Elements of Reason: Cognition, Choice, and the Bounds of Rationality. New York: Cambridge University Press. Luskin, Robert C. (1987) ‘Measuring Political Sophistication,’ American Journal of Political Science, 31: 856–99. Lyons, Martyn (1975) France Under the Directory. Cambridge: Cambridge University Press. McCuan, David, Shaun Bowler, Todd Donovan, and Ken Fernandez (1998) ‘California’s Political Warriors: Campaign Professionals and the Initiative

266

Bibliography

Process,’ in S. Bowler, T. Donovan, and C. J. Tolbert (eds), Citizens as Legislators: Direct Democracy in the United States. Columbus: Ohio State University Press. McGarvie, Richard E. (1999) Democracy: Choosing Australia’s Republic. Melbourne: Melbourne University Press. MacIntyre, Alasdair (1984) After Virtue: A Study in Moral Theory, 2nd edn. Notre Dame, IN: University of Notre Dame Press. Macpherson, C. B. (1953) Democracy in Alberta: Social Credit and the Party System, 2nd edn. Toronto: University of Toronto Press. McRoberts, Kenneth and Monahan, Patrick (eds) (1993) The Charlottetown Accord, the Referendum, and the Future of Canada. Toronto: University of Toronto Press. Madison, James ([1787]1961) ‘Federalist #10 and #51,’ The Federalist Papers. New York: American New Library of World Literature. Madison, James, Alexander Hamilton, and John Jay (1987) The Federalist Papers. New York: Penguin Books. Magleby, David (1984) Direct Legislation: Voting on Ballot Propositions in the United States. Baltimore, MD: Johns Hopkins University Press. Magleby, David (1994) ‘Direct Legislation in the American States,’ in D. Butler and A. Ranney (eds), Referendums Around the World. Washington: American Enterprise Institute, pp. 218–57. Mair, Peter (1992) ‘La trasformazione del partito di massa in Europa,’ in M. Calise (ed.), Come cambiano i partiti. Bologna: Il Mulino: 99–122. Manin, Bernard (1987) ‘On Legitimacy and Political Deliberation,’ Political Theory, 15: 338–68. Marcus, G. E. and R. L. Hanson (eds) (1993) Reconsidering the Democratic Public. University Park, PA: University of Pennsylvania Press. Matsusaka, John (1992) ‘Economics of Direct Legislation,’ Quarterly Journal of Economics, 107: 541–71. Matsusaka, John (1995) ‘Fiscal Effects of the Voter Initiative: Evidence from the last 30 Years,’ Journal of Political Economy, 103: 587–623. Matsusaka, John (1998) ‘Fiscal Effects of the Voter Initiative in the First Half of the Twentieth Century.’ Manuscript, Marshall School of Business, University of Southern California. Matsusaka, John (1999) Problems with a Methodology Used to Test whether Policy is More or Less Responsive to Public Opinion in States with Voter Intiatives. Marshall School of Business, University of Southern California. Matsusaka, John (2000) ‘Fiscal Effects of the Voter Initiative in the First Half of the Twentieth Century,’ Journal of Law and Economics, 43: 619–44. Matsusaka, John and Nolan M. McCarty (1998) Political Resource Allocation: The Benefits and Costs of Voter Initiatives. Marshall School of Business, University of Southern California. Mendelsohn, Matthew (1996) ‘Introducing Deliberative Democracy in Canada: Learning from the American Experience,’ American Review of Canadian Studies, 26: 449–68. Mendelsohn, Matthew (2000) ‘Public Brokerage: Constitutional Reform and Accommodation of Mass Publics,’ Canadian Journal of Political Science, 33(2): 245–73. Mendelsohn, Matthew and Fred Cutler (2000) ‘The Effect of Referendums on Democratic Citizens: Information, Politicization, Efficacy and Tolerance,’ British Journal of Political Science, 30(4): 669–98.

Bibliography

267

Mendes, Errol (1993) Sinking Again into the Quagmire of Conflicting Visions, Groups, Underinclusion, and Death by Referendum, in K. McRoberts and P. Monahan (eds), The Charlottetown Accord, the Referendum, and the Future of Canada. Toronto: University of Toronto Press. Miller, David (1992) ‘Deliberative Democracy and Social Choice,’ Political Studies, XL: p. 6. Miller, Kenneth P. (1999) The Role of the Courts in the Initiative Process: A Search for Standards. Paper presented at the meeting of the American Political Science Association, Atlanta. Miller, Warren E., Donald R. Kinder, Steven J. Rosenstone, and the National Election Studies (1993) American National Election Study: Pooled Senate Election Study 1988, 1990, 1992. Ann Arbor, MI: Inter-University Consortium for Political and Social Research. Mishler, William and Richard Rose (1997) ‘Trust, Distrust and Skepticism: Popular Evaluations of Civil and Political Institutions in Post-Communist Societies,’ Journal of Politics, 59: 418–51. Moen, Kathrine L., Irene Øyangen, and Anders Todal Jenssen (1996) The 1994 Nordic EU Referendum Study. Study Description and Codebook. Department of Sociology and Political Science, Norwegian University of Science and Technology. Monière, Denis and Jean H. Guay (1996) La bataille du Québec: Troisième épisode: 30 jours qui ébranlèrent le Canada. Quebec: Fides. Morel, Laurence (1996) ‘France: Towards a Less Controversial Use of the Referendum?’ in M. Gallagher and P. V. Uleri (eds), The Referendum Experience in Europe. London: Macmillan (now Palgrave). Morel, Laurence (1998) ‘Referendums and the Evolution of Party Government in Liberal Democracies,’ European Review, 6: 2. Morton, F. L. (1985) ‘Group Rights versus Individual Rights in the Charter,’ in N. Nevitte and A. Kornberg (eds), Minorities and the Canadian State. Oakville, ON: Mosaic Press, pp. 71–84. Moser, Peter (1996) ‘Why is Swiss Politics so Stable?’ Zeitschrift für Volkswirtschaft und Statistik, 132(1): 31–60. Mueller, Dennis C. (1989) Public Choice II. Cambridge: Cambridge University Press. Mughan, Anthony (1995) ‘Television and Presidentialism: Australian and U.S. Legislative Elections Compared,’ Political Communication, 12: 327–42. Nathan, Howard (1998) ‘The Constitutional Convention II: The Convention from the Inside,’ Quadrant, 42(4): 21–5. Neidhart, Leonard (1970) Plebiszit und pluralitäre Demokratie. Bern: Francke. Neumann, W. Russell (1986) The Paradox of Mass Politics: Knowledge and Opinion in the American Electorate. Cambridge, MA: Harvard University Press. Nevitte, Neil (1996) The Decline of Deference: Canadian Value Change in CrossNational Perspective. Peterborough, ON: Broadview Press. Newell, Allen (1990) Unified Theories of Cognition. Cambridge, MA: Harvard University Press. Nical, Mike (1997) The Making of the Constitution: The Story of South Africa’s Constitutional Assembly, May 1994 to December 1996. Cape Town: Churchill Murray Publication. Nie, Norman, Jane Junn, and Kenneth Stehlik-Barry (1996) Education and Democratic Citizenship in America. Chicago: University of Chicago Press.

268

Bibliography

Nilsen, Tonje Raaen (1996) ‘Elitedominans i EU-strid? Et innholdsanalytisk perspektiv på NRK og TV2s nyhetsdekning av EU-striden 24.10.–27.11.94,’ Working Paper No. 9, The Referendum Project. Department of Sociology and Political Science, Norwegian University of Science and Technology. Nozick, Robert (1974) Anarchy, State and Utopia. New York: Basic Books. Nye, Joseph S., Jr (1997) ‘Introduction: The Decline of Confidence in Government,’ in J. S. Nye, Jr, P. D. Zelikow, and D. C. King (eds), Why People Don’t Trust Government. Cambridge, MA: Harvard University Press. Oskarson, Maria and Kristen Ringdal (1998) ‘The Arguments,’ in A. T. Jenssen, P. Pesonen, and M. Gilljam (eds), To Join or Not to Join: Three Nordic Referendums on Membership in the European Union. Oslo: Scandinavian University Press. Page, Benjamin I. and Robert Y. Shapiro (1992) The Rational Public: Fifty Years of Trends in America’s Policy Preferences. Chicago: University of Chicago Press. Papadopoulos, Yannis (1995) ‘Analysis of Functions and Dysfunctions of Direct Democracy: Top-Down and Bottom-Up Perspectives,’ Politics and Society, 23: 421–48. Papadopoulos, Yannis (1998) Démocratie directe. Paris: Economica. Parodi, Jean-Luc (1973) La Cinquième République et le système majoritaire. Doctoral thesis: Paris, Institut d’Etudes Politiques. Patterson, Thomas E. (1993) Out of Order. New York: Knopf. Pettersen, Per Arnt, Anders Todal Jenssen, and Ola Listhaug (1996) ‘The 1994 EU Referendum in Norway: Continuity and Change,’ Scandinavian Political Studies, 19: 257–81. Pierce Roy, Henry Valen, and Ola Listhaug (1983) ‘Referendum Voting Behavior: The Norwegian and British Referenda on Membership in the European Community,’ American Journal of Political Science, 27: 43–63. Pinker, Steven (1998) How the Mind Works. New York: Norton. Pommerehne, Werner W. (1978) ‘Institutional Approaches to Public Expenditure: Empirical Evidence from Swiss Municipalities,’ Journal of Public Economics, 9: 255–80. Pomper, Gerald (ed.) (1980) Party Renewal in America. New York: Praeger. Price, Vincent and John Zaller (1993) ‘Who Gets the News? Alternative Measures of News Reception and Their Implications for Research,’ Public Opinion Quarterly, 57: 133–64. Putnam, Robert (1976) The Comparative Study of Political Elites. Englewood Cliffs, NJ: Prentice-Hall. Quermonne, Jean-Louis (1985) ‘Le référendum: essai de typologie prospective,’ Revue du droit public et de la science politique en France et à l’étranger, May–June: 577–90. Qvortrup, Mads (1999) Constitutional Implications of the Use of the Referendum. DPhil. thesis: University of Oxford. Ramberg, Inge (1995) ‘EU-striden i tabloidformat,’ Working Paper No. 7, The Referendum Project. Department of Sociology and Political Science, Norwegian University of Science and Technology. Rappard, William E. (1912) ‘The Initiative, Referendum and Recall in Switzerland,’ American Political Science Review, 6(3): 345–66. Rawls, John (1980) ‘Kantian Constructivism in Moral Theory,’ Journal of Philosophy, 77: 515–72. Rawls, John (1993) Political Liberalism. New York: Columbia University Press.

Bibliography

269

Ray, Leonard (1999) ‘Conversion, Acquiescence, or Delusion: The Contingent Nature of the Party-Voter Connection,’ Political Behavior, 21: 325–47. Riker, W. H. (1982) Liberalism against Populism. San Francisco: Freeman. Ringdal, Kristen and Henry Valen (1998) ‘Structural Divisions in the EU Referendums,’ in A. T. Jenssen, P. Pesonen, and M. Gilljam (eds), To Join or Not to Join: Three Nordic Referendums on Membership in the European Union. Oslo: Scandinavian University Press. Robertson, David (1976) A Theory of Party Competition. New York: Wiley. Robinson, Gertrude (1998) Constructing the Quebec Referendum: French and English Media Voices. Toronto: University of Toronto Press. Robinson, Michael J. and Margaret A. Sheehan (1983) Over the Wire and on TV: CBS and UPI in Campaign ’80. New York: Russell Sage. Rokkan, Stein (1967) ‘Geography, Religion and Social Class: Crosscutting Cleavages in Norwegian Politics,’ in S. M. Lipset and S. Rokkan (eds), Party Systems and Voter Alignments. New York: Free Press. Rokkan, Stein (1970) Citizens, Elections, Parties. Oslo: Universitetsforlaget. Rokkan, Stein and Henry Valen (1964) ‘Regional Contrasts in Norwegian Politics,’ in E. Allardt and Y. Littunen (eds), Cleavages, Ideologies and Party Systems. Helsinki: Westermarck Society. Rokkan, Stein and Henry Valen (1970) ‘The Mobilization of the Periphery,’ in S. Rokkan (ed.), Citizens, Elections, Parties. Oslo: Universitetsforlaget. Romer, Thomas and Howard Rosenthal (1978) ‘Political Resource Allocation, Controlled Agendas, and the Status Quo,’ Public Choice, 33(4): 27–43. Romer, Thomas and Howard Rosenthal (1979a) ‘Bureaucrats versus Voters: On the Political Economy of Resource Allocation by Direct Democracy,’ Quarterly Journal of Economics, 93(4): 561–87. Romer, Thomas and Howard Rosenthal (1979b) ‘The Elusive Median Voter,’ Journal of Public Economics, 12: 143–70. Rourke, John T., Richard P. Hiskes, and Cyrus Ernesto Zarakzadeh (1992) Direct Democracy and International Politics: Deciding International Issues Through Referendums. Boulder, CO: Lynne Rienner. Rousseau, J. J. (1913) The Social Contract, trans. G. D. H. Cole. Chicago: Great Books Foundation. Rousseau, J. J. ([1762] 1973) The Social Contract and Discourses, trans. G. D. H. Cole, J. Brunfitt, and J. C. Hall. London: Dent. Ruin, Olof (1996) ‘Referendums as Instruments for Defusing Political Issues,’ in M. Gallagher et al. (eds), The Referendum Experience in Europe. London: Macmillan (now Palgrave). Russell, Peter (1993) Constitutional Odyssey: Can Canada Become a Sovereign People? Toronto: University of Toronto Press. Saltzburg, Stephen A. (1993) ‘Improving the Quality of Jury Decision-making,’ in R. E. Litan (ed.), Verdict: Assessing the Civil Jury System. Washington: Brookings Institution. Sandel, Michael (1984) ‘The Procedural Republic and the Unencumbered Self,’ Political Theory, 12(1): 81–96. Särlvik, Bo (1959) ‘Opinionbildningen vid folkomröstningen 1957’, Stockholm, Department of Justice, SOU 10. Sartori, Giovanni (1987) The Theory of Democracy Revisited. Chatham, NJ: Chatham House.

270

Bibliography

Savoie, Donald (1999) Governing From the Centre: the Concentration of Political Power in Canadian Politics. Toronto: University of Toronto Press. Sawer, Geoffrey (1967) Australian Federalism in the Courts. Melbourne: Melbourne University Press. Schmidt, David D. (1989) Citizen Lawmakers: The Ballot Initiative Revolution. Philadelphia: Temple University Press. Schrag, Peter (1998) Paradise Lost: California’s Present, America’s Future. Berkeley: University of California Press. Sears, David O. and Jack Citrin (1982) Tax Revolt: Something for Nothing in California. Cambridge, MA: Harvard University Press. Semetko, Holli A., Jay G. Blumler, Michael Gurevitch, and David H. Weaver (1991) The Formation of Campaign Agendas: A Comparative Analysis of Party and Media Roles in Recent American and British Elections. Hillsdale, NJ: Erlbaum. Setälä, Maija (1999a) Referendums in Western Europe – A Wave of Direct Democracy? Paper presented at the 1999 Annual Meeting of the American Political Science Association, Atlanta. Setälä, Maija (1999b) Referendums and Democratic Government: Normative Theory and the Analysis of Institutions. London: Macmillan (now Palgrave). Seymour-Ure, Colin (1976) ‘Press,’ in D. Butler and U. Kitzinger (eds), The 1975 Referendum. London: Macmillan Press (now Palgrave), pp. 214–45. Siune, Karen, Svensson, Palle, and Tonsgaard, Ole (1994) ‘The European Union: The Danes Said “No” in 1992 but “Yes” in 1993: How and Why?’ Electoral Studies 13(2): 107–16. Smith, Anthony (1976) ‘Broadcasting,’ in D. Butler and U. Kitzinger (eds), The 1975 Referendum. London: Macmillan Press (now Palgrave), pp. 190–213. Smith, Daniel (1998) Tax Crusaders and the Politics of Direct Democracy. New York: Routledge. Smith, Daniel (1999) The Initiative to Party: The Role of Political Parties in State Ballot Initiatives. Paper presented at the Annual Meeting of the Western Political Science Association, Seattle. Smith, Gordon (1976) ‘The Functional Properties of the Referendum,’ European Journal of Political Research, 4(1): 31–45. Smith, Mark A. (1999) Ballot Initiatives, Voter Interest, and Turnout. Paper presented at the Annual Meeting of the Western Political Science Association, Seattle. Sniderman, Paul M., Richard A. Brody, and Philip E. Tetlock (1991) Reasoning and Choice: Explorations in Political Psychology. New York: Cambridge University Press. Sobel, Joel (1985) ‘A Theory of Credibility,’ Review of Economic Studies, 52: 557–73. Somin, Ilya (1999) ‘Voter Ignorance and the Democratic Ideal,’ Critical Review, 12: 413–58. Soysal, Yasemin Nuhoglu (1994) Limits of Citizenship. Migrants and Postnational Membership in Europe. Chicago: University of Chicago Press. Steiner, Jurg (1974) Amicable Agreement Versus Majority Rule: Conflict Resolution in Switzerland. Chapel Hill: University of North Carolina Press. Steiner, Jurg (1995) European Democracies, 3rd edn. White Plains, NY: Longman. Steunenberg, Bernard (1992) ‘Referendum, Initiative, and Veto Power,’ Kyklos, 45(4): 501–29. Stouffer, Samuel (1955) Communism, Conformity and Civil Liberties. New York: Doubleday.

Bibliography

271

Suksi, Markku (1996) Bringing in the People. Dordrecht: Martinus Nijhoff. Sunstein, Cass (1984) ‘Naked Preferences and the Constitution,’ Columbia Law Review, 84: 1689–733. Svensson, Palle (1984) ‘Class, Party and Ideology: A Danish Case Study of Electoral Behaviour in Referendums.’ Scandinavian Political Studies 7(3): 175–196. Svensson, Palle (1996) ‘Denmark: The Referendum as a Minority Protection,’ in M. Gallagher, P. Uleri, and P. Vincenzo (eds), The Referendum Experience in Europe. London: Macmillan (now Palgrave), pp. 33–51. Tolbert, Caroline (1998) ‘Changing Rules for State Legislatures: Direct Democracy and Governance Policies,’ in S. Bowler, T. Donovan, and C. Tolbert (eds), Citizens as Legislators: Direct Democracy in the United States. Columbus: Ohio State University Press. Tolbert, Caroline and Rodney Hero (1998) ‘Race, Ethnicity and Direct Democracy: Basis of Support for Anti-Immigrant and Official English Measures,’ in S. Bowler, T. Donovan, and C. Tolbert (eds), Citizens as Legislators: Direct Democracy in the United States. Columbus: Ohio State University Press. Tolbert, Caroline and Rodney Hero (1999) Dealing with Diversity: Racial/Ethnic Context and Policy Change. Paper presented at the Annual Meeting of the Western Political Science Association, Seattle. Tolbert, Caroline, John Grummel, and Ridge Bowman (1999) Facing Diversity: Responses to Economic and Demographic Change. Paper presented at the Midwest Political Science Association, Chicago. Trechsel, Alexander H. and Hanspeter Kriesi (1996) ‘Switzerland,’ in M. Gallagher and P. V. Uleri (eds), The Referendum Experience in Europe. New York: St. Martin’s Press (now Palgrave), pp. 185–208. Tsebelis, George (forthcoming) ‘Veto Players in Political Analysis,’ Governance. Tully, James (1995) Strange Multiplicity: Constitutionalism in an Age of Diversity. Montreal: McGill-Queen’s University Press. Tully, James (1994) ‘Diversity’s Gambit Declined,’ in C. Cooke (ed.), Constitutional Predicament: Canada After the Referendum of 1992. Montreal: McGillQueen’s University Press. Turnbull, Malcolm (1999) Fighting for the Republic: the Ultimate Insider’s Account. South Yarra: Hardie Grant. Uhr, John (1999) The Case for Yes. Sydney: Federation Press. Uleri, Pier Vincenzo (1994) ‘Dall’instaurazione alla crisi democratica. Un’analisi in chiave comparata del fenomeno referendario in Italia (1946–93),’ in M. Caciagli, P. Uleri, and P. Vincenzo (eds), Democrazie e referendum. Roma: Laterza: 390–429. Uleri, Pier Vincenzo (1996) ‘Italy: Referendums and Initiatives from the Origins to the Crisis of a Democratic Regime,’ in M. Gallagher and P. V. Uleri (eds), The Referendum Experience in Europe. London: Macmillan Press (now Palgrave): 106–25. Valen, Henry (1973) ‘Norway: “No” to EEC,’ Scandinavian Political Studies, 8: 214–26. Valen, Henry (1974) ‘Valgforskning og politikk. Tilsløring eller avklaring,’ in M. S. Motensen (ed.), I forskningens lys. NAVF 1949–74. Oslo: Lyche. Valen, Henry and Per Arnt Pettersen (1995) ‘Rekordhøy valgdeltakelse’ in A. T. Jenssen and H. Valen (eds), Brussel midt imot (Opposing Brussels). Oslo: Ad Notam Gyldendal. Van Mill, David (1996) ‘The Possibility of Rational Outcomes from Democratic Discourse and Procedures,’ Journal of Politics, 58(3): 734–52.

272

Bibliography

von Beyme, G. (1985) Political Parties in Western Democracies. Aldershot: Gower. Walker, Geoffrey de Q. (1988) ‘The People’s Law: Initiative and Referendum,’ University of Queensland Law Journal, 15(1): 33–45. Walker, Geoffrey de Q. (1992) ‘Participation by Initiative and Referendum: Could It Work in Australia,’ in M. M. Clark (ed.), Citizen Participation in Government. Sydney: Hale & Iremonger. Walzer, Michael (1982) ‘Pluralism in Political Perspective,’ in M. Walzer (ed.), The Politics of Ethnicity. Cambridge, MA: Harvard University Press. Walzer, Michael (1999) ‘Deliberation, and What Else,’ in S. Macedo (ed.), Deliberative Politics: Essays on Democracy and Disagreement. Oxford: Oxford University Press. Waters, M. Dane (1999) ‘A Century Later – The Experiment with Citizen-Initiated Legislation Continues,’ in E. C. Ladd (ed.), America at the Polls 1998. Storrs, CT: Roper Center. Wattenberg, Martin P. (1990) The Decline of American Political Parties. Cambridge, MA: Harvard University Press. Weatherford, M. Stephen (1987) ‘How Does Government Performance Influence Political Support?’ Political Behavior, 9 (1): 5–28. Weiser, Philip (1993) ‘Ackerman’s Proposal for Popular Constitutional Lawmaking: Can It Realize His Aspirations for Dualist Democracy?’ New York University Law Review, 68: 907–59. Wenzel, James, Todd Donovan, and Shaun Bowler (1998) ‘Direct Democracy and Minorities: Changing Attitudes about Minorities Targeted by Initiatives,’ in S. Bowler, T. Donovan, and C. Tolbert (eds), Citizens as Legislators: Direct Democracy in the United States. Columbus: Ohio State University Press. Werder, Hans (1978) Die Bedeutung der Volksinitiative in der Nachkriegszeit. Bern: Francke. White, Stephen and Ronald J. Hill (1996) ‘Russia, the former Soviet Union and Eastern Europe: The Referendum as a Flexible Political Instrument,’ in M. Gallagher, Michael, P. Uleri, and P. Vincenzo (eds), The Referendum Experience in Europe. London: Macmillan (now Palgrave), pp. 153–70. Wittman, Donald (1995) The Myth of Democratic Failure: Why Political Institutions are Efficient. Chicago: University of Chicago Press. Wolfinger, Ray and Fred Greenstein (1968) ‘The Repeal of Fair Housing in California: An Analysis of Referendum Voting,’ American Political Science Review, 62: 753–69. Wyller, Thomas Christian (1996) ‘Norway: Six Exceptions to the Rule,’ in M. Gallagher et al. (eds), The Referendum Experience in Europe. London: Macmillan (now Palgrave). Young, Iris (1990) Justice and the Politics of Difference. Princeton, NJ: Princeton University Press. Zaller, John R. (1992) The Nature and Origins of Mass Opinion. New York: Cambridge University Press. Zimmerman, Joseph F. (1986) Participatory Democracy: Populism Revived. New York: Praeger. Zisk, Betty (1987) Money, Media and the Grassroots: State Ballot Issues and the Electoral Process. Newbury Park, CA: Sage.

Index Aboriginal peoples 147, 221, 225 abortion 18, 94, 103–4, 106, 161 abrogative referendums 27, 70; see also referendum abrogativo accountability 68, 75, 234, 243, 249–50 acid rain 93 Ackerman, Bruce 111–12, 238, 240, 250 ad hoc organizations 7, 16, 172–3, 177, 188–9, 217 see also interest groups adversarial approaches 234 advisory initiatives and referendums 128–30, 174 affirmative action 132, 135, 137 African National Congress (ANC) 240 agenda-setting 75, 83–4, 90, 98, 128–31, 178 Algeria 58, 60 Amsterdam Treaty 63 anti-Semitism 126 Arizona 93 Australia 5, 7–8, 10–11, 51, 110–24, 129, 237–8 constitution and constitutional change 115–19 Constitutional Convention and republican referendum 120–4, 217 Austria 55–6, 59 Baker, Lynn 233 balance in referendum campaigns 216–17, 221 Banaszak, Lee Ann 94, 97 Barber, Benjamin 155 Beer, Samuel 112 Belarus 128 Belgium 58 bills of rights 119 Blair, Tony 62

Blais, André 156–7, 196 Bogdanor, Vernon 20, 94–5, 174 Bonaparte, Napoleon 192, 204–5 British Columbia 147, 164 Bromley Spoils System Bill 93 Brown, Michael 225–6 Brundtland, Gro Harlem 180 Bush, Jeb 32 Butler, David 2, 12, 175 California 11–12, 28, 69–70, 74, 91–3, 130–41 Camobreco, John F. 102–5 Campbell, Stuart L. 206 Canada 1, 5, 8, 10, 18, 59, 61, 72, 114, 116, 118, 122, 129–30, 147, 154, 156, 161, 164, 196, 212–13, 220–1, 225, 237–8, 244–7 Charlottetown Accord 59, 122, 129–30, 154, 196 Chile 237 Citrin, Jack 30 civil rights 136, 142–3, 147 Clark, Glen 147 cleavage model 181–2, 190 Clinton, Bill 32, 39 closure, prospect of 246–8 coalitions 21, 76 Cobban, Alfred 204–5 cognitive mobilization 25–6, 29–32, 37–40 Colorado 126, 162, 238 communitarianism 155–63 ‘competence’ of voters 16, 191–6, 200–3, 207–8, 211 definition of 194 compromise, consensus 9–10, 20, 76–7, 130–1, 234 Condorcet Paradox 141–2 conscientious objectors 126 conscription 61 consent of the governed 111 273

274

Index

constitutional change 5–7, 10–11, 18, 72, 109–124, 174, 231–42, 247–51 in Australia 115–24 ongoing process of consultation on 250–1 packages of measures 251 ratification of 113, 115, 234, 242, 249–50 counterproposals 93 Cronin, Thomas 93–4, 97, 148 Crouch, Winston 93, 95–7, 100 Dahl, Robert 153 death penalty 104, 106 deliberation 16–21, 69, 76–7, 155, 211, 214, 226–30, 231–7, 240–4, 247–52 definition of 233 insulated from decision-making 249–50 by juries 241, 248–9 and legitimacy 243 and majoritarianism 242 and voting 231, 247–8 democratic deficit 15 Denmark 1, 20–1, 49, 58, 62–3, 95, 129–30, 174, 190, 217 devolution 62, 129, 221–2, 225–6 Dewey, John 243 dialogic model of deliberation 16, 232–4 direct democracy, direct legislation 1–4, 15, 25–31, 47, 67–9, 88–106, 110–11 alternative forms of 70–1, 83–5, 128–31 appealing features of 155–8 economic effects of 100–3, 106 effects on political parties 68, 75–87 elites in 127–32 functions of parties in 85–7 indirect effects of 90–3, 99–100, 105–6 objections to 71, 191 policy consequences of 103–5 public support for 33–7, 40–2 as a threat to minorities 125–7, 130–1, 136, 142–4, 148, 152, 239

disaffection towards government 25–6, 29–32, 37–40 discrimination 150–1, 162 discursive action 235–6, 241–2, 245 divisiveness 242, 244 divorce law 129, 132 ‘dualist democracy’ 111–12 egalitarianism 237 electoral reform 18, 70, 94–5 elites 2–4, 6, 10–11, 14, 18–19, 27, 119, 123, 125, 135–6, 142–4, 148, 160, 173, 191–3, 203, 207–9, 216, 229, 238–9, 251 in direct democracy 127–32 distrust of 156–7, 162, 186 equality, conceptions of 13–14, 149–52, 156–62 Erikson, Robert S. 102 Erlander, Tagus 59 Ethiopia 237 European integration and the European Union 7, 15, 21, 51–6, 62, 129–30, 170–5, 179–82, 186–7, 217–18, 221–2 evidence-driven conversations 248–50 exit polls 196, 200 factionalism 74, 77–8, 83 faux populism 26–9, 33, 37, 42 Fearon, James 97 Feld, Lars P. 101, 106 Finland 55, 63, 128, 170–4, 179–82, 187–90 Florida 31–3, 36–41 France 57–63, 92, 128–30, 192, 204–7, 217 Frey, Bruno 88, 136 Gallagher, Michael 2 Gamble, Barbara 88, 133–4, 147–8, 152 game theory 201 Gamson, William 30–1, 37, 39 de Gaulle, Charles 57, 60 gay rights 126, 132–4, 137, 153, 161–2, 238 general will 155 Germany 51, 128 Gidengill, Elisabeth 156–7, 196

Index Goette, Lorenz 88, 136 Gonzalez, Felipe 55, 59 government-initiated referendums 48–51, 62–4, 72, 76, 83–4, 113–19 reasons for 53–5 Guay, Jean H. 222

Kiewiet, D. Roderick 103, 106 King, Gary 97 Kirchgässner, Gebhard 101 Kirghizia 161 Kobach, Kris 71 Kreisky, Bruno 56, 59 Kriesi, Hanspeter 92

Hagen, Michael 104–5 Hallin, Daniel C. 213 health policies 92, 96, 99 hegemony 48–9, 113, 116 Hero, Rodney 88 Hitler, Adolf 192 Hjelseth, Arve 172 Hobbesian predicament 237 Hofer, R. 93

Lascher, Edward L. 104–5 legislative referendums 27, 48, 56–9, 128–30 legitimation 60–1, 109, 149, 243, 249–50 Leopold III, King of Belgium 58 Lévesque, René 157 libertarianism 68 Lijphart, Arend 110, 113, 116 Locke, John 111 logrolling 9 Luxembourg 51

Immergut, Ellen M. 92, 96–7, 99 immigration policy 93, 97 information needs of voters 98–9, 218 information short-cuts 196–203, 218 initiative process 8, 12–13, 15, 19, 26–8, 32–3, 36, 48–51, 56, 72–6, 86, 88–92, 113, 123, 125–7, 130–7 direct and indirect 20 indirect effects of 135 in individual American states 133–4, 143, 235 public opinion on 137–43 tension within 143 Inkatha Freedom Party 240 insurance industry 196–8, 219 interest groups 11–12, 68–9, 75, 84–6, 91–2, 98, 105, 138, 219, 222, 225 see also ad hoc organizations Ireland 7, 51, 62, 129, 175, 240 Italy 1, 5, 7, 11, 18–21, 27–8, 49, 51, 62, 69–70, 72, 74, 76–7, 85, 94–5, 129, 132, 135, 175 Japan 74, 77 judicial review 116, 235 jury deliberations 241, 248–9 Kazakhstan 161 Keohane, Robert O. 97 Key, V. O. 93, 95–7, 100

275

Maastricht Treaty 56, 61, 63, 129–30, 174, 217–18 see also European integration and the European Union McCarty, Nolan M. 99 McCubbins, Mathew D. 200–2 McGarvie, Richard 121 McIver, John P. 102 Madison, James 133 Mair, Peter 80 majoritarianism 17–18, 88, 127, 134, 155, 231–2, 238–50 and the spirit of deliberation 242 Massachusetts 93 Matignon Agreements 61 Matsusaka, John 99, 101–6 ‘median voter’ concept 12, 99–104 Medicaid 93 Michigan 77, 97 ‘minoritarian sensibility’ 152–5, 159–63 minorities, identification of 143–4 minority rights 13–14, 105, 125–7, 133–6, 142, 144, 147–53, 158–64, 239 protected by majorities 152–5 see also direct democracy as a threat to minorities

276

Index

Mitterrand, François 60 Monière, Denis 222 moral issues 5, 137 Moser, Peter 98–9 motivations for holding referendums 7, 48, 53–5, 63–4 Mulroney, Brian 59 Nader, Ralph 197–9 Napoleon III 192, 204, 206–7 nationalism 157, 160–1 Neidhart, Leonard 91–2, 96 Nevitte, Neil 156–7, 196 New Caledonia 57, 60–1 New Zealand 128–30, 238 news media 17, 31–2, 177 coverage of election and referendum campaigns 211–30 Nisga’a people 147, 163–4 North Atlantic Treaty Organization (NATO) 53, 55, 128 Northern Ireland 62 Norway 16, 54, 61, 170–6, 179–81, 186–8 nuclear power 53, 55–6, 62, 128, 174 obligatory referendums 60–3, 129, 171

financing of 171–2 ‘hollowing-out’ of 69, 74 influence on voting 178–89 levels of activity in 78–9 links with ad hoc organizations 172–3 role and functions of 15–16, 67, 85–7, 169–70, 188–9 voters’ identification with 181, 190 see also party discipline; party leaders politicization 15 pollution 93 Pommerehne, Werner W. 100–1 Pomper, Gerald 81 Pompidou, Georges 60 populism 27, 70–1, 75, 169, 238; see also faux populism precedent 7, 62–3 ‘professionalization of the initiative industry’ 26, 28–9 Propositions 187 and 209 135 ‘public brokerage’ 20 public opinion 131, 169–70, 211, 239 on the initiative process 137–43 publicity test applied to majority interests 233, 235

48, 51, 53,

Palmer, R. R. 112 Papadopoulos, Yannis 132 Parizeau, Jacques 247 partisanship 16–17, 32, 115, 119–20, 123, 132, 143, 161 party discipline 77 party leaders 212–13 Patterson, Samuel C. 4 Patterson, Thomas E. 213 pension policy 174 petitions 75–6 plebiscites 48, 57–62, 128, 192, 204, 238, 240, 242 pluralist democracy 153–4 political entrepreneurs 6, 26, 28, 69 political parties divisions within 53–6, 62–3, 77–8, 83, 131, 171, 179–80 effect of referendums on 7–8, 68, 75–87

Quebec 8, 17, 59, 74, 80, 84, 114, 118, 154, 157–8, 161, 211, 219–22, 225–6, 244, 247 questions, wording of 149, 160–3 Ranney, Austin 2, 12, 175 Ray, Leonard 187–8 referendum abrogativo 18–21, 94 referendum democracy, definition 3–5, 211 referendums appealing features of 156, 163 differences from elections 215–16 religious attitudes 161 representative democracy 1–4, 10–11, 14–20, 68, 242 rights see bills of rights; civil rights; gay rights; minority rights; women’s rights risk aversion 11 Robinson, Gertrude 223–4, 226 Rocard, Michel 57–8

Index Rochlin, Steven 104–5 Romania 128 Romer, Thomas 98, 103 Rosenthal, Howard 98, 103 Rousseau, Jean-Jacques 67, 111, 155 Särlvik, Bo 173 Sartori, Giovanni 68 Savioz, Marcel 101, 106 Sawer, Geoffrey 115 Schluter, Poul 58 Schmidt, David D. 93 Schrag, Peter 88 Sciarini, Pascal 190 Scotland 62, 129, 221–2, 225–6 separation of powers, doctrine of 67, 75–6 Serbia 128 Single European Act 58 Smith, Daniel A. 27–8 Smith, Gordon 48–51, 113 Somin, Ilya 193–4 South Africa 237–41, 249 sovereignty of the people 109–10, 112, 229, 238–40 Soviet Union 161 Soysal, Yasemin Nuhoglu 97 Spain 54–5, 59, 128 sponsorship of referendums 4–7, 48–9 status quo, defence of 92, 207 Steiner, Jurg 126 Steunenberg, Bernard 98–9 Stouffer, Samuel 127 Suksi, Markku 171 Svensson, Palle 20, 95 Sweden 55–9, 63, 92, 128, 170–5, 179–82, 186–90 Switzerland 1, 5, 9, 13, 15, 18, 27–8, 49, 51, 62, 69–77, 81, 86, 91–101, 106, 113, 125–9, 132–6, 212 Szakaly, Kristin 103, 106 threats of referendums 94–5, 98–9, 240 Tolbert, Caroline 88, 219

277

Tully, James 20, 250 Turnbull, Malcolm 121 turnout 175, 205–6 Uganda 237 Uleri, Pier Vincenzo 2, 94–5 umbrella committees 21, 74, 220–1, 250 United Kingdom 1, 51, 62, 221–2 United States 1, 5, 11, 13, 15, 19–20, 25–9, 42, 67, 72, 75, 77, 80–3, 86, 91–7, 101–2, 106, 116, 123, 125–9, 132–6, 143, 147, 161, 235 constitutionalism 112, 237–8, 242 Democratic Party 77, 81 Republican Party 75, 77, 81 state-level initiatives 133–4, 143 Supreme Court 126, 238 see also Arizona; California; Colorado; Florida; Massachusetts; Michigan; Washington; Wyoming Uzbekistan 161 Verba, Sidney 97 Vernehmlassung process 9, 91 von Beyme, G. 79 votes of confidence 57–60, 86 voting procedures 231, 241, 245–50 Wales 62, 129 Washington 137 Wattenberg, Martin P. 81 Weiser, Philip 238 Wilson, Harold 54–5 Wilson, Pete 132 women’s rights 126 women’s suffrage 94, 97 Wright, Gerald C. 102 Wyoming 93 Yugoslavia

157–8

zero sum games Zurich 135–6

246–7

Referendum Democracy: Citizens, Elites and Deliberation in Referendum Campaigns

Financing Referendum Campaigns

Direct Legislation by The Citizenship Through The Initiative And Referendum

Political Campaigning in Referendums: Framing the Referendum Issue

Public deliberation: pluralism, complexity, and democracy

Giving Voters a Voice: The Origins of the Initiative and Referendum in America

Creating Citizens: Political Education and Liberal Democracy

Democracy and deliberation: new directions for democratic reform

Intentionality, Deliberation and Autonomy

Democracy in the European Union: Integration Through Deliberation?

The Logic of Democracy: Reconciling Equality, Deliberation, and Minority Protection

Investing in Democracy: Engaging Citizens in Collaborative Governance

Deliberation Day

Deliberation Plantation

Deliberation, Participation and Democracy: Can the People Govern?

Citizens

Napoleon's Campaigns in Italy

Democracy without Citizens: Media and the Decay of American Politics

Napoleons Campaigns in Italy

Referendums and Representative Democracy: Responsiveness, Accountability and Deliberation (Routledge ECPR Studies in European Political Science)

Public Deliberation: Pluralism, Complexity, and Democracy (Studies in Contemporary German Social Thought)

Deliberation, Social choice and Absolutist Democracy (Routledge Innovations in Political Theory, 22)

Public Deliberation: Pluralism, Complexity, and Democracy (Studies in Contemporary German Social Thought)

Friends and Citizens

Deliberation and Decision: Economics, Constitutional Theory and Deliberative Democracy (Law, Ethics and Economics)

The Logic of Democracy: Reconciling Equality, Deliberation, and Minority Protection (Michigan Studies in Political Analysis)

Citizens and the State (Beliefs in Government)

The Global Commonwealth of Citizens: Toward Cosmopolitan Democracy

Power, Protest and Participation: Local Elites and Development in India

Referendum Democracy: Citizens, Elites and Deliberation in Referendum Campaigns

Financing Referendum Campaigns

Direct Legislation by The Citizenship Through The Initiative And Referendum

Political Campaigning in Referendums: Framing the Referendum Issue

Public deliberation: pluralism, complexity, and democracy

Giving Voters a Voice: The Origins of the Initiative and Referendum in America

Creating Citizens: Political Education and Liberal Democracy

Democracy and deliberation: new directions for democratic reform

Intentionality, Deliberation and Autonomy

Democracy in the European Union: Integration Through Deliberation?

The Logic of Democracy: Reconciling Equality, Deliberation, and Minority Protection

Investing in Democracy: Engaging Citizens in Collaborative Governance

Deliberation Day

Deliberation Plantation

Deliberation, Participation and Democracy: Can the People Govern?

Citizens

Napoleon's Campaigns in Italy

Democracy without Citizens: Media and the Decay of American Politics

Napoleons Campaigns in Italy

Referendums and Representative Democracy: Responsiveness, Accountability and Deliberation (Routledge ECPR Studies in European Political Science)

Public Deliberation: Pluralism, Complexity, and Democracy (Studies in Contemporary German Social Thought)

Deliberation, Social choice and Absolutist Democracy (Routledge Innovations in Political Theory, 22)

Public Deliberation: Pluralism, Complexity, and Democracy (Studies in Contemporary German Social Thought)

Friends and Citizens

Deliberation and Decision: Economics, Constitutional Theory and Deliberative Democracy (Law, Ethics and Economics)

The Logic of Democracy: Reconciling Equality, Deliberation, and Minority Protection (Michigan Studies in Political Analysis)

Citizens and the State (Beliefs in Government)

The Global Commonwealth of Citizens: Toward Cosmopolitan Democracy

Power, Protest and Participation: Local Elites and Development in India

Recommend Documents