European Integration and Political Conflict (Themes in European Governance)

This page intentionally left blank European Integration and Political Conflict Over the past half-century, Europe has...

Author: Gary Marks | Marco R. Steenbergen

36 downloads 1102 Views 1MB 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 PDF

This page intentionally left blank

European Integration and Political Conflict

Over the past half-century, Europe has experienced the most radical reallocation of authority that has ever taken place in peace-time, yet the ideological conflicts that will emerge from this are only now becoming apparent. The editors of this volume, Gary Marks and Marco Steenbergen, have brought together a formidable group of scholars of European and comparative politics to investigate patterns of conflict that are arising in the European Union. Using diverse sources of data, and examining a range of actors, including citizens, political parties, members of the European Parliament, social movements, and interest groups, the authors of this volume conclude that political contestation concerning European integration is indeed rooted in the basic conflicts that have shaped political life in Western Europe for many years. This volume provides a comprehensive analysis of political conflict in the European Union that will shape the field for years to come. gary marks is Professor of Political Science at the University of North Carolina, Chapel Hill, and founding Director of the UNC Center for European Studies. Marks’ recent books include Multi-Level Governance and European Integration (with Liesbet Hooghe; 2001), and It Didn’t Happen Here: Why Socialism Failed in the United States (with Seymour Martin Lipset; 2000). marco r. steenbergen is Associate Professor at the Department of Political Science, University of North Carolina, Chapel Hill. His recent articles include “Measuring Political Deliberation: A Discourse Quality Index,” Comparative European Politics, 1 (2003): 20–48 (with André Bächtiger, Markus Sporndli, ¨ and Jurg ¨ Steiner) and “Modelling Multilevel Data Structures,” American Journal of Political Science, 46 (2002): 218–37 (with Bradford Jones).

Themes in European Governance Series Editors Andrea Føllesdal Johan P. Olsen

Editorial Board Stefano Bartolini Andrew Moravcsik Fritz W. Scharpf Albert Weale

Beate Kohler-Koch Ulrich Preuss Philip Schlesinger J. H. H. Weiler

Percy Lehning Thomas Risse Helen Wallace

The evolving European systems of governance, in particular the European Union, challenge and transform the state, the most important locus of governance and political identity and loyalty over the past 200 years. The series Themes in European Governance aims to publish the best theoretical and analytical scholarship on the impact of European governance on the core institutions, policies and identities of nation-states. It focuses upon the implications for issues such as citizenship, welfare, political decision-making, and economic, monetary and fiscal policies. An initiative of Cambridge University Press and the Programme on Advanced Research on the Europeanization of the Nation-State (ARENA), Norway, the series includes contributions in the social sciences, humanities and law. The series aims to provide theoretically informed studies analysing key issues at the European level and within European states. Volumes in the series will be of interest to scholars and students of Europe both within Europe and worldwide. They will be of particular relevance to those interested in the development of sovereignty and governance of European states and in the issues raised by multi-level governance and multi-national integration throughout the world.

Other books in the series: Paulette Kurzer Markets and Moral Regulation: Cultural Change in the European Union Christoph Knill The Europeanisation of National Administrations: Patterns of Institutional Change and Persistence Tanja Borzel ¨ States and Regions in the European Union: Institutional Adaptation in Germany and Spain Liesbet Hooghe The European Commission and the Integration of Europe: Images of Governance Gallya Lahav Immigration and Politics in the New Europe: Reinventing Europe Frank Schimmelfennig The EU, NATO and the Integration of Europe: Rules and Rhetoric

European Integration and Political Conflict Edited by

Gary Marks Marco R. Steenbergen

   Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo Cambridge University Press The Edinburgh Building, Cambridge  , UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9780521827799 © Cambridge University Press, 2004 This publication is in copyright. Subject to statutory exception and to the provision of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published in print format 2004 - -

---- eBook (EBL) --- eBook (EBL)

- -

---- hardback --- hardback

- -

---- paperback --- paperback

Cambridge University Press has no responsibility for the persistence or accuracy of s for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

Contents

List of figures List of tables List of contributors Preface Introduction: Models of political conflict in the European Union marco r. steenbergen and gary marks Part I: 1

2

3

4

Part II: 5

page vii ix xii xiii

1

Citizens The structure of citizen attitudes and the European political space mat thew j. gabel and christopher j. anderson

13

Potential for contestation on European matters at national elections in Europe cees van der eijk and mark n. franklin

32

Don’t rock the boat: expectations, fears, and opposition to EU-level policy-making leonard ray

51

Varieties of capitalism and political divides over European integration adam p. brinegar, seth k. jolly, and herbert kitschelt

62

Political parties Defining the EU political space: an empirical study of the European election manifestos, 1979–1999 mat thew j. gabel and simon hix

93 v

vi

Contents

6

7

8

Does left/right structure party positions on European integration? liesbet hooghe , gary marks, and carole j. wilson Political competition in the European Parliament: evidence from roll call and survey analyses jacques j. a. thomassen, abdul g. noury, and erik voeten Contesting Europe? The salience of European integration as a party issue marco r. steenbergen and david j. scot t

Part III: 9

10

11

120

141

165

Groups

Contestation potential of interest groups in the EU: emergence, structure, and political alliances bernhard wessels

195

Contestation in the streets: European protest and the emerging Euro-polity doug imig

216

Conclusion: European integration and political conflict gary marks

235

References Index

260 275

Figures

0.1 0.2 0.3 0.4 2.1 2.2 2.3 2.4

3.1 3.2 3.3 5.1 5.2 6.1 6.2 6.3 7.1 8.1 8.2 8.3

The international relations model The Hix–Lord model The regulation model The Hooghe–Marks model Voter positions in left/right (horizontal) and less/more EU integration (vertical) terms Parties’ positions on left/right (horizontal) and anti-/pro-integration (vertical) dimensions Party positions in left/right (horizontal) and less/more EU integration (vertical) terms Party positions in left/right (horizontal) and less/more EU integration (vertical) terms; parties weighted by size Expectation of loss of social benefits as a function of current social protection Fear of loss of benefits by income level and social protection status quo The contingent relationship between ideology and support for the EU Euro-party locations over time Euro-party positions in a two-dimensional space Support for European integration by party family since 1984 Positioning on selected EU policies by left/right dimension Positioning on selected EU policies by new politics dimension A typology of democratic regimes The salience of European integration over time The salience of European integration by party family The salience of European integration by member state

page 6 7 8 9 40 44 45

46 54 56 59 108 109 123 124 132 143 176 177 178 vii

viii

List of figures

8.4 The relationship between salience and dissent (1992–6) 9.1 Alternative routes for contestation 9.2 The scope and character of contestation 9.3 The “circle of institutionalization” of interest groups at the supranational level 9.4 The dynamics of European interest group formation: empirical values and estimates of “reaction” and “anticipation” hypotheses 9.5 Size of the economy and number of countries’ member organizations in European umbrella organizations 9.6 Trade dependency and standardized number of countries’ member organizations in European umbrella organizations 9.7 Differentiation and degree of “European encompassiveness” of interest domains 9.8 Fragmentation of interest group systems of different domains 9.9 Alliances between interest groups and political parties at the national level 9.10 Alliances between interest groups and political parties at the European level 9.11 Regulated capitalism vs. neoliberalism – political positions of members of the European Parliament with frequent interest group contact 10.1 Frequency and percentage of Western European contentious events provoked by EU policies and institutions, 1984–1997 10.2 Dimensions of contestation in Euro-protests, 1984–1997 10.3 Distribution of Euro-protests along six dimensions of contestation, 1984–1997 11.1 A model of coalition formation 11.2 Patterns of contestation 11.3 Party positions on European issues 11.4 Public opinion on European issues

186 198 198 200

204 205

206 208 209 212 213

215

224 230 231 249 250 254 255

Tables

1.1 Policy areas included in the survey page 18 1.2 Correlations between responses to policy questions 20 1.3 Factor patterns (standardized solutions) for single-factor models 24 1.4 Factor patterns (standardized solutions) for Hix–Lord model 26 1.5 Factor patterns (standardized solutions) for Hooghe–Marks model 27 2.1 Correlations between left/right and pro-/anti-EU measures 34 2.2 Characteristics of voters’ self-placement on EU integration and left/right scales 38 2.3 Variance in party positions on left/right and EU orientation 42 3.1 Fear of a loss of social benefits as a function of the national status quo 55 3.2 Probability that respondent prefers policy made at the national level only (logit results) 57 3.3 The contingent effect of ideology on support for the European Union 58 4.1 Five specifications of varieties of capitalism 70 4.2 Two measures of dispositions to European integration and their national mean scores 75 4.3 Varieties of capitalism and national evaluations of European integration 78 4.4 Contextual models with aggregate and individual-level data 80 4.5 Ideology and contextual determinants of dispositions toward the EU integration process 82 4.6 Left/right self-placement and endorsement of European integration: bivariate correlations by country 84 ix

x

List of tables

4.7 Direct effects and interactions between ideology and context as determinants of European integration views 5.1 Number of political statements (“raw scores”) in the manifestos 5.2 Percentage of all the parties’ manifestos dedicated to each issue category 5.3 Correlation matrix (Pearson correlation coefficients) 5.4 Factor patterns (standardized solution) 6.1 Multiple regression analysis for party positioning on European integration 6.2 Multiple regression analysis for party positioning on European integration and EU policies 7.1 The transnationality of European party groups 7.2 Mean and variance of political groups’ positions in the EP (fourth parliament) 7.3 Impact of party and nationality on ideal point locations (entries are eta-squared) 7.4 Issue dimensions in the European Parliament (factor loadings > 0.4 are in bold) 7.5 Mean and variance of party groups on the issue dimensions 7.6 Influence of party and nationality on MEP attitudes (entries are eta-squared) 8.1 Descriptive statistics 8.2 Salience and the political environment 8.3 Salience and vote-seeking – model without election effects 8.4 Salience and vote-seeking – pooled model with election effects 8.5 Salience and office-seeking – model without election effects 8.6 Salience and office-seeking – pooled model with election effects 8.7 Salience and cohesion-seeking – model without election effects 8.8 Salience and cohesion-seeking – pooled model with election effects 8.9 Salience and party goals in 1999 9.1 Anticipation or reaction? Founding of European interest groups

85 98 101 102 104 125 127 146 149 151 153 155 157 173 179 180 182 183 184 185 187 192 203

List of tables

10.1 The domestic and European repertoires of contentious action for occupational and non-occupational groups, 1984–1997 11.1 Country and ideology

xi

226 257

Contributors

christopher j. anderson, Syracuse University adam p. brinegar, Duke University cees van der eijk, University of Amsterdam mark n. franklin, Trinity College, Connecticut and University of Amsterdam mat thew j. gabel, University of Kentucky (Lexington) simon hix, London School of Economics and Political Science liesbet hooghe, University of North Carolina at Chapel Hill doug imig, University of Memphis seth k. jolly, Duke University herbert kitschelt, Duke University gary marks, University of North Carolina at Chapel Hill abdul g. noury, European Center for Advanced Research in Economics (Brussels) leonard ray, Louisiana State University (Baton Rouge) david j. scot t, Virginia Commonwealth University (Richmond) marco r. steenbergen, University of North Carolina at Chapel Hill jacques j. a. thomassen, University of Twente erik voeten, George Washington University (Washington, DC) bernhard wessels, Wissenschaftszentrum Berlin für Sozialforschung (WZB) carole j. wilson, University of Texas at Dallas

xii

Preface

This book grew out of three workshops held at the University of North Carolina Center for European Studies between October 1998 and May 2000 organized by Gary Marks, Marco Steenbergen, David Scott, and Carole J. Wilson. By the late 1990s the notion that the European Union was part of an overarching, multilayered polity was commonplace, as was the conviction that comparative politics provided a powerful set of tools for analyzing that polity. Scholars of social movements, interest groups, political parties, mass publics, legislatures, elites, and bureaucracies were drawn to the study of the European Union both to encompass it within existing theories and to refine those theories. This is the intellectual background to our project. Our goal was to bring together comparativists who could shed light on the underlying structure of conflict in the European Union and who could relate this to the conflicts that shaped politics within European countries. The project draws on two scholarly traditions: the analysis of cleavages and dimensions of contestation that stems from the work of Stein Rokkan and Seymour Martin Lipset, and the analysis of political conflict in the European Union that originated in the writings of Ernst Haas and Philippe Schmitter. At the time we were formulating this project we had the sense that we were engaging fundamentally new questions, or combining old questions in novel ways. The dimensionality of European integration and its connection to domestic contestation is indeed a relatively new topic. But we were keenly aware that the line of inquiry that we were pursuing – investigating the underlying structure of political preferences and of political conflict – is well established in political science. We seek to probe a single, fundamental, question in depth: how is political contestation at the European level connected to that in domestic arenas? We engage this precise question on a broad front. We examine how citizens, national political parties, members of the European Parliament, European political parties, social movements, and interest groups conceive the issues arising from European integration and how this is linked to dimensions of contestation in national politics. This research xiii

xiv

Preface

is only possible because of the efforts of numerous scholars, including several authors of chapters in this volume, to build data sets. We analyze Eurobarometer data, manifesto data for national and European political parties, roll call data for the European Parliament, a survey of MEPs, European Election Study data, interest group data, Reuters’ reports, and two expert data sets on national political parties. The project has been funded by two extraordinary government sources: a US Department of Education grant for a National Resource Center in European Studies at the University of North Carolina; and a grant from the European Union for a North Carolina European Union Center. This book is truly a transatlantic endeavor, both in funding and in authorship. We wish to thank Ruth Pitts, Associate Director of CES, Carrie and Lauren Lovelace, for their tireless organizational efforts, and Erica Edwards, Susan Glover, Heather Mbaye, Moira Nelson, and Sarah Hutchison for valuable research assistance. Gary Marks wishes to thank the Wissenschaftszentrum Berlin fur ¨ Sozialforschung for providing the time to work on the conclusion. Ian Budge, Hans-Dieter Klingemann, Paul Taggart, and Sid Tarrow contributed to our debates at various points, and John Haslam of CUP expertly guided this project into book form.

Introduction: Models of political conflict in the European Union Marco R. Steenbergen and Gary Marks In the era following the Single European Act and the Maastricht Treaty, the European Union has been transformed into a multilevel polity in which European issues have become important not only for national governments, but also for citizens, political parties, interest groups, and social movements. How is conflict over European integration structured? This is the question that this book addresses. The question of contestation over European integration has two related components. First, how do domestic and European political actors conceive the basic alternatives? Can debates over European integration, despite their complexity, be reduced to a relatively small number of dimensions? Does contestation over European integration resolve itself into a single underlying dimension, or does it involve two or more separate dimensions? Second, how is contestation over European integration related, if at all, to the issues that have characterized political life in Western Europe over the past century or more? In particular, how is contestation over European integration related to the left/right divide concerning the role of the state and equality vs. economic freedom? These topics were first raised by neofunctionalists writing in the early days of European integration. Ernst Haas paid close attention to the domestic sources of opposition and support for European integration in his classic study, The Uniting of Europe, published in 1958. However, most scholars continued to view European integration as the result of foreign policies conducted by government elites acting on a “permissive consensus” (Lindberg and Scheingold 1970). European integration was conceived to take place among, but not within, countries. This view became untenable as the European Union became a more openly contested arena for political parties, interest groups, subnational governments, and social movements after the Maastricht Accord of 1991 (Ansell, Parsons, and Darden 1997; Bache forthcoming; Caporaso 1996; Hooghe and Marks 1999; Imig and Tarrow 2001; Peterson and Bomberg 1999; Pollack 2000; Taggart 1998). Comparativists began, once again, 1

2

Marco R. Steenbergen and Gary Marks

to explore European integration as an extension of domestic politics (Laffan 1996; Niedermayer and Sinnott 1995; Risse 1996; Sbragia 1992; Schmitter 1996; Tsebelis 2000; Wessels 1997). While economic models of preference formation appeared promising for explaining trade policy, many comparativists were sensitive to the way in which ideology frames preferences, and wished to broaden the study of European integration to public opinion, social movements, and party politics (Banchoff and Smith 1999; Bell and Lord 1998; Caporaso and Keeler 1995; Jachtenfuchs and Kohler-Koch 1996; Katz and Wessels 1999; Ladrech and Marlière 1999; Mair 1999; Raunio 1996; Ray 1997; Schmitt and Thomassen 1999). They therefore made the connection with the tradition of scholarship on dimensions of contestation, to which we turn next. Dimensions of contestation The literature on dimensions of contestation goes back at least to Lipset and Rokkan’s (1967) pioneering work on cleavage structures. Lipset and Rokkan’s theory of social cleavages hypothesizes how macro developments – the national revolution, the Protestant Reformation, and the industrial revolution – produce enduring structures of conflict that shape organizational formation and perceptions of the political world. The question we pose in this volume is whether and how the issues arising from European integration are linked to these structures and, in particular, to the ubiquitous left/right divide. In order to make progress with this question, we need to amend the Lipset/Rokkan framework in three respects. These amendments lead us to examine the response of particular actors and organizations to European integration; whether there are different patterns of contestation for different issues; and how the interaction of territorial arenas at different levels shapes contention. Actors First, we examine how the issues arising from European integration connect to existing structures of conflict in existing democratic systems. The historical transformations analyzed by Lipset and Rokkan took place prior to democratization. European integration, in contrast, plays out within democracies having highly institutionalized systems of contestation. Political parties, interest groups, and social movements actively mediate the impact of European issues. While it is clear that modern systems of interest mediation are not frozen, they are deep-rooted, and this reinforces continuity of

Models of political conflict in the European Union

3

contestation within them. Political actors have an incentive to interpret new issues in the light of existing cleavages such as the left/right ideological dimension. Why should this be so? First, one can invoke the concept of path-dependence to argue that it is extremely costly for political parties to abandon existing cleavage structures (Marks and Wilson 2000; Scott 2001). Parties attract ideologically motivated activists, they build strong institutional ties to particular constituencies, and they develop reputations for particular programs and policies. One can make a similar case from a rational choice perspective. Whereas stable political equilibria exist in a political space that is dominated by a single dimension (due to the median voter theorem), stable equilibria in multidimensional political spaces are much more difficult to arrive at (see, for example, Hinich and Munger 1997). The problem is already quite complex in a two-dimensional space, but becomes intractable in spaces of higher dimensionality. Thus, introducing new dimensions of conflict is costly to those benefiting from the status quo because it will entail instability. Finally, one can invoke cognitive political psychology to make the argument. The more complex the political space, the more difficult it becomes for actors to operate in this space due to cognitive limitations. Moreover, information-processing theory argues that decision-makers interpret new information in light of what they already know, suggesting a tendency to accommodate new issues to old cleavages (Steenbergen and Lodge forthcoming).

Issues Second, we shift from analysis of cleavages to more fine-grained analysis of issues. Lipset and Rokkan’s concept of cleavage ties together social structure, the organization of political conflict, and the substantive character of that conflict. Cleavages arise to the extent that social structure – chiefly, occupation, religion, and spatial location – determines political preferences. In contemporary European democracies, this causal connection has weakened. The links between social structure and political preferences that Lipset and Rokkan diagnosed in 1967 appear looser today. In this volume we relax the assumption that European integration can be conceptualized as a cleavage. In order to explore the structure of preferences on European integration, we therefore operate at the level of issues, not cleavages. To what extent, we ask, do the issues arising from European integration hang together as a single dimension, and to what extent is this dimension (or dimensions) connected to existing structures of conflict?

4


Arenas Finally, one of our core concerns is to examine dimensions of contestation at the national and European levels. We attempt not only to explain variation across national polities, but also to analyze dimensions of contestation among European political parties and members of the European Parliament. We cannot therefore follow Lipset and Rokkan in taking the national polity as the unit for the analysis for party systems (Allardt 2001: 22–3). Models In the Lipset/Rokkan framework, dimensions of contestation that arose from the class cleavage, the religious cleavage, and the center–periphery cleavage are enduring because they are rooted in social structure and political organization. From their standpoint in the mid-1960s, Lipset and Rokkan believed that the resulting pattern of cleavages was “frozen” into place. Few scholars still argue that these cleavages are permanent (Dalton, Flanagan and Beck 1984; Franklin 1992; Hix and Lord 1997; Karvonen and Kuhnle 2001). However, their ideological residue is still visible in the politics of the EU member states. The ideological continuum from left to right is a central organizing dimension in Western Europe (Barnes and Kaase 1979; Hix 1999a, 1999b; MacDonald, Listhaug, and Rabinowitz 1991; Hix and Lord 1998; van der Eijk and Niemoller ¨ 1983). There is some debate whether this continuum can be divided into an economic and a libertarian-authoritarian dimension (Kitschelt 1994), or an old politics versus new politics dimension (Franklin et al. 1992). But no one doubts the importance of the categories left and right in Europe. While few comparativists take issue with the maxim that ideology – and in particular the left/right divide – constrains how individuals and organizations interpret new issues, the tightness of fit is debated. There are four logical possibilities:1 r Contestation takes place on a single anti-integration vs. pro-integration dimension; the left/right continuum is irrelevant for understanding contestation on European integration. r The dimensions are unrelated (i.e., they are orthogonal to each other). r The dimensions are fused in a single dimension. r The dimensions are related to each other in two-dimensional space, but they are not necessarily fused (i.e., the dimensions are oblique). 1

The following discussion draws on an earlier version of the chapter by Matthew Gabel and Simon Hix presented in Chapel Hill, North Carolina, April/May 2000.


5

Each of these possibilities reflects a particular scholarly view on the subject. The first is consistent with an IR (international relations) model of contestation; the second is developed by Simon Hix and Christopher Lord (1997); the third is consistent with a speculation articulated by George Tsebelis and Geoffrey Garrett (1996); and the fourth corresponds to Liesbet Hooghe and Gary Marks’ model of regulated capitalism vs. market liberalism (1999). The international relations model International relations approaches imply that European integration, and contestation about European integration, are independent from the left/right dimension of domestic contestation. This assumption is shared by realism, intergovernmentalism, and neofunctionalism. Realist scholars hypothesize that national leaders take positions on the European Union in response to geopolitical pressures (Hoffmann 1966). Pursuit of the national interest, rather than domestic ideological differences, determines whether national leaders support or oppose further European integration. Liberal intergovernmentalists argue that economic pressures bearing on national producer groups are primary. Exporters favor deeper economic integration; import-competing producer groups are opposed (Moravcsik 1998). The link between economic integration and supranational decision-making is functional. Governments will create supranational institutions to the extent that they cannot use national institutions to solve the credible commitment problem of international contracting. Liberal intergovernmentalists maintain, against realists, that domestic conflict is important in explaining why some governments support and others oppose integration. But that conflict is about the gains and losses from trade, and is independent from the left/right dimension that structures much domestic contestation. Neofunctionalists emphasize that functional pressures for political integration are mediated by cross-national coalitions and supranational entrepreneurs. The integration process – which may lead national elites to redefine policy problems and even their identities – is independent of the conventional left/right dimension of domestic political contestation. Support of and opposition to further integration arises as elite actors – bureaucrats and politicians – assess costs and benefits in a dynamic context of problem-solving, spillover, and learning (Haas 1958). The international relations model that we draw from these literatures conceives of contestation across the range of EU issues along a single dimension. Conflict on this dimension involves diverse sets of

6


Less integration (defend national sovereignty)

More integration (promote supranational governance)

Figure 0.1 The international relations model.

actors – national governments for realists, governments and producer groups for intergovernmentalists, and bureaucrats and politicians for neofunctionalists. However, in each case, contestation is independent from the left/right concerns that frame domestic politics. Hence in figure 0.1 we represent the international relations model as an autonomous dimension ranging from “less integration” to “more integration.” The Hix–Lord model: two orthogonal dimensions Simon Hix and Christopher Lord have argued that contemporary EU politics is increasingly two-dimensional (Hix and Lord 1997; Hix 1999a, 1999b). A left/right dimension, summarizing diverse economic and sociopolitical issues in the domestic arena, remains orthogonal to a “national sovereignty” dimension ranging from less integration to more integration (figure 0.2). The orthogonality of these dimensions reflects contrasting pressures. Major political parties have a big stake in the existing left/right pattern of contestation and they seek to extend it to new issues that arise on the political agenda. When challenging parties and groups raise new issues, the major parties respond by trying to force them to compete as well on the left/right dimension. But national sovereignty issues are difficult to assimilate into the left/right dimension. Hix and Lord argue that the major parties are divided on national sovereignty. They therefore prefer to compete on the left/right dimension while bottling up competition on issues of European integration. “Traditional parties distinguish . . . [among] themselves in the domestic arena over the role of state authority in the making of social and economic policies, and not on the question of the institutional design of the emerging supranational political system in Europe” (Hix and Lord 1997: 26). They do this by taking a generally pro-integration stance, leaving opposition to minor parties at the ideological extremes. According to Simon Hix, the independence/integration and left/right dimensions cannot be collapsed into a new single dimension because they mobilize cross-cutting political coalitions (1999a). The left/right dimension involves the allocation of resources and values between functional groups, whereas the national sovereignty dimension involves


7

More integration

Left

Right

Less integration Figure 0.2 The Hix–Lord model.

the allocation of resources and values between territorial groups. The upshot of this is that social classes and political parties are internally divided over European integration. Correspondingly, the four possible dichotomous alternatives – left/more integration, left/less integration, right/more integration, and right/less integration – are all feasible policy positions, and are likely to be pursued by different forces. Hence, these two dimensions co-exist orthogonally in EU political space in the same way that functional and territorial cleavages persist in other territorially divided polities (such as Belgium, Switzerland, Canada, or the United States).

The regulation model George Tsebelis and Geoffrey Garrett have speculated that European integration might be subsumed into the left/right dimension (2000). In this scenario, EU politics is fused to domestic competition between the left, which pushes for common economic regulation across Europe, and the right, which favors less EU regulation. Both the international relations model and the regulation model conceive of a single dimension of contestation for European issues, but they take sharply opposing positions on

8


Left (high regulation)

Right (low regulation)

Figure 0.3 The regulation model.

how this dimension relates to the left/right dimension in domestic politics. Whereas the international relations model implies that European integration is autonomous from the conflicts that have historically structured domestic contestation, the regulation scenario hypothesizes that European contestation is an expression of such conflicts. The widely recognized primacy of national politics in EU elections establishes the plausibility of the regulation model. European elections are “second-order” elections in which political parties and voters are chiefly motivated by national issues (van der Eijk and Franklin 1995; Reif and Schmitt 1980). Moreover, “European elections usually occur between national elections, when the popularity of the party/coalition in government is typically relatively low” (Tsebelis and Garrett 2000: 31). National political parties dominate voting in the European Parliament, and their performance in national elections shapes the composition of national governments, which send delegates to the Council of Ministers. Consequently, legislative actors in the Parliament and national representatives in the Council may be constrained by the same left/right dimension that structures national politics in EU member states. Kreppel and Tsebelis find evidence that traditional left/right divisions characterize many issues that are debated in the European Parliament and the EU political process (1999), while Garrett has argued that bargaining over institutional reform in the Single European Act was shaped by national preferences concerning the extent of intervention in the European economy (1992). In both cases, one may argue that political parties in the European Parliament and in national governments will only support reforms that shift the status quo closer to their ideal position along the left/right dimension. Whereas the international relations model conceives European integration as a single dimension that is entirely independent of the left/right dimension, the regulation model hypothesizes a single dimension that is subsumed into the left/right dimension. This is represented in figure 0.3.

The Hooghe–Marks model: “regulated capitalism” vs. “neoliberalism” Liesbet Hooghe and Gary Marks (1999, 2001) identify a left/right dimension ranging from social democracy to market liberalism and a


9

European integration dimension from nationalism to supranationalism. In their view, these dimensions are neither fused together nor orthogonal to each other. Rather, Hooghe and Marks claim that certain aspects of European integration are likely to be absorbed into the left/right dimension. To the extent that this is the case, pro- and anti-EU and left and right become indistinguishable. However, not all aspects of integration are easily incorporated into the left/right dimension and to the extent that they cannot be, a distinct pro-/anti-integration dimension emerges. More integration

Regulated capitalism

Left

Right

Neoliberalism

Less integration Figure 0.4 The Hooghe–Marks model.

Hooghe and Marks hypothesize that the center-left is likely to become more pro-European as the debate over European integration focuses on market regulation rather than market-making. The center-left, including in particular social democrats, support regulated capitalism, a project to build environmental, social, infrastructural, and redistributive policy at the European level (1999).2 As regulatory issues are taken up at the 2

Mark Pollack suggests that Prime Minister Tony Blair’s “third way” approach is an alternative to Hooghe and Marks’ regulated capitalism (2000). However, Blair’s “third way” and “regulated capitalism” both reject government ownership, Keynesian demand-side policy, and heavy-handed government regulation in favor of “market-enhancing or market supporting – rather than market-replacing or even market-correcting – policies” (Hooghe and Marks 1999: 87).

10


European level, social democrats become more favorably disposed to further integration. Those on the political right, in contrast, become more opposed to European integration. They wish to combine European-wide markets with minimal European regulation, and once economic and monetary integration is in place, they become skeptical of the benefits of further European integration. The neoliberal project rejects supranational authority, and strives instead to provoke regulatory competition among national governments within an encompassing market. Contestation in EU policy space is therefore structured in two camps. Figure 0.4 depicts this hypothesis through ellipses in two of the quadrants. One can think of a line that passes through these ellipses, sloping down from regulated capitalism on the center-left to neoliberal capitalism on the right. Thus, Hooghe and Marks propose a two-dimensional model of the EU political space where the left/right dimension and the national sovereignty dimension jointly structure actors’ policy positions in EU political space. Plan of the book These models generate conflicting expectations about the connection between the issues raised by European integration and the left/right divide. To evaluate their validity we must disaggregate, and examine variation among actors, issues, and arenas. Is the dimensionality of EU political space, and its connection with domestic structures of conflict, actorspecific or arena-specific? To what extent do the issues arising from European integration hold together as a single dimension? Do particular subsets of issues connect to domestic dimensions of contestation? We gain leverage with these questions by investigating a range of issues for several kinds of actors at the national and European levels. We have chapters by Gabel and Anderson, van der Eijk and Franklin, Ray, and Brinegar, Jolly, and Kitschelt on citizens; Hooghe, Marks, and Wilson and Steenbergen and Scott on national political parties; Gabel and Hix on European political parties; Thomassen, Noury, and Voeten on European members of parliament; Wessels on interest groups; and Imig on social movements. We seek a general understanding of European contestation on the basis of a wide-ranging analysis of actors, issues, and arenas.

Part I

Citizens

1

The structure of citizen attitudes and the European political space∗ Matthew J. Gabel and Christopher J. Anderson

As Steenbergen and Marks (Introduction) argue, the definition and substantive content of the political space is crucial for understanding the nature of political competition in the European Union. In the study of industrialized democracies, including the EU member states, scholars often define the policy space in terms of voter preferences over policy (see Gabel and Huber 2000). Since parties and representatives compete before an electorate, the ideological structure of voters’ preferences is fundamental to understanding political contestation. Consequently, one common approach to defining the political space is empirically to examine how an electorate structures its policy preferences. In this chapter, we attempt to describe the EU policy space in the same manner: by examining the structure (or lack thereof) of EU citizens’ preferences over EU policy. However, it is important to note that voter preferences do not play exactly the same role in EU politics as in representative democracies. For one, the links between policy-makers and citizens are different in the European Union than in a typical representative democracy. The Council of Ministers – arguably the most important legislative body in the EU – consists of representatives of national governments elected in national, not EU, elections. It is relatively uncommon that national governments fall or lose elections due to their positions taken in the Council of Ministers. As a result, past research has, at least implicitly, dismissed the EU electorate as a salient constituent for these national representatives. Instead, citizens provided a “permissive consensus” that allowed national officials to pursue their own objectives, not a policy space that systematically constrained their behavior. ∗

Earlier versions of this paper were presented at conferences on “Contestation in the European Union,” University of North Carolina, Chapel Hill, August 1999 and April 2000. The data used in this study come from Eurobarometer 44.2 Bis and are available as ICPSR Study No. 6748. The original collector of the data, ICPSR, and the relevant funding agency, bear no responsibility for the uses of this collection or for interpretations or inferences based upon such uses. We would like to thank Andrew LoTempio for his help with handling the data.

13

14

Matthew J. Gabel and Christopher J. Anderson

Increasingly, however, institutional reforms of the legislative process have altered the EU in the direction of a representative democracy. As a result, a growing body of research now addresses questions of representation, policy-making, and future political dynamics in the European Union (Schmitt and Thomassen 1999). To evaluate the character and quality of representation in EU policy-making, we need to compare the preferences of citizens with the preferences and behavior of representatives. Thus, an accurate description of how voters structure their preferences over EU policy is fundamental to such an undertaking. Since much of the empirical research on representation in the EU uses spatial models (e.g., Schmitt and Thomassen 1999), this requires a description of citizens’ EU policy space. In addition, the structure of citizens’ preferences over EU policy remains an important issue for understanding two aspects of European integration. First, calls for reforms of the EU to fill the “democratic deficit” are criticized on the grounds that the EU is not sufficiently integrated at the mass level to support functioning democratic institutions. According to Scharpf (1999: 187), the democratic deficit is, in large part, due to the lack of a common identity or a European-wide policy discourse. This deficit, it is argued, cannot be filled simply by institutional reforms designed to enhance the opportunity for popular influence through a more powerful European Parliament, for example. While certainly the existing evidence indicates that EU citizens lack a strong European identity (e.g., Gabel 1998b), we know very little empirically about the potential viability of Europe-wide policy discourse. In national contexts, this discourse and political space are usually defined by a simplifying language – often referred to in terms of ideology – that facilitates political communication and competition. For example, the left/right ideological dimension is crucial to how voters choose between parties, parties compete for voters, and policy positions are packaged in party platforms. In the absence of this structure in the policy space, citizens would lack a central component of political discourse, undermining meaningful political participation. In contrast, where a fairly simple ideological structure underlies political discourse, voters can identify policy packages that cross-cut ethnic or geographic differences, facilitate compromise, and generate stable policy outcomes. Thus, one key question for the success of institutional reforms of the EU is whether or not mass attitudes on European integration are ideologically structured. Second, the development of current and proposed institutional reforms suggests that the European electorate may play an increasingly important role in structuring political competition in the EU. A variety

Citizens attitudes and the European political space

15

of recent institutional reforms have increased the opportunity for mass political behavior to influence EU policy-making. For example, the gradual development of the legislative power of the European Parliament provides a direct channel for citizens’ preferences over EU policy to influence EU policy-making. In addition, the creation of the Committee of the Regions and the Economic and Social Committee provides citizens with alternative representative channels into EU policy-making. Finally, if current concerns about the democratic deficit are answered by further institutionalizing popular input and control over EU policy-making, then the preferences of the EU electorate will undoubtedly become an important determinant of the policy space of EU competition. Under such reforms, voters’ preferences over EU policy might also cause particular cross-national political coalitions to form to influence policy-making. For example, consider the scenario in which the traditional left/right dimension from the domestic political arena also structures public preferences over EU policies. In this case, the parties of the left at the domestic level will form a natural coalition at the European level, as their constituents’ policy interests converge. This would serve to buttress the existing coalitions of national parties in the European Parliament. However, if citizens’ policy space is different from this traditional left/right dimension, then the current party coalitions at the EU level may unravel. In sum, to address questions of representation, the viability of institutional reforms, and the structure of EU politics in a reformed EU we need to understand the structure of the EU political space at the mass level. Unfortunately, previous research on citizens’ attitudes toward the EU offers little evidence relevant to these three questions. Instead, scholars have usually focused on support for European unification and a country’s membership in the EU (e.g., Eichenberg and Dalton 1993; Eichenberg 1998; Gabel 1998b), as well as support for particular institutions, such as the European Court of Justice or the President of the European Commission (Caldeira and Gibson 1995; Gelleny and Anderson 2000). The remainder of the chapter is therefore designed to provide a first step toward mapping out this structure of the EU political space at the mass level. Specifically, we are interested in whether any of the four models of the EU policy space described in Steenbergen and Marks (Introduction) are accurate models of the EU policy space at the mass level. To this end, we use confirmatory factor analysis to examine how well these models of the policy space account for the observed positions of EU citizens on EU policy issues. In the following section we describe the data and methods used to examine this question. We then discuss the findings and their implications for our understanding of the EU.

16


Models of the EU policy space How do citizens structure their preferences over EU policies? In the study of citizens’ policy preferences in representative democracies, this question is typically answered by developing an empirical model that simplifies voters’ preferences across a host of issues into a small number of dimensions. Fundamental to this approach is the assumption that policy positions are structured by underlying ideological dimensions that account for covariation in these positions. These ideological dimensions represent the structure of political discourse, representing a linguistic shorthand for political communication and competition. Consistent with a long tradition of research on mass political behavior, previous studies have conceived of these dimensions as ideological constraints on citizens’ policy positions, such that citizens’ positions on a broad range of issues are related to each other in consistent and identifiable ways (e.g., Converse 1964; Kinder and Sears 1985: 664). Ideology therefore reduces differences in citizens’ positions over many policies to differences in positions on a small number of dimensions. This implies that if one can uncover the relationships between specific policies and the ideological dimensions (i.e., the policy content of the dimension), then one can infer citizens’ positions on the ideological dimension from their positions on the specific policies. Put another way, if we know how policy positions are structured, we can infer a citizen’s position on one issue from his or her position on another issue. Again, if such an ideological structure exists, this represents the structure of political discourse. If such a structure does not exist, political competition and communication may be inefficient at best, and impossibly complicated at worst. In the absence of theoretical or conceptual guidance, such a model of the ideological space could be created inductively by searching for patterns in policy positions. However, exploratory analyses of this sort come at a price. Usually, if the policy space has more than one dimension, statistical techniques for identifying the policy space do not generate unique solutions – i.e., there are multiple structures that fit the data equally well (Long 1983: 34). Consequently, we prefer to use conceptual models to specify ex ante the relationships between citizens’ policy positions and the policy content of the ideological dimension underlying these positions. We can then examine this underlying ideological structure through confirmatory analysis. As described in Steenbergen and Marks (Introduction), previous research on EU politics proposes four models of the EU policy space. None of these models has been explicitly specified for the mass public. Instead,


17

these studies have modeled the policy space of competition between national governments and national political parties over EU policy (e.g., Garrett and Tsebelis 1999). However, these models of the political space are generally based on a rationale that also applies to the electorate. For example, Marks and Wilson (2000) argue that the space of partisan competition in the EU is based on the cleavages that structure domestic politics because these cleavages structure the way parties view policy at the national and the supranational level. Like party leaders, voters also have cognitive maps and cues (e.g., the left/right dimension) that they use to organize the political world in the domestic arena, and these maps and cues provide a prism through which they can organize their preferences over EU policies. Data and analysis Steenbergen and Marks (Introduction) describe four models of the EU policy space: the international relations model, the regulation model, the Hix–Lord model, and the Hooghe–Marks model. To investigate whether EU citizens’ attitudes to EU policy are structured according to these models, we conducted several confirmatory factor analyses of Eurobarometer survey data. Ideally, we would analyze a data set consisting of EU citizens’ preferences over all the policies that constitute the EU policy space – i.e., all policy areas under EU authority. Using confirmatory factor analysis, we would then estimate how well the hypothesized underlying structure of attitudes accounts for the observed covariation among respondents’ policy preferences. Unfortunately, such a data set does not exist. The Eurobarometer survey asks a uniform set of questions to respondents across all EU member states, but rarely asks respondents about their preferences over a large number of policies under EU authority.1 However, one notable exception is the question presented in table 1.1 which was asked in Eurobarometer 44.2 bis (spring 1996).2 This question, applied to the list of policy statements (a–y), largely meets our data needs. First, the list of policy statements covers a broad range of areas of EU policy authority. Furthermore, while the list does not exhaust all EU policies, it does include 1 2

The Eurobarometer does include a variety of questions concerning the creation of a new EU policy authority, but our focus is on preferences regarding existing EU policy areas. The respondent could answer “key priority,” “not a key priority,” or “don’t know.” We coded these responses as (0) not a key priority, (1) don’t know, and (2) key priority. We included “don’t know” as an intermediate category because we expected it to capture indifference regarding the policy statement and because no indifference category was made available. Also, a dichotomous variable would be problematic for conducting factor analysis.

18


Table 1.1 Policy areas included in the survey Some people expect the European Union to become (even) more active than now in certain policy areas. For each of the following, please tell me if you consider it a key priority or not. a. b. c. d. e. f. g. h. i. j. k. l. m. n. o. p. q. r. s. t. u. v. w. x. y.

Keeping peace by intervening more firmly in possible conflicts Dealing with the immigration problem Protecting our European cultures in all their expressions: art, cinema, etc. Setting up a European army for a common defense Paying less attention to the economy and more to social justice Protecting European Union products from non-member countries’ products Making joint efforts to protect the environment Giving more information about decisions taken at the European level and their practical consequences Developing joint programs to fight unemployment Protecting us from non-European competition, from the USA, Japan, etc. Having a strong European currency Fighting drug-trafficking Fighting against international crime Preventing the import of manufactured goods from countries where working conditions are unacceptable Promoting the defense of human rights Increasing controls at all external borders of the European Union Supporting the poorer regions of the European Union Protecting consumers Being closer to the citizens Improving equality of opportunity between men and women Creating better opportunities for smaller European Union member countries Ensuring that every European Union citizen can live, work, or study anywhere in the European Union Giving more help to the poor and socially excluded in the European Union Giving more help to people in Third World countries Improving equality of opportunities for minorities

Note: This is not an exhaustive list of all policy areas in the survey. We only list the policy areas that are relevant for testing the theoretical models.

policy areas that are central to distinguishing between the models of the EU policy space we seek to test. Second, the policy statements generally indicate a policy direction, not simply a policy area. For example, respondents were not asked whether EU activity in the area of international intervention is a priority. They were asked whether intervening “more firmly” is a priority. This directional component is important for our data analysis, since we want to examine how citizens’ policy positions across policy areas covary so as to define the policy space. The one clear weakness of this survey for our purposes is that the respondent is not simply asked whether he or she agrees with each policy


19

statement. The respondent is asked whether the policy statement is a priority. Thus, it is possible that a respondent might support EU policies that fight against drug-trafficking but not consider them a priority, resulting in a response of “not a key priority.” If this is the case, we would not be able to distinguish such a respondent from one who was against more EU activity in the area of fighting drug-trafficking. Put differently, the question is probably tapping the salience and direction of each policy statement, and we simply want to extract the respondents’ preferred policy direction. Lacking a better survey, we have no solution to this problem. However, we should note that the survey question design helps minimize this concern. The survey question does not limit the number of policies the respondent can identify as a priority. Consequently, respondents are not forced explicitly to prioritize among policy statements and can state their directional preference regarding each policy statement. We will return to this issue when interpreting the results of the analysis. Table 1.2 presents the correlations between the responses to the policy statements (a–y) in table 1.1. There were 54,004 respondents used in the analysis. The focus of our analysis is whether these correlations between the policy positions are structured according to the systematic patterns identified in the models described in the previous section. We used confirmatory factor analysis to bring evidence to bear on this question. Confirmatory factor analysis evaluates the performance of a particular factor structure in accounting for the covariation among a set of variables – in this case, the set of policy statements. The factor structure specifies (a) how many dimensions or factors underlie the covariation in policy positions, and (b) how specific policy positions relate to particular underlying dimensions and to each other. The results of the confirmatory factor analysis for each factor structure will then allow us to compare the fit of each structure to the observed survey data. Each of the four models specifies a particular factor structure underlying the survey responses related to the set of policy positions. In order to test these models, we need to specify specific hypotheses about how these dimensions relate to particular policy statements and, in the case of two dimensions, how these dimensions are related. We can then assess which factor structure best accounts for the relationships in citizens’ policy positions. Note that all of the models purport to define the policy space, which involves all issues. Thus we expect, for each model, that the underlying structure is associated with all policy statements. However, to clearly define the character of the dimension, we must impose some constraints on how particular policies relate to particular dimensions. That is, we must force particular policy statements that we identify as “markers”

– 0.15 0.12 0.20 0.11 0.15 0.16 0.13 0.17 0.14 0.17 0.16 0.17 0.14 0.20 0.16 0.17 0.14 0.13 0.15 0.17 0.17 0.17 0.15 0.17 – 0.24 0.17 0.12 0.19 0.21 0.21 0.21 0.19 0.17 0.17 0.20 0.19 0.22 0.25 0.17 0.22 0.18 0.19 0.17 0.19 0.18 0.15 0.18

b

– 0.23 0.17 0.25 0.20 0.26 0.17 0.24 0.19 0.12 0.14 0.22 0.22 0.23 0.25 0.28 0.25 0.27 0.26 0.23 0.23 0.22 0.25

c

– 0.14 0.23 0.08 0.15 0.14 0.24 0.31 0.09 0.08 0.15 0.12 0.23 0.17 0.17 0.16 0.15 0.18 0.19 0.14 0.14 0.16

d

– 0.13 0.16 0.16 0.19 0.12 0.09 0.12 0.13 0.19 0.232 0.13 0.21 0.19 0.21 0.23 0.20 0.16 0.25 0.22 0.23

e

– 0.14 0.21 0.20 0.50 0.27 0.13 0.13 0.27 0.16 0.31 0.20 0.24 0.18 0.17 0.22 0.21 0.17 0.10 0.16

f

– 0.27 0.25 0.11 0.14 0.22 0.28 0.21 0.35 0.17 0.25 0.24 0.23 0.26 0.23 0.23 0.26 0.21 0.24

g

– 0.22 0.20 0.19 0.15 0.21 0.26 0.30 0.24 0.30 0.28 0.33 0.28 0.32 0.30 0.27 0.20 0.29

h

– 0.21 0.21 0.21 0.22 0.18 0.28 0.16 0.26 0.26 0.24 0.24 0.24 0.29 0.30 0.20 0.26

i

– 0.26 0.12 0.13 0.28 0.15 0.33 0.18 0.24 0.19 0.17 0.22 0.28 0.20 0.16 0.21

j

Note: All correlation coefficients are significant at the 0.01 level.

a b c d e f g h i j k l m n o p q r s t u v w x y

a

– 0.11 0.12 0.14 0.17 0.19 0.24 0.20 0.18 0.18 0.22 0.28 0.20 0.16 0.21

k

Table 1.2 Correlations between responses to policy questions

– 0.58 0.18 0.24 0.22 0.16 0.21 0.18 0.19 0.18 0.16 0.20 0.15 0.17

l

– 0.22 0.31 0.22 0.18 0.24 0.21 0.22 0.19 0.19 0.23 0.16 0.20

m

– 0.27 0.29 0.25 0.28 0.25 0.27 0.27 0.22 0.26 0.21 0.26

n

– 0.21 0.34 0.33 0.31 0.38 0.32 0.30 0.38 0.29 0.37

o

– 0.22 0.28 0.25 0.22 0.23 0.20 0.18 0.13 0.18

p

– 0.33 0.30 0.34 0.46 0.34 0.48 0.42 0.44

q

– 0.46 0.39 0.34 0.30 0.32 0.24 0.33

r

– 0.41 0.34 0.28 0.32 0.25 0.33

s

– 0.40 0.31 0.37 0.32 0.41

t

– 0.35 0.40 0.35 0.42

u

w

x

– 0.34 – 0.26 0.47 – 0.35 0.52 0.52

v


21

for specific dimensions. These markers serve to define the dimensions. We want to choose only enough markers for each dimension to capture the character of that dimension. The international relations model posits that the EU policy space has one dimension that accounts for respondents’ positions on the policy statements. In particular, policy statements that raise issues of national sovereignty should load particularly strongly on that dimension. The set of policy statements provided in the survey does not include many of these statements. Partly, that is because we sought a survey question involving policy areas that the EU already exercises supranational authority over, not issues that are still largely controlled by national governments. However, several policy statements address issue areas over which, at the time of the survey (1996), the EU institutions did not explicitly exercise supranational authority. Specifically, policy statements (d) and (k) concern EU activity in policy areas that are traditionally central to national sovereignty (military and monetary authority) and that were not under independent EU authority. We also suspect that several other policies involved national sovereignty.3 Policies (b), (l), (m), and (p) all involved issue areas where the EU did not yet exercise concrete authority. At the time of the survey issues of immigration, international crime, drug-trafficking, and border controls were decided through intergovernmental cooperation under the EU third pillar. Consequently, these policy statements refer to activity by the EU that would serve to increase supranational authority over these issues at the expense of national sovereignty. If the national sovereignty model structures citizens’ EU policy space, we would expect these policy statements to all relate to the dimension in the same direction (positive or negative) and to have particularly strong loadings on that dimension. If citizens strictly view EU policy based on their willingness to give up national sovereignty, then this dimension should dominate citizens’ policy attitudes. The regulation model also assumes one dimension that underlies citizens’ positions on the policy statements, but assumes that the dimension is characterized by economic concerns regarding regulation of the economy. The policy statements (i) and (r) most closely capture these concerns, as they call for active policies at the EU level to intervene in economic markets (employment and consumer protection), presumably through regulation in the market place. Thus, if this model defines citizens’ EU policy space, we would expect these statements to relate to the dimension in the same direction and to demonstrate strong loadings on that dimension. 3

Note that we have also conducted the analysis with subsets of these statements as markers for this dimension. The results are consistent with those presented here.

22


The Hix–Lord model posits that two dimensions underlie the covariation in respondents’ policy positions. The first dimension is basically that of the international relations model described above. The second dimension is the primary dimension of domestic contestation in the EU member states, which captures salient social and economic issues. One component of this dimension involves conflict over authoritarian vs. libertarian values. This includes questions of individual freedom vs. traditional practices and “new politics” issues regarding feminism, minority rights, and ecology (Hix 1999a: 73). The other component is the economic conflict regarding market intervention vs. laissez-faire policies toward employment, public ownership, and economic welfare (Hix 1999a: 73). These combine to form a socio-economic left (intervention/libertarian)/right (free-market/authoritarian) dimension. In the set of policy areas available, we expect four policies to serve as markers for the two components of the left/right dimension. The issues of fighting unemployment and giving more help to the poor and socially excluded should identify the economic component and the issues involving equal opportunities should identify the authoritarian/libertarian component. We expect these four markers to relate to the left/right dimension in the same direction and to have strong loadings on this dimension. This model also assumes that these two dimensions are unrelated (orthogonal) to each other. The Hooghe–Marks model also posits basically the same two dimensions as the Hix–Lord model, except they expect the two dimensions to be related to each other. Specifically, they expect left positions on the socio-economic dimension to be related to supranational positions on the sovereignty dimension to form a policy combination characterized by “regulated capitalism.” They also expect positions on the right to be associated with positions on the nationalism end of the sovereignty dimension to form a policy combination characterized by “neoliberalism” (Hooghe and Marks 1999: 77). To capture this distinction from the Hix– Lord model, we estimate exactly the same model as described for the Hix–Lord model but allow the dimensions to relate to each other. Results Tables 1.3 through 1.5 present maximum likelihood results of the confirmatory factor analyses designed to estimate how well the four models of the EU policy space account for the observed structure of policy positions among the EU mass public.4 In addition to the standardized results for 4

We used EQS statistical software.


23

the loadings of the policy statements on each dimension, we also report measures of model fit. Consistent with recommendations by Hoyle and Panter (1995), we use the following fit indices: the goodness-of-fit index (GFI, Joreskog ¨ and Sorbom ¨ 1981), the non-normed fit index (NNFI, Bentler and Bonnett 1980), the incremental fit index (IFI, Bollen, 1989), and the comparative fit index (CFI, Bentler 1990). The values of GFI, NNFI, IFI, and CFI range from 0 to 1.0, with higher values indicating better model fit (Bentler and Bonnet 1980; Hoyle and Panter 1995). These measures are sensitive to the model complexity, so that greater model fit due simply to fewer constraints on the interrelationships of the variables is discounted. Note that the null hypothesis – that there is no structure to mass attitudes toward EU politics – is viable. Certainly the level of public sophistication and knowledge regarding EU policy issues would suggest that citizens lack any structure to their attitudes on EU policy. Model 1 in table 1.3 provides evidence regarding the international relations model. Consistent with expectations, the underlying dimension relates in the same direction (positive) to the policy statements calling for loss of national sovereignty over policy (e.g., creating an EU army). That is, respondents who felt that “having a strong European currency” was a key priority were also likely to consider “setting up a European army for common defense” a key priority. However, the marker policy statements regarding national sovereignty are not among the strongest loading policy statements on this dimension. Thus, we would not consider this evidence of a single national sovereignty defining the EU political space for the mass public. The results in model 1 in table 1.3 also test the regulation model. The marker policy statements for this economic dimension load in the same direction (positive), as expected. For example, respondents who considered developing joint programs to fight unemployment as a key priority were also likely to identify consumer protection as a key priority for more EU activity. However, while some of the marker statements have strong loadings, the statements with the strongest loadings are difficult to characterize as economic in nature. For example, promoting human rights and improving EU transparency are certainly more representative of “new/old politics” issues than economic regulation. Thus, while the substantive character of this dimension appears much closer to the regulation model than the national sovereignty model, the one-dimensional policy space has a broader character than simply regulation. One reason this dimension might capture more than the regulation model is that it reflects the domestic left/right dimension that voters use to organize their domestic political space. As we will discuss shortly, concerns over regulation are an important component, but so are

24


Table 1.3 Factor patterns (standardized solutions) for single-factor models

Key priority for more EU activity Intervene in international conflicts Immigration Protect EU cultures Create EU army Promote social justice Protect EU from imports Protect environment Increase EU transparency Fight unemployment Protect from non-EU competition Adopt a strong European currency Fight drug-trafficking Fight international crime Prevent import of goods made by child labor Promote human rights Increase control of EU external borders Support poorer EU regions Protect consumers Be closer to the people Improve equal opportunities between men and women Create greater opportunities for smaller member states Ensure free movement of EU citizens Give more help to poor and socially excluded Give more help to Third World Improve equality of opportunities for minorities Left/right self-placement GFI NNFI IFI CFI

Model 1

Model 2

One factor

One factor

0.306∗ 0.367∗ 0.444∗ 0.320∗ 0.359∗ 0.390∗ 0.439∗ 0.499∗ 0.453∗ 0.389∗ 0.378∗ 0.360∗ 0.409∗ 0.464∗ 0.571∗ 0.423∗ 0.613∗ 0.575∗ 0.555∗ 0.592∗ 0.609∗ 0.533∗ 0.631∗ 0.530∗ 0.635∗ – 0.884 0.756 0.776 0.776

0.306∗ 0.367∗ 0.445∗ 0.320∗ 0.360∗ 0.390∗ 0.439∗ 0.499∗ 0.454∗ 0.389∗ 0.378∗ 0.360∗ 0.408∗ 0.464∗ 0.571∗ 0.423∗ 0.613∗ 0.575∗ 0.555∗ 0.593∗ 0.609∗ 0.532∗ 0.631∗ 0.530∗ 0.635∗ 0.065∗ 0.885 0.753 0.773 0.772

Note: ∗ p < 0.05.

concerns over social justice and equal opportunity – issue areas that load strongly in model 1. Thus, we were interested in whether this dimension is simply capturing the domestic left/right dimension. To test this, we reestimated model 1 but added a new indicator: respondents’ left/right self-placement. If the dimension in model 1 is capturing the domestic left/right dimension, then we expect left/right self-placements to be very highly correlated with this dimension. Clearly, this is not the case. Left/ right self-placements are very weakly related to this dimension. Model 3


25

in table 1.4 is designed to capture the constraints imposed by the Hix– Lord model. Consistent with expectations, the sovereignty dimension (factor 1) underlies responses to the marker statements in the expected way. Respondents who consider creating an EU army as a key priority also consider having a strong EU currency as a key priority. Put more generally, citizens who support (oppose) increasing supranational political authority at the expense of national sovereignty in one policy area tend to support (oppose) it in other areas.5 Other policies that load strongly on this dimension also support this interpretation of the dimension. The statement about increasing control of external borders refers to an aspect of EU activity that, at the time of the survey, was not part of the first pillar of the EU. Interestingly, this dimension also has a strong impact on respondents’ attitudes toward protecting the EU market from nonEU goods and increasing control of EU borders. Respondents who were pro-integration were pro-protection. The combination of these policy priorities suggests that this dimension is capturing support for building a stronger supranational authority with a “fortress Europe” character. In addition, the domestic left/right dimension (factor 2) has a substantive policy character that is consistent with most of the expectations of the Hix–Lord model. Respondents who supported economic redistribution and government involvement in the economy also set as priorities policies like environmental protection, aid to the poor and socially excluded, and equality of opportunity. However, the dimension does not appear to capture important aspects of the “new politics” component of the socio-economic left/right dimension. The dimension is positively related to both the priorities of environmental protection and increased governmental transparency – “new politics” issues – and the priorities of fighting crime and dealing with immigration problems – two indicators of authoritarian or “old politics” concerns. Thus, this dimension appears to capture the economic component and only part of the social values component of left/right dimension. Finally, the model fit of this factor structure is a small improvement on the one-factor model presented in table 1.3. Model 4 in table 1.5 provides a test of the Hooghe–Marks model. As in model 3, the results are generally consistent with expectations about the sovereignty dimension (factor 1). Also, as in model 2, the results are consistent with most of the expectations regarding the socio-economic left/right dimension (factor 2). But the issues distinguishing “new” from “old” politics are again puzzling. 5

Note that if only the statements that most obviously involve a decrease in national sovereignty (creating an army and having an EU currency) are used to mark this dimension, the results are consistent with expectations.

26


Table 1.4 Factor pattern (standardized solutions) for Hix–Lord model Model 3: two factor

Key priority for more EU activity Intervene in international conflicts Immigration Protect EU cultures Create EU army Promote social justice Protect EU from imports Protect environment Increase EU transparency Fight unemployment Protect from non-EU competition Adopt a strong European currency Fight drug-trafficking Fight international crime Prevent import of goods made by child labor Promote human rights Increase control of EU external borders Support poorer EU regions Protect consumers Be closer to the people Improve equal opportunities between men and women Create greater opportunities for smaller member states Ensure free movement of EU citizens Give more help to poor and socially excluded Give more help to Third World Improve equality of opportunities for minorities GFI NNFI IFI CFI Note: ∗ p < 0.05.

Factor 1

Factor 2

Sovereignty/integration

Socio-economic L/R

0.200∗ 0.279∗ 0.304∗ 0.409∗ 0.086∗ 0.620∗ 0.116∗ 0.237∗ – 0.613∗ 0.402∗ 0.171∗ 0.168∗ 0.316∗

0.221∗ 0.247∗ 0.323∗ – 0.341∗ 0.117∗ 0.406∗ 0.414∗ 0.421∗ 0.120∗ – 0.293∗ 0.347∗ 0.344∗

0.101∗ 0.469∗ 0.087∗ 0.257∗ 0.187∗ –

0.560∗ 0.220∗ 0.620∗ 0.488∗ 0.498∗ 0.601∗

0.148∗

0.582∗

0.190∗ –

0.472∗ 0.686∗

−0.053∗ –

0.604∗ 0.691∗ 0.922 0.811 0.839 0.839


27

Table 1.5 Factor pattern (standardized solutions) for Hooghe–Marks model Model 4: two factors

Key priority for more EU activity Intervene in international conflicts Immigration Protect EU cultures Create EU army Promote social justice Protect EU from imports Protect environment Increase EU transparency Fight unemployment Protect from non-EU competition Adopt a strong European currency Fight drug-trafficking Fight international crime Prevent import of goods made by child labor Promote human rights Increase control of EU external borders Support poorer EU regions Protect consumers Be closer to the people Improve equal opportunities between men and women Create greater opportunities for smaller member states Ensure free movement of EU citizens Give more help to poor and socially excluded Give more help to Third World Improve equality of opportunities for minorities Inter-factor correlation GFI NNFI IFI CFI Note: ∗ p < 0.05.

Factor 1

Factor 2

Sovereignty/integration

Socio-economic L/R

0.211∗ 0.315∗ 0.343∗ 0.398∗ 0.068∗ 0.835∗ 0.086∗ 0.242∗ – 0.827∗ 0.433∗ 0.169∗ 0.157∗ 0.375∗

0.131∗ 0.101∗ 0.160∗ – 0.313∗ −0.309∗ 0.374∗ 0.300∗ 0.435∗ −0.304∗ – 0.219∗ 0.278∗ 0.154∗

0.052∗ 0.568∗ 0.033∗ 0.266∗ 0.172∗ –

0.543∗ −0.056∗ 0.613∗ 0.357∗ 0.419∗ 0.604∗

0.117∗

0.530∗

0.174∗ –

0.397∗ 0.676∗

−0.136∗ –

0.683∗ 0.681∗

0.671∗

– 0.929 0.831 0.857 0.857

28


The interrelationship between these two dimensions is consistent with expectations. The correlation between factors is high (0.671) and positive. This means that priorities of further integration on factor 1 are positively related to priorities for “left” policies on factor 2. Thus, as one’s priorities become less strongly “left” on factor 2 they also become less integrationist on factor 1. This is consistent with the Hooghe–Marks conceptions of two fundamental poles in the EU political space: regulated capitalism and neoliberalism. The model fit indices all indicate that this two-factor model performs better than models 1, 2 and 3. But with such a strong inter-factor correlation, we do not want to exaggerate the substantive or statistical superiority of this model over model 1. In essence there is basically one dimension to the political space. Clearly, model 4 provides evidence of subtle and systematic distinctions in the structure of the EU policy space that would be missed by assuming only one dimension to the policy space. These distinctions are meaningful and provide a clearer interpretation of the policy content of this dimension than one would glean from a onefactor model. These distinctions also help us understand why the onefactor model does not correlate well with citizens’ self-reported left/right positions. As shown in model 2 of table 1.3, respondents’ left/right positions are not strongly related to the dimension structuring their priorities over EU policy. This makes sense if we understand that the primary dimension of the policy space only captures part of the domestic left/right dimension – particularly the economic component. Given that the EU policy dimension also includes concerns regarding national sovereignty and lacks a clear “new politics” component, the dimension is only weakly related to the domestic left/right. Alternative explanations Before concluding, we want to return to our concern about the survey question and how it might influence the results of our analysis. Recall that the survey question asked respondents to identify whether or not they considered each policy statement a “key priority” for more EU activity. The potential problem with this question design is that it asks both whether the respondent agrees with the policy statement and whether the respondent considers it a priority. Thus the question taps salience as well as policy position. Our hope is that, because the question did not limit the number of priorities the respondent could choose, differences in the responses largely reflect policy positions. However, the variation in responses may actually represent differences in citizens’ priorities over areas of EU governance. If respondents do not read the statements closely


29

enough to discern policy direction, they may simply be responding based on their preferences over which issues should be conducted at the EU level. Thus we want to consider whether such preferences could explain the results of the analyses before we accept the interpretations based on the models of the EU policy space. Dalton and Eichenberg (1998) developed three hypotheses regarding how citizens vary in their preferences for EU governance over particular policies. We will examine whether the results of the confirmatory factor analysis are consistent with any of these hypotheses. First, based on functionalist theory, Dalton and Eichenberg (1998: 254) argued that support for EU governance of policy should be greatest (weakest) for issues that are difficult (easy) to solve at the national level and which have clear (dubious) benefits from international coordination. For example, citizens should support EU policies that involve protecting the environment, fighting international crime, and managing immigration, and oppose policies designed to address equal opportunity or protecting consumers. The evidence from model 1 in table 1.3 is inconsistent with this hypothesis. Respondents who considered environmental protection or fighting international crime as priorities also considered improving equal opportunities and protecting consumers as priorities. Second, Dalton and Eichenberg (1998: 255) hypothesized that citizens would support EU authority in areas of “low politics” but oppose EU authority over “high politics” issues. “High politics” includes foreign policy, defense, and control of the national economy. “Low politics” includes welfare policies and tariff policies. Thus, we would expect respondents to identify issues involving the creation of a single EU currency and an EU army as “not a key priority” and issues such as trade policy and environmental policy as a “key priority.” This hypothesis is not consistent with the results of the confirmatory factor analyses. In model 1, the low and high politics policies are positively related. Respondents who consider protecting the EU from imports and protecting the environment as key priorities also tend to feel that the creation of an EU army and having a common currency are key priorities. Third, Dalton and Eichenberg (1998: 255) hypothesized that citizens would support EU governance of policies that they personally benefit from and oppose those policies that decrease their welfare in a utilitarian sense. While it is difficult to identify what this hypothesis implies about variations in respondents’ priorities across policy statements, it seems highly unlikely that the results of the factor analysis are due to such calculations by the respondents. Note that all the policy statements load positively on a single dimension (see model 1). The only way that the utilitarian hypothesis could account for this is if respondents

30


generally felt they benefited from all of these policy statements. This seems very unlikely, particularly since some of the policy statements concern redistribution of resources. Thus we are confident that the results of the confirmatory factor analysis cannot be explained by alternative explanations related to these three hypotheses about citizens’ preferences over EU policy authority. Discussion Returning to our original concern, the results presented here indicate that citizens’ attitudes toward EU policies are far from random. In fact, we find systematic evidence that the cognitive map of European voters has a meaningful structure that is theoretically interpretable. Taken together, our results indicate that the European political space is basically one-dimensional. However, the substantive meaning of this dimension requires an appreciation for the underlying two-dimensional structure. There appears to be both a left/right and a supranational component to citizens’ policy preferences. These dimensions are strongly related, so they effectively merge to form one dimension of the EU policy space. This interpretation of the EU policy space at the mass level is very similar to the interpretation of the EU policy space at the party group level presented in Gabel and Hix (chapter 5). There are several questions this analysis was not designed for and therefore could not address. For example, we do not know whether the attitudinal structure uncovered here is stable over time; that is, whether this structure is a recent phenomenon or whether Europeans’ attitudes have been structured this way for a long time. Moreover, this study was not designed to examine the structure of attitudes within member states. Our working assumption has been that the analysis of a truly European policy space requires the analysis of the European electorate as a whole. It is possible, if not likely, that there are cross-national differences that may be worth investigating. However, these should not detract from the finding that Europeans as a whole conceptualize the space of contestation in the EU in one dimension along a European left/right dimension. And it is difficult to explain the results reported here with a story about widely varying structures underlying national publics’ views on EU policy. We accept that the degree to which the findings here speak to any national public may vary in degree, but it is unlikely that national publics differ dramatically in the structure of the EU policy space from this model. The results presented here should also be reassuring for those worried about representation in the EU. Even when people are not extensively informed about the details involved in governing the Union, they use the


31

shortcuts they know to make sense of EU policy issues. These shortcuts are akin to the shortcuts they use to make sense of politics at the national level, but they come with a distinctly European flavor. Thus people simplify the complexities of European-level politics in a way that makes sense to them. In addition, our results should be comforting to those interested in reforming the democratic institutions of the united Europe. They suggest that voters’ conceptual map is not random, and that the considerable political heterogeneity that exists across the member states does not necessarily pose an obstacle to achieving meaningful democratic representation in the European Union. Recall that the mass policy space identified is the same policy space that Hix (1999b) contends structures competition among political parties at the EU level. If this structure endures at the mass level, existing parties will not have to reinvent themselves to contest European elections when truly European issues come to the forefront of voters’ interests. As a result, building a European party system may not be as difficult as it may seem at first glance. However, a fairly stable set of ideological constraints on how people view European policy issues also means that parties may have a difficult time manipulating the policy space at the mass level.

2

Potential for contestation on European matters at national elections in Europe Cees van der Eijk and Mark N. Franklin

That citizens of European Union countries differ in their attitudes regarding Europe is a commonplace of political commentary. Some favor their country’s membership in the EU, others oppose it. Some, while thinking that membership is generally a good thing, feel that steps toward unification have gone far enough – or even too far. Others believe that further steps should be taken. Citizens of EU countries also differ in terms of more traditional political orientations – attitudes to the proper role of government in society, welfare provision, and other matters which have increasingly over the past half-century come to be subsumed within a single orientation towards government action, generally referred to as the left/right orientation (Lipset 1960; Lijphart 1980; Franklin, Mackie, Valen, et al., 1992). These two orientations are often assumed to be orthogonal, with the newer pro-/anti-EU orientation cutting across the more traditional left/right orientation (see, e.g., Hicks and Lord 1998; Hooghe and Marks 1999). Our own research (van der Eijk and Franklin 1996; van der Eijk, Franklin, and van der Brug 1999; van der Brug, Franklin, and van der Eijk 2000) demonstrates that EU orientation does not currently have much impact on party choice at EU elections.1 Elections to the European Parliament have been described as “second-order national” elections at which the arena supposedly at issue (the European arena) takes second place to the national arena as a focus for issue and representational concerns (Reif and Schmitt 1980; Reif 1984, 1985; Marsh and Franklin 1996); and the national arena is quintessentially one in which left/ right orientations dominate. Nevertheless, in this chapter we will argue that the pro-/anti-EU orientation, despite its apparent irrelevance for political behavior, constitutes something of a “sleeping giant” that has the potential, if awakened, to impel voters to political behavior that (because of its degree of orthogonality with left/right orientations) 1

The analyses referred to pertain to the European Parliament elections of 1999, and are based on surveys from the electorates of all member states (see below for details).

32

European matters at national elections

33

undercuts the bases for contemporary party mobilization in many, if not most, European polities. We will also show that the prerequisites are present for these consequences to show themselves under certain specifiable conditions. The structure of this chapter is as follows. After introducing our topic and the data we employ, we describe the orientations of voters in the two dimensions of concern to us in this chapter. We will show that, at least at the level of voter orientations, the pro-/anti-EU dimension (no less than the left/right dimension) represents definite differences between voters. Such attitudes may as yet have little political effect, but they are not non-attitudes. Then we will turn to parties, and describe how parties are positioned in terms of these two dimensions. Here we will see that parties (except for some small ones) offer only limited choice in terms of pro- and anti-European orientations. We will argue that this is one of the reasons why the differences between voters on this dimension have no chance of influencing policy-making. In the final section we will describe the conditions that we feel would have to pertain in order for the EU dimension to become influential and the circumstances under which we believe that these conditions will occur. In our opinion the sleeping giant is stirring even now. Initially we employ two measures of EU orientation (which will be described in some detail in the next section), both of which are found to be largely unrelated to the standard measure of left/right position (see table 2.1). The middle panel in table 2.1 (derived from the question “Do you think that the European Union is a good/bad thing?”) shows least correlation with left/right position in any country, with only four significant relationships. The left panel (derived from the “Some people say European unification has gone too far. What do you think?”) shows slightly higher correlations. Nevertheless, even those correlations are generally very low, with none even as strong as 0.3.2 The good/bad variable is one that has been used ever since the earliest days of research into public attitudes toward Europe. The other variable is of more recent vintage. In this chapter we make greater use of the newer variable because it is a more sensitive measure (using a ten-point scale rather than a three-fold distinction) and because we have corresponding data on where voters perceive the parties to be, which we do not have for the other variable. Table 2.1 demonstrates that the pattern of relationships for this newer variable across countries is very similar to the pattern found for the more established measure. The final two columns show the intercorrelation 2

The – by and large – opposite signs of relationships with these two variables reflect the fact that their codings run in opposite directions.

34

Cees van der Eijk and Mark N. Franklin

Table 2.1 Correlations between left/right and pro-/anti-EU measures

Left/right and more/less EU integration

Left/right and EU good/bad thing

More/less EU integration and EU good/bad thing

Correlation p-value

Correlation p-value Correlation p-value

Austria Belgium (Flanders) Belgium (Wallonia) Britain Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Sweden

−0.136 −0.286 −0.079 −0.099 0.145 0.023 −0.102 0.019 0.012 0.045 0.122 −0.033 −0.100 0.067 0.020 0.253

0.000 0.000 0.313 0.005 0.000 0.663 0.005 0.573 0.805 0.348 0.000 0.599 0.003 0.213 0.565 0.000

0.056 0.196 0.034 0.099 −0.176 −0.029 0.058 0.03 −0.074 0.077 −0.011 0.060 0.049 −0.051 −0.104 −0.274

0.886 0.751 0.244 0.063 0.000 0.536 0.481 0.549 0.833 0.090 0.463 0.462 0.669 0.926 0.023 0.000

−0.489 −0.586 −0.303 −0.474 −0.485 −0.391 −0.358 −0.445 −0.300 −0.290 −0.269 −0.340 −0.338 −0.297 −0.323 −0.543

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

EU-15

−0.01

ns

−0.01

ns

−0.423

ns

between the two EU measures in each country. Over all countries taken together the correlation is 0.42.3 Of course, the low correlations shown in table 2.1 could result from the EU measure having no meaningful content, and we will need to start by considering this possibility; but, assuming for the moment that the correlations are not just manifestations of random data, they indicate whether and how, in various systems, EU attitudes are related to left/right ones. In most systems the correlations are negligible, as already pointed out. Exceptions are Flanders and to a lesser extent Austria (and possibly France, the Netherlands, and Britain), where negative feelings toward the EU are mainly harbored by right-wing people. In Denmark and Sweden (and to a much lesser extent in Spain), however, negative feelings toward the EU are mainly harbored by left-wing people. The differences between these two groups of countries point to the fact that, to the extent that differences in opinion about EU integration 3

Correlations are quite low, because about 70 percent of respondents answer “good” on the second of these variables, whereas these responses are distributed across the higher values of a ten-point scale on the first variable.


35

are being politicized, there is no “natural” connection of “what goes with what.” It is the dynamics of the domestic political arena that here and there brings forth a connection with either the right or the left (there is also the possibility of a connection with both right and left – a curvilinear relationship – which we will return to in a later section of this chapter). This means that any such “colonization” of contestation over EU integration threatens the perspective that we sketched in Choosing Europe? (van der Eijk and Franklin 1996), namely that of a common European electorate that responds in common ways to political problems and political stimuli.4 We will have reason in later sections to refer back to the occasionally non-zero correlations between left/right and EU orientations, and the fact that these correlations are not all in the same direction. Data Our data are taken from the voters’ component of the European Election Study, 1999.5 These data contain two measures of EU orientation. As already mentioned above, one is a ten-point scale on which respondents were asked to indicate whether European unification had gone too far, or 4

5

Although we have in the European Election Study of 1999 (EES99) no other issue variables of relevance, there are no indications from other comparative research (or from previous EESs) that the contents of the left/right dimension are fundamentally different between the countries of the EU (everywhere, left is associated with equality, state intervention in the economy, support for the welfare state, feminism, ecology; whereas right is everywhere associated with freedom, support for restricted government, free markets, etc.). This study was designed and organized by an international group of scholars that initiated similar studies on the occasion of the 1989 and 1994 European Parliament elections. Members of the 1999 group were Wouter van der Brug (Amsterdam), Pilar del Castillo (Madrid), Roland Cayrol (Paris), Cees van der Eijk (Amsterdam), Mark Franklin (Hartford CT), Soren ¨ Holmberg (Goteborg), ¨ Renato Mannheimer (Genoa), Michael Marsh (Dublin), Jacques Thomassen (Enschede), Colette Ysmal (Paris), Bernhard Wessels (Berlin), and Hermann Schmitt (Mannheim) who coordinated the efforts of the group. The study consisted of a single cross-sectional survey fielded in each member state of the EU immediately after the European Parliament elections of 1999. The questionnaires were administered in the language of each country by telephone interviews (in Italy by way of telepanel) with sample sizes of roughly 1,000 in some countries (Britain, Denmark, France, Germany, Netherlands, and Spain), roughly 500 in others (Austria, Belgium, Finland, Greece, Ireland, Portugal, and Sweden), of 300 in Luxembourg and of 3,707 in Italy. The study was funded largely by the University of Amsterdam and the Dutch National Science Foundation, with important additional contributions from the Spanish CIS, the universities of Mannheim and Genoa and Trinity College, Hartford CT. The data were deposited in 2002 at the ICPSR, University of Michigan, and at the Steinmetz Archives in Amsterdam.The study is extensively documented on the European Elections Studies web site (http://www2.trincoll.edu/∼mfrankli/EES.html). This site not only describes this and earlier European Parliament election studies, but also contains an extensive list of publications emanating from these studies.

36


had to be pushed further. The other is a three-point scale on which respondents were asked whether their country’s membership of the EU was a good thing, a bad thing, or neither good nor bad. The second of these questions, suitably recoded to place “neither good nor bad” in the middle of the scale, is the same question as has been asked in Eurobarometer studies since 1970 and which, in other research (Franklin and Wlezien 1997), we have found to perform as a valid measure of preferred position on European integration. That is, people who say membership is a good thing favor a high level of integration, those who say it is neither good nor bad favor a medium level and those who say it is a bad thing want a low level or no integration at all. So, in terms of measuring a preferred position, the variable is closely equivalent to our measure of left/right position, though measured only on a three-point scale in contrast to the ten-point scale that we have for left/right. Note that someone who says they want a low level of unification might or might not want actions to be taken that would change the extent of unification currently in place, as is true of those who say they want a high level of unification. The other variable (responses to the “gone too far” question), though measured on a ten-point scale, is a measure of policy preference. Those who pick a low value want actions to be taken to reduce the current level of unification, while those who pick a high value want actions to be taken to increase the extent of unification. Because this variable contains a policy implication, we would expect it to be somewhat related to the left/right measure, at least to the extent that EU integration has been politicized in party-political terms. After all, left/right position is the dominant cleavage in European politics, and political scientists have argued since at least the time of Schattschneider (1960) that for a concern to have policy implications it needs to have been incorporated at least to some extent into the dominant dimension of current political contestation. Thus significant correlations were to be expected in table 2.1, where left/right orientation was correlated with this variable. Indeed, it is surprising that these correlations were so weak and that so few of them were significant. In addition to these two measures we have a set of variables that are very similar to the “gone too far” measure but which relate not to respondents’ own positions but to their perception of the positions taken in EU integration by the parties that compete for their votes – a measure for each party, scored on the same ten-point scale as the measure of the respondents’ own orientations. We also have a set of measures of the perceived left/right position of each party, which correspond to the left/right position of respondents in much the same way. These are the variables we will employ in the present chapter.


37

The nature of voters’ EU orientations Table 2.2 shows a number of univariate statistics relating to the left/right and “gone too far” measures, using the former as a point of reference since it is a ubiquitous scale which is understood to tap a real-world phenomenon. For each variable the table shows three such statistics: the percentage of respondents who failed to answer the question and, for those who did answer, the extent to which they are in agreement about this answer (the peakedness of the distribution)6 and the average extremity of their answers (the difference between the median answer and the midpoint of the scale). Several things stand out from this table. Most notably, the percentage of missing data for the EU integration scale is little more than half what it is for the left/right scale. This implies that what is tapped by the scale is not so moot that people have difficulty placing themselves. Quite the contrary. The second thing to note is that agreement in terms of where people place themselves on the left/right scale is much greater than for the “gone too far” scale. Third, on average respondents place themselves further away from the midpoint of the ten-point response scale (i.e., from 5.5) on the “gone too far” scale than on the left/right one. In other words, their positions with respect to EU integration are (on average) more “extreme” than for left/right.7 These coefficients make two things clear. First, the measures we have of EU orientation are unlikely to have emanated from the kind of random response process that would be triggered by non-attitudes: that is, when people have no “real” orientations. If people were pulling their preferences regarding the EU out of thin air then we would see much more missing data for the EU variable than for the left/right variable. In fact we see less. Moreover, random response would result in a combination of low agreement and low extremity. In fact, we find that this combination is more characteristic of the left/right scale, the validity of which is not in question. The combination of lesser agreement together with greater extremity is particularly indicative of real attitudes rather than random response. Second, the low degree of agreement regarding their self-placements signifies that people do differ considerably among themselves as to whether EU integration has gone too far or not. The 6

7

This coefficient attains a value of 1 where all respondents place themselves at the same point on the scale; it takes on a value of zero if they are uniformly distributed over the entire scale; and it takes on a value of −1 when half the respondents place themselves at each extreme of the scale. See van der Eijk (2001). One reason for this is precisely the fact that they do not make party choices based on this variable. Not having to make choices, they have found no need to temper the extremity of their positions. Thus the differences between the two variables would likely be attenuated if voters started to make real choices based on pro- and anti-EU attitudes.

Across-system average

Austria Belgium (Flanders) Belgium (Wallonia) Britain Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Sweden

501 274 226 978 1001 501 1023 1001 500 502 3707 301 1001 499 998 505

N

8.82

3.39 7.66 10.18 7.57 6.69 16.97 12.41 1.40 7.60 2.99 22.85 9.30 4.40 13.83 6.91 6.93

Percentage missing

0.18

0.15 0.19 −0.07 0.15 0.13 0.19 0.22 0.25 0.29 0.20 0.27 0.23 0.35 0.07 0.09 0.13

Agreement

– EU integration scale –

0.86

0.54 0.39 0.04 0.74 0.43 0.76 1.12 0.85 2.50 0.08 1.10 1.62 1.14 0.86 1.20 0.46

Extremity

14.68

12.57 25.55 23.45 11.96 8.79 16.77 18.77 6.49 8.80 10.76 25.65 11.63 11.29 22.04 13.83 6.53

0.35

0.52 0.45 0.33 0.44 0.36 0.34 0.31 0.45 0.07 0.45 0.22 0.43 0.34 0.48 0.22 0.17

Agreement

– Left/right scale – Percentage missing

Table 2.2 Characteristics of voters’ self-placement on EU integration and left/right scales

0.33

0.38 0.03 0.50 0.46 0.09 0.01 0.56 0.54 0.25 0.24 0.33 0.40 0.12 0.33 0.63 0.32

Extremity


39

smaller this coefficient, the higher the potential for political contestation. After all, if agreement is high in terms of where people put themselves, then there is not much to fight about politically. On top of this, the higher extremity scores for the EU measure (more than twice as high as for left/right) indicate on average a greater intensity of feeling regarding this issue. Evidently, the ingredients for contestation over EU integration are even more powerful than over the more tradional issues that are subsumed under the left/right divide. The greater dispersion of attitudes regarding Europe as compared to left/right attitudes can be seen particularly clearly if we graph the joint distribution, which is done in figure 2.1. In this graph, voters are located according to their support for different political parties, which provides us with a “quasi-party space.” The greater dispersion of opinions regarding Europe is clear in this graph from the fact that points are much more spread out in the vertical than in the horizontal dimension; actually the picture shows a bimodality in the vertical dimension as opposed to a more unimodal clustering in the middle of the horizontal one. It is also evident that there is no clear curvilinear relationship between the two dimensions. Extreme attitudes toward Europe are not more evident at any particular point on the left/right scale.8 The potential for contestation over issues regarding Europe is clear from these findings. The important question is why such potential for contestation has not so far been realized in elections to the European Parliament, or, for that matter, in the domestic political arenas of the EU member states. Variety of choices on offer One possible reason for lack of contestation would be that political parties do not provide the vehicles for contestation. If voters are not offered a choice between different visions of Europe, whatever the differences between voters, these cannot be expressed in their choices of parties to support. One of the questions we may therefore ask is to what extent voters have a choice between parties in terms of the differences in their positions. Is there a large variety of choices, or is there only a choice between Tweedledum and Tweedledee? 8

The same picture is obtained by graphing the prevalence of each conjunction of positions regarding Europe without reference to parties, but that graph consists of equally spaced points that differ in their weight. However we represent these different weights (by shading or by size of points), the picture is an unfamiliar one that is harder to interpret than the picture we have chosen to present in figure 2.1.

40


Figure 2.1 Voter positions in left/right (horizontal) and less/more EU integration (vertical) terms. Note: All EU countries together. Local linear fit line added. Each symbol in the graph signifies the median positions on the two dimensions of the potential electoral supporters of a particular party. Who is counted as a supporter and who is not depends on the response to the question “How likely is it that you would ever vote for the party in question?” Answers to these questions range from 1 (“never”) to 10 (“will certainly vote for this party”). Respondents were weighted according to these responses: answering 1 (“never”) resulted in a weight of 0, as these respondents cannot in any way be considered as electoral supporters of the party in question. An answer of 10 (“will certainly vote for this party”) yields a weight of 1. Intermediate answers were associated with weights that were linearly interpolated between these extremes.

This question is of relevance in the context of political contestation, as parties are important vehicles for expressing political disagreements, and for mobilizing citizens in this respect. Of course, parties are no longer the exclusive actors in this regard, as social movements, interest groups, and media also play a role in mobilizing and expressing contestation. However, for the “authoritative allocation of values” and public policy,


41

these other actors depend on political parties (and the governments that they populate). So when there is little difference between parties, there is also little room for expressing differences in opinion in a politically potent manner. Perhaps existing parties should in such circumstances fear the potential lure of new parties catering to these different preferences; but how easy it is for newcomers to gain access and representation depends on institutional arrangements. The easiest way to discover whether parties are offering a relevant choice regarding European integration is by considering the range of party positions in terms of this issue. We can measure these positions on the basis of voters’ perceptions of where the parties stand.9 Following the same logic as before, we will consider these measures both for EU position and also for left/right party location. The dispersion of party locations across each of these dimensions in the member states of the European Union tells us how much choice is on offer on that dimension in that country. The degree of dispersion (which we measure in terms of variance) is not, however, immediately comparable between political systems, owing to differences in numbers of parties whose locations we know; but within each system the dispersion of the parties can easily be compared as between the two dimensions. Moreover, these variances can be computed in two different ways. One way is to have each party contribute equally to them, which is sensible from the perspective of choices on offer (irrespective of whether a party is small or large, it offers a choice to the voters). The other way is to weight parties according to their size (measured as percentage votes in the last national election), which makes sense from the perspective of political culture, political communication, and “mobilization of bias” (Schattschneider 1960). The reasoning that leads us to employ the second measure in addition to the first is that larger parties contribute (on average, and other things being equal) more than smaller ones to the dominant climate of opinion and to what is considered “mainstream.” Table 2.3 displays these variances, computed in both these ways, for each system of the EU separately as well as for the EU as a whole (meaning for all parties for which these perceptions have been obtained in all EU countries). The table shows that, in general (and regardless of whether 9

Of course, one may wonder how well these perceptions reflect actual positions of parties regarding policy. In terms of the motivations of voters to choose a particular party, this question is of little relevance. Analysis of the 1994 election study indicates that such perceptions may differ considerably from other indicators of party positions, particularly in the case of issues that are hardly politicized (cf. van der Brug and van der Eijk 1999). In all analyses to follow, when parties are characterized by perceptions, it is the (interpolated) median of individual respondents’ perceptions that is used.

42


Table 2.3 Variance in party positions on left/right and EU orientation – Parties weighted on size –

– Parties weighted equally –

Integration

Left/right

Difference

Integration

Left/right

Difference

Austria Belgium (Fl.) Belgium (W.) Britain Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Sweden

1.57 0.16 2.31 0.55 4.03 0.08 1.02 1.52 2.52 0.60 0.48 0.91 0.53 3.27 1.16 4.17

1.89 4.07 5.46 1.14 3.66 4.18 5.82 2.80 6.69 1.28 4.88 1.87 2.36 3.28 5.88 4.60

0.32 3.91 3.15 0.59 −0.37 4.10 4.80 1.28 4.17 0.68 4.40 0.96 1.83 0.01 4.72 0.43

1.17 0.15 5.08 1.46 6.36 0.68 1.46 4.50 3.51 0.99 0.43 1.67 2.16 7.33 3.22 5.05

2.15 4.65 7.18 1.44 5.80 3.43 8.90 7.33 6.73 2.02 3.53 3.36 3.92 9.40 6.27 5.77

0.98 4.50 2.10 −0.02 −0.56 2.75 7.44 2.83 3.22 1.03 3.10 1.69 1.76 2.07 3.05 0.72

Entire EU

2.13

3.50

1.37

2.53

4.54

2.01

they are weighted by size) parties offer voters more choice on the left/ right dimension than on the EU dimension (see the columns labeled “difference” where positive values signify a larger dispersion of parties on the left/right dimension, while negative values show a larger dispersion on the EU dimension). In each subtable there is only one country (Denmark) in which parties differ more on the EU dimension than on the left/right dimension.10 This finding helps us to understand why, in most countries, there is so little contestation on European issues at European Parliament elections, but it makes it even more remarkable that the preferences of individuals were found in the previous section to be so much more diverse on the EU dimension than on the left/right dimension. Comparing the weighted and unweighted variances by country shows to what extent the palette of choice on each of these dimensions is dependent on small parties. On the EU dimension, small parties are responsible 10

This is owing to the inclusion of the two Euroskeptic movements ( Juni Bevaegelsen and Folke Bevaegelsen) that compete only in European Parliament elections. When weighting parties by size, Portugal shows no real differences in variance between the two dimensions. Unweighted, there is more of a difference in Portugal, but less of a difference in Britain (owing to the effect of the tiny UKIP, positioned extremely against EU integration in 1999).


43

for a lot of the variance in party positions in Wallonia, Denmark, Germany, Portugal, and Spain. On the left/right dimension, small parties are responsible for a lot of the variance in party positions in Portugal, Germany, France, and Denmark. In all other countries, the smaller parties usually offer some, but not very many more choices (in terms of variance of party positions) than the larger parties do. A second way to look at the contribution of parties to contestation over European matters is to consider the extent to which the party space corresponds to the attitude space in two dimensions. Although we saw in our introductory section that attitudes on European matters are largely orthogonal to attitudes on matters bound up with left/right orientation, this might not be true for parties. Figure 2.2 displays a number of thumbnail graphs, one for each country, in which each party is plotted according to its location in left/right and EU terms. As can readily be seen, for most countries the dispersion of parties in terms of left/right (across the graph) is considerably greater than the dispersion of parties in terms of attitudes to Europe (up and down the graph). In Finland, Flanders, France, and Italy, indeed, the parties are spread out in virtually a straight horizontal line across the center of the graph. In several other countries (Austria, Britain, the Netherlands, and Wallonia) there is a virtually amorphous scattering along the center of the graph. In some cases (Denmark, Greece, Ireland, Luxembourg, Spain, and Sweden) the willing eye can discern a positive relationship between the two dimensions, such that left parties are more likely to be Euroskeptical while right parties are more likely to be in favor of European integration. In Portugal, by contrast, the pattern seems to resemble an inverted U, with extreme parties of both left and right being more likely to oppose European unification. If we look hard for this pattern we seem to see it also in some of the other countries: France, Germany, the Netherlands, and even Denmark. Indeed, traces of an upside-down U can be seen in most of the countries where the party spread encompasses both extremes of the left/right scale. This suggests that the apparently different patterns in the different countries are in reality due to the fact that specific countries seldom have a complete set of all the parties that might have occurred there. Some countries have no party of the extreme right; other countries have no party of the extreme left. Consequently, what we could be seeing in these thumbnail graphs are country-specific projections of a common pattern filtered in each country by the party system that exists there. To assess this conjecture, in figure 2.3 we superimpose the thumbnail graphs from figure 2.2 on top of each other, producing a single graph for all countries of the EU in 1999. This single graph plots the location in left/right and pro-/anti-EU terms of all parties about which we have

44


Figure 2.2 Parties’ positions on left/right (horizontal) and anti-/prointegration (vertical) dimensions.

the necessary information. There we can see that our conjecture to some extent holds good, though there seem to be two patterns in that graph. One pattern is of a positive relationship between the two dimensions, as seen more clearly in some of the thumbnail graphs of figure 2.2, where more right-wing parties are more inclined to favor European integration. The other pattern is of the inverted U, where extreme parties at both ends of the left/right spectrum oppose integration. The latter pattern is the clearer one. Except for one party in each of Greece and Sweden (see figure 2.2), the inverted U would have been quite clear. Earlier we saw no evidence of a tendency among voters for pro-/antiEuropean attitudes to be related in any particular way to left/right attitudes. The positive relationship and inverted U that we see among parties


45

Figure 2.3 Party positions in left/right (horizontal) and less/more EU integration (vertical) terms.

was not similarly evident among voters. More specifically, we found that over the full range of the left/right scale voters vary considerably in terms of their European attitudes. This is not true for parties – and especially not within any particular political system, as was apparent in the thumbnail graphs presented in figure 2.2. For whatever reasons, voters at all positions of the left/right scale are offered little choice with respect to EU integration by the parties in their systems. So voters are forced to choose between either expressing in their party choice their left/right ideological concerns, which forces them to ignore their preferences regarding European integration, or the other way around: choosing parties on the basis of their positions regarding Europe, at the cost of not being able to select the party they would have preferred in left/right terms. Indeed, the situation is even worse than initially appears, since all parties are not equal. Most European countries have a very few large parties together with a larger number of small ones; and voters often have reservations about supporting parties that have little chance of influencing government policy. Parties that are anti-Europe are particularly likely to

46


Figure 2.4 Party positions in left/right (horizontal) and less/more EU integration (vertical) terms; parties weighted by size.

be small parties. Thus figures 2.2 and 2.3 mislead to some extent by portraying all parties as though they were equally viable as choices for all voters. In figure 2.4 we provide an alternative perspective by portraying the parties in terms not only of their relative locations but also in terms of their relative weight measured as share of the vote in the most recent national election. Larger parties are indicated by larger circles. In figure 2.4 the prominence of the inverted U is much less than in figure 2.3. However, the figure still shows a slight positive relationship between left/right and EU position for political parties. Evidently political parties of the left are less likely to be pro-EU, consistent with the findings of other chapters in this volume, though the relationship is not a strong one. The extremities of the inverted U (where most anti-European parties are found) turn out to be populated largely by small parties. In view of the variation between their actual and potential voters, it is something of a riddle to find that large, moderate (in left/right terms) parties do not compete on the basis of the EU dimension. To the extent that they fail to do so, each of them is supported by a constituency of voters that contains deep rifts on matters European, as pointed out in previous research


47

(van der Eijk, Franklin et al., 1996). For the moment those constituencies appear to be docile in the face of their lack of choice but, given the strongly held nature of these opinions, the question is how long this state of affairs can endure. This question relates to the image of the “sleeping giant” that we introduced early in this chapter. The Sleeping Giant Since we view European parties through the lens provided to us by European voters (our measures of party position are those provided by respondents to our survey), it is clear that European voters feel that they cannot generally choose a party on the basis of its EU positions while also maintaining their ability to choose a party on the basis of its position in left/right terms. At present in most countries voters are willing to put their preferences regarding the EU on ice and make their choices between parties on other grounds; but how long this state of affairs can continue is a question that does not have an unequivocal answer. Certainly it would be a brave commentator who would assert that the present state of affairs can continue indefinitely. Already the pro-/ anti-EU policy dimension appears much more ripe for politicization (in terms of the number of voters who hold positions and the extremity of these positions) than does the left/right policy dimension. This being the case, it is surely only a matter of time before policy entrepreneurs in some countries seize the opportunity, presented to their parties by these quite polarized opinions, to differentiate themselves from other parties in EU terms. Indeed, this already appears to have happened in some countries where small parties of the far left or far right have already taken up distinctly pro- or (more often) anti-EU stances. In some countries these stances even appear to have paid electoral dividends by attracting voters who would not otherwise have voted for a party at the extreme of the left/right spectrum. As EU matters become more politicized, this could result in still more voters being pulled toward party choices they would not have made on the basis of left/right concerns. This could have unfortunate consequences for the development of a European party system. We saw in table 2.1 that the more policy-relevant measure of preference for unification was somewhat more orientated in left/right terms than was the less policyrelevant measure. We also saw that there was considerable incoherence (as between different countries) in the left/right and pro-/anti-EU linkage (some of the non-zero correlations in table 2.1 were positive, some negative). So the increasing politicization of EU issues might ironically interfere with the development of a truly European party system – see the discussion in van der Eijk and Franklin (1996, chapter 21). It will

48


also make it more and more difficult for party groups in the European Parliament to maintain the cohesion that they have so far been able to maintain on the basis of similarities within groups in left/right terms (cf. Thomassen and Schmitt 1997). What are the preconditions for greater politicization of EU matters? We can think of four different sets: the activities of elites, external events, chance, and shifts in voter orientations. The last of these has already happened. As documented extensively in this chapter, the preconditions for voter attention to pro- or anti-EU arguments are already present. But voter readiness is not enough. Some policy entrepreneur or entrepreneurs will have to come along who are willing to capitalize on these preconditions in order to win votes that otherwise would have gone elsewhere. The only reason this has been slow to happen is because established elites have so far been largely able to prevent it from happening. Existing parties are very resilient. Existing party elites know which side their bread is buttered on. As argued extensively in van der Eijk and Franklin (1996: 370–1), all party leaders understand from long experience the danger to them of an issue that is not already integrated into the left/right spectrum of concerns. European issues have as a matter of fact proved easier to manage than other new issues in most European countries (the exception is probably Denmark). Environmental, feminist, gay rights, and energy issues have provided far greater headaches to party elites in most countries than European issues, no doubt mainly because the EU has in most countries successfully been presented as a matter of foreign policy, which is in all European countries a branch of policy-making traditionally left in the hands of governments. But one only has to read the newspapers to see how issues that were once viewed as purely foreign have become domesticated in recent decades. Moreover, as the EU moves more and more into areas that impinge on the lives of individuals, it becomes harder and harder for governments to argue successfully that these are areas in which they should be allowed free rein. Government ministers do not make matters easier for themselves in this regard when they return from Brussels blaming the EU for domestic policy repercussions that were in fact the consequences of their own policies. By hiding behind the EU as a shield from voters, governments increase the likelihood that voters will respond to any proposals that may be made to address that shield directly. From where might such proposals come? We have to remember that political elites are not monolithic. Though the leaders of large parties have an interest in keeping European matters off the political agenda, the same is increasingly untrue of the leaders of small parties (accounting for the


49

anti-European policies of so many small parties in figure 2.4), or of challengers from within large parties, as we saw in the British Conservative Party in 1997, when the pro-European leadership of John Major was successfully challenged by an anti-European upstart, William Hague, who succeeded to the leadership after the party was decimated in the 1997 British general election. Such challengers, where successful, can capture a major party and use it as a vehicle for gathering anti-European support. This attempt has so far been unsuccessful in Britain because the antiEuropean leader is inept and is repeatedly outmaneuvered by his Labour counterpart. But it is only a matter of time before a chance occurrence or an external event changes the equation in favor of the Tory leader. And there is plenty of time for this to happen. The problems generated by the evolution of the European Union are not going to go away, much as the established leaders of large parties might wish that they would. On the contrary, the evolution of the EU in terms of its internal market and in terms of its external enlargement is guaranteed to generate contentious political issues within the member countries for years (probably generations) to come. Think only of the domestic political impact of large numbers of legal migrants into Germany, Holland or Britain from Poland or Slovenia. When the British Prime Minister of the late 1950s, Harold Macmillan, was asked by a young reporter what had been the greatest challenges to his leadership, his answer was “Events, dear boy; events.” Events are going to be the greatest challenge to existing major party leaders and to their success in keeping European issues off the domestic political agenda in their countries. Some of these might be chance events that we cannot possibly anticipate. Others are only too easy to anticipate: labor market stresses, as already mentioned; taxation problems; energy crises; knockon effects from the single currency in most EU countries. The potential sources of issues to politicize the pro-/anti-EU dimension are limitless. In some country at some time they will result in a new alignment of political forces, either by giving support to what is now a minor party (or to a party that arises in response to the new issue) or by giving support to opposition forces within an existing large party. And if this happens in one country, the contagion effect on other countries can be expected to be virtually immediate. One development that has already changed the balance of forces in Europe is precisely the fact that developments in any EU country that might be of interest to another member country are immediately reported there. One has only to think of the wildfire spread of direct action in response to the gasoline crisis in the fall of the year 2000.

50


Of course, even when the European dimension becomes politicized in this way, the effect will not be permanent. The left/right dimension, because it subsumes so many issues of continuing importance, will certainly retain its dominance. The role we expect to see for the pro-/anti-EU dimension in Europe is one similar to the role played in the United States by the dimension of states’ rights versus equal application of the laws. By shifting attention to this issue dimension, American politicians can often change a minority position to a majority position. Thus the battle for supremacy in that nation is often a battle about how an issue will be framed. We expect the same thing to happen in Europe – perhaps not immediately and perhaps not everywhere, but sometime soon and somewhere not a thousand miles from Brussels.

3

Don’t rock the boat: expectations, fears, and opposition to EU-level policy-making Leonard Ray

The proper extension of the power of the EU – the exact domains in which decisions should be taken at the European level rather than by national or subnational authorities – remains a contentious issue. While the existence of the EU is uncontested, the scope of EU authority is an element of political dispute. Some scholars have examined this new political question and tried to make sense of it as a novel political issue, weakly or not related to traditional ideas and ideologies. Public opinion research on the European electorate has found support for European integration to be organized along two separate dimensions, one left/right, the other pro-/anti-Europe (Hix 1999b; Gabel and Anderson, this volume). Other research on political parties has demonstrated that the issue of European integration fits into existing political cleavages and Weltanschauung. Marks, Wilson, and Ray find that support for the EU is related to traditional party families (2002). Hooghe, Marks, and Wilson (chapter 6, this volume) find a relationship with the “new politics” dimension. Data on the European Parliament also suggest the importance of the left/right dimension in structuring contestation within this eminently European body (Thomassen, Noury, and Voeten, chapter 7). I argue that a new European dimension in the political landscape is neither unrelated (orthogonal) to the traditional ideologies which have structured European politics, nor coterminous with them. Instead, the relationship between ideology and support for EU decision-making varies systematically according to the expected impact of EU decision-making on policies voters hold dear. A policy expectations account of support for EU-level policy-making Previous work has noted a shift over time in the ideological cast of European integration. Hooghe and Marks (1999) argue that the post-EMU era will witness a political struggle between rightist partisans of a “neoliberal market” and leftist supporters of European “regulated capitalism,” 51

52

Leonard Ray

implying that support for the EU will be associated with the political left. Implicit in their argument is the idea that support for the EU is derivative of individual preferences over regulation and the market. A similar argument about the role of anticipated policy outcomes and integration can be found in Gabel (1998b). This chapter applies that intuition to the problem of spatial, rather than temporal or class differences. I begin with the notion that preferences for policy outcomes are related to beliefs about the appropriate venue for policy-making. Given the diversity of prevailing conditions across Europe, the expected impact of EU-level decision-making in a given policy area will differ across nations. It is this difference in expectations that drives my account of the protean relationship between EU support and ideology. Only if European nations were homogeneous in their current political, economic and social conditions could the institutional dimension be either independent of, or consistently related to, the left/right ideological divide. This chapter employs a very simple spatial model of policy preferences in order to derive the disutility of EU-level policy for individuals. Members of a group are assumed to prefer EU-level policy-making when the disutility of the expected outcome of EU-level policy-making is smaller than the disutility of the prevailing national policy. The specific policy I examine is social protection, and the relevant social groups are defined by income quartiles. I assume, for the sake of argument, that the transfer of responsibility for social protection to the European Union is expected to lead to a convergence of social protection standards across the EU. The generosity of this anticipated common policy is not relevant to the argument, although the data presented below do suggest that Europeans fear that a lowest common denominator policy would be the result. I further argue that support for an EU-level social policy is not divorced from support for further integration in general. If European integration is expected to lead to a convergence of policy outcomes, then support for the EU will come from the right side of the ideological spectrum in those nations with a “leftist” national policy regime. Conversely, nations with a more “rightist” set of national policies will see support for integration focused on the left side of the ideological spectrum. Data and results Data to test these hypotheses were drawn from the Eurobarometer 44.2 bis “mega” survey conducted in the spring of 1996. In addition to the usual demographic and trend questions about support for European integration, this survey contains questions about the policy priorities of the European Union, about respondents’ fears and expectations for the

Don’t rock the boat

53

future, and about the appropriate level for decision-making in several policy areas. I assume that Europeans expect that European integration will lead to a convergence of policy outcomes across the EU. Inferences about these expectations can be drawn from cross-national differences in responses to a question which asks respondents whether they expect a loss of social benefits as a result of European integration.1 If respondents expect a convergence of social benefits and wages to a common level, then the expectation of a loss of benefits or jobs should be a function of the prevailing national level of social benefits. As figure 3.1 indicates, this is the case. The proportion of respondents who think that a loss of benefits is likely increases with the percentage of GDP devoted to social protection.2 A test of this relationship at the individual, rather than the national, level allows the incorporation of income as a control variable. The likelihood that a respondent fears the loss of social benefits should be a function of both the current national level of benefits (which influences the probability that a loss is expected) and the respondent’s income which is related to the likely dependence on these benefits. Indeed, the sensitivity to the current level of social protection should be greatest for those most likely to rely on such aid. Fears of the loss of social protection are measured using one of a battery of Eurobarometer questions probing respondents’ fears associated with European integration.3 The fear of the loss of benefits is modeled with the following regression equation: P (fear of loss of benefits) = a + b 1 social protection + b 2 income + b 3 social protection ∗ income Because the dependent variable is dichotomous (a respondent either fears the outcome or does not), these equations are estimated using logistic regression. The results, presented in table 3.1, conform to theoretical expectations. Respondents in nations with higher social protection spending are more likely to fear the loss of benefits, but this effect weakens for higher income groups. The substantive significance of logistic regression 1

2 3

The expectations questions were worded as follows. “Regarding the building of Europe, the European Union, some people may have fears. Here is a list of things which some people say they are afraid of. For each one, please tell me if you – personally – think that is likely to happen, or not?” A table presenting the cross-national data used in this chapter is included as an appendix. This suite of questions (Q 24a) is worded as follows “Regarding the building of Europe, the European Union, some people may have fears. Here is a list of things which some people say they are afraid of. For each one, please tell me if you – personally – are currently afraid of it, or not?”

% Expecting loss of social benefits

0.00

10.00

20.00

30.00

40.00

50.00

60.00

70.00

80.00

POR

20

IRL

25

ITA

UK

30

Social protection spending as % of GDP

SPA

LUX

BEL DK

35

SWE

y = 1.5356x + 11.814 R 2 = 0.505

FIN

Figure 3.1 Expectation of loss of social benefits as a function of current social protection.

15

GRE

AUS

GER FRA

NET

40


55

Table 3.1 Fear of a loss of social benefits as a function of the national status quo Likelihood that a respondent fears the loss of social benefits. (logit results) Social protection as percentage of GDP 0.075∗∗ Income 0.223∗∗ Social protection as percentage of GDP ∗ income −0.012∗∗ Constant −1.517∗∗ Percentage predicted correctly 58.1 N 48,408 Percentage of respondents fearing outcome 56.4 Note: ∗∗ p < 0.001, ∗ p < 0.01.

results is notoriously difficult to discern from the raw estimates, a problem compounded by the use of interaction terms. Figure 3.2 presents these results in a more user-friendly graphical format. This expectation of a convergence of social protection under continued European integration implies that an individual’s preferences for policymaking at the national or European level will be a function of their satisfaction with current national social protection policy. The better the prevailing national policy, the more the respondent will prefer that decision-making over that policy remain strictly at the national level. The question used for this analysis is drawn from a battery which asked respondents whether certain policies should be decided at the national level, or jointly within the European Union.4 Support for national-level decision-making about health and social policy is modeled as a positive function of a nation’s level of social protection spending. This cross-national effect is interacted with income levels and, as argued above, should diminish as income increases. The regression equation is similar to the one used above, and the estimation is again a logistic regression procedure. For this analysis, we add controls for two possible confounding ideological variables. Controls for a respondent’s position on the pro-/anti-integration and left/right dimensions were added to ensure that the results from the policy expectations model are not spurious effects of these ideological influences.5 The results are reported in table 3.2. 4

5

This suite of questions (Q 26) was worded as follows: “Some people believe that certain areas of policy should be decided by the (national ) government, while other areas of policy should be decided jointly within the European Union. Which of the following areas of policy do you think should be decided by the (national ) government, and which should be decided jointly within the European Union?” The pro-/anti-integration dimension is measured with an additive index of responses to the question “In general, are you for or against efforts being made to unify Europe?”

56

Leonard Ray

Probability that respondent fears loss of social benefits

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 10

15

20

25

30

35

40

45

Social protection spending as % of GDP lowest quartile 3rd quartile 2nd quartile highest quartile

Figure 3.2 Fear of loss of benefits by income level and social protection status quo.

Preferences about the appropriate governing level for social policy are strongly related to the current levels of social protection. The substantive effects are quite substantial, with a shift from the minimum level to the maximum level of social protection spending increasing the probability that a respondent prefers national-level policy-making by 0.28. The two ideological dimensions, pro-/anti-EU orientations and left/ right ideology do have consistently significant effects on preferences over the appropriate level for policy-making. Not surprisingly, pro-European orientations are associated with lower levels of support for national policymaking. This effect is moderately strong, decreasing the likelihood of and the question “Generally speaking, do you think that [our country’s] membership of the European Union is . . . a good thing, a bad thing, neither good nor bad?” Left/right ideology is measured with the following question “In political matters people talk of ‘the left’ and ‘the right.’ How would you place your views on this scale?”


57

Table 3.2 Probability that respondent prefers policy made at the national level only (logit results) Health and social welfare

β

Social protection as percentage of GDP Social protection as percentage of GDP ∗ income quartile Income quartile Pro-EU orientation Left (1) to right (10) ideology Constant Percentage predicted correctly N Percentage of respondents preferring national policy

0.041∗∗ 0.005∗

0.25 0.23

−0.169∗ −0.226∗∗ 0.051∗∗ −0.353∗∗ 62.2 38,823 58

−0.02 −0.18 0.10

Change in probability

Note: ∗∗ p < 0.001, ∗ p < 0.01.

support for national policy by 0.18. Left/right ideology has a weaker relationship, with respondents on the right being more supportive of national policy-making. If support for EU-level policy-making is related to support for the EU as a whole or for European integration as a project, then the model sketched out above can be adapted to explain how and why the relationship between left/right ideology and the pro-/anti-EU dimension varies across national contexts. In those nations where the prevailing national policy regime is closer to the ideal policy preferences of “leftist” individuals, support for integration should be concentrated on the right side of the political spectrum. Conversely, in nations where national policy is relatively far from the preferences of “leftists,” the left will be more supportive of integration as a way to achieve, at the European level, outcomes unobtainable under a purely national regime. In order to test this hypothesis, support for European integration6 was regressed on the Eurobarometer measure of left/right ideology, and on an interaction term where left/right placement was multiplied by the “leftism” of current national policy. For this analysis, spending on social protection was used as a proxy for policy “leftism.”7 The resulting OLS regression equation has the following form: 6 7

Measured by the additive index described earlier. Some may argue that the Christian democratic center-right is not averse to high social protection spending. A second test, more in keeping with Esping-Andersen’s analysis, would involve interacting ideology with government employment.

58

Leonard Ray

Table 3.3 The contingent effect of ideology on support for the European Union Predicting orientations towards European integration Left (1) to right (10) ideology Social protection as percentage of GDP Ideology ∗ social protection spending Constant Adjusted R2 N

β −0.127∗∗ −0.092∗∗ 0.005∗∗ 3.555∗∗ 0.035 49,569

Note: ∗∗ p < 0.01, ∗ p < 0.01.

EU support = a + b 1 ideology + b 2 ideology ∗ social protection + b 3 Social protection results for the estimation of this equation are presented in table 3.3. As these results indicate, support for European integration would be associated with the ideological left in a nation with no spending on social protection. As social protection spending increases, support for integration declines overall, while the relationship between leftism and support for integration weakens. Eventually, when social protection spending constitutes 25 percent of GDP, the slope reverses sign, and the right becomes more supportive of integration. These results are presented graphically in figure 3.3. Implications The policy expectations approach helps to explain the protean relationship between left/right ideology and support for European integration. The model does suggest that those groups who benefit most from the national status quo are reluctant to cede power to the supranational level. The conclusion that the “winners” in European societies are Euroskeptics may appear to fly in the face of some conventional wisdom about public support for the European Union. However, this argument elucidates a source of friction slowing the process of European integration. Punctuated deadlock Those groups which enjoy or have been able to secure favorable policies at the national level may tend to occupy privileged positions within the national polity. If this is the case, then they should be able to exercise disproportionate influence over their nation’s policies towards European

Attitude towards European integration

0

0.5

1

1.5

2

2.5

1

3

4

6

left /right self-placement

5

7

8

9

Figure 3.3 The contingent relationship between ideology and support for the EU.

2

10

maximum social spending

mean social spending

minimum social spending

60

Leonard Ray

integration. These groups will also tend to oppose the expansion of EUlevel policy-making. This produces a natural drag on the process of integration, and a brake on the process of spillover. This may not, however, mean that EU policy tends towards lowest common denominator outcomes. If policy-making at the EU level must overcome the opposition of entrenched groups which are protecting favorable national-level policies, then EU policy, while rare, may tend towards highest common denominator outcomes in order to placate these strategically located veto actors. The result would be periods of apparent stagnation with minimal policy output, punctuated by the sudden adoption of apparently ambitious common policies. Conclusion: the location of a common European political space Policy expectations do appear to inform opinions as to the appropriate level of decision-making in Europe, and evaluations of the EU as a whole. At the mass level, questions of desired policy outcomes are confounded (both in the debates of politicians and in the survey instruments used by social scientists) with the question of the appropriate level for decisionmaking. If preferences over concrete policy outcomes are similarly structured along broadly left/right lines across Europe, then one would expect to see evidence of a coherent European political space and common discourse only in those situations where the question of the appropriate level of policy-making is moot. For example, policy-making within the institutions of the European Union is presumably a matter of concrete policy outcomes, rather than the appropriateness of EU involvement in the policy area under discussion. Thus we should not be surprised that within the European Parliament the left/right dimension is a powerful predictor of MEP voting behavior (Thomassen, Noury, and Voeten, chapter 7 this volume). This research helps illuminate why, at the mass level, the left/right ideological dimension may appear “orthogonal” to support for decisionmaking at the European level, while within European institutions (such as the European Parliament) a left/right ideological dimension captures much more of the scope of political conflict. If debate within European institutions centers around the concrete policies to be adopted by the EU, then the classic left/right division should structure the preferences of political elites and party groups. However, at the mass level, much public debate about the European Union still swirls around the institutional questions of how much power supranational bodies should exercise, and


61

over what policy areas. So long as this is the case, similar preferences over policy outcomes will lead to disagreement on the institutional plane. Thus when individual Europeans are queried about the European Union, the common European political space is fractured by national differences in policy expectations, and fifteen distinct political spaces emerge. Appendix: cross-national data and sources Social protection spending as percentage of GDP Austria Belgium Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Sweden United Kingdom

30.2 27.1 33.7 34.8 30.5 30.8 16.0 21.1 25.3 24.9 32.3 19.5 23.6 37.7 28.1

Standard deviation Mean

5.83 27.71

Source: Eurostat (1997). Data are from 1994.

4

Varieties of capitalism and political divides over European integration Adam P. Brinegar, Seth K. Jolly, and Herbert Kitschelt

Under its umbrella, the European Union covers countries with highly diverse configurations of capitalist political-economic institutions. In the macro-level political economy literature these differences have led scholars to generate a number of hypotheses about the relative gains or losses of individual member countries from important institutional innovations that advance integration, such as the formation of the European Central Bank and a common currency (cf. Hall and Franzese 1998; Iversen 1998). Moreover, individual citizens and labor market participants may perceive costs and benefits differently, contingent upon national wagebargaining systems or welfare state policies. Domestic political divides between advocates and opponents of EU integration may play out differently and yield contrasting partisan alignments if polities are embedded in different institutional “varieties” of capitalism. In this chapter, we explore how the diversity of capitalist institutions affects political contestation over EU integration in two respects. First, capitalist institutions affect the proportion of voters in each country who have an incentive to challenge EU integration. In other words, political economy shapes the “grievance level” that may provide the raw material of patterns of domestic contestation. Contingent upon existing national economic institutions, citizens calculate how their benefits (in terms of jobs, income growth, etc.) are likely to be affected collectively for most voters (“sociotropic” calculations). Second, they also may focus on their potential individual benefits and costs that result from changes in the expected economic payoffs induced by the consequences of European integration for national political-economic institutions. Here citizens’ general political ideology and their individual asset endowments in labor markets may produce domestic alignments of conflict over European integration (“egocentric” voting). How ideology and assets affect such alignments may be contingent upon domestic political-economic institutions. Whether citizens are leaning toward or away from further European integration does not simply depend on whether they are “left” or “right,” but whether they are left or right within a particular national 62

Varieties of capitalism and political divides

63

political-economic context. Consequently, political alignments among parties over the EU issue may vary across member countries. In order to probe into the empirical plausibility of this line of reasoning, we report here first how domestic varieties of capitalism affect central national public opinion trends toward European integration. Against this baseline, we then estimate the relative effect of individual citizens’ ideological orientations on cross-nationally varying domestic alignments about EU integration and the interaction between individual and contextual varieties of capitalism conditions. In our unreported research we also controlled for citizens’ human capital endowments (skills, professions) and interactions between capital endowments and national varieties of capitalism. These latter controls do not affect the empirical robustness of our main propositions. Because such human capital factors have been explored in the past (Gabel 1998a and 1998b; Hix 1999b: chapters 5 and 6) and elsewhere in this volume by Leonard Ray’s analytically elegant chapter, we ignore such additional factors in our chapter.1 Our chapter is not concerned with partisan vote choice in elections. Yet it leads us to the following hypothesis: whether and how parties can politicize EU integration depends very much on their national politicaleconomic context. Such contexts affect not only the magnitude of perceived grievances, but also their ideological embeddedness into domestic partisan alignments. The explanatory value of interacting contextual and individual-level variables to account for alignments over European integration highlights the plausibility of a multilevel model of the European polity. On the one hand, national differences of interests are captured by the direct contextual effect on national varieties of capitalism. On the other, the interaction between contextual conditions and individuals’ political ideologies demonstrates the relevance of domestic divides over EU integration inserted into the broader national competitive alignments.2 Varieties of capitalism and EU integration: theoretical hypotheses Let us distinguish three interrelated, but empirically non-identical dimensions of capitalist political-economic arrangements. First, there are macroeconomic features of wage-bargaining regimes and monetary policy governance (e.g., Scharpf 1991; Iversen 1999; Soskice and Iversen 1 2

For a further exploration of the interaction between human capital and wage-bargaining systems in citizens’ evaluation of the European integration process, see Scheve (2000). For an outline of alternative polity models of the EU see Moravcsik (1998), chapters 1 and 2, and Hooghe and Marks (1999, 2001) as well as Peterson and Bomberg (1999).

64

Adam P. Brinegar, Seth K. Jolly, Herbert Kitschelt

1998; Franzese 2001, 2002). Second, there is the institutional microeconomics of different corporate governance structures and labor regimes elaborated by David Soskice (1999) and empirically fleshed out in Hall and Gingerich (2001). In ways relevant for our analysis, this approach has recently been applied to a new interpretation of the welfare state developed by Iversen and Wren (1998), Mares (2001), Estevez-Abe, Iversen, and Soskice (2001), and Iversen and Soskice (2001). Third, there are types of welfare states resulting from distributive social conflict and partisan competition as originally proposed by Esping-Andersen (1990) and further developed by Hicks (1999), Esping-Andersen (1999), and Huber and Stephens (2001). Based on these dimensions of advanced capitalist democracies we infer a total of five propositions about contextual and interactive relations predicting citizens’ predispositions toward EU integration. The mechanism linking institutions and citizens’ assessments of EU integration is the perception of costs and benefits accruing from integration in light of domestic capitalist institutions. Whether or not these perceptions accurately reflect the consequences of integration, however, is irrelevant for our paper. For example, the logic of popular sociotropic cost-benefit perceptions that appears to explain a substantial share of cross-national variance in support of further European integration presumes that this process would make social policies converge toward the patterns of the currently most widespread conservative welfare state institutions. But there are serious scholarly models with plausible scenarios that European integration will generalize liberal-residual welfare state policies (cf. Scharpf 2002). Conversely, others may object that European integration will not challenge microeconomic institutions of the welfare state, as long as countries abide by common macroeconomic monetary and fiscal regimes. For our purposes, it is irrelevant which of these positions turns out to be right. What matters is whether we can link citizens’ perceptions of the costs and benefits of European integration to actual patterns of welfare state institutions. Wage-bargaining regimes and central bank autonomy Centralized wage-bargaining and the autonomy of central banks in setting monetary policy provide mechanisms that cumulatively contribute to macroeconomic stability (low inflation and unemployment, high growth). Centralized wage-bargaining at the sectoral or the national level enables unions to achieve wage moderation and employers not to offer wage drift in a situation of tight labor markets. Centralization produces the collective good of lower inflation because employers and unions can make credible


65

commitments to enforce moderate agreements against opponents in their own ranks. Autonomous central banks charged with maintaining the stability of the currency tend to punish undisciplined wage settlements by ratcheting up real interest rates, thus triggering lower investment and higher unemployment. Sectoral or national wage-bargaining centralization may anticipate and avert this negative outcome.3 In the absence of an independent central bank, centralized wage-bargaining may lead to wage moderation only if leftist governments with tight links to labor unions convince the latter that such policies provide a leftist government with a sufficiently favorable economic performance record to boost its chances of being reelected.4 Wage-bargaining institutions and central banks may affect citizens’ sociotropic calculations of the costs and benefits of European integration based on two premises, the first of which we also employ in subsequent propositions about the impact of national capitalist institutions on public opinion. This first general premise states that the median national voter is concerned with the collective good of the country (“sociotropic voting”) and controls domestic policy outcomes. Indeed, there is evidence to believe that policy outcomes are close to the ideal points of median voters, especially in multi-party systems (Powell 2000). Our second specific premise is that citizens know that European integration institutes an independent central bank regime capable of punishing inflationary wage policies. Under these conditions, the median voters in the countries with sectorally and nationally coordinated wage-bargaining systems may be more opposed to EU integration than those situated in fragmented bargaining systems (proposition 1). If countries with coordinated wage-bargaining commit to moderate wage policies, business and labor in fragmented bargaining systems may free-ride and boost wages and prices. This triggers inflation and a redistribution of assets toward such countries. If the European Central Bank punishes such moves with high real interest rates, the economic consequences will hurt everyone, including citizens in countries that have shown wage moderation. Therefore citizens in such systems will be more opposed to European economic integration than in liberal decentralized systems (proposition 1). 3

4

In line with Soskice and Iversen (2000), but not Iversen (1998; 1999), we presume that both sectorally and nationally coordinated wage-bargaining regimes benefit from independent central banks. Empirically, by the late 1990s, Western Europe no longer had nationally centralized wage-bargaining systems. In a strictly game-theoretic formulation, wage moderation is not a Nash equilibrium of the interaction between labor unions and leftist governments, as Scharpf (1991) explains. It requires political goodwill on the part of the labor unions.

66


The microeconomics of capitalist institutions, skill formation, and welfare states The distinction between uncoordinated, sectorally coordinated, and nationally coordinated market economies also has a microeconomic formulation. Coordination encompasses (1) the nature of labor contracts (short- or long-term), (2) the nature of capital markets (equity- or debtbased), (3) techniques of skill formation (general or specific skills), and (4) patterns of technological innovation (venture capitalist firm-centric or collaboration across networks of firms) (Soskice 1999). Hall and Gingerich (2001) collected empirical indicators for all of these dimensions and determined through factor analysis that national patterns of shareholder power in publicly traded companies, the dispersion of corporate control beyond the chief executive officer, the size of the stock market, the level and degree of wage coordination, and labor turnover indeed all relate to a single underlying unobserved variable we call here the continuum from “stakeholder” to “shareholder” capitalism. Each system may have its own advantages, but wage equality and the lot of the less affluent have clearly been better under stakeholder capitalism. In the perception of mass publics, European integration may encourage the erosion of stakeholder capitalism and of its distributive benefits. Hence European integration should be opposed more intensely in countries that have enjoyed the distributive benefits of stakeholder capitalism (proposition 2). Estevez-Abe et al. (2001) and Iversen and Soskice (2001) also link varieties of capitalist institutions to welfare state regimes. Coordinated economies institute the possibility of generalized reciprocity and longterm relational bargaining among firms and between employers and wageearners. If market actors agree to underspecified contracts, but do not opportunistically exploit the contractual vagueness of their obligations, all actors will benefit by lowering their transaction costs and encouraging business and wage-earners to make otherwise risky investments. For example, “patient” capital provided by banks allows industrial firms to invest without watching short-term profit performance. And long-term labor contracts enable wage-earners to commit to investments in assetspecific skills that bear a return only if applied in a particular sector or a particular firm. The existence of comprehensive welfare states that lower the risk of unemployment and/or increase the compensation for job loss facilitates asset-specific human capital investments. If European integration is perceived as a process that advances market liberalization and undermines welfare state arrangements encouraging asset-specific human capital investments, then mass publics, particularly in those countries


67

that have instituted such systems, should be averse to further European integration (proposition 3). The welfare state as a distributive settlement Esping-Andersen (1990; 1999) proposes to distinguish three types of welfare states. First, there are residual welfare states in countries dominated by market-liberal parties. Such welfare states provide means-tested benefits and income-related social insurance entitlements. Their redistributive capacity is quite limited and they provide strong incentives for the (re)commodification of male and female labor. Second, nonliberal, Christian democratic and other “centrist” parties have developed conservative-Christian welfare states based on comprehensive coverage and stratified income-related and family-based cash benefits with moderate to strong redistributive impact. They set disincentives for women and older workers to stay in or enter labor markets. Third, comprehensive social democratic welfare states in countries with long-term moderate-left party rule in the formative and expansionary development of the welfare state involve comprehensive coverage, flat-rate entitlements in cash and services, sometimes supplemented by income-stratified benefits, with an overall strong redistributive capacity favoring low-income citizens, but intensive incentives for all citizens, including women, to participate in labor markets. The upshot of the distributive welfare state argument for citizens’ evaluations of European integration today may develop along one of the following two lines. In each variant, domestic median voters of a country compare their national social policy status quo to the most common national social policy practices within the entire set of European Union member countries, especially the most powerful among them, such as France and Germany. That domestic median voter then calculates her costs and benefits of adopting the predominant European welfare state pattern domestically according to one of two rationales. In the first variant of the argument, median voters always prefer the status quo in their own country and view change from the status quo as a net “cost” imposed on them to move away from the national “ideal point” of social policy. Given that conservative-Christian patterns of the welfare state prevail among the EU members with the greatest seniority in the association and in countries that have often served as agenda-setters for the European integration process, the domestic median voter in countries with conservative welfare states might be quite happy with the European integration process in the belief that such integration will unleash few tendencies to dislodge the established domestic welfare state

68


policies. Thus, high satisfaction with further EU integration should prevail in Austria, Belgium, France, Germany, Italy, and the Netherlands. By contrast, median voters in both liberal-residual (i.e., Britain) and social democratic-comprehensive welfare states (i.e., Denmark, Finland, Sweden) may harbor distinctly fewer pro-European integration attitudes because they may anticipate higher national adjustment costs if forced to conform with European social policy norms (proposition 4). According to an alternative logic of perception, median voters always prefer more redistribution and thus the Scandinavian social democratic model of welfare state. After all, the median voter earns less than the average voter and therefore may always harbor some appetite for additional redistribution. The question, then, is in which EU countries domestic median voters have reason to believe that further EU integration will advance their redistributive aspirations most. Citizens in countries with already comprehensive, redistributive welfare states may not see much possibility for further redistribution through EU integration and therefore will be more predisposed to opposing it (Denmark, Finland, Sweden). But the median voter in a country with a residual welfare state will most keenly favor EU integration in order to enhance the redistributive impact of social policy. Conservative, Christian democratic welfare states exhibit populations with intermediate enthusiasm for further EU integration (proposition 5). How to deal with economic “laggards” in the varieties of capitalism framework So far, our discussion of varieties of capitalism has conspicuously left out any mention of the relatively poor, laggard economies of Greece and Portugal, as well as the less starkly, but still significantly, trailing economies of Spain and Ireland.5 Except for Ireland, the same omission characterizes almost the entire literature on the comparative political economy of advanced capitalist democracies.6 Macro-comparative data available on these countries are thin, particularly for variables that directly pertain to the comparison of varieties of capitalism. An even more serious problem is that data for these countries require a different interpretation than for other countries – something that could be taken care of only with elaborate quantitative correction procedures. For example, 5

6

We have left out Luxembourg, simply because a country/city-state of less than 0.5 million citizens that is also the seat of major European Union institutions constitutes the most clear-cut violation of the assumption of unit homogeneity and independence in comparative research. Exceptions are Boix (1998) and Hall and Gingerich (2001).


69

lower social expenditures may be a consequence simply of less economic development, not of adoption of a residual welfare state pattern. Indeed, the microeconomic corporate governance structures of the southern European countries show little adherence to the market-liberal model (see Hall and Gingerich 2001). Given the lack of rigorous data, we propose the following scoring conventions. Case studies suggest that the less-developed EU members have neither strong wage-bargaining centralization nor high investment in specific human capital assets. We therefore score them low on that dimension. For advanced countries, we adopt Iversen’s (1999) operationalization of liberal uncoordinated and sectorally coordinated systems as a dummy variable. On corporate governance, we adopt Hall and Gingerich’s (2001) index of stakeholder capitalism, showing moderately high values in Portugal and Spain and lower values in Ireland. When it comes to the welfare state, the median voters inside developmental laggards may be at least as interested in EU integration as those of Christian democratic, conservative welfare states. None of the laggard countries has strong market liberal parties that would push them toward a residual welfare state trajectory. Moreover, their social policy schemes and expenditure levels, relative to their achieved affluence, appear to set them on tracks leading to coverage levels and schemes at least as encompassing and redistributive as those of Christian democratic, conservative welfare states. At the same time, none of these countries has strong, encompassing social democratic parties and union movements that could press for universalistic redistributive welfare states. For these reasons, we postulate that median voters in these countries anticipate a domestic propensity to promote conservative welfare states. They therefore expect European unification not to affect the domestic logic of social policy development significantly (proposition 4) or to exercise only moderate pressure to cut back on welfare states (proposition 5). One further important consideration should be added, however, to illuminate the distinctiveness of the laggard countries. The EU budget supplies them with economic aid through so-called structural funds to a vastly greater extent (absolutely and as a percentage of GDP) than the remaining countries. In order to capture the specific attractiveness of European integration for economic laggards, we employed the disbursement of structural funds per capita, in thousands of dollars in 1996, as an independent variable from European statistics (EUROSTAT data). Sweden is at the low end, Ireland at the top end.7 Because national receipts of EU 7

This variable is taken from Hix (1999b), table 9.5, and is also employed in Ray (this volume).

0 [0] [0] [0]

II. Economic laggards Ireland Greece Portugal Spain 0 [0] [0] [0]

1 1 1 1 1 1 1 1 1 0

Sources: (1) Iversen (1999), pp. 84–5, own coding in brackets. (2) Estevez-Abe et al. (2001), p. 179, own coding in brackets. (3) Hall and Gingerich (2001), p. 45, own coding in brackets. (4) own coding from Esping-Andersen (1990; 1999) see text. (5) own coding, see text.

1 1 1 1 0 1 0 1 1 0

I. Advanced EU members Austria Belgium Denmark Finland France Germany Italy Netherlands Sweden United Kingdom

Centralization of wage bargaining (1)

Labor market regime: bias towards specific skills (2)

0.29 [0.50] 0.72 0.58

1.0 0.74 0.70 0.72 0.69 0.95 0.87 0.66 0.69 0.07

Overall index of corporate coordination: “stakeholder” capitalism (3)

Table 4.1 Five specifications of varieties of capitalism

[1] [1] [1] 1

1 1 0 0 1 1 1 1 0 0

Dominance of the conservative Christian democratic welfare state (4)

[2] [2] [2] 2

2 2 1 1 2 2 2 2 1 3

Limits to national redistribution: from encompassing to residual welfare state (5)


71

structural funds are negatively related to the wage centralization variable (r = −0.64), it is quite possible that only one of the two variables has a statistically significant effect on endorsement of European integration. Table 4.1 lists values for the five dimensions of varieties of capitalism chosen to explore our propositions about popular support for EU integration. Figures in brackets indicate our own judgmental assignment of values to economic laggard countries based on plausibility considerations presented above. Interestingly, while all measure varieties of capitalism, in most instances the five indicators are only weakly linked to each other, regardless of whether we examine the eleven “core” countries on which firm data are available for the first three columns or add the Mediterranean economic laggard countries as well. Only two bivariate correlations exceed 0.60, namely those for wage-bargaining in column 1 and propensity toward specific skill formation in column 2 (r = 0.62 for N = 11 and 0.75 for N = 14) and those for specific skill formation (column 2) and stakeholder capitalism (r = 0.89 for N = 11 and 0.66 for N = 14). What is important for the empirical testing of separate propositions, however, is the weak correlation between wage-bargaining centralization and the conservative Christian democratic welfare state (r = −0.18, N = 11 and −0.32 for N = 14). Also, features of labor market organization and corporate governance correlate only moderately with the national welfare state redistribution variable in column 5. Political ideology and European integration As a first cut, the varieties of capitalism literature help to generate hypotheses only about the central tendency in the disposition of national electorates toward further EU integration. But national settings may interact with individual political preferences within each country. Before we turn to these interactions, let us first consider the potential direct, linear effect of two major dimensions of political ideology in an advanced capitalist democracy on preferences over European integration, orientations toward economic (re)distribution, and libertarian or authoritarian socio-cultural regulation. Both of these orientations affect a generalized measure of ideology, citizens’ left/right self-placement, even though these patterns vary cross-nationally (Knutsen 1995). Individual-level ideologies Dispositions toward economic redistribution Are citizens who would like to restrain politically authorized income redistribution through

72


social policy disposed more toward EU integration or national autonomy? We find that no a priori response to this question is possible. Liberals may see EU integration as a way to create a common market free of government interference. But they may also perceive the EU as a regional club that interferes with markets at its outer boundaries and undercuts global trade openness. Moreover, they may fear additional layers of social policy and business regulation beyond levels reached by national policy intervention. Disposition toward libertarian or authoritarian socio-political arrangements Libertarians value individual autonomy to determine lifestyles and values as long as it is compatible with other citizens’ corresponding liberties. A most salient issue for libertarians in this regard is the full, equal acceptance of women’s participation and self-determination in all arenas of social life. Moreover, they call for participatory, democratic processes to determine the provision of collective goods. Authoritarians, by contrast, prioritize normative cultural conformity and political authority and demand that individuals subordinate themselves to the imperatives of collectively binding norms and especially a paternalist family order. Furthermore, whereas libertarians tend to embrace a universalistic conception of political community, authoritarians insist on boundaries between distinctive cultural and political communities defined by internal norms and authority patterns. Again, no a priori deductive expectation about the relationship of sociopolitical preferences to European integration is possible. Libertarians may favor it as a realization of multicultural, universalistic values. Or they may reject it as a process in which a faceless administrative technocracy dilutes democratic electoral participation (“democratic deficit”) and levels the cultural distinctiveness of national and subnational groups. In a similar vein, authoritarians may be pro-European if they expect that regional economic competition reasserts traditional values of thrift, diligence, and conformism with traditional family relations. They may oppose it as a negation of national communities and because it represents a permissive, relativist, multicultural social order. Left/right self-placements If left/right self-placements capture elements of economic and socio-cultural orientation, but such orientations affect attitudes over European integration in a theoretically indeterminate way, the same a priori indeterminacy applies to the relationship between citizens’ left/right self-placements and approval of EU integration. Without consideration of national anchor points and varieties of capitalism, little of interest at a systematic, theoretical level can be said. If European integration is salient, but left/right self-placements and other ideological


73

dispositions do not explain citizens’ evaluation of it, we may suspect that Europe constitutes a cross-cutting issue dimension. Ideologies and national varieties of capitalism: contextual interaction effects The varieties of capitalism literature generate few hypotheses pertaining to the more socio-cultural aspects of citizens’ ideological dispositions. Let us therefore focus on the economic dimension of ideology and move the rationalist collective cost-benefit perspective of evaluating European integration to an individualist egocentric perspective. Not a median voter, but each individual with her own economic interests may assess the prospects of European integration in light of expectations over the likely policy outcomes of such a process. As also argued in Ray (this volume), voters compare their personal economic ideal points with the national status quo and the expected European status quo resulting from further integration. In residual, liberal welfare states, leftists who would like to see more economic redistribution would obviously view European integration as a benefit, if it moves national conditions from the status quo to at least a conservative, but more encompassing and redistributive welfare state. Rightists, in that setting, will be opposed. Conversely, in encompassing, egalitarian, universalistic, social democratic welfare states, leftists who are fond of the national status quo can only fear that European integration will lead away from their personal ideal point. In such countries, the left is likely to be more anti-European and the right more pro-European. Especially women and elderly people who benefit from the encompassing welfare states should oppose European integration. In countries with conservative encompassing welfare states, EU integration should not be significantly related to left/right ideology, but cross-cutting. The varieties of capitalism approach is analytically powerful, particularly if direct individual-level effects of economic ideology or socio-cultural beliefs on the evaluation of European Union integration weaken or vanish once we interact components of personal ideology with conditions of the welfare state. Model specifications: dependent and further independent variables Dependent variables Many of the models of public EU approval employ rather primitive survey instruments to measure citizens’ dispositions toward European

74


integration. The bluntest and most commonly used indicator is a threetier response to the question whether the European Union is good, bad, or neither good nor bad. Others draw on a four-point scale on which respondents indicate how much they are in favor of (or against) efforts to unify Europe. Ray (in this volume, n. 5) employs an index that combines both these questions, thereby generating a somewhat more refined dependent variable with a better distribution of response values than each of the indicators taken separately. We call it the Current Evaluation of European Integration (CEEI) and employ it here as well. In addition, some have used a question asking respondents to assess whether European integration benefits their country. In our research, we have analyzed all of these dependent variables with similar results. We therefore report here only results for the CEEI. We actually began our research, however, by devising a more sophisticated and realistic index of support for further European integration. The practical question confronting public policy and citizens as spectators is the speed and scope of European integration, not all-or-nothing questions of EU membership and approval. We therefore focus on a seven-point scale in which respondents indicate the desired speed of European integration (slower – faster). We then reasoned that desired speed may be relative to respondents’ anchor points, the perceived actual speed of integration, also measured on a seven-point scale. Those who perceive an actually fast speed of integration but desire a slow speed are most opposed to further integration measures. Conversely, those who perceive a currently slow speed, but desire high speed, are most in favor of accelerating European integration. Survey response patterns show, however, that perceptions of actual and desired speed are all but uncorrelated. Moreover, the perceived actual speed of integration is also uncorrelated to our independent variables. We do employ here an index that interacts perceived actual and desired speed. The construction of this index of Overall European Integration View (OEIV) is explained in appendix 1. Accelerators of European integration score a maximum of twenty-one points, those who want to apply brakes to the process, a minimum of three points. We also ran our entire empirical analysis with the simple seven-point scale for respondents’ desired speed of European integration, but the results are basically identical to those of the more refined OEIV index that controls for perceptual anchor point. Table 4.2 presents the mean national values of our two dependent variables, OEIV and CEEI, arranged by groups of countries descending from high to low enthusiasm for European integration. We can see here that, at the aggregate level, the national values on the two measures of disposition toward EU integration correlate quite robustly (r = 0.73).


75

Table 4.2 Two measures of dispositions to European integration and their national mean scores Overall European integration view (OEIV) (range 3–21) (1)

Current evaluation of European integration (CEEI) (range 2–8) (2)

High enthusiasm for European integration Italy 16.67 (2.54) Portugal 16.06 (2.89) Greece 15.93 (3.20) Spain 15.54 (3.09) Ireland 14.38 (2.88)

6.67 (1.39) 6.22 (1.55) 6.30 (1.64) 6.20 (1.54) 6.78 (1.35)

Muted enthusiasm for European integration France 14.56 (3.59) Austria 14.28 (3.94) Netherlands 13.51 (3.37) Belgium 13.12 (4.07)

5.94 (1.72) 5.51 (1.87) 6.46 (1.51) 5.78 (1.73)

Feeble enthusiasm for European integration United Kingdom 12.77 (4.39) Sweden 12.20 (3.66) Germany 12.11 (3.94) Denmark 11.77 (3.80) Finland 11.71 (3.53) Cross-national average 13.78 (3.92)

5.40 (1.84) 5.48 (1.80) 5.53 (1.92) 5.72 (1.90) 5.35 (1.83) 5.93 (1.77)

Notes: Standard deviations in parentheses. Desired speed and OEIV: Eurobarometer 44.3 ovr. Composite index, based on adding scores for respondents’ evaluation of EU integration as “good or bad” and opting “for or against”: Eurobarometer 44.2 bis. OEIV: cross-national mean of standard deviation: 3.50 (s.d. = 0.51).

Interestingly, at the individual level, OEIV and CEEI are only moderately related to each other. The same is the case with other measures of public endorsement of European integration.8 This finding could indicate that (1) the various measures tap different underlying concepts of EU assessment or that (2) the measures tap the same underlying concepts, but in a noisy fashion with high measurement error. The second interpretation becomes more plausible if each of the dependent variables relates to the independent variables in similar ways. This is indeed what can be established.9 8 9

See appendix 2. The lack of a strong interrelation among independent variables may also suggest that the European integration issue simply was not terribly salient to voters at the time of the surveys, at least not in most countries. Hence they “make up” their opinions as the questionnaire proceeds.

76


Independent variables: individual level For individual-level variables, Eurobarometer provides few surveys measuring both attitudes toward economic redistribution and libertarianauthoritarian political-cultural views. We have therefore chosen Eurobarometer 44.3OVR for our analysis because it permits the construction of an index of preferences over the comprehensiveness of the welfare state, with high values indicating endorsement of a residual, liberal welfare state (for the items and index construction, see appendix 1). The survey also contains a battery of questions gauging respondents’ views on women’s inclusion and equality in economic, social, and political life (for items and construction of the gender equality index, see appendix 1), as an instrument to tap libertarian-authoritarian orientations. Unfortunately, Eurobarometer 44.3OVR includes OEIV but not CEEI. For that, we employ Eurobarometer 44.2OVR, which then forces us to drop the attitudes toward welfare state and gender equality on the independent variable side. To tap ideology in all models, we can only employ left/right self-placement as a summary term.10 We also control for age, which may or may not be related to socio-cultural conservatism and social policy (support of the welfare state). Finally, based on Ray’s (1999) expert judgment of parties’ positions on EU integration, we include a variable on the effect of “issue leadership” on individual respondents’ evaluation of the integration process. For each respondent, we score her favorite party’s elite EU position, as reported in Ray (1999) for 1996, on a seven-point scale, running from anti- to pro-European preference. The interaction between varieties of capitalism and individual ideological dispositions results in a straightforward fashion from the multiplication of individual-level ideological preferences with the contextual residual welfare state variable. This allows us to test whether the preference of ideologically left or right voters for European integration depends on political economic context. In comprehensive welfare states, the support of the left for European integration should be lower than that of the right because integration may bring a reduction of redistributive social policy effort. In residual welfare states, by contrast, the left may favor European integration as a way to upgrade domestic social policies toward more 10

Within each country and across the entire set of countries, the instruments for the economic substantive ideological dimension, rejection of an encompassing redistributive welfare state, correlates with more rightist self-placements (r = +0.16, N = 16,780); the instrument for the second dimension, support for gender equality in all spheres of social life, correlates with more leftist self-placements (r = −0.08), based on the 1–10 left/right scale. The correlations between the substantive ideological preferences and general ideology are slight enough to include all preferences in regressions.


77

redistribution, if the conservative comprehensive welfare state becomes the norm. The presence of these conditional effects would be indicated by a negative coefficient associated with the interaction term. An analogous logic applies to the interaction effects with age. In comprehensive welfare states, older people should be particularly skeptical about EU integration, but not in residual liberal welfare states where they stand to gain from European integration, if they expect such welfare states to be upgraded to a more redistributive European norm.

The aggregate-contextual model Table 4.3 reports simple bivariate correlations between attributes of capitalist institutions and the two measures of dispositions toward EU integration in national mass publics at the aggregate level. In order to explore the robustness of our initial estimates, we calculate the relations both for the advanced core EU members for which we have reliable data on the independent variables and for all members, including the economic laggards. The patterns remain the same. The simple bivariate correlations confirm only propositions 1 and 4 with correlations that explain at least 30 percent of the variance (shaded cells). Countries with more centralized wage-bargaining regimes have a distinct tendency toward stronger opposition to EU integration, presumably because they fear that economic integration undercuts such arrangements and produces higher unemployment when the autonomous central bank punishes lack of wage discipline in member states with fragmented industrial bargaining systems (proposition 1). Citizens from countries with conservative welfare states are on average considerably more favorable to EU integration than citizens from countries either with residual or with encompassing, highly redistributive social democratic welfare states, presumably because median voters in both cases fear a displacement of the national social policy status quo by EU integration (proposition 4). Before we move on and analyze EU approval and related national political alignments in greater detail, let us reflect on why the more specific microeconomic versions of the varieties of capitalism theory do not evidence clear relations to public opinion about EU integration. For average citizens, the practical salience and the intelligibility of redistributive welfare state policies are likely to be much greater than those of the microeconomic features of varieties of capitalism, such as corporate governance structures and wage-earners’ skill formation. Because issues of social policy and wage-bargaining are at the center of labor union concerns, powerful unions may serve as the intellectual transmission belts to

78


Table 4.3 Varieties of capitalism and national evaluations of European integration Overall European integration view (OEIV)

Current evaluation of European integration (CEEI)

Core EU members (N = 11)

All EU members (N = 14)

Core EU members (N = 11)

All EU members (N = 14)

1. Centralization of wage-bargaining

−0.64

−0.75

−0.47

−0.56

2. Bias toward specific skills

−0.07

−0.47

−0.20

−0.36

3. “Stakeholder” capitalism

+0.07

+0.00

−0.12

−0.15

4. Conservative welfare state

+0.66

+0.69

+0.59

+0.63

5. Limits to national redistribution

+0.42

+0.42

+0.19

+0.22

spread skepticism about the EU in countries where the unions’ political elites tend to perceive the EU as a threat to established achievements and practices. Mass publics and labor leaders may perceive the merits of EU integration quite differently than social democratic politicians, even though the latter are usually seen as close to the unions. Party strategists may pay more attention to the complicated and hard-to-communicate consequences of central bank autonomy and its benefits for an incumbent government in pursuit of wage moderation. If we were to examine the position of partisan elites rather than of mass publics on European integration in each country, as measured by Ray (1999) from 1982 to 1996, we would probably find much less consistency with proposition 1 and probably a broadly favorable reception of a European monetary regime that makes wage inflation costly to the industrial bargaining parties. While the average popular opinions on the merits of EU integration diverge across countries in line with wage-bargaining and welfare state regimes, those of political leaders may converge across most parties and countries because they have greater incentives to value the monetary regime and its consequences. Indeed, left party leaders, as advocates and representatives of working-class constituencies and labor unions, give up opposition to the European integration process precisely in those countries characterized by weak monetary policy autonomy combined with either liberal, uncoordinated wagebargaining (Britain) or in originally nationally coordinated, but later sectorally coordinated wage-bargaining systems (Austria, Denmark, Finland, Sweden). The case study literature confirms the importance of monetary policy


79

considerations, especially in the economic policies of the British Labour Party and the Swedish Social Democrats (cf. Aylott 2001: 154–66; Gamble and Kelly 2001: 62–7). The left party elites’ adoption of proEuropean integration positions is plausibly related to the imperatives of wage moderation in fragmented bargaining systems as well as nationally coordinated bargaining regimes with considerable wage drift, such as Sweden in the 1980s. To probe further into political alignments about European politics, let us develop multivariate statistical tests that determine the relative importance of national context, domestic divides, and the interaction between them. We begin with a simple model that employs only contextual varieties of capitalism effects as independent variables (table 4.4). This model comes in two versions, first as a pure country-level small-N specification that yields a very high proportion of explained variance for OEIV and CEEI. For purposes of comparison with the nested, hierarchical models, we also produce an individual-level version in which the contextual variables are regressed on individual citizens’ preferences over European integration. Because of much individual-level variance within countries, these models have less, though still substantial, explanatory power. Aggregate small-N models (N = 14) for both measures of propensity toward European integration (OEIV and CEEI) tell essentially the same story. Sectorally centralized wage-bargaining regimes as well as nonconservative welfare states of either the market-liberal or universalisticredistributive type reduce support for EU integration. With wagebargaining added as a key variable, structural fund disbursements do not have an independent effect on national patterns of EU support.11 The same applies to the presence of a residual welfare state hypothesized to give median voters an intense desire for further redistribution of powers and a hope that further European integration would accomplish this. The individual-level models of European integration support reported in the lower half of table 4.4 generally confirm the results of the aggregatelevel analysis. Because of the large number of observations, now also structural fund disbursal and limits of welfare state redistribution have significant, but weak, substantive effects. These effects are opposite to the hypothesized direction. Thus, in individual-level model II, the presence of a residual welfare state makes respondents less likely to express support for European integration than those in comprehensive, redistributive welfare states, in contrast to the sign of the bivariate correlations reported 11

Without wage-bargaining, higher structural fund receipts are a significant net contributor to a country’s level of approval for EU integration, but the overall explained variance of such equations is lower than in the more fully specified versions presented here.

80


Table 4.4 Contextual models with aggregate and individual-level data Overall European integration view (OEIV) (range 3–21) (1) S.E.

t-value


Dependent variables

Coeff.

Coeff.

S.E.

t-value

Aggregate model I (N = 14) Constant Conservative welfare state (1–2) Coordinated wage-bargaining system (0–1) Structural funds ($ 1000/cap; 0.16–1.70) Adj. R-square

13.5 0.70 19.3 1.8 0.57 3.2 −1.9 0.61 −3.1 0.08 0.58 0.13 0.73 ( p < 0.001)

5.60 0.28 19.9 0.45 0.23 2.0 −0.22 0.25 −0.90 0.25 0.23 1.1 0.46 ( p < 0.03)

Aggregate model II (N = 14) Constant Conservative welfare state Wage-bargaining system Structural funds Limited national redistribution (1–3) Adj. R-square

15.2 1.5 10.3 2.15 0.61 3.5 −2.46 0.74 −3.3 −0.24 0.61 −0.39 −0.79 0.61 −1.3 0.75 ( p < 0.002)

5.9 0.51 11.5 0.53 0.28 1.9 −0.33 0.31 −1.1 0.18 0.26 0.69 −0.12 0.21 −0.60 0.43 ( p < 0.06)

Individual-level model I (N = 16,946 (column 1) and 55,028 (column 2)) Constant 13.7 0.08 161.2 Conservative welfare state 1.61 0.07 23.8 Wage-bargaining system −2.14 0.07 −29.4 Structural funds 0.10 0.07 1.4 Adj. R-square 0.139 ( p < 0.000)

5.61 0.02 277.3 0.46 0.02 25.5 −0.30 0.02 −17.0 0.22 0.02 11.5 0.037 ( p < 0.000)

Individual-level model II (N = 16,946 (column 1) and 55,028 (column 2)) Constant 15.6 0.18 87.0 Conservative welfare state 2.1 0.08 26.8 Wage-bargaining system −2.8 0.09 −31.2 Structural funds −0.27 0.08 3.4 Limited national redistribution −0.94 0.08 −12.2 Adj. R-square 0.146 ( p < 0.000)

6.3 0.05 137.0 0.54 0.02 29.0 −0.50 0.02 −23.6 −0.11 0.02 5.4 −0.31 0.02 −16.8 0.042 ( p < 0.000)

in table 4.3. But the small substantive size of the coefficients here and in subsequent tables allows us to ignore this inconsistency.12 Comparing the performance of the statistical models for the two different dependent variables, they explain greater variance for our OEIV measure than for CEEI. OEIV is the more sophisticated and realistic 12

It appears that the wage-bargaining system “overpredicts” lower levels of EU integration support among sectorally coordinated bargaining systems. A removal of that variable generates regressions in which the structural fund variable and the limited national redistribution variables consistently display the correct signs.


81

measure of support for further EU integration. It may therefore be the more valid measure of actual popular dispositions toward the integration process. The “varieties of capitalism” framework in explaining popular dispositions toward EU integration is strengthened by the fact that its variables provide the best explanation for public opinion precisely when the latter is measured with the most refined available operationalization. Individual-level models Statistical models so far only account for cross-national differences, but not intra-country conflictual alignments about EU integration or the interaction between the two. The upper part of table 4.5 displays pure individual-level models of EU integration support, featuring respondents’ ideology (support of comprehensive welfare states, endorsement of women’s equal participation in all walks of life), age, and the issue leadership of each respondent’s favorite party as the independent variables. Opponents of encompassing welfare states and of gender equalization are more opposed to European integration, as are older people and those following leadership cues from parties opposed to the integration process. Net of these effects, there is a very slight tendency that respondents who place themselves further on the right are more pro-European, but this effect is substantively too small to deserve interpretation. The main effect goes through age, welfare state opinions, and gender views in the OEIV model and age with party leadership cues in the CEEI model. The robust, but small, negative age effect may capture in part the life cycle position of older people (unwillingness to change); in part ideological dispositions, for example negative affects toward cultural diversity; in part labor market position, e.g., that of older workers or of retirees fearing a reduction of social security benefits in liberalized European markets. While these models confirm the presence of significant intra-country disagreements on the merits of EU integration, the overall explained variance of a pure individual-level model and its substantive effects are very small. A left-wing ideologue supporting a comprehensive welfare state (value 7) and gender equality (value 18) and placing herself clearly on the left (value 2) is only 0.94 units more for EU integration on the twenty-one-point OEIV scale than a right-wing ideologue opposed to expansive welfare policies (value 11) and gender equality (value 6) with a clear rightist self-placement (value 9). The lower part of table 4.5 therefore adds the already familiar contextual variables in an additive fashion and reveals that both sets of factors work together. A comparison of coefficients in that part with those of the previous context-level specifications (last part of table 4.4) and the pure individual-level specification (first part of table 4.5) reveals that

82


Table 4.5 Ideology and contextual determinants of dispositions toward the EU integration process Overall European integration view (OEIV) (range 3–21) (1) Dependent variables

Coeff.

S.E.

t-value

Current evaluation of European integration (CEEI) (range 2–8) (2) Coeff.

S.E.

t-value

Ideology only, individual-level model (N = 16,870 (column 1) and 55,028 (column 2)) Constant 14.8 0.27 55.5 5.19 0.04 139.0 Prefers less welfare state (6–12) −0.14 0.01 −12.3 N/A Soc. + pol. gender equality (4–20) +0.07 0.01 6.9 N/A Age (18–80+) −0.034 0.002 −19.1 −0.007 0.0004 −15.5 Left-right ideology (1–10) 0.057 0.016 3.5 0.01 0.004 2.7 Party leadership cue (1–7) 0.10 0.02 4.6 0.19 0.005 36.3 Adj. R-square 0.036 ( p < 0.000) 0.026 ( p < 0.000) Ideology and context, individual-level model (N = 16,870 (column 1) and 55,028 (column 2)) Constant 14.76 0.32 46.8 5.58 0.06 99.7 Prefers less welfare state −0.03 0.01 −2.9 N/A Gender equality 0.07 0.01 6.7 N/A Age −0.03 0.002 −17.0 −0.005 0.0004 −12.7 LR – ideology 0.07 0.02 4.3 0.02 0.004 5.6 Party leadership cue (1–7) 0.17 0.02 8.7 0.21 0.01 39.2 Conservative welfare state 2.05 0.08 26.9 0.56 0.018 30.4 Structural funds −0.37 0.08 −4.8 −0.001 0.02 −0.07 Wage-bargaining system −2.69 0.09 −30.4 −0.62 0.02 −29.2 Limited national redistribution −0.86 0.08 −11.2 −0.36 0.02 −19.7 Adj. R-square 0.167 ( p < 0.000) 0.07 ( p < 0.000)

they invariably remain almost identical. The cross-national differences in dispositions toward EU integration accounted for by “varieties of capitalism” variables account for much more variance than the individual-level variables. Does this mean that there is definitely more contestation of the European integration issue between European countries than within them? We should not rule out domestic alignments over EU integration until we have considered the interaction effect between political-economic context and citizens’ individual ideological orientations. Nesting individual-level effects into a contextual analysis Interaction effects postulate that the personal ideology of a respondent affects views of EU integration contingent upon the prevailing variety of


83

capitalism in a country. If in comprehensive redistributive social democratic welfare states the left has the most to fear from European integration, because it could mean a rollback of social policy, it is the left in residual welfare states that should be most hopeful about European integration, because the latter can make things only better for social policy rather than worse. The personal ideology of respondents, as captured by left/right self-placements and the other indicators available, thus needs to be interacted with our “limited national redistribution” variable characterizing different welfare states (table 4.1, column 5). Before embarking on a complicated multivariate equation with interactive terms, we can demonstrate the interaction effect between context and respondents’ ideology in a simpler fashion. Table 4.6 gives for each country, grouped by their values on the “limited national redistribution” variable, the within-country correlation between left/right self-placement and views of EU integration. In comprehensive welfare states (scored 1 on context), we expect this correlation to be positive (rightists are more pro-European); in residual welfare states (scored 3 on context), it should be negative (leftists like EU integration). Welfare state status should thus predict the sign and the strength of the domestic correlations between political ideology and EU integration. And that is indeed the case in an almost perfect fashion. The correlations between national context and national left/right/EU correlations are −0.90 for the OEIV variable and −0.78 for CEEI. To test the efficacy of causal interaction effects on citizens’ views of European integration with greater precision, we now interact the ideology variables with the three-level welfare state variable (residual welfare state = 3) and keep direct contextual and individual-level effects as controls (table 4.7). The coefficients for our main contextual effects (conservative welfare state, wage-bargaining system) remain just about undisturbed, when compared to tables 4.4 and 4.5.13 But the individual-level determinants now either change through the interaction specifications, or stay about the same (party leadership cues), or shrink to marginality (views on gender equality, age). The interactive effects between ideology and political-economic context, however, are quite substantial in the predicted fashion. In residual welfare states, the left is more pro-European; in encompassing redistributive welfare states, the right is more pro-European. To employ once more 13

But note that the minor effect of the limited national redistribution variable (LNR), the context variable interacted with individuals’ ideological propensities, now switches signs and has again the “correct” interpretation that the median voter in residual welfare states is much more pro-European than the median voter in a social democratic welfare state.

84


Table 4.6 Left/right self-placement and endorsement of European integration: bivariate correlations by country Overall European integration view (OEIV) (range 3–21) (1)


Right prefers more European integration: egalitarian social democratic welfare states Sweden 0.14 Denmark 0.16 Finland 0.10

0.34 0.30 0.15

Weak left/right pattern of preferences over European integration I: conservative welfare states Austria −0.07 Belgium 0.01 France 0.01 Germany 0.006 Italy 0.05 Netherlands −0.001

−0.05 0.02 −0.015 −0.055 −0.07 −0.07

Weak left/right pattern over European integration II: economic laggards Greece 0.02 Ireland −0.02 Portugal −0.04 Spain 0.07

0.22 0.046 0.02 −0.07

Left prefers more European integration: residual welfare state United Kingdom Correlation of welfare state context (welfare state redistribution) with the constraint between left/right self-placement and support of EU integration*)

−0.15

−0.10

−0.90

−0.78

Note: * social democratic welfare state = 1; conservative welfare state or economic laggard = 2; residual welfare state = 3.

the example of ideological left-wingers and right-wingers with the same variable values specified above, a rightist in Denmark, Finland, or Sweden is on average 1.61 points more for EU integration on the twenty-one-point OEIV scale than a left-winger, but 1.67 points less for EU integration in a residual welfare state such as Britain! This effect dwarfs the net effects of ideology in the simple additive and the pure individual-level ideology models of support for EU integration with maximal effects of less than one unit on the twenty-one-point OEIV scale. Of course, the direct contextual effects of national varieties of capitalism, also in this model specification, remain larger than the interactive


85

Table 4.7 Direct effects and interactions between ideology and context as determinants of European integration views Overall European integration view (OEIV) (range 3–21) (1) N = 16,870 S.E.

Current evaluation of European integration (CEEI) (range 2–8) (2) N = 55,028

Dependent variables

Coeff.

t-value

Coeff.

Constant Prefers less welfare state Gender equality Age LR – ideology Party leadership cue (1–7) Conservative welfare state Structural funds Wage-bargaining system Limited national redistribution (LNR) Prefers less W.S. * LNR Gender equality * LNR Age * LNR LR ideology * LNR Adj. R-square

11.51 0.87 13.3 0.13 0.04 3.3 0.013 0.04 0.3 −0.016 0.006 −2.6 0.41 0.06 6.5 0.14 0.02 7.1 2.10 0.08 26.6 −0.37 0.08 4.7 −2.65 0.09 −29.9 0.96 0.44 2.2 −0.09 0.02 −4.4 0.03 0.02 1.4 −0.006 0.003 −2.0 −0.18 0.032 −5.7 0.171 ( p < 0.000)

4.04

S.E.

t-value

0.12 33.8 N/A N/A −0.005 0.002 −2.8 0.31 0.02 18.8 0.19 0.005 35.0 0.58 0.02 31.4 0.01 0.02 0.5 −0.61 0.02 −28.7 0.46 0.06 7.8 N/A N/A −0.00 0.001 −0.3 −0.15 0.008 −18.0 0.075 ( p < 0.000)

effects of domestic alignments. On average, citizens in conservative welfare states with fragmented wage-bargaining are 5.71 units on the twentyone-point OIEV scale more in favor of European integration than citizens in encompassing redistributive welfare states with sectorally centralized wage-bargaining! A right-winger in Sweden is more pro-European than his leftist domestic adversary, but still less pro-European than the average conservative in Britain who is much more Euroskeptical than his leftwing antagonist. And right-wingers in both residual and encompassing redistributive welfare states are less pro-European than their ideological counterparts in conservative welfare states, regardless of whether they live under centralized or decentralized wage-bargaining.

Conclusion We draw three lessons from our analysis. First, varieties of capitalism matter for public views of European integration. Contingent upon the national political-economic status quo, citizens have more or less

86


favorable dispositions toward EU integration (the direct “sociotropic” mechanism). Both prevailing wage-bargaining institutions and the presence or absence of conservative Christian democratic welfare states leave a distinctive imprint on political opinions about EU integration. Furthermore, political-economic institutions also shape the domestic political alignments among citizens over the EU issue. The critical contextual condition that interacts with domestic ideological divides is the extent of social redistribution achieved by national welfare states. In redistributive welfare states it is the left that opposes further EU integration, in liberal-residual welfare states, the right. In conservative Christian democratic welfare states, there is not much left/right polarization of the issue. The two-level conditioning of conflict over European integration lends additional plausibility to theories that consider the European Union a multilevel polity. Second, where salient, the politicization of European integration can be integrated into prevailing patterns of left-libertarian versus rightauthoritarian party competition, though in contextually contingent configurations (cf. Marks and Wilson 2000; Marks, Wilson, and Ray 2002). In liberal-residual and in encompassing, redistributive welfare states, the EU issue clearly reinforces political alignments already established in the polity. In the remaining countries with conservative welfare states or laggard economic development, a politicization of Europe is more likely to cross-cut existing left-libertarian versus right-authoritarian alignments, but here it is less plausible that the EU issue might become highly salient and divisive. As long as countries with such political-economic configurations – e.g., France, Italy, Germany, or Spain and associated smaller countries – dominate the EU polity, voters may not fear that the policy consequences of integration could dramatically dislocate the domestic status quo in social and economic policy. The further extension of the EU to east-central Europe, however, may signal the end of such confidence in the hegemony of conservative Christian democratic politics and economic policy-making. In light of this process, citizens in all countries may revise their estimates of the costs and benefits of European integration with consequences for political alignments that cannot yet be anticipated. Third, for the time being, the rather loose empirical association among citizens’ responses to various measures of support for European integration suggests that this issue basket is not highly salient in most polities. There is a certain arbitrariness with which most respondents treat questions relating to European integration, even though the central tendencies of their evaluations can be subjected to systematic analysis. People’s limited attribution of salience to the European integration issue may derive


87

from the fact that European governance is more important as a catalyst of (de-regulating) markets than directly shaping resource allocation through social policies. As long as the size of the EU budget remains modest and constraints on national social policies are confined to general fiscal and monetary policy parameters, national political parties will overwhelmingly focus on national party agendas (Mair 2000). Even in Britain, where contextual conditions favor the politicization of the European issue by the partisan right and the left, the dismal electoral fortunes of the British Conservatives with stridently anti-European appeals in recent elections show that domestic policy issues cannot be easily replaced by a new political agenda. Nevertheless, questions of European integration leave an impact on domestic politics via their association with national patterns of political-economic institutions and governance processes.

Appendix 1: Variable construction Index of overall evaluation of European integration (OEIV) Desired speed: slow

Desired speed: fast

Perceived speed: slow

2 Status quo contented I: Slow, but adequate

4 Militant advocates: Want to speed up the process

Perceived speed: fast

1 Militant opponents: Want to slow down integration

3 Status quo contented II: Fast and adequate

Desired speed

Perceived speed

14 = high

12

10

8

6

4

2 = low

1 = high

15

13

11

9

7

5

3

2

16

14

12

10

8

6

4

3

17

15

13

11

9

7

5

4

18

16

14

12

10

8

6

5

19

17

15

13

11

9

7

6

20

18

16

14

12

10

8

7 = low

21

20

18

16

14

12

10

88


Index of citizens’ preference for less welfare state redistribution The index is constructed using six questions (Q105 a, b, c, f, g, h) from the Eurobarometer 44.3OVR questionnaire. Each question uses a threepoint scale in which we recoded the answers so that the higher the score, the less welfare redistribution was preferred. It is simply an additive index with a range of 6–18, with higher scores indicating a preference for less welfare. Q.105. Do you tend to agree or tend to disagree with each of the following statements? Original scale ranges from: 1 = tend to agree, 2 = tend to disagree, and 3 = don’t know. Q105a: I would be ready to pay more tax if I were sure it would be devoted to creating new jobs (1–3, with 3 = tend to disagree). Q105b: Public ownership of industry should be expanded (1–3, with 3 = tend to disagree). Q105c: The welfare state costs too much to be maintained in its present form (1–3, with 3 = tend to agree). Q105f: Government should play a greater role in the management of the economy (1–3, with 3 = tend to disagree). Q105g: The welfare state makes for a fairer society (1–3, with 3 = tend to disagree). Q105h: The welfare state reduces the will to work (1–3, with 3 = tend to agree). Index of citizens’ preference for social and political gender equality The index is constructed using four questions (Q98 a, b, c, d) from the Eurobarometer 44.3OVR data set. The variables were recoded so that the higher the score, the more gender equality was accepted, with don’t know respondents in the middle. It is an additive index with a range from 4 to 20 with higher scores indicating more support for gender equality. Q.98. Could you please tell me if you agree totally, agree, disagree or disagree totally with the following opinions. Equality between women and men . . .? Original scale ranges from: 1 = agree totally, 2 = agree, 3 = disagree, 4 = disagree totally, 5 = don’t know. Q98a: strengthens democracy (1–5, with 5 = agree totally). Q98b: makes the personal development of women and men easier (1–5, with 5 = agree totally).


89

Q98c: brings more harm than good in daily life (1–5, with 5 = disagree totally). Q98d: improves the quality of human relationships (1–5, with 5 = agree totally). Appendix 2 The following correlation matrix is derived from Eurobarometer 44.2BIS. We did not employ this survey in many of our statistical estimations, because it misses variables that would allow us to measure the ideological positions of our respondents (market liberalism; gender equality) at the individual level. Nevertheless, 44.2BIS shows the limited relationship among the variables: Perceived Desired Overall EU Index of EU actual speed European membership In favor EU support membership speed of EU of EU integration good, bad of EU (4) + (5) benefits integration integration view (OEIV) or neither integration = my country (1) (2) (3) (4) (5) (6) (7) Perceived actual speed of EU integration Desired speed of EU integration

1.000

−0.04

1.000

Overall European integration view (OEIV)

0.30

0.94

1.000

EU membership good, bad or neither

−0.08

0.27

0.23

1.000

In favor of EU integration

−0.06

0.49

0.45

0.44

1.000

Index of EU support

−0.08

0.47

0.42

0.79

0.90

1.000

EU benefits [my country]

−0.12

0.32

0.27

0.44

0.48

0.54

1.000

Shaded areas indicate a mathematical-definitional association between concepts (individual indicators and aggregative indices of attitudes toward European unification).

Part II

Political parties

5

Defining the EU political space: an empirical study of the European election manifestos, 1979–1999 Matthew J. Gabel and Simon Hix

As Steenbergen and Marks (Introduction) describe, scholars of EU policy-making have adopted conflicting assumptions about the dimensionality and character of the EU policy space. Since the shape of the political space – the number of dimensions, the policy content of these dimensions, and the location of actors in this space – is a central determinant of political competition and outcomes, these conflicting assumptions often lead to different conclusions about and interpretations of EU policy-making. This is a serious impediment to advancing our theoretical understanding of EU politics. A resolution of this theoretical conflict depends on assessing the relative value of the conflicting assumptions about the character of the policy space. To help address this problem, we attempt to examine empirically whether the structure of the EU political space is consistent with these existing models. Specifically, we investigate whether the existing models described by Marks and Steenbergen (in this book) account for the EU policy space as defined by the European party federations – or “Europarties.” These Euro-parties bring together the domestic and Europeanlevel political elites in the four main European party families – socialists, Christian democrats/conservatives, liberals, and greens. In the European elections of 1979, 1984, 1989, 1994, and 1999, the Euro-parties drafted manifestos describing their positions across a broad range of policies involving the EU. We use an established content analysis technique to turn these text documents into numerical data representing Euro-parties’ positions on specific political issues. Through confirmatory factor analysis, we then examine whether these policy positions are related in ways consistent with existing models of the EU political space. Finally, we describe how the positions of the “Euro-parties” in the EU political space have changed over the last twenty years.

93

94

Matthew J. Gabel and Simon Hix

The shape of the EU policy space: four rival models Steenbergen and Marks (Introduction) describe four extant models of the EU policy space: the international relations model, the regulation model, the Hix–Lord model, and the Hooghe–Marks model. To examine the empirical accuracy of these models of the EU political space, we adopt a common analytical approach in the empirical study of the political space in national contexts (e.g., Budge, Robertson, and Hearl 1987; Gabel and Huber 2000; Budge et al. 2001). A traditional approach to characterizing a policy space is to describe its dimensionality, where dimensions represent constraints on the policy positions of political actors. For example, to understand the political space of national party competition, scholars examine how parties’ positions on a variety of salient policy questions are interrelated (e.g., Budge, Robertson, and Hearl 1987; Budge et al. 2001; Laver 2001). A dimension reduces differences in parties’ positions across many issues to differences on that dimension. To take one common example, the left/right dimension simplifies differences in parties’ positions on a variety of socio-economic issues into differences along a single dimension. As a result, if one can identify the relationship between specific policies and the underlying dimensions, then one can infer parties’ positions on individual policies from their positions on the underlying dimensions (Gabel and Huber 2000: 95). Each of the aforementioned models of the EU political space specifies a distinct dimensionality to the EU political space. To examine their accuracy, we need information about the positions of relevant political actors on a variety of EU political issues. In the national context, a common source of such information would be party manifestos, which present parties’ promises and positions regarding a vast array of issues (see esp. Budge et al. 2001). Since national parties compete for office and adopt policies, these actors’ policy positions are an obvious object of study for defining the national political space. However, in the EU context, no comparable source of data exists. There are no elections determining control of EU governance that are contested by political parties who issue manifestos detailing their positions on a broad array of EU issues.1 But there is one source of data that offers a reasonable approximation: the manifestos of the Euro-parties.

1

National representatives chosen in national elections constitute the Council of Ministers – which must approve all legislation – and the European Council.

Defining the EU political space

95

Data: the Euro-parties’ election manifestos, 1979–1999 Euro-parties are federations of national parties that organize according to political family into the following groups: socialist, Christian democrat/ conservative, liberal, and green. Like their constituent national parties, the Euro-parties have their own administrative organization, budget, and secretariat (Hix and Lord 1997: 63). The Euro-parties are active in setting and coordinating the EU policy agenda for their members. In addition, “party leaders’ meetings” – which consist of national party leaders, European Commissioners, and party group leaders in the European Parliament (EP) from the same party family – help coordinate policy positions and proposals across different EU policy-making institutions (Hix and Lord 1997: 64). Hix and Lord (1997) argue that these Euro-parties serve as effective umbrella organizations for structuring the policy agendas of their members in EU policy-making. Thus, while representatives of national parties are the primary actors in EU policy-making (in the Council of Ministers and the European Parliament), their behavior is structured by the overarching Euro-parties. These Euro-parties state their positions publicly in “manifestos” that are adopted immediately before the elections for the EP (which have been held every five years since 1979). Specifically, we have collected, coded, and analyzed the manifestos of the socialist, Christian democrat/ conservative, liberal, and green “Euro-parties.” We will use this information about Euro-party policy positions to examine the aforementioned models of the EU political space. These manifestos are a valuable source of information about the EU policy space for several reasons. First, these documents are a reasonably accurate statement of the positions the European political elites take on issues on the EU agenda at different points in time. Granted, these documents are rarely used in the parties’ election campaigns in the domestic arena, and are not well known by European citizens. However, the domestic media like to point out the inconsistencies between the commitments made by national party leaders in these European manifestos and in their national party programs or manifestos. Consequently, to monitor commitments made at the European level, national party leaders have increasingly delegated senior party officials to the task of negotiating these documents. For example, in 1998 the British Labour Party sent the Foreign Minister, Robin Cook, to lead the working group drafting the European socialist party family’s manifesto for the 1999 European elections. Similarly highranking figures were sent by the French, German, and Italian socialist

96


parties, all of whom were in government in the domestic arena at the time. And, once these negotiations are complete, the manifestos are signed at the highest political level, by the national party leaders. These signatories are the key political actors in the EU system, competing for the highest offices in the domestic arena, and scrutinizing the behavior of their representative in the EU institutions – ministers in the Council, Commissioners, and the national delegations of the parties in the EP. Second, since their emergence around the time of the first EP elections in the mid-1970s, the member parties of the four transnational party federations in our analysis have dominated politics in the EU (Hix 1996; Hix and Lord 1997: 30–9, 167–97). The Party of European Socialists (PES) incorporates the major party on the center-left in every EU member state, who together receive the support of about 30 percent of voters in the EU. In November 2001 they made up 181 of the 626 seats in the EP, ten of the twenty Commissioners, ten of the fifteen EU Prime Ministers, and were in government in eleven of the fifteen EU states.2 The European People’s Party (EPP) involves twenty-three national Christian democratic and conservative parties and the EPP-European Democrats (EPP-ED) group in the EP, who together incorporate the main center-right party in every EU state except Ireland. These parties command the support of approximately 30 percent of voters in the EU and, in November 2001, EPP politicians made up 232 of the 626 seats in the EP, six of the twenty Commissioners, four of the fifteen EU Prime Ministers, and were in government in five of the fifteen EU states.3 The European Liberal, Democrat and Reform Party (ELDR) incorporates eighteen liberal, centrist, reform, radical, and agrarian parties and the ELDR group in the EP. In the mid-1990s the ELDR lost two of its most influential parties to the EPP: the Portuguese social democrats and the French republicans. Nevertheless, the ELDR remains present in every EU state, and these parties command the support of approximately 10 percent of voters in the EU. In November 2001, ELDR politicians made up 51 of the 626 seats in the EP, two of the twenty Commissioners, one of the fifteen EU Prime Ministers, and were in government in eight of the fifteen EU states.4 Finally, the European Federation of Green Parties (EFGP) incorporates eighteen environmentalist, ecologist, and green-left parties and the green members of the Green/European Free Alliance group in the EP. These parties command the support of approximately 5 percent of voters 2 3 4

This information was taken from the web site of the PES: http://www.eurosocialists.org. This information was taken from the web site of the EPP: http://www.eppe.org/home/ default.asp. This information was taken from the web site of the ELDR: http://www.eldr.org.


97

in the EU, and in November 2001 EFGP politicians made up 38 of the 626 seats in the EP, one of the twenty Commissioners, one of the EU Foreign Ministers (from Germany), and were in government in four of the fifteen EU states.5 Overall, these four Euro-parties dominate politics at the national and European levels in the EU, commanding the support of over 75 percent of voters, controlling over 80 percent of seats in the European Parliament, 100 percent of EU Commissioners, over 90 percent of EU Prime Ministers, and over 85 percent of parties in government. Hence, if we look at how these parties position themselves against each other in their manifestos for European elections, we can gain a fairly accurate picture of the terrain of politics in, and relating to, the European Union. Coding and measurement: from manifesto promises to policy locations The EPP, PES, and ELDR have adopted manifestos in every European election since 1979, whereas the EFGP has only done so since the 1989 elections. This leads to eighteen Euro-party manifestos in total – three from the 1979 and 1984 elections, and four from the 1989, 1994, and 1999 elections. To convert these texts into numerical data for analysis, we used a standard content analysis technique which has been widely used to examine national party manifestos (Budge, Robertson, and Hearl 1987; Budge et al. 2001). This technique involves the following steps. First, we drew up a coding frame, which is set out in the appendix. Our coding frame is based on Michael Laver’s reformed version of the ECPR party manifestos project coding frame (Laver and Garry 2000; Laver 2001). This scheme differs from the ECPR scheme in three ways. First, it is more hierarchical than the ECPR scheme, which allows for categories to be amalgamated more systematically. Second, the scheme includes exclusively directional categories. The ECPR scheme also included categories that were simply designed to tap issue saliency – e.g., the category “free enterprise” – without any clear directional character to the category. As we are interested in the relative positions of Euro-parties on each issue rather than the relative saliency, this coding is appropriate to our analysis. Third, to Laver’s scheme we add a whole new set of issue categories (categories beginning with the number 6), relating to positions on the general question of European integration, on specific views on EU institutions and EU policies, and on a number of general issues on the EU agenda. 5

This information was taken from the web site of the EFGP: http://www.europeangreens. org.

98


Table 5.1 Number of political statements (“raw scores”) in the manifestos

PES EPP ELDR EFGP Average

1979

1984

1989

1994

1999

Average

246 76 760 – 361

429 26 659 – 371

389 75 339 134 234

258 82 273 512 281

173 96 257 220 187

247 71 458 289 278

Second, each document is then “coded,” using the same contentanalysis technique as that used in the ECPR party manifestos project (Budge, Robertson, and Hearl 1987; Budge et al. 2001: 215–18). This process involves coding every “statement” (sentence or part-sentence) in each manifesto for one and only one of the issue categories in the coding frame. From these “raw scores,” we then calculated the proportion (percentage) of sentences in each document dedicated to each policy category. These percentages are the data we use in the analysis.6 For example, if a manifesto contains two sentences stating that the party supports the EU single market, followed by one statement qualifying this support (for example, by criticizing the failure of the EU to complete the single market effectively), the first two sentences are coded as “support for the single market” (category 63121) and the third sentence is coded as “criticism of the single market” (category 63123). Then, if the manifesto contains 100 statements and there are no other references to the single market in the document, the proportional score for category 63121 is 2.0 percent and for category 63123 is 1.0 percent. The result is a valuable data set which could have a number of applications in the study of EU politics. To give the reader an idea of the extent of this data set, table 5.1 shows the number of political statements (raw scores) in each manifesto. As the table shows, the EPP tended to issue shorter manifestos than the other parties. However, this does not bias the results. The EPP manifestos covered as many issues as the other parties’ manifestos, simply in less depth. Because we are interested in the 6

One potential problem with this coding method is that, by using percentages of total statements, the percentages in each category are not independent. However, since manifestos vary – sometimes dramatically – in length, we need to adjust the categories to control for the level of verbosity in each text. We should also note that this coding method has proven effective in identifying the dimensionality of the political space in the national context (e.g., Gabel and Huber 2000).


99

proportion of the manifesto dedicated to each issue category rather than the absolute raw scores, the length of the manifesto is irrelevant for our analysis. Third, to create a smaller number of variables for statistical analysis we combined the percentage scores into a number of consolidated issue categories.7 These categories are designed both to capture the breadth of the content in the manifestos and to allow us to differentiate among the empirical claims of the theoretical models of the EU political space discussed earlier. We focus on these categories because they involve issues of both EU policy and integration, which are the issues hypothesized to characterize the EU political space. Note that each category has a directional opposite, which is necessary for coding purposes.8 Consistent with this expectation, these categories represent, on average, about 72 percent of all statements in a manifesto.9 Below, we list the categories (the subcategories from the appendix are in parentheses): 1. Pro state/economic intervention (11, 3311, 3332) 2. Anti state/economic intervention (13, 3331, 3312) 3. Non-traditional political/social values (21, 141, 311, 3314, 3316, 3318, 3333) 4. Traditional political/social values (23, 142, 313, 3313, 3334, 3336, 3338) 5. Pro democratic EU institutional reform (62311, 62321, 62331, 6421, 6431, 6441, 64721, 64731, 64741, 64751, 6491, 62513, 62521, 63323, 62611, 6112) 6. Anti democratic EU institutional reform (62313, 62323, 62333, 6423, 6433, 6443, 64723, 64733, 64743, 64753, 6493, 62511, 62523, 63321, 62613, 6132) 7. Pro new EU regulatory/redistributive policies (63111, 63131, 63141, 63343, 63251, 63333, 63361, 63371, 63431) 8. Anti new EU regulatory/redistributive policies (63113, 63133, 63143, 63341, 63253, 63331, 63363, 63373, 63433) 9. Pro EU enlargement (64101) 7

8

9

Due to low levels of covariation across many of the categories, it was difficult to identify the confirmatory factor analysis models estimated here. Also, given the empirical implications of the models of the EU political space, we needed to create super-categories consistent with the hypothesized policy areas. Note that the hypotheses we test identify specific relationships between positive and negative statements across issue areas. We therefore want to test these claims directly, not through some aggregate pro/con score. Note that the EU reform categories were omitted because they average about 1 percent of statements in a manifesto.

100


10. Anti EU enlargement (64103) 11. Pro EU economic integration (63121, 63311, 63221, 63161, 63411, 6111) 12. Anti EU economic integration (63123, 63313, 63223, 63163, 63413, 6131) 13. Pro new EU political integration (63511, 63521, 63611, 63621, 63631, 63641, 63651) 14. Anti new EU political integration (63513, 63523, 63613, 63623, 63633, 63643, 63653) Some of these categories deserve further discussion. The categories concerning institutional reform of the EU (numbers 5 and 6) include statements regarding (a) strengthening the European Parliament vis-à-vis other EU institutions, (b) increased transparency and reduction in the democratic deficit, and (c) increased interest representation in the EU. The categories concerning “new EU regulatory/redistributive policies” include statements about new policies that have a specific expected policy outcome. Policy outcomes that involved regulation of business or the economy or which generally favored labor at the expense of capital were coded as “pro.” For example, support for more EU regulation of the economy or environment was coded “pro” in this category. A statement supporting “the single goal of price stability” was coded as “anti.” The categories for “EU economic integration” include statements referring to the fundamental economic components of EU membership: the internal market, competition policy, common external trade policy, and the Common Agricultural Policy. The categories for “new EU political integration” include statements referring to further policy integration, but with no policy outcome specified. For example, a statement in support of common immigration policies does not specify whether that policy would be restrictive or liberal. Table 5.2 presents the proportion of all parties’ manifesto commitments in each election dedicated to these consolidated issue categories. Furthermore, most of these categories had broadly the same level of saliency across each election. The few exceptions are easily explainable. For example, “pro new EU regulatory/redistributive policies” was highly salient in the 1989 election following the delegation of new regulatory competences to the EU level in the 1987 Single European Act, which included the program to complete the EU single market by the end of 1992. Conversely, the issue of EU enlargement (either “pro” or “anti”) was not very salient in the 1989 election since, following the membership of Spain and Portugal in 1986, it was clear that the next round of enlargement (which did not come until 1995) was several years away.


101

Table 5.2 Percentage of all the parties’ manifestos dedicated to each issue category 1979

1984

1989

1994

1999

Avg.

Pro state/economic intervention Anti state/economic intervention Non-traditional political/social values Traditional political/social values Pro democratic EU institutional reform Anti democratic EU institutional reform Pro new EU regulatory/redistributive policies Anti new EU regulatory/redistributive policies Pro EU enlargement Anti EU enlargement Pro EU economic integration Anti EU economic integration Pro new EU political integration Anti new EU political integration

16.7 3.5 13.8 2.6 4.9 0.6 8.9

11.2 3.5 12.1 2.1 7.2 0.2 9.5

7.1 3.3 14.0 2.7 5.9 0.1 11.3

7.9 4.5 13.2 1.9 6.1 0.8 8.5

10.9 7.3 13.1 1.0 7.0 0.5 4.3

10.7 4.4 13.3 2.1 6.2 0.4 8.5

0.5

0.6

0.9

1.4

0.7

0.8

1.5 0.2 3.3 0.8 2.0 0.5

2.0 0.2 6.1 1.7 5.3 0.4

0.2 0.1 6.4 4.1 4.8 2.4

1.6 0.6 4.0 2.9 8.2 1.7

2.9 0.5 3.8 1.2 6.5 1.5

1.6 0.3 4.7 2.2 5.4 1.3

Total

59.8

62.1

63.2

63.5

61.2

62.0

Finally, table 5.3 presents the correlation matrix for these twelve variables. Note that with only eighteen observations, many correlations are insignificant at standard levels of statistical significance. Statistical analysis Method To test the rival models of the shape of the EU political space in a statistical analysis, we use confirmatory factor analysis of these fourteen issue categories. Factor analysis examines how the covariation among manifesto emphases across issue categories is structured by latent or underlying factors (dimensions). Confirmatory factor analysis evaluates the performance of a particular factor structure in explaining the covariation among these parties’ emphases across policy categories. The factor structure specifies the number of factors, their interrelationships, and their relationships to the categories, which serve as indicators of the factors. The performance of the factor structure is evaluated through a measurement model that assesses the fit of the structure with the data. Note that we estimate the measurement models with the full set of manifesto data for all five European elections. With all manifestos included, we

Note: ∗ p < 0.10.

+statist −statist +trad. value −trad. value +dem. reform −dem. reform +enlarge −enlarge +EU statist −EU statist +econ. integ. −econ. integ. +political integ. −political integ. – −0.49∗ −0.12 −0.42∗ 0.07 0.30 0.18 0.33 −0.16 0.30 0.19 −0.12 0.45∗ −0.28

+statist

– −0.24 0.50∗ −0.44∗ −0.20 −0.04 −0.32 −0.28 −0.21 −0.69∗ 0.25 −0.55∗ 0.35

−statist

– 0.18 0.33∗ −0.14 0.27 −0.28 −0.17 −0.29 0.03 −0.43∗ 0.02 −0.18

+trad. value

– −0.40∗ 0.04 0.04 −0.28 −0.38 0.06 −0.50∗ 0.37 −0.63∗ 0.28

−trad. value

– −0.10 0.19 0.20 0.05 −0.25 0.42∗ −0.47∗ 0.29 0.09

+dem. reform

– 0.32 0.48∗ −0.18 0.32 −0.13 −0.26 0.42∗ −0.36

−dem. reform

Table 5.3 Correlation matrix (Pearson correlation coefficients)

– 0.05 −0.41∗ −0.21 −0.18 0.31 0.10 −0.46∗

+enlarge

– −0.37 0.50∗ 0.06 −0.19 0.65∗ 0.06

−enlarge

– −0.37 0.48∗ −0.12 −0.10 −0.31

+EU statist

– −0.13 0.51∗ 0.39 0.33

−EU statist

– −0.18 0.33 −0.31

+econ. integ.

– −0.31 0.61∗

−econ. integ.

– −0.20

+political integ.

–

−political integ.


103

have eighteen observations (manifestos) for each policy category, which is a very small number for estimation purposes. We therefore cannot estimate measurement models for subsets of the data by year or by party federation. The four models predict different factor structures for the covariation in the issue category variables: 1. International relations model: All variables should load on one factor (dimension), which will be identified by the variables measuring European integration (numbers 5 to 14). For these variables, the parameter estimates for the “pro” variables should be in the same direction (positive or negative), with the parameter estimates for the “anti” variables sharing the opposite direction. 2. Regulation model: All variables should load on one dimension, with the variables for left/right (numbers 1 to 4) defining the dimension. The pro-state/intervention variable should load in the opposite direction to that of the anti-state/intervention variable and the traditional values variables should load in the opposite direction to the non-traditional values dimension. Also, the variables for EU-level regulation should load consistently with the left/right variables. 3. Hix–Lord model: The variables should load on two distinct (orthogonal) dimensions. The European integration variables should load on one dimension and the traditional left/right socio-economic variables (numbers 1 to 4) should load on the other dimension. On the European integration dimension, the parameter estimates for the “pro” variables should be in the same direction (positive or negative), with the parameter estimates for the “anti” variables sharing the opposite direction. On the left/right dimension, the pro-state/intervention variable should load in the opposite direction to that of the anti-state/intervention variable and the traditional values variables should load in the opposite direction to the non-traditional values dimension. These two dimensions should be unrelated. 4. Hooghe–Marks model: The variables should load on two dimensions as specified in the Hix–Lord model. However, these dimensions should be strongly related to each other such that pro-integration positions are associated with left positions and anti-integration positions are associated with right positions. This would result in a policy space where basically two camps compete: those supporting “regulated capitalism” and those supporting “neoliberalism” at the EU level. Results Table 5.4 presents the maximum likelihood results from the three measurement models designed to evaluate the aforementioned models of the

Note: ∗ p < 0.05.

0.780∗ −0.629∗ 0.641∗ −0.054

−0.761∗ 0.545∗ −0.708∗ 0.049 0.447∗ 0.349 0.128 0.542∗ 0.118 0.196 0.550∗ −0.435∗ 0.812∗ −0.393 – 0.499 0.061 0.218 0.138

Pro-state/intervention Anti-state/intervention Non-traditional values Traditional values Pro EU democratic reform Anti EU democratic reform Pro new EU regulatory/redistributive policies Anti new EU regulatory/redistributive policies Pro enlargement Anti enlargement Pro EU economic integration Anti EU economic integration Pro new EU political integration Anti new EU political integration Factor correlation GFI NNFI IFI CFI – 0.492 −0.018 0.218 0.205

Factor 1

Factor 1

0.181 0.542∗ −0.108 0.916∗ −0.367 0.525∗ 0.069 −0.201 0.719∗ −0.010

Factor 2

Model 2: Two Factors (orthogonal)

Observed indicator

Model 1: Single Factor

Table 5.4 Factor patterns (standardized solution)

−0.838 0.505 0.045 0.280 0.202

0.718∗ −0.604∗ 0.722∗ 0.002

Factor 1

0.333 0.449∗ 0.119 0.674∗ −0.082 0.368 0.359 −0.345 0.952∗ −0.241

Factor 2

Model 3: Two Factors (interrelated)


105

EU political space. We report the standardized coefficients for these models as well as several measures of model fit. As recommended by Hoyle and Panter (1995), we report the following fit indices: the goodness-offit index (GFI, Joreskog ¨ and Sorbom ¨ 1981), the non-normed fit index (NNFI, Bentler and Bonnett 1980), the incremental fit index (IFI, Bollen 1989), and the comparative fit index (CFI, Bentler 1990). The values of GFI, NNFI, IFI, and CFI range from 0 to 1.0, with higher values indicating better model fit (Bentler and Bonnett 1980; Hoyle and Panter 1995). The first model assumes one factor underlies all fourteen categories. For the non-integration variables (1–4), opposing categories load on this dimension in opposite directions (one positive, one negative). This is not true for the integration variables. Of these, only the variables for new political integration and economic integration load in opposing directions by topic (and only the economic variables are both significant). The basic interpretation of the dimension is that parties that are anti-regulation in general express traditional values and support new political integration but also support economic integration. Note that parties that are against new EU regulatory policies are for economic integration and new EU political integration. These results are inconsistent with the international relations model, as the integration variables do not load consistent with expectations. The results are roughly consistent with the regulation model, as the traditional left/right variables load consistent with expectations and the antistate/intervention and the anti-new EU regulatory policies variables load in the same direction. Thus, of the one-factor conceptual models, the regulation model performs best. Model 2 presents a two-factor model designed to test the Hix–Lord model. The results are consistent with expectations on the left/right dimension (factor 1). However, on the national sovereignty dimension (factor 2), the loadings are not clearly indicative of a pro-/anti-integration dimension. The loadings indicate that parties that are anti-enlargement are also anti new EU regulatory policies and anti EU democratic reform. But these parties would be favorable to new political integration. In addition, this factor does not account for parties’ positions on EU economic integration. It is also important to note that the goodness of fit of this model is generally inferior to the one-factor model presented in model 1. Model 3 attempts to capture the specifications of the Hooghe–Marks model by relaxing the assumption of orthogonality. In terms of interpretation, the factor loadings in this model are basically the same as those in model 2. The inter-factor correlation is quite high (−0.838), suggesting

106


that these two dimensions are strongly related. The factors interrelate such that positions that are pro state intervention and anti traditional values on the first dimension are negatively related to anti EU democratic reform positions and anti new EU regulatory policy positions on the second dimension. This is consistent with expectations, representing a regulated capitalism camp. However, these same “left” positions on the first dimensions are negatively related to pro EU political integration positions and have an ambiguous relationship to positions on EU economic integration. This is inconsistent with a “regulated capitalism” agenda. As for the “neoliberalism” camp, there is some supporting evidence. Positions against state intervention were positively related to positions opposed to EU democratic reform and new EU regulatory/redistributive policies. But these “right” positions on the first dimension were positively related to positions favoring new EU political integration. This is not consistent with a neoliberalism agenda. One reason for the unexpected results regarding positions on new EU political integration is that these categories did not specify the direction of policy in a proposed area of new EU authority. Support for new political integration simply indicated that a manifesto supported the extension of EU authority to a new policy area – irrespective of whether that policy might be consistent with a left or right ideological position. Thus, the relationship between the left/right dimension and these positions is not particularly important for the Hooghe–Marks model. The more important categories are those regarding institutional reform and regulation of the EU market. For those categories the model performs generally as expected. Consequently, we find the results, on balance, supporting the Hooghe–Marks model. The model fit for model 3 is an improvement on model 2 but not clearly better than model 1. The Bentler–Bonnett non-normed fit index is smaller for model 3 than for model 1 while the other goodness-of-fit measures favor model 3 slightly over model 1. Thus, particularly with the high factor correlation in model 3, we cannot conclude that the Hooghe– Marks model is a better representation of the EU policy space than the one-factor model. Our interpretation of these results is that the policy space is basically one-dimensional, consistent with the regulation model, where European economic integration represents an anti-regulation enterprise. That is, Euro-parties that are against state intervention in the economy are also for economic integration in the EU. Certainly there is evidence that this space has a multidimensional character and is basically consistent with the Hooghe–Marks model, but the correlation between these dimensions is very high.


107

Descriptive analysis Method Finally, we also used these manifesto data to evaluate how the parties differentiate themselves from each other and over time in the policy areas and whether this is in accordance with the expectations of the four models of the EU political space. As mentioned above, for the confirmatory factor analysis, we did not have sufficient degrees of freedom to examine the temporal dynamics of the EU policy space or Euro-party policy positions. However, we can give some descriptive account of these changes by simply tracking Euro-party policy positions in particular areas over time. We use the raw score data from the manifestos to calculate the location of the parties in each election in different policy areas. Following the ECPR party manifestos project method, the location of a party on a particular issue is the proportion (percentage) of a manifesto dedicated to one side of the issue minus the proportion dedicated to the other side of the issue (Budge and Klingemann 2001; Budge 2001). For example, if a manifesto contains 10 percent of “pro state/economic intervention” statements and 5 percent of “anti state/economic intervention” statements, the manifesto is located at +5.0 on the state/economic intervention dimension. Results Figure 5.1 shows that the Euro-parties were relatively distinct on economic left/right issues (figure 5.1a). The socialists moved dramatically to the right in the 1980s, and came close to the positions of the EPP and the liberals in the 1989 election. But in 1994 and 1999 the socialists again presented a distinct “center-left” agenda. The EPP (who at that time had few explicitly “conservative” member parties) also moved from a more centrist economic agenda in the late 1970s, to the left of the liberals, to a clear free-market agenda in the late 1990s, very close to the position of the liberals. Also, the EFGP party emerged as a new force – clearly to the left of the socialists in the 1999 election. The results consequently support the Hix–Lord, Hooghe–Marks, and regulation model claims about party distinction on economic issues relating to the traditional left/right conflict. On European integration issues, figure 5.1b shows that the positions of the two main parties have almost completely reversed. At the end of the 1970s and in the early 1980s, the EPP were generally pro-integration,

108


(a) 30%

20% PES EPP

10%

ELDR EFGP

0%

−10% 1979

1984

1989

1994

1999

(b) 15% 10% 5%

PES EPP

0%

ELDR EFGP

−5% −10% −15% 1979

1984

1989

1994

1999

Figure 5.1 Euro-party locations over time: (a) economic left/right (pro- vs. anti-state intervention; (b) economic integration (pro- vs. anti-).

whereas the socialists were generally more skeptical. Conversely, at the end of the 1990s, the socialists were the main champions of economic integration and the EPP were one of the more skeptical forces. In between, the liberals followed the EPP on issues of EU economic integration. And the greens emerged as the only anti-integrationist force at the European level. Since the parties crossed paths on many of these issue dimensions,

EU economic integration


109

PES EPP ELDR EFGP

Economic left-right

Figure 5.2 Euro-party positions in a two-dimensional space.

the results refute the claims of all the models that parties should maintain clear and distinct positions on European integration issues. However, if we map the party positions at the beginning and end of the period in a two-dimensional space, we can modify this conclusion. As figure 5.2 shows, across the whole period, as the Hix–Lord model predicts, all four positions were present: left/anti-integration (socialists in 1979 and greens in 1999), left/pro-integration (socialists in 1999), right/ anti-integration (EPP and liberals in 1999), and right/pro-integration (liberals and EPP in 1979). As a result, the Hooghe–Marks claim about the relationship between the two dimensions does not hold in 1979, when the center-left PES was more anti-European than both the EPP and ELDR. But their claims do hold in 1999. The socialists became more pro-European as they began to endorse regulatory capitalism at the European and national levels (instead of welfare capitalism at the domestic level). On the other side, the EPP became more anti-European, as they began to advocate neoliberal economic policies, and the already neoliberal ELDR became more anti-European as the EU became more regulatory. Nevertheless, the regulation model prediction that the party line-up on the economic left/right should map directly onto the pro-/anti-Europe dimension does not hold. In 1979, whereas it was PES–EPP–ELDR on the left/right, it was the reverse (ELDR–EPP–PES) from pro- to antiintegration. In other words, at the beginning of the period, the correlation

110


between the directions was in the opposite direction to the regulation model prediction, with the left mapping onto anti-Europe, and the right mapping onto pro-Europe. The predictions were closer in 1999 – when it was PES vs. EPP/ELDR from left to right and PES vs. ELDR/EPP from pro- to anti-integration. But this only holds if the greens are excluded as outliers (perhaps on the grounds that they are weaker at the European level than the other party families), and the liberals, conservatives, and Christian democrats are treated as a single center-right bloc. Conclusion As in Gabel and Anderson (chapter 1) and Hooghe, Marks, and Wilson (chapter 6), we have explored empirically whether political agents take positions on EU policy that are structured by ideological dimensions that define a coherent EU policy space. We have focused on the policy positions of “Euro-parties” as presented in their electoral manifestos. Our results indicate that the positions of these parties are structured in systematic ways interpretable through the lens of existing models of the policy space. First of all, the manifesto data reveal that the Euro-parties have differentiated themselves on some major policy issues, but this differentiation is not simply a copy of the domestic ideological “family” distinctions between socialists, Christian democrats/conservatives, and liberals. On economic issues on the EU agenda and relating to European integration, Euro-parties consisting of national parties on the domestic ideological left or right took positions that were distinct and consistent with these ideological positions. But on the questions of the speed and nature of economic integration in Europe (such as what powers should be delegated to the European level), the Euro-parties were less consistent in how they differentiated themselves from one another in the positions they took. In fact, across the twenty-year period, the left and right parties swapped positions on the question of European integration: with the socialists moving from being the least pro-integration party to being the most pro-integration, and the Christian democrats/conservatives moving in the reverse direction. As a result, if the twenty-year period is taken as a whole, the Hix–Lord model of a two-dimensional political space seems to hold, where the left/right and EU integration dimensions are orthogonal, which produces four distinct and enduring positions in the space. Alternatively, one could argue that the Hooghe–Marks or regulation models capture the evolution from a two-dimensional space to a new one-dimensional space: between a “regulatory capitalist” position (on the left and pro-integration) and a neoliberal position (on the right and less pro-integration). From the


111

Hix–Lord perspective, though, this new structure is only temporary, as left/anti-integration and right/pro-integration positions are sure to emerge at some point in the future. Second, the results of the confirmatory factor analyses provide systematic empirical evidence as to the accuracy of four prominent models of the EU political space. Interestingly, arguably the most common model – the international relations model – is inconsistent with the pattern of policy positions adopted by Euro-parties in their election manifestos. In addition, the evidence provides only modest support for the Hix–Lord model and the Hooghe–Marks model. Instead, the Euro-parties appear to structure their policy space more in terms of traditional socio-economic left/right – i.e., the regulation model. The best performing model was one in which the traditional socio-economic left/right determined the positions of the Euro-parties on EU socio-economic policies and on issues relating to further economic integration in Europe. But a related yet distinct dimension structures issues of further political integration, new EU roles in regulating the economy, and EU territorial enlargement. However, given the strong relationship between these two dimensions, for practical purposes we would conclude that the single-dimensional model is a reasonable depiction of the EU and a reasonable assumption for spatial modeling. That single dimension would best be characterized as reflecting a traditional socio-economic left/right dimension. We should emphasize that we consider these as tentative conclusions. While the Euro-party manifesto data are valuable as a source of information about relevant political actors’ positions on EU policy, they have some limitations. First, not all political actors involved in EU policymaking are members of Euro-parties. In particular, those actors (e.g., political parties) with policy positions outside the mainstream of the traditional party families were ignored in our analysis. To the extent that these actors are important to EU policy-making and take considerably different positions on EU policy from those of the Euro-parties, our results will misrepresent the EU policy space. However, as we argued above, most of the key players in EU politics are members of these Euro-parties. Second, and possibly of greater concern, the Euro-parties may have significant internal conflict over EU policies and yet they must write a common manifesto. Such internal dissent could result in manifestos that avoid sensitive issues or ones that are ambivalent. For example, if all Euro-parties suffer from an internal division over further institutional and territorial integration of Europe, then this “integration” dimension of conflict may simply be ignored in the Euro-party manifestos. If this concern over integration is an important determinant of national parties’ behavior in EU policy-making, our analysis would fail to capture this dimension of the EU policy space (Hix 1999a).

112


Put more generally, this problem is important for our results if (a) national parties largely act independently of their Euro-parties when they participate in EU policy-making, and (b) national parties within Europarties are divided on important issues of EU policy. We can only speculate on the likelihood that either of these conditions accurately describes EU politics. As we argued above, Euro-parties appear to influence the behavior of their members and to shape the agenda of EU politics. Also, the Euro-parties are not simply arbitrary groupings of national parties. The Euro-parties consist of parties with similar constituencies and ideological/policy concerns. However, to the extent that national parties pursue policy agendas distinct from their party federations, the results of our analysis may mis-characterize the dimensionality of the EU political space. But lacking an alternative and better source of data about national parties’ positions on EU policies, we hope that the results reported here provide some empirical guidance on how to describe and model EU politics. Appendix: Coding frame 1 ECONOMY (including environment) 11 ECONOMY/+state+increased role of the state (e.g., social justice/ solidarity/redistribution) 111 increased budget (pro public spending/increased taxes/budget deficit) 112 public ownership (increase or defend against privatization) 113 increased state regulation of private sector 1134 trade protectionism 114 direct action (i.e., concern for unemployment) 1144 corporatism 115 efficiency and value for money not a priority 13 ECONOMY/−state−reduced role of the state (e.g., free market/individual freedom) 131 reduced budget (e.g., tax reform) 132 increased privatization 133 decreased state regulation of private sector/deregulation 1334 trade protectionism – opposed (pro free trade) 134 direct action – opposed 1344 corporatism – opposed 135 efficiency, thrift, and value for money a priority 14 ECONOMY/environment vs. growth 141 favorable mention of environment/anti-science and technology 143 favor growth over environmental protection/pro-science and technology 2 POLITICAL SYSTEM 21 POLITICAL SYSTEM/radical 211 promote constitutional reform 213 promote reform of bureaucracy (e.g., openness and transparency)


113

214 positive discussion of democratic decision-making/human rights 23 POLITICAL SYSTEM/conservative 231 oppose constitutional reform 233 oppose reform of bureaucracy 234 negative discussion of democratic decision-making/human rights 3 SOCIETY 31 SOCIETY/values (public action on social norms and values) 311 liberal, permissive, or non-traditional social values 3111 liberal, permissive sexual values (e.g., gay rights, pro-abortion) 3112 liberal discussion of role of the family 3113 liberal attitudes on crime/policing 3114 promote equal treatment of individuals 313 conservative, restrictive, and/or traditional social values 3131 traditional sexual values (e.g., gay rights, pro-abortion) 3132 traditional discussion of role of the family 3133 conservative attitudes on crime/policing 3134 conservative attitudes towards equal treatment of individuals 33 SOCIETY/sectional interests 331 promote interests of 3311 workers and organized labor 3312 employers and employers’ federations 3313 farmers and other rural interests 3314 women and women’s groups 3315 particular religious groups 3316 particular ethnic groups (e.g., immigrant/migrant workers) 3318 consumers 3319 elderly 33110 children and young persons 33111 disabled 333 opposed to interests of 3331 workers and organized labor 3332 employers and employers’ federations 3333 farmers and other rural interests 3334 women and women’s groups 3335 particular religious groups 3336 particular ethnic groups 3338 consumers 3339 elderly 33310 children and young persons 33311 disabled 4 EXTERNAL – role of the state in external relations 41 EXTERNAL/bilateral 411 Pro – support closer relations with 4111 United States 4112 (Former) Soviet Union/Communist bloc 413 Con – oppose closer relations with 4131 United States 4132 (Former) Soviet Union/Communist bloc 42 EXTERNAL/multilateral

114


421 Pro multilateralism – support closer relations with 4211 United Nations 4214 GATT/WTO 4215 NATO 423 Con – oppose closer relations with 4231 United Nations 4234 GATT/WTO 4235 NATO 43 EXTERNAL/security Total “dove” policies (4311, 4321, 4331, 4341) Total “hawk” policies (4313, 4323, 4333, 4343) 431 nuclear weapons 4311 oppose nuclear weapons (pro-disarmament) 4313 support nuclear weapons (anti-disarmament) 432 conventional weapons 4321 oppose conventional weapons (pro-disarmament) 4323 support conventional weapons (anti-disarmament) 433 war/peace 4331 oppose military engagement 4333 support military engagement 434 neutrality 4341 support neutrality 4343 oppose neutrality 44 EXTERNAL/aid (and preferential trade with Second/Third World) 441 Pro – support increases/oppose cuts 4411 to former Communist states 4412 to developing world 443 Con – oppose increases/support cuts 4431 to former Communist states 4432 to developing world 45 EXTERNAL/talks and peace resolution 451 Pro – positive discussion of talks in trouble spots/peace resolution 453 Con – negative discussion of talks in trouble spots/peace resolution 5 GENERAL 51 GENERAL/partisan (negative campaigning against other parties, or pro own party) 52 GENERAL/policy (substantive policy positions not classified elsewhere) 53 GENERAL/personal (personal attacks or personal eulogies) 54 GENERAL/pap (general empty discussion – “the bullshit box”) 6 EUROPEAN UNION/INTEGRATION 61 EU/integration – general 611 Pro – general statement of support for EU/goal of European integration 6111 for economic reasons (e.g., prosperity, social justice, welfare) 6112 for political reasons (e.g., peace, democracy/freedom, security) 613 Con – general statement of opposition to EU/goal of European integration


115

6131 for economic reasons (e.g., prosperity, social justice, welfare) 6132 for political reasons (e.g., peace, democracy/freedom, security) 62 EU/institutions Total pro-institutions (62111, 62211, 62311, 62411, 62511, 62611) Total anti-institutions (62113, 62213, 62313, 62413, 62513, 62613) Total pro-reform (62121, 62131, 62141, 62221, 62231, 62241, 62321, 62331, 62521) Total anti-reform (62123, 62133, 62143, 62223, 62233, 62243, 62323, 62333, 62523) 621 Commission 6211 general 62111 support (more powers) for Commission 62113 opposition/criticism of Commission 6212 reform of Commission administration 62121 support for reform of the Commission 62123 opposition to reform of Commission 6213 Commission President powers 62131 support for more power for Commission President 62133 opposition to more power for Commission President 6214 number of Commissioners 62141 support for fewer Commissioners 62143 opposition to fewer Commissioners 622 Council/European Council 6221 general 62211 support (more powers) for Council 62213 opposition/criticism of Council 6222 more qualified majority voting (QMV) 62221 support for more QMV 62223 opposition to more QMV 6223 change vote weighting (increase votes of larger states) 62231 support for more votes for larger states 62233 opposition to more votes for larger states 6224 reform of Council presidency 62241 support for reform of presidency 62243 opposition to reform of presidency 623 European Parliament (EP) 6231 general 62311 support (more powers) for European Parliament 62313 opposition/criticism of European Parliament 6232 more legislative/budgetary powers of EP 62321 support for more powers for EP vis-à-vis Council 62323 opposition to more powers for EP vis-à-vis Council 6233 more powers of EP over Commission 62331 support for more powers of EP vis-à-vis Commission 62333 opposition to more powers for EP vis-à-vis Commission 624 Court of Justice 6241 general 62411 support (more powers) for Court of Justice 62413 opposition/criticism of Court of Justice

116


625 European Central Bank (ECB) 6251 general 62511 support for ECB 62513 opposition/criticism of ECB 6252 reform of ECB decision-making 62521 support for reform of ECB decision-making 62523 opposition to reform of ECB decision-making 626 Court of Auditors 6261 general 62611 support for Court of Auditors 62613 opposition/criticism of Court of Auditors 63 EU/policies Total anti-EU policies (63113–63, ’213–53, ’313–73, ’413–53, ’513–23, ’613–53, ’73, ’83) Total pro-EU policies (63111–61, ’211–51, ’311–71, ’411–51, ’511–21, ’611–51, ’71, ’81) 631 regulatory policies 6311 general regulatory policies 63111 support for (more) EU regulation of economy/society 63113 opposition to (more) EU regulation of economy/society 6312 single market 63121 support for the single market 63123 opposition to/criticism of the single market 6313 environmental regulation 63131 support for (more) EU environmental regulation 63133 opposition to (more) EU environmental regulation 6314 social regulation/social policy – workers 63141 support for (more) EU social regulation/policy for workers 63143 oppositionto(more)EUsocial regulation/policyfor workers 6315 social regulation/social policy – women 63151 support for (more) EU social regulation/policy – women 63153 opposition to (more) EU social regulation/policy – women 6316 competition policy/state aid 63161 support for EU competition/state aid policies 63163 opposition to EU competition/state aid policies 632 budgetary policies 6321 general budget/increase vs. decrease (budget reform) 63211 support increase in EU budget 63213 oppose increase in EU budget/support decrease 6322 Common Agricultural Policy (CAP) 63221 support CAP/increased CAP budget 63223 oppose CAP/increased CAP budget/support reform 6323 structural funds/cohesion policy 63231 support increase in structural funds/cohesion policy


117

63233 oppose increase in structural funds/cohesion policy 6324 research and development 63241 support increase in EU budget on R&D 63243 oppose increase in EU budget/support decrease on R&D 6325 development/humanitarian aid 63251 support increase in EU budget on development/ humanitarian aid 63253 oppose increase/support decrease on development/ humanitarian aid 633 Economic and Monetary Union (EMU) 6331 general 63311 support ERM/EMU/single currency 63313 oppose ERM/EMU/single currency 6332 independent central bank (vs. political control) 63321 support ECB independent from political control 63323 oppose ECB independence vs. support political influence on EMU 6333 goal of price stability (vs. growth/employment) 63331 support single goal of price stability 63333 oppose single goal of price stability/support of growth goal 6334 convergence criteria 63341 support “economic” convergence criteria 63343 oppose “economic” convergence criteria/support other criteria 6335 stability pact/constraints on national budget deficits 63351 support stability pact/constraints on national budget deficits 63353 oppose stability pact/constraints on national budget deficits 6336 fiscal federalism 63361 support increased EU budget as part of EMU 63363 oppose increased EU budget as part of EMU 6337 macroeconomic policy coordination/tax harmonization/common employment policies 63371 support macroeconomic policy coordination 63373 oppose macroeconomic policy coordination 634 external trade/common commercial policy 6341 EU trade policy 63411 support for EU external trade policy 63413 opposition to/criticism of EU external trade policy 6342 bilateral free trade with USA 63421 support for more free trade with USA 63423 oppose free trade with USA 6343 preferential trade with developing world/development policy 63431 support for more trade policies with developing world

118


63433 oppose/critical of preferential trade with developing world 6344 preferential trade agreements – Eastern Europe 63441 support for preferential trade with Eastern Europe 63443 oppose preferential trade with Eastern Europe 6345 preferential trade agreements – other 63451 support for preferential trade with other states/in general 63453 oppose preferential trade with other states/in general 635 Common Foreign and Security Policy (CFSP)/defense policy 6351 CFSP – general/integration 63511 support for more integration in area of CFSP 63513 oppose more integration in area of CFSP 6352 independent EU defense capability 63521 support independent EU defense capability 63523 oppose independent EU defense capability/support NATO 636 Justice and Home Affairs (JHA) 6361 JHA, citizenship – general/integration 63611 support for more integration in area of JHA/EU citizenship 63613 oppose more integration in area of JHA 6362 removal of internal borders/free movement of persons 63621 support for removal of internal borders/free movement 63623 oppose removal of internal borders/free movement 6363 common immigration/asylum policies 63631 support for common immigration/asylum policies 63633 oppose common immigration/asylum policies 6364 common policing/policies to fight drug-trafficking, terrorism, crime 63641 support for more common policing 63643 oppose more common policing 6365 judicial cooperation 63651 support for more judicial cooperation 63653 oppose more judicial cooperation 637 other EU policies 6371 support for any other existing EU policy 6373 oppose/critical statement on any other existing policy 638 new EU policy competences 6381 support for a new EU policy competence 6383 oppose a new EU policy competence in a particular area 64 EU/issues 641 institutional reform – general (i.e., IGCs) 6411 general statement of support for institutional reform/a new IGC 6413 opposition to further institutional reform/a new IGC 642 transparency/accountability 6421 support for more accountability/transparency in the EU 6423 opposition to more accountability/transparency in the EU


119

643 democratic deficit – reduction 6431 support for efforts to reduce the democratic deficit 6433 opposition to reduction of the democratic deficit 644 national parliamentary control of the EU 6441 support increased role for national parliaments in EU policymaking 6443 oppose increased role for national parliaments in EU policymaking 645 subsidiarity 6451 support for more subsidiarity/codification of subsidiarity 6453 opposition to subsidiarity/codification of subsidiarity 646 flexible integration/opt-outs 6461 support for flexible integration/opt-outs 6463 opposition to flexible integration/opt-outs 647 interest representation 6471 business interests 64711 support more say for business interests in EU 64713 oppose more say for business interests in EU 6472 trade unions (or other diffuse interests – e.g., women, migrants) 64721 support more say for trade unions in EU 64723 oppose role of trade unions in EU 6473 environmental groups 64731 support more say for environmental groups in EU 64733 oppose role of environmental groups in EU 6474 consumers 64741 support more say for consumer groups in EU 64743 oppose role of consumer groups in EU 6475 regions 64751 support more say for regions in EU policy process 64753 oppose role of regions in EU policy process 648 European elections – reform 6481 support reform of European elections/reform (e.g., uniform procedure) 6483 opposition to reform of European elections 649 Pan-European parties 6491 support for pan-European political parties 6493 opposition to pan-European political parties 6410 enlargement 64101 support for EU enlargement 64103 oppose/delay EU enlargement

6

Does left/right structure party positions on European integration?∗ Liesbet Hooghe, Gary Marks, and Carole J. Wilson

How is contestation on European integration structured among political parties competing in the member states? How is it related, if at all, to the political conflicts that have shaped political life in Western Europe? The framework within which we pose these questions is the standard model of European party system dynamics consisting of the following elements: r Contestation among political parties is limited to one or two dimensions. This renders competition among parties institutionally and intellectually tractable. r These dimensions are, first, a left/right dimension tapping greater vs. lesser government regulation of market outcomes and, in many party systems, a related new politics dimension tapping communal, environmental, and cultural issues. The general question we ask in this chapter is whether issues arising from European integration are assimilated into these existing dimensions of domestic contestation. Can the positions that political parties take on European issues be read from their positions on the left/right and new politics dimensions? Or are these European issues unrelated – orthogonal – to these dimensions? Does European integration put a new and potentially disruptive set of issues on the agenda that cannot be swallowed within existing patterns of political contestation? If these issues are assimilated, how are they assimilated? What, in other words, are the substantive connections between party positioning on European integration and party positioning on the dimensions that structure domestic politics? ∗

We wish to thank the participants in workshops held at the University of North Carolina at Chapel Hill on “Citizenship, Parties and Elections in the European Union,” in October 1998, August 1999, and April/May 2000; participants in a seminar at the Department of Government and Foreign Affairs, University of Virginia, at which this paper was presented in September 2000; Paul Taggart for organizing panels on Euroskepticism at the European Community Studies Association meeting, Madison, May–June 2001, and Mark Aspinwall, Jim Caporaso, Leonard Ray, David Scott, and Marco Steenbergen for helpful comments. We are grateful to Susan Glover and Heather Mbaye for research assistance.

120

Party positions on European integration

121

One must, we believe, disaggregate European integration into its particular policies (e.g., environmental, cohesion, and fiscal policy) in order to answer these questions accurately. Using expert data on 125 parties in fourteen countries for a range of policy areas, we are able to do this. We show that there is a strong relationship between the left/right dimension that chiefly structures party competition in European societies and European integration. We do this by taking a fresh look at the inverted U-curve describing pro-integration centrist parties and antiintegration peripheral parties, and by specifying the residual linear association between left/right position and position on European integration issues. Second, we investigate the influence of a second dimension, a new politics dimension that we conceive as ranging from GAL (green/alternative/ libertarian) to TAN (traditional/authoritarian/nationalist). We find that this dimension is the most general and powerful predictor of party positioning on the issues that arise from European integration.

Data The analysis undertaken in this chapter is based on a new data set gathered under the auspices of the UNC-Chapel Hill Center for European Studies. An expert survey conducted in 1999 by Gary Marks, Marco Steenbergen, David Scott, and Carole Wilson asked country experts to evaluate the positions of national political parties on European integration on a seven-point scale (ranging from 1 to 7) with the lowest score representing strong opposition to European integration and the highest score indicating strong support for European integration. One hundred and twenty-three country experts reporting on their country of expertise evaluated the positions of a total of 125 political parties in all EU member states except Luxembourg. The survey replicates Leonard Ray’s 1997 expert survey of party positioning on European integration at four different time points: 1984, 1988, 1992, and 1996 (Ray 1999). The 1999 survey offers two additional kinds of data that allow us to simultaneously widen and sharpen our view of party positioning. First are a set of questions that tap the degree of support across parties for European integration in seven policy areas: EU environmental policy, EU cohesion policy, EU asylum policy, EU employment policy, EU fiscal policy, EU foreign policy, and expanding the European Parliament’s

122

Liesbet Hooghe, Gary Marks, and Carole J. Wilson

power.1 Second are items that tap party positions on basic dimensions of political contestation, including an economic left/right scale and a new politics scale.2 Our estimate of each party’s position on each of these questions is the mean of the country experts’ evaluations. On average, there are nine experts for each country. Statistical tests indicate that these data are reliable within conventional limits and are consistent with alternative sources of available data (Marks and Steenbergen forthcoming). The left/right dimension Whether and how the basic left/right dimension of contestation structures party positioning on European integration is the central topic of this volume. What is the relationship between a party’s left/right placement and its position on issues arising from European integration? The inverted U-curve Let us begin with some uncontested facts. Extreme left parties and extreme right parties share Euroskepticism; parties in the middle, including most social democratic, Christian democratic, liberal, and conservative 1

For example, the item for cohesion policy is as follows: Now consider the issue of structural or cohesion policy. This policy transfers resources to the poorest regions of the EU and is the second largest item in the EU’s budget. Some political parties wish to rein in or eliminate the EU’s cohesion policy, whereas others wish to sustain or expand it. What about the leadership of the parties listed below?

2

Both scales range from 0 (extreme left) to 10 (extreme right). The wording of the questions for the economic left/right and new politics dimensions is as follows: Political scientists often classify parties in terms of the ideological stance they take on economic issues. Parties to the right on economic issues tend to emphasize a reduction of the economic role of the government; they want lower taxes, less regulation, privatization, reduced government spending, and a leaner welfare state that imposes fewer burdens on employers. Parties to the left on economic issues want the government to retain an active role in the economy. Using these criteria, please indicate where the parties are located in terms of their economic ideology. Another way parties are sometimes classified is in terms of their views of democratic freedoms and rights. “Libertarian” or “post-materialist” parties tend to favor expanded personal freedoms and rights. Such parties, for example, support abortion, doctor-assisted suicide, same-sex marriages. They favor increased democratic participation and freedom of speech. At the same time, they oppose discrimination on ethnic, religious, political or sexual grounds. In sum, these parties want government to stay out of the life choices that people make and they promote widespread democracy. “Traditional” or “authoritarian” parties often reject these ideas. These parties believe that the government should be a firm authority that expresses moral voice. To these parties, order is preferable to unbridled participation and freedom. On the scale below, please indicate where parties are located in terms of their ideological views of freedoms and rights.


123

7

6

General support for European integration

Party family Rad right 5 Conserv 4

Liberal Christdem

3 Socialist Rad left

2

Green 1 1984

1988

1992

1996

1999

Year Figure 6.1 Support for European integration by party family since 1984. (Note: This excludes Protestant, agrarian, regionalist, and non-affiliated parties. Only parties in EU member states that obtained votes in the national elections prior to the time of evaluation are included. Parties number fifty-seven in 1984, seventy-four in 1988, seventy in 1992, ninetyone in 1996, and ninety-three in 1999. Scores vary between 1 (strongly opposed to European integration) and 7 (strongly in favor of European integration).)

parties, are generally much more supportive of European integration (figure 6.1). When one charts the positions of party families combining a left/right dimension with an orthogonal dimension indicating level of support for European integration, the result is an inverted U. Figure 6.2 displays just such a curve describing left/right placement and support for European integration and three selected policies for all political parties in the EU-14 in 1999. How can one explain the U-curve? One explanation engages the strategic responses of parties and draws its inspiration from William Riker’s work (1982, 1986). “Within the constraints imposed upon them . . . each party attempts to strategically manipulate the European integration

124

Liesbet Hooghe, Gary Marks, and Carole J. Wilson 7

6

5

4 European integration 3 EU employment EU cohesion policy

2

Support EU fiscal policy 1 1

2

3

4

5

6

7

8

9

Left/Right

Figure 6.2 Positioning on selected EU policies by left/right dimension. (Note: All parties in 1999 (N = 125).)

issue . . . to meet its goals” (Scott 2001: 6). Parties that are successful in the existing structure of contestation have little incentive to rock the boat, while unsuccessful parties, i.e., parties with weak electoral support or those that are locked out of government, have an interest in restructuring contestation.3 The same strategic logic that leads mainstream parties to assimilate the issues raised by European integration into the left/right dimension of party competition leads peripheral parties to exploit European integration in an effort to shake up the party system (Hix 1999a, 1999b). In Paul Taggart’s words, European integration is a “touchstone of domestic dissent” for peripheral parties (1998: 384). A second explanation emphasizes ideology, and the ways in which extremism on the left/right dimension leads to extremism on new issues arising on the agenda. As Ernst Haas made clear, the European Union is the product of party-political actors on the center-right, center, and, 3

Geoffrey Evans raises an interesting counterexample: the decision on the part of the British Conservative Party to campaign against further European integration, despite the fact that this issue was clearly orthogonal to the main left/right dimension (Evans 1999). To date, the Conservative Party has failed to exploit this electorally.


125

Table 6.1 Multiple regression analysis for party positioning on European integration Model

Standardized coefficients (β)

t-value

Left/right extremism Government participation Electoral vote Left/right position GAL/TAN position R2

−0.41 0.25 0.10 0.23 −0.48 0.52

−5.73 3.08 1.32 2.58 −5.74

Note: N = 125.

to a lesser extent, the center-left who have dominated decision-making in Europe over the past half-century (Haas 1958). European integration is primarily a market-liberal project mitigated by some measure of regulated capitalism. The Euroskepticism of extreme parties arises, therefore, not only from their opposition to the EU’s policies, but also because they reject the ideology of the EU’s construction. So we expect a party’s support for European integration to decline with its distance from the center of the left/right dimension. Both explanations find confirmation in our data. We measure the stake that political parties have in existing dimensions of contestation with the variables electoral vote, which is the proportion of the electoral vote captured by a party in 1999 or at the previous national election, and government participation, a dummy variable with a value of 1 if the party participated in government in the period 1965–95. We measure left/right extremism by squaring the distance between a party’s left/right location and that of the median party. We control for left/right position and GAL/TAN position. Table 6.1 reveals that government participation is a strong and highly significant influence on party positioning on European integration. Electoral vote is insignificant in the presence of government participation, but becomes highly significant when government participation is dropped from the analysis. These two variables are quite highly correlated (0.58). Exclusion from government leads to Euroskepticism, and excluded parties tend to be those that are electorally weak. However, weak electoral performance and exclusion from government only partially explain why parties oppose European integration. The effect of left/right extremism survives the controls we apply in the model represented in table 6.1. The ideological positioning of parties towards the extremes of the left/right

126


dimension exerts a powerful influence on EU positioning independently of electoral performance or government participation. A linear relationship? Beyond the non-linear relationship captured by left/right extremism, is there a linear association between a party’s left/right location and its support for policies connected with European integration? There are three distinct lines of response (Steenbergen and Marks in this volume; Gabel and Hix in this volume; Hix 1999a, 1999b; Hooghe and Marks 1999: 88; Kreppel and Tsebelis 1999; Tsebelis and Garrett 2000):4 r The regulation model. European integration is subsumed into the left/right dimension. European politics is fused to the basic domestic competition between the left, which pushes for common economic regulation across Europe, and the right, which favors less regulation. Party positioning on left/right and European issues coincide. r The Hix–Lord model. European integration and left/right contestation are independent of each other. European integration engages national sovereignty and mobilizes territorial groups, which compete on where authority should be located. Left/right contestation involves the allocation of values among functional interests. Hence, party positioning on domestic and European issues are orthogonal to each other. r The Hooghe–Marks model. Left/right contestation shapes positioning only on European policies that are concerned with redistribution and regulating capitalism. Hence, the center-left supports European integration in cohesion policy, social policy, unemployment policy, environmental regulation, and upgrading the European Parliament, while the right supports market integration but opposes European re-regulation. Left/right location is related to a subset of European issues. By and large, our data confirm the third model. Table 6.2 presents results of multiple OLS regressions for positioning on general European integration and on six EU policies. On general European integration, the more right-wing a party is, the more it favors European integration. The regulation model incorrectly predicts the sign of the coefficient for left/right position. When we control for GAL /TAN and left/right extremism, there is no consistent linear relationship between left/right and various components of European integration. This result is consistent with the Hix–Lord model and with the Hooghe–Marks model, which distinguishes among the kinds of policy positions that are constrained by left/right positioning. 4

The international relations model, which posits no connection between European integration and the left/right dimension, is clearly invalid for our purposes.


127

Table 6.2 Multiple regression analysis for party positioning on European integration and EU policies Standardized coefficients (β) (t-value)

Parties’ positions on General EU integration European Parliament powers EU asylum policy EU environment policy EU employment policy EU cohesion policy EU fiscal policy

Notes: N = 125,

∗p

Left/right extremism

Left/right position

GAL/TAN position

−0.50 (−6.74) −0.38 (−4.86) −0.37 (−4.95) −0.42 (−6.49) −0.48 (−6.37) −0.47 (−6.56) −0.31 (−3.57)

0.28 (3.02) 0.15 (1.47) −0.15 (−1.56) −0.24 (−2.98) −0.29 (−3.01) −0.35 (−3.88) 0.12 (1.08)

−0.53 (−6.00) −0.61 (−6.55) −0.54 (−6.08) −0.58 (−7.47) −0.38 (−4.21) −0.40 (−4.63) −0.43 (−4.13)

R2 0.43∗ 0.37∗ 0.43∗ 0.57∗ 0.41∗ 0.47∗ 0.21∗

< 0.01.

When one examines policies to achieve European regulated capitalism, including environmental policy, employment policy, and cohesion policy, the effect of left/right position is quite strong and highly significant. Consistent with this, the coefficient for left/right position is insignificant, and the sign unstable, for policies that are distant from egalitarian and regulatory concerns, such as EU asylum policy and the power of the European Parliament. There appears to be an explicable pattern of orthogonality and association between left/right position and positions on the various elements of European integration.5 5

Fiscal policy is an exception. Fiscal policy is at the core of the struggle between neoliberals and proponents of European regulated capitalism. Yet party positions on this policy are not significantly explained by left/right location. The reason for this is that economic left and right both want fiscal coordination, but for very different reasons. Proponents of regulated capitalism wish to shift authority to the European Union to redress the current imbalance between fiscal and monetary policy under EMU. EU fiscal capacity should, they believe, be strengthened to cope with asymmetrical shocks. Neoliberals, on the other hand, wish to maintain the mismatch between fiscal policy capacity and monetary policy. With monetary policy securely hived off to an independent central bank, national governments are induced to compete for investment by reducing the overall tax burden and shifting its incidence from mobile capital to less mobile factors of production. Greater EU fiscal coordination could accelerate this market-driven process by subjecting tax

128


The linear pull of left/right location on parties’ positions concerning European regulated capitalism is sharper when we examine the subset of mainstream parties. Our initial attempt to formulate the third model focused exclusively on the moderate left, the center, and the moderate right: that is, the social democratic, liberal, Christian democratic, and conservative party families, which dominate national governments and represent 80 percent of the electoral vote across the EU (Hooghe and Marks 1999). There is no inverted U-curve here, but a strong and highly significant downward sloping line from a pro-integrationist left to a less integrationist right for cohesion, employment, and environment policy. This finding is robust across controls for the variables listed in table 6.2. Discussion The Euroskepticism of the extreme left is decisive in creating the inverted U-curve, while mainstream parties underpin the linear association between left/right position and support for European integration issues. We discuss these in turn. Radical left parties are highly Euroskeptical and this accounts for the inverted U-shape describing all parties on the left/right dimension Opposition to European integration is deeply entrenched among the radical left. This opposition is not merely anti-system. It is rooted in the perception that European integration fundamentally threatens cherished radical left goals and that the European Union has been co-opted by mobile capital. For example, when the Swedish social democratic government applied for EU membership, its competitor to the left – Venstre – responded that “the message in the Maastricht Treaty was the construction of a capitalist block” (Christensen 1996: 534). In the same vein, the Greek communist KKE has virulently opposed membership in a European Union of “monopolists” (Hermet et al. 1998: 245). incentives for investors to EU competition authorities and by imposing budget discipline, e.g., through the Growth and Stability Pact. Jan Beyers and Bart Kerremans summarize the position of Belgian liberal parties on EU fiscal policy: The Francophone liberals expressed the hope that EMU would spill over into a European fiscal policy. But “fiscal” did not refer to a larger EU budget. It rather referred to the harmonization of national fiscal policies so as to avoid fiscal competition. The [harmonized] level, however, would have to be lower than the current level of taxation in Belgium. So liberals would like to reconcile two objectives: 1) avoiding that competition would erode the budgetary basis of the Belgian social security system, and 2) avoiding that a European fiscal policy would open the door for a European Keynesian state or for ever higher taxes (Beyers and Kerremans 2001: 23).


129

Underlying this reaction is the hard fact that the European Union is fundamentally inhospitable to radical left policy goals. Fritz Scharpf (1996, 1999) has made the point that there is an in-built institutional asymmetry in the European treaties favoring market-deepening over market-correcting policies. Social democrats expect that this asymmetry can be countered to achieve moderate reform. For the radical left, the European Union is biased beyond repair. Even EU institutions that facilitate regulated capitalism will not provide the kind of policies radical leftists deem essential to curb market forces: public control over capital flows, extensive public investment in industrial policy, a statutory right to work, and a statutory reduction of the working week. And so the radical left comes to the same conclusion as the radical right, but for very different reasons. Figure 6.1 illustrates that, in contrast to the social democrats, the radical left has persisted in its opposition to European integration. This opposition is primarily motivated by ideological differences in economic left/right placement. Among the subset of radical left parties in 1999, the association between left/right position and position on European integration is powerful and highly significant (R = 0.74). The association remains strong when we expand the subset to include all parties on the economic left, that is, green parties and social democrats (R = 0.68). The more left-wing parties are, the more Euroskeptical. Social democratic parties have become distinctly more pro-integration as regulated capitalism has come on the European agenda. Parties of the right oppose regulating capitalism at the national and European levels, and this leads them to selectively oppose European integration European integration is double-edged for center-left parties (Hix 1995a, 1995b; Hooghe and Marks 1996; Hix and Lord 1997; Hix 1999a; Ladrech 1997; Ladrech and Marlière 1999; Marks and Wilson 2000; Schmitt and Thomassen 1999). On the one hand, it threatens left achievements at the national level because it intensifies international economic competition while undermining Keynesian responses to it. On the other hand, deeper integration may enhance the possibilities for social democracy by creating democratic, authoritative institutions capable of pursuing employment, environmental, or cohesion policy at the European level. As a Flemish socialist exclaimed during a parliamentary debate on Belgian participation in EMU in 1996: “Why do you think that the German labor unions hope that the third stage of EMU will succeed? . . . They know that EMU will create the foundations for a Rhineland model on a European scale, for a project that will meet the needs of all Belgians and Europeans. That will be a model that will preserve our social welfare in a globalizing economy” (quoted in Beyers and Kerremans 2001: 144).

130


Social democratic parties are not monolithically in favor of deeper integration. Factions in some parties, particularly in Sweden, Denmark, and Germany, remain doubtful about the potential for European regulated capitalism, and argue that while European legislation may ratchet up social democracy in poorer countries, it stands in the way of higher standards in the social democratic heartland of Europe. But this is a minority view. Majorities in one party after another have come to perceive European integration as a means for projecting social democratic goals in a liberalizing world economy (Hooghe and Marks 1999; Ladrech 1997; Katz and Wessels 1999). Figure 6.1 shows that social democratic parties have shifted in favor of European integration over the past fifteen years. In 1984, the largest pool of Euroskepticism – measured according to electoral strength – was social democratic. Four social democratic parties were rated below 5 on our scale of general support for European integration – the Greek PASOK (4.0), the Irish Labour Party (4.0), the British Labour Party (4.5) and the Danish social democrats (4.7). In 1999, no social democratic party had a score below 5, and the average position of all social democratic parties had increased to 6.3, from 5.5 in 1984. Similarly, the left/right dimension structures the positions of right-ofcenter parties to European integration. They support market integration – which means that they support European integration in general terms – but they oppose policies, particularly concerning the environment, cohesion, or employment, that regulate capitalism. When one restricts the analysis to political parties in the four mainstream party families, left/right position is far more powerful than left/right extremism and GAL/TAN position in explaining party support for EU cohesion policy (β for left/right position = −0.30, p = 0.075) and EU employment policy (β = −0.72, p < 0.000). On employment policy, left/right position alone explains 55 percent of the variance.6 Social democratic parties form the bedrock of the coalition in favor of EU employment policy. Support tapers off sharply as one moves to the right. The strongest opponents tend to be parties toward the right on the left/right scale: that is, neoliberal parties. The new politics dimension The past two decades have seen the rise of issues concerned with lifestyle, ecology, cultural diversity, nationalism, and immigration. This dimension 6

When left/right extremism and GAL/TAN position are introduced into the equation, the variance explained is 57 percent. But neither of these variables is significant in explaining mainstream party positions on EU employment policy or cohesion policy.


131

of contestation has been labeled post-materialist/materialist (Inglehart 1990), new politics/old politics (Franklin 1992; Muller-Rommel ¨ 1989), green/traditionalist, and left-libertarian/right-authoritarian (Kitschelt 1994, 1995). One pole combines ecology (or greenness), alternative politics (including participatory democracy), and libertarianism. We summarize this as the green/alternative/libertarian or GAL pole. The opposite pole combines support for traditional values, opposition to immigration, and defense of the national community. We summarize this as the traditional/authoritarian/nationalism or TAN pole. The argument Does a party’s location on this dimension tell us where the party stands on European integration? Our expectation, as before, is that the greater the proximity of a new issue to a political party’s established programmatic concerns, the more that issue will be assimilated into the party’s ideology (Marks and Wilson 2000; Marks, Wilson, and Ray 2001; KLMS 1999). To what extent, then, is the new politics dimension relevant for interpreting European issues? It takes little imagination to see that certain substantive issues, including EU asylum policy and environmental policy, are closely tied to existing new politics concerns. But the connection is deeper and more general. At its core is national sovereignty. New right parties react against a series of perceived threats to the national community. The threats are many: immigrants, foreign cultural influences, cosmopolitan elites, international agencies. European integration combines several of these threats, and poses one more: it undermines national sovereignty. As scholars of European integration have long argued, one of the chief consequences of European integration is to weaken the authority of national states (Borzel ¨ and Risse 2000; Caporaso 1996, 2000; Hooghe and Marks 2001; Jachtenfuchs and Kohler-Koch 1996; Marks, Hooghe, and Blank 1996; Risse 1996; Schmitter 1996; StoneSweet and Sandholtz 1997). This is a core idea of multilevel governance, and it is anathema to those on the radical right, which is why parties towards the TAN pole strenuously oppose European integration. Parties towards the GAL pole are not so deeply motivated. They are driven by specific policy goals rather than profound antipathies. On the one hand, European integration offers a prospect of more encompassing environmental regulation and enhanced locational choice for individuals. On the other hand, it threatens to intensify regulatory competition, weaken democracy, and disempower public interest groups. The new data we have at our disposal reveal how powerful the connection between new politics and European integration is. The standardized

132

Liesbet Hooghe, Gary Marks, and Carole J. Wilson 7

6

5

4 European integration 3 EU environment

Support

EP powers

2

EU asylum policy 1 1

2

3

4

5

6

7

8

9

10

New politics (GAL/ TAN)

Figure 6.3 Positioning on selected EU policies by new politics dimension. (Note: All parties in 1999 (N = 125).)

coefficients for the GAL/TAN dimension in table 6.2 are uniformly large and highly significant. GAL/TAN exerts a strong, consistent, and, it must be said, largely overlooked effect on party positioning on European issues. Figures 6.1 and 6.3 begin to tell the story underlying these associations. Parties on the radical right have become by far the most Euroskeptic of any party family. In 1988 their average level of support fell below that of the radical left, and the gap widened in the late 1990s. Support for European integration falls off a cliff at the right-hand side of figure 6.3 for parties that are positioned at 7 or more on our new politics scale. If one isolates the right-hand side of figure 6.3 (i.e., taking parties that score 5 or more on the new politics scale), the association between support for European integration and new politics is −0.63. With the exception of environmental policy (R = −0.34), the association between new politics and European integration is weak on the left side of figure 6.3. Clearly, the TAN side of the new politics dimension drives the overall relationship.


133

Discussion The relationship between the new politics dimension and support for European integration combines the following three subplots: Parties near the TAN pole, i.e., radical right and right-populist parties, are, without exception, highly Euroskeptical and their relative electoral weight within their national party systems has grown considerably over the past two decades The radical right and populist-right is the most Euroskeptical party family. Their average score in 1999 on our seven-point scale is 1.9 with 4.0 being neutral. Common to all is a rejection of further political integration. They champion national sovereignty, and this leads them to favor at most an intergovernmental EU with retention of the national veto. The Flemish Bloc wants the European Union to be “a confederate body for cooperation in economic matters, the fight against crime, defence, foreign policy, and other matters of common interest,” and it “rejects a European superstate that is too centralist and too bureaucratic” (http://www.vlaams-blok.be, accessed on August 8, 2001). In the slogan of the French National Front, the radical right supports a “Europe des patries.” Most radical right parties reject the supremacy of EU law over national law, and they oppose EU parliamentary powers. The National Front wants to “restore the supremacy of French law over European law so that no European text can be imposed on the French without previously [being] voted as a national law” (http://www.frontnational.com/programme, accessed August 8, 2001). The Danish People’s Party “would like decisions in the EU to be made by the Council of Ministers in order that the European Parliament’s own powers are reduced. The Danish People’s Party would preferably like to see the European Parliament closed down” (program of the Danish People’s Party, quoted in Benedetto 2001). Euroskepticism sometimes encompasses economic integration as well. While the Danish People’s Party, the Austrian Freedom Party, and the Flemish Bloc support free trade and the single market in Europe, others such as the French National Front and the Italian National Alliance have reservations. In the language of the National Front: The Europe of Brussels and Maastricht is a machine that crushes nations and people; it generates unemployment, a fixation with fiscal discipline, bureaucracy, and recession. Power is in the clutches of a handful of anonymous and irresponsible senior officials, who impose the homogenization of our legislation, put downwards pressure on our social systems, open borders to low-price imports and world-wide immigration, and force our integration in a new economic and

134


political world order wholly dominated by the United States (http://www.frontnational.com/programme, accessed on August 8, 2001).

During the 2002 French presidential elections, Jean-Marie Le Pen announced that, if he were elected, France would negotiate withdrawal from the European Union (program of the FN, http://195.246.155.17/, accessed on May 15, 2002). The views of the radical right on Europe are an extension of their basic orientations. In the words of Le Pen, “My European program is an exact extrapolation of the national program of the Front National” (quoted in Shields 1995). Since the 1980s, the Austrian Freedom Party’s electoral campaigns have consistently forged connections between the threat ¨ to national identity, over-foreignization (Uberfremdung) and immigrant criminality, political and social corruption in Austria, and the arrogance of the European Union (Luther 2001). Nationalism, anti-immigration, and traditionalism go hand in hand. European integration has mobilized the nationalist proclivities of the Austrian Freedom Party, the National Front, and new right parties generally. These parties claim to speak for national pride and the silent majority of the nation – in the electorally successful claim of the Austrian Freedom Party, “Wir Sagen Was Ihr Denkt” (we say what you think). The commitment of these parties to nationalism is prior to their commitment to authoritarianism. Radical right and right-populist parties are proud to call themselves “national” parties, whereas they reject the epithet “authoritarian.” The Euroskepticism of these parties is strongly linked to their opposition to immigration. They see themselves as defending national community and culture against foreigners (Betz 1994; Betz and Immerfall 1998; Hainsworth 1992; Karapin 1998; Kitschelt 1995). In the June 1999 parliamentary elections, the Flemish Bloc’s campaign theme was “Baas in eigen land” (in charge of our own country) – an update of their earlier slogan “Eigen Volk Eerst” (own people first). In the Netherlands, the radicalright List Pim Fortuyn charged that “Nederland is vol” (the Netherlands is full). Anti-immigrant sentiment leads these parties to oppose free movement of persons in the single market, a concern that has intensified with prospective enlargement to the east. Jorg ¨ Haider, leader of the Austrian Freedom Party, has opposed enlargement to the countries on Austria’s eastern border: “From the moment we open our borders, 200,000 people will come here, settle, and look for jobs.”7 Others, like the National Front, tie cooperation with former communist countries to a strict immigration policy. The National Front rejects the Schengen rules, and wants 7

Economist, July 11, 1998, p. 55.


135

to reestablish full control over France’s borders to keep out international crime, immigration, and terrorism. This is also the line of the two rightpopulist breakaway Gaullist parties – former RPR heavyweight Charles Pasqua’s Rassemblement pour la France (RPF),8 and Philippe de Villiers’ Mouvement pour la France (MPF), a coalition of Euroskeptics from the RPR and the UDF (Flood 1997; Hermet et al. 1998; Messina 2001). The MPF proposes to abolish Schengen, reestablish French border controls, and repatriate non-European immigrants: because a “multicultural society gravely menaces national identity . . . non-European foreigners in France should choose between assimilation or gradual repatriation” (quoted in Benedetto 2001). The Flemish Bloc, too, demands stricter European border controls to prevent immigration from outside Europe (especially North Africa and Turkey), and it wants to send back nonEuropean immigrants (http://www.vlaams-blok.be; Messina 2001). Their suspicion of European integration is also rooted in traditional values. The National Alliance, for example, criticized the EU Charter of Fundamental Rights approved at the Nice intergovernmental conference in December 2000 for emphasizing rights rather than duties, and for implicitly supporting “alternative family patterns” that “threaten the natural family, composed of man and woman, and this should be safeguarded in its integrity.”9 Radical right and right-populist parties make up the largest reservoir of opposition to European integration. They anchor the association between GAL/TAN position and European integration. But GAL/TAN location does more than predict the position of parties on the extreme right. It also predicts the positioning of conservative parties on European issues, as we describe next. Conservative parties with a TAN inclination tend to be Euroskeptical. The new politics dimension efficiently distinguishes between anti- and prointegration mainstream party families Unlike radical right parties, which are defined by their extreme location on the new politics dimension, conservative parties are distinguished mainly by their stance on the left/right dimension. However, several conservative parties, including the British Conservative Party, the Portuguese Popular Party, the Irish Fianna Fail, the French Gaullists (RPR), and, more recently, Forza Italia, have a TAN orientation. The RPR and Forza Italia have, from their founding, been expressly national parties. The Conservative Party, building on its Tory 8 9

Not coincidentally, RPF is the name of the party created around Charles de Gaulle in the first days of the Fourth Republic. “Documento della Delegazione di Alleanza Nazionale al Parlamento Europeo relativo alla Carta dei diritti fondamentali, alla riforma istituzionale ed all’allargamento dell’Unione Europea” (quoted in Benedetto 2001).

136


heritage, has always emphasized the unity of the British nation against peripheral nationalism, disestablished churches, and, during the twentieth century, class conflict analyses (Flood 1997; Baker et al 1997; Sowemimo 1996). While they are not so extreme as radical right and right-populist parties, these conservative parties defend national culture, the national community, and national sovereignty against the influx of immigrants, against competing sources of identity within the state, and against external pressures from other countries and international organizations. The French conservative right has consistently condemned European Union immigration and asylum policy as being too lax (Flood 1997). Forza Italia explicitly rejects “a universal, multi-racial society that is rooted in the markets” in favor of a “Christian” model of society based on the “primacy of the nation understood in the romantic sense, as a nucleus and base of values, religion, culture, language, dress and tradition” (quoted in the Financial Times, April 1, 2000). Philippe Séguin, a leading voice of the French RPR, put it this way: “A European consciousness does exist, and it can serve as a basis for a broad community of peoples. What does not exist is a single, homogeneous European people with some collective will to found a vast multinational state. We do have a European consciousness, but we do not have a European national feeling. So much the better” (Séguin 1993, quoted in Flood 1997: 16). The national orientation of these parties has an unambiguous bottom line for their position on European integration: the national state should defend its legitimate sovereign right to govern persons living in its territory. Euroskeptical voices in conservative parties rarely seek withdrawal from the European Union, but they argue for a looser confederation that includes as much of Europe as possible, “from the Atlantic to the Urals” in Charles de Gaulle’s words. Conservative parties in Britain and France are riven by a struggle between nationalists and neoliberals about the future of the European Union.10 Nationalists oppose any dilution of national sovereignty, but neoliberals are prepared to limit national control if necessary for economic integration. This clash has dominated the internal politics of the British Conservative Party since the Maastricht Treaty (Baker et al. 1997, 10

Other conservative parties have been less nationalist. Scandinavian conservative parties have defined themselves mainly in left/right terms, in opposition to social democracy, rather than as national parties (Ljunggren 1988). These parties were unable to develop a strong national base in the countryside, with the result that the left, not the right, has been most successful in appropriating national symbols. Conservatism in Spain and Greece has striven to distance itself from its pre-democratic tradition, which combined reaction and nationalism.


137

1999; George 1998; Whiteley et al. 1994; Whiteley 1998). After the 1997 national election, anti-Europeans gained the initiative, but proEuropeans nipped at their heels. One result was to alienate the party from its traditional constituency – affluent, educated, middle-class voters – whose pragmatic pro-European attitudes fit uncomfortably with the party’s principled Euroskepticism (Evans 1998, 1999). Yet, in the wake of the Conservatives’ decisive defeat in the June 2001 national elections, a pro-European neoliberal candidate lost to a Euroskeptic, nationalist candidate in the party’s leadership contest. Similar disagreements in the Gaullist RPR propelled two staunchly anti-Europeanist factions to break away in the early 1990s (Flood 1997). According to Peter Mair, these cases “typif[y] a more general phenomenon, whereby Europe forces fissures inside parties which may then eventually be released in further fractionalization within national parliaments and national electoral arenas” (Mair 2000: 36). Given the endemic conflict between neoliberalism, oriented on the left/right dimension, and nationalism, oriented on the GAL/TAN dimension, we hypothesize that conservative parties are particularly prone to such fissures.11 One must refer to the GAL/TAN dimension to explain how conservative parties position themselves on issues arising from European integration. Our data reveal that the EU positioning of mainstream parties generally is strongly associated with the location of these parties on the GAL/TAN dimension. For 1999, the correlation between mainstream parties’ positions on the GAL/TAN dimension and European integration is −0.45. As one would expect, the relationship is yet stronger when one examines policy areas where integration involves loss of national authority, such as strengthening the European Parliament and developing a European asylum policy. The associations remain very strong and significant when one controls for the left/right dimension of contestation. Hence, the new politics dimension not only structures orientations to European integration among parties on the TAN side; it also structures party positioning in the major party families. Green parties, located towards the GAL pole, have become more integrationist. This strengthens the association between the new politics dimension and support for European integration Opposition to European integration among TAN-leaning parties is not reciprocated by enthusiastic support among GAL parties. The mainstay of the GAL side, green parties, are equivocal about European integration. Green wariness about 11

Conflict within political parties is a vital element of party response to European integration which we cannot pursue in the scope of this chapter.

138


the lack of democratic transparency in the EU impedes general support (4.6 on a seven-point scale). As Elizabeth Bomberg observes: “Greens in Europe . . . face a strategic paradox: the incentives to work through the EU are great, yet how can they work through institutions that inherently violate green principles?” (Bomberg 1998: 4; Rudig ¨ 1996: 268). However, it is the green parties that have become the strongest supporters of European environmental policy (6.6 on a seven-point scale). Green parties also generally support strengthening the European Parliament (5.6 on a seven-point scale, second only to the liberals). A major determinant of green party positioning on European integration is the relative weight of pragmatic (“realo”) versus principled (“fundi”) tendencies. A second, related, influence is whether the party is purely environmentalist or combines green and radical-left views (Christensen 1996; Bomberg 1998). The more reformist and environmentalist a green party is, the more likely it is to support European integration. In recent years, reformism has been ascendant in the larger green parties, including the influential German greens. Back in 1984, the German greens condemned European integration in sweeping terms as an attempt to create a European superpower. By the early 1990s, their position was more complex. Four green MEPs challenged the Maastricht Treaty in the German Constitutional Court on the grounds that it violated “eternal” German constitutional principles of “federalism” and “democracy” (Rudig ¨ 1996: 264).12 But the party also made clear that it supported European integration in principle: “Especially in view of increasing nationalistic and racist opinions and attacks in Germany and elsewhere, the Greens emphasize the importance and necessity of European integration.”13 The French greens have moved in the same direction. The Maastricht referendum of September 1992 precipitated an intense debate in Les Verts and Génération Ecologiste – the former straddling the issue, and the latter recommending a “yes” vote (Rudig ¨ 1996). Poor results in subsequent national and European elections strengthened the pragmatic, proEuropean wings of both parties (Bomberg 1998; Hermet et al. 1998). Daniel Cohn-Bendit led Les Verts in the 1999 European Parliamentary elections on a manifesto that was expressly pro-integration. All green parties, except the Irish and Swedish parties, have moderated their Euroskepticism. It is no coincidence that these two small parties have a strong new left current. Increasing support for European integration 12 13

The Court rejected their claim, but stated that further development of the European Union had to be in step with the strengthening of democratic institutions. Policy statement of the Land Council, October 1992 (quoted in Rudig ¨ 1996: 263).


139

has been most pronounced in the larger green parties, particularly the Austrian, Dutch, Finnish, French, and German greens. Green parties are most favorably disposed to European integration where it touches their substantive commitment to protect the environment, strengthen democracy, and liberalize immigration law. These parties take pro-integration positions on the environment, strengthening the European parliament, and on asylum, and it is on precisely these issues that GAL/TAN location exerts the strongest influence relative to variables reflecting a party’s left/right location (see table 6.2). The new politics dimension of contestation structures party positions on European integration because TAN parties are so deeply opposed. On the GAL side, green parties have become more favorably disposed to European integration in the 1990s, and we hypothesize that they have done so to the extent they are pragmatic, rather than fundamentalist, and green, not red-green. Conclusion We began by asking whether support for European integration is structured by two dimensions of contestation that predominate within EU member countries, and we can now answer that indeed it is. By disaggregating European integration into component policies, a detailed, explicable pattern of support and opposition comes into view. The programmatic convictions that constrain party positioning in domestic political arenas also constrain party positioning on European integration. We find that the conventional left/right dimension – summarizing contestation about political control of the economy – structures European integration. Political parties located toward the left extreme and the right extreme – peripheral parties – are significantly more Euroskeptical than parties towards the center. The influence of left/right extremism is demonstrably not a facet of either weak electoral performance or exclusion from government. Does left/right location influence support for European integration in linear fashion? We find that it does, but the effect is strongest on issues that involve the basic choice between a neoliberal Europe and European regulated capitalism. Party positions on EU environmental policy, EU cohesion policy, and EU employment policy are constrained by location on the left/right dimension. Social democratic parties that were merely lukewarm on the market-making project of the 1980s provide strong support for European integration in these policy areas. The most important finding of our research is that the new politics dimension of party competition, ranging from green/alternative/libertarian

140


politics (GAL) to traditional/authoritarian/nationalist politics (TAN), powerfully structures variation on issues arising from European integration. The association is strongest at the TAN pole, reflecting the strength of nationalism among radical right parties. However, the new politics dimension also influences the positions of major parties. We find strong associations between party score on the GAL/TAN dimension, overall support for European integration, and support for particular aspects of European integration, including environmental policy, asylum policy, and strengthening the European Parliament. Given the lack of attention to this factor in our own previously published work and in the work of others, the strength of the new politics dimension is unexpected. We included a question on the new politics dimension in our survey on the hunch that it might be related to EU party positioning, but we had no idea that it would overshadow the left/right dimension. Perhaps we were not listening clearly enough to the insight that first sparked our interest in the effect of the left/right dimension: European politics is domestic politics by other means. Our analysis of this phenomenon is no more than a first cut. There is much more to find out, particularly about how European integration plays back on domestic contestation. Does European integration bring some aspects of the new politics dimension to the fore while de-emphasizing others? Will European integration inject the issue of national sovereignty into party competition? Will this strengthen political parties near the TAN pole? We can only speculate about these basic questions arising from our analysis. What we can predict, however, is that future research on support for European integration will investigate the new politics dimension of contestation as intently as it has investigated the left/right dimension.

7

Political competition in the European Parliament: evidence from roll call and survey analyses Jacques J. A. Thomassen, Abdul G. Noury, and Erik Voeten

Ever since Schumpeter (1942) defined democracy in terms of a competition of political leaders for the votes of the people, public contestation or political competition has been generally recognized as one of the most essential characteristics of modern democracy (Dahl 1971). As modern democracy is hardly conceivable without political parties, political competition implies a major function for mass political parties. As Bingham Powell (1982: 3) puts it: “The competitive electoral context, with several political parties organizing the alternatives that face the voters, is the identifying property of the contemporary democratic process.” It is in this respect that the European Union is often said to be failing. There is no competitive electoral context at the European level. European elections are basically fought by national political parties on national rather than European issues. Because national party systems are based on national cleavages, they fail to organize the alternatives that are relevant to the voters in European elections, i.e., alternatives with respect to the development of the European Union as such. Even worse, any debate on these issues is suppressed by the leadership of the major political parties because they are internally divided on these issues and would risk being split apart when these issues were politicized. In order to remedy this aspect of the democratic deficit, it has been argued that in order to face the European electorate with a relevant choice, the party system should be reshuffled in such a way that parties organize themselves along the continuum pro- vs. anti-European integration. Elsewhere (Thomassen 1998; Thomassen and Schmitt 1999a, 1999b) we have questioned this line of argument. To the extent that the Union develops into a higher level of government, the major policy issues it is faced with are not necessarily different from those on the agenda of national governments, only on a larger scale. From this perspective the effectiveness of the European party system does not depend on the extent to which 141

142

Jacques J. A. Thomassen, Abdul G. Noury, Erik Voeten

it is able to politicize the issue of European integration but rather on the extent to which it is able to aggregate national political agendas and the national cleavage structures at the European level. Moreover, there is a second element in this discussion that is often overlooked. As much as “[t]he system of managing the major political conflicts of a society by allowing one or more opposition parties to compete with the governing parties for votes in elections and in parliament” might be “one of the greatest and most unexpected social discoveries that man has ever stumbled upon” (Dahl 1966: xvii), there is a precious balance between competition and political stability. A supposition as old as Aristotle is that a more or less homogenous society is an important condition for a stable political system. Heterogeneity in terms of race, language, or religion puts the stability of a political system at risk. A more refined version of this argument is that it is not heterogeneity as such that threatens the stability of a political system, but the extent to which they are cross-cutting rather than mutually reinforcing. As Lipset (1960: 89) argues: the chances for stable democracy are enhanced to the extent that groups and individuals have a number of crosscutting, politically relevant affiliations. To the degree that a significant proportion of the population is pulled among conflicting forces, its members have an interest in reducing the intensity of political conflict . . . If the issues of religion, citizenship, and “collective bargaining” are allowed to accumulate, they reinforce each other, and the more reinforced and correlated the sources of cleavage, the less likelihood for political tolerance.

The belief in the validity of this general law of political stability in the Anglo-Saxon literature of the 1960s and 1970s was such that Robert Dahl, one of its advocates, once confronted a Dutch colleague with the statement: “You know, your country theoretically cannot exist” (Daalder 1974: 606). The Netherlands with its typical verzuiling, or pillarized society, had become the prototype of a society where major social groups were totally segmented and therefore, according to Lipset’s argument, vulnerable to political instability. Since this controversy occurred, the proponents of the model of consociational democracy, and Lijphart in particular, have made it sufficiently clear that a plural, heterogeneous society is compatible with a stable democratic political regime, not only in real life, but even in theory. A plural society is not necessarily unfit for democracy, but only for a particular kind of democracy, a majoritarian model of democracy. Different societies need different regimes (Lijphart 1984). In his earlier work, Lijphart illustrated his argument on the basis of the typology in figure 7.1. His main argument is that the danger of adversarial politics in a plural society can be neutralized by prudent political leaders

Political competition in the European Parliament

143

Society Homogeneous

Plural

Coalescent

Depoliticized democracy

Consociational democracy

Adversarial

Centripetal democracy

Centrifugal democracy

Elite behavior

Figure 7.1 A typology of democratic regimes. Note: Lijphart (1977: 106).

who recognize this danger just as well as political scientists do and opt for a different type of democracy – a consociational type of democracy where political elites, in order to prevent the political system from falling apart, tend to cooperate rather than to fight each other. According to Lijphart, two types of regime are stable, while the other two are not. A centripetal democracy is the classic type of democracy to which much of the Anglo-Saxon literature refers: a homogenous society where two (groups of) parties compete for political power. If both homogeneity and cooperating elites are enhancing political stability, one might expect that the combination of the two might lead to the most stable type of democratic regime. However, this is not the case because of the inherent oligarchic and therefore undemocratic characteristics of such a system. It will mobilize its own opposition and is therefore inherently unstable. The centrifugal type of democratic regime is unstable because unrestrained political contestation in a heterogeneous society might make political controversies too hot to handle. Recognition of this problem by prudent political elites makes the difference between an unstable centrifugal and a stable consociational type of democracy (Lijphart 1977: 105–9). Political parties as a factor of integration If even some of the smaller member states of the Union hardly meet the requirements for a competitive party system to work as an integrating rather than disintegrating factor, it defies our imagination to see how such a system could work at the European level and contribute to an ever closer Union. And yet, no less an ambition is expressed in article 138a of the Maastricht treaty: “Political parties at European level are important

144


as a factor for integration within the Union. They contribute to forming a European awareness and to expressing the political will of the citizens of the Union.” Elsewhere (Thomassen and Schmitt 1999a, 1999b), we have argued that the second function, expressing the will of the citizens of the Union requires a competitive party system at the European level. In order to play an important role as a factor for integration within the Union, a European party system needs to meet further requirements. Our contention is that a competitive party system at the European level will only be a factor of integration when the major cleavage dimensions are cross-cutting. We limit our search for these dimensions to the cleavage structure that is becoming evident in the European Parliament. Former research, based on different data sources, revealed two or three main dimensions of conflict in European Union politics. The first one is the left/right dimension. In most European countries the argument can and has been made that a single left/right dimension has absorbed all traditional cleavage dimensions. We also distinguish a traditional/libertarian dimension that is presented either as an element of the left/right dimension (Hix 1999a) or as a dimension that is highly correlated with the left/right dimension but does not coincide with it (Thomassen 1998; Thomassen and Schmitt 1999a, 1999b; Hooghe, Marks, and Wilson chapter 6). More important than the distinction between these two dimensions is the existence of an integration/independence dimension (Hix 1999a; Thomassen and Schmitt 1999a, 1999b) that is independent of the left/right dimension in its broader meaning. According to our argument, a cross-cutting of these two conflict dimensions enhances the stability of the European political system. This argument can be interpreted in two different ways. The first interpretation is that attitudes and opinions on European integration of individual people and politicians and the positions of political parties should not be constrained by the left/right dimension. However, a second interpretation is far more important. The major challenge for an effective democratic political system at the European level is to counteract the traditional dividing lines in Europe, the national borders. Therefore, our contention is that political parties will be more important as a factor of integration within the Union, the more they are based on cross-national cleavages rather than on national identity. Similarly, European political integration is enhanced when members of the European Parliament take positions according to their European party affiliation rather than their national identity. This holds true for both policy dimensions. The more political differences coincide with national borders, the more disruptive politicization of these differences will be.


145

In order to preserve the Union, political elites would be forced to resort to an accommodating style of politics, which at the European level could hardly be anything else than a purely intergovernmental regime. Also, some issue domains might be more sensitive for national differences than others. If, on “normal” left/right issues, the political majority at the European level is different than in some member states, this most certainly does not contribute to the process of integration, but as long as it does not reflect a more or less permanent difference between countries this at least is a situation that most federal states have learned to cope with. However, the issue domain of integration/independence is far more sensitive. Not a single European nation would be impressed by the argument that it should give up part of its sovereignty because a majority of the rest of Europe has decided that it should. Against this background the objective of this chapter is to determine the relative importance of cross-national political affiliations versus national identity in different policy domains. In the next section we describe how cross-national the major party groups in the European Parliament are. In the following section we explain why we use two different research methods: roll call analysis and survey research. Then we present the results of our analysis using these separate methods. In the concluding section we reflect on the meaning of our findings for the prospect of representative democracy in the European Union. Political parties at the European level A simple test of the transnationality of European party groups is how many countries are represented in each group. Raunio (1996: 125) used an index of transnationality (IT), based on Rae’s index of party fractionalization. The formula for its calculation is: IT = 1 −

n

SC 2

i =1

where SCi is the share of members from country i within a given political group. By definition, I T will be close to 0 if the political group is mainly composed of members from a single country (i.e., SC is close to 1). On the other hand, I T will be close to 1 if there are many members from various countries (i.e., SC is close to 0). The results for the 1989–94 parliament are presented in table 7.1. It is clear that in a formal sense the transnationality of at least the major party groups is high. Parties from all or almost all member states take part

146


Table 7.1 The transnationality of European party groups Party groups

IT

PES EPP ELDR EDG V EUL EDA ER LU Rainbow

0.852 0.841 0.851 0.111 0.790 0.360 0.566 0.526 0.653 0.824

Source: Raunio (1996: 127).

in the PES, the EPP, and the ELDR. At least as important is that a clear majority of the representatives from each country are members of these three major party groups. Only in France and Italy are sizeable numbers of MEPs represented in smaller and far less internationalized groups (UPE). Research approach Above we have argued that European political integration is enhanced when members of the European parliament take positions according to their European party affiliation rather than their national identity. This assumption dictates our empirical approach. Therefore, we address the following research question: What is the relative effect of party group membership and national identity on the positions MEPs take in different issue domains? The analysis in table 7.1 shows that the organization of the European party system is transnational in a formal sense. However, our research question is posed at the level of the individual behavior and attitudes of MEPs. We therefore rely on two different research methods: roll call analysis and a survey among MEPs. Roll call analysis is the traditional way of measuring the cohesiveness of party groups in the European Parliament (Attina 1990; Bardi 1994; Brzinski 1995; Raunio 1996 and 1998). The ability of political groups to achieve internal party cohesion and discipline in parliamentary votes is an indication of their success or failure (Brzinski


147

1995: 137). In general these analyses have revealed a surprisingly high level of cohesiveness within party groups. This finding has caused some skepticism with regard to the validity of roll call behavior as a measure of the cohesiveness of party groups in the European Parliament. According to Bardi, the findings of a very high degree of intra-group cohesiveness in the analysis of EP roll call data should be treated very cautiously: Intra-group quasi-unanimity is usually guaranteed by long preparatory committee and party-group sessions which have the stated purpose of smoothing out most disagreements. What is even more important, however, is that most MEPs consider quasi-unanimity necessary for the projection of a strong image vis-à-vis other EC or national institutions, and it is quite possible that MEP voting behavior is distorted by the paramount importance which is attributed to this goal. In such circumstances, intra-group cohesiveness as such may lose most of its meaning, at least as an indicator of Europarty institutionalization, particularly because inter-group differences would themselves be blurred (Bardi 1994: 367–8).

In spite of this, one might consider the high level of party group discipline in voting as an indicator of the success of European parliamentary institutions in framing European politics according to ideological and party lines rather than national interests. However, if we want to get a better understanding of the prospects for a transnational party system, we also need to analyze the attitudes of MEPs. If voting by unanimity is the result of an endless massage of different policy views, we need to know how much massage was needed and how much tension is camouflaged by a unanimous vote. Asking the opinion of individual MEPs on important issue dimensions can reveal this. The extent to which these opinions are constrained by party affiliation rather than national identity is an important indicator of the potential of transnational political parties as a factor of integration. In this chapter we use both research strategies. In the next section we report the results of roll call analysis, and in the following section the results of a survey among MEPs. Roll call analysis Data and method One way to examine both the relative importance of nationality compared to political party affiliation and the cohesiveness of the European party system is to examine the voting behavior of members of the European Parliament (MEPs). An analysis of the dimensions underlying MEPs’ votes on various proposals during the plenary sessions of the European Parliament provides us with valuable indications of how coalitions form. The analyses presented in this section are based on Noury (2002). Our

148


database contains a large number of MEPs’ voting decisions during the EP’s third (1989–94) and fourth (1994–9) legislatures. We used legislators’ “yes” and “no” decisions on all votes during those periods. Abstentions are treated as missing values. Following the literature on roll call voting, we discarded all votes with less than 2.5 percent on the minority side (i.e., unanimous or “near unanimous” votes).1 To scale legislators, we used Poole and Rosenthal’s (1997) spatial model of roll call voting, called NOMINATE.2 Similar to standard factor models, NOMINATE is a scaling procedure that can be used to summarize the structure underlying data. Unlike factor models, NOMINATE has a foundation in a popular political-economic model of legislative decision-making: the spatial model (e.g., Downs 1956). The underlying behavioral model assumes that each legislator, characterized by a bell-shaped utility function, has an ideal position in a multidimensional policy space and acts rationally based on it. NOMINATE uses a (quasi-) maximum-likelihood procedure to estimate the unobserved ideal points of legislators from their observed vote choices. Dimensions of contestation In analyzing roll call votes, one important question is the number of dimensions needed to explain legislators’ voting behavior. We can address this question by evaluating the extent to which NOMINATE models of different dimensionality account for the observed roll call votes. Noury (2002) identifies four effective dimensions to MEPs’ voting behavior during the EP’s fourth legislature. The first and by far the most important dimension classifies about 90 percent of legislators’ votes correctly. The explanatory power of each additional dimension decreases smoothly until the fourth dimension. Additional dimensions merely capture random noise. The second, the third, and the fourth dimensions correctly classify a small but a non-negligible share of MEPs’ votes. More precisely, adding the second dimension to the model increases the correct classification scores by about 2 percent on average. As far as the third and fourth dimensions are concerned, they increase the correct classifications by about 1.8 percent and 1.2 percent, respectively. Although the third and fourth dimensions increase the classification scores only slightly, they are non-negligible because they have substantive interpretations. Moreover, 1

2

This research project was undertaken with support from the European Commission’s Phare ACE program 1996, project no. P96-6252-R, and has been supervised by Gérard Roland from ECARE, Brussels. NOMINATE stands for NOMINAl Three-step Estimation; see Poole and Rosenthal (1997) for a detailed presentation and discussion of its statistical properties.


149

Table 7.2 Mean and variance of political groups’ positions in the EP (fourth parliament) PSE (N = 241)

EPP (N = 199)

ELDR (N = 52)

UPE (N = 60)

GUE (N = 23)

Mean Variance Mean Variance Mean Variance Mean Variance Mean Variance Dim. 1 −0.17 0.010 Dim. 2 0.63 0.04 Dim. 3 −0.23 0.04 Dim. 4 −0.29 0.036

0.47 0.54 0.39 0.20

0.014 0.049 0.055 0.036

0.18 −0.21 −0.31 0.68

0.016 0.041 0.040 0.084

0.66 −0.39 −0.09 −0.42

0.010 0.039 0.010 0.016

−0.13 −0.35 0.47 −0.31

0.014 0.020 0.034 0.047

other goodness-of-fit statistics frequently used in the literature on roll call voting, namely the proportional reduction in errors (PREs), show that these dimensions should not be discarded from the analysis (see Noury 2002).3 After determining the dimensionality, we discuss the substantive meaning of those dimensions. As with exploratory factor analysis, attaching a substantive interpretation to the uncovered dimensions is a subjective matter. We can get some leverage on this question by analyzing how votes on different issues fit in the uncovered policy space and/or by relating observed characteristics of legislators (e.g., party, nationality) to their ideal points. We pursue the latter route here. Table 7.2 reports the average ideal points of the main political groups in the EP on all four dimensions. The table also gives the within-party variance in these ideal points. The first dimension is the traditional left/right dimension. All members affiliated with a left-wing political group (i.e., PSE and GUE) are located on one extreme of the first dimension. That is, they all have negative average ideal points. Members of right-wing political groups (e.g., EPP and UPE) have their ideal positions on the opposite side. Their average ideal points are positive on the first dimension (as reported in table 7.2). Liberals (ELDR) are located midway between these two groups of legislators on the first dimension. This finding is in concordance with the results from the other chapters in this volume. Left/right is the dominant dimension of contestation in the European Parliament, just as it is the dominant constraint underlying attitudes of European citizens, party manifestos, social movements, and expert evaluations of party positions. 3

The relative importance of the first dimension is exaggerated somewhat because of the large number of lopsided votes. The PRE evaluates the explained variance relative to the simple model that everyone always votes with the majority. Using this criterion the additional dimensions become relatively more important, although the first dimension still dominates. See Noury (2002) for more detail.

150


The second dimension is the integration/independence dimension. On this dimension, positive coordinates characterize the two largest political groups (i.e., PSE and EPP). As far as European issues are concerned these two parties have almost identical average ideal points! Smaller, extremist, and anti-European political groups, together with greens, are located on the opposite side of the second dimension. As Hooghe, Marks, and Wilson observed in the previous chapter, the parties that are on opposite extreme ends of the left/right ideological spectrum take similar positions on this integration/independence dimension. It should be noted that many European policies concerned with redistribution are captured by the left/right dimension of contestation in the European Parliament. The third and fourth dimensions are related not only to political party affiliation but also to legislators’ nationalities. To be precise, the third dimension mostly separates MEPs from the UK from other members. The fourth dimension captures conflict between nationalities along a north– south division. We refer the interested reader to Noury (2002) for more detail. To analyze the voting cohesion of party groups, we also report the intra-party variance of ideal points in table 7.2. A political party group that is cohesive on an issue dimension should show little variance in its members’ ideal points. From table 7.2 we can see that some political groups are more cohesive than others and that political party groups vary in their cohesiveness by issue dimension. Legislators of all political groups vote most cohesively on issues that fit conflict on the first dimension, the left/right dimension. This implies that party group cohesiveness is highest on issues that fit best within the traditional national party system in most European countries. Party cohesiveness on the second dimension is much less than on the first. This tells us that the party groups vote less cohesively on issues that relate to European integration than on issues that fit the left/right dimension. In general, the intra-party variance of MEPs’ ideal points for most parties decreases as we move to higher dimensions.4 This indicates that nationality or factors other than party adherence might explain more of the variance between MEP ideal points on the higher dimensions. In general, party cohesiveness is large on most issue dimensions. Moreover, Noury (2002) shows compelling evidence that party groups have become increasingly cohesive actors over time. To be precise, the intra-party variance on the first dimension has become twice as small over an eight-year period. 4

The only exception is the UPE, whose variance on the second dimension is considerably larger than its variances on the third and fourth dimensions.


151

Table 7.3 Impact of party and nationality on ideal point locations (entries are eta-squared) Party

1st dimension 2nd dimension 3rd dimension 4th dimension

Nationality

All

Biggest five∗

All

Biggest five∗∗

0.892 0.875 0.668 0.729

0.897 0.781 0.685 0.746

0.083 0.276 0.307 0.273

0.086 0.405 0.381 0.183

Notes: ∗ This analysis is limited to members of the five largest party groups: PSE, EPP, ELDR, GUE, UPE, which together account for 81 percent of MEPs. ∗∗ This analysis is limited to members from Germany, France, UK, Italy, and Spain which together account for 64 percent of all MEPs.

Party groups and nationality Our discussion of the results in the preceding subsection suggests that there are four dimensions to MEPs’ voting behavior. Each dimension has a distinct substantive meaning. We now investigate to what extent MEPs vote according to their political group affiliation or nationality and how this depends on the issue dimension. In Noury (2002), we used analysis of variance (ANOVA) to discriminate between the two competing factors, namely party and nationality. The ANOVA results show which factors explain the largest share of variance in MEP ideal points along the different issue dimensions. If MEPs vote in accordance with their political party affiliation rather than their nationality, the share of variance explained by the former should be higher. The ANOVA results presented in table 7.3 show that on all dimensions the share of variance explained by MEPs’ political party affiliation is substantially higher than the share of variance explained by MEPs’ national identity. Note, however, that the variation in MEP ideal points explained by political party affiliation declines in higher dimensions. Especially on the third dimension, nationality becomes an important explanatory factor. The results differ slightly when we limit our focus to the five largest parties and countries. As far as the biggest five parties are concerned, the results remain virtually unchanged, except that the five largest parties appear to be somewhat less cohesive on the integration/independence dimension than the smaller parties. The results change to a larger extent when we focus on the biggest five countries. The share of variance explained by nationality on the integration/independence

152


dimension is much larger for the biggest five countries than for the sample with all countries included. That is, MEPs of the largest five countries vote more according to their nationality on European issues than other MEPs.

Analysis of survey data Data and method In this section we address the same research questions as in the previous section, but now by using survey data from the project “Political Representation and Legitimacy in the European Union.”5 This project is an extension of the European Elections Study (EES) 1994 and consists among others of the following components: a. a survey among members of the European Parliament (face-to-face interviews);6 b. a mail survey among members of the national parliaments of the member states.7 We use data from the survey among the members of the European Parliament (MEPs). In this survey, MEPs were asked to give their opinion on a great number of issues. In this section we analyze the structure underlying the attitudes of MEPs on these issues and compare this to the structure uncovered by our analysis of roll call votes. Survey data provide a nice complement to roll call data. First, the substantive interpretation of the dimensions extracted from a scaling analysis of hundreds of roll calls is often problematic. By using scales that are extracted from a limited and specific set of issue questions, we are better equipped to analyze the impact of nationality, party, and party cohesiveness on substantive issue dimensions. Second, survey data provide an opportunity to better evaluate the role of party groups in establishing the dimensions of contestation in the European Parliament. Is the cohesiveness of party groups in actual roll call voting merely a product of pre-floor organizational efforts or can 5 6

7

The main publications from this project are Marsh and Norris (1997), Katz and Wessels (1999), and Schmitt and Thomassen (1999). The fieldwork was done by INFRATEST. Fieldwork costs were covered by a grant from the Deutsche Forschungsgemeinschaft. The survey was coordinated by Bernhard Wessels, Wissenschaftszentrum Berlin. These surveys were conducted in all member states, except Austria, Denmark, and Finland. This part of the study was coordinated by Bernhard Wessels, Wissenschaftszentrum Berlin, and Jacques Thomassen, University of Twente.


153

Table 7.4 Issue dimensions in the European Parliament (factor loadings > 0.4 are in bold) Integration/ independence Increase range of responsibilities of EU Democratic legitimization of EU based on EP/NP EP power to pass laws that directly apply to all members National/European currency Decisions at national/European level∗ Remove national borders/border control Reduce unemployment/limit inflation EU employment program/concentration on single market Reduce inequality of incomes Maintain levels of welfare even if taxes rise Government given greater role in economy Tougher action against criminals Decriminalize use of marijuana Stronger measures to restrict immigration Women free to decide on abortion Eigenvalues (after VARIMAX rotation) Percent of variance explained

Socio-economic left/right

Libertarian/ traditional

0.88

0.02

0.05

0.87

−0.02

0.08

0.85 −0.83 −0.81 0.76 −0.09

0.05 0.11 −0.13 −0.06 0.83

0.12 0.09 0.13 0.17 0.12

0.11 −0.03 0.08 −0.04 0.05 0.28 0.01 0.01

0.82 0.79 0.78 0.76 −0.02 0.11 −0.24 0.23

0.09 0.31 0.24 0.01 −0.73 0.72 −0.71 0.68

4.3 28.5

3.3 22.3

2.3 15.1

Note: ∗ A scale from questions v6 1 to v6 17 of the MEP survey on the appropriate level of decision-making for different policy areas. Cronbach’s alpha is 0.93. It holds for all items that if they are removed from the scale, the reliability of the scale decreases.

it also be found in the structure underlying the attitudes of MEPs? A related issue is that we may expect that national differences matter more in the attitudes of MEPs than in their voting behavior where differences may have been blurred through long preparatory committee meetings and party group sessions. Issue dimensions Table 7.4 reports the results from a factor analysis on the set of issue questions included in the MEP survey. The results reveal three distinct issue domains: an integration/independence, a (socio-economic) left/right, and a traditional/libertarian dimension. The issues in the integration/independence domain all refer to the power and competencies

154


of the Union vs. the member states, rather than to the distribution of power between different European institutions. The second dimension consists of socio-economic issues that all relate to the classic left/right dimension: social equality, the role of the state, and maintaining the welfare state vs. a free market economy. The third dimension, which can be labeled traditional/libertarian (cf. Kriesi 1989; Middendorp 1992), consists of a moral issue (abortion), the issue of immigration, and two issues in the domain of law and order. This dimension is similar to the new politics dimension discussed in the previous chapter. It should be noted that the relative amount of variance explained by the individual factors provides little indication of the relative importance of the factors in the European Parliament. It merely reflects the proportion of variance explained in the items included in the analysis. We should thus not attach any significance to the observation that the integration/independence dimension reflects the first dimension of analysis in the survey analysis whereas it represents the second dimension in the roll call analysis. It is, however, significant that the same substantive dimensions show up as important dimensions that underlie both attitudes as well as the roll call voting behavior of MEPs. The three-dimensional structure found in table 7.4 is robust in comparison with results that are found in studies that have used alternative methods. Hix (1999a) found a very similar structure in an analysis of European leaders’ statements. He presents the socio-economic left/right and libertarian/traditional dimensions as two different components of the left/right dimension. It is unmistakable that these two components are highly related to each other (see also Hooghe, Marks, and Wilson in the previous chapter). The self-placement of MEPs on a left/right scale correlates strongly with the factor scores of MEPs on the socio-economic factor (Pearson correlation of 0.56) as well as on the libertarian/traditional factor (correlation of 0.52). This does, however, not imply that the two factors represent the same substantive content, nor that political conflict on the two factors is necessarily structured in the same way. The factor solution in table 7.4 is the result of a VARIMAX rotation. Therefore the three factors by definition are uncorrelated. An OBLIMIN rotation of the factor solution shows that the two factors are correlated, but only moderately (0.30).8 Moreover we can see from the table that most items with a high loading on one factor have low or modest loadings on the other factor. Table 7.5 reports the mean and variance of factor scores by European party group. The table is equivalent to table 7.2. It demonstrates that the 8

The European factor has zero correlation with the other two factors.

0.62 0.38 0.42 0.06

−0.32 −0.56 −0.45 −0.66

−0.16 0.68 0.51 0.09

Mean 0.59 0.84 0.47 0.12

Variance

EPP (N = 89)

Notes: ∗ Ten-point scale that is rescaled to –1 (left) to 1 (right).

Integration/independence Left/right Libertarian/traditional Left/right self-placement∗

Variance

Mean

PSE (N = 98)

−0.15 0.59 −0.18 −0.07 0.41 0.96 0.99 0.12

Variance

ELDR (N = 42) Mean

Table 7.5 Mean and variance of party groups on the issue dimensions

0.63 −0.99 −0.72 −0.89

Mean

1.43 0.11 0.46 0.02

Variance

GUE (N = 19)

0.36 −0.40 0.97 0.16

Mean

0.80 0.66 0.61 0.10

Variance

UPE (N = 18)

156


ELDR is located on the “right wing” of the socio-economic factor and on the “left wing” of the libertarian/traditional factor (cf. Thomassen and Schmitt 1999b). The UPE (conservatives) represent exactly the opposite pattern: they are strongly “right-wing” on issues such as crime and immigration, but more centrist on socio-economic issues. While both factors thus clearly relate to the left/right division that is often so conveniently used to characterize party competition, each factor also has a distinctive substantive interpretation that would be obscured by relying on a simple one-dimensional interpretation of left/right. Moreover, it is noteworthy to observe that the structure of conflict between the three main parties on the libertarian/traditional dimension resembles partisan conflict on the third dimension found in the roll call analysis.9 Table 7.5 also presents the intra-party group variance of the factor scores on the three issue dimensions. The variance of the entire sample on these factor scores is, by assumption, 1. This implies that a variance of 0.5 within a party group means that the variance in scores within the party is about twice as low as within the entire sample. A variance of larger than 1 thus implies that the variance within a party group exceeds the average variance in the sample, and thus that partisanship is an unimportant predictor of attitudes. An example is the GUE on the integration/ independence dimension. Against this point of reference, most variances indicate less than impressive party cohesiveness underlying MEP attitudes. This holds not only on the integration/independence dimension, but on the other two dimensions as well. The high intra-party variances for the ELDR on the left/right and libertarian/traditional dimensions imply that the variance within this party equals the variance in the entire sample on these dimensions. We can thus reject the notion that the ELDR is a cohesive party on these two attitudinal dimensions. The EPP has a large amount of intra-group variation in attitudes on the socio-economic left/right dimension. The limited cohesiveness of opinions on socio-economic issues from EPP and ELDR MEPs stands in glaring contrast to the cohesive PSE on this dimension. This result indicates that the homogenous composition of the PSE of purely socialist parties is not without consequences when compared to the other two major party groups who seem to pay a logical price for the 9

However we should note that libertarian/traditional issues are likely to be absorbed in a combination of dimensions in the NOMINATE solution if these issues are, as we argue, related to the socio-economic left/right dimension. Since a NOMINATE solution represents a multidimensional space rather than a set of uncorrelated factor solutions, a separate substantive issue dimension does not necessarily show up as an independent dimension, but may simply be represented by different locations of cutting lines in the multidimensional space (see Voeten [2000] for a non-technical explanation of this phenomenon).


157

Table 7.6 Influence of party and nationality on MEP attitudes (entries are eta-squared) Party

Nationality

Factors

All groups

Biggest five∗

All countries

Biggest five∗∗

Integration/independence Socio-economic left/right Libertarian/traditional Left/right self-placement

0.31 0.38 0.42 0.67

0.10 0.43 0.37 0.65

0.38 0.14 0.11 0.04

0.21 0.11 0.04 0.01

Notes: ∗ This analysis is limited to members of the five largest party groups: PSE, EPP, ELDR, GUE, UPE, which together account for 86 percent of all valid responses. ∗∗ This analysis is limited to respondents from Germany, France, UK, Italy, and Spain which together comprise 61 percent of all valid responses.

diversity of the parties that were admitted in the past. If this interpretation is correct, we expect to find large nationality-based differences within the ELDR and the EPP, but not within the PSE on this dimension. We investigate this hypothesis in the next subsection. It is interesting to observe that the substantial differences between the cohesiveness of major party groups do not translate into similar differences in cohesiveness in voting behavior (compare table 7.2). This finding may indicate that the EPP and the ELDR leaderships are successful in orchestrating cohesive behavior among their heterogeneous members. On the other hand, it may simply reflect the relative absence of divisive issues that concern socio-economic redistribution in the European Parliament. For example, most redistribution carried out by the EU is sectoral or geographic in nature rather than class-based. Party groups and nationality We have argued that the more the behavior and attitudes of MEPs are constrained by their transnational party affiliation rather than their national background, the more political parties at the European level can be considered as a factor for integration. The outcome of our roll call analysis was that party affiliation is far more important than national background. We now investigate to what extent this outcome is sustained by the survey data. Table 7.6 presents the results from an analysis of variance (ANOVA). The table is the equivalent of table 7.3 in the roll call analysis. The entries indicate the percentage of the variance in the factor scores that is explained by party group membership and national background

158


respectively. In the second column the analysis is limited to the five largest party groups and countries respectively. A number of conclusions can be drawn from this table. First, there is a clear difference in the results for the different factors. Party group membership is far more important than nationality in explaining the variance in issue positions on the left/right and the libertarian/traditional dimension than on the integration/independence dimension. This result is consistent with the findings from the roll call analysis (see table 7.3). At first glance, party group is almost as important as nation in explaining the variance on the integration/independence dimension. However, the major explanatory factor here is the difference between a few small antiEuropean party groups and the five largest party groups. The number of MEPs of smaller groups in our sample is too small to be reliable. Once we limit the analysis to the five major party groups, the variance explained by party group is reduced to a mere 10 percent. This clearly confirms our finding from the roll call analysis that the large parties, especially the PSE and the EPP, are not competitive on the integration/independence dimension. Further evidence for this lack of competition between the large party groups on the integration/independence dimension is provided by a linear regression with the five party dummies as explanatory variables and the factor scores on the integration/independence factor as dependent variables. Here we found that the coefficients for the three biggest parties (the PSE, the EPP, and the ELDR) are statistically indistinguishable from each other at any reasonable level of significance.10 Simply replacing these three dummy variables with one variable that indicates membership in either of the three biggest party groups does not decrease the explanatory value of the model.11 This implies that MEPs of the PSE, the EPP, and the ELDR have indistinguishable attitudes on the integration/independence factor. This may explain why Gabel and Hix in chapter 5 only found one dimension of contestation in their evaluation of party manifestos. Among the main parties, there simply is not much competition on the integration/independence dimension. So if we reduce intra-party variance by assumption (as in all analyses of party manifestos), this dimension will not appear as an independent dimension of contestation. On the integration/independence factor, national background is more important as an explanatory factor, although the explanatory power of national background is reduced when the analysis is limited to the five 10 11

Even at a 10 percent significance level. An F-test of the difference in explanatory value of these two models generates an F-value of 0.82 (Pr (F) 0.44).


159

major countries. This finding is contrary to our findings in the roll call analysis. This may be the result of the relatively low number of respondents from various small countries in our sample. These results cannot come as a total surprise. The formation of party groups, with all its anomalies, is a more or less successful aggregation to the European level of more or less similar cleavage structures in the member states. The major dimension of conflict in most of these countries nowadays is the left/right dimension. Aggregating this dimension to the European level will make it as important at the European as on the national level. However, neither in theory nor empirically can we establish a clear relation between the left/right dimension and the integration/independence dimension. Therefore, it is no wonder that the structure of the party system is unrelated to issues that are related to deferring decision-making authority to the supranational level. As Hooghe, Marks, and Wilson pointed out in the previous chapter, this may be different for European issues that more directly correspond to the redistributional issues upon which most party systems are based. This does not necessarily imply that opinions on issues of European integration should differ from country to country. But as far as these issues are related to the issue of national sovereignty, it should not come as too much of a surprise either. Nationality within party groups In the previous subsection we hypothesized that the lack of party cohesiveness may be due to national differences within party groups. We now investigate this thesis for the two main dimensions underlying attitudes of MEPs: the integration/independence and the socio-economic left/right dimensions. In order to create subgroups of sufficient sample size we have split countries into different groups that are expected to share similar attitudes on the different issue dimensions. There is no lack of hypotheses on the question of how countries might differ on the integration/independence dimension. First, there is the hypothesis that support for European integration would be related to length of membership. According to this hypothesis the original six members should show the highest level of support, followed by the successive extensions. A second hypothesis is that there would be a direct relation between the net economic benefit of a country from the EU and its support for European integration. Third, the tension between center and periphery would have its effect on the level of support. The problem with all these hypotheses is that they partly overlap. In an attempt to do justice to these different insights we grouped the member states in four categories: the “original six” (the Benelux countries,

160


Germany, France, and Italy), the UK as very much a case of its own, the “Northern” extension countries (Denmark, Sweden, Finland, and Austria) and the “Southern” extension countries (Spain, Portugal, Greece, and Ireland). This configuration explains the variance within the three largest party groups well: eta-squared of 0.30 for the PSE, 0.47 for the EPP and 0.32 for the ELDR.12 Within all party groups the UK and Nordic representatives are less “Europhile” than representatives of the other countries are. There is little distinction between MEPs from the “original six” and the southern extension countries, except that, within the EPP southerners are slightly less “European.” On the left/right dimension we group countries following a modified “north–south division.” The rationale behind this division is that differences in attitudes on socio-economic issues are likely to be created by differences in economic conditions (captured by the general north– south division) as well as by differences in domestic welfare state experiences. Within separate country groups in the northern part of the EU, we follow Esping-Andersen’s distinction between different welfare state regimes. This implies separating the Scandinavian countries from Christian democratic regimes and the (in Europe) unique British model.13 We also create a specific category in order to separate France from the southern European countries, recognizing France’s unique political and economic system. As expected, given the homogenous ideological nature of the PSE, we find that nationality explains much more variance within the EPP and the ELDR than within the PSE. The ANOVA analysis results in eta-squareds of 0.10 for the PSE, 0.30 for the EPP14 and 0.42 for the ELDR.15 It is interesting that both within the EPP as well as within the PSE the Scandinavian MEPs are least “leftist.” Another interesting finding is that the difference between the average position of the EPP and the PSE on this dimension is extremely large for the UK, much smaller for southern European countries and the northern European Christian democracies and by far the smallest for France.16 Finally, in both northern and southern Christian democracies the ELDR is considerably more right-wing on

12 13 14 15 16

Even when the UK is eliminated from the analysis due to small N (2) within the ELDR. Scandinavian countries: Sweden, Denmark, Finland; Christian democratic: Austria, Germany and Benelux countries. See Esping-Andersen (1990). 0.29 when the only French EPP member in the sample is eliminated. Also 0.42 when the UK is eliminated from the analysis due to small N (2) within the ELDR. Difference between the positions of the EPP and the PSE for the UK is 2.25, for Christian democracies 1.29, for southern European countries 0.74, and for France only 0.27!


161

socio-economic issues than the EPP, whereas this party is centrist in the UK and Scandinavia.17 The national differences in attitudes towards European integration seem to be similar across party lines, whereas these differences on the socio-economic dimension point to distortions in the aggregation of national parties caused by fundamental differences in the relative position of national parties along this dimension. These differences become especially apparent when we analyze the relative positions of the ELDR and the EPP on this dimension. In some countries the liberals take a moderate position in between socialists and conservatives on socio-economic issues; in others they form the most right-wing party on these issues. This translates into relatively little party cohesiveness in attitudes of its members on socio-economic issues. However, it does not translate into less cohesive voting on left/right issues, as table 7.2 illustrates.18 Perhaps the extreme position and relatively high variance of the ELDR on the fourth dimension in table 7.2 captures some of this internal conflict. Another explanation for the relative cohesiveness of the voting behavior of the ELDR might be that social issues of a redistributional nature are relatively unimportant in the European Parliament. Summary and conclusions The research question addressed in this chapter is to what extent MEPs’ positions are defined by their membership of an EP party group rather than their national background. In order to answer that question we used two different methods, roll call analysis and a survey among MEPs. Rather than considering these as two alternative methods for measuring the same phenomenon, we used the survey data as a background for a better understanding of MEP voting behavior. In our analysis of MEPs’ voting behavior we first examined the dimensionality of the issue space of voting behavior. By far the most important issue dimension is the traditional left/right dimension, classifying about 90 percent of the legislators’ votes correctly. Far less important, but not negligible, are three additional dimensions. The most important and robust of these additional dimensions is the pro-/anti-European integration 17

18

The ELDR is 0.4 to the “right” of the EPP in northern Christian democratic countries and 0.29 in southern European countries. In Scandinavian countries, the ELDR is 0.38 to the “left” of the EPP; in the UK, this is 1.09. Note that the argument reflects differences in the relative position of parties within countries, not differences in the absolute positions of parties along the scales. This result can thus not be explained by national differences in perception of the scales.

162


dimension. That left/right and European integration are the most significant dimensions of contestation in the EP is confirmed both by the analysis of survey data in this chapter and by an analysis of policy statements by political leaders (Hix 1999a). The third and fourth issue dimensions are much less important and mostly seem to capture nationalist variations within party groups. On all issue dimensions, party group is a far more important explanatory factor of roll call behavior than nationality, although the explanatory power of party affiliation is higher for the left/right than the other dimensions. Also, the cohesiveness of party groups as expressed in roll call behavior is impressive. Again, cohesiveness on the left/right dimension is higher than on the other dimensions, including the European integration dimension. High levels of cohesiveness are a consistent finding in studies of roll call behavior in the European Parliament. It has become a matter of dispute in the literature whether roll call behavior is a valid indicator of EP party cohesion (Bardi 1994; Raunio 1996). Bardi is a representative of the skeptical view that the level of group cohesion shown by roll call analyses is due to long preparatory negotiations and compromises manufactured within the groups before the vote is taken. A survey among MEPs might be a better method to assess the level of cohesiveness. However, we are inclined to argue that it is wrong to define this debate in terms of the validity of instruments of measurement. Roll call behavior and attitudes are substantively different. But, of course, it is important to know what is behind these cohesive votes. Can cohesiveness in behavior be explained by a high level of consensus on policy matters within the party groups or is it due to the institutional context of the EP, like the decision rules of the Parliament requiring absolute majorities of the EP membership for the exercise of key powers (Hix and Lord 1998: 97)? Again the answer to this question is somewhat different for different issue dimensions and also depends on which findings one wants to emphasize. On both the left/right and the libertarian/traditional dimensions the attitudes of MEPs are far more strongly constrained by their party group membership than by their national background. This finding justifies the conclusion that the major party groups in the EP are more than accidental coalitions, but are based on a common ideology, leading to cohesive roll call behavior. However, at the same time, the variance in attitudes within party groups indicates a party cohesiveness that is far from impressive. The PSE seems to be the only party group that is relatively cohesive on the left/right dimension. Both the EPP and the ELDR have paid a price for admitting political parties with a different ideological background. Moreover, part of the variance within these two party


163

groups is related to national background, suggesting that in many cases it will indeed take quite a bit of massage to reach a common standpoint, not just between individual members of parliament but between national delegations. European party groups lack the hierarchical structure to impose parliamentary discipline. Therefore, there is no alternative to a consensus type of decision-making within these party groups (Lord 1998: 205). For the European integration dimension, our findings are totally different. On this dimension there are no systematic differences between the major party groups. Not party group, but national background is the major explanatory factor here. This observation may explain an apparent disparity between our findings and those in several of the other chapters which do not uncover a separate integration/independence dimension. The existence of this dimension is the combined result of a few small and cohesive party groups (the extremists) and of variation within the main party groups. Analyses that treat these party groups as cohesive units (e.g., those based on party manifestos) will thus not uncover this separate dimension. These findings can shed some light on the discussion of party formation in the European Union. As mentioned in the introduction, several objections against the present party system and the way it functions are brought into the discussion about the democratic deficit in the Union. A first objection is that European elections are not fought between real European parties competing for the votes of a European electorate, but by national parties who compete for the votes of a national electorate. A second objection is that European elections are not fought on European issues (i.e., on the dimension of pro-/anti-European integration), but on national issues. Our findings might lead to the conclusion that both problems cannot be solved simultaneously. Notwithstanding all the qualifications one can make, the European party system is not an unsuccessful aggregation of the major cleavage structure dominating politics in most European countries. From this perspective, a system of European political parties competing in a European electoral arena is not unthinkable. Given the broad array of interests each of these parties would have to take into account, their message to the voters would be general and hardly exciting. Whether that is good or bad is a matter of perspective. The argument of broad and diffuse parties is made both in defense of and against a two-party system (Lijphart 1984: 107–14). It depends on whether one wants to emphasize the moderating effect of such a system, forcing parties to accommodate the median voter, or the lack of a clear policy program due to all the compromises to be made. Such broad parties might still have a problem

164


arousing the European voter, but they certainly might be a factor of integration. However, it is out of the question that such parties might compete on the European dimension. Not only is there hardly any difference between the major party groups on this issue dimension, but at least as important is that party groups are internally divided on this issue. Any attempt to profile themselves on this dimension might split them apart. Therefore, the major parties deliberately try to keep the issue of European integration from the political agenda, preventing European voters from expressing themselves on this single most important issue dimension in European politics (Bogdanor 1989; Hix 1999a). For this reason, several observers have argued for reshuffling the European party system into a system that would allow the electorate to express their opinion on the future of the European Union. This would be the case when different parties represented different positions on the simple continuum of anti-Europe, in favor of the status quo, in favor of further European integration (Andeweg 1995: 69; Bogdanor 1989: 209). However, if such a party system became dominant in European politics, the remedy might be worse than the disease. Not only would such parties be internally divided on almost every other political issue, but given that policy views on this dimension are strongly related to national background, it would articulate national rather than cross-national political cleavages. Therefore, such a party system would most likely be a factor of disintegration rather than of integration in the Union.

8

Contesting Europe? The salience of European integration as a party issue Marco R. Steenbergen and David J. Scott

The ongoing process of integration in Europe has fundamentally altered the political environment in which the political parties of the EU member states find themselves. European integration has produced new political issues, which cannot always be easily accommodated into existing cleavage structures, as the preceding chapters reveal. It has also changed the political opportunity structure – parties may play these new issues up or they may play them down. In this chapter, we analyze why some national political parties have stressed European integration, while others have refrained from doing so. An analysis of the salience of European integration at the party level is important for several reasons. First, it speaks directly to the topic of contestation. A prerequisite for contestation is that political actors are willing to debate an issue – there is a willingness to give the issue a modicum of salience. To what extent do political parties show such willingness? Second, salience is also critical for understanding representation in the EU. Van der Eijk and Franklin (1996; this volume; see also Reif and Schmitt 1980) have observed that European elections are rarely about the scope and nature of integration, even though at the level of the electorate a contestation potential exists. This may be a contributing factor to the so-called democratic deficit of the EU. The lack of European content in European elections may be due to an unwillingness or inability of parties to raise integration above a critical salience threshold. If this is true, it is essential to understand why this happens. Finally, an exploration of issue salience sheds light on the adaptive responses of political actors to European integration. An important theme of this volume is that integration has changed the political environment in critical ways. Political actors have to come to grips with these changes. As the previous chapters show, one way they adapt is by trying to accommodate their positions vis-à-vis integration in existing cleavage structures such as the left/right dimension. However, another mode of adaptation is to manipulate the salience of European integration in a way that best suits the needs of the actor. 165

166

Marco R. Steenbergen and David J. Scott

We believe that the political parties of the EU member states have adapted to integration in part by differentially stressing it as an issue (or, perhaps better, bundle of issues). Our thesis is that variation in the issue salience of European integration across parties can be attributed to a considerable extent to the strategic behavior of those parties. Put simply, parties that stand to gain from the issue, in whatever sense, try to emphasize the issue, while parties that stand to lose try to de-emphasize it. These different strategic motivations produce cross-party variation in the salience of European integration. Our task in this chapter, then, is two-fold. First, we explore the extent to which political parties vary in their emphasis on European integration. Our assumption is that variation exists, but is this actually true? Second, we develop an explanation of the variation in the issue salience of European integration across parties. Specifically, we explore different strategic motivations that may cause parties to place more or less emphasis on European integration. The exploration of these topics can provide important insights into the nature of contestation over European integration as well as issue salience more generally. The remainder of this chapter is organized as follows. We begin by discussing a theory of issue salience in which parties are viewed as strategic actors that seek to influence salience in order to achieve certain goals (subject to constraints imposed by the political system). Next, we derive a series of testable hypotheses from this theory. We then discuss our data and present our results. We conclude by discussing the implications of our findings. An appendix includes additional results on the salience of European integration. Issue salience and strategic parties Why does one political party emphasize an issue such as European integration, while another party shuns it? In this chapter, we adopt the position, derived from salience theory (Budge et al. 2001; Budge and Farlie 1983; Klingemann, Hoffebert, and Budge 1994; Robertson 1976), that parties strategically manipulate issue salience. They do so to accomplish certain goals, which include electoral success, holding office, and maintaining party cohesion. However, the manipulation of issue salience is constrained by factors endemic to the political system in which a party operates. Let us discuss each of these theoretical assumptions in turn. Salience theory Traditional models of strategic party behavior suggest that parties compete by changing their position on issues (Downs 1957; Enelow and

The salience of European integration

167

Hinich 1984; Hinich and Munger 1997; MacDonald, Listhaug, and Rabinowitz 1991; Shepsle 1992). However, positional shifting may be very difficult to accomplish. Parties are often committed to particular issue positions; for example, because of their ideology or reputation (see Kollman, Miller, and Page 1992; Robertson 1976). If positional shifts are difficult, then parties have to look for alternative modes of competition. Salience theory states that party competition often takes the form of a battle over agenda control.1 Parties try to affect the issues that are on the agenda, selectively emphasizing those issues that are favorable to them, however this is defined, and de-emphasizing those that are unfavorable. Thus party competition involves the definition of the political space – i.e., what is salient and what is not (Budge et al. 2001; Budge and Farlie 1983; Klingemann, Hoffebert, and Budge 1994; Robertson 1976; see also Riker 1982). There is a great deal of anecdotal and systematic support for the claims of salience theory. For example, in the most recent British parliamentary elections the Conservatives unsuccessfully tried to introduce European integration as a campaign issue, in part to drive a wedge into the support for Labour and in part to cover up (or de-emphasize) the ineffective and unpopular oppositional stances of their own party. In the United States, the 1992 presidential elections serve as a good illustration of salience theory: while President George H. W. Bush tried to emphasize foreign policy, the domain of his policy successes, Bill Clinton successfully focused the campaign on the economy, where Bush was weak and the Democrats had an advantage. The manipulation of issue salience can lead to issue ownership in the long run (e.g., Petrocik 1996). That is, certain issues may become associated with a particular party (e.g., the environment with parties from the left and opposition to immigration with parties from the right). Importantly, parties can also influence issue salience to affect their short-term success. In this sense, issue salience is an adaptive mechanism that allows parties to adjust to a changing political world. Parties as seekers of multiple goals Inherent in salience theory is the assumption that political parties influence issue salience strategically, i.e., in ways that allow them to accomplish 1

In reality, party competition involves strategic choices on both issue salience and issue positions, and a fully specified analysis should treat both of these elements as endogenous. Unfortunately, it is difficult to identify such a model. Thus we shall treat a party’s EU stance as exogenous, while treating its level of EU salience as endogenous. This contrasts with conventional models of party competition, which treat issue salience as exogenous and issue positions as endogenous (Downs 1957; Enelow and Hinich 1984; Hinich and Munger 1997; Shepsle 1992).

168


certain goals. While salience theorists have typically focused on the goal of electoral success, we assume that parties seek multiple goals that may or may not be mutually compatible (see Austen-Smith and Banks 1988; Kollman, Miller, and Page 1992; Sened 1996; Strøm and Muller ¨ 1999). Specifically, we consider three sets of party goals: (1) vote-seeking, (2) office-seeking (in conjunction with policy-seeking), and (3) cohesionseeking. Vote-seeking Electoral success – be it defined as vote maximization (Downs 1957), majority-seeking (Robertson 1976), or pluralityseeking (Hinich and Ordeshook 1970) – is often viewed as the overriding objective of all political parties. According to Downs (1957), parties are, in the end, nothing but teams composed of individuals who want to win an election. Other objectives are subservient to the electoral goals of the party and, as such, are only of minor importance. Electoral considerations can cause parties to shift the emphasis on issues. If a party’s position on an issue is far removed from the electorate, as represented by the mean or median voter, then the party benefits from de-emphasizing the issue. On the other hand, if the party’s position on an issue is close to that of the electorate, then the party should emphasize the issue in order to maximize electoral success. These predictions follow from both proximity (Enelow and Hinich 1984; Hinich and Munger 1997) and directional (MacDonald, Listhaug, and Rabinowitz 1991) models of issue voting. Office-seeking One argument against the preeminence of electoral objectives is that votes ultimately mean little if political parties cannot secure political office. Parties are interested in political power and this comes with office-holding, not with votes (Laver and Schofield 1990; Riker 1962). The distinction between vote-seeking and office-seeking is especially relevant in multi-party systems where winning an election does not automatically mean securing a majority in the assembly. In many of these systems, coalitions are inevitable and electoral success does not automatically translate into participation in the government. The pursuit of political office may also produce strategic choices concerning issue salience. Parties that wish to partake in government want to avoid issues that will alienate them from potential coalition partners. By contrast, they should emphasize issues on which they have common ground with potential coalition parties. This is not to say that parties wish to downplay broad ideological differences among themselves. Theorists of coalition government have argued that parties do not just seek office; they want to accomplish policy goals


169

while in office. Consequently, not just any party will do as a potential coalition partner. On the contrary, parties tend to look for ideologically compatible partners (Axelrod 1970; De Swaan 1973; Laver and Schofield 1990). Nonetheless, within this ideologically compatible subset, parties may have an incentive to downplay specific issues that divide them. After all, if they do not do this, the subset of possible coalition parties may well become an empty set, and this would effectively exclude a party from office. Cohesion-seeking A third important goal is party cohesion. Electoral success and office-holding have little meaning if a party is about to fall apart. A particularly important concern is to keep the party activists in the fold, because they are critical to the survival of the party (Aldrich 1983; Robertson 1976). Again, the manipulation of issue salience may be an important mechanism to achieve this goal. Specifically, to the extent that certain divisive issues threaten party cohesion, the party leadership may have an incentive to downplay those issues. By contrast, the leadership may be interested in playing up those issues on which the party is united since this may build party cohesion. However, there are limitations to the strategic manipulation of issue salience in the face of internal dissent. While parties that experience modest levels of dissent may be able to de-emphasize an issue, parties experiencing major dissent may be unable to do this. Here a significant faction disagrees with the mainstream of the party on an issue. It is likely that such disagreement will inspire extensive debate within the party and, as a consequence, the divisive issue will receive a great deal of emphasis. That is, the issue becomes salient in spite of strategic party objectives (Scott 2001). Thus we expect the relationship between salience and internal dissent to be non-linear. The political environment While parties have different incentives for emphasizing an issue, they rarely have monopolistic agenda control. In the end, the salience of an issue is determined by many factors, not least of which is the behavior of other parties.2 A party can ignore those factors only at the risk of removing itself from the mainstream political debate. Thus a party that hopes to de-emphasize an issue may be unable to do so when other parties have 2

In turn, the party system may be responding to the demands of citizens or of interest groups and social movements. Additionally, the system may respond to political events and to the emphasis that an issue receives in the mass media (see Cook 1998; Kleinnijenhuis and Rietberg 1995).

170


an incentive to emphasize it. Likewise, it may be exceedingly difficult for a party to emphasize an issue if other parties refuse to do so. The political system thus constrains agenda control by a single party (Ray 1997). It is important to stress, however, that the system usually does not completely remove agenda control from a single party. At the margins, parties have considerable leverage over issue salience. They can give an issue a more or less prominent place on their platform (Budge et al. 2001; Budge and Farlie 1983). Party leaders can discuss or neglect issues during campaign speeches and interviews. Party activists may use issues to generate support, or they may ignore them. We may not find massive party differences in issue salience, but there is ample room for politically meaningful variation in salience across parties. The salience of European integration How do these theoretical ideas apply to the issue of European integration? Four testable hypotheses are suggested, all of which predict that parties adjust the salience weight they associate with the issue. The first hypothesis highlights the importance of the political environment and the constraints that it places on the salience of European integration. Hypothesis 1: The salience of European integration to a party increases as other parties in the political system emphasize this issue. We dub this the systemic salience hypothesis. This is a useful baseline hypothesis because it suggests that parties have no leeway in determining the salience of European integration as a political issue. In the analysis, we compare the performance of models that include strategic party motivations with this baseline model. The second hypothesis relates the salience of European integration to the electoral objectives of political parties. Hypothesis 2: The greater the distance between the party position on European integration and the position of voters, the lower the party’s salience weight for European integration. This hypothesis follows immediately from salience theory, which claims that parties try to de-emphasize issues that could prove costly in elections. It should be noted that there are actually two sets of voters that parties care about. First, parties want to retain their core support base. Consequently, they want to adjust the salience of European integration in response to


171

the distance on this issue between the party and its base of supporters. Second, parties want to reach out to the electorate as a whole in order to increase their vote share. Consequently, they have an incentive to adjust the salience of European integration in response to the distance on this issue between the party and the mean voter (Ray 1997). The third hypothesis stems from the view of parties as office-seekers. Hypothesis 3: When a party is ideologically compatible with other parties but holds an inconsistent view on European integration, then the party will de-emphasize this issue in order to enhance the opportunity for coalition formation. This hypothesis has a limited scope; it applies only to EU member states with coalition governments. It suggests that a party that finds an ideological match in another party has found itself a potential coalition partner. However, an incompatibility in the parties’ viewpoints concerning European integration could undermine this potential. This puts pressure on a party to de-emphasize this issue. Finally, the goal of party cohesion suggests the following hypothesis. Hypothesis 4: The greater a party’s internal division over European integration, the less salient this issue will be for the party. However, when the internal divisions become too great, then the salience of European integration will increase. De-emphasizing issues that have the potential to become explosive in the party is rational from the perspective of party maintenance. Thus we should find that internally divided parties place less emphasis on European integration than parties that are undivided over this issue (see Ray 1997; van der Eijk and Franklin 1996). However, as stated before, we expect that deeply divided parties cannot suppress the issue, so that the relationship between the salience of European integration and internal dissent is expected to be non-linear (see also Scott 2001). With respect to hypotheses 2–4 we note that the postulated salience mechanisms may be particularly powerful when a national election is imminent. During elections, political parties have to pay particularly careful attention to their objectives. This holds true of course for vote-seeking. However, it also holds true for office-seeking (the election will trigger a new coalition formation process) and for cohesion-seeking (party divisions can be especially costly during elections). Consequently, parties will be particularly responsive in adjusting the salience of European integration when an election is on the horizon.

172


Data and measures Two sources provide the data needed to test the various hypotheses. For information about party characteristics, most importantly salience, we rely on Ray’s (1999) expert survey data, which cover the years 1984, 1988, 1992, and 1996. For information about public opinion we rely on the cumulative file of the Eurobarometer survey series, focusing on the same years (Scholz and Schmitt 2001). Dependent variable Past studies of issue salience have usually relied on party manifesto data (Budge et al. 2001; Budge and Farlie 1983; Klingemann, Hoffebert, and Budge 1994). Unfortunately, these data are insufficient for our goals. Party manifestos reflect the salience of an issue such as European integration at the time of an election. However, we want to compare issue salience across election and non-election years. That is, we are interested in the question of whether the impact of party goals on the salience of European integration is greater in election than in non-election years. This requires data that are obtained before, during, and after elections. Ray’s (1999) expert survey data satisfy this requirement. Ray conducted a survey among party experts for each member state in the EU-15.3 These experts were asked to evaluate the position with respect to European integration of the political parties of a particular member state in 1984, 1988, 1992, and 1996. In addition, the experts were asked to judge “the relative importance of this issue in the party’s public stance” for the same years. These judgments were provided on a five-point scale: 1 = “European integration is of no importance, never mentioned by the party”; 2 = “European integration is a minor issue for the party”; 3 = “European integration is an important issue for the party”; 4 = “European integration is one of the most important issues for the party”; and 5 = “European integration is the most important issue for the party.” As a measure of the issue salience of European integration we take the average rating of a party across experts.4 Descriptive statistics for our salience measure are provided in table 8.1. 3 4

The number of experts per member state ranged between five and thirteen, with an average of 8.2 experts per country (Ray 1999). The Spearman rank correlation between the mean expert salience ratings and a manifestobased measure of salience is 0.312 ( p < 0.01). While this is not impressive, it suggests that the expert measure moves in the same direction as the manifesto-based measure. The latter measure is defined as the sum of the percentages of positive and negative statements about European integration in a party’s manifesto (see Budge et al. 2001).


173

Table 8.1 Descriptive statistics Variable

Mean

S.D.

Minimum

Maximum

N

Salience Systemic salience Distance from mean supporter Distance from national mean Mean distance from coalition parties Dissent

3.04 3.04 0.74 1.30 4.32 1.74

0.67 0.45 1.30 1.85 7.23 0.57

1.00 1.74 0.00 0.00 0.00 1.00

5.00 4.04 9.80 4.00 27.04 4.00

586 586 387 542 64 583

Note: The descriptive statistics are based on pooled data for 1984, 1988, 1992, and 1996.

Predictors There are four key predictors: systemic salience, vote-seeking, officeseeking, and cohesion-seeking. Systemic salience is measured as the mean salience score of all of the parties in a member state, excluding the party under consideration. For instance, given three parties, A, B, and C, the systemic salience level of party A is defined as the mean salience level of parties B and C.5 Electoral considerations are captured via two predictors. The first of these predictors measures the discrepancy between the stance on European integration of a party and the mean stance on this issue of party supporters. This predictor is called distance from mean supporter. The integration stances of the parties are obtained from Ray’s expert survey. This survey asked the experts to evaluate “the overall orientation of the party leadership towards European integration.” The response scale ran from (1) “strongly opposed to European integration” to (7) “strongly in favor of European integration.” The mean expert rating of a party is taken as the indicator of the party’s stance on European integration. The integration stance of supporters is measured as the mean response to the following question in the Eurobarometer surveys: “Generally speaking, do you think that [your country’s] membership in the European Community (Common Market) is . . . a good thing, a bad thing, neither good 5

We exclude A’s salience level from the computation of systemic salience in order to avoid endogeneity problems. Our measure of systemic salience implies that parties B and C hold equal control over the political agenda. While this assumption has the benefit of simplicity, it may not be reasonable if the parties are considerably different in size: for example, if B attracts many more voters than C. To address this problem, we also performed our analyses using a vote-weighted measure of systemic salience. In this measure, parties influence systemic salience in proportion to the percentage of the vote that they gathered in the previous national election. Inclusion of this measure does not significantly alter the results (which are available upon request).

174


nor bad.”6 We count as party supporters anyone of voting age that indicated they would likely vote for that party if elections were held tomorrow. Since the response scales for the parties and party supporters are different, we standardized them first. We then subtracted the mean integration stance of the party supporters from the integration stance of the party and squared the result, thus creating a Euclidean distance metric. Our second electoral predictor measures the discrepancy between the integration stance of a party and the mean integration stance in the mass public, i.e., all voting-age Eurobarometer respondents. For this measure, we trichotomize the party positions into pro-integration, anti-integration, and neutral positions. The same was done for the mean integration stance in the mass public.7 These trichotomized measures were subtracted and then squared to create a Euclidian distance metric. This measure is called distance from national mean. The objective of office-seeking is captured through the following threestep approach. First, we selected potential coalition partners for each party in the subset of EU member states that have coalition government. The set of potential coalition partners was defined as the smallest set of parties that, along with the party of interest, will carry a majority in the legislature based on the most recent election results. Second, we narrowed down the set of coalition partners by considering their ideological proximity to a party. Using party ideology measures for 1984 (Castles and Mair 1984) and 1996 (Hix and Lord 1997), we defined the average ideological distance between a party and all of the partners that would produce a majority government. We then selected as the party’s preferred coalition those partners that were ideologically most proximate (while still generating a majority coalition). Finally, we calculated the squared difference in the European integration stance of this party and the coalition partners, where the integration stances come from Ray’s (1999) expert survey. The resulting variable is called mean distance from coalition parties. Note that the limited availability of ideological ratings of parties means that we only consider the impact of office-seeking on salience for two years (1984 and 1996).8 6

7

8

This is the only measure of support for European integration in the Eurobarometer surveys that is available for the entire time period (1984, 1988, 1992, and 1996). Despite its limited scope, this measure has proved to be one of the best indicators of support for integration policies (Gabel 1998b). We used trichotomization to obtain comparable metrics for the integration stances of the mass public and the parties. (This could not be done via standardization because there is no cross-party variance in the mass public’s integration stance, so that z-scores are undefined.) While the trichotomization can be carried out in different ways, it does not matter much for the results how this is done. If ideological ratings were missing for large parties (i.e., those controlling ten seats or more), then we excluded the entire country from the computation of the office-seeking


175

Finally, party cohesion considerations are captured through another measure from Ray’s (1999) expert survey. The experts rated “the degree of dissent within the party over the party leadership’s position [on European integration].” The response scale included five categories: 1 = “complete unity”; 2 = “minor dissent”; 3 = “significant dissent”; 4 = “party evenly split on issue”; and 5 = “leadership position opposed by a majority of the party activists.” We use the mean expert rating as an indicator of a party’s internal dissent over European integration. We also include the square of this rating in order to capture the possible non-linear effect of dissent on salience. Descriptive statistics for the different predictors can be found in table 8.1. Results We discuss the results in two parts. First, we consider a series of descriptive statistics in order to demonstrate that significant variation in the salience of European integration exists at the national party level. Next, we discuss whether the four hypotheses that we developed can account for this variation. Since our hypotheses are all directional in nature, we will be using one-tailed tests for this purpose. Salience patterns Do national political parties differ in the salience they attach to European integration? The answer to this question is affirmative. As table 8.1 shows, there is considerable variance in the salience ratings for European integration. The minimum and maximum reveal that we can find political parties across the entire salience scale, with some placing no emphasis on European integration and others viewing it as a critical issue. It is instructive to show how salience varies over time, across party families, and across member states. Figure 8.1 depicts the mean salience level of European integration across parties in each of the four years that we analyze. The figure also depicts a bar that stretches 1.4 standard deviations on each side of the mean.9 This error bar graph shows that the mean salience level of European integration is around 3. This means that the issue is an important one for the average party in all four years, although there is a slight upward trend in the salience until 1996 (this trend is statistically significant: F [3, 582] = 17.15, p < 0.05). Importantly, however,

9

measure. This measure was computed only if a party controlled at least one seat in the assembly. For between-group (e.g., between-year) comparisons setting the error bars at ±1.4 standard deviations is preferable to the conventional ±1.96 standard deviations (Goldstein and Healy 1995).

176

Marco R. Steenbergen and David J. Scott 5

Salience

4

3

2

1 84

88

92

96

Year Figure 8.1 The salience of European integration over time. (Note: The black dots indicate the means. The bars around these dots represent the mean plus or minus 1.4 standard deviations.)

the error bars show considerable variation in the salience scores across parties, with salience scores spanning almost the entire response scale in all years. Figure 8.1 also shows that there is no evidence of declining variation in the salience of European integration over time. Given the monumental changes in the EU, one might expect that over time all parties would have come to consider European integration an important issue. However, this is not what we find. Indeed, even in 1996 – well into the post-Maastricht era – many parties considered integration at best an issue of minor importance. These parties stand in stark contrast with other parties that viewed (and continued to view) integration as a major issue. The salience of European integration varies in interesting ways across party families, as figure 8.2 illustrates. On average, the salience was lowest for parties that operate at the extremes of the political spectrum: M = 2.83 for the greens, the extreme right, and the extreme left; M = 3.14


177

5

Salience

4

3

2

1 Conservatives

Liberals

Christian democrats

Social democrats

Regional

Extreme left

Extreme right

Greens

Party family Figure 8.2 The salience of European integration by party family. (Note: The black dots indicate the means. The bars around these dots represent the mean plus or minus 1.4 standard deviations.)

for the regional parties, the social democrats, the Christian democrats, the liberals, and the conservatives (F [1, 584] = 29.67, p = 0.01). However, while these differences are important, they are overshadowed by the enormous differences in salience that exist within party families. In each of these families one can find parties that give relatively little weight to European integration, but one can also find parties that place a great deal of weight on this issue. Consequently, the cross-family overlap between the error bars in figure 8.2 is quite extensive. A similar story can be told for cross-national differences in the salience of European integration. Figure 8.3 shows the mean salience scores in the EU-15. It is clear from this figure that the member states vary in how

178

Marco R. Steenbergen and David J. Scott 5

Salience

4

3

2

1 NE FI

BE

IR UK IT

GE SW FR LU PO AU SP DK GR

Country Figure 8.3 The salience of European integration by member state. (Notes: The black dots indicate the means. The bars around these dots represent the mean plus or minus 1.4 standard deviations. NE = the Netherlands, BE = Belgium, FI = Finland, IT = Italy, IR = Ireland, UK = the United Kingdom, GE = Germany, LU = Luxembourg, SW = Sweden, FR = France, PO = Portugal, AU = Austria, SP = Spain, GR = Greece, and DK = Denmark.)

much salience their parties give to European integration (these differences are statistically significant: F [14, 571] = 10.49, p < 0.01). However, these cross-national differences again pale in comparison with the within-country differences in the salience of European integration. While in some countries there exists considerable agreement among the parties in their emphasis on this issue (e.g., Luxembourg and the Netherlands), the more typical situation is that a country’s parties differ greatly in their salience scores (e.g., Austria, Finland, Germany, and Sweden). The overall picture, then, is one that shows a great deal of cross-party variation in the salience of European integration. For some parties this issue is only of minor importance, while for other parties it is a major issue or even the most important issue. The question is now whether


179

Table 8.2 Salience and the political environment Predictor

1984

1988

1992

1996

Systemic salience

0.867∗∗ (0.043) 0.370∗∗ (0.127)

0.828∗∗ (0.073) 0.499∗ (0.222)

0.710∗∗ (0.101) 0.935∗∗ (0.334)

0.627∗∗ (0.148) 1.120∗ (0.488)

0.385 141

0.308 146

0.158 148

0.102 151

Constant R2 N

Notes: Table entries are OLS estimates with robust, cluster-corrected standard errors in parentheses. ∗∗ p < 0.01, ∗ p < 0.05 (one-tailed tests).

this variation in salience can be accounted for in terms of our salience theoretic framework. We turn to this question next. Explaining salience Systemic salience We begin our exploration of the determinants of the issue salience of European integration by considering the political environment in which a party operates. Our systemic salience hypothesis states that the emphasis that a party places on European integration is to some extent a function of the salience of this issue to other parties. This hypothesis is supported by the data. Table 8.2 shows the results of a regression analysis of the party salience scores on systemic salience.10 As predicted, the effect of systemic salience is statistically significant and positive: when other parties place greater emphasis on European integration, then the party under consideration tends to follow suit. Thus parties appear to be somewhat constrained in determining the emphasis they want to place on the issue. This constraint is far from perfect, however. The results in table 8.2 reveal that systemic salience explains only a small portion of the variance in the party salience scores. Moreover, this explained variance has declined over time. In 1984, systemic salience still explained about 39 percent 10

The analysis was conducted using ordinary least squares (OLS). Because parties are clustered inside the EU member states, the standard errors were corrected to eliminate bias (Lee, Forthofer, and Lorimor 1989). This correction is particularly important when the intra-class correlation is positive, as is likely to be the case here. In this situation, the OLS standard errors are underestimated, which makes it too easy to find statistical significance.

180


Table 8.3 Salience and vote-seeking – model without election effects Predictor Distance from mean party supporter Distance from national mean Systemic salience Constant R2 Baseline R 2 N

1984

1988

1992

1996

0.030 (0.021) −0.085+ (0.050) 0.832∗∗ (0.119) 0.633+ (0.371)

0.008 (0.049) −0.095∗ (0.040) 0.869∗∗ (0.089) 0.567∗ (0.295)

−0.060∗∗ (0.019) −0.099∗ (0.043) 0.729∗∗ (0.125) 1.053∗ (0.426)

−0.063∗ (0.025) −0.100∗∗ (0.031) 0.761∗∗ (0.146) 1.020∗ (0.485)

0.299 0.231

0.349 0.267

0.252 0.125

0.304 0.154

72

90

109

116

Notes: Table entries are OLS estimates with robust, cluster-corrected standard errors in parentheses. The baseline R 2 is for a model including systemic salience only. ∗∗ p < 0.01, ∗ p < 0.05, +p < 0.10 (one-tailed tests).

of the variance, but by 1996 this percentage had declined to about 10 percent.11 This leaves considerable room for other factors to affect salience, a result that is consistent with our prediction that the political environment can only imperfectly account for the salience of European integration to a party. Vote-seeking One of the other factors that may explain salience is vote-seeking. Table 8.3 shows the effect of this predictor and reveals several interesting patterns. First, electoral considerations appear to have an effect on the salience of European integration for parties. This is particularly true when we consider the distance from the national mean, which has a statistically significant effect (at the 0.05 level or better) in three of the four years. The distance from the mean party supporter is significant at the 0.05 level or better in only two years. Overall, however, electoral considerations appear to have an effect on salience. This effect is in the expected direction: as a party’s stance on European integration is further removed from the electorate, the party tends to place less emphasis on this issue. Moreover, the effect improves the explanatory power of the model, which is evident when we compare the R 2 to a baseline R 2 , 11

We can rule out another explanation for the declining R 2 . It is well known that explained variance can decrease as the variance in the dependent variable or a predictor decreases (Achen 1991). Neither of these explanations holds true in the present situation.


181

pertaining to a model that includes systemic salience only (for the same set of parties). Second, there is a clear breaking point in the time series. Up to 1992, the electoral effects tend to be relatively weak, with one of the measures never reaching statistical significance at conventional levels. This is the pre-Maastricht era, and we expect that many citizens perceived European integration to be a remote issue at best. This meant that parties did not have to worry too much about the electoral fallout of their stance toward European integration. By 1992, however, things had changed. The Maastricht Treaty spelled important changes in the European Union, which received much discussion in the mass media and drew the attention of citizens. Moreover, the referenda in Denmark and elsewhere demonstrated to political elites that public opinion on European integration had to be reckoned with. It is not surprising, then, to find that salience began to respond to electoral considerations. In 1992 and 1996, salience appears to have responded to a party’s distance from both the national mean and the mean supporter. This suggests that parties placed greater emphasis on European integration if their position was in line with the position of the mass public and that of the core supporters. In contrast, parties with an unpopular stance on European integration tended to downplay this issue. The results in table 8.3 do not distinguish between parties facing imminent elections and other parties. For obvious reasons, electoral concerns weigh more heavily during elections than at other times. This may affect the relationship between these considerations and the salience of European integration, and it may qualify the conclusions drawn above. To explore the moderating effect of elections, we created interaction terms between the two electoral variables and an election dummy variable that measures the proximity of an election.12 Since the sample sizes in each of the years are somewhat small, we estimated the interaction effects in a pooled analysis that also includes systemic salience and dummy variables for years as controls (using 1984 as the baseline). The results are shown in table 8.4. As this table indicates, the presence of an election does not significantly alter the impact of the distance from the national mean on salience (the interaction with the election dummy is not statistically significant). Apparently, political parties pay attention to national public opinion regardless of whether or not an election is on the horizon. However, the interaction between the election dummy and a party’s distance from the mean supporter is statistically significant. When an election is 12

An election was considered proximate or imminent if it took place in the same year as the salience measure was taken, or in the next year.

182


Table 8.4 Salience and vote-seeking – pooled model with election effects Predictor Systemic salience Distance from mean party supporter Distance from national mean Distance from mean party supporter × election Distance from national mean × election Election 1988 1992 1996 Constant R2 Baseline R 2 N

Estimate 0.746∗∗ (0.081) −0.024 (0.024) −0.105∗∗ (0.032) −0.085∗ (0.040) 0.043 (0.048) 0.034 (0.054) 0.030 (0.041) 0.068 (0.056) 0.131∗ (0.069) 0.913∗∗ (0.248) 0.337 0.223 387

Notes: Table entries are OLS estimates with robust, cluster-corrected standard errors in parentheses. The baseline R 2 is based on a model with systemic salience and the year dummy variables only. ∗∗ p < 0.01, ∗ p < 0.05 (one-tailed tests).

not imminent, parties do not appear to adjust issue salience in response to their supporters (b = −0.024, not significant). When facing an election, however, parties do pay attention to their supporters and lower the salience of European integration the further they are removed from their support base (b = −0.109, p < 0.01). The results on vote-seeking can help us make sense of the descriptive statistics that we reported earlier. One of the striking differences that we noted was that between the parties at the extremes of the ideological spectrum – the greens, the extreme left, and the extreme right – and those located more at the center. The extreme parties score lowest in the


183

Table 8.5 Salience and office-seeking – model without election effects Predictor

1984

Mean distance to potential coalition parties

0.029 (0.072) 0.962∗ (0.266) 0.205 (0.787)

−0.005 (0.011) 0.956∗∗ (0.192) 0.380 (0.629)

0.319 0.290

0.285 0.282

17

47

Systemic salience Constant R2 Baseline R2 N

1996

Notes: Table entries are OLS estimates with robust, cluster-corrected standard errors in parentheses. The baseline R2 is for a model including systemic salience only. ∗∗ p < 0.01, ∗ p < 0.05 (one-tailed tests).

salience they attach to European integration, but why is this? It turns out that these parties embrace integration stances that are much further removed from the national electorate than the stances of the other parties. The average distance from the national mean is 2.49 for the extreme parties. In contrast, the average distance from the national mean is only 0.67 for the other parties. This is a significant difference (F [1, 540] = 152.08, p < 0.01). In light of the coefficients in table 8.4, it suggests a difference in salience of about 0.19 points between the extreme parties and all other parties. This goes a long way in explaining the observed difference between the parties.13 Office-seeking Electoral success is not the only motive of political parties. Another important motive is political office, and in parliamentary systems this typically entails coalition-building. Potential coalition parties may have differences of opinion over the issue of European integration, even if on the whole the ideological stances of these parties are compatible. How does this affect the salience of integration for such parties? Table 8.5 provides the answer to this question for 1984 and 1996, the two years for which we have party ideology scores that allowed us to 13

Another contributing factor is that the systemic salience is higher for the extreme parties: M = 3.12 for the greens, extreme left, and extreme right; M = 3.00 for all other parties (F[1, 584] = 10.10, p < 0.01).

184


Table 8.6 Salience and office-seeking – pooled model with election effects Predictor Mean distance to potential coalition parties Mean distance to potential coalition parties × election Systemic salience Election 1996 Constant R2 Baseline R 2 N

Estimate −0.001 (0.012) −0.067∗∗ (0.012) 0.900∗∗ (0.150) 0.266∗ (0.096) 0.104 (0.173) 0.419 (0.367) 0.498 0.492 64

Notes: Table entries are OLS estimates with robust, cluster-corrected standard errors in parentheses. The baseline R 2 is for a model that includes systemic salience and the year dummy variable only. ∗∗ p < 0.01, ∗ p < 0.05 (one-tailed tests).

designate potential coalition partners.14 This table shows that discrepancies in the European integration stances of potential coalition parties did not significantly affect the salience of integration in 1984 or 1996. This suggests that office-seeking was not a relevant consideration in the determination of issue salience. As in the case of electoral goals, it may be that coalition goals become salient to a party only when an election is on the horizon. Thus we reestimated the model including an interaction between a dummy measuring the imminence of an election and the measure of the distance in the party positions on European integration. As before, we pooled the data to increase the statistical power. The results of this analysis can be found in table 8.6. These results are consistent with our expectations. When the next election is still far away, coalition considerations do not influence the salience that a party attaches to European integration (b = −0.001, not significant). However, if an election is near, then parties 14

This analysis is restricted to EU member states that have coalition governments, because a distinction between vote-seeking and office-seeking is difficult to make otherwise. This excludes Greece and the United Kingdom from the analysis.


185

Table 8.7 Salience and cohesion-seeking – model without election effects Predictor Dissent Dissent2 Systemic salience Constant R2 Baseline R 2 N

1984

1988

1992

1996

0.074 (0.543) −0.073 (0.128) 0.915∗∗ (0.052) 0.338 (0.522)

−0.029 (0.529) −0.030 (0.120) 0.851∗∗ (0.083) 0.566 (0.550)

−0.982∗∗ (0.327) 0.211∗∗ (0.069) 0.670∗∗ (0.131) 2.084∗∗ (0.589)

−0.752∗ (0.407) 0.165∗ (0.083) 0.569∗∗ (0.170) 2.148∗∗ (0.672)

0.412 0.394

0.318 0.308

0.194 0.157

0.123 0.098

140

146

147

150

Notes: Table entries are OLS estimates with robust, cluster-corrected standard errors in parentheses. The baseline R 2 is for a model that includes systemic salience only. ∗∗ p < 0.01, ∗ p < 0.05 (one-tailed tests).

tend to pay closer attention to the difference between their own stance toward European integration and that of potential coalition parties, adjusting the issue salience in response to this (b = −0.068, p < 0.01). Cohesion-seeking European integration has proved to be a remarkably divisive issue for parties (Hix 1999a; Hix and Lord 1997; van der Eijk and Franklin 1996). This may provide a strong incentive for parties not to emphasize this issue (Ray 1997; van der Eijk and Franklin 1996). After all, emphasizing one’s internal divisions on an issue could cause the breakup of the party. At the same time, severe intra-party conflicts over European integration may make it impossible for the party to de-emphasize the issue. Thus, we expect a non-linear relationship between internal dissent and the salience of European integration for a party. The empirical evidence presented in table 8.7 suggests that internal dissent matters only in two years, 1992 and 1996. In the era before the Maastricht Treaty, internal party divisions appear to have been no reason for parties to de-emphasize European integration as an issue. However, in the period following the Maastricht Treaty, internal dissent had a prominent effect on salience. Moreover, this effect was non-linear, as we had expected. Figure 8.4 illustrates the relationship between internal dissent and salience; this figure is based on the estimates in a combined analysis

186


3.6

Predicted salience

3.5

3.4

3.3

3.2

3.1

1.0

1.5

2.0

2.5 Dissent

3.0

3.5

4.0

Figure 8.4 The relationship between salience and dissent (1992–6). (Note: The line shows the predicted relationship between internal dissent and salience, assuming that systemic salience is at the mean (3.22).)

of 1992 and 1996.15 It is clear that modest levels of dissent are associated with a decreased salience of European integration. Higher levels of dissent, by contrast, are associated with a greater salience of this issue. The tipping point is around 2.3, which lies between minor dissent and significant dissent. This suggests that parties may be unsuccessful at de-emphasizing the issue at all but the slightest levels of dissent. Elections moderate the impact of dissent on salience. Table 8.8 presents a pooled analysis that includes interactions between dissent, its square, and an election dummy that indicates the presence of an imminent election. When elections are far away, there is no significant effect of dissent or the square of dissent. However, with elections approaching, the effect of dissent is statistically significant: b = −1.736, p < 0.01. Likewise, the effect of the squared dissent term is statistically significant: b = 0.377, p < 0.01. This means that parties try to de-emphasize European integration when there is dissent over this issue and an election is near. However, 15

This analysis includes dissent, the square of dissent, systemic salience, and a year dummy variable for 1996. The figure assumes that the year is 1996 and that systemic salience is at the mean (3.22).


187

Table 8.8 Salience and cohesion-seeking – pooled model with election effects Predictor Dissent Dissent2 Dissent × election Dissent2 × election Election Systemic salience 1988 1992 1996 Constant R2 Baseline R 2 N

Estimate −0.378 (0.324) 0.079 (0.066) −1.358∗∗ (0.386) 0.298∗∗ (0.100) 1.412∗∗ (0.349) 0.780∗∗ (0.069) 0.021+ (0.016) 0.115∗∗ (0.038) 0.086∗ (0.039) 1.003∗∗ (0.356) 0.330 0.307 583

Notes: Table entries are OLS estimates with robust, clustercorrected standard errors in parentheses. The baseline R 2 is for a model that includes systemic salience and the dummy variables for years only. ∗∗ p < 0.01, ∗ p < 0.05, + p < 0.10 (one-tailed tests).

it is also clear that significant levels of dissent will cause the issue to become salient at the time of an election. Discussion What do these results tell us about the salience of European integration at the national party level? With respect to hypothesis 1 (the systemic salience hypothesis), the results suggest that the salience of integration is determined in an important way by the weight that other parties in a country give to this issue. Systemic salience is the only predictor that consistently attains statistical significance in the analyses, and its effect is always in the expected direction. Parties cannot decide on

188


issue salience willy-nilly. Lest they find themselves marginalized from the mainstream national political debate, they will have to consider the importance that other parties attach to the issue of European integration. Systemic salience thus serves as an important constraint on parties. This constraint is far from perfect, however. It leaves considerable leeway for parties to fine-tune the emphasis they want to place on European integration. The cross-party variance in the salience of integration is impressive. Our analyses show that this variance is not random noise. On the contrary, we find that it can be accounted for in terms of party objectives. First, it appears that parties adjust the salience of European integration to meet electoral objectives, especially in the post-Maastricht era. Thus there is considerable support for hypothesis 2. Second, parties appear to be responsive to party cohesion needs: modest levels of internal dissent significantly reduce the salience of European integration. However, the effect of dissent is non-linear, and at higher levels of dissent the salience of European integration actually increases. Thus, hypothesis 4 also receives support, at least for two of the years (1992 and 1996). Finally, officeseeking does not appear to be a significant influence on salience. Indeed, there is evidence of an effect only when an election is imminent. Thus, hypothesis 3 receives the weakest support.16 It is important to note that the party objectives are frequently at loggerheads. Parties may receive electoral gains from emphasizing the issue of European integration, but they may suffer in terms of party cohesion or office-seeking. This creates difficult decision problems for parties that are not unique to the issue of integration (Strøm and Muller ¨ 1999) but have proved to be particularly challenging in this context. The longitudinal patterns in the data suggest that these difficult choices will not become any easier in the future. As European integration has become more politicized (Ray 1997), our analyses have shown that more of the party objectives come into play in determining the salience of integration to a party. Thus, by 1996, parties – at least those facing an election – reacted to electoral concerns and party cohesion concerns (and possibly coalition concerns) in deciding whether to emphasize the issue of integration. Balancing these different needs is a formidable task, which may not always prove possible. This became evident in the 2001 British elections, when the Conservative Party played up European integration to cater to a Euroskeptic British electorate. The strategy failed to pay electoral dividends and once again stirred up the internal debate in the party over 16

Of course we should also point out that the sample sizes are quite small when we analyze the implications of office-seeking for salience. This lowers our statistical ability to find an effect.


189

European integration, contributing to a divisive leadership election after the general election (Garnett and Lynch 2002; Kelly 2002). Conclusion In this chapter, we have analyzed the salience of European integration for the political parties of the EU member states. We have observed crossparty variance in issue salience and we have offered several systematic explanations for this variance. Our results speak to a number of important questions. Contestation potential The first question concerns the potential for contestation over European integration. Van der Eijk and Franklin (1996; this volume) have noted that European elections are seldom about European integration. One can speculate about the causes of this, but at least one possibility is that political parties do not want to discuss European politics; they would rather make European elections a forum about domestic issues. While we have not studied European political parties, our findings concerning national parties can shed important light on this speculation. These findings suggest a dual answer. On the one hand, we have found ample evidence that certain parties downplay European integration. These tend to be parties that take an unpopular stance on this issue or experience modest amounts of internal dissent. These parties would rather discuss other political issues, as they may fear the consequences of discussing integration politics. And if their number accumulates, then European integration may become a “forgotten” issue in an entire nation. There is a countervailing force, however, in the form of parties that stand to gain from addressing European integration. These parties will try to keep the issue alive and on the political agenda. They want elections, including those at the European level, to be about integration politics. In sum, these parties favor contestation of the EU. Until 1996, the number of such parties did not grow much. The number of parties that, according to our data, consider integration to be one of the most important issues or the most important issue was twentyeight in 1984, twenty-six in 1988, thirty-one in 1992, and thirty in 1996. However, the number of parties that consider integration to be of no importance or only a minor issue has definitely declined: this number was twenty-four in 1984, seventeen in 1988, seven in 1992, and eight in 1996. This suggests that the contestation potential has grown – more

190


parties consider European integration to be a salient issue, as is evident from figure 8.1. We can only speculate about what this will mean for the content of future elections. The issue of European integration has to compete with many other issues, which may be even more important than integration.17 Even parties that would like to discuss integration during an election campaign may find themselves discussing other things much more frequently. Such parties may have to draw connections between those other issues and the EU in order to make integration a component of the campaign. This task should become easier every day, as the EU impinges on more and more aspects of politics, including those issues that have traditionally been labeled “domestic.” Strategic salience While our focus on the EU is important, it is possible to consider more broadly the question of how parties determine the salience of a political issue. That is, we can think of Europe as a “laboratory” (van der Eijk and Franklin 1996) in which we can test general theories about politics. What makes this “laboratory” so effective is that the EU is an evolving polity – we are in the luxurious position of testing theories in a dynamic context in which many fundamental variables are continuously changing. What have we learnt about issue salience in this laboratory? We know that salience is not determined completely by exogenous factors such as the developments in the EU. While these factors are important, political parties can deviate from the trends that these factors set. They do so by acting strategically, positioning themselves above the salience trend when the issue is advantageous to them, and positioning themselves below this trend when the issue is disadvantageous. This finding is consistent with salience theory (Budge and Farlie 1983; Klingemann, Hoffebert, and Budge 1994; Robertson 1976). However, we analyzed salience both during and after elections, whereas salience theorists have typically focused on elections. Moreover, we considered a larger set of motives than has typically been considered by salience theorists, focusing not only on electoral considerations but also on party cohesion and political office. 17

In this connection, we should point out a limitation of our study. We have looked at the salience of European integration by itself, not in a broader multi-issue context, in which issues compete for attention (see Jones 1994).


191

Depending on how these objectives are aligned, parties may have very different incentives for raising an issue on the political agenda. The question of whether or not to raise an issue looms large in party politics. Any account of contestation should consider the way in which parties answer that question. Appendix: An updated analysis Ray’s (1999) expert survey data end in 1996. The question is, what has happened to the salience of European integration since that time? We can shed light on this issue by considering data from an expert survey that was conducted in 1999 (Marks, Steenbergen, and Hooghe 1999). Unfortunately, this survey used a slightly different question format for salience, which is the reason we did not include it in the previous analyses. Experts in the 1999 survey were asked to indicate the salience of European integration for a party using a four-point scale, with response categories 1 = “European integration is of no importance at all,” 2 = “European integration is of little importance,” 3 = “European integration is of some importance,” and 4 = “European integration is of great importance.” As before, we use the mean expert rating as an indicator of salience. We explain this indicator as a function of systemic salience, vote-seeking, office-seeking, and cohesion-seeking. These predictors are included one at a time and also collectively. Table 8.9 displays the results from this analysis. Model 1 includes the electoral variables along with systemic salience. We see that the distance from the national mean is a significant predictor of salience, which replicates our earlier results. Model 2 focuses on the distance from the mean coalition party as a predictor of salience. This predictor has a significant effect on salience: the further a party’s stance on European integration is removed from potential coalition parties, the less likely the party is to emphasize this issue. In model 3, the impact of dissent is considered. This variable does not have a significant effect on the salience of European integration, at least not when it is entered only with systemic salience. In model 4, the joint effect of different party goals is considered. The effect of the distance from the national mean remains strong and significant. By contrast, office-seeking drops out as a significant predictor. Apparently, this consideration is of little relevance to parties when they decide on the salience of European integration once other goals are taken into account. Dissent is statistically significant in model 4 and has an effect that is consistent with our hypothesis.

192


Table 8.9 Salience and party goals in 1999 Predictor Distance from mean supporter Distance from national mean

Model 1 −0.018 (0.032) −0.119∗∗ (0.044)

Mean distance from coalition parties

Model 2

R2 Baseline R 2 N

−0.027 (0.044) −0.163∗∗ (0.059) −0.000 (0.013) −1.033∗ (0.442) 0.189∗ (0.099) 1.056∗∗ (0.166) 1.285∗ (0.616)

0.798∗∗ (0.136) 0.787∗ (0.419)

0.978∗∗ (0.119) 0.252 (0.321)

−0.274 (0.571) 0.050 (0.130) 0.768∗∗ (0.084) 1.002+ (0.637)

0.324 0.187

0.336 0.254

0.189 0.179

0.551 0.278

115

97

148

90

Dissent2

Constant

Model 4

−0.028∗ (0.013)

Dissent

Systemic salience

Model 3

Notes: Table entries are OLS estimates with robust, cluster-corrected standard errors in parentheses. The baseline R 2 is for a model that includes systemic salience only. ∗∗ p < 0.01, ∗ p < 0.05, + p < 0.10 (one-tailed tests).

On the whole, the results in table 8.9 suggest that salience behaved more or less the same in 1999 as it had in 1992 and 1996. That is, it is responsive to electoral considerations and to internal dissent. Thus strategic considerations continued to influence dissent in 1999, as our theory would expect.

Part III

Groups

9

Contestation potential of interest groups in the EU: emergence, structure, and political alliances Bernhard Wessels

According to Dahl (1971: 6), two central dimensions characterize democracy: the right to public contestation and the right to participate. Conceptually, these dimensions are independent. Empirically, they are connected. Political competition combined with broad participation, i.e., competition and inclusiveness together, determine the degree to which public contestation is possible (Dahl 1971: 20). With the project of European integration, a political order emerged at the supranational level for which this was – and still is – a critical issue. Discussion about the democratic deficit indicates that the Euro-political system continues to lack some central elements essential for democratic processes. This topic has received much scholarly attention. There is a vast literature on European elections (van der Eijk and Franklin 1996), the European party system (Hix and Lord 1998; Hix 1996), the working of European institutions (Rometsch and Wessels 1996), political representation (Schmitt and Thomassen 1999), and decision rules (Konig ¨ 1997). In contrast to research on parties, voluntary associations and interest groups have been researched less, in particular with regard to the democratic question and empirical analyses. Early exceptions are the studies of Kirchner and Schwaiger (Kirchner 1978; Kirchner and Schwaiger 1981); more recent ones are Greenwood, Grote, and Ronit (1992) and the works of Kohler-Koch, but the question of the role of interest groups for contestation is not their topic. Here I wish to draw attention to this question and in particular to the dimension of inclusiveness of the EU system beyond electoral and formalized participation, i.e., interest intermediation in its broad sense. In terms of democratic process, I am concerned with what Rokkan called the second tier (Rokkan 1966) and Offe the second circuit (Offe 1981) of political representation, i.e., the role of interest groups in the political process. 195

196

Bernhard Wessels

This chapter has three objectives. First, it outlines and tests a model of the genesis of interest groups at the European level. The crucial question here is whether the emergence of the European system of interest groups follows the patterns of anticipation or reaction in the context of statebuilding. Both hypotheses can be found in the literature. From the underlying pattern one might speculate about the character of contestation of interest groups in the EU context. Second, some hypotheses in analogy to the population ecology approach will be tested. The major question here is the degree to which there is imbalance in interest group participation across different countries. This leads to the question of encompassiveness and sectoral as well as country-based fragmentation of the European system of interest groups. Third, the question of contestation will be addressed. Here I begin by investigating alliances between interest groups and political parties at the European and national levels. Some conclusions can be drawn from these findings for the nature of cleavages and thus the type of contestation to be expected from interest group participation. The major question is whether national cleavage systems are reproduced at the European level or whether the supranational character of politics produces a different cleavage structure. I then address the question of whether, when, and along which dimensions there is struggle over European integration. Finally, I draw some tentative conclusions about contestation in the EU. My analyses are primarily based on a data set encompassing almost 900 European-level interest groups – self-registered at the European Commission – and their roughly 8,800 national members. Interest groups and political contestation Traditional pluralistic approaches emphasize that authoritative decisionmaking is determined by interest group competition (Bentley 1908; Truman 1951). However, this view neglects at least partly the importance of the state for the genesis and role of interest groups. Representing a modified approach to pluralism, Fraenkel has pointed to the fact that the state cannot be subsumed under the competition model of pluralism. Monopoly of legitimate power gives the state a special role (Fraenkel 1968: 45–6). Interest groups act vis-à-vis the state. One can go further and argue that many interest groups came into existence only because there is a state. The contrast between functional and territorial representation bears directly on interest group contestation. Interest groups are generally defined as organizations separate from government which seek to influence public policy. They link the state and major social interests (Wilson 1990: 1). This is why many authors have observed that interest intermediation adds

Contestation potential of interest groups

197

a “second circuit” to the “machinery of the democratic representative polity” (Offe 1981: 141). In 1966, Stein Rokkan described Norway as a “two-tier system” in which “votes count but resources decide” – with the vote potential constituting “only one among many different power resources in . . . bargaining processes” and associations being part of them (Rokkan 1966: 105). Such “corporate pluralism” limits the force of plurality representation. But it complements more than limits the “system of public contestation” in representative institutions (Dahl 1971: 11). Given that interest groups are concerned with functional representation, one might argue that they are driven by specific rather than general demands. That is, they are concerned with specific legislation and policies affecting their interests and those of their clientele. In terms of cleavages, they represent particular interests on a conflict dimension, whereas political parties structure conflict dimensions. But that is not to say that interest groups do not influence the political order. If, for example, they demand regulations guaranteeing equal competition between economic actors, this does affect the order. However, this is rather a matter of operationalizing general principles and applying them to institutional settings. In the context of European integration, this might be different. According to Hix and Lord (1997: 209), the cleavage system in the European Parliament is defined by the traditional left/right dimension alongside an anti- and pro-integration dimension. Hooghe and Marks (1999) explicitly argue that contestation in the EU is characterized by social democracy vs. market liberalism and nationalism vs. supranationalism. The European party system appears to be embedded in a two-dimensional cleavage space. Does this apply also to interest groups? Are they rooted exclusively on the traditional dimension of political conflict, i.e., the left/right divide, or do they also contest European integration directly? Dahl’s four-fold table on opportunities available for contestation in nation-states (Dahl 1971: 13) raises the question of how to understand contestation in the EU. Given the policy capacity of national and European decision-making, one might expect contestation at both levels. Conventionally, democracy involves the efforts of national interest groups at the national level. Two routes are available for contestation at the European level: the national and the European. The national route is characterized by activities of national interest groups aiming to influence European decisions, either directly at the European level or via their national governments. The European route, that is European-level interest articulation and intermediation, can only exist if there are effective interest groups at the European level (figure 9.1). Several questions arise concerning the role of interest groups for political contestation in the European Union. When, why, and to what degree has a European system of interest groups emerged? Only if there

198

Bernhard Wessels Contestation National National

Supranational

Traditional

National Route

European Route — Down

European Route — Up

Participation Supranational

Figure 9.1 Alternative routes for contestation. Interest group system

Interest group system

Fragmented

Concentrated

Inclusive

Pluralistic

Encompassive

Exclusive

Oligarchic

Selective

Figure 9.2 The scope and character of contestation.

is a sufficient number of effective interest groups at the European level is direct contestation possible. The answer to this question has implications for the character of contestation. If the emergence of European interest groups reflects established policy competencies, it is likely that contestation will primarily be concerned with particular policies along the left/right dimension. If interest groups form in anticipation of institutional change, they are more likely to be oriented to institutions as well as policy. A second distinction refers to the character of contestation, and depends on interest group structure. Taking up Dahl’s dimension of inclusiveness, interest group systems can be inclusive or exclusive. We also need to ask how interests are represented. This depends on the degree to which the interest group system is fragmented. The typology in figure 9.2 is intended to serve as a heuristic. This typology refers to the quality or scope of contestation. It can be described in empirical terms with regard to inclusiveness and the emergent structure of the interest group system. This raises the question of the contribution of interest groups to democratic representation. Since contestation is possible only through political participation and since meaningful participation demands interest group representation, it is safe to say that interest groups are particularly important at the European level. The sheer number of citizens represented by a member of the European Parliament compared to the number represented by a member of a national


199

parliament in one of the member states – the “number problem of political representation” (Wessels 1999) – makes the problem of linkage between citizens and decision-makers clear. It is obvious that European democracy cannot be based on authoritative institutions alone. As Alois Pfeiffer, member of the Commission of the European Communities put it: “[The EU] would indeed be condemned to superficiality were it not supported by representative socio-political forces” (foreword to Barnouin 1986). Linked with this is the question of equal opportunity for different interests within and among member states at the European level. On the one hand, this problem can be related to Offe (1969) and Olson (1971): not all interests can organize or articulate their demands with equal ease. But at the European level another dimension of inequality arises having to do with inequality in opportunities of interest groups among nations, given the huge differences in economic power, population size, and national political structures. I have placed four big questions on the table. Do interest groups emerge at the European level and why? Is there inequality in the contestation potential of interest groups between countries? Is the system of interest groups selective or encompassive, and pluralistic or oligarchic? And, finally, how can one describe alliances between interest groups and political parties in terms of cleavage structure, and how does that structure at the national level translate to the European level? Anticipation or reaction: the co-evolution of the European state and interest groups The history of interest group systems in nation-states shows that interest group formation has often responded to changes in the allocation of authoritative competencies. Trade associations, for example, were established to protect national capital against foreign capital (see Feldman and Homburg 1977). In some cases, governments have invited social interests to organize in order to have a counterpart to speak to (e.g., British employers in the 1960s). The reaction hypothesis appears to cover labor unions also. Labor unions were established to gain legal recognition of workers’ interests and to introduce collective bargaining in the context of state-mandated rules of the game (see, for example, Armingeon 1994). In each of these cases, interest group formation can be conceived in terms of reaction. The implication is that only if there is something to be contested, does a motive for organization arise. But this is just half of the story of group political mobilization. At least as important have been the efforts of groups to mobilize for the right to contest basic institutions. These include, for example, the bourgeois and

200

Bernhard Wessels

Growing policy competence of European political institutions

Increasing harmonization at the EC level

Demands of European interest groups

Decisions affecting national polities

Formation of European interest groups

Figure 9.3 The “circle of institutionalization” of interest groups at the supranational level. Note: Adapted from Kirchner (1978: 4).

labor movements in the first wave of democratization, and civil rights movements in the most recent wave. Hence, another pattern is that of interest groups forming in anticipation of increasing state competencies, either to influence the process or to be prepared to react to it. Anticipation and reaction can lead to the co-evolution of interest groups and the state. For example, a series of fatal steam engine explosions in late nineteenth-century Germany brought about a system of self-governance in which manufacturers mandated technical standards and engine inspections. The state transformed these standards into law and privileged the business associations with regulatory competencies. Interest group formation and state regulative capacity often increase in tandem. Such co-evolution is often based on a division of labor between state and interest groups characterized by subsidiarity (i.e. self-governance), delegation, or corporatism (i.e., concertation) (Glagow and Schimank 1983: 541). It is plausible to believe that these processes may also underlie interest group formation at the European level. A straightforward model that encompasses the three ways in which interest groups and public authorities interact at the European level has been developed by Kirchner (1978: 4) under the rubric “circle of institutionalization” (figure 9.3). Most observers of European interest groups would probably agree with Sidjanski’s hypothesis that “the development of the power of the European Economic Community has given rise to reaction from those interests, which are most directly affected” (Sidjanski 1972: 401). However,


201

Kohler-Koch (1994: 171) has found some evidence for the opposite hypothesis. She observed that interest group formation did not parallel an increase in the EC’s competencies but that “the anticipation of a growing importance of the EC in a rather vague sense . . . stimulated the establishment of transnational organizations.” Given that they have contrasting implications for the character of contestation, let us test the reaction and anticipation hypotheses against each other. This requires that we identify the points in time when interest groups formed at the European level. If the anticipation hypothesis is valid, interest groups should form prior to the allocation of new competencies to the European level. If the reaction hypothesis is valid, interest group formation should take place following such change. This requires that we identify shifts in authoritative competences to the EC/EU. Sidjanski (1972) distinguishes several phases of interest group formation at the European level, each caused by authoritative reallocation. The first wave is associated with the Marshall plan and the launch of the OEEC (Organization for European Economic Cooperation). The resulting interest groups had a very loose structure, mirroring in this sense the loose power with which European institutions were vested. The next wave, leading to the creation of about ten major organizations, began in 1951 with the establishment of the ECSC (European Coal and Steel Community). The major subsequent shifts of authority are as follows. In 1958, the EEC (European Economic Community) was launched. A European polity was institutionalized with the fusion of the ECSC, the EAEC (EURATOM), and the EEC in 1967, enlargement to include Denmark, Great Britain, and Ireland in 1973, the first direct elections to the European Parliament in 1979, the Single European Act in 1987, a European Monetary System in 1990, the Maastricht Treaty in 1992, its implementation, and the introduction of a currency union, both in 1994. It is not so easy to formulate precise expectations concerning the effects of these events. But it seems plausible that major institutional reform matters more than institutional fusion, enlargement, or policy change. Two competing interpretations are possible for the first direct European elections. On the one hand, the growing legitimacy and importance of the European Parliament could be expected to precipitate interest group formation and lobbying; on the other hand, one might emphasize that the direct election of MEPs from 1979 was only of symbolic importance. I evaluate the competing hypotheses of anticipation and reaction in two ways: first, by using the annual number of all interest groups founded in each year since 1945 as the dependent variable, second, on a stacked data matrix of twenty-four different domains of interest groups in order to test whether sector makes a difference. The reaction hypothesis is modeled by

202

Bernhard Wessels

introducing dummy variables for the year of institutional change and the two years after. The anticipation hypothesis is modeled by introducing dummy variables for the three years prior to institutional change. Turning first to aggregate models which do not differentiate interest domains, regression analysis reveals that institutional reforms correlate with interest group formation. However, it disconfirms the anticipation hypothesis. The predictions of the reaction model, including particularly large increases in the number of interest groups after the launch of the EEC in 1958, the Single European Act in 1987, the first level of the currency union, and the Maastricht Treaty, are confirmed. These effects are highly significant and differ strongly from those of the other events introduced in the model. Only two of the coefficients predicted by the anticipation model are statistically significant – the signing of the Maastricht Treaty and its implementation. But, due to overlap, the dummies cannot distinguish between reaction and anticipation in these periods. Since neither 1987 nor 1958 show significant effects in the anticipation model and almost all coefficients have the wrong sign, the evidence here strongly supports the reaction hypothesis (table 9.1). A plot of estimated and actual numbers of organizations founded each year since 1946 is revealing. There is a little peak in 1953/4, a rather larger one in 1958/9, and an even larger one in 1988/9 that continues for several years. Clearly, annual predictions derived from the reaction hypothesis are much nearer the mark for the period up to 1988 than those based on the anticipation hypothesis (figure 9.4). As already mentioned, the period after 1988 is one of concentrated institutional change where the predicted effects of anticipation and reaction cannot be disentangled. Regressions based on the stacked data matrix constructed for twenty-four specific domains do not challenge these results. Sectoral differences with regard to the impact of institutional changes are only minor. The institutionalization of an interest group system at the European level appears to have much in common with that at the national level. Supranational interest organizations have come into being as the power of European political institutions has grown. The conclusion is straightforward: since waves of interest group formation follow institutional reform, interests appear to be driven by their desire to influence policy in the European polity. This would mean that specialized functional representation rather than generalized contestation motivates interest group activity. With regard to cleavages, the implication is that the left/right (Hix and Lord 1998) or social democracy/market liberalism dimension (Hooghe and Marks 1999) dominates the interest group arena. But this interpretation is based on foundation patterns alone. Later we will return to this issue. What we


203

Table 9.1 Anticipation or reaction? Founding of European interest groups Annual “Reaction” Variable

B

Stacked

“Anticipation”

“Reaction”

SE B

B

SE B

B

SE B

“Anticipation” B

SE B

P1951 P1958 P1967 P1973 P1979 P1987 P1990 P1992 P1994 AGRAR INDUS SERVI POLSO

2.27 19.94 1.61 4.61 1.28 10.94 24.19 11.26 −3.14

3.53 3.53 3.53 3.53 3.53 3.53 3.71 3.84 3.71

−6.74 −5.74 −3.16 −2.74 −0.08 −2.41 1.89 17.11 15.55

4.72 4.72 3.93 4.47 4.47 4.47 4.70 4.86 4.70

0.25 2.99 −0.49 −0.37 0.37 1.40 1.92 0.93 −0.32 −3.45 −4.46 −2.89 −3.42

0.60 0.53 0.61 0.55 0.66 0.55 0.47 0.51 0.65 0.53 0.46 0.46 0.54

−0.86 −0.98 −1.17 −0.89 −0.37 −0.70 0.38 1.09 0.87 −3.56 −4.53 −2.97 −3.69

0.78 0.76 0.57 −0.64 0.60 0.64 0.58 0.50 0.49 −0.54 0.47 0.47 0.56

(Constant)

11.06

1.07

14.41

1.38

7.16

0.42

7.85

0.42

R2 R 2 (adj.) N

0.69 0.63 53.00

0.51 0.41 53.00

0.25 0.23 497.00

0.21 0.18 497.00

Notes: Annual: based on all interest groups founded in the respective year. Stacked: based on stacked data matrix of foundation by twenty-four interest group domains. “Reaction”: assumption: interest groups founded in reaction to competence shift. Period dummies are therefore coded 1 for the respective year and the two after. “Anticipation”: assumption: interest groups founded in anticipation of competence shift. Period dummies are therefore coded 1 for the three years before.

can say with confidence is that, with the broad-based mobilization of interests, a significant potential for contestation has arisen at the European level. Nations and domains: diversity and inequality in interest group participation in the EU Most interest groups at the European level do not have direct memberships, but are umbrella organizations, often of national groups. Thus, it might well be that interest group participation varies across countries. This might be due to differences in resources or differences in preferences.

204

Bernhard Wessels

50

No. of EU interest groups founded

45 Empirical

40

"Reaction" 35

"Anticipation"

30 25 20 15 10 5 1997

1994

1991

1988

1985

1982

1979

1976

1973

1970

1967

1964

1961

1958

1955

1952

1949

1946

0

Figure 9.4 The dynamics of European interest group formation: empirical values and estimates of “reaction” and “anticipation” hypotheses.

Whereas the former results from inequality, the latter may simply be a matter of free choice. It may also be that national interest group participation at the European level reflects diverse organizational cultures. Membership rates in interest groups vary from 20 percent in Spain to more than 80 percent in Sweden (Wessels 1997). Diversity is even greater with respect to organizational structure. Trade union movements and business associations vary with respect to their external power, their degree of centralization, and the extent to which they are fragmented (Wessels 1996). Visser and Ebbinghaus (1992), Streeck (1996), and others have discussed the implications of such variation for interest group representation at the European level. The message of these studies is that national patterns persist and constrain the formation of European interest groups. Turning to the resource aspect first, it is well known that national interest group systems vary according to population size and the size of the economy (Lowery and Gray 1993; Murrell 1984). Whereas this does not pose problems within the individual countries, the situation is somewhat different at the European level. If differences of scale result in different participation rates, this would build inequality into the interest group system. A population ecology approach to organization can be applied to the European level. Population size is strongly correlated with the number of organizations from that country active at the European level (logged;

No. of national member organizations in EU umbrellas


205

1200

1000 FR

BE

GE

GB

800

NE

IT SP

600 DE AU

400 PO IR

FI GR

LU

200

0 3.5

SW

4

4.5

5

5.5

6

6.5

GNP (log)

Figure 9.5 Size of the economy and number of countries’ member organizations in European umbrella organizations. (Adjusted R2 = 0.74; p < 0.001.)

adjusted R2 = 0.57; p < 0.001). The same is true for size of the economy measured by GNP (R2 = 0.74, figure 9.5), and exports (R2 = 0.87). Belgium, France, and Germany are each represented by more than 900 interest groups in European umbrella organizations, followed by Great Britain, the Netherlands, Italy, and Spain with between 650 and 850. Further down the scale come Denmark (ca. 500), Austria, Sweden, Finland, Portugal, Ireland, and Greece (between 300 and 500), followed by Luxembourg (ca. 200). Interests from different countries have different chances of articulation. This need not necessarily imply inequality of representation if the scope of interests in smaller societies and economies were smaller. Functional differentiation might vary across countries. However, this is difficult to measure. There is no relation, though, between the size of economic sectors and relative European representation across countries. Standardizing the number of national organizations present in European umbrella organizations for the size of the economy shows that the number of organizations is quite closely related to trade dependency.

No. of national member organizations in EU umbrellas standardized by GNP

206

Bernhard Wessels 2.5 LU

2 IR FI GR

PO

BE

DE

1.5 AU

SW

NE

SP

1 IT

FR

GB GE

0.5

0 1.5

1.6

1.7 1.8 1.9 2 Import and export in % of GNP (log)

2.1

2.2

Figure 9.6 Trade dependency and standardized number of countries’ member organizations in European umbrella organizations. (Adjusted R2 = 0.55; p < 0.001.)

Standardizing for the size of an economy reveals that smaller economies are overrepresented at the European level (figure 9.6). This suggests that the relative incidence of participation in Europe could be a mere consequence of different national interest structures. In this case, unequal participation would not mean unequal representation. It is claimed that many European interest groups lack the capacity not only to bind their members but to act effectively on their behalf. Schendelen argues that the problem of creating representative linkages is “even bigger at the transnational level of the European Community” (Schendelen 1991: 359). Some interest groups, however, have increased their powers vis-à-vis their members considerably during recent years. The ETUC began to reform its internal structure in 1991 and succeeded in amending its constitution in 1995. Now the executive committee can determine who should represent it, with which mandate, in negotiations with employers.


207

UNICE changed its statute in 1992 and became responsible for representing its members in negotiations on the Social Protocol. With the help of the European Commission and its criteria for inclusion in the process, some interest groups have come to exercise something approximating a monopoly of representation in their respective areas (Falkner and Nentwich 1997). The Social Protocol, which was implemented in 1995, was a collective agreement signed by the peak organizations of employees and employers. This might represent an ideal typical co-evolution of interest groups and the state: interest groups gain internal and external power, while the state gains regulatory capacity. Comprehensive research of interest group activity is possible only for a sample of organizations and thus must be generalized cautiously. Moreover, the data available for self-registered interest groups is rather thin with respect to organizational characteristics. Nonetheless, two pieces of information allow us to capture the degree of differentiation of interest group domains, the degree of “encompassiveness,” and in consequence the degree of fragmentation. The assumptions here are as follows: r The degree of differentiation of an interest domain (e.g., communications or agriculture) is equivalent to the number of European interest groups present in the domain. r Encompassiveness is harder to define. This concept, introduced by Olson (1982), characterizes organizations that more or less completely represent a particular domain. At the European level encompassiveness requires representation of the interests of all member states. “European encompassiveness” can be measured as the proportion of interest groups in a domain representing all fifteen member states. Plotting interest group domains on these two measures reveals wide variation. Most differentiated are the metal and electric industries; least differentiated are employers’ organizations, labor unions, and part of the service sector. Least encompassive from a European perspective are human rights interests and the energy and extractive sectors. European encompassiveness is highest among unions and employers (figure 9.7). Looking somewhat closer at sectors, industry shows on average the highest level of differentiation and the lowest level of European encompassiveness. Less differentiated and more encompassing is agriculture, followed by services and public interests. Economic interest groups (unions, employers) have the lowest level of differentiation and the highest level of European encompassiveness. This suggests that the industrial sector is most oriented to specific contestation, while socio-economic interest groups are most oriented to general contestation. Interestingly, the structure of national-level organizations’ and EU-umbrellas’ fragmentation is

208

Bernhard Wessels

Mean of "European encompassiveness"

0.85 Employ

0.80

Unions

0.75

Commu Profess

0.70

Health Finance

Publ oth, pol. Consum/Envir Science/Educ Agric oth Leisure Bever Regions Chem/Plast/Gla

Trade

0.65 0.60

Transp

0.55

Construc

Metal/Elec

Wood/Pap/Text

Service-Misc Agri+Fish

0.50

SMEs

0.45

HumanR

Energy/Extr.

0.40 0

10

20

30

40

50

60

70

Number of EU organizations in domain Agriculture

Industry

Services

Public interests

Econ IG

Figure 9.7 Differentiation and degree of “European encompassiveness” of interest domains.

very similar. This means that the diversity of interests translates quite directly from the national to the European level (figure 9.8). Interest groups that are concerned with the distribution and redistribution of public goods, i.e., in labor relations, environment, and consumer affairs, are less differentiated and more encompassing than those dealing with particular industrial interests. Interest groups concerned with movable goods (e.g., money, services) are less fragmented than industry groups but more so than economic interest groups concerned with public goods. Other public interests are either highly diverse or not very encompassing (human rights groups, for example). Industry by and large is organized pluralistically, interest groups concerned with public goods and distribution are quite encompassive, public interest groups quite selective (from a European perspective), services and agriculture somewhere in between.


209

National-level organization fragmentation index ((1-Encomp)*NoOrg)

0

200

400

600

800

1000

1200

1400

AGRICULTURE......... Agri+Fish Bever Agric oth INDUSTRY................... Energy/Extr. Chem/Plast/Glass Wood/Pap/Text Construc Metal/Elec SERVICES................... Transp Finance Commu Leisure Science/Educ Trade Health Service-Misc ECON. INTER. GROUPS... EU interest groups

Unions

National interest groups

Profess Employ SMEs PUBLIC INTEREST..... Regions Consum /Envir Publ oth , pol, reli. HumanR

0

10

20

30

40

50

60

70

EU-level organization fragmentation index ((1-Encomp)*NoOrg)

Figure 9.8 Fragmentation of interest group systems of different domains.

Political alliances and cleavages: the content of structure So far, I have analyzed the emergence of interest groups at the European level and have engaged evidence bearing on their structure and their contestation potential. I have not yet addressed the content of their

210

Bernhard Wessels

demands. Obviously, interest groups are purposive organizations. However, it is hardly possible to map the entire interest group1 landscape. I intend to address the substantive character of contestation via the alliance structure between interest groups and political parties. This will allow me to sketch a general picture, building on an established literature dealing with cleavages and party systems. In their seminal analysis, Lipset and Rokkan (1967) demonstrate that the cleavage system in Western Europe in the 1960s is consistent with that of the 1920s. Labor–capital conflict and the religious dimension play a decisive role across European societies. However, some change has taken place since the late 1960s and 1970s, when a new cleavage dimension connected to so-called new social movements and the environment emerged in most Western democracies. This dimension has been named in various ways: old vs. new politics (Hildenbrandt and Dalton 1976), authoritarianism vs. left-libertarianism (Kitschelt 1988), or in terms of ecological conflict. There is general agreement that the cleavage system of modern societies has become more differentiated and that this has been translated into the party system by green or left-libertarian parties. Does this dimension translate to the European level via voluntary associations and social movements? If so, one would expect that a triangular coalition structure would be cross-cut by a pro- vs. anti-integration dimension. The triangle consists of labor–capital–environment: in terms of political actors, unions–employers–environmental organizations (Wessels 1991). Each pole generates a particular distributional conflict. The traditional left/right conflict is concerned with the distribution of wealth. The new politics dimension is concerned with quality of life and participation, i.e., with the allocation of state resources for providing a clean environment and other regulations. Concrete collective interests along these two cleavage dimensions can be thought of as being defined and articulated by interest groups. Political parties generalize these interests and structure the cleavage system. The observable pattern of social alliances (Stinchcombe 1975) which has emerged is straightforward and demonstrates that the party system is consistent with the interest group system: 1

Single events provide opportunity to gain a clearer view: for example the case of the 1996 Inter-Governmental Conference (IGC) where hundreds of organizations took the chance and followed the invitation of the Institutional Affairs Committee of the European Parliament to participate in its hearings (Mazey and Richardson 1997). But even with this kind of single event, the variety of the articulated interests and rival coalitions even in the same domain make it difficult to come to clear conclusions about the dimensions of contest, except that they can be defined in policy terms, i.e., competitiveness, sustainability, and also on procedural matters like consultation obligations.


211

unions are with social democratic and communist parties, employers’ and industries’ associations are with liberal, conservative, or Christian democratic parties, and environmental organizations and so-called new social movements are with green or left-libertarian parties. When we map the frequency of contacts between interest groups and members of national parliaments, the outcome is exactly as expected. An eleven-country average of surveys among members of national parliaments2 shows that there is dense communication in three quite separate arenas: the labor alliance (unions, left parties), the bourgeois alliance (industry and trade associations, employers, and liberal, conservative, and Christian democratic parties), and the alliance of the weak (environmental organizations and green parties, but also regionalist and nationalist parties). Consumer associations lie in between (figure 9.9). Given the political affinities of the parties, one can conclude that interest groups at the national level are embedded in a triangular conflict consisting of the traditional poles – labor and capital – alongside a third pole on a separate dimension – environmental interests. Turning to the European level, based on findings of a representative survey of the members of the European Parliament about their contacts with interest groups,3 and including also European umbrella interest groups, a very similar structure can be observed. Again, a labor, a bourgeois, and a green alliance can be identified. European interest groups are also embedded in this structure. The national pattern seems to translate quite easily into the party system at the European level (figure 9.10). While this triangular cleavage structure context is rooted in the left/ right divide, it also appears to connect with the integration dimension. According to both Hooghe and Marks (1999) and Hix and Lord (1998), two dimensions structure the struggle over European integration: a left/right dimension, representing mainly the traditional socio-economic cleavage about the role of the state, and a supranationalism/nationalism dimension. Hooghe and Marks identify two political projects across these dimensions. The neoliberal project “attempts to insulate markets from political interference by combining European-wide market integration with minimal European regulation.” The project of regulated capitalism “proposes a variety of market-enhancing and market-supporting legislation to create a social democratic dimension of European governance. This project attempts to deepen the European Union and increase its 2

3

Missing EU members: Austria, Denmark, Finland, Great Britain. The surveys among the members of parliament were conducted by the project “Political Representation in Europe.” For documentation see the appendix in Katz and Wessels (1999). Data came from the same project but cover MEPs of all fifteen EU member states. See ibid.

212

Bernhard Wessels

2

1.5

Labor alliance Unions

1 Rad. Left Soc. Dem. 0.5

Prof Bank Lib Region Chr. Dem. T ransp Indus T rade Cons. Agric

0

Consum Environ

−0.5

Greens

−1

−1.5

−2 −2

Nation.

Alliance of the weak −1.5

−1

−0.5

Bourgeois alliance 0

0.5

1

1.5

Figure 9.9 Alliances between interest groups and political parties at the national level. (Notes: Correspondence analysis of the frequency of contacts of members of eleven national parliaments with national interest groups. The smaller the distance between a party and an interest group, the denser the contacts between the interest group and party representatives, on average.)

capacity for regulation, by among other things, upgrading the European Parliament” (Hooghe and Marks 1999: 75). Evidence in this regard can only be observed indirectly. Since we know from prior research that MEPs who report very frequent contact with a particular interest group also report being responsive to that group (Wessels 1999), it is plausible to assume that frequency of contact might tell us something about the political orientation of an interest group in relation to that of the MEP. The positions of MEPs on the left/right dimension are measured by self-placement. Their positioning on the nationalism/supranationalism


213

1.2

Labor alliance NUnio

1 0.8

EUnio

GUE (COM)

PSE (SOC)

0.6

ETran ETrad

0.4 ECons 0.2

Alliance of the weak

UPE (CON) EProf EIndu

0 −0.2

EAgra NTran

NCons

EPP (CHR)

EDN (NAT)

NIndu

NBank NI (No PG)

NEnvi

−0.4

NAgra ELDR (LIB)

EEnvi

−0.6

V (GRE)

NTrad

ARE (RAD) −0.8 −1.5

EBank

NProf

−1

−0.5

0

0.5

Bourgeois alliance 1

1.5

Figure 9.10 Alliances between interest groups and political parties at the European Level. (Notes: Correspondence analysis of the frequency of contacts of members of the European Parliament with national and European interest groups. Closeness represents the contact density. The smaller the distance between a party and an interest group, the more contacts with the respective interest group representatives of the respective party have on average.)

dimension is measured by their response to a question tapping their preference for a parliamentary EU system in which the European Parliament, the European Commission, and the Council of Ministers are empowered at the expense of national governments and national parliaments. When we plot the positions of MEPs with frequent contact to particular European interest groups in this two-dimensional space, several groups emerge. The “alliance of the weak,” i.e., unions, environmental, and consumer organizations, are located close to the project of regulated capitalism. They are left of center on the left/right dimension and strongly

214

Bernhard Wessels

support a parliamentary order of the EU. In contrast, MEPs who are responsive to interest groups representing industry, trade, and commerce are located in the neoliberal quadrant: right of center with very little support for a European parliamentary order. MEPs responsive to professional and transport associations are also in the neoliberal quadrant, but on its leftside. Agriculture and fishing as well as banking and insurance are located in the middle ground between the project of regulated capitalism and that of neoliberalism (figure 9.11). These findings are consistent with the existence of a cleavage between organizations representing producers and those concerned with the distribution and redistribution of public goods. One should keep in mind that these findings are not derived from interest groups directly but from their alliance partners. This probably means that the underlying contrast between these groups is understressed rather than overstressed. A “struggle over European integration” appears to take place between proponents of two political projects as observed by Hooghe and Marks (1999), in the arena of interest groups as well as political parties.

Interest groups and political contestation in Europe Several tentative conclusions can be drawn from this chapter. One obvious, but not unimportant finding is that an understanding of contestation in the EU must come to grips with interest groups. The sheer number of interest groups has risen from ten in 1945 to more than 900 today. The increase can be explained in terms of reaction rather than anticipation. In general, European interest groups are more concerned with particular policy issues than with the political order at large. But there are counterexamples, such as the unions, which also contest European integration as a process (see Kirchner 1978). My analysis of differentiation and European encompassiveness suggests that the extent to which one might speak of a European system of interest mediation and contestation is a matter of normative judgment. Socio-economic interest groups like unions and employers are most European in this regard; those representing particular industries are least European. Sectoral interest groups tend to be more concerned with highly specialized contestation, while public and cross-sectoral interest groups tend to have broader demands that may also concern institutional reform in its broadest sense. Substantively, we have seen that producers’ interests tend to favor a weak European state and strong market forces, whereas consumer interests tend to favor a strong democratic European state with redistributing capacity.


215

80

Net support for a parliamentary order of the EU

Consumer

"Regulated capitalism" quadrant

70

60 Trade unions

Environment

Banking/Insurance

50 Agricult./Fishing Industry

40 Transport Trade/Commerce

30 Profess. Assoc.

"Neoliberalism" quadrant

20 3

3.5

4

4.5

5

5.5

6

6.5

7

7.5

8

Left / right self-placement

Figure 9.11 Regulated capitalism vs. neoliberalism – political positions of members of the European Parliament with frequent interest group contact. (Source: MEP Survey; see n. 3.)

One implication of this is that interest group contestation at the European level is more closely related to left vs. right than with supranationalism vs. nationalism. However, the evidence presented here is not consistent with a bipolar structure of conflict. Instead, I have uncovered a triangular structure based on labor, capital, and environment. Interestingly, this is what we find at the national level, and it translates fully to the European level.

10

Contestation in the streets: European protest and the emerging Euro-polity∗ Doug Imig

By the second week of September 2000, truckers, farmers, and fishermen protesting against high fuel prices had paralyzed the economies and governments of Western Europe – cutting off the fuel supply of businesses and private citizens alike. Linked together through trade unions and engaging in wildcat actions, the protesters made good their threats to raise blockades in a matter of hours, armed with little more than their vehicles and cell-phones.1 This round of protests against the high cost of fuel had started several weeks earlier in France, where fishermen barricaded fishing ports and turned away cross-channel ferries. Joined by truckers, farmers, and taxi drivers, they paralyzed refineries and oil depots, brought traffic on the Calais–Paris motorway to a standstill, and barricaded the streets of Marseille with tons of sardines and anchovies. The fishermen’s actions had the desired effect: gaining concessions from the French government in the form of subsidy payments to offset fuel costs. But the government’s capitulation, conversely, failed to stem the protests. Emboldened by the success of the fishermen, French truckers seized the initiative, blockading some eighty petrol depots and oil refineries across the country, shutting down the airport in Nice, and closing the channel tunnel, prompting the European Commission to launch an investigation into whether the French government was failing to uphold its obligation under EU rules to facilitate the free flow of goods and services. The French government responded to the new round of protests with a broad package of concessions to fuel-dependent professions, prompting ∗

1

This chapter is based on research conducted in collaboration with Sidney Tarrow, and supported by the National Science Foundation (#SBR-961819). I thank Gary Marks, Marco Steenbergen, Sidney Tarrow, and the participants in the Dimensions of Contestation in the European Union Workshops – sponsored by the Center for European Studies at the University of North Carolina – for their extremely helpful suggestions on earlier drafts. This account is drawn from media accounts reported by Reuters, BBC News, PR Newswire, the Associated Press, Agence France Presse, and Deutsche Presse-Agentur.

216

Contestation in the streets

217

José Huleux, president of the Calais union of fishermen to announce: “It’s over and we have won.” But even as the barricades came down in France, truckers, farmers, and fishermen across the European Union mounted the barricades – demanding that their governments lower fuel taxes that in some cases amounted to more than 80 percent of the price of fuel. During the first three weeks of September 2000, almost no EU member was spared blockades. In the Netherlands, truckers brought traffic around Amsterdam and Rotterdam to a standstill. Across the border, in Belgium, a week of blockades shut down Charleroi and Nivelles, and closed the government quarter in Brussels. Angry truckers vowed to barricade Belgian and EU government offices “for weeks if that was what it would take” to gain concessions. In Germany, truckers, farmers, and taxi drivers skirted Germany’s strict laws against unauthorized protests, staging a series of “go-slow” convoys, paralyzing traffic from the French border to Berlin. In Ireland, lorry drivers vowed to shut down the country if their scheduled meetings with the government failed to yield sufficient concessions. In Spain, truckers, fishermen, and farmers mounted blockades of Madrid, Barcelona, and Merida. Similar actions were taken in Norway, Sweden, and Denmark. Meanwhile, to the east, truckers launched protests in Poland, Slovenia, and Hungary. Perhaps no country was as hard hit by the fuel protests as Britain, where nationwide blockades of refineries and fuel depots caused some 3,000 petrol stations to run dry. The resulting fuel shortage, in turn, led to waves of panic-buying, prompting Sir Peter Davis, the chief executive of Sainsbury’s (Britain’s largest supermarket chain) to announce that without the quick resumption of trucking services the chain was likely to be out of food in “days rather than weeks.” The standoff paralyzed school bus services, leaving thousands of school children stranded, and even the Premier Football League was forced to cancel soccer matches. Across Europe, protesters’ demands were remarkably similar: national governments must bring down the high price of fuel. In particular, the protesters’ ire was directed toward high gasoline taxes, which – when added to the rising cost of crude oil and the slipping value of the euro – had caused the price of fuel to soar. But European governments had grown reliant on fuel taxes and were reluctant to cut them, and the fuel protests nearly splintered a number of governing coalitions. The German greens, for example, were outraged by the suggestion of their social democratic coalition partners that the environmental tax on fuel should be cut. Likewise, French Prime Minister Lionel Jospin found himself caught between his communist partners (who demanded swift cuts in petrol prices) and his green colleagues (who derided any such concessions as handouts

218

Doug Imig

to polluters). British Prime Minister Tony Blair, meanwhile, found his popular support waning, and Conservatives vowed that – if elected – they would rapidly move to cut fuel taxes. Where was the European Union in all of this? Many national governments were eager to divert the anger of their constituents away from themselves, and might have welcomed the intervention of the EU. Voicing the position of government leaders across the continent, French Transport Minister Jean-Claude Gayssot gamely argued that European nations should respond collectively to the protests through the European Union, forcefully signaling to OPEC their collective discontent over high oil prices. By mid-September, the European Commission announced that it would seek a Europe-wide response to domestic dismay over fuel prices. The irony of the French presidency, and France’s leading role in crafting the EC stance, was not lost on other national delegates, who quickly pointed out that France was the first EU country to capitulate to protesters’ demands. European issues and European protests? As the European fuel protests suggest, contentious political action has increasingly become part of the repertoire of European political claimsmaking. From the 1960s through the 1980s, a growing number of Europeans reported that they had taken part in contentious politics (Dalton 1988), and the use of disruptive forms of political expression is by no means limited to European truckers, farmers, and fishermen. Over the past three decades a wide range of social and political actors, from antinuclear activists to anti-immigration skinheads, have taken to the streets. Even as farmers and fishermen fill the headlines, the main actors in collective action from the 1960s – students and workers – continue to engage in protest (Kreisi et al. 1995). Post-1968 Europe in many respects seems to have become a “social movement society,” in which previously unacceptable forms of behavior have entered the conventional repertoire of contention (Meyer and Tarrow 1997; Tarrow 1998a). To what extent have Europeans employed contentious political action to respond to the issues involved in European integration? Is there an emerging realm of European protest, in which grievances as well as repertoires of action are diffusing across national borders? More broadly, are protesters beginning to develop shared identities that cross national borders, and engaging in collective or coordinated transnational protests surrounding European issues? If so, is there a way in which we can include evaluations of contentious political action alongside evaluations of public opinion and party manifestos in our efforts to ascertain and measure the


219

degree to which European integration is being contested in the streets by ordinary European citizens? As the fuel protests suggest, ascertaining the European dimensions of contentious political events and campaigns can be a complex process. We can point to the following European dimensions of this campaign: r The wave of blockades that began in France diffused quickly across national borders, spreading across the EU. r Parallel groups of actors in different national settings employed parallel repertoires of action in pursuit of parallel concessions from their national governments. r Those national governments, in turn, were forced to modulate their responses to domestic contention in order to ensure that they were perceived – in domestic politics – as protecting national competitiveness, and – in European political space – as promoting free trade. At the same time, the fuel protests also illustrate the need for clarity when it comes to discussing European contention. After all, while these actions were European in certain ways, there are at least three significant reasons to argue the fuel protests were examples of domestic rather than European protest. First, this is a story about domestic, rather than European actors. The fishermen, farmers, and truckers that joined together in protest did so within the confines of domestic political space. They were not linked together through the growing network of transnational social movement organizations (cf. Smith 2001) or through European Works Councils (Martin and Ross 2001). Their national pattern of organization followed the domestic and national pattern of trade union organization, and the tenuous transnational links between national unions (Turner 1996). Further, the most active transnational movement organizations, including Greenpeace and Friends of the Earth, would have been improbable allies in efforts to reduce environmental taxes and otherwise ease constraints on fuel consumption. In these respects, the fuel protests fit a much more familiar pattern of protest, in which domestic actors demand accommodation from their national governments over what they perceive to be domestic issues. While the broad pattern of occupational groups taking to the barricades to demand fuel price reductions was a multinational phenomenon, groups of national protesters took action in opposition to – rather than in accord with – protesters from other nations, demanding that their governments act to level the playing field with foreign competitors. This is a familiar pattern as well. In those instances where European protest does cross national borders, it often involves confrontation – rather than collaboration – between similar occupational groups on opposite sides of the

220

Doug Imig

border. These confrontational protests find occupational groups appealing to their national governments for protection against the incursion of foreign competitors, against EU rulings they believe to be unfair, or demanding that the EU take action against slack enforcement by other national governments (Bush and Simi 2001). Second, this is a story about domestic rather than European targets. The fuel protesters targeted national governments and national fuel taxes. Even in those instances where national fuel taxes were low, as in the case of the gazoil used by French fishermen, protesters still demanded that national governments do something about high fuel prices, and largely ignored the European Union. When the EU finally did become involved, it was in an effort to resolve the dispute by coordinating a European fuel strategy that would prevent national concessions from distorting the single market and, in the process, to craft a unified European voice. Third, the fuel protests unfolded through domestic rather than European processes of contentious politics. Truckers, farmers, and fishermen blockaded refineries and fuel depots, highways, ports, and office districts. While their blockades – particularly of ferry terminals, the Channel tunnel, or border crossings – had dramatic repercussions for international trade, the protesters did not engage in transnational protests, either by coordinating their actions across national borders, or by banding together to mount transnational movement events, or by leveling their demands at the supranational institutions of the European Union. Except in the case of Brussels, where proximity made ready targets of EU institutions, protesters largely ignored them. Certainly this is a case of massive numbers of ordinary citizens taking extreme and potentially costly action in pursuit of synchronous political and economic concessions. But it is not an example of European contention, if by that phrase we mean one that involves citizens bound together as Europeans rather than as national citizens, or that targets European rather than national targets, or that occurs through transnational rather than domestic processes of political engagement. How, then, should we decipher the ways in which contentious political engagement speaks to the question of European integration? Put differently, how representative is the case of the fuel protests to the broad range of protest occurring in Europe; and to what degree are Europeans expressing their concerns with the processes and direction of European integration through contentious political action? It may be that the EU will simply prove to be beyond the reach of contentious political action, or that its institutional logic will advantage institutional approaches and discourage contentious ones (Marks and McAdam 1996, 1999). In the absence of significant transaction costs or


221

variations in resources, citizens would be expected to automatically bring their claims before the agents most directly positioned to respond to their grievances. But when – as in the case of the EU – these agents are distant, indirect, and often obscure, claims are more likely to be directed to where people possess dense social networks, organizational resources, and visible political opportunities (Imig and Tarrow 1999). The interrelationship of available structures of political opportunity and the agency and reach of ordinary citizens suggests that the predominant form of contentious political engagement will continue to take the form of: r routine domestic protests, in which domestic actors will organize domestically and petition, march, strike, demonstrate, sit in, obstruct premises, and more rarely commit acts of violence against intra-national targets. Meanwhile, where protest does take a European form, it is unclear whether we will see: r cooperative transnationalism: in which parallel protests make claims on different national targets in cooperative but recognizably separate acts of contentious politics; r collective transnationalism: in which protest is organized across borders against common European targets; or r the domestication of conflict, in which national actors protest at home against policies of the European Union, demanding that their own national governments act as interlocutors on their behalf before the EU. An overview of European protest politics To gain some purchase on these questions, it is necessary to look beyond the details of individual protests and headline-grabbing campaigns, such as the fuel protests, in an attempt to observe trends and inflections in the changing pattern of contentious political action across Europe. If we can follow the development of contentious politics in Europe across the recent period of integration, and if we can characterize the shifting proportion of those protests which are European rather than domestic in character, we should be able to address the following questions. First, to what degree does contentious political action in Europe involve the EU? Second, what groups of claimants are most likely to undertake contentious collective action over European issues? And, third, does this realm of European protest align along identifiable dimensions of political contestation? Data Beginning in the mid-1990s, Sidney Tarrow and I began working to construct a database on European protest events. Drawn from wire service

222

Doug Imig

reports on the then twelve members of the EU, the database catalogs nearly 10,000 protests that occurred in Europe between 1984 and 1997, and provides an empirical basis for the quantitative analysis of European contentious politics. In building this database, we attempted to construct a longitudinal and cross-national source of information on European protest events. The empirical record provided by the resulting data set allows us to begin to talk about both the degree to which Europe is becoming the focus of protest politics, and the forms that protest takes when citizens respond to EU issues. Moreover, it allows us another vantage point from which to analyze the dimensionality of popular contestation over European integration. In order to construct this data set, we reviewed all of the political media reports filed by Reuters between January 1, 1984 and December 31, 1997. Across this fourteen-year period, Reuters carried accounts of some 9,872 protest events in Europe. As we would expect, these actions were launched by a broad range of social actors, including farmers and workers, environmentalists and peace activists, as well as students, skinheads, immigrants, and many others. Moreover, this data set catalogs a broad and evolving spectrum of routine forms of political engagement, including strikes, marches, sit-ins, and rallies, as well as more violent forms, like rock-throwing, hunger strikes, and soccer hooliganism.2 Europeanization and the domestication of protest What can this data set tell us about the degree to which Europeans have come to employ contentious political action to respond to the EU? To answer this question, it was first necessary to differentiate between domestic and European protest – a task complicated by the composite relationship between national governments and European institutions. While some claimants who are directly affected by EU decisions are likely to frame their grievances in terms of European institutions and policies, many others are more likely to continue to frame their grievances in national and 2

For a complete account of the database and its construction, see Imig and Tarrow (2001) (especially chapter 2 and the methodological appendix). In brief: we used an automated coding system to “read” and code the location, source, event type, and target of every political event reported in our Reuters sources, and isolated those political events that occurred in one of the twelve selected EU member states between January 1, 1984 and December 31, 1997; that were initiated by a private actor or group; and that were contentious rather than routine interaction events. In this way, we generated the subset of 9,872 contentious political events that are the empirical basis for the findings reported here. Finally, we hand-checked and hand-coded each of these events to compile the smaller subset of (490) events in which there was a direct relationship between protesters and the policies or institutions of the European Union.


223

domestic terms. This makes it difficult – short of generous and possibly subjective interpretations of the press data – to readily isolate the subset of European protest. In response to this complexity, we chose a conservative operationalization of European protest, in which a media report had to link – within the first sentence – a protest with the EU in order to categorize it as a Euro-protest. Within the set of 9,872 contentious events recorded for this fourteen-year period, some 490 events (roughly 5 percent of the total) fit this strict definition of European protest. As even these rough figures suggest, the predominant source of grievances and the dominant venue for contentious political action in Europe continue to be within domestic politics. Almost 95 percent of recent protests in Western Europe are instances where private actors make claims against national or subnational targets over purely domestic grievances. This finding alone is a healthy corrective to the notion that the nation-state is withering rapidly as a focus of citizens’ claims and that Western Europe will become a transnational polity in short order. For students of social movements, this finding should caution that, at least where Europe is concerned, a truly transnational sphere of social movements is proving to be slow in coming. But for students of European integration, the other side of this equation is equally interesting. Whereas analyses of public opinion data and party manifestos suggest that European integration has little salience for most Europeans (van der Eijk and Franklin 1995; Steenbergen and Marks this volume), the record of recent contentious political events suggests that a small but significant share of politics in the streets – i.e., contentious events – is motivated by the policies and institutions of the European Union. The implication is that for at least some groups of actors the growing policy competence of the EU is extremely salient, outweighing the high transaction costs of engaging in collective action.

Euro-protest: a small but growing phenomenon There is also evidence in the data set that European conflict patterns are changing. Figure 10.1 presents the distribution of European protests – those explicitly concerned with the EU over time – alongside the percentage of all Western European protests that they represent. As the figure suggests, while the share of all events motivated by EU institutions and policies remains small, their frequency as well as their share of the total has risen rapidly post-Maastricht. In other words, Europeans are increasingly likely to take to the streets in protest against the European Union,

Doug Imig

100

160

90

140

80

120

Percentage

70 60

100

50

80

40

60

30

# of Events

224

40

20

20

10 0

0 84

85

86

87

88

89

90

EU events as a percentage of total

91

92

93

94

95

96

97

Frequency count of EU events

Figure 10.1 Frequency and percentage of Western European contentious events provoked by EU policies and institutions, 1984–1997.

its agents, and its policies – though they are still much more likely to launch protests against domestic grievances. The trends represented in figure 10.1 intimate that European contestation is changing. As Gary Marks and Marco Steenbergen suggest (this volume), we would expect to find considerable variation over time in the political conflicts that are engendered by European integration, as the competence of the EU grows, and as ordinary citizens find new and powerful venues for expressing their concerns with that process of development. Figure 10.1 provides evidence of a general upward trend in the incidence of European protest, as well as an erratic pattern of engagement of ordinary citizens and the European Union, suggesting that two processes may be at work simultaneously. The gradual upward trend suggests domestic actors may be learning to engage supranational policy-makers incrementally. At the same time, individual peaks of heightened protest indicate that specific policy disputes may selectively heighten the incentives for certain aggrieved groups to make claims involving the EU – underscoring the importance of critical junctures in the timing and frequency of contentious politics (McAdam and Sewell 2001). European actors, processes, and targets To the extent that Europe is emerging as a focus of contentious politics, is it the focus of particular groups of social and political actors, or is it


225

drawing European citizens at large into the streets? Are certain groups better positioned to surmount the obstacles to transnational mobilization; or are European grievances leading to other processes of contention (e.g., leading to the domestication of European issues within domestic political venues and against national political institutions)? We would suspect that the high transaction costs of transnational collective action are easier for some groups to surmount, resulting in differential access to supranational institutions. As a result, European integration would encourage those social and political groups that are already comparatively resource-rich to more readily engage European institutions and policies, while less well-armed claimants will remain locked within national political space. At the same time, those groups that are already well organized, such as workers, are not likely to abandon their well-worn national opportunity structures to organize at the European level (Streeck 1996). Such a change may require the development of a cross-national mobilizing capacity, and an infrastructure for transnational mobilization (Turner 1996). For the time being, this would leave workers further disconnected from the venues in which Europeanization is occurring (Ross and Martin 2001). On the other hand, the range of non-occupational groups taking to the streets on behalf of post-materialist concerns may find the transition to the European level easier to make for several reasons. First, they draw on a comparatively mobile, educated, professional constituency whose members can talk more easily to Eurocrats or to others like themselves from other member states. Second, their concerns – such as clean air and water – cross national boundaries (Smith 2001). On the other hand, once they reach Brussels, according to scholars like Marks and McAdam (1996; 1999), such groups are increasingly likely to take advantage of the lobbying opportunities that the European Union makes available rather than engage in protest. Thus we may find them in Europe but less likely to engage in contentious politics against Europe. What does the empirical evidence suggest about the likelihood of occupational and non-occupational groups across Europe mounting protests over EU issues? First, our findings hint at the importance of an existing organizational infrastructure combined with direct contact with EU policy in provoking contentious politics. We see this, for example, in table 10.1, which indicates that the largest share of Euro-protests (more than 82 percent of the total) have been launched by groups organized along occupational lines, suggesting that it is workers who have been first to protest against the EU. This finding is particularly striking when we

226

Doug Imig

Table 10.1 The domestic and European repertoires of contentious action for occupational and non-occupational groups, 1984–1997 (%)

Occupational groups, N = 3,439 Euro-protests, N = 402 Domestic protests, N = 3,037

Violence against property

Violence against persons

Peaceful protests Strikes

Confrontational protests

27.6

9.3

45.3

8.6

9.1

18.0

45.7

27.8

3.3

5.2

0.0

36.2

3.7

10.0

0.0

33.2

10.7

25.2∗

Non-occupational groups, N =6,433 Euro-protests, 50.0 N = 88 Domestic protests, 30.9 N = 6,345

Notes: ∗ 38.8 percent of this category is composed of Basque, Catalan, and Irish Nationalist actions, and events that escalated from more peaceful protests (33.2 percent of the total). If these two types of protest were removed, violence against persons would account for 9.1 percent of the contentious political action of non-occupational groups in domestic politics.

compare it with who is most likely to protest around domestic issues. Across the twelve nations studied, non-occupational groups accounted for more than twice as many domestic protests as did labor unions, farmers, and other groups of workers. Our data, then, suggest that the preponderance of Euro-protest is likely to have another feature in common with the recent fuel protest campaign: primarily it has been initiated by occupational groups, and organized along the lines of existing domestic unions and professional associations. Who are the occupational groups that are most active in protests against EU policies? To date, farmers alone account for roughly half of the protest launched by occupational groups and targeting the EU. Following their long involvement with the Common Agricultural Policy, farmers are more attuned to European policy than are many other groups. Even among farmers, those with long EU involvement – such as the Dutch – are far more likely to see Brussels at the center of their concerns than are more recent entrants to the EU (Klandermans et al. 2001). In addition to farmers, we find a vigorous range of contentious actions launched by other occupational groups, including fishermen, construction workers, and miners. Not coincidentally, it is these same


227

groups who are confronting the painful realities of integration first hand: through reductions in agricultural subsidies and production quotas, shifting trade restrictions, limitations on net sizes and fishing territories, and layoffs and closures in the name of fiscal austerity and monetary union. Alongside the activism of occupational groups, other social actors have been much slower to engage in transnational contentious politics. NGO representatives in the environmental, migrant and women’s sectors interviewed in Brussels underscore this finding, noting the indifference of their grassroots memberships to the growing importance of European decision-making. Without grassroots support, their lobbying efforts lack the clout of the better-financed business and professional groups they oppose (Imig and Tarrow 2001). The picture of Euro-protest that emerges, then, is characterized by a handful of dramatic cases of contentious action on the part of the environmental, student, anti-nuclear, animal rights, and anti-racist movements, but the largest proportion of contentious politics surrounding European integration continues to involve farmers and workers. Europe may be developing at the summit as the “Europe of the Banks,” but the Europe of those who work appears to be developing through a much more contentious process. But what are we to make of the thesis that the expanding competence of the EU in itself advantages more institutional lobbying approaches, and marginalizes contentious politics altogether? One way to gain some insight into this issue is by comparing the repertoire of contentious claimsmaking employed by groups of actors in domestic and European political space. Is the development of transnational politics leading to the adoption of less contentious tactics? Table 10.1 compares the repertoires employed by occupational and non-occupational groups when they protest against domestic and against European issues, and suggests a number of striking differences in the repertoires employed by these two sets of actors when they launch domestic and European protests. For occupational groups, the most notable differences parallel the presence or absence of more institutionalized and routinized opportunities for political expression. This is most evident, for example, when it comes to occupational strikes, the mainstay of contentious labor engagement in domestic politics (accounting for nearly half of the domestic protests in the sample). In comparison, strikes have been called much less often over European issues, where unions cannot resort to the traditional range of tactics such as strikes and related forms of engagement that have long served labor well.

228

Doug Imig

Second, these data suggest that non-occupation groups are more peaceful in protesting against European issues than are their counterparts in labor (peaceful protests account for half of the European actions of non-occupational groups, compared to a little over a quarter of the European protests launched by occupational groups). In sum, while non-occupational groups account for the majority of protests in Western Europe today, they are much less likely to engage in protests about European issues and institutions than they are to mobilize against domestic targets over domestic issues. Additionally, when we place the popular accounts of violent clashes between protesters and the police at World Trade Organization or G-7 meetings into the context provided by our data, it appears that transnational protests launched by groups other than workers have proven to be both more scarce and more peaceful than those launched by workers, farmers, and trade unions. It is these latter groups that have been quicker both to launch Euro-protests and to engage in more confrontational and even violent action. The dimensionality of protest As the discussion to this point has demonstrated, occupationally based groups are at the forefront of protests concerning Europe. But is it safe to assume that workers – or for that matter environmentalists or members of the women’s movement – have presented uniform demands concerning European issues? Can we identify the dimensions of these groups’ contentious political engagement with Europe? To address these issues directly, each of the 490 events in the sample of European protests was assigned a code to indicate the central issue underlying the conflict. Drawing from research on the dimensionality of political contestation (cf. Inglehart 1978; 1997; Hix 1998; and Kitschelt 1999), these protest events were assigned to one of six categories. Each of these categories, in turn, represents the extreme poles of the three most often discussed – and potentially distinct – continuums along which contestation over European may be occurring (cf. Steenbergen and Marks, this volume). The first pair of categories corresponds with the materialist/post-materialist distinction: r Protest events were categorized as materialist if they concerned issues involving economic controls, general wage policies, and policy debates concerning banking and currency. r Events were categorized as post-materialist, following Inglehart, if they involved any one of a broad set of non-market dimensions of European


229

public policy, including the environment, animal rights, and human rights. The second dimension of contestation relates to the traditional left/right political continuum: r Protests were coded as left political/economic concerns if they involved either demands for greater domestic economic protection or called for greater governmental intervention to protect industrial groups, or subsidy or production quota arrangements. r Events classified as right political/economic concerns included demands for increased trade or for opening Europe-wide markets. Finally, the third pair of categories corresponds with the prointegration/anti-integration dimension: r Protests were categorized as pro-integration if they called for increased EU power or policy competence. Protests in this category usually took the form of calls for the EU to intervene in disputes between groups of competitors from different member nations. r The anti-integration dimension, in contrast, included those protests that were launched against growing EU policy-making authority, proposals, and action. If we employ this filter to interpret our data, what picture emerges concerning the dimensions of popular contestation over European integration? Figure 10.2 presents the distribution of European protest events over time, coded by their underlying dimension of contestation. As the figure indicates, the largest share of European protests fell along the left/right dimension. Additionally, a robust share of these protests occurred along the pro- and anti-European dimension. But are these two dimensions largely capturing overlapping phenomena, with occupationally based groups seeking greater domestic economic protection and more favorable subsidy and quota arrangements on the one hand, and decrying proposed or implemented EU policy initiatives on the other? Certainly we would suspect that there would be a strong congruence between the left politics and anti-EU dimensions, with the same groups demanding both greater national sovereignty and greater domestic economic protections. Figure 10.3 provides additional insight into this relationship by grouping Euro-protests according to their underlying dimension of contestation. This view of the data helps to accentuate the relative frequency and placement of protest along each of these six dimensions of contestation. Figure 10.3 underscores the clustering of these Euro-protests around a specific range of issues, centered on market protectionism. Across these

0

10

20

30

40

50

60

70

80

90

82

86

Materialist / Post-materialist

88

Left / Right

90

Pro- / Anti-integration

92

Figure 10.2 Dimensions of contestation in Euro-protests, 1984–1997.

84

94

96

98


231

6

5

4

3

2

1

15 Dec 82

10 Sep 85

06 Jun 88

03 Mar 91

27 Nov 93

23 Aug 96

Figure 10.3 Distribution of Euro-protests along six dimensions of contestation, 1984–1997. Notes: 1 = materialist; 2 = post-materialist; 3 = left politics/intervention; 4 = right politics/free trade; 5 = pro-integration; 6 = anti-integration.

fourteen years, the great majority of European protests advanced left political/economic and anti-integration demands. This clustering may offer reasons to rethink the strong version of the thesis that attitudes toward the EU are orthogonal to other dimensions of contestation. At the very least, the clustering of protests around the left politics and anti-integration poles suggests that, in the eyes of those groups who are launching direct action, the expansion of the EU is extremely relevant so far as it influences domestic political and economic life. At the same time, the figure also suggests a degree of change over time in terms of the dimensions along which protesters are contesting Europe. Over time, we see other issues, including post-materialist and new political concerns – provoking Euro-protests. The concerns of these emerging groups of European claimants have extended well beyond economic protectionism. But even as these groups have emerged in transnational politics, they have not displaced those farmers and unionists who have been active on the left politics and anti-integration dimensions. In this respect, the breadth of contestation over Europe appears to be expanding beyond materialist concerns, without displacing the central position of unions and workers. While environmentalists and supporters

232

Doug Imig

of migrants’ rights, for example, may be developing a range of Eurocentered collective actions, the realm of transnational collective action is still notable for the range of interests that are not represented. Preliminary findings: evidence of the growing popular salience of Europe? The picture that emerges from the longitudinal and cross-national examination of European protest suggests that European integration is highly salient to a growing range of citizens across the continent. While Europefocused contention still constitutes a small share of all protest, that share has been rising since the mid-1980s. As we would expect, it has been those groups that fear their livelihood is threatened by integration – groups largely organized by occupation – that have launched the largest share of these Euro-protests. In consequence, when protests are mounted over Europe, they most often oppose the specifics of integration and call upon national governments for greater domestic market protection. One implication of this finding is that – rather than being orthogonal to domestic political concerns – there is a close correspondence between opposition to the project of Europeanization and more traditionally domestic left political concerns. The development of a supranational realm of European government presents a series of new opportunities and constraints for domestic social actors. Although transnational political engagement and supranational governance increasingly characterize politics in Europe, the realm of social movement mobilization – to date – has largely remained focused on domestic politics. But protesters are becoming attuned to European issues and are slowly building a European repertoire of contentious action. They are starting to band together in cross-border actions with like-minded actors from across the continent, and to focus their demands on the European Union. But set against the rapid development of supranational policies and institutions, substantial barriers remain in the way of movements eager to join the transnational realm. As a result, national governments continue to play a primary role in policy-making before the EU, and tried-and-true routines of collective action and familiar institutional patterns continue to attach citizens to their national political systems. In this respect, it is not surprising that Europe is increasingly the source of citizens’ grievances, while the majority of efforts to express political preferences about Europe occur in domestic rather than transnational venues. But it is within the 5 percent of the events in our data set that target European rather than national issues and institutions that we begin


233

to see the reasons why private actors come to focus on Europe. As in the fuel protests, farmers, fishermen, construction workers, miners, and truckers have launched most of the Euro-protests in our sample. It is these same groups that have confronted the painful realities of integration first hand and, consequently, have had the greatest incentive to surmount the barriers to transnational engagement. In the same vein, claimants are proving to be just as likely to employ contentious forms of political engagement in European as they are in domestic policy space. For those concerned with the democratic deficit at the European level, this finding emphasizes the continuing struggle of ordinary citizens to find appropriate ways to approach the supranational institutions of the EU. Moreover, these findings also suggest that, rather than seeing an immediate and direct displacement of contentious politics from the national to the supranational levels, we are more likely to see a range of social movement approaches to the European level, including transnational cooperation against domestic actors, collective European protests, and the domestication of European issues within national politics. To date, the largest proportion of protests concerning the European Union have taken domestic rather than transnational shape, indicating that while Europeans are increasingly troubled by the policy incursions of the EU, they continue to vent their grievances close to home – demanding that their national governments act on their behalf. The Euro-protests in our data have largely proven to be the project of workers. Those groups that have been most likely to march, picket, and strike against Europe have done so over issues including jobs and welfare. As integration has progressed, occupational groups have been joined in the streets by a growing range of claimants, including students, peace marchers, environmentalists, and anti-racist forces. But the empirical record also suggests that the emergence of Euro-protests over post-material issues has been both uneven – with a number of notable absences, including representatives of territorial politics – and accompanied by more, rather than fewer, protests around traditional materialist concerns. What are the implications of these findings for our understanding of popular contestation surrounding European integration? The increasing presence of the EU as the target of Western European protests suggests that a growing number of Europeans are ascribing to the EU responsibility for their grievances, and that they are seeking ways to effectively voice their concerns. For most citizens, the process of confronting Europe has followed directly from the ways in which integration has affected their own existences. This has meant that those groups most immediately

234

Doug Imig

affected by EU actions have been the first to mobilize in response. The result has been a clustering of European protests around left political causes and in opposition to the project of integration, as ordinary people have fought to protect their jobs and families against restructurings that they fear will threaten their way of life.

11

Conclusion: European integration and political conflict∗ Gary Marks

Over the past half-century, Europe has experienced the most radical reallocation of authority that has ever taken place in peace-time, yet the ideological conflicts that will emerge from this are only now becoming apparent. This book originated in the efforts of a group of scholars to investigate the patterns of conflict – dimensions of contestation – that have arisen from European integration. The question that motivates us is a broad one: how does European integration play into the domestic politics of the member states?1 In this volume, we resolve this abstract question into a more precise and empirical one: to what extent and how are the issues arising from European integration connected to the dimensions of contestation that structure domestic politics? Is European integration assimilated within the major lines of conflict, above all the competition between left and right, or is it unrelated? Rather than divide Europe by country, each of us examines one kind of group – citizens, national political parties, social movements, interest groups, members of the European Parliament, and European political parties – for the EU as a whole. We engage several kinds of data, including Eurobarometer surveys, party manifestos, expert evaluations of party positions, and elite interviews. We cannot claim to be of a single mind, but we do claim that we arrive at broadly consistent answers to our question. The aim of this chapter is to convey their substantive thrust. That our conclusions are based on analysis of several independent sets of data for diverse national and European actors reinforces, we think, their plausibility. ∗

1

I would like to thank Liesbet Hooghe for inspiration and ideas, and Simon Hix, Herbert Kitschelt, Hans-Dieter Klingemann, Fritz Scharpf, Wolfgang Streeck, Bernhard Wessels, and members of the research unit on Institutions and Social Change at the Wissenschaftszentrum Berlin for perceptive comments on this chapter. I would also like to thank seminar participants at the Technische Universität Munchen ¨ and the Max Planck Institute for Social Research in Cologne to whom this chapter was presented. Our focus is on patterns of contestation (Katz and Wessels 1999; Schmitt and Thomassen 1999), rather than levels of support for European integration.

235

236

Gary Marks

I propose to set out our findings in steps, beginning, in the next section, at the aggregate level. To what extent does the left/right divide constrain views of European integration in general, across the European Union as a whole? We can answer simply: there is no robust linear relationship. The relationship we find is curvilinear: an inverted U-curve, pronounced for political parties, weak for the general public, created by centrist support for European integration, and opposition from both left and right extremes. When we probe beyond the aggregate level, things begin to get more interesting and more controversial. How has the ideological fit between European integration and domestic dimensions of conflict changed over time? Are ideological patterns of support and opposition to European integration visible at the level of particular policy areas? Does the ideological fit between domestic contestation and the issues arising from European integration vary from country to country, and if so, why? To answer these questions we have to cut three ways: r By time. European integration is not merely a moving target; its ideological bearings have shifted 180 degrees over the past two decades. In the 1980s, European integration was essentially a market-making project, favored by the right, less so by the left. By the turn of the century, the situation was reversed, as left-leaning policies, such as environmental policy, social policy, and employment policy came on the agenda. r By issue. Some European issues connect closely to domestic politics; others do not connect at all. European issues having distributional consequences within countries are most closely related to left/right contestation. European issues that affect national sovereignty relate more closely to new politics contestation. r By territory. National institutions frame how European issues map on to domestic cleavages. Some issues play the same way across countries – and give rise to pure ideological cleavages – while others evoke contrasting patterns of support and opposition in different countries. Issues that have consequences for the allocation of values across countries give rise to national coalitions. Our approach is fine-grained because the questions we ask require it. The devil is in the details. But our aim is not to confound the reader with complexity. By looking more precisely, we can observe – and generalize about – patterns of political conflict that are invisible at the aggregate level. We have had to abandon our original ambition to arrive at one simple model that can describe how left/right contestation is related to support for European integration. We began by setting out four simple and

European integration and political conflict

237

logically distinctive models, and we continue to find them useful benchmarks. But no one model is valid for the EU at every resolution of detail. The thrust of this book is to examine the conditions under which one or the other model is valid, and this leads us to inquire into variation across issues and variation across territory. Aggregate findings At the aggregate level – that is to say, when we treat European integration as a single dimension – the model that best describes the relationship of European integration to the left/right dimension over the past two decades is the Hix–Lord model, in which European integration and left/right positioning are orthogonal to each other. According to this model, left/right conflict allocates values among functional groups, whereas European integration allocates values among territorial groups. Hence, the position that a person takes on one dimension does not constrain her position on the other dimension. As we described this model in the Introduction, all four quadrants are feasible policy options: left/pro-integration, left/antiintegration, right/pro-integration, and right/anti-integration. This is confirmed by the chapters in this volume that are concerned with individual citizens. Matthew Gabel and Christopher Anderson (chapter 1) find that citizens’ views on European integration are weakly associated with left/right self-placement. Left/right self-placement has a factor loading of 0.065 in a one-factor model (model 2) of attitudes towards more EU activity, far lower than any other item. Cees van der Eijk and Mark Franklin (chapter 2) find essentially the same thing. Pro-/anti-EU orientations of voters bear almost no systematic relation to their left/right self-placement, as illustrated in figure 2.1. We reject one possible explanation for this non-association, namely that orientations toward European integration are unstructured and, consequently, random. Gabel and Anderson find that attitudes towards European integration in the public at large are quite well structured. Van der Eijk and Franklin note that respondents appear to have little difficulty placing themselves on an EU integration scale. The percentage of missing data for this scale in their survey of the European electorate is little more than half that for the left/right scale.2 Respondents are not at a loss to place themselves on a pro-/anti-European integration scale. Moreover, as van der Eijk and Franklin point out, respondents locate themselves further toward the extremes on the EU integration scale than they do on 2

For a contrasting view of the extent to which individual citizens have structured opinions about European integration see Sinnott (1997) and Green (2001).

238

Gary Marks

the left/right scale – an indication that real attitudes towards the EU are being tapped. Analyses of political parties also conclude that there is no strong and durable relationship between left/right positioning and support or opposition to European integration in general. Liesbet Hooghe, Gary Marks, and Carole J. Wilson (chapter 6) find a positive linear association between left/right party position and overall support for European integration. Right parties are more likely than left parties to support European integration in general. However, as we discuss below, they are less likely to support further integration on issues such as the environment and employment policy. So there is no overall significant linear association between European integration, conceived as a whole, and left/right contestation. This leads van der Eijk and Franklin to describe European integration as a “sleeping giant.” European integration is orthogonal to the left/right divide, yet it is difficult to overestimate its substantive importance. European integration has transformed Europe economically and politically, yet orientations to it are not constrained by the dimension that chiefly structures contestation across European societies. If European integration were to become highly salient, it might therefore become a combustible issue (Evans 1999). The inverted U-curve The most powerful association that we find at the aggregate level between left/right position and European integration is an inverted U-curve describing support for European integration among centrist parties, and opposition among parties toward the extremes of both left and right (Aspinwall 2002; Hix and Lord 1997; Marks, Wilson, and Ray 2002; Taggart 1998). Doug Imig (chapter 10) finds that the bulk of popular contestation oriented directly or indirectly toward the European Union is anti-integration. Most of the groups that have organized protests are on the left, but all are outside the centrist mainstream that controls the levers of authority (Imig and Tarrow 2000, 2001). Hooghe, Marks, and Wilson find that national political parties towards the left and right extremes take Euroskeptical positions on European integration at the aggregate level and across the board on individual issues. Van der Eijk and Franklin also find an inverted U-curve in party positions as imputed by voters (figure 2.3).3 3

To pick up the inverted U-curve one must include parties or groups at the extreme, but these tend to have small memberships and support, and as a result tend to be underrepresented in data sets such as the Manifesto Research Group data or the European elections survey.


239

There is a substantive explanation for this and a strategic one. Substantively, the European Union is a centrist project for the simple reason that mainstream parties – Christian democrats, liberals, social democrats, and conservatives – have dominated national governments, national parliaments, the European Parliament, and the European Commission. Parties on the extreme left and extreme right, along with contentious social movements, have little love for institutions they have done almost nothing to create. They attack European integration as an extension of their domestic opposition. The extreme left views European integration as an elitist capitalist project that isolates decision-making from citizens in the interests of powerful corporations. The extreme right views European integration as an elitist supranational project that weakens national autonomy and traditional values. Strategically, positions on European integration are framed with an eye to sustaining or challenging existing dimensions of domestic conflict (Steenbergen and Scott, chapter 8). Centrist political parties converge in support of European integration because they want to bottle up a potential new dimension of conflict (Hix 1999a; Scott 2001). They cannot assimilate European integration into the dominant left/right dimension that structures national competition, and so they try to avoid competing on it. This has the considerable advantage of dampening an issue that could otherwise fracture mainstream parties. Conversely, parties that are toward the left and right extremes want to raise the heat by taking anti positions on European integration. While such parties are minor contenders on the established left/right dimension, they may be far more successful if they can impose a cross-cutting conflict on which they are more united than their mainstream competitors. Time The relationship between left/right orientations and the degree of support for European integration depends on when one is asking the question. European integration is a swiftly moving target. Two decades ago, in the early to mid-1980s, the chief issue on the agenda had to do with sweeping away non-tariff barriers to trade. This meant limiting public subsidies to industry, bypassing protectionist product standards, opening up public procurement, and reducing red tape – all of which was music to the ears of those on the right. It was no coincidence that the Single European Act of 1986 was negotiated by nine center-right governments and three left governments, the most important of which was the Mitterrand government in France which had tried, and failed, to implement an interventionist socialist program. A majority of those on the left were prepared to go

240

Gary Marks

along with the market project, but they regarded it merely as a first step to a social democratic “Citizens’ Europe” (Hooghe and Marks 1999; Ladrech and Marlière 1999). Following the Maastricht Treaty (1993) and currency union (2002), the creation of a single European market is no longer a topic of debate in Euroland. Yet, as Steenbergen and Scott show (figure 8.1), the salience of European integration increased from the 1980s to the 1990s. On the agenda now are a wide variety of proposals for further integration, several of which are more popular with the left than with the right. These include market-flanking policies, such as employment and environmental policy, which draw disproportional support from the left in every EU member state. On the basis of content analysis of European party manifestos, Matthew Gabel and Simon Hix (chapter 5), find that left and right have switched positions on European integration over the past two decades. The center-right European Peoples’ Party was more pro-integration than the party of European Socialists in the 1970s; by the 1990s, the situation was reversed. There is no immutable relationship between left/right positioning and support for European integration. As in the American federal system, sometimes it is the left and sometimes the right that supports more centralization (or more decentralization). It all depends on what the status quo is and what one wants to defend or achieve by reallocating authority. Digging deeper We refine the analysis of the connection between domestic and European contestation in two ways, by examining variation between issues and variation across countries. We discover that particular aspects of European integration evoke responses that are indeed constrained by ideology. The search for such connections leads us to analyze new politics contestation alongside left/right contestation. We also examine how territory mediates ideology across the member states of the European Union. We discover that national institutions frame how European integration plays on the left/right divide. Disaggregating by issue Left/right The connections between domestic and European contestation come into sharp view when one examines specific sets of issues. European


241

integration is diverse in the particular sense that the issues it raises are more weakly intercorrelated than the issues that make up the left/right dimension (see Brinegar, Jolly, and Kitschelt, chapter 4, n. 8).4 Both left and right can support more European integration. It depends on what issue one is talking about. So, for example, the left favors more integration in employment policy, while the right favors market integration. At the aggregate level, when one asks about European integration in general, such contrasts wash out. There is not much difference in the degree to which left and right support European integration as a whole. This is reflected in the weak association between left/right and the standard Eurobarometer question concerning “more or less integration” (Gabel and Anderson, chapter 1).5 An issue-based approach tells a different story. When Gabel and Anderson examine citizens’ views on “what kind of integration,” rather than “more or less integration,” they find that a left/right dimension underlies public attitudes. The items that load most heavily on this dimension are “improving equality of opportunity” (for minorities and women), “more help to the poor and socially excluded” (and to the Third World), “support for poorer EU regions,” and “protect[ing] consumers.” European issues that have to do with the political regulation of the market are most closely connected to the left/right dimension. According to the Hooghe–Marks model set out in the Introduction, the centerleft supports political integration in order to create European regulated capitalism with the capacity to regulate markets, redistribute resources, and sustain partnership among public and private actors (1999). The project of regulated capitalism at the European level is rooted in Jacques Delors’ decade-long presidency of the European Commission (1985– 94), and his effort to build an espace organisé around social and cohesion policy. Regulated capitalism is an ideological project – and is opposed by those on the right who consider market integration a worthy goal, rather than a point of departure for further integration. This is consistent with Bernhard Wessels’ findings for interest groups (chapter 9). He discovers three coalitions at the European level: a bourgeois alliance of Christian democrats, conservatives, and industrial groups; a labor alliance of social democrats and labor unions; and an alliance of “the weak,” composed of greens, regionalists, and environmental 4

5

Kris Deschouwer makes the point that differences across policy sectors are particularly pronounced under multilevel governance (2000: 11). On complexities of the left/right dimension see Elff (2002). Similarly, van der Eijk and Franklin note that left/right placement is correlated only with their more policy-relevant measure of preference for unification. Gabel and Hix note that left/right distinguishes European political parties on economic issues, but not on basic constitutional questions such as what powers should be delegated to the European level.

242

Gary Marks

and consumer groups. The latter coalitions are most in favor of further Europeanization and strengthening the European Parliament. We find that the location of national political parties on the left/right divide constrains whether they support or oppose European integration on policies related to regulated capitalism. Employment policy is a prime example. The further to the left a party is located on the left/right dimension, the greater its support for a European employment policy (Hooghe, Marks, and Wilson, chapter 6; Thomassen and Schmitt 1997: 172). The relationship is strongest for mainstream parties – social democratic, Christian democratic, liberal, and conservative parties (r = 0.75) – but it is significant across all parties, despite the fact, as noted above, that extreme left and extreme right parties tend to be opposed to just about any shift of competence to the European level. Left/right positioning exerts a similar constraint on cohesion policy, which funds infrastructural and training programs in poorer regions in an effort to increase their economic growth. The issues that motivate the classic left/right divide have to do with liberty versus equity, free markets versus government steering, and individual economic freedom versus collective rights. These encompass a fair share of conflicts in capitalist society, but they do not bear directly on questions relating to the territorial allocation of authority in a multilevel polity. In the past, socialists have fought for state centralization to counterbalance the power of property and concentrated private ownership of industry, but this does not translate into the demand for more authority at the European level. The reason for this is that social democrats are also defenders of the national institutions they have done so much to create. To the extent that social democrats have successfully created national systems of welfare, industrial relations, and health care, they fear that European integration may undermine them by intensifying regulatory competition. Many social democrats echo Fritz Scharpf in stressing that the EU is biased towards negative integration, that is, towards market-creating and market-enabling policies, rather than market regulation (Scharpf 1996; 1999). Only if further integration were to undo this bias could one be sure that shifting competencies to the European level would be a step toward regulated capitalism. If the bias remains – and it is deeply rooted in the European Court of Justice – then Europeanizing public policy will be self-defeating from a social democratic perspective.6 Hence, the left/right divide does not speak directly to the territorial allocation of authority. 6

Euroskepticism on the left is reinforced by the view that the weakness of a European identity precludes redistributive policy at the European level. According to this line of argument, the absence of a meaningful European demos limits the legitimacy of the European Parliament, and hence its effectiveness for European regulated capitalism.


243

This is what Jacques Thomassen, Abdul Noury, and Erik Voeten find (chapter 7). They diagnose three distinct issue dimensions for members of the European Parliament. The first is an integration/independence dimension composed mainly of constitutional issues that engage the territorial allocation of authority (including the general question of increasing the range of responsibilities of the EU, and strengthening the European Parliament). The second is a left/right dimension that extends the concern with state and markets to the European level (including whether to create an EU employment program versus concentrating on the single market), and the third is a libertarian/traditional or new politics dimension based on law and order and lifestyle issues. The left/right dimension constrains support for European issues that have distributional consequences within, rather than among, member states. This is consistent with a bounded rationality explanation of response to European integration. The strategic response of an organization to new issues depends on its prior ideology, which acts as a lens through which it interprets new opportunities or challenges arising on the political agenda (Kitschelt, Lange, Marks, and Stephens 1999). Groups that mobilize functional interests within national states – political parties and functional interest groups – are particularly responsive to the distributional effects of a European issue across domestic groups. They are primed, so to speak, to interpret European integration in the light of their ideological concerns. Conversely, organizations (such as national and regional governments) that mobilize territorial interests are particularly responsive to the distributional effects of issues among territorial units, as I discuss below.

New politics By the same logic, one would expect that EU issues engaging lifestyle, gender, environment, participatory decision-making, and national culture to be most closely associated with the new politics dimension within member states. Items that load strongly on the new politics dimension include those that ask whether protecting consumers, controlling immigration, increasing EU transparency, protecting human rights, and protecting national cultures should be key priorities for EU activity (see Gabel and Anderson, chapter 1). Hooghe, Marks, and Wilson (chapter 6) find that the position of political parties on the new politics dimension is strongly associated with their support for an EU environmental policy (r = 0.61; sig. > 0.01) and for an EU asylum policy (r = 0.57; sig. > 0.01).

244

Gary Marks

In some respects, the new politics dimension is more intimately connected to European integration than is the left/right dimension. New politics conflicts engage the “nation” and its defense, alongside individual choice versus traditional values, the environment, and participation versus hierarchy. Those on the right of this dimension oppose European integration for essentially the same reasons that they oppose immigration: both infuse foreigners into the society; both threaten the national community. The Flemish Block’s campaign slogan in the 1999 Belgian election was “In charge of our own country,” an update on their earlier “Our own people first.” The defense of national sovereignty lies close to the heart of those on the TAN (traditional–authoritarian–nationalist) side of this divide, not because national sovereignty is useful for other ends, but because it is intrinsically valued. This distinguishes the new right from market liberals. Market liberals view national sovereignty in terms of its implications for economic exchange. They are opposed to barriers to trade, and they therefore support strong international regimes that can facilitate market integration, even if this eviscerates key national state competencies, including monetary control. At the same time, market liberals oppose the creation of a powerful and legitimate continental authority that could be used to control markets. The European orientations of those on the right of the left/right divide are nuanced, unlike those on the new right. Hooghe, Marks, and Wilson (chapter 6) conclude that a party’s position on the new politics dimension is considerably more powerful than its position on the left/right dimension in predicting its support for integration across each of the seven issue areas they examine. Radical right parties are now by far the most Euroskeptical of any of the eight party family groupings in Europe, including the radical left. Conservative parties that lean to the TAN side of the new politics dimension – emphasizing traditional or authoritarian values – tend to be more Euroskeptical than those that do not. The relationship is weaker for new politics/green parties on the other side of this dimension, except on issues, such as the environment and asylum, that relate directly to their core concerns. This is consistent with Wessels’ finding that of the three alliances he distinguishes, it is the new politics alliance that is most supportive of European integration. Models Of the four models that we set out in the Introduction, the one that appears most valid at the level of issues is the Hooghe–Marks model. Several contributors to this volume stress that the moderate left has become more


245

supportive of European integration on issues related to regulated capitalism, and that the moderate right has become skeptical of integration beyond market-making. But the pro-EU orientation of the moderate left is not written in stone. It was, for example, not evident before the great market reforms of the 1980s. In 1984, according to Leonard Ray’s data on party positions, social democrats supplied the largest reservoir of opposition to European integration (1999). However, since the early 1990s, debates about the balance of European and national policy-making are intelligible in left/right terms. The research reported in this book goes considerably beyond the models set out in the Introduction. First, we have discovered a connection between new politics and European integration. There is good reason – and some evidence – to believe that new politics contestation is intimately related not only to particular policy choices, but also to fundamental constitutional issues raised by European integration. Such a connection is consistent with the dogged opposition to European integration on the part of radical right parties in recent national elections, including those in France and the Netherlands in 2002, where the National Front and the List Pim Fortuyn raised the heat on European integration. Conservative parties in the UK and France have been deeply riven by conflicts between market liberals who are pragmatic on issues of national sovereignty and new politics rightists who reject European supranationalism. Furthermore, the models we set out in our Introduction say nothing about territorial variation in the ideological bases of European contestation. The Hooghe–Marks model hypothesizes variation among the issues that make up European integration, but assumes that the European Union is ideologically homogeneous. We need now to relax this assumption and theorize about how left/right structures positioning on European integration in different countries. Territorial variation is as fundamental to an understanding of European contestation as variation across issues, and I turn to this topic next. Disaggregating by territory The European Union both tames and intensifies territorial politics. It tames territorial politics by creating a web of mutual dependencies that reduce – and perhaps eliminate – the possibility of war among member states. The EU routinizes international relations within a system of multilevel governance. It internalizes – and domesticates – territorial relations by transforming diplomacy among states into the making, implementation, and adjudication of laws. I have already described one decisive outcome of this process: ideological conflicts that cross-cut

246

Gary Marks

territorial conflicts and that meld domestic groups of one ideological stripe or another into transnational coalitions. This was Jean Monnet’s hope and goal, but it is one side only of European integration. European integration intensifies territorial politics and intensifies the national. It does this both within and among countries. It does so within countries by generating insecurities that provoke a nationalist reaction. Chapters in this volume detail the way in which the radical right mobilizes anti-European feelings in defense of national authenticity. European integration undermines national sovereignty – and citizens understand this. Barriers to economic competition within Europe have been dismantled at the same time as the capacity of national states to ameliorate the effects of competition – either through welfare or fiscal subvention – has narrowed. Citizens who have the least to gain from economic integration because they lack mobile skills and capital, and who feel personally vulnerable, are most likely to support a TAN ideology: traditionalism, authoritarianism, nationalism. Root-and-branch opposition to European integration fits comfortably with reactionary defense of the nation, and radical right parties now make up the largest voting bloc of outright opposition to European integration across the EU. European integration exacerbates territorial conflict among countries because it engages national (and regional/local) governments in a process of ongoing bargaining over a range of policies that formerly were determined within, rather than among, national states. To the extent that such policy-making involves redistribution and the territories in question have a capacity for strategy – that is to say, they are collective actors, not merely aggregations – so the outcome will be intense territorial bargaining. Intergovernmental institutions – the European Council and the Council of Ministers – are key venues for such bargaining. In the European Parliament such territorial conflicts fragment ideologically based party fractions. Territorial variation may refract ideological coalitions. The allocation of a particular policy competence to the European level may have dissimilar – or even contrary – consequences in different countries. Distinctive political and economic institutions filter how actors apply their ideological preferences to European issues (cf. Eichenberg and Dalton 1998; Hall and Gingerich 2001; Hall and Soskice 2001; Hix 1999b; Kitschelt, Lange, Marks, and Stephens 1999). This is the point of departure for the chapters by Cees van der Eijk and Mark Franklin (chapter 2), Leonard Ray (chapter 3), and Adam Brinegar, Seth Jolly, and Herbert Kitschelt (chapter 4). Van der Eijk and Franklin observe that “It is the dynamics of the domestic political arena that here and there brings forth a connection with either the right or the left.” Ray’s thesis is that:


247

In those nations where the prevailing national policy regime is closer to the ideal preferences of “leftist” individuals, support for integration should be concentrated on the right side of the political spectrum. Conversely, in nations where national policy is relatively far from the preferences of “leftists,” the left will be more supportive of integration as a way to achieve, at the European level, outcomes unobtainable under a purely national regime (chapter 3).

Brinegar, Jolly, and Kitschelt find compelling evidence that “In redistributive welfare states it is the left that opposes further EU integration, in liberal-residual welfare states the right” (chapter 4). The chapters in this book analyze the domestic ideological underpinnings of the debate concerning European integration. We have gone furthest in probing how European integration is connected – or not connected – to the left/right divide in a variety of arenas for several kinds of actors. This was our main objective when we began this project. We have raised the question of how European integration is related to the new politics dimension. And we have begun to inquire into the way that territory mediates ideology across the European Union. The final section of this chapter no longer encapsulates the findings of previous chapters, but takes some tentative steps in linking two lines of analysis we discuss in some detail throughout the book. Readers who seek in this conclusion a concise overview of our project and its conclusions can stop here. What follows is meant, at best, to suggest some fruitful avenues for further inquiry.

Combining issues and territory Let me begin by combining two basic strategies adopted by the contributors to this book: an issue-based approach, and an appreciation of the way in which territory (e.g., via national institutions) mediates ideology. When one analyzes the territorial dimension of European contestation from the standpoint of variation across issues, it is useful to consider the mediating effect of national institutions as a variable rather than a constant. At one extreme, there are issues that have decisively different distributional consequences from country to country. As Brinegar, Jolly, and Kitschelt inform us, the distributional impact of a European welfare policy in a liberal uncoordinated economy, such as the UK, is very different from that in a social democratic coordinated economy, such as Sweden. Hence, left and right take different positions in different countries. At the other extreme are issues – the decision whether to expand EU competencies in higher education, for example – that have consistent domestic distributional implications across EU countries. In such cases, one would

248

Gary Marks

expect the pattern of support and opposition from left and right to be the same across countries. Figure 11.1 hypothesizes how this variation constrains the positions that individuals and groups adopt on an issue and the coalitions they form. The idea that motivates this model is simple: the greater the territorial congruence of a policy’s distributional impact, the more one can expect domestic actors to line up the same way.7 The extent to which one finds ideologically based coalitions depends on whether a policy has the same distributional effects across its territorial subunits.8 Conversely, the greater the territorial heterogeneity of distributional effects across territory, the more one can expect ideologically inconsistent coalitions. In this case the outcome will be “unholy alliances” – ideologically mixed coalitions that combine left and right groups on both sides of the issue.9 A policy may also engage territory directly by distributing values across constituent units. What might one expect when European integration allocates values not only within countries, but across them? This question is logically independent from the question of the homogeneity of domestic distributional impacts. The reasoning here follows the same format as in figure 11.1, but with a twist: how does the allocation of values across constituent units affect coalition building?10 The logic of this analysis is multilevel. It can apply to states in a transnational polity, regions within a state, or localities within a region. Figure 11.2 summarizes these ideas. It explains coalitions in terms of the interaction of territory and ideology. The y-axis in figure 11.2 represents the extent to which the domestic distribution of costs and benefits is homogeneous across territory, in this case, across the member states of the European Union. If domestic costs/benefits are similar across countries, two things follow: first, left and right can be expected to line up in a consistent way across countries; second, European-wide, ideologically pure, coalitions will emerge. The x-axis represents the extent to which an issue involves distribution across territory (i.e., across EU member states). If there is extensive distribution of values across territory, 7

8 9

10

“Homogeneity/heterogeneity of distributional impact” and “consistency of ideological positioning” may serve as useful concepts for evaluating political cohesion in federal political systems. These concepts provide analytical leverage on the following question: “To what extent is there a common ideological playing field across a particular territory?” In the European Union, the chief territorial subunits are the member states, but the logic of the model applies within as well as among countries. Figure 11.1 spells out these pure types as answers to the questions posed in the thicklined boxes. The model is recursive in that coalitions make policy, which affects the allocation of values (the dotted arrows in figure 11.1). I define “values” in their broad, Eastonian, sense. Distribution (or allocation) of values involves who is allowed to do what as well as who gets what.


249

POLICY- MAKING

ALLOCATION OF VALUES

POLITICAL PREFERENCES

Costs and benefits of the policy within units (e.g., countries)

The political positions taken by individuals/groups

Is the internal distribution of costs and benefits the same or different across countries?

Are ideological positions (e.g., of left and right) consistent or inconsistent across territory?

COALITIONS Transnational patterns of support and opposition Are coalitions ideologically coherent?

The distributional effects of the policy are the same across countries.

Consistent across countries

IDEOLOGICAL

The distributional effects of the policy vary across countries.

Inconsistent across countries

“UNHOLY”

Figure 11.1 A model of coalition formation.

two things follow: first, this will give rise to territorial conflict; second, coalition-building will be territorial. Keep in mind that the two logics of allocation, preferences, and coalition-formation – ideology and territory – are intrinsically independent of each other. Hence, particular policies may give rise to one, both, or neither of the patterns of coalition-formation. Figure 11.2 illustrates four possibilities: r Ideology trumps territory in the lower-left quadrant. Here coalitions arise from a consistent pattern of distributional conflicts within countries in the absence of distributional conflicts among countries. r Ideology and territory are both powerful sources of coalition-building in the bottom-right quadrant. This quadrant describes a consistent pattern of distributional conflicts within countries, but one that is cross-cut by territorial coalitions arising from high levels of redistribution among countries. r Neither ideological nor inter-state conflict structures coalitions in the top-left quadrant. Territorial distribution is low, and national institutions filter the impact of issues so that the positioning of left and right varies from country to country – creating “unholy alliances.”

250

Gary Marks Is the domestic distributional impact of an EU policy similar or different across countries? DIFFERENT

IDEOLOGY MEDIATED BY TERRITORY: UNHOLY ALLIANCES Is the intercountry distributional impact of an SMALL EU policy large or small?

TERRITORY DOMINANT: NATIONAL COALITIONS

LARGE

IDEOLOGY DOMINANT: EUROPEAN-WIDE COALITIONS

SIMILAR

Figure 11.2 Patterns of contestation.

r Territory, but not ideology, structures coalition-building in the top-right quadrant. This quadrant combines unholy alliances with national coalitions arising from high levels of inter-country distribution. Ideological lines of conflict are muddied by national institutions; inter-state conflict is dominant. The analyses of this volume generate a set of expectations about the location of actors, institutions, and issues within this two-dimensional space. With respect to actors, we expect the positioning of political parties on European issues to be more ideologically consistent in left/right terms than the positioning of citizens. That is, on any given issue, political parties should be located further towards the bottom of figure 11.2. Contributions to this volume indicate that the preferences of national political parties on a subset of issues arising from European integration


251

are sometimes quite closely related to left/right position – certainly far more so than for the public at large. I have already noted that Gabel and Anderson and van der Eijk and Franklin discover only a weak and insignificant linear association between left/right positioning and support for European integration among the general public. In contrast, Thomassen, Noury, and Voeten find that the left/right dimension explains 23 percent of the variance in MEP preferences across fifteen issues. Hooghe, Marks, and Wilson conclude that the effect of left/right positioning for national parties is significant across a subset of European issues, including employment and environmental policy. We therefore expect that left/right constrains party positions on European integration more strongly than it constrains the positions of citizens. This expectation can be grounded in political psychology. One of the best-established generalizations in the study of political attitudes is that there is wide variance among citizens in their political knowledge and sophistication, and that this is associated with the extent to which citizens structure their views of the political world (Jennings 1992; Kinder 1998). Elites are more likely to understand and use political abstractions, such as “left” and “right.” Correspondingly, their attitudes towards political objects are usually more stable and more structured. Party leaders do not just structure their views coherently, they also inform the public about new issues that arise and how they should be evaluated (Steenbergen and Scott, chapter 8). In terms of arenas, our expectation is that those dominated by political parties – e.g., national parliaments and the European Parliament – should be located further toward the ideological quadrants at the bottom of figure 11.2 than arenas dominated by territorial organizations, such as national governments. Political parties compete on ideology; national governments compete by representing distinct territorial communities.11 The logical implication is that the European Council and the EU’s Council of Ministers will be biased towards the top right-hand quadrant of figure 11.2, while European and national parliaments will be biased towards the bottom left-hand quadrant. Coalition-building in these arenas is likely to be mixed if both sources of distribution are present. Each arena exhibits this in a characteristic way. European-wide ideological coalitions are most visible in the European Parliament, but in the face of territorial redistribution one can usually trace national tensions within party fractions. Territorial coalitions 11

Political parties representing the demands of particular territorial minorities are an interesting exception.

252

Gary Marks

are most visible in the Council of Ministers, but on ideologically salient issues, these are modulated by the party composition of national governments.12 Finally, the authors of this volume find wide variations in the interaction of ideology and territory across issues. There are two big questions here. What kinds of issues are likely to generate ideological (in)consistency across the left/right divide? How are countries grouped on whether the left or the right is most supportive of European integration? Let us take these questions in turn. Our expectations on ideological consistency across the left/right divide are informed by the regulated capitalism hypothesis. In recent years the left has come to support further integration on issues that flank market integration. These are policy areas – employment, the environment, social policy, cohesion policy – that were part of Jacques Delors’ project to create a citizens’ Europe. We would therefore expect these policies to be placed in the lower-left quadrant of figure 11.2. At the opposite extreme (in the top-right quadrant) are constitutional issues. These often allocate values (power or resources) across countries, but have murky consequences for the allocation of values within countries. So, for example, the distribution of voting weights in the Council of Ministers engages countries as contending players, but has no clear consequences for who gets what within countries. Thomassen, Noury, and Voeten find that membership in a particular party family shapes the views of MEPs on socio-economic left/right issues and libertarian-traditional issues, while nationality is the most powerful influence for constitutional issues. Foreign and defense policy are similar in that they allocate values among countries more transparently than they allocate values within countries. Our expectation concerning national patterns of conflict is rooted in the varieties of capitalism literature (Hall and Soskice 2001; Soskice 1999). Assuming that European integration leads to policy convergence, integration should be supported by those on the right and opposed by those on the left in leftist policy regimes, and supported by those on the 12

On the role of territoriality in the EU see Egeberg (2001), Sbragia (1993) and Ansell and Di Palma (forthcoming), particularly chapters by Christopher Ansell, Stefano Bartolini, Giuseppe Di Palma, and Sidney Tarrow. The effects of cross-cutting pressures surface in research on the European Parliament, where voting can be measured. Thomassen, Noury, and Voeten find a high level of party group discipline in EP voting, “an indicator of the success of European parliamentary institutions in framing European politics according to ideological and party lines rather than national interests” (Thomassen, Noury, and Voeten, this volume; Hix, Noury, and Roland 2003).


253

left and opposed by those on the right in rightist policy regimes (Ray, chapter 3).13 A second line of theorizing explores how variation among party systems structures debate (Steenbergen and Scott, chapter 8). I have noted that there is a strong curvilinear relationship between party positioning on the left/right divide and support for European integration. Extreme parties of the left and right mobilize publics against European integration, and where only one or the other is present, one can expect this to have a significant impact on the sign of the coefficient for left/right positioning on European integration. Where an extreme left party exists in the absence of an extreme right party, I expect a positive association between left/right position and support for European integration. Conversely, where an extreme right party exists in the absence of an extreme left party, this should give rise to a negative association. Locating issues in the schema Where are European issues located in the two-dimensional space conceptualized here? Figure 11.3 illustrates where national political parties stand on seven policies, and figure 11.4 illustrates public opinion positions on thirteen policies.14 13

14

As Brinegar, Jolly, and Kitschelt argue: “In residual, liberal welfare states, leftists who would like to see more economic redistribution would obviously see European integration as a benefit, if it moves national conditions from the status quo to at least a conservative, but more encompassing and redistributive welfare state. Rightists, in that setting, will be opposed. Conversely, in encompassing, egalitarian, universalistic, social democratic welfare states, leftists who are fond of the national status quo can only fear that European integration will lead away from their personal ideal point. In such countries, the left is likely to be more anti-European and the right more pro-European . . . In countries with conservative encompassing welfare states, EU integration should not be significantly related to left/right ideology, but cross-cutting” (chapter 4). Figure 11.3 is derived from expert evaluations of positions of national parties on a sevenpoint scale that ranges from strongly opposes integration to strongly in favor of integration (Marks and Steenbergen 1999). For example, our question on the EP is as follows: “We would like to start with the party leadership’s position on the powers of the European Parliament. Some parties advocate that the powers of the European Parliament should be drastically expanded, to remove the so-called ‘democratic deficit.’ Other parties argue that the powers of the European Parliament are already extensive and that there is no need to expand these powers further. In your judgment, where does the leadership of the parties listed below stand vis-à-vis expansion of the powers of the EP?” Figure 11.4 is based on question 30 in Eurobarometer 54.1, conducted in the fall of 2000, which reads: “For each of the following areas, do you think that decisions should be made by the [nationality] government, or made jointly within the European Union? 1 = nationality, 2 = jointly within the European Union, 3 = don’t know.” These scores are recoded in analysis.

254

Gary Marks 50

IDEOLOGICAL CONSISTENCY (LEFT/RIGHT)

Foreign policy

Fiscal policy

60 EP powers Regional policy

70

Asylum policy 80

90

100

0.6

Employment policy

0.8

Environment policy

1.0

1.2

1.4

1.6

1.8

2.0

DIVERSITY OF NATIONAL PREFERENCES Figure 11.3 Party positions on European issues.

Ideological consistency across territory (the y-axis) is operationalized as the proportion of countries where left and right line up in the same way. I measure this by estimating regressions for left/right position on support for European integration for each policy in each country, and comparing the signs of the coefficients (i.e., the slope of the regression line). A policy located at the bottom-most point on the y-axis indicates that left and right take consistent positions with respect to each other in all countries. A policy located at the top-most point indicates that in 50 percent of countries the left takes one position with respect to the right, and in the remaining 50 percent the positions of left and right are reversed. The reference line in figures 11.3 and 11.4 is located at the point where 75 percent of EU countries have a consistent pattern of left/right support. This may be a conservative benchmark from which to evaluate ideology in the European Union: it demands that a given pattern of left/right support is three times as frequent as the alternative. I measure diversity of national preferences for each policy by calculating the mean score for political parties (figure 11.3) or for individuals (figure 11.4) in each country, and then calculating the interquartile range


255

50 Agricultural policy

Foreign policy

IDEOLOGICAL CONSISTENCY (LEFT/RIGHT)

Environmental policy 60

Regional policy

Health policy

Defense policy

R&D

Currency policy

70

Education policy 80

Social policy Aid policy

Employment policy Asylum policy

90

100

0.0

0.1

0.2

0.3

DIVERSITY OF NATIONAL PREFERENCES Figure 11.4 Public opinion on European issues.

of these country averages. The interquartile range for each policy is represented on the x-axis.15 The results for ideological consistency are roughly in keeping with the expectations set out above. In figure 11.3, two policies have almost complete ideological consistency: employment and environment policy. These policies are integral to regulated capitalism, and directly engage the question of the role of the state in the economy. Foreign policy, fiscal policy, and the powers of the European Parliament are, as we expect, 15

The scale in figure 11.3 is support for integration measured in six intervals from 1 (strongly opposed to European integration in this policy area) to 7 (strongly in favor of integration in this policy area). The scale in figure 11.4 is from zero (strongly opposed to European integration in this policy area) to 1 (strongly in favor of integration in this policy area). The simple correlation of interquartile range scores for the five common policy areas in figures 11.3 and 11.4 is 0.55, which is not statistically significant given the small number of cases. It is interesting to note that the standard deviations of national preferences aggregated for national political parties are significantly associated (r = 0.83, sig. = 0.081) with the standard deviations among government positions across five policy areas that are common to the Marks and Steenbergen data set and the one compiled by Mark Aspinwall (2002).

256

Gary Marks

the least consistent. Regional policy, however, is further north than we expect. Comparison of figures 11.3 and 11.4 reveals that political parties are more ideologically consistent across the EU than is the general public. Of the five policies represented in both figures, political parties are more consistent on employment, environmental, regional, and foreign policy, and less consistent only on asylum policy. The simple correlation between public and party scores for ideological consistency on these five policies is 0.46: insignificant given the small number of cases. In figure 11.4, as in figure 11.3, policies that have to do with regulated capitalism exhibit relatively high levels of ideological consistency. Education, employment, and social policy are in this camp, while Third World aid and asylum policy have obvious connections to it. However, the public shows much less ideological consistency on environmental policy than do political parties. As we expect, foreign policy and defense policy have little ideological consistency. Before concluding, let us take a brief look at how ideology plays out across the territory of the EU. Certain countries stand out as exceptions to the patterns of ideological consistency illustrated in figures 11.3 and 11.4. Across the seven policy areas and fourteen countries for which we have data for national political parties (108 cases in total), there are twenty-seven cases where the right is more pro-integration than the left. Three countries – Denmark, Greece, and Sweden – account for seventeen of these.16 In the remaining countries, the connection between left/right and European integration is consistent: the left is most favorably oriented to European integration in 89 percent of country/policy cases. For the general public, 59 cases out of a total of 182 (thirteen policy areas across fourteen countries) have either no left/right constraint on European positioning, or have the right more favorably oriented than the left. Three countries stand out. Sweden, Denmark, and Finland account for thirty-four, or a little over half, of these fifty-nine cases. In the remaining countries, the left is favorably oriented to European integration in 83 percent of country/policy cases. Variation in the articulation of left and right on European integration appears as great, or greater, across territory as across policy areas. So, clearly, there are strong national patterns in the data. They appear to broadly match expectations derived from variation among types 16

The concentration of contrary cases in Denmark, Greece, and Sweden is even greater when we undertake the same analysis for the consistency of new politics ideology. These three countries account for eighteen of twenty-five contrary cases.


257

Table 11.1 Country and ideology Proportion of policies in a country where the left supports European integration and the right opposes National political parties Presence of extreme left party in absence of extreme right party Scandinavia R-squared N (EU minus Luxembourg)

2.57∗∗ 0.32 14

Public 2.79∗ −9.17∗∗∗ 0.68 14

Note: ∗ p < 0.1, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

of capitalism and variation among party systems. For public attitudes, the simplest explanation of national variation is to invoke a proper name: Scandinavia (Sv˚asand and Lindstrom ¨ 1996). A dichotomous variable that splices the Scandinavian countries apart from the rest explains almost 60 percent of variance where the dependent variable is the number of policies for each country in which the left is more supportive of European integration than the right. This is in line with Ray’s and Brinegar, Jolly, and Kitschelt’s argument that the prevailing type of capitalism in a country interacts with personal ideology, and that in redistributive social democratic welfare states, it is the left that has most to lose from European integration, while the right has most to gain. This variable does not have much strength for political parties. To explain the territorial pattern of ideology for parties it may be more fruitful to examine a more proximate factor: the presence/absence of extreme parties. A dichotomous variable with the value 1 for a country having an extreme left party but no extreme right party is significantly associated with the ideological consistency of public support for European integration, explaining an estimated 32 percent of the variance (see table 11.1). Concluding remarks Our focus in this volume has been on how conflict over European integration connects – or does not connect – with the dimensions of contestation that structure politics within European societies. Our endeavor is “second generation” in that it departs from the analysis of levels of support for

258

Gary Marks

European integration, which has dominated the field of EU opinion research for the past twenty years. Whereas levels of support can change quite quickly over time (Eichenberg and Dalton 1998), patterns of conflict are far more stable. Lipset and Rokkan stressed that the interaction of social cleavages, rooted in the national revolution, Protestant Reformation, and industrial revolution in the sixteenth to early twentieth centuries, maintained a tight grip on party competition through the 1960s and, they speculated, beyond. An adjective that comes to their minds when discussing change is “glacial.” By following them in investigating patterns of conflict, we seem to be on firm ground. Or are we? Recall that in left/right terms, the character of support and opposition for European integration has changed decisively over the past two decades. The European Union is a moving target, not just because it evokes quite rapidly changing levels of support, but because the essential nature of the beast has been transformed from a market-making to a polity-making process. Twenty years ago, the right supported further European integration in order to achieve an integrated market. The bulk of the opposition came from the moderate and far left. Today, it is the right, particularly the radical right, that opposes further integration, especially in the areas most favored by the moderate left – e.g., in employment policy, environmental policy, and asylum policy. The left is more favorably inclined to integration in every single policy represented in figures 11.3 and 11.4 but one: foreign policy. As economic and monetary integration have passed from contentious issues into accomplished facts, so the focus of debate has shifted from creating a market to regulating it. As a result, conflict about the future of the EU more closely resembles conflict within member states, pitting a left in favor of a more active, caring government against a right defending markets and economic freedom. If one takes the period as a whole, it is clear that there is no intrinsic connection between being on the left or right and being pro- or antiintegration. In principle, as the Hix–Lord model assumes, the territorial organization of authority is orthogonal to functional conflicts that motivate left vs. right at the domestic level. Alternative architectures of multilevel governance do not translate, once and for all, into left vs. right conflict. The Hooghe–Marks model is most appropriate for the post-1980s when one breaks open European integration into its constituent issues. A significant subset of European issues involves distributional choices that are closely related to left/right conflict. But our efforts to understand precisely how European integration is linked to domestic conflicts have raised questions that go beyond the models we began with. Three stand out as challenges for future research.


259

We need to theorize territorial variation in ideological positioning as well as variation across issues and across time. We need to explore links between new politics and European integration alongside those between the left/right divide and European integration. And we need a theory of coalition-building that encompasses conflict among constituent polities as well as conflict within them.

References

Achen, Christopher H. 1991, “What Does ‘Explained Variance’ Explain? Reply,” Political Analysis 2: 173–84. Allardt, Erik 2001, “Party Systems and Voter Alignments in the Tradition of Political Sociology,” in Karvonen, L. and Kuhnle, S. (eds.), Party Systems and Voter Alignments Revisited, Routledge, pp. 13–26. Andeweg, Rudy 1995, “The Reshaping of National Party Systems,” West European Politics 18: 58–79. Anon. 1998, “Jorg ¨ Haider, Austria’s (and Europe’s) Border Guard,” Economist, July 11, p. 55. Ansell, Christopher K. and Di Palma, Giuseppe, forthcoming, On Restructuring Territoriality: Europe and North America, Cambridge University Press. Ansell, Christopher K., Parsons, Craig A., and Darden, Keith A. 1997, “Dual Networks in European Regional Development Policy,” Journal of Common Market Studies 35: 347–75. Armingeon, Klaus 1994, Staat und Arbeitsbeziehungen: Ein internationaler Vergleich, Westdeutscher Verlag. Aspinwall, Mark 2002, “Preferring Europe: Ideology and National Preferences on European Integration,” European Union Politics 3: 81–111. Attina, Fulvio 1990, “The Voting Behavior of European Parliament Members and the Problem of Europarties,” European Journal of Political Research 18: 557–79. Austen-Smith, David and Banks, Jeffrey 1988, “Elections, Coalitions, and Legislative Outcomes,” American Political Science Review 82: 405–22. Axelrod, Robert 1970, Conflict of Interests: A Theory of Divergent Goals with Applications to Politics, Markham. Aylott, Nicholas 2001, “The Swedish Social Democratic Party,” in Notermans, T. (ed.), Social Democracy and Monetary Union, Berghahn Books, pp. 149– 74. Bache, Ian (ed.) Forthcoming, Themes and Issues in Multi-Level Governance, Oxford University Press. Baker, David, Gamble, Andrew, Ludlam, Steve, and Seawright, David 1997, “The 1994/96 Conservative and Labour Members of Parliament Surveys on Europe: The Data Compared,” presented at American Political Science Association Meeting, Washington, DC. Baker, David, Gamble, Andrew, and Seawright, David 1999, “Open Regionalism or National Political Economy? New Labour and a Conservative Consensus 260

References

261

on the EU?” presented at American Political Science Association Meeting, Atlanta. Banchoff, Thomas F. and Smith, Mitchell P. 1999, Legitimacy and the European Union: The Contested Polity, Routledge. Bardi, Luciano 1994, “Transnational Party Federations, European Parliamentary Party Groups and the Building of Europarties,” in Katz, R. S. and Mair, P. (eds.), How Parties Organize: Change and Adaptation in Party Organizations in Western Democracies, Sage, pp. 357–72. Barnes, Samuel H. and Kaase, Max 1979, Political Action: Mass Participation in Five Western Democracies, Sage. Barnouin, Barbara 1986, European Labour Movement and European Integration, Frances Pinter. Bell, David S. and Lord, Christopher (eds.) 1998, Transnational Parties in the European Union, Ashgate. Benedetto, Giacomo 2001, “Scepticism and its Antithesis: The Case of Eurosceptics in the European Parliament,” presented at European Community Studies Association Meetings, Madison, WI. Bentler, Peter M. 1990, “Fit Indexes in Structural Models,” Psychological Bulletin 107: 238–46. Bentler, Peter M. and Bonnett, David G. 1980, “Significance Tests and Goodness-of-Fit in the Analysis of Covariance Structures,” Psychological Bulletin 88: 588–606. Bentley, Arthur F. 1908, The Process of Government, University of Chicago Press. Betz, Hans-Georg 1994, Radical Right-Wing Populism in Western Europe, St. Martin’s Press. Betz, Hans-Georg and Immerfall, Stefan (eds.) 1998, The New Politics of the Right: Neo-Populist Parties and Movements in Established Democracies, St. Martin’s Press. Beyers, Jan and Kerremans, Bart 2001, “Diverging Images of Consensus: Belgium and its View on European Integration,” in Goldmann, K. and Gilland, K. (eds.), Nationality versus Europeanisation: The National View of the Nation in Four EU Countries, Department of Political Science, Stockholm University, pp. 126–65. Bogdanor, Vernon 1989, “Direct Elections, Representative Democracy and European Integration,” Electoral Studies 8: 205–16. Boix, Carles 1998, Political Parties, Growth and Equality: Conservative and Social Democratic Economic Strategies in the World Economy, Cambridge University Press. Bollen, Kenneth A. 1989, “A New Incremental Fit Index for General Structural Equation Models,” Sociological Methods and Research 17: 303–16. Bomberg, Elizabeth E. 1998, Green Parties and Politics in the European Union, Routledge. Borzel, ¨ Tanja and Risse, Thomas 2000, “Who is Afraid of a European Federation? How to Constitutionalize a Multi-Level Governance System?” in Joerges, C., Mény, Y., and Weiler, J. H. H. (eds.), What Kind of Constitution for What Kind of Polity?: Responses to Joschka Fischer, European University Institute, pp. 45–59.

262

References

Brug, Wouter van der, Franklin, Mark N., and Eijk, Cees van der 2000, “The Economy and the Vote: Electoral Responses to Economic Conditions in 15 Countries,” presented at American Political Science Association Meetings, Washington, DC. Brzinski, Joanne Bay 1995, “Political Group Cohesion in the European Parliament,” in Rhodes, C. and Mazey, S. (eds.), The State of the European Union, vol. III: Building a European Polity?, Lynne Rienner Publishers, pp. 135–58. Budge, Ian and Farlie, Dennis 1983, “Party Competition: Selective Emphasis or Direct Confrontation? An Alternative View with Data,” in Daalder, H. and Mair, P. (eds.), Western European Party Systems: Continuity and Change, Sage, pp. 267–305. Budge, Ian and Klingemann, Hans-Dieter 2001, “Finally! Comparative OverTime Mapping of Party Policy Movement,” in Budge, I., Klingemann, H.-D., Volkens, A., Bara, J., and Tanenbaum, E. (eds.), Mapping Policy Preferences: Estimates for Parties, Electors, and Governments 1945–1998, Oxford University Press, pp. 19–50. Budge, Ian, Robertson, David, and Hearl, Derek (eds.) 1987, Ideology, Strategy and Party Change: Spatial Analysis of Post-War Election Programmes in 19 Democracies, Cambridge University Press. Budge, Ian, Klingemann, Hans-Dieter, Volkens, Andrea, Bara, Judith, and Tanenbaum, Eric (eds.) 2001, Mapping Policy Preferences: Estimates for Parties, Electors, and Governments 1945–1998, Oxford University Press. Bush, Evelyn and Simi, Pete 2001, “European Farmers and their Protests,” in Imig, D. and Tarrow, S. (eds.), Contentious Europeans: Protest and Politics in an Emerging Polity, Rowman and Littlefield, pp. 97–121. Caporaso, James A. 1996, “The European Union and Forms of State: Westphalian, Regulatory or Post-Modern?” Journal of Common Market Studies 34: 29–52. Caporaso, James A. 2000, “Changes in the Westphalian Order: Territory, Public Authority, and Sovereignty,” International Studies Review 2: 1–28. Caporaso, James A. and Keeler, John S. 1995, “The European Community and Regional Integration Theory,” in Rhodes, C. and Mazey, S. (eds.), The State of the European Union, vol. III: Building a European Polity?, Lynne Rienner Publishers, pp. 29–62. Castles, Francis M. and Mair, Peter 1984, “Left-Right Political Scales: Some ‘Expert’ Judgments,” European Journal of Political Research 12: 73–88. Christensen, Dag Arne 1996, “The Left-Wing Opposition in Denmark, Norway, and Sweden: Cases of Europhobia?” West European Politics 19: 526–46. Converse, Philip 1964, “The Nature of Belief Systems in Mass Publics,” in Apter, D. E. (ed.), Ideology and Discontent, Free Press, pp. 206–61. Cook, Timothy 1998, Governing with the News: The News Media as a Political Institution, University of Chicago Press. Daalder, Hans 1974, “The Consociational Democracy Theme,” World Politics 26: 604–21. Dahl, Robert A. (ed.) 1966, Political Oppositions in Western Democracies, Yale University Press.

References

263

Dahl, Robert A. 1971, Polyarchy: Participation and Opposition, Yale University Press. Dalton, Russell J. and Eichenberg, Richard C. 1998, “Citizen Support for Policy Integration,” in Sandholtz, W. and Sweet, A. S. (eds.), European Integration and Supranational Governance, Oxford University Press, pp. 250–82. Dalton, Russell J., Flanagan, Scott C., and Beck, Paul Allen 1984, Electoral Change in Advanced Industrial Democracies: Realignment or Dealignment?, Princeton University Press. De Swaan, Abram 1973, Coalition Theory and Cabinet Formation, Elsevier. Deschouwer, Kris 2000, “The European Multi-Level Party Systems: Towards a Framework for Analysis,” working paper, European University Institute. Downs, Anthony 1957, An Economic Theory of Democracy, Harper and Row. Egeberg, Morten 2001, “How Federal? The Organizational Dimension of Integration in the EU (and Elsewhere),” Journal of European Public Policy 8: 728–46. Eichenberg, Richard C. 1998, “Measures, Methods, and Models in the Study of Public Opinion and European Integration, 1973–1997,” presented at American Political Science Association Meetings, Boston. Eichenberg, Richard C. and Dalton, Russell J. 1998, “Post-Maastricht Blues: Do Political Economy Models Still Explain Citizen Support for European Integration?” working paper, Department of Political Science, Tufts University. Eichener, Volker and Voelzkow, Helmut 1994, “Ko-Evolution politischadministrativer und verbandlicher Strukturen: Am Beispiel der technischen Harmonisierung des europäischen Arbeits-, Verbraucher- und Umweltschutzes,” Politische Vierteljahresschrift 25: 256–91. Eijk, Cees van der 2001, “Measuring Agreement in Ordered Rating Scales. Quality and Quantity,” Quality and Quantity 35: 325–41. Eijk, Cees van der and Franklin, Mark N. 1996, Choosing Europe?: The European Electorate and National Politics in the Face of Union, University of Michigan Press. Eijk, Cees van der and Niemoller, ¨ Broer 1983, Electoral Change in the Netherlands, CT Press. Eijk, Cees van der, Franklin, Mark N., and Brug, Wouter van de 1999, “Policy Preferences and Party Choice,” in Schmitt, H. and Thomassen, J. J. A. (eds.), Political Representation and Legitimacy in the European Union, Oxford University Press, pp. 161–85. Elff, Martin 2002, “An Integrated Perspective on Party Platforms and Electoral Choice,” presented at American Political Science Association Meetings, Boston. Enelow, James M. and Hinich, Melvin J. 1984, The Spatial Theory of Voting: An Introduction, Cambridge University Press. Esping-Andersen, Gøsta 1990, The Three Worlds of Welfare Capitalism, Princeton University Press. Esping-Andersen, Gøsta 1999, Social Foundations of Postindustrial Economies, Oxford University Press. Estevez-Abe, Margarita, Iversen, Torben, and Soskice, David 2001, “Social Protection and the Formation of Skills: A Reinterpretation of the Welfare State,”

264

References

in Hall, P. A. and Soskice, D. (eds.), Varieties of Capitalism. The Institutional Foundations of Comparative Advantage, Oxford University Press, pp. 145–83. Eurostat 1997, Eurostat Yearbook ’97, Office for Official Publications of the European Communities. Evans, Geoffrey 1998, “Euroscepticism and Conservative Electoral Support: How an Asset Became a Liability,” British Journal of Political Science 28: 573–90. Evans, Geoffrey 1999, “Europe: A New Electoral Cleavage?” in Evans, G. and Norris, P. (eds.), Critical Elections: British Parties and Voters in Long-Term Perspective, Sage, pp. 207–22. Falkner, Gerda and Nentwich, Michael 1997, “The Treaty of Amsterdam: Towards a New Institutional Balance,” European Integration Online Papers 1 (15), available at http://eiop.or.at/eiop/texte/1997-015a.htm. Feldman, Gerald D. and Homburg, Heidrun 1977, Industrie und Inflation: Studien und Dokumente zur Politik der deutschen Unternehmer 1916–1923, Hoffmann und Campe. Flood, Christopher G. 1997, “Euroscepticism in the Politics of the British and the French Right: A Comparison,” presented at American Political Science Association Meetings, Washington, DC. Fraenkel, Ernst 1968, Deutschland und die westlichen Demokratien, Kohlhammer. Franklin, Mark N. 1992, “The Decline of Cleavage Politics,” in Franklin, M. N., Mackie, T. T., and Valen, H. (eds.), Electoral Change: Responses to Evolving Social and Attitudinal Structures in Western Countries, Cambridge University Press, pp. 383–405. Franklin, Mark N. and Wlezien, Christopher 1997, “The Responsive Public: Issue Salience, Policy Change, and Preferences for European Unification,” Journal of Theoretical Politics 9: 347–63. Franklin, Mark N., Mackie, Thomas T., and Valen, Henry 1992, Electoral Change: Responses to Evolving Social and Attitudinal Structures in Western Countries, Cambridge University Press. Franzese, Robert J., Jr. 2001, “Institutional and Sectoral Interactions in Monetary Policy and Wage/Price-Bargaining,” in Hall, P. A. and Soskice, D. (eds.), Varieties of Capitalism. The Institutional Foundations of Comparative Advantage, Oxford University Press, pp. 104–44. Franzese, Robert J., Jr. 2002, Explaining Macroeconomic Policies. Institutional Interactions in the Evolving Political Economies of Developed Democracies, Cambridge University Press. Gabel, Matthew J. 1998a, “The Endurance of Supranational Governance: A Consociational Interpretation of the European Union,” Comparative Politics 30: 463–77. Gabel, Matthew J. 1998b, Interests and Integration: Market Liberalization, Public Opinion, and European Union, University of Michigan Press. Gabel, Matthew J. and Huber, John D. 2000, “Putting Parties in their Place: Inferring Party Left-Right Ideological Positions from Party Manifestos Data,” American Journal of Political Science 44: 94–103. Gamble, Andrew and Kelly, Gavin 2001, “The British Labour Party and Monetary Union,” in Notermans, T. (ed.), Social Democracy and Monetary Union, Berghahn Books, pp. 47–70.

References

265

Garnett, Mark and Lynch, Philip 2002, “Bandwagon Blues: The Tory Fightback Fails,” Political Quarterly 73: 29–37. Gelleny, Ronald D. and Anderson, Christopher J. 2000, “The Economy, Accountability, and Support for the President of the European Commission,” European Union Politics 1: 173–200. George, Stephen 1998, “The Intellectual Debate in Britain on the European Union,” working paper, Notre Europe. Glagow, Manfred and Schimank, Uwe 1983, “Gesellschaftssteuerung durch korporatistische Verhandlungssysteme. Zur begrifflichen Klärung,” in Falter, J. W., Fenner, C., and Greven, M. T. (eds.), Politische Willensbildung und Interessenvermittlung, Westdeutscher Verlag, pp. 539–46. Goldstein, Harvey and Healy, Michael J. R. 1995, “The Graphical Presentation of a Collection of Means,” Journal of the Royal Statistical Society, A 158: 175–7. Green, David 2001, “Support, (Mostly) Yes – But For What? Multilevel Governance, Policy Competencies and European Public Opinion,” presented at European Community Studies Association Meetings, Madison, WI. Greenwood, Justin, Grote, Jurgen ¨ R., and Ronit, Karsten (eds.) 1992, Organized Interests and the European Community, Sage. Haas, Ernst B. 1958, The Uniting of Europe: Political, Social, and Economic Forces, 1950–1957, Stanford University Press. Haas, Ernst B. 1964, Beyond the Nation-State: Functionalism and International Organization, Stanford University Press. Hainsworth, Paul (ed.) 1992, The Extreme Right in Europe and the USA, St. Martin’s Press. Hall, Peter A. and Franzese, Robert J., Jr. 1998, “Mixed Signals: Central Bank Independence, Coordinated Wage Bargaining, and European Monetary Union,” International Organization 52: 502–36. Hall, Peter A. and Gingerich, Daniel W. 2001, “Varieties of Capitalism and Institutional Complementarities in the Macroeconomy: An Empirical Analysis,” presented at American Political Science Association Meetings, San Francisco. Hall, Peter A. and Soskice, David (eds.) 2001, Varieties of Capitalism: The Institutional Foundations of Comparative Advantage, Oxford University Press. Hermet, Guy, Hottinger, Julian Thomas, and Seiler, Daniel-Louis (eds.) 1998, Les partis politiques en Europe de l’Ouest, Economica. Hicks, Alexander 1999, Social Democracy and Welfare Capitalism. A Century of Income Security Policies, Ithaca, NY: Cornell University Press. Hildenbrandt, Kai and Dalton, Russell J. 1976, “Die neue Politik – Politischer Wandel oder Schonwetterpolitik?” ¨ Politische Vierteljahresschrift 7: 230– 56. Hinich, Melvin J. and Munger, Michael C. 1997, Analytical Politics, Cambridge University Press. Hinich, Melvin J. and Ordeshook, Peter C. 1970, “Plurality Maximization vs. Vote Maximization: A Spatial Analysis with Variable Participation,” American Political Science Review 64: 772–91. Hix, Simon 1995a, “Parties at the European Level and the Legitimacy of EU Socio-Economic Policy,” Journal of Common Market Studies 33: 527–53.

266

References

Hix, Simon 1995b, “Political Parties in the European Union System: A ‘Comparative Politics Approach’ to the Development of Party Federations,” Ph.D. thesis, European University Institute. Hix, Simon 1996, “The Transnational Party Federations,” in Gaffney, J. (ed.), Political Parties and the European Union, Routledge, pp. 308–31. Hix, Simon 1999a, “Dimensions and Alignments in European Union Politics: Cognitive Constraints and Partisan Responses,” European Journal of Political Research 35: 69–125. Hix, Simon 1999b, The Political System of the European Union, St. Martin’s Press. Hix, Simon and Lord, Christopher 1997, Political Parties in the European Union, St. Martin’s Press. Hix, Simon and Lord, Christopher 1998, “A Model Transnational Party? The Party of European Socialists,” in Bell, D. S. and Lord, C. (eds.), Transnational Parties in the European Union, Ashgate, pp. 86–101. Hix, Simon, Noury, Abdul, and Roland, Gérard 2003, “Power to the Parties: Cohesion and Competition in the European Parliament, 1979–2001,” Unpublished manuscript. Hoffmann, Stanley 1966, “Obstinate or Obsolete? The Fate of the Nation-State and the Case of Western Europe,” Daedalus 95: 862–915. Hooghe, Liesbet and Marks, Gary 1996, “Birth of a Polity: The Struggle over European Integration,” presented at Tenth International Conference of Europeanists, Chicago. Hooghe, Liesbet and Marks, Gary 1999, “The Making of a Polity: The Struggle Over European Integration,” in Kitschelt, H., Lange, P., Marks, G., and Stephens, J. (eds.), Continuity and Change in Contemporary Capitalism, Cambridge University Press, pp. 70–97. Hooghe, Liesbet and Marks, Gary 2001, Multi-Level Governance and European Integration, Rowman and Littlefield. Hoyle, Rick H. and Panter, Abigail T. 1995, “Writing about Structural Equation Models,” in Hoyle, R. H. (ed.), Structural Equation Modeling: Concepts, Issues, and Applications, Sage, pp. 158–76. Huber, Evelyne and Stephens, John D. 2001, Development and Crisis of the Welfare State. Parties and Policies in Global Markets, University of Chicago Press. Imig, Doug and Tarrow, Sidney 1999, “The Europeanization of Movements? A New Approach to Transnational Contention,” in Porta, D. d., Kriesi, H., and Rucht, D. (eds.), Social Movements in a Globalizing World, St. Martin’s Press, pp. 112–33. Imig, Doug and Tarrow, Sidney 2000, “Political Contention in Europeanizing Societies,” West European Politics 23: 73–93. Imig, Doug and Tarrow, Sidney (eds.) 2001, Contentious Europeans: Protest and Politics in the New Europe, Rowman and Littlefield. Inglehart, Ronald 1978, The Silent Revolution: Changing Values and Political Styles among Western Publics, Princeton University Press. Inglehart, Ronald 1990, Culture Shift in Advanced Industrial Society, Princeton University Press. Inglehart, Ronald 1997, Modernization and Postmodernization: Cultural, Economic, and Political Change in 43 Societies, Princeton University Press.

References

267

Iversen, Torben 1998, “Wage Bargaining, Central Bank Independence and the Real Effects of Money,” International Organization 52: 469–504. Iversen, Torben 1999, Contested Economic Institutions: The Politics of Macroeconomics and Wage Bargaining in Advanced Democracies, Cambridge University Press. Iversen, Torben and Suskice, David 2001, “An Asset Theory of Social Policy Preferences,” American Political Science Review 95.4: 875–94. Iversen, Torben and Wren, Ann 1998, “Equality, Employment, and Budgetary Restraint: The Trilemma of the Service Economy,” World Politics 50: 507–46. Jachtenfuchs, Markus and Kohler-Koch, Beate (eds.) 1996, Europäische Integration, Leske und Budrich. Jennings, M. Kent 1992, “Ideological Thinking Among Mass Publics and Political Elites,” Public Opinion Quarterly 56: 419–41. Jones, Bryan D. 1994, Reconceiving Decision-Making in Democratic Politics: Attention, Choice, and Public Policy, University of Chicago Press. Joreskog, ¨ Karl and Sorbom, ¨ Dag 1981, LISREL V: Analysis of Linear Structural Relationships by the Method of Maximum Likelihood, National Educational Resources. Karapin, Roger 1998, “Explaining Far-Right Electoral Success in Germany,” German Politics and Society 16: 24–61. Karvonen, Lauri and Kuhnle, Stein (eds.) 2001, Party Systems and Voter Alignments Revisited, Routledge. Katz, Richard S. and Wessels, Bernhard (eds.) 1999, The European Parliament, the National Parliaments, and European Integration, Oxford University Press. Kelly, Richard 2002, “The Party Didn’t Work: Conservative Reorganisation and Electoral Failure,” Political Quarterly 73: 38–43. Kinder, Donald and Sears, David O. 1985, “Public Opinion and Political Action,” in Lindzey, G. and Aronson, E. (eds.), The Handbook of Social Psychology, vol. II, Random House, pp. 659–741. Kirchner, Emil J. 1978, Trade Unions as a Pressure Group in the European Community, Saxonhouse. Kirchner, Emil J. and Schwaiger, Konrad 1981, The Role of Interest Groups in the European Community, Gower. Kitschelt, Herbert 1988, “Left-Libertarian Parties,” World Politics 40: 194–234. Kitschelt, Herbert 1994, The Transformation of European Social Democracy, Cambridge University Press. Kitschelt, Herbert 1999, “European Social Democracy between Political Economy and Electoral Competition,” in Kitschelt, H., Lange, P., Marks, G., and Stephens, J. D. (eds.), Continuity and Change in Contemporary Capitalism, Cambridge University Press, pp. 317–45. Kitschelt, Herbert and McGann, Anthony J. 1995, The Radical Right in Western Europe: A Comparative Analysis, University of Michigan Press. Kitschelt, Herbert, Lange, Peter, Marks, Gary, and Stephens, John 1999, “Conclusion: Convergence and Divergence in Advanced Capitalist Democracies,” in Kitschelt, H., Lange, P., Marks, G., and Stephens, J. (eds.), Continuity and Change in Contemporary Capitalism, Cambridge University Press, pp. 427– 60.

268

References

Klandermans, Bert, Weerd, Marga de, Sabucedo, Jose Manuel, and Rodriquez, Mauro 2001, “Framing Contention: Dutch and Spanish Farmers Confront the EU,” in Imig, D. and Tarrow, S. (eds.), Contentious Europeans: Protest and Politics in an Emerging Polity, Rowman and Littlefield, pp. 77–95. Kleinnijenhuis, Jan and Rietberg, Edward M. 1995, “Parties, Media, the Public and the Economy: Patterns of Societal Agenda-Setting,” European Journal of Political Research 28: 95–118. Klingemann, Hans-Dieter, Hoffebert, Richard I., and Budge, Ian 1994, Parties, Policies, and Democracy, Westview Press. Knutsen, Oddbjorn 1995, “Value Orientations, Political Conflicts and Left-Right Identification: A Comparative Study,” European Journal of Political Research 28: 63–93. Kohler-Koch, Beate 1994, “Changing Patterns of Interest Intermediation in the European Union,” Government and Opposition 29: 167–80. Kollman, Ken, Miller, John H., and Page, Scott E. 1992, “Adaptive Parties in Spatial Elections,” American Political Science Review 86: 929–37. Konig, ¨ Thomas 1997, Europa auf dem Weg zum Mehrheitssystem, Westdeutscher Verlag. Kreppel, Amie and Tsebelis, George 1999, “Coalition Formation in the European Parliament,” Comparative Political Studies 32: 933–66. Kriesi, Hanspeter 1989, “The Mobilization Potential of the New Social Movements in the Netherlands in 1986 and 1987: Its Structural and Cultural Basis,” in Ganzeboom, H. B. and Flap, H. (eds.), New Social Movements and Value Change, SISWO, pp. 51–88. Kriesi, Hanspeter, Koopmans, Ruud, Duyvendak, Jan W., and Giugni, Marco G. 1995, New Social Movements in Western Europe, University of Minnesota Press. Ladrech, Robert 1997, “Social Democratic Parties and EC Integration: Transnational Party Responses to Europe 1992,” European Journal of Political Research 24: 195–210. Ladrech, Robert and Marlière, Philippe (eds.) 1999, Social Democratic Parties in the European Union. History, Organization, Policies, Macmillan. Laffan, Brigid 1996, “Politics of Identity and Political Order in Europe,” Journal of Common Market Studies 34: 81–102. Laver, Michael (ed.) 2001, Estimating the Policy Position of Political Actors, Routledge. Laver, Michael and Garry, John 2000, “Estimating Policy Positions From Political Texts,” American Journal of Political Science 44: 619–34. Laver, Michael and Schofield, Norman 1990, Multiparty Government: The Politics of Coalition in Europe, Oxford University Press. Lee, Eun Sul, Forthofer, Ron N., and Lorimor, Ronald J. 1989, Analyzing Complex Survey Data, Sage. Lijphart, Arend 1977, Democracy in Plural Societies: A Comparative Exploration, Yale University Press. Lijphart, Arend 1984, Democracies: Patterns of Majoritarian and Consensus Government in Twenty-One Countries, Yale University Press. Lindberg, Leon and Scheingold, Stuart 1970, Europe’s Would-Be Polity: Patterns of Change in the European Community, Prentice Hall.

References

269

Lipset, Seymour Martin 1960, Political Man: The Social Bases of Politics, Doubleday. Lipset, Seymour Martin and Rokkan, Stein 1967, “Cleavage Structures, Party Systems, and Voter Alignments: An Introduction,” in Lipset, S. M., and Rokkan, S. (eds.), Party Systems and Voter Alignments, Free Press, pp. 1–64. Ljunggren, S. B. 1988, “Conservatism in Norway and Sweden,” in Girvin, B. (ed.), The Transformation of Contemporary Conservatism, Sage, pp. 120–44. Long, J. Scott 1983, Confirmatory Factor Analysis, Sage. Lord, Christopher 1998, “Party Groups, EP Committees and Consensus Democracy,” in Bell, D. S. and Lord, C. (eds.), Transnational Parties in the European Union, Ashgate, pp. 204–17. Lowery, David and Gray, Virginia 1993, “The Density of State Interest Group Systems,” Journal of Politics 55: 191–206. Luther, Kurt Richard 2001, “From Populist Protest to Incumbency: The Strate¨ gic Challenges Facing Jorg ¨ Haider’s Freedom Party of Austria (FPO),” working paper, Keele European Parties Research Unit (KEPRU), Keele University. MacDonald, Stuart Elaine, Listhaug, Ola, and Rabinowitz, George 1991, “Issues and Party Support in Multiparty Systems,” American Political Science Review 85: 1107–31. Mair, Peter 2000, “The Limited Impact of Europe on National Party Systems,” West European Politics 23: 27–51. Mares, Isabela 2001, “Firms and the Welfare State: When, Why, and How Does Social Policy Matter to Employers?” in Hall, P. A. and Soskice, D. (eds.), Varieties of Capitalism. The Institutional Foundations of Comparative Advantage, Oxford University Press, pp. 184–212. Marks, Gary and McAdam, Doug 1996, “Social Movements and the Changing Structure of Political Opportunity in the European Union,” West European Politics 20: 111–33. Marks, Gary and McAdam, Doug 1999, “On the Relationship of Political Opportunities to the Form of Collective Action: The Case of the European Union,” in Porta, D. d., Kriesi, H., and Rucht, D. (eds.), Social Movements in a Globalizing World, St. Martin’s Press, pp. 97–111. Marks, Gary, and Steenbergen, Marco R. 1999, “1999 Expert Survey on National Parties and the European Union,” available at http://www. unc.edu/∼gwmarks. Marks, Gary and Wilson, Carole J. 2000, “The Past in the Present: A Cleavage Theory of Party Response to European Integration,” British Journal of Political Science 30: 433–59. Marks, Gary, Hooghe, Liesbet, and Blank, Kermit 1996, “European Integration since the 1980s: State-Centric versus Multi-Level Governance,” Journal of Common Market Studies 34: 341–78. Marks, Gary, Wilson, Carole J., and Ray, Leonard 2002, “National Political Parties and European Integration,” American Journal of Political Science 46: 585–94. Marsh, Michael and Franklin, Mark N. 1996, “The Foundations: Unanswered Questions from the Study of European Elections, 1979–1994,” in Eijk, C. v. d. and Franklin, M. N. (eds.), Choosing Europe? The European Electorate

270

References

and National Politics in the Face of Unification, University of Michigan Press, pp. 11–32. Martin, Andrew and Ross, George 2001, “Trade Union Organizing at the European Level: The Dilemma of Borrowed Resources,” in Imig, D. and Tarrow, S. (eds.), Contentious Europeans: Protest and Politics in an Emerging Polity, Rowman and Littlefield, pp. 53–76. Mazey, Sonia and Richardson, Jeremy 1997, “Policy Framing: Interest Groups and the 1996 Inter-Governmental Conference,” West European Politics 20: 111–33. McAdam, Doug and Sewell, William, Jr. 2001, “Temporality in the Study of Social Movements and Revolution,” in Aminzade, R., Goldstone, J., McAdam, D., and Perry, E. J. (eds.), Silence and Voice in the Study of Contentious Politics, Cambridge University Press, pp. 89–125. Messina, Anthony M. 2001, “Globalization and the Surge of Anti-Immigration Groups in Western Europe,” presented at European Community Studies Association Meetings, Madison. Meyer, David and Tarrow, Sidney (eds.) 1998, The Movement Society: Contentious Politics for a New Century, Rowman and Littlefield. Middendorp, Cees P. 1992, Ideology in Dutch Politics: The Democratic System Reconsidered, 1970–1985, Van Gorcum. Moravcsik, Andrew 1998, The Choice for Europe. Social Purpose and State Power from Messina to Maastricht, Cornell University Press. Muller-Rommel, ¨ Ferdinand (ed.) 1989, New Politics in Western Europe: The Rise and Success of Green Parties and Alternative Lists, Westview Press. Murrell, Peter 1984, “An Examination of the Factors Affecting the Formation of Interest Groups in OECD Countries,” Public Choice 43: 151–71. Niedermayer, Oskar and Sinnott, Richard (eds.) 1995, Public Opinion and Internationalized Governance, Oxford University Press. Noury, Abdul G. 2002, “Ideology, Nationality and Euro-Parliamentarians,” European Union Politics 3: 33–58. Offe, Claus 1969, “Politische Herrschaft und Klassenstrukturen: Zur Analyse spätkapitalistischer Gesellschaftssysteme,” in Kress, G. and Senghaas, D. (eds.), Politikwissenschaft, Fischer, pp. 155–89. Offe, Claus 1981, “The Attribution of Public Status to Interest Groups: Observations on the West German Case,” in Berger, S. (ed.), Organizing Interests in Western Europe, Cambridge University Press, pp. 123–58. Olson, Mancur 1971, The Logic of Collective Action: Public Goods and the Theory of Groups, Harvard University Press. Olson, Mancur 1982, The Rise and Decline of Nations: Economic Growth, Stagflation, and Social Rigidities, Yale University Press. Peterson, John and Bomberg, Elizabeth E. 1999, Decision-Making in the European Union, Macmillan. Petrocik, John R. 1996, “Issue Ownership in Presidential Elections, with a 1980 Case Study,” American Journal of Political Science 40: 825–50. Pollack, Mark A. 2000, “Blairism in Brussels: The ‘Third Way’ in Europe since Amsterdam,” in Cowles, M. G. and Smith, M. (eds.), The State of the European Union, vol. V: Risks, Reform, Resistance, and Revival, Oxford University Press, pp. 266–91.

References

271

Poole, Keith T. and Rosenthal, Howard 1997, Congress: A Political-Economic History of Roll-Call Voting, Oxford University Press. Powell, G. Bingham 2000, Elections as Instruments of Democracy: Majoritarian and Proportional Visions, Yale University Press. Raunio, Tapio 1996, “Party Group Behavior in the European Parliament,” working paper, University of Tampere. Raunio, Tapio 1998, “Cleavages and Alignments in the European Parliament: MEP Voting Behavior, 1989–1994,” in Bell, D. S. and Lord, C. (eds.), Transnational Parties in the European Union, Ashgate, pp. 168–88. Ray, Leonard 1997, “Politicizing Europe: Political Parties and the Changing Nature of Public Opinion about the EU,” Ph.D. thesis, University of North Carolina. Ray, Leonard 1999, “Measuring Party Orientations Towards European Integration. Results from an Expert Survey,” European Journal of Political Research 36: 283–306. Reif, Karlheinz 1984, European Elections 1979/81 and 1984: Conclusions and Perspectives from Empirical Research, Quorum. Reif, Karlheinz and Schmitt, Hermann 1980, “Nine Second Order National Elections: A Conceptual Framework for the Analysis of European Election Results,” European Journal of Political Research 8: 3–44. Riker, William H. 1962, Liberalism against Populism: A Confrontation between the Theory of Democracy and the Theory of Social Choice, W. H. Freeman. Riker, William H. 1986, The Art of Political Manipulation, Yale University Press. Risse, Thomas 1996, “Exploring the Nature of the Beast: International Relations Theory and Comparative Policy Analysis Meet the European Union,” Journal of Common Market Studies 34: 53–80. Robertson, David 1976, A Theory of Party Competition, John Wiley and Sons. Rokkan, Stein 1966, “Norway: Numerical Democracy and Corporate Pluralism,” in Dahl, R. A. (ed.), Political Oppositions in Western Democracies, Yale University Press, pp. 70–115. Rometsch, Dietrich and Wessels, Wolfgang (eds.) 1996, The European Union and Member States: Towards Institutional Fusion?, Manchester University Press. Rudig, ¨ Wolfgang 1996, “Green Parties and the European Union,” in Gaffney, J. (ed.), Political Parties and the European Union, Routledge, pp. 254–74. Sbragia, Alberta (ed.) 1992, Euro-Politics: Institutions and Policy-Making in the “New” European Community, Brookings Institution. Sbragia, Alberta 1993, “The European Community: A Balancing Act,” Publius 23: 23–38. Scharpf, Fritz W. 1991, Crisis and Choice in European Social Democracy, Cornell University Press. Scharpf, Fritz W. 1996, “Negative and Positive Integration in the Political Economy of European Welfare States,” in Marks, G., Scharpf, F. W., Schmitter, P., and Streeck, W. (eds.), Governance in the European Union, Sage, pp. 15–39. Scharpf, Fritz W. 1999, Governing in Europe: Effective and Democratic?, Oxford University Press. Scharpf, Fritz W. 2002, “The European Social Model: Coping With the Challenges of Diversity,” working paper, Max Planck Institute for the Study of Societies.

272

References

Schattschneider, Elmer E. 1960, The Semisovereign People: A Realist’s View of Democracy in America, Holt, Rinehart, and Winston. Schendelen, Rinus van 1991, “Images of Democratic Representation in the European Community,” in Klingemann, H.-D., Stoss, ¨ R., and Wessels, B. (eds.), Politische Klasse und politische Institutionen, Westdeutscher Verlag, pp. 357–71. Scheve, Kenneth 2000, “Comparative Context and Public Preferences over Regional Economic Integration,” presented at American Political Science Association Meetings, Washington, DC. Schmitt, Hermann and Thomassen, Jacques J. A. (eds.) 1999, Political Representation and Legitimacy in the European Union, Oxford University Press. Schmitter, Philippe 1996, “Examining the Present Euro-Polity with the Help of Past Theories,” in Marks, G., Scharpf, F. W., Schmitter, P., and Streeck, W. (eds.), Governance in the European Union, Sage, pp. 1–14. Scott, David J. 2001, “The Salience of Integration: The Strategic Behavior of National Political Parties in the European Union,” Ph.D. thesis, University of North Carolina. Sened, Itai 1996, “A Model of Coalition Formation: Theory and Evidence,” Journal of Politics 58: 350–72. Shepsle, Kenneth A. 1992, Models of Multiparty Electoral Competition, Harwood. Shields, James 1995, “Le Pen and the Progression of the Far-Right Vote in France,” French Politics and Society 13: 21–39. Sidjanski, Dusan 1972, “Pressure Groups and the European Economic Community,” in Hodges, M. (eds.), European Integration, Penguin Books, pp. 401– 20. Sinnott, Richard 1997, “European Public Opinion and the European Union: The Knowledge Gap,” working paper 126, Barcelona. Sitter, Nick 2001, “The Politics of Opposition and European Integration in Scandinavia: Is Euro-Skepticism a Government-Opposition Dynamic?” West European Politics 24: 22–39. Smith, Jackie G. 2001, “Globalizing Resistance: The Battle of Seattle and the Future of Social Movements,” Mobilization 6: 1–20. Soskice, David 1999, “Divergent Production Regimes: Coordinated and Uncoordinated Market Economies in the 1980s and 1990s,” in Kitschelt, H., Lange, P., Marks, G., and Stephens, J. D. (eds.), Continuity and Change in Contemporary Capitalism, Cambridge University Press, pp. 101–34. Soskice, David and Iversen, Torben 1998, “Multiple Wage-Bargaining Systems in the Single European Currency Area,” Oxford Review of Economic Policy 14: 110–24. Soskice, David and Iversen, Torben 2000, “The Non-Neutrality of Monetary Policy with Large Price or Wage Setters,” Quarterly Journal of Economics 115: 265–84. Sowemimo, Matthew 1996, “The Conservative Party and European Integration: 1988–95,” Party Politics 2: 77–97. Steenbergen, Marco R. and Lodge, Milton Forthcoming, “Process Matters: Cognitive Models of Candidate Evaluation,” in MacKuen, M. and Rabinowitz, G. (eds.), Electoral Democracy, University of Michigan Press. Stinchcombe, Arthur L. 1975, “Social Structure and Politics,” in Greenstein, F. I. and Polsby, N. W. (eds.), Handbook of Political Science, vol. III: Macropolitical Theory, Addison-Wesley, pp. 557–622.

References

273

Stone-Sweet, Alec and Sandholtz, Wayne 1997, “European Integration and Supranational Governance,” Journal of European Public Policy 4: 297–317. Streeck, Wolfgang 1996, “Neo-Voluntarism: A New European Social Policy Regime?” in Marks, G., Scharpf, F. W., Schmitter, P., and Streeck, W. (eds.), Governance in the European Union, Sage, pp. 64–94. Strøm, Kaare and Muller, ¨ Wolfgang C. 1999, “Political Parties and Hard Choices,” in Strøm, K. and Muller, ¨ W. C. (eds.), Policy, Office, or Votes: How Political Parties in Western Europe Make Hard Decisions, Cambridge University Press, pp. 1–35. Sv˚asand, Lars and Lindstrom, ¨ Ulf 1996, “Scandinavian Political Parties and the European Union,” in Gaffney, J. (ed.), Political Parties and the European Union, Routledge, pp. 205–19. Thomassen, Jacques J. A. 1998, “Political Representation and Legitimacy in the European Union,” presented at American Political Science Association Meetings, Boston. Thomassen, Jacques J. A. and Schmitt, Hermann 1997, “Policy Representation,” European Journal of Political Research 32: 165–84. Thomassen, Jacques J. A. and Schmitt, Hermann 1999a, “In Conclusion: Political Representation and Legitimacy in the European Union,” in Schmitt, H. and Thomassen, J. J. A. (eds.), Political Representation and Legitimacy in the European Union, Oxford University Press, pp. 257–69. Thomassen, Jacques J. A. and Schmitt, Hermann 1999b, “Partisan Structures in the European Parliament,” in Katz, R. S. and Wessels, B. (eds.), The European Parliament, the National Parliaments and European Integration, Oxford University Press, pp. 129–48. Truman, David B. 1951, The Governmental Process: Political Interests and Public Opinion, Knopf. Tsebelis, George and Garrett, Geoffrey 1996, “Agenda Setting Power, Power Indices, and Decision Making in the European Union,” International Review of Law and Economics 16: 345–61. Tsebelis, George and Garrett, Geoffrey 2000, “Legislative Politics in the European Union,” European Union Politics 1: 9–36. Turner, Lowell 1996, “The Europeanization of Labour: Structure Before Action,” European Journal of Industrial Relations 2: 325–44. Visser, Jelle and Ebbinghaus, Bernhard 1992, “Making the Most of Diversity? European Integration and Transnational Organization of Labour,” in Greenwood, J., Grote, J. R., and Ronit, K. (eds.), Organized Interests and the European Community, Sage, pp. 206–37. Voeten, Erik 2000, “Clashes in the Assembly,” International Organization 54: 185–216. Wessels, Bernhard 1991, “Vielfalt oder strukturierte Komplexität? Zur Institutionalisierung politischer Spannungslinien im Verbände- und Parteiensystem in der Bundesrepublik,” Kölner Zeitschrift für Soziologie und Sozialpsychologie 43: 454–75. Wessels, Bernhard 1996, “Systems of Economic Interest Groups and SocioEconomic Performance,” presented at American Political Science Association Meetings, San Francisco. Wessels, Bernhard 1997, “Organizing Capacity of Societies and Modernity,” in Deth, J. W. v. (ed.), Groups and Public Life: Social Participation, Voluntary

274

References

Associations, and Political Involvement in Representative Democracies, Routledge, pp. 198–219. Wessels, Bernhard 1999, “European Parliament and Interest Groups,” in Katz, R. S. and Wessels, B. (eds.), The European Parliament, the National Parliaments, and European Integration, Oxford University Press, pp. 105–28. Whiteley, Paul 1998, “The Conservative Campaign,” Parliamentary Affairs 50: 542–54. Whiteley, Paul, Seyd, Patrick, Richardson, Jeremy, and Bissell, Paul 1994, “Thatcherism and the Conservative Party,” Political Studies 42: 185–202. Wilson, Graham K. 1990, Interest Groups, Basil Blackwell.

Index

Achen, Christopher 180 Aldrich, John 169 Allardt, Erik 4 Anderson, Christopher 13, 15, 51, 110, 237, 241, 243, 251 Andeweg, Rudy 164 Ansell, Christopher K. 1 Armingeon, Klaus 199 Aspinwall, Mark 238, 255 Attina, Fulvio 146 Austen-Smith, David 168 Austria 34, 35, 38, 42, 43, 61, 68, 75, 78, 84, 152, 160, 178, 205, 211 authoritarian (parties) 72, 86, 122, 210 Axelrod, Robert 169 Aylott, Nicholas 79 Bache, Ian 1 Baker, David 136 Banchoff, Thomas F. 2 Banks, Jeffrey 168 Bardi, Luciano 146–7, 162 Barnes, Samuel 4 Barnouin, Barbara 199 Beck, Paul Allen 4 Belgium 7, 34, 35, 38, 42, 61, 68, 70, 75, 84, 128, 205, 217, 244 Bell, David 2 Benedetto, Giacomo 133, 135 Benelux countries 159, 160 Bentler, Peter 23, 105 Bentley, Arthur 196 Betz, Hans-Georg 134 Beyers, Jan 128, 129 Blair, Tony 9, 218 Bogdanor, Vernon 164 Boix, Carles 68 Bollen, Kenneth 23 Bomberg, Elizabeth 1, 63, 138 Bonnett, David 23, 105 Borzel, ¨ Tanja 131

Brinegar, Adam 10, 62, 241, 246, 247, 253, 257 Britain see United Kingdom Brzinski, Joanne Bay 146 Budge, Ian 94, 97, 98, 107, 166, 167, 170, 172, 190 Bush, Evelyn 220 Bush, George H. W. 167 Caldeira, Lester 15 Canada 7 Caporaso, James 1, 2, 120, 131 Castillo, Pilar del 35 Castles, Frank 174 Cayrol, Roland 35 centralized wage bargaining 64–5, 69, 70, 77, 78, 79, 80, 82, 85 Choosing Europe? 35, 47, 48 Christensen, Dag Arne 128, 138 Christian democracy 57, 86 Christian democratic parties 96, 110 Christian democratic (conservative) welfare state 67, 68, 69, 77, 78, 80, 82, 85, 86, 160, 253 cleavages, social 2–5, 141–2, 144, 159, 163, 164, 165, 196, 197, 202, 210–11, 236, 258 Clinton, Bill 167 Cohn-Bendit, Daniel 138 Committee of Regions 15 Common Agricultural Policy (CAP) 100, 226 Communist Party (Kommounistiko Komma Ellados, Greece) 128 Communist Party (Parti Communiste Fran¸cais, France) 217 conservative parties 96, 110 Conservative Party (United Kingdom) 48–9, 87, 135, 136, 167, 188, 218 consociational democracy 142–3 convergence 53, 55 Converse, Philip 16

275

276

Index

Cook, Robin 95 Cook, Timothy 169 Council of Ministers 8, 13, 94, 95, 133, 213, 246, 251, 252 currency union 240 Daalder, Hans 142 Dahl, Robert 141, 142, 195, 197, 198 Dalton, Russell 4, 15, 29, 210, 218, 246, 258 Darden, Keith A. 1 de Gaulle, Charles 135, 136 Delors, Jacques 241, 252 democratic deficit 14, 15, 72, 141, 163, 253 Democratic Party (United States) 167 demos 242 Denmark 34, 35, 38, 42, 43, 48–9, 61, 68, 70, 75, 78, 84, 130, 152, 157, 160, 256 Deschouwer, Kris 241 De Swaan, Abram 169 dissent, within political parties 169, 171, 175, 185–7 Downs, Anthony 148, 166, 167, 168 Ebbinghaus, Bernhard 204 Economic and Social Committee 15 Egeberg, Morten 252 Eichenberg, Richard 15, 29, 246, 258 Elff, Martin 241 Enelow, James 166, 167, 168 EPP-ED, Group of the European People’s Party and European Democrats 96 Esping-Andersen, Gøsta 57, 64, 67, 70, 160 Estevez-Abe, Margarita 64, 66, 70 EU Charter of Fundamental Rights 135 Eurobarometer surveys 173, 174, 241, 253 Euro-parties 93, 95–7, 110–12; see also European party federations European Atomic Energy Community (EURATOM) 201 European Central Bank (ECB) 62, 65 European Coal and Steel Community (ECSC) 201 European Commission 196, 199, 207, 213, 216, 218, 239, 241 European Council 94, 246, 251 European Court of Justice (ECJ) 15, 242 European Economic Community (EEC) 200, 201, 202 European elections 93, 95, 97, 141, 163, 195

European Federation of Green Parties (EFGP) 96–7, 107, 110 European Liberal, Democrat, and Reform Party (ELDR) 96, 97, 109–10, 146, 149, 151, 155, 156–7, 158, 160–1, 162–3 European Parliament (EP) 8, 14, 15, 32, 35, 42, 48, 51, 95, 100, 127, 133, 137, 138, 140, 141, 146–53, 162–4, 212, 215, 235, 239, 242, 243, 246, 251, 252, 253 European party federations 93, 96; see also Euro-parties European party groups 145–64 European People’s Party (EPP) 96, 97, 107–10, 146, 149, 150, 151, 155, 156–7, 158, 160, 161, 162–3, 240 European Trade Union Confederation (ETUC) 206 European Works Councils 219 Euroskepticism 125, 128, 130, 132, 133–9, 188, 238, 242, 244 Evans, Geoffrey 124, 238 Falkner, Gerda 207 Farlie, Dennis 166, 167, 170, 172, 190 Feldman, Gerald 199 Fianna Fail (Ireland) 96, 135 Finland 34, 35, 38, 42, 43, 61, 68, 70, 75, 78, 84, 152, 160, 178, 205, 211, 256 Flanagan, Scott C. 4 Flanders 34, 38, 42, 43 Flemish Bloc (Vlaams Blok, Belgium) 133, 134, 135, 244 Flood, Christopher 135, 136 Folke Bevaegelsen 42 Forthofer, Ron 179 Forza Italia (Italy) 135, 136 Fraenkel, Ernst 196 France 34, 35, 38, 42, 43, 61, 67, 68, 70, 75, 84, 86, 136, 146, 151, 157, 160, 205, 216–17, 239, 245 Franklin, Mark 4, 8, 10, 32, 35, 36, 47, 48, 131, 165, 171, 185, 189, 190, 195, 223, 237–8, 241, 246, 251 Franzese, Robert 62, 63 Freedom Party (Freiheitliche Partei ¨ Osterreichs, Austria) 133, 134 Friends of the Earth 219 G-7 228 Gabel, Matthew J. 4, 10, 13, 15, 30, 51, 63, 93, 94, 98, 110, 126, 158, 174, 235, 237, 240, 241, 243, 251

Index GAL (green/alternative/libertarian) 121, 125, 127, 131, 137, 140 Gamble, Andrew 79 Garnett, Mark 189 Garrett, Geoffrey 5, 7, 8, 17, 126; see also regulation model Garry, John 97 Gaullists (Rassemblement pour la République, France) 135 Gayssot, Jean-Claude 218 Gelleny, Ronald 15 George, Stephen 136 German Constitutional Court 138 Germany 34, 35, 38, 42, 43, 61, 67, 68, 70, 75, 84, 86, 130, 151, 157, 160–217 Gibson, Gregory 15 Gingerich, Daniel 64, 66, 68, 69, 70, 246 Glagow, Manfred 200 Goldstein, Harvey 179 Gray, Virginia 204 Greece 34, 35, 38, 42, 43, 44, 61, 68, 70, 75, 84, 136, 160, 205, 235, 256 Green, David 237 green parties 96, 138–9, 150, 211, 244 Green Party (Les Verts, France) 138, 217 Green Party (Génération Ecologiste, France) 138 Green Party (Die Grunen, ¨ Germany) 138, 217 Green/Free Alliance Group (in European Parliament) 96 Greenpeace 219 Greenwood, Justin 195 Grote, Jurgen ¨ 195 Growth and Stability Pact 128 Haas, Ernst 1, 5, 124–5 Hague, William 49 Haider, Jorg ¨ 134 Hainsworth, Paul 134 Hall, Peter 62, 64, 66, 68, 69, 70, 246, 252 Healy, Michael 175 Hearl, Derek 94, 97, 98 Hermet, Guy 128, 135, 138 Hicks, Alexander 64 Hildenbrandt, Kai 210 Hinich, Melvin 3, 166, 167, 168–75 Hix, Simon 4, 5, 6–7, 10, 22, 30, 31, 51, 63, 69, 93, 95, 96, 111, 124, 126, 129, 144, 154, 158, 162, 164, 174, 185, 195, 197, 202, 228, 238, 239, 240, 241, 246, 252 see Hix-Lord Model

277 Hix–Lord model 6–7, 17, 22, 25, 26, 94, 103, 105, 107, 109, 110–11, 126, 237, 244, 245, 258 Hoffebert, Richard 166, 167, 168, 172, 190 Hoffmann, Stanley 5 Holmberg, ¨ Soren ¨ 35 Homburg, Heidrun 199 Hooghe, Liesbet 1, 5, 8, 9, 10, 22, 51, 63, 120, 126, 128, 129, 130, 131, 144, 150, 154, 159, 174, 191, 197, 202, 214, 235, 238, 240, 242, 243, 244, 251 Hooghe–Marks model 9, 17, 22, 27, 94, 103, 105, 106, 107, 109, 110–11, 126, 241, 252, 258 Hoyle, Rick 23 Huber, Evelyne 64 Huber, John 13, 94, 98 Hungary 217 Imig, Doug 1, 10, 216, 221, 227–8, 238 Immerfall, Stefan 134 Inglehart, Ronald 131, 228 integration/independence dimension 144–5, 150, 151–2, 153–7, 158–60, 162–3, 243 interest groups 195–6 Intergovernmental Conference (IGC) 210; in Nice 135 intergovernmentalism 5 international relations model 5–6, 17, 21, 23, 94, 103, 105, 111, 126 Ireland 34, 35, 38, 42, 43, 61, 68, 69, 70, 75, 84, 201, 205, 217 Italy 34, 35, 38, 42, 43, 61, 68, 70, 75, 84, 86, 146, 151, 152, 157, 160, 205 Iversen, Torben 62, 63, 64, 65, 66, 69, 70 Jachtenfuchs, Markus 2, 131 Jennings, M. Kent 251 Jolly, Seth 10, 62, 241, 246, 247, 253, 257 Jones, Bryan 190 Joreskog, ¨ Karl 23, 105 Jospin, Lionel 217 June Movement (Juni Bevaegelsen, Denmark) 42 Kaase, Max 4 Karapin, Roger 134 Karvonen, Lauri 4 Katz, Richard 2, 211, 235 Keeler, John S. 2 Kelly, Gavin 79, 189 Kerremans, Bart 128, 129 Kinder, Donald 16, 251

278

Index

Kirchner, Emil 195, 200, 214 Kitschelt, Herbert 4, 10, 62, 131, 134, 210, 228, 235, 241, 243, 246, 247, 253, 257 Klandermans, Bert 226 Kleinnijenhuis, Jan 169 Klingemann, Hans-Dieter 107, 166, 167, 172, 190, 235 KLMS 131 Knutsen, Oddbjorn 71 Kohler-Koch, Beate 2, 131, 195, 201 Kollman, Ken 167, 168 Konig, ¨ Thomas 195 Kreppel, Amie 8, 126 Kriesi, Hanspeter 154, 218 Kuhnle, Stein 4 Labour Party (Ireland) 130 Labour Party (United Kingdom) 49, 79, 95, 130, 167 Ladrech, Robert 2, 129, 130, 240 Laffan, Brigid 2 Lange, Peter 243, 246 Laver, Michael 94, 97, 168, 169 Lee, Eun Sul 179 Le Pen, Jean Marie 134 left 52, 57–8, 73, 78, 81, 83 left/right dimension 2, 4–8, 14, 15, 22, 23–8, 30, 35, 42, 50, 56, 60, 71, 72, 81, 94, 105, 106, 107, 120–1, 122, 127, 139–40, 197, 210, 229, 236, 237–8, 239, 240–2, 244–5, 250–3, 254 in European Parliament 144–5, 149, 150, 153–7, 158–9, 160, 162–3, 243 liberal parties 96, 110 libertarian 72, 86, 122, 210, 211 Lijphart, Arend 32, 142, 143, 163 Lindberg, Leon 1 Lindstrom, ¨ Ulf 257 Lipset, Seymour Martin 2–4, 32, 142, 210, 248, 258 List Pim Fortuyn (Netherlands) 134, 245 Listhaug, Ola 4, 167, 168 Ljunggren, S. B. 136 Lodge, Milton 3 Long, Scott 16 Lord, Christopher 2, 4, 5, 6–7, 32, 95, 129, 162, 163, 174, 185, 195, 197, 202, 238; see also Hix–Lord model Lorimor, Ronald 179 Lowery, David 204

lowest common denominator 60 Luxembourg 38, 42, 43, 61, 68, 121, 178, 205 Lynch, Philip 189 Maastricht Treaty 1, 128, 136, 138, 143, 185, 201, 202, 240 McAdam, Doug 220, 224, 225 MacDonald, Stuart Elaine 4, 167, 168 Mackie, Thomas 32 Macmillan, Harold 49 Mair, Peter 2, 87, 174 Major, John 49 majoritarian democracy 142 Mannheimer, Renato 35 Mares, Isabela 64 Marks, Gary 1, 3, 5, 8, 9, 10, 15, 16, 17, 32, 51, 86, 93, 94, 110, 120, 121, 122, 126, 128, 129, 130, 131, 144, 150, 154, 159, 174, 191, 197, 202, 214, 216, 220, 223, 224, 225, 228, 235, 238, 240, 242, 243, 244, 245, 246, 251, 253, 255 Marlière, Philippe 2, 129, 240 Marsh, Michael 32, 35, 152 Marshall Plan 201 Martin, Andrew 219, 225 materialist/post-materialist dimension 228–9 Mazey, Sonia 210 Messina, Anthony 135 Meyer, David 218 Middendorp, Cees 154 Miller, John 167, 168 Mitterrand, Fran¸cois 239 Monnet, Jean 246 Moravcsik, Andrew 5, 63 Mouvement pour la France 135 Muller, ¨ Wolfgang 168, 188 multilevel governance 241 Munger, Michael 3, 167, 168, 175 Murrell, Peter 204 National Alliance (Alleanza Nazionale, Italy) 133, 135 National Front (Front National, France) 133–4, 245 neofunctionalism 1, 5 neoliberalism 8–10, 22, 28, 106, 127, 214, 215 neoliberal market 51 Netherlands 34, 35, 38, 42, 43, 49, 61, 68, 70, 75, 84, 142, 152, 178, 205, 217, 245

Index new politics 22, 23, 25, 51, 122, 130, 133, 139, 210, 236, 243–5, 256 Niedermayer, Oskar 2 Niemoller, ¨ Broer 4 Norris, P. 152 Norway 217 Noury, Abdul 10, 51, 60, 141, 147, 148, 149, 156, 243, 252 Offe, Claus 195, 197, 199 Olson, Mancur 199, 207 OPEC 218 Ordeshook, Peter 168 Organization for European Economic Cooperation (OEEC) 201 Page, Scott 167, 168 Panter, Abigail 23, 105 Parsons, Craig A. 1 Party of European Socialists (PES) 96, 97, 109–10, 146, 149, 150–8, 160, 161, 240 party families 93, 110 party manifestos 93, 95–101 PASOK (Panellino Socialistiko Kinima, Greece) 130 Pasqua, Charles 135 People’s Party (Dansk Folkeparti, Denmark) 133 People’s Party (Partido Popular, Portugal) 135 Peterson, John 1, 63 Petrocik, John 167 Pfeiffer, Alois 199 Poland 49, 217 policy space 13–14, 15, 16, 17, 18, 21, 22, 23, 28–31, 93, 94, 110–12 political parties 51 Pollack, Mark 1, 9 Poole, Keith 148 Portugal 35, 38, 42, 43, 61, 68, 69, 70, 75, 84, 100, 160, 205 Powell, Bingham 65, 141 President, European Commission 15 pro-/anti-integration dimension 105, 197, 229, 237 public opinion 51 Rabinowitz, George 4, 167, 168 radical left parties 124 radical right parties 124 Rae index 145 Rally for France (Rassemblement pour la France, France) 135

279 Raunio, Tapio 2, 145, 146, 162 Ray, Leonard 2, 10, 51, 63, 69, 73, 74, 76, 78, 86, 120, 121, 131, 170, 171, 172, 173–5, 185, 188, 238, 245, 246, 253, 257 realism 5 regulated capitalism 8–10, 22, 28, 51, 106, 109, 127–8, 130, 213, 215, 241, 242, 252 regulation model 7–8, 17, 21, 23, 94, 103, 105, 106, 107, 109, 110–11, 126 Reif, Karlheinz 8, 32, 165 residual (liberal) welfare states 67, 68, 73, 76, 79 Richardson, Jeremy 210 Rietberg, Edward 169 right 73, 76, 84–5 Riker, William 123, 167, 168 Risse, Thomas 2, 131 Robertson, David 94, 97, 98, 166, 167, 168, 169, 190 Rokkan, Stein 2–4, 195, 197, 210, 248, 258 Roland, Gérard 148, 252 Rometsch, Dietrich 195 Ronit, Karsten 195 Rosenthal, Howard 148 Ross, George 219, 225 Rudig, ¨ Wolfgang 138 salience 72, 77, 86, 165–6 of European integration 169–71, 173, 179, 180, 182, 183, 184, 185, 187, 189–91, 192 theory 166–70, 176, 177, 178, 186, 232 Sandholtz, Wayne 131 Sbragia, Alberta 2, 252 Scharpf, Fritz 63, 64, 65, 129, 242 Schattschneider, Elmer 36, 41 Scheingold, Stuart 1 Schendelen, Rinus van 206 Scheve, Kenneth 63 Schimank, Uwe 200 Schmitt, Hermann 2, 8, 14, 32, 35, 129, 165, 172, 195, 235, 242 Schmitter, Philippe 2, 131 Scholz, Evi 172 Schofield, Norman 168, 169 Schumpeter, Joseph 141 Schwaiger, Konrad 195 Scott, David 3, 10, 120, 121, 124, 126, 165, 169, 192, 239, 251 Sears, Davis 16 second-order elections 8, 32

280

Index

Séguin, Philippe 136 Sened, Itai 168 Sewell, William 224 Shepsle, Kenneth 167 Shields, J. G. 134 Sidjanski, Dusan 200, 201 Simi, Pete 220 Single European Act (SEA) 1, 201, 202, 239 Sinnott, Richard 2, 235, 237 Slovenia 49, 217 Smith, Mitchell P. 2, 219, 225 social democratic parties 129 Social Democratic Party (Socialdemokratiet i Danmark, Denmark) 130 Social Democratic Party (Sozialdemokratische Partei Deutschlands, Germany) 217 Social Democratic Party (Arbetarepartiet-Socialdemokraterna, Sweden) 79, 128 social democratic welfare state 67, 68, 73, 257 socialist parties 96, 110 Sorbom, ¨ Dag 23, 105 Soskice, David 63, 64, 65, 66, 246, 252 Sowemimo, Matthew 136 Spain 34, 38, 42, 43, 61, 68, 69, 70, 75, 84, 86, 100, 136, 151, 152, 157, 160, 204, 205, 217 Steenbergen, Marco 1, 3, 10, 13, 15, 16, 17, 93, 94, 120, 121, 122, 126, 165, 174, 191, 216, 223, 224, 228, 239, 251, 253, 255 Stephens, John 64, 243, 246 Stinchcombe, Arthur 210 Stone-Sweet, Alec 131 Streeck, Wolfgang 204, 224 Strøm, Kaare 168, 188 structural funds 69, 79, 80, 82, 85 Sv˚asand, Lars 257 Sweden 34, 35, 38, 42, 43, 44, 61, 68, 70, 75, 78, 79, 84, 85, 130, 160, 178, 204, 205, 217, 247, 256 Switzerland 7

Taggart, Paul 1, 120, 124, 238 TAN (traditional/authoritarian/nationalist) 121, 125, 131, 132, 133, 135, 137–40, 244, 246 Tarrow, Sidney 1, 216, 218, 221, 227–8 Thomassen, Jacques 2, 10, 14, 35, 51, 60, 129, 152, 195, 235, 242, 243, 252 traditional/libertarian dimension 144, 153–7, 158–9, 243 Truman, David 196 Tsebelis, George 2, 5, 7, 8, 17, 126; see also regulation model Turkey 135 Turner, Lowell 219, 225 UKIP (United Kingdom Independence Party) 42 UNICE 207 United Kingdom 34, 35, 38, 42, 43, 49, 61, 68, 70, 78, 84–5, 87, 136, 150, 151, 152, 160–1, 167, 188, 201, 205, 211, 217, 245, 247 United States 7, 167 Van der Brug, Wouter 32, 35, 41 Van der Eijk, Cees 4, 8, 10, 32, 35, 37, 41, 46, 47, 48, 165, 171, 185, 189, 190, 195, 223, 237–8, 241, 246, 251 Villiers, Philippe de 135 Visser, Jelle 204 Voeten, Erik 10, 51, 60, 141, 156, 243, 252 Wallonia 34, 38, 42, 43 Wessels, Bernard 2, 10, 35, 130, 152, 195, 199, 204, 210, 211, 212, 235, 241 Wessels, Wolfgang 2, 195 Whiteley, Paul 136 Wilson, Carole 3, 10, 17, 51, 86, 110, 120, 121, 129, 131, 144, 150, 154, 159, 238, 240, 242, 243, 244, 245, 251 Wilson, Graham 196 Wlezien, Christopher 36 World Trade Organization (WTO) 227–8 Wren, Ann 64 Ysmal, Collette 35

European Integration and Political Conflict (Themes in European Governance)

European Integration and Supranational Governance

The European Court and Civil Society: Litigation, Mobilization and Governance (Themes in European Governance)

Leadership and Negotiation in the European Union (Themes in European Governance)

Markets and Moral Regulation: Cultural Change in the European Union (Themes in European Governance)

The EU, NATO and the Integration of Europe: Rules and Rhetoric (Themes in European Governance)

European Political Cultures: Conflict or Convergence?

Governance in the European Union

European Union governance: effectiveness and legitimacy in European Commission committees

Theorizing European Integration

European Financial Integration

Theorizing European Integration

Theorizing European Integration

European Integration Theory

University dynamics and European integration

European Integration: Scope and Limits

Resources, Governance and Civil Conflict (Routledge Ecpr Studies in European Political Science)

Themes in Modern European History 1830-1890 (Themes in Modern European History)

States and Regions in the European Union: Institutional Adaptation in Germany and Spain (Themes in European Governance)

Themes in Modern European History since 1945 (Themes in Modern European History Series)

Themes in Modern European History, 18901945

The European Parliament, the National Parliaments, and European Integration

European Cities: Social Conflicts and Governance (European Societies)

Political Economy of European Integration, The: Arguments and Analysis

European Union History: Themes and Debates

Interlocking Dimensions of European Integration

The economics of European integration

The State of European Integration

The Economics of European Integration

European Integration and Disintegration: East and West

The Political Economy of European Employment: European Integration and the Transnationalization of the

European Integration and Political Conflict (Themes in European Governance)

European Integration and Supranational Governance

The European Court and Civil Society: Litigation, Mobilization and Governance (Themes in European Governance)

Leadership and Negotiation in the European Union (Themes in European Governance)

Markets and Moral Regulation: Cultural Change in the European Union (Themes in European Governance)

The EU, NATO and the Integration of Europe: Rules and Rhetoric (Themes in European Governance)

European Political Cultures: Conflict or Convergence?

Governance in the European Union

European Union governance: effectiveness and legitimacy in European Commission committees

Theorizing European Integration

European Financial Integration

Theorizing European Integration

Theorizing European Integration

European Integration Theory

University dynamics and European integration

European Integration: Scope and Limits

Resources, Governance and Civil Conflict (Routledge Ecpr Studies in European Political Science)

Themes in Modern European History 1830-1890 (Themes in Modern European History)

States and Regions in the European Union: Institutional Adaptation in Germany and Spain (Themes in European Governance)

Themes in Modern European History since 1945 (Themes in Modern European History Series)

Themes in Modern European History, 18901945

The European Parliament, the National Parliaments, and European Integration

European Cities: Social Conflicts and Governance (European Societies)

Political Economy of European Integration, The: Arguments and Analysis

European Union History: Themes and Debates

Interlocking Dimensions of European Integration

The economics of European integration

The State of European Integration

The Economics of European Integration

European Integration and Disintegration: East and West

The Political Economy of European Employment: European Integration and the Transnationalization of the

Recommend Documents