Widening the European Union: The Politics of Institutional Change and Reform (Routledge Advances in European Politics, 5)

RECTO RUNNING HEAD

Widening the European Union

The enlargement of the European Union to include countries from Central and Eastern Europe has become an important political issue. Widening the European Union focuses on those institutional reforms of the union that may be necessary to make the enlargement possible. The institutional structure, originally designed for a union of just six states, might not be suitable for a EU of twentyseven or more member states. An overview of rational choice theories of institutional change is provided before the volume focuses on several aspects of institutional reforms in the Union. Issues discussed include: • The EU’s decision-making process. • The role of the Council in connection with the Euopean Commission and the European Parliament. • The EU’s budget procedure and the arrangements used in the European Monetary Union. • How the EU could maintain the momentum of integration after enlargement. Widening the European Union contains contributions from a distinguished team of European academics. This makes this book a valuable resource to students of the European Union with an interest in its politics and policy, enlargement and institutional reform. Bernard Steunenberg is Professor of Public Administration at Leiden University, the Netherlands. His research interests include the role which political institutions play in the initiation, adoption and implementation of public policy, particularly in the context of the European Union.

Routledge advances in European politics

1

Russian Messianism Third Rome, Revolution, Communism and after Peter J. S. Duncan

2

European Integration and the Postmodern Condition Governance, democracy, identity Peter van Ham

3

Nationalism in Italian Politics The stories of the Northern League, 1980–2000 Damian Tambini

4

Uncertain Europe Building a new European security order Edited by Graham Timmins and Martin Smith

5

Widening the European Union The politics of institutional change and reform Edited by Bernard Steunenberg

6

Institutional Challenges in the European Union Edited by Madeleine Hosli, Adrian van Deemen and Mika Widgrén

7

Europe Unbound Enlarging and reshaping the boundaries of the European Union Edited by Jan Zielonka

Widening the European Union The politics of institutional change and reform Edited by

Bernard Steunenberg

London and New York

First published 2002 by Routledge 11 New Fetter Lane, London EC4P 4EE Simultaneously published in the USA and Canada by Routledge 29 West 35th Street, New York, NY 10001 Routledge is an imprint of the Taylor & Francis Group This edition published in the Taylor & Francis e-Library, 2003. © 2002 Bernard Steunenberg All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data Widening the European Union: the politics of institutional change and reform / edited by Bernard Steunenberg. p. cm. Includes bibliographical references and index. 1. European Union—Membership. 2. European Union—Europe, Central. 3. European Union—Europe, Eastern. I. Steunenberg, Bernard. JN30.W53 2002 341.242´2–dc21 ISBN 0-203-16629-9 Master e-book ISBN

ISBN 0-203-26090-2 (Adobe eReader Format) ISBN 0–415–26835–4 (Print Edition)

2002068074

Contents

List of tables List of figures List of contributors Preface

vii ix x xv

PART I

Introduction 1 Enlargement and reform in the European Union

1 3

BERNARD STEUNENBERG

2 Rational choice and institutional change: an overview of current theories

21

KEITH DOWDING

PART II

Constitutional aspects of enlargement 3 Constitutional issues regarding European Union expansion

39 41

DENNIS C. MUELLER

4 Enlargement of the European Union: determining the boundaries of an optimal legal area

58

HANS-JÖRG SCHMIDT-TRENZ AND DIETER SCHMIDTCHEN

5 The politics of institutional change: diverging goals for reforming the European legislature

80

THOMA S KÖNIG AND THOMA S BRÄUNINGER

6 An even wider Union: the effects of enlargement on EU decision-making BERNARD STEUNENBERG

97

vi

Contents

PART III

Reforming European Union decision-making 7 Enlargement and the Council of Ministers

119 121

JAN-ERIK L ANE AND REINERT MÆL AND

8 Enlargement and the European budget: budgetary decision-making and fiscal constraints

144

L ARS P. FELD, GEBHARD KIRCHGÄS SNER AND HANNELORE WECK-HANNEMANN

9 Enlargement and the European Central Bank

163

JORDI BACARIA, GEORGIOS E. CHORTAREA S AND ANDREA S P. KYRIACOU

PART IV

Dynamics of widening

181

10 A never-ending success story? The dynamics of widening and deepening European integration

183

GERALD SCHNEIDER

11 How political parties, rather than member states, are building the European Union

202

JOSEP M. COLOMER

12 A federal European constitution for an enlarged European Union: insights from constitutional economics

224

FRIEDRICH SCHNEIDER

PART V

Conclusion

239

13 Changing times: the future of enlargement of the European Union and the theoretical challenges ahead

241

BERNARD STEUNENBERG AND ANTOANETA DIMITROVA

Index

249

Chapter Title

vii

Tables

3.1 5.1 5.2

6.1 6.2 6.3

6.4

6.5

6.6

6.7

6.8 7.1 7.2 7.3 7.4 7.5

Distribution of European Union expenditures by budget category, 1985 and 1995 (in percentages) Decision probability in the codecision and consultation procedure compared to 1986 baseline Individual efficacy in the consultation and modified codecision procedure under different settings (inclusiveness indices) Comparative analysis of Council voting rules: probabilities and inertia index for the Commission–Council game Expected benefits of different voting rules in the Council: the strategic power index for the Commission–Council game Comparative analysis of Council voting rules: probabilities and inertia index for the Commission–Council game under a uniform distribution Expected benefits of different voting rules in the Council: the strategic power index for the Commission–Council game under a uniform distribution The effects of Nice: probabilities and inertia index for the Commission–Council game for a uniform distribution and after enlargement (EU-27) The effects of Nice: the strategic power index for individual players in an enlarged Union for the Commission– Council game and a uniform distribution Comparative analysis of Council voting rules after enlargement: probabilities for the Commission–Council game under different distributions Comparative analysis of procedures with Parliament’s involvement: probabilities for different distributions Member state voting power in the Council: Banzhaf power The capacity to block in the EU Council Banzhaf voting power related to population Voting power 2005 The attractiveness of simple majority

50 89

91 101 102

104

105

107

108

110 111 131 131 132 134 135

viii 9.1 10.1 10.2 11.1 11.2 11.3

Tables Regional representation in three monetary unions The length between the application and accession Effects of independent variables on the depth of a unilateral integration stance Fragmentation of the European Union (1979–1999) Seats and voting power of European Political Groups (1979–1999) Voting cohesion of European Political Groups (1984–2004)

167 187 193 206 208 217

Tables

ix

Figures

1.1 4.1 4.2 4.3 5.1 6.1 7.1 8.1 8.2 8.3 8.4 8.5 10.1 10.2 10.3

Interdependencies between enlargement and reforms Optimal size of the state Optimal enlargement Net gains of enlargement Consequences of reform options and the Nice reform on member states’ inclusiveness The game trees for both legislative game forms Unanimity power: individual Banzhaf power indices Decision structure of EU compulsory expenditures Decision structure of EU non-compulsory expenditures EU budgetary policymaking without a fiscal referendum EU budgetary policymaking with a fiscal referendum EU budgetary policymaking with a fiscal referendum after enlargement by twelve new member states The depth and scope of regional integration The impact of the number of member states and the size of the regional system on the depth of integration Enforcement and compensation in enlargement negotiations

7 64 67 69 92 99 129 148 150 152 155 156 185 193 195

x

Authors name

Contributors

Jordi Bacaria is Professor of Economics at the Universitat Autònoma de Barcelona, and Co-director at the European Institute in Mexico, on behalf of the European Commission. His research interests are economic policy, public choice and European integration. Recent publications include ‘EuroMediterranean free trade areas: commercial implications’, Mediterranean Politics 4 (1999) (with A. Tovias), and The Catalan Innovation System: Governing Rapid Changes in Regional Innovation Systems, Cambridge: UCL Press, 1998 (with S. Borràs). Thomas Bräuninger is Emmy Noether-Fellow of the German Science Foundation and Research Associate at the European Union Center of the University of Pittsburgh. His main research interests are in the comparative analysis of legislative decision-making and institutional design in international politics. His recent publications include ‘Making rules for governing global commons. The case of deep-sea mining’, Journal of Conflict Resolution 44 (2000): 604–29 (with Thomas König) and ‘The checks and balances of party federalism’, European Journal for Political Research 36 (1999): 207–34 (with Thomas König). Georgios E. Chortareas is an economist at the Bank of England. He has taught at the University of Connecticut, and he was a post-doctoral Fellow at the Minda De Gunzburg Center for European Studies, Harvard University. His research interests are in monetary policy rules, international macroeconomics and the institutional design of monetary policy. Recent publications includes ‘Central banker contracts, contract costs, and contract targets’, Bulletin of Economic Research, forthcoming (with S. Miller), and ‘Does it pay to be transparent? International evidence from the publication of central bank forecasts’, Federal Reserve Bank of St Louis Economic Review, forthcoming (with D. Stasavage and G. Sterne). Josep M. Colomer is Research Professor in Political Science at CSIC in Barcelona, and CIDE, Mexico. He has taught at the Pompeu Fabra University,

Contributors

xi

as well as the universities of Georgetown, New York, Chicago, and SciencesPoParis. He is the editor of Political Institutions in Europe, London: Routledge, 2002, as well as the author of Political Institutions: Democracy and Social Choice, Oxford: Oxford University Press, 2001; Strategic Transitions: Game Theory and Democratization, Baltimore: Johns Hopkins University Press, 2000; and Game Theory and the Transition to Democracy: The Spanish Mode, Aldershot: Edward Elgar, 1995. Antoaneta Dimitrova is Lecturer at the Department of Public Administration of Leiden University. Her main research interests are democratisation in Central and Eastern Europe and governance and institutional change in the context of the European Union enlargement to the East. Recent publications include ‘The search for convergence of national policies in the European Union: An impossible quest?’, European Union Politics 1 (2000): 201-26 (with Bernard Steunenberg), and ‘Discourses about democracy in Bulgaria: the sharp winds of crisis and the steady course of the ship at sea’ and ‘Slovakia’ (with J. Dryzek and L. Holmes), in J. Dryzek and L. Holmes (eds) Postcommunist Democratization, Cambridge: Cambridge University Press, 2002. Keith Dowding is Professor of Political Science at the London School of Economics and Political Science. He has published Rational Choice and Political Power, Aldershot: Edward Elgar, 1991; The Civil Service, London: Routledge, 1995; Preferences, Institutions and Rational Choice, Oxford: Clarendon Press, 1995 (edited with Desmond King); Power, Buckingham: Open University Press, 1996; and Challenges to Democracy, London: Palgrave, 2001 (edited with James Hughes and Helen Margetts). He has written numerous articles in the fields of urban politics, public administration, British politics, social choice and political theory. He is co-editor of the Journal of Theoretical Politics. Lars P. Feld is Senior Lecturer in Economics at the University of St Gallen (Switzerland). He has been Visiting Fellow at the University of Southern California at Los Angeles and at the University of Rennes, France. His main research interests are public finance and political economics. His publications include Die Direkte Demokratie: Modern, Erfolgreich, Entwicklungs- und Exportfähig‚ Basel: Helbing und Lichtenhahn, 1999 (with Gebhard Kirchgässner and Marcel R. Savioz), Steuerwettbewerb und seine Auswirkungen auf Allokation und Distribution, Tübingen: Mohr Siebeck, 2000; ‘The political economy of direct legislation’, Economic Policy 33 (2001): 329-67 (with Gebhard Kirchgässner); and ‘The impact of corporate and personal income taxes on the location of firms and on employment’, Journal of Public Economics, forthcoming (with Gebhard Kirchgässner). Gebhard Kirchgässner is Professor of Economics and Econometrics and Director of the Swiss Institute for International Economics and Applied Economic Research at the University of St Gallen. He is a member of the Swiss Federal Committee of Business Cycle Questions (‘Kommission für Konjunkturfragen’) and Research Fellow of the CESifo Research Network. His main research

xii

Contributors interests are new political economy, public finance, energy and environmental economics and applied econometrics. He is the author of Homo Economicus, Tübingen: Mohr (Siebeck), 1991 (2nd edition 2000); Demokratische Wirtschaftspolitik: Theorie und Anwendung, München: Vahlen, 1994 (with Bruno S. Frey), and Die Direkte Demokratie: Modern, Erfolgreich, Entwicklungs- und Exportfähig, Basel: Helbing und Lichtenhahn, 1999 (with Lars P. Feld and Marcel R. Savioz).

Thomas König is Professor at the Department of Politics and Management of the University of Konstanz. He recently published ‘Examining the EU legislative process: the relative importance of agenda and veto power’, European Union Politics 2 (2001): 329-51 (with Mirja Pöter) and ‘In view of ratification: Governmental preferences and domestic constraints at the Amsterdam Intergovernmental Conference’, International Organization, forthcoming (with Simon Hug). Andreas P. Kyriacou is Associate Professor at the Department of Applied Economics of the Universitat Autònoma de Barcelona, Spain. His research interests are rational choice, constitutional economics, European studies and conflict resolution. Recent publications of his include ‘A ‘just and viable’ solution to the Cyprus problem: in search of institutional viability’, Mediterranean Politics 5 (2000): 54-75, ‘An ethnically based federal and bicameral system: the case of Cyprus’, International Review of Law and Economics 20 (2000): 261-8, and ‘A viable solution to the Cyprus problem in the context of European Union accession’, The Cyprus Review 12 (2000): 35-59. Jan-Erik Lane is Professor of Political Science at the University of Geneva. His research interests include political theory, comparative politics and public administration. He has published widely in this field, including New Public Management, London: Routledge, 2000; New Institutional Politics, London: Routledge, 2000; and Culture and Politics, London: Aldershot, 2002 (with Svante Ersson). Reinert Mæland is a statistician at Lund University, Sweden. His main research interests are computational statistics, econometrics, and cooperative game theory. Recent publications include ‘Constitutional analysis: The power index approach’, European Journal of Political Research 37 (2000): 31-56 (with Jan-Erik Lane), and ‘The growth of the public sector in Switzerland’, West European Politics 24 (2001): 169–90 (with Jan-Erik Lane). Dennis C. Mueller is Professor of Economics at the University of Vienna. His research lies in the broad areas of industrial organisation, public choice and constitutional political economy. Recent publications include Constitutional Democracy, Oxford: Oxford University Press, 1996; Perspectives on Public Choice, Cambridge: Cambridge University Press, 1997; and Public Choice III, Cambridge: Cambridge University Press, forthcoming in 2002.

Contributors

xiii

Hans-Jörg Schmidt-Trenz started his academic career at the University of Saarland and the University of Michigan (1980). He completed his doctoral thesis (Foreign Trade and the Territoriality of Law: Foundations of a New Institutional Economics of International Transactions, Baden-Baden: Nomos,1990) in 1988 and his habilitation (The Logic of Collective Action By Delegation, Tübingen: Mohr-Siebeck, 1996) in 1995. Since 2000, he has been Professor of Economics at the University of Saarland, Saarbrücken, Germany, and gives lectures on the new institutional economics of international transactions and the logic of collective action. In 1988 he started a parallel career as a Director of the Chamber of Commerce of Karlsruhe. Since 1996 he has been CEO of the Hamburg Chamber of Commerce. Dieter Schmidtchen is Professor of Economics and Director of the Center for the Study of Law and Economics at the Universität des Saarlandes, Saarbrücken. His research interests are in the fields of antitrust regulation, law and economics, and institutional economics. His recent publications include: Constitutional Law and Economics of the European Union, Cheltenham: Edward Elgar, 1997 (with Robert Cooter); ‘Strategic power in the European Union: evaluating the distribution of power in policy games’, Journal of Theoretical Politics 11 (1999): 339-66 (with Bernard Steunenberg and Christian Koboldt); and ‘Beyond comitology: a comparative analysis of implementation procedures with parliamentary involvement’, in Peter Moser, Gerald Schneider and Gebhard Kirchgässner (eds) Decision Rules in the European Union: A Rational Choice Perspective, Houndmills/Basingstoke: Macmillan, 2000: 72–98 (with Bernard Steunenberg and Christian Koboldt). Friedrich Schneider is Professor of Economics and Vice-President for Foreign Affairs at the Johannes Kepler University of Linz, Austria. He has held visiting positions at Carnegie-Mellon University, Pittsburgh (USA), Aarhus University (Denmark), La Trobe University in Melbourne (Australia), and the University of Saarland in Saarbrücken (Germany). His research interests include economic policy, shadow economy, environmental economics, privatisation and deregulation policies. Recent publications include ‘No change for incentive-orientated environmental policies in representative democracies? A public choice analysis’, Ecological Economics 31 (1999): 123-38 (with Juergen Volkert), and The Shadow Economy: Theoretical Approaches, Empirical Studies, and Political Implications, Cambridge: Cambridge University Press, 2001 (with Dominik Enste). Gerald Schneider is a Professor of Political Science in the Department of Politics and Management at the University of Konstanz and Executive Editor of European Union Politics. His research focuses on decision-making, political reform, armed conflict and foreign policy. His most recent publications include The Rules of Integration, Manchester: Manchester University Press, 2001 (co-edited with Mark D. Aspinwall), Asylpolitik auf Abwegen, Opladen: Leske & Budrichlisher, 2001 (co-authored with Thomas Holzer), and articles

xiv

Contributors

published in the Journal of Conflict Resolution, International Migration Review, European Journal of Political Research and European Urban and Regional Studies. Bernard Steunenberg is Professor of Public Administration at Leiden University, the Netherlands. His main research interest is the analysis of decision-making at the European Union and neo-institutional approaches to social interaction. Recent publications include ‘Strategic power in the European Union: evaluating the distribution of power in policy games’, Journal of Theoretical Politics 11 (1999): 339–66 (with Dieter Schmidtchen and Christian Koboldt), and Political Institutions and Public Policy: Perspectives on European Decision Making, 1997, Dordrecht: Kluwer (with Frans van Vught). Hannelore Weck-Hannemann is Professor of Political Economy at the Institute of Public Economics at the University of Innsbruck, Austria. Her main research interests are the analysis of non-market decision-making, public choice theory, international political economy and environmental economics. Her publications include Politische Ökonomie des Protektionismus: Eine Institutionelle und Empirische Analyse, Frankfurt and New York: Campus, 1992; ‘Globalization as a challenge for public choice theory’, Public Choice 106 (2001): 77–92; and ‘Do people care about democracy? An experiment exploring the value of voting rights’, Public Choice 91 (1997): 27–47 (with Werner Güth).

Contributors

xv

Preface

This book is the result of the activities conducted in the framework of the Network on Enlargement and New Membership of the European Union (NEMEU), which was supported by the European Science Foundation (ESF) from November 1995 until June 1998. The network aimed to contribute to our understanding of the enlargement of the EU and the extent to which enlargement would affect the European Union. The chapters in this volume are the result of the seminars of the NEMEU network. As the network coordinator I would like to thank ESF for supporting our efforts to shed light on institutional changes related to enlargement of the Union. I would also like to thank the Netherlands Organisation for Scientific Research (NWO) whose support under the project ‘Enlarging the European Union: A comparative analysis of constitutional change’ has made further research on enlargement possible. Finally, I would like to thank Monique Snippers and Marcia Clifford for excellent support in the organisation of NEMEU meetings, and Anna Kruithof and Angelique Steen for editorial assistance. BERNARD STEUNENBERG Leiden, December 2001

xvi

Authors name

RECTO RUNNING HEAD

Part I

Introduction

2

Authors name

Chapter Title

1

3

Enlargement and reform in the European Union Bernard Steunenberg

Introduction The next enlargement with countries from Central and Eastern Europe has become an important political issue in the European Union. The general view is that this enlargement will substantially affect the functioning of the Union. The institutional structure of the European Union, which was initially designed for a Union of only six member states, might not be suitable for a European Union of twenty-seven or more member states. The problems associated with enlargement and their possible solutions have been discussed at two intergovernmental conferences (in 1996–7 and in 2000) leading to treaty reforms made at Amsterdam and Nice. Concerns about the consequences of the Eastern enlargement were already expressed in the early 1990s. The Copenhagen European Council, which provided the general criteria for EU membership, stressed that ‘[t]he Union’s capacity to absorb new members, while maintaining the momentum of European integration, is also an important consideration in the general interest of both the Union and the candidate countries’ (European Council 1993: 6). Besides reforms within the candidate countries, enlargement would also require adaptations on the side of the Union. In the subsequent years, concerns about the consequences of enlargement and calls for changing the Union’s institutions have become louder. At the beginning of the preparations for the 1996–7 Intergovernmental Conference, a working group of the European Council declared that [t]he next enlargement will be a great opportunity for Europe and will also be different from the previous ones because of the large number of applicant countries and the heterogeneity of their political, economic and social situations. To ensure that the next enlargement does not weaken, change the nature of or actually break up the Union, the reforms needed to cope with the challenges involved must first be made. (Reflection Group 1995: 3) Clearly, as indicated by this statement, reform of the Union is seen as a condition for further enlargement. In the same vein, the European Commission (1996: 2)

4

B. Steunenberg

suggested that ‘[a]s the number of Union members increases, it creates a risk of the Union being watered down’. The Commission proposed several reforms including changes in the decision-making rules and procedures of the Union. The European Parliament adopted recommendations based on similar concerns, and called for a reform of the Union’s decision-making process (European Parliament 1995, 1996). More recently, Jean-Luc Dehaene, Richard von Weizsäcker and David Simon expressed these fears in their report on the institutional implications of enlargement. In this report, the ‘three wise men’ stated that ‘[a] significant increase in the number of participants automatically increases problems of decision-making and management’ (Dehaene et al. 1999: 6). Finding agreement among the Union’s member states on institutional reform has been difficult. The 1996–7 Intergovernmental Conference was intended to settle this and many other institutional matters. However, the European heads of state and government decided at the Amsterdam European Council in June 1997 to postpone the decision on the voting rules in the Council and other institutional matters. The delegations from Belgium, France and Italy protested, but they were not able to convince the others to go ahead with new and substantial changes. The Nice Council in December 2000 was more successful in this respect. At Nice the member states resolved more problems. The Treaty of Nice (2001) includes a re-weighting of the vote shares of the member states in the Council and a reduction of the number of Commissioners to one per member state for the next years before moving to a system in which these posts might rotate between the members.1 However, the Nice Treaty still has to be ratified by the member states. The recent ‘no’ of the Irish population in a referendum indicates that the coming into force of this treaty could turn out to be problematic. In addition, some issues are again postponed and are scheduled for yet another intergovernmental conference in 2004 and a possible treaty reform. Enlargement does not only raise numerical problems. It also raises questions about how decisions are made and are accepted by the European citizens. Creating a Union of twenty-seven member states with a total population of about 481 millions has the potential to reinforce the existing democratic deficit in the Union and put further pressure on the Union’s legitimacy. The European Parliament could play an important role in representing the views of the European citizens. Although the European Parliament has become a co-legislator over the last years as part of the Union’s codecision procedure, its position in the Union’s decision-making process is still relatively weak. Besides codecision, various other decision-making procedures still exist, especially in the second and third pillars of the Union, in which Parliament hardly plays a role. Some procedures allow the Council to take decisions without any parliamentarian involvement, while others require the Council just to consult Parliament on its proposed measures. Consultation, of course, is not similar to making a decision. It does not allow Parliament to fully voice its opinion in the final decision. In addition, Parliament lacks other decision-making rights national parliaments in Europe have, including the right of initiative, which is reserved to the European Commission, and full budgetary power.

Enlargement and reform in the EU

5

Even if the member states could agree on reforming the Union’s decision-making institutions, including an even more prominent role for the European Parliament, the question remains whether this would improve the Union’s legitimacy. As Weiler (1999: 80) indicates, changes in the Union’s institutions affect ‘formal legitimacy’, since they ensure that decisions are made through a democratic process. However, ‘social legitimacy’, that is, the general acceptance of the system of collective decision-making, is a different matter and may remain unaffected by these measures. Making the rather complicated codecision procedure more transparent, or increasing Parliament’s role in the political and legislative process, may not reduce the gap that already exists between the Union’s decision-making bodies, including the European Parliament, and the European citizens. The everdecreasing turnout at European Parliament elections is revealing. To improve the Union’s legitimacy we need much more than only a transfer of power between the Council and Parliament. We also need the acceptance by the European citizens of the decisions taken in Brussels. A major challenge for research on the European Union today is to critically examine various calls and proposals for reform and the process of institutional change related to the enlargement of the Union. How will enlargement affect decision-making in the European Union or, more generally, European integration? What will be the effects on the current and the future member states? How can some of the perceived disadvantages of enlargement be overcome by changing the current EU institutions? Which reforms are politically feasible and which are not? And what will be the possible side effects of these reforms? To answer these questions, we need a better understanding of the main characteristics of institutions and the way in which they change. This requires theoretical and empirical analyses of institutions and institutional reform. The contributions in this volume aim to address these issues, which are highly relevant for the future of Europe.

The process of enlargement Currently the associated countries of Central and Eastern Europe as well as Cyprus and Malta are negotiating the terms and timing of their membership. The first round of accession negotiations started after the Luxembourg European Council in December 1997. At this meeting the current member states decided to open formal negotiations with all applicant countries. At the same time, enlargement negotiations were only started with a first group of six countries. The remaining countries were offered annual screenings of the progress made in view of the well-known Copenhagen criteria, which set out the general conditions for EU membership.2 As Friis (1998) describes, the outcome reached at the Luxembourg summit was a result of political bargaining between the member states in which the Commission proposals, set out in Agenda 2000, played an important role. The first group of countries that were invited to the negotiation table were the Czech Republic, Estonia, Hungary, Poland and Slovenia. Cyprus was also part of this group since it already submitted its application for membership in 1990,

6

B. Steunenberg

while Greece had achieved an agreement earlier on Cyprus’ membership. At that stage the Union indicated that Bulgaria, Latvia, Lithuania, Rumania and Slovakia experienced more difficulties in satisfying the Copenhagen criteria. The Luxembourg Council decided to delay negotiations with these countries to provide them with more time to introduce the necessary changes. The Council asked the Commission to deliver annual progress reports on all the candidates, which would allow the second group of applicant countries to catch up. Malta, which applied for membership in 1990 and froze its application when Malta’s Labour Party came into power, rejoined the accession process in 1998. Two years later, at the Helsinki European Council in December 1999, the Union followed up its earlier decision made in Luxembourg. It decided to start negotiations with the second group of countries as well, sometimes referred to as the Helsinki group. This decision marked a change in the Union’s accession policy from dealing with groups to the so-called ‘regatta’ approach. Entry seems to depend more on the progress made by individual countries than on some predetermined order.3 At the same summit, the member states agreed to grant Turkey, which had already applied for membership in 1987, candidacy status. However, the Union has not yet opened negotiations with Turkey. The opening of the accession negotiations has created expectations that the applicant countries will join the Union in the coming years. These expectations form the driving force behind substantial transformations in the applicant countries that are a result of the conditions set out for EU membership. Transformations that aimed to bring these countries closer to stable democratic political systems with a market economy, able to implement the acquis communautaire.

The interdependency of enlargement and reform A driving factor in the current enlargement process is the Union’s credibility in pursuing enlargement based on its ‘historical’ promise to the former socialists countries in Central and Eastern Europe after the fall of the Berlin wall. The criteria for membership require reforms within the applicant countries that have to be implemented before EU membership can be granted. This involves substantial efforts in the domestic political arena, which are only feasible if they are recognised by the Union. Without the prospect of entry, the governments of the applicant countries would not be able to defend and implement all of these reforms. However, the Union reserves the right to take a decision on entry only after successful reform. This makes the ‘exchange’ between the Union and the applicant country problematic. Only when an applicant country has delivered its part of the ‘deal’, will the Union decide on its membership. The sequential nature of this process, in which both parties engage in nonsimultaneous exchange, allows for the possibility that the Union goes back on its promise. A decision on entry could, in principle, be delayed or even postponed for many years. Knowing this, an applicant country may not be that eager to push for domestic reform, mitigating the changes the Union wants to accomplish. As a consequence, the Union will be less eager to consider this country as member, which reduces the Union’s credibility. This downward spiral could lead to post-

Enlargement and reform in the EU

7

poning and finally giving up enlargement. To avoid this, the Union has provided safeguards and possible rewards that strengthen its credibility. The Association Agreements can be regarded as an initial expression of the Union’s willingness to pursue enlargement.4 The opening of accession negotiations, as decided by the Luxembourg and Helsinki European Councils, the opening of new chapters during these negotiations, and the creation of special support programs for applicant countries (ISPA, SAPARD) signal the Union’s intentions to complete this enlargement. Despite these steps, which aim to improve the applicant’s confidence in the Union, the Union’s credibility can still be easily disrupted due to the length of the process. An important factor that could damage the Union’s credibility is the progress of the Union’s internal reforms. Since reforms are regarded as a condition for further enlargement, their progress affects the expectations of the applicant countries. To the current member states, enlargement provides an opportunity to renegotiate some of the political ‘deals’ that were struck in the past. This may concern changes of the Union’s decision-making rules and procedures as well as policies, such as the common agricultural policy and structural funds. All current member states need to accept the outcome of these reforms – or the fact that some institutions or policies will not be changed – in order to reach a stage at which they will unanimously support enlargement. If some members disagree, or feel that yet another solution might be more preferable, it will trigger another round of negotiations, which will slow down enlargement. The applicant countries could interpret such delays as lack of commitment for enlargement. This reduces the Union’s credibility and as a result the applicant’s willingness to pursuit further domestic change may drop. Enlargement thus critically depends on the efforts made in both the applicant countries and the current Union. The discussed interdependencies between enlargement and reform are illustrated in Figure 1.1. If an applicant country does not initiate or slows down its reforms, due to lack of support in the domestic political arena, it will not satisfy the conditions for entry. Possible entry to the Union is delayed until these reforms have been made. However, this situation may not endanger the entry of other applicant countries, which may satisfy the memberSupport for EU reforms by the current member states Support for domestic reforms in the applicant country insufficient

sufficient

Figure 1.1

insufficient

sufficient

no enlargement

no enlargement due to insufficient reforms in the applicant country

no enlargement due to EU credibility problem

enlargement of the Union

Interdependencies between enlargement and reforms.

8

B. Steunenberg

ship criteria. If, on the other hand, the Union fails to introduce sufficient reform within its policies and institutions as a result of disagreement between the current member states, enlargement will be endangered. A downward spiral may occur in which an applicant country questions the Union’s credibility leading to a slowing down of domestic reforms and postponement of enlargement. Enlargement is only feasible when both the external and the internal reforms are handled in such a way that applicant countries have successfully implemented their reforms and the current member states have reached agreement on institutional reform of the Union. The bulk of the current literature on EU enlargement pays little attention to the importance of reforms within the Union. Some studies approach enlargement as one of the EU policy sectors (Falkner 1996; Sedelmeier and Wallace 2000), like agriculture or transport, and discuss its costs and benefits (Baldwin et al. 1997). Others stress the importance of the negotiations between the Union and the applicant countries, which are expected to affect the process of enlargement (Preston 1997; Redmond 1997; Avery and Cameron 1998). The main perspective taken in this book, which differs from these studies, is that the Eastern enlargement very much depends on a successful reform of the Union itself.

Analysing enlargement as a process of institutional reform A different perspective on enlargement Analysing enlargement as a process of institutional reform requires also a different approach to research. So far, most research on enlargement draws upon extensive cases studies, in which past enlargements of the Union have been thoroughly analysed (Arter 1993; Redmond 1997). Some of this work takes a comparative perspective and analyses past enlargements next to previous ones (Preston 1997), or contrasts enlargements of the European Union with enlargements of other international organisations in Europe (Croft et al. 1999). All these studies try to answer the question ‘why things happened as they happened’ by focusing in great detail on the perceived course of events. Despite their empirical richness, most of these studies do not employ a welldeveloped theoretical framework. Such a framework would point at the underlying causal mechanisms that could account for possible variations in the outcomes of decision-making affecting enlargement. Without a theoretical framework, the identified reasons for why a specific course of events may have occurred – however well documented by the empirical material – do not enhance our insights on future enlargement processes. It is not clear whether and to what extent the identified reasons for a specific course of events will be effective in other circumstances. And even when these ‘reasons’ again play a role under other circumstances, it is not clear whether they will lead to similar consequences. Based on a descriptive approach on enlargement, it is impossible to tell whether a next enlargement will develop along the same lines as one of the previous enlargements, since the underlying mechanisms remain unclear.

Enlargement and reform in the EU

9

The often-heard statement that the Eastern enlargement is ‘unique’ and ‘cannot be compared to previous enlargements’ illustrates the fact that our general knowledge about these decision-making processes is still rather limited. As indicated, this lack of understanding is partly due to the strong emphasis on case study research in this field and lack of a clear theoretical orientation. This problem with enlargement studies has also been noted by Helen Wallace (2000: 149). She writes in a recent review that ‘it is an oddity of the literature on the EU, and more generally on (west) European integration, that so little effort has been made to theorize about the enlargement of the EU’. In this volume we therefore approach enlargement from a different perspective. We present theory-driven analyses of the institutional consequences of the Union’s wish for enlargement. Since our main focus is on the institutional adaptations of the Union, the contributions are embedded in institutional rational choice. This analytical framework allows us to analyse and explain the process and outcomes of decision-making, which are taken as the result of the interactions between political actors within the context of a specific set of institutions. Institutional rational choice Institutional rational choice can be regarded as a response to some of the rational or public choice-literature in the late 1970s. Searching for the stability of political decision-making, which initially has been set on research agenda by Arrow (1951), several scholars have examined whether stability could be based on the preferences of political actors involved. With the exception of decisions that only involve one policy dimension,5 or are based on an extremely symmetrical distribution of preferences,6 the answer was as simple as it was staggering: in a more dimensional policy space and using majority rule, actors cannot reach a stable outcome. The result is chaos in the sense that any option in the policy space can be the outcome, which could be manipulated by the actors who control the political agenda (McKelvey 1976; Schofield 1978). At the same time, this result also shows that, from a theoretical perspective, collective decisions cannot be based on preferences alone. Particularly Tullock (1981) has raised the question how this result relates to daily life of politics in which decisions are made using majority rule and preserved for many years. The discrepancy between the theoretical results derived from rational choice models and every-day life has led to a shift in perspective. Instead of focusing on preferences, scholars began to analyse how institutions may affect political decision-making. That is, how institutions could provide stability and, at the same time, how institutions shape the outcome of collective decision-making. This approach is called institutional rational choice, which initially started to analyse the impact of existing institutions, mostly taken as rules or procedures, on the possibilities of making collective decisions (Shepsle 1979, 1989; Shepsle and Weingast 1981; Ostrom 1986). From this work, a large number of studies evolved analysing the extent to which political institutions influence the interactions between different political actors and thus affect the making and contents of public policy.

10

B. Steunenberg

European Union politics In the area of European Union politics, institutional rational choice has led to a large number of studies analysing the Union’s legislative process. Since this process is regulated by several different legislative procedures and voting rules, it offers an excellent opportunity to compare the possible effects of different institutions. Depending on the different emphasis on ways in which political actors are inclined to engage in collective decision-making, at least two different areas of study have emerged. The first area is voting power studies, which emphasise the power of different actors and how different institutions affect the distribution of power among these actors. In these studies power reflects the ability of an actor to decide on a new policy by joining a ‘winning’ coalition, which could successfully endorse this new policy. Institutions in this area are often taken as voting rules, or rules that connect voting processes in different decision-making bodies so that a ‘winning’ coalition over these different bodies can be computed. In the context of the Union, a large number of studies have assessed the power of the member states in the Council of the European Union.7 Not surprisingly, the member states hold the same voting power under unanimity rule since each of them has the possibility of a veto. In case of qualified majority voting, the voting weights as provided by the Treaty (Article 205.2 EC) lead to a distribution of power between member states that differs from what one would expect based on key indicators such as GNP or population size. In addition, the expected voting outcomes in the European Parliament have been analysed presuming that ‘nationality’ or ‘ideology’ determines the parliamentarians’ voting behaviour.8 With an enlargement leading to more ministers in the Council or more representatives in Parliament, the power scores for individual countries would drop. In the case of ideologybased power scores, the effect will be less straightforward, since it depends on the political preferences of the newly elected MEPs. A ‘natural’ extension to these studies combines voting in the Council and Parliament and provides ‘overall’ power indices by assuming that representatives in both bodies are motivated in the same way (Hosli 1995; Bindseil and Hantke 1997; Colomer and Hosli 1997). However, this link appears to be problematic, since ministers and MEPs are differently motivated and are not necessarily motivated by only nationality or ideology. Their preferences, which guide their behaviour, seem to be at least a mixture of both elements (see Scully 2002). An interesting insight which stems from this literature is that institutions taken as voting rules could have two distinct effects on power.9 On the one hand, a voting rule has an effect on an actor’s ability to block change, which provides a safeguard against unfavourable proposals. An extreme example of such a rule is unanimity rule, which only allows an actor to stop change by using its veto power. On the other hand, voting rules differ to the extent to which they allow for change and thus provide actors the opportunity to move to a more preferred outcome. In this way, voting rules may also function as a safeguard against inertia, that is, the impossibility of changing the current state of affairs. Combining both effects, it is not evident that political actors would only prefer a

Enlargement and reform in the EU

11

rule increasing their ability to block policy change. Especially when the same rule is used for a large number of different decision-making problems – which is typical for institutional choice problems – the prospect of change could increase the actors potential to be part of a ‘winning’ coalition. As a result, political actors might expect to receive relatively higher benefits on average than under the circumstance in which change could be more easily blocked. Voting power studies mainly focus on the players’ abilities to affect the outcome of voting based on the formation of ‘winning’ coalitions. Other institutional aspects of the process of decision-making, such as differences in authority or limited access to the political agenda, are not addressed. This forms a limitation of these studies in understanding the role of institutions in the political process (Garrett and Tsebelis 1996; Tsebelis and Garrett 1996; Steunenberg et al. 1999). The procedural studies, which form the second area of research, focus on the institutional features that shape decision-making. They aim to understand how the Union decides on policies as a result of the preferences of actors and institutions. These institutions are now primarily regarded as the procedures that govern the Union’s decision-making process. Starting with the legislative procedures, such as the older consultation and cooperation procedures (Tsebelis 1994), this work quickly expanded to the different versions of codecision when new treaties were adopted (Crombez 1997, 2000; Steunenberg 1997).10 In a similar way the implementation procedures of the Union (Steunenberg et al. 1996, 1997), which are known as comitology, as well as the budget procedure (Feld et al. in Chapter 8 of this volume) have been modelled. The procedural studies show that institutions matter in the sense that they affect the outcomes political actors can obtain. With regard to enlargement, these studies show that its possible effects are the combined result of institutions and preferences, which makes matters slightly more complicated. If enlargement leads to more preference heterogeneity, it would complicate decision-making. If the Union’s legislative institutions, such as the frequent use of unanimity rule, are kept the same, enlargement would lead to more instances in which the Union cannot take a decision. Enlargement then leads to less policy flexibility. If the institutions are adapted, this effect could be mitigated and the Union’s ability to respond to new policy challenges would be maintained. The procedural studies reveal a strong sensitivity to the way in which the political process is structured. This has led to various theoretical debates on the consequences of different model specifications pointing to the limitations of such studies.11 First, most of the work is theoretical and the models that are developed have not yet been put to a test in the sense that predicted outcomes are confronted with the actual outcomes of decision-making. Second, most models of the Union’s legislative procedures are based on some interpretation of the formal procedures as indicated by the EC Treaty. This perspective creates a kind of ‘formalistic bias’ that neglects ‘informal’ practices and other working methods. For instance, the role of the Presidency in Council discussions and the informal consultations between the Council, Parliament and the Commission in preparation of a conciliation committee procedure may also matter.

12

B. Steunenberg

Both limitations require further work, especially empirical research in order to sift between ‘useful’ and ‘not useful’ models, which would increase our insights in the importance of institutions. Yet voting power studies as well as procedural studies have started to open the ‘black box’ of decision-making in the Union. The contributions in this volume continue this line of research and explore the extent to which institutional change and reform determine the prospects of EU enlargement towards the East.

Structure of the book After an overview and further discussion of current rational choice theories on institutional change by Keith Dowding in Chapter 2, the chapters in the book will be organised into three main parts, each covering an important issue area with regard to the widening of the European Union. Following the institutional perspective taken in this book, we start with constitutional aspects of enlargement. The second part of the book discusses specific reforms of the Union’s decision-making process in view of enlargement, including decision-making by the Council, budgetary decision-making and decision-making on monetary policy. The third part focuses on the dynamics of enlargement, that is, how the Union could maintain European integration after enlargement. Chapter 13 concludes the book. Constitutional aspects of enlargement The first part starts by focusing on the constitutional aspects of enlargement and deals with some of its problems and possible solutions. Starting from the premise that any governmental structure should advance the interests of its citizens, the question arises whether enlargement of the Union could be ‘optimal’ from the perspective of both the current members and the applicant countries. As the fiscal federalism literature in economics suggests, a ‘local’ provision of collective goods could be more efficient when groups of voters prefer different kinds or quantities of these commodities. A ‘centralised provision’, like the European Union, would lead to inefficiency, unless local provisions carry with it disadvantages that outweigh this gain (Oates 1972: 11–12). Furthermore, a ‘local’ provision may increase the legitimacy of government, since it allows voters to better select goods and services that match their preferences. Dennis Mueller, in Chapter 3, starts from this perspective and examines the procedures that could be used in order to admit new members to the European Union, both from the perspective of the current member states and the newly entering states. This contribution takes up issues regarding entry requirements, cross-national subsidies, voting rules, secession clauses and the free mobility of citizens in the Union. Hans-Jörg Schmidt-Trenz and Dieter Schmidtchen continue in the same vein in Chapter 4 and focus on the limitations to the scope of the acquis communautaire. Their approach is to seek whether a uniform set of legal rules, such as the acquis, is restricted to a group of individuals or citizens of a certain size. Increasing this

Enlargement and reform in the EU

13

size beyond some point could lead to ‘inefficiencies’ in the sense that both current and newly admitted members could become ‘worse off ’. Especially in the case of a ‘second wave of enlargement’, the paradoxical situation could arise that the member states’ willingness to admit a second group of applicant countries would be reduced if these countries are successful in adopting the acquis. This is what Schmidt-Trenz and Schmidtchen call the applicant’s dilemma. In the contribution of Thomas König and Thomas Bräuninger in Chapter 5, the focus is shifted from the effects of enlargement on citizens to the ability of political actors, such as ministers or MEPs, to affect the outcome of decisionmaking in a future Union. Using insights from voting power studies, they analyse the proposed changes of the Union’s decision-making rules in order to allow for enlargement from the perspective of inclusiveness. This concept indicates the extent to which political actors in the Union could be part of a coalition that supports new policy measures. The continuous decrease of the Union’s effectiveness in making decisions since the mid-1980s could be addressed by lowering the voting quota and eventually introducing simple majority voting. However, this would affect the current distribution of power between member states based on their inclusiveness in ‘winning’ coalitions. König and Bräuninger indicate, parallel to what happened in Nice, that member states would only accept a lower voting quota to increase the Union’s effectiveness provided the actual balance between member states is preserved by changing the member states’ voting weights. Bernard Steunenberg in Chapter 6 discusses the effects of enlargement in terms of the Union’s ability to make new decisions (‘policy flexibility,’ or effectiveness in König and Bräuninger’s terminology), and the power to the individual member states. Contrary to the contribution of König and Bräuninger, which is embedded in the voting power literature, his analysis is an example of the procedural studies on European Union decision-making. Using non-cooperative game theory, he models the Union’s decision-making process as a sequential game in which political actors are allowed to make specific decisions in different stages. Steunenberg employs simulation techniques to show how changes in the number of member states as well as their preferences may affect decision-making. His analysis shows that the Union’s decision-making capacities critically depend on the voting rules that are used and on the heterogeneity of the preferences of the member states. Policy flexibility already appears to be very limited in the current Union when the Council has to apply unanimity rule. Moreover, when qualified majority rule is applied, enlargement does not affect flexibility when the preferences of the current and new members are based on similar distributions.12 In addition, the analysis shows that the new qualified majority rule as introduced by the Nice Treaty will lead to more instances of deadlock, which reduces the Union’s flexibility in making new policy while the opposite was intended. Focusing on individual member states, the analysis illustrates the effects of enlargement and changes in voting rules. It appears that the introduction of qualified majority voting in areas where unanimity rule applies could have the neteffect that all current member states would gain in power, also after enlargement.

14

B. Steunenberg

Reforming European Union decision-making The second part discusses the institutional reforms of the Union in order to allow for further enlargement. These reforms include the decision-making in the Council of the European Union – formerly the ‘Council of Ministers’13 – which is the main focus of Chapter 7 by Jan-Erik Lane and Reinert Mæland. Using insights from voting power studies, Lane and Mæland argue that in the Council the voting power of a member state is a function of its individual power to block and the group’s collective power to decide. This individual ‘power to block’ is often valued by constitutional economists who take this as a safeguard against group decisions that violate the preferences of individual members. The ‘power to decide’ is mostly appreciated by public administration scholars, who stress the importance of making new policies in order to react swiftly on changing circumstances so that these policies do not become dysfunctional. Lane and Mæland focus on both aspects and show how enlargement will affect the power positions of the current member states. As the voting power of the current members drops when the Union is enlarged, they suggest transforming the voting in the Council towards simple majority voting. They also demonstrate that, based on their methodology, the reforms introduced by the Nice European Council gave priority to the member states’ blocking power. The ability to make new decisions – or the Union’s policy flexibility – will be reduced by these reforms, as has been shown in Chapter 6 as well. Lars Feld, Gebhard Kirchgässner and Hannelore Weck-Hannemann discuss the Union’s budgetary decision-making process in Chapter 8 and analyse how the size of the Union’s budget is affected by aspects of this procedure. They show that the current procedure tends to lead to a higher budget than the one a substantial number of member states prefer. In view of this result, Feld, Kirchgässner and Weck-Hannemann discuss the possibility of introducing a budgetary referendum at the EU level in order to realign the outcomes of decision-making to the preferences of voters. Their analysis shows that a referendum, which gives voters veto power, only prevents outcomes that strongly deviate from the median voter’s most preferred budget. It is therefore an ‘imperfect’ safeguard against too much spending. At the same time, the introduction of a referendum restricts the range of feasible budgets, so that political decision-makers in the Union cannot deviate too strongly from voter preferences. These properties stay when the number of member states increases due to enlargement. In addition, the introduction of budgetary referenda in the EU could also be a way to involve more European citizens in the Union’s decision-making process. In Chapter 9 Jordi Bacaria, Georgios Chortareas and Andreas Kyriacou focus on the making of monetary policy, which, as part of the European Monetary Union (EMU), has been delegated to the European System of Central Banks. This ‘system’ consists of the European Central Bank and the National Central Banks of the participating countries. Based on the importance of credibility in monetary policymaking, Bacaria, Chortareas and Kyriacou indicate that possible differences in participating national economies and thus of the preferred policy measures to be taken by the ECB could be a threat. As more countries may want

Enlargement and reform in the EU

15

to participate in EMU, in connection with a further enlargement of the EU, possible dissent about the measures to be taken could increase as the diversity of the participating economies increases. At the same time, the authors indicate that some regional representation in EMU is necessary as a check against too much concentration of decision-making power at the ECB. In response to this trade-off between credibility and decision-making power, Bacaria, Chortareas and Kyriacou suggest alternative institutional arrangements for monetary policymaking in Europe. These arrangements include denationalisation of regional representation and rotating membership of states in the monetary policy-setting body. Dynamics of widening The last part of the book focuses on the dynamics of European integration and discusses alternative arrangements that may facilitate further cooperation in Europe. As indicated, many observers as well as the main political actors in the Union contend that the connection between widening and deepening poses a major problem for the European integration process. In Chapter 10 Gerald Schneider argues that these fears might be overstated. Analysing the dynamics of European integration in a model that combines aspects of trade theory and non-cooperative bargaining games, he shows that most observers fail to differentiate between the beneficial effects of system change and the detrimental impact that enlargement has on the effectiveness of the organisation. As European integration gains momentum, adding more members to the Union becomes less desirable as it increases transaction costs. Focusing on the accession negotiations, Schneider notes that the success of the current Union could be a reason why both enlargement and internal reform have become so complicated. In addition, he shows that in accession negotiations with unequal partners, patience pays off, which is an additional reason for the current delays. Josep Colomer in Chapter 11 presents a rather different view on the process of European integration. His analysis is based on recent developments within the current Union, particularly in the European Parliament. He claims that political party formation and coalition building in the European Parliament is the driving force of further European integration, allowing the Union to govern relatively effectively a highly pluralistic society. In spite of successive enlargements and the very high number of electoral parties obtaining representation in the European Union institutions, the number of effective European Political Groups in the European Parliament has decreased. Using voting power analysis, Colomer explains why national parties have an incentive to join large European Political Groups instead of forming smaller nationalistic groupings. Furthermore, he also shows empirically that the cohesion of European Political Groups has increased and that the European Parliament has gained in importance in the Union’s legislative decision-making process. As a consequence of these developments, intergovernmentalism is being replaced with federalising relations. Therefore Colomer is rather optimistic about the governability of the Union after enlarge-

16

B. Steunenberg

ment provided that new member states have party systems fitting the European Political Groups. Finally, Friedrich Schneider argues in Chapter 12 that the Union needs more explicitly grounded institutional arrangements in order to maintain a legitimate governmental body which is able to serve the interests of citizens. His analysis, which differs from Colomer’s, is grounded in the constitutional political economy literature, which analyses different arrangements of collective decisionmaking. Schneider argues that in order to provide a successful enlargement a Federal European Constitution is needed, which consists of four basic elements. The first element is that the European legislation should consist of a twochamber system with full responsibility over all federal items. The next two elements are the subsidiarity principle and federalism. Both are, according to Schneider, best suited to limiting the domain of the central European authority to which certain tasks are given. A fourth element is direct democracy, which provides the possibility for European voters to participate actively in the political decision-making, to break political and interest group cartels, and to prevent unwanted shifts in responsibilities from EU member states to the European Union.

Notes 1 Although the Nice European Council was held in December 2000, the final text for a new treaty was renegotiated afterwards and finally signed on 26 February 2001. 2 As indicated at the Copenhagen European Council of June 1993, new member states need to satisfy the following conditions: (a) stability of democratic and judicial institutions that guarantee democracy, the rule of law, human rights and respect for and protection of minorities; (b) the existence of a functioning market economy, as well as the ability to cope with competitive pressures and market forces within the Union; and, (c) the ability to take on the obligations of membership, including adherence to the aims of political economic and monetary union (European Council 1993: 6). See Nicolaides et al. (1999) for a further discussion of these criteria and other requirements to applicant countries. 3 The agreement reached at the Laeken summit (December 2001) suggests that enlargement takes place in two waves. This is the ‘10+2’ schedule: a first group of ten applicant countries and a second group consisting of Bulgaria and Romania. 4 Despite the insistence of the candidates, the Association Agreements did not include a clause specifying that these will lead to membership. 5 See Black’s median voter theorem, which is based on a single policy dimension and single peaked preferences (Black 1987 [1958]). 6 See the results derived by Plott (1967), which presume an extreme symmetry in the distribution of preferences to allow for a stable outcome. 7 See, for instance, the studies by Johnston and Hunt (1977), Hosli (1993, 1994, 1995), Widgrén (1994), Johnston (1995), and Berg and Lane (1996). 8 See, for instance, Taylor and Johnston (1978), Fowler et al. (1983), Lane et al. (1995), and Hosli (1997). 9 See König and Bräuninger (1998), and Berg and Lane (2001). See also the contributions of König and Bräuninger in this volume (Chapter 5), and of Lane and Mæland (Chapter 7). 10 See Steunenberg (1994) and Crombez (1996) for comparative analyses of these different procedures. 11 See the debates between Moser (1996) and Tsebelis (1996) on Parliament’s role in the

Enlargement and reform in the EU

17

cooperation procedure, between Scully (1997a, 1997b) and Tsebelis and Garrett (1997) on similar issues, and between Crombez (2001), Steunenberg (2001), and Tsebelis and Garrett (2000) on Tsebelis and Garrett’s claim that the change, introduced by the Maastricht Treaty, from the cooperation procedure to the codecision procedure was a ‘bad deal’ for Parliament. 12 The preferences of the member states are thus not similar, only their distributions are. Therefore, in a specific case, the current and the new member states may have substantial differences in opinion. 13 The term ‘Council of Ministers’ is still broadly used, also by authors in this book. We therefore use both terms synonymously.

References Arrow, K. (1951) Social Choice and Individual Values, New York: Wiley. Arter, D. (1993) The Politics of European Integration in the Twentieth Century, Aldershot: Dartmouth. Avery, G. and F. Cameron (1998) The Enlargement of the European Union, Sheffield: Sheffield Academic Press. Baldwin, R., J.F. Francois, and R. Portes (1997) ‘The costs and benefits of Eastern enlargement: the impact of the EU and Central Europe’, Economic Policy 24: 125–76. Berg, S. and J.-E. Lane (1996) ‘Game theory, power, and the European Community’, in: M. Wiberg (ed.) Multidisciplinary Views on Strategic Interactions, Turku, Finland: University of Turku Studies on Political Science no 14: 17–34. Berg, S., and J.-E. Lane (2001) ‘Measurements of voting power: individual and collective properties with applications to the EU’, in: H. Holler and G. Owen (eds) Power Indices and Coalition Formation, Dordrecht: Kluwer, forthcoming. Bindseil, U. and C. Hantke (1997) ‘The power distribution in decision making among EU member states’, European Journal of Political Economy 13: 171–85. Black, D. (1987 [1958]) The Theory of Committees and Elections, Boston/Dordrecht/Lancaster: Kluwer. Brennan, G. and J.M. Buchanan (1985) The Reason of Rules: Constitutional Political Economy, Cambridge: Cambridge University Press. Colomer, J.M. and M.O. Hosli (1997) ‘Decision-making in the European Union: the power of political parties’, Aussenwirtschaft 52: 255–80. Croft, S., J. Redmond, G.W. Rees and M. Webber (eds) (1999) The Enlargement of Europe, Manchester: Manchester University Press. Crombez, C. (1996) ‘Legislative procedures in the European Community’, British Journal of Political Science 26: 199–228. Crombez, C. (1997) ‘The co-decision procedure in the European Union’, Legislative Studies Quarterly 22: 97–119. Crombez, C. (2000) ‘Institutional reform and co-decision in the European Union’, Constitutional Political Economy 11: 41–57. Crombez, C. (2001) ‘Codecision: towards a bicameral European Union’, European Union Politics 1: 363–8. Dehaene, J.L, R. von Weizsäcker, and D. Simon (1999) The Institutional Implications of Enlargement: Report to the European Commission, Brussels, 18 October 1999. European Commission (1996) Reinforcing Political Union and Preparing for Enlargement, Commission Opinion, Brussels, 28 February 1996. European Council (1993) Conclusions of the Presidency: European Council in Copenhagen, 21–22 June 1993 (http://europa.eu.int/rapid/start/cgi/guesten.ksh?p_actiongettxt= gt&doc=DOC/93/3/0/AGED&lg=EN).

18

B. Steunenberg

European Parliament (1995) Report on the Functioning of the Treaty on European Union with a view to the 1996 Intergovernmental Conference: Implementation and Development of the Union, Committee on Institutional Affairs (Bourlanges/Martin report), 4 May 1995 (A4–0102/95). European Parliament (1996) Report on (i) Parliament’s Opinion on the Convening of the IGC; and (ii) Evaluation of the Work of the Reflection Group and Definition of the Political Priorities of the European Parliament with a View to the Intergovernmental Conference, Committee on Institutional Affairs (Dury/Maij-Weggen report), 5 March 1996 (A4–0068/96). Falkner, G. (1996) ‘Enlarging the European Union’, in J.J. Richardson (ed.) European Union: Power and policy-making, London: Routledge, 233–46. Fowler, L.L., P.L. Polhuis and S.C. Paine (1983) ‘Changing patterns of voting strength in the European Parliament’, Comparative Politics 15: 159–75. Friis, L. (1998) ‘ “The end of the beginning” of Eastern enlargement: Luxembourg summit and agenda-setting’, European Integration online Papers 2(7):1–15 (http://eiop.or.at/ eiop/texte/1998–007a.htm). Garrett, G. and G. Tsebelis (1996) ‘An institutional critique of intergovernmentalism’, International Organization 50: 269–99. Hosli, M.O. (1993) ‘Admission of European Free Trade Association states to the European Community: effects on voting power in the European Community Council of Ministers’, International Organization 47: 629–43. Hosli, M.O. (1994) ‘Der potentielle einfluß der Schweiz im EG-Ministerrat: eine Analyse aufgrund des Banzhaf-machtindexes’, Swiss Journal of Economics and Statistics 130: 89–105. Hosli, M.O. (1995) ‘The balance between small and large: effects of a double majority system on voting power in the European Union’, International Studies Quarterly 39: 352–70. Hosli, M.O. (1997) ‘Voting strength in the European Parliament: the influence of national and of partisan actors’, European Journal of Political Research 31: 351–66. Johnston, R.J. (1995) ‘The conflict over qualified majority voting in the European Union Council of Ministers: an analysis of the UK negotiating stance using power indices’, British Journal of Political Science 25: 245–54. Johnston, R.J. and A.J. Hunt (1977) ‘Voting power in the E.E.C.’s Council of Ministers: an essay on method in political geography’, Geoforum 8(1): 1–9. König, T. and T. Bräuninger (1998) ‘The inclusiveness of European decision rules’, Journal of Theoretical Politics 10: 125–42. Lane, J.-E., R. Mæland, and S. Berg (1995) ‘The EU parliament: seats, states and political parties’, Journal of Theoretical Politics 7: 395–400. McKelvey, R.D. (1976) ‘Intransitivities in multidimensional voting models’, Journal of Economic Theory 12: 472–82. Moser, P. (1996) ‘The European Parliament as a conditional agenda setter: what are the conditions? A critique of Tsebelis (1994)’, American Political Science Review 90: 834–8. Nicolaides, P., S.R. Boean, F. Bollen and P. Pezaros (1999) A Guide to the Enlargement of the European Union (II): A Review of the Process, Negotiations, Policy Reforms and Enforcement Capacity, Maastricht: European Institute of Public Administration (revised edition). Oates, W. E. (1972) Fiscal Federalism, New York: Harcourt, Brace and Jovanovich. Ostrom, E. (1986) ‘An agenda for the study of institutions’, Public Choice 48: 3–25. Plott, C.R. (1967) ‘A notion of equilibrium and its possibility under majority rule’, American Economic Review 57: 787–806. Preston, C. (1997) Enlargement and Integration in the European Union, London: Routledge. Redmond, J. (ed.) (1997) The 1995 Enlargement of the European Union, Ashgate: Aldershot.

Enlargement and reform in the EU

19

Reflection Group (1995) Reflection Group’s Report, Brussels, 5 December 1995 (SN 520/95). Schofield, N. (1978) ‘Instability in simple dynamic games’, Review of Economic Studies 45: 575–94. Scully, R.M. (1997a) ‘The European Parliament and the co-decision procedure: a reassessment’, Legislative Studies 3: 58–73. Scully, R.M. (1997b) ‘The European Parliament and co-decision: a rejoinder to Tsebelis and Garrett’, Legislative Studies 3: 93–103. Scully, R.M. (2002) ‘Going native? Institutional and partisan loyalty in the European Parliament’, in: B. Steunenberg and J. Thomassen (eds) The European Parliament: Moving Toward Democracy in the EU, Boulder: Rowman and Littlefield, forthcoming. Sedelmeier, U. and H. Wallace (2000) ‘Eastern enlargement: strategy or second thoughts?’ in H. Wallace and W. Wallace (eds) Policy-making in the European Union, Oxford: Oxford University Press (4th edition), 427–60. Shepsle, K.A. (1979) ‘Institutional arrangements and equilibrium in multidimensional voting models’, American Journal of Political Science 23: 27–60. Shepsle, K.A. (1989) ‘Studying institutions: Some lessons from the rational choice approach’, Journal of Theoretical Politics 1: 131–47. Shepsle, K.A. and B.R. Weingast (1981) ‘Structure-induced equilibrium and legislative choice’, Public Choice 37: 503–19. Steunenberg, B. (1994) ‘Decision making under different institutional arrangements: Legislation by the European Community’, Journal of Theoretical and Institutional Economics 150: 642–69. Steunenberg, B. (1997) ‘Codecision and its reform: a comparative analysis of decision making rules in the European Union’, in: B. Steunenberg and F.A. van Vught (eds) Political Institutions and Public Policy: Perspectives on European Decision Making, Dordrecht: Kluwer, 205–29. Steunenberg, B. (2001) ‘Seeing what you want to see: the limits of current modelling on the European Union’, European Union Politics 1: 368–73. Steunenberg, B., C. Koboldt, and D. Schmidtchen (1996) ‘Policymaking, comitology, and the balance of power in the European Union’, International Review of Law and Economics 16: 329–44. Steunenberg, B., C. Koboldt, and D. Schmidtchen (1997) ‘Beyond comitology: European policymaking with parliamentary involvement’, Aussenwirtschaft 52: 87–112. Steunenberg, B., D. Schmidtchen, and C. Koboldt (1999) ‘Strategic power in the European Union: evaluating the distribution of power in policy games’, Journal of Theoretical Politics 11: 339–66. Taylor, P.J. and R.J. Johnston (1978) ‘Population distributions and political power in the European Parliament’, Regional Studies 12: 61–8. Tsebelis, G. (1994) ‘The power of the European Parliament as a conditional agenda setter’, American Political Science Review 88: 128–42. Tsebelis, G. (1996) ‘More on the European Parliament as a conditional agenda setter: response to Moser’, American Political Science Review 90: 839–44. Tsebelis, G. and G. Garrett (1996) ‘Agenda setting power, power indices, and decision making in the European Union’, International Review of Law and Economics 16: 345–61. Tsebelis, G. and G. Garrett (1997) ‘Agenda setting, vetoes and the European Union’s codecision procedure’, Legislative Studies 3: 74–92. Tsebelis, G. and G. Garrett (2000) ‘Legislative politics in the European Union’, European Union Politics 1: 5–32. Tullock, G. (1981) ‘Why so much stability?’, Public Choice 37:189–202.

20

B. Steunenberg

Wallace, H. (2000) ‘EU enlargement: A neglected subject’, in: M. Green Cowles and M. Smith (eds) The State of the European Union, Volume 5: Risks, Reform, Resistance, and Revival, Oxford: Oxford University Press, 149–63. Weiler, Joseph H.H. (1999) The Constitution of Europe. ‘Do the new clothes have an emperor?’ and other Essays on European integration, Cambridge: Cambridge University Press. Widgrén, M. (1994) ‘Voting power in the EC decision making and the consequences of two different enlargements’, European Economic Review 38: 1153–70.

Chapter Title

2

21

Rational choice and institutional change An overview of current theories Keith Dowding

Institutional rules and political battles Rational choice institutional models are an important new paradigm for studying institutions and institutional change in the EU. We can trace the beginnings of the institutional approach in both political science and economics to a set of institution-free theorems. In political science these go back to Arrow’s General Possibility Theorem (Arrow 1951a, 1963), in economics to the fundamental theorems of welfare economics (Arrow 1951b) and also Coase’s (1937, 1960) work. Descriptive institutional analysis grew up in apparent ignorance of formal institutional analysis. The importance of institutions was first formally demonstrated in Arrow’s Theorem which shows the outcome of any preference-aggregating procedure depends upon the procedural principles adopted as much as the preferences of the collectivity. In other words, different preference-aggregating procedures may produce different results from the same collection of ordinal preferences. Similarly, much of institutional rational choice has taken off from a set of social choice theorems sometimes called ‘chaos theorems’. These follow from the famous Condorcet problem of majority rule where for three voters and three alternatives, each alternative can be majority defeated by another. The existence of such intransitivity depends on the complexity of the decision problem and the likelihood of the relevant preference profiles occurring. McKelvey (1976, 1979) demonstrates that unless preferences are symmetrical around a median point there is a global cycle involving all choices. Schofield (1978, 1983) shows that multidimensional voting games typically have an empty core in which players have no equilibrium strategies. Many have suggested that these theorems predict ‘policy chaos’ (Riker 1982), or rather no predictions (Austen-Smith and Banks 1998). Either way they suggest that institutions may reduce the set of viable options and the possibility of cycles, or ensure they occur around a restricted set. It also helps agenda-setters to control policy outcomes. Work developing from Arrow’s Theorem teaches that the rules by which decisions are made partially determine outcomes. From the fundamental theorems of welfare economics we learn that the distribution of property rights determines the terms of efficient trade. All political battles must therefore be over the distribution of property rights or, more generally, rights to control certain transactions.

22

K. Dowding

If the rules that exist were created by the winning coalition of earlier political battles they will remain stable as long as they further the interests of that winning coalition, or until a new winning coalition can form to overturn them. In practice, rent seeking occurs when different coalitions form in different issue areas in order to change those rules that only apply in that issue area. The overarching constitutional rules upon which political battle is engaged are much harder to change and alter rarely in most states. One reason why the European Union (EU) has proved so interesting is that here we have a new set of institutional rules that have been modified in apparently minor, though in fact fundamental ways to affect the relative powers of different coalitions of actors.

Institutions for stability1 ‘Why so much stability?’ has been a concurrent theme of much of formal political science. Significant progress in a markedly institutional form came from the rational choice modelling of the US legislature, leading to the concept of ‘structure-induced equilibrium’. The examination of chaotic outcomes was used to examine the nature of politics in the legislature of the US Congress, which suggested that often the outcomes were arbitrary or manipulated (Ferejohn and Fiorinna 1975). However, while these results – often described as structure-free – were not entirely non-institutional, since the nature of successful manipulation and potential arbitrariness depend on the precise rules governing the translation of votes into outcomes, they suggest a degree of cycling that is not apparent. The response to what Shepsle and Weingast (1995) describe as the ‘first-generation models’ was to examine ways in which institutions channel legislators’ selfinterest in structure-induced equilibrium to end possible voting cycles. The second-generation models propose a series of rules and regulations creating ‘structure-induced equilibrium’. These models suggest that while Congress is about the politics of distribution, it is not a simple account of majority rule, rather a continuing game of strategic interaction played within a certain institutional environment. This environment is a set of smaller committees, covering limited jurisdictions. Reducing the number of players, and reducing the dimension of conflict, reduces the probability of cycling and the opportunities for vote trading. Furthermore, members of congressional committees tend to be legislators with an interest in the issues within the purview of the committee, leading to a greater homogeneity of perspective, or at least to make the disputes more onedimensional. The procedural routines of these committees further restrict the opportunities for the emergence of cycling and other instabilities. These exogenous institutions are built into the formal model because they are known to exist in the real world but are not explained by it. Nevertheless they suggest why there is more stability than first-generation models predict. Generally in second-generation models, committees are seen as preferencebased or demand-side institutions, which increase the welfare of legislators (by enabling them to direct goods to their constituents) through vote trading. However, committees may be conceived not as agents of distribution economising on transaction costs, but rather as specialised factors of production. They could act

Rational choice and institutional change

23

as welfare-increasing institutions by increasing productive efficiency on the supply side, thereby raising the quality of legislative policy deliberation, debate and decision-making by creating expert legislators who specialise in a given policy arena. They are able to ask for, and examine, more detailed proposals from the executive and bureaucracy and thereby increase the amount of control that the legislature has over government agencies. Third-generation models developed from these ideas, but more importantly they tried also to explain how institutional rules arise. Krehbiel (1991) argues that since all legislative choices are majoritarian they should determine not only what happens now but also what rules will operate in the future. Any legislature is therefore run under ‘remote majoritarian choices’. Despite relative longevity institutions can be changed by majority coalitions. Krehbiel presumes therefore that current institutions must serve today’s majority interests. This assumption makes institutions endogenous. It explains institutions as the result of previous coalition decisions that survive because in some sense they serve current coalition objectives. Changing rules thus requires bringing a new majority coalition into being, which may involve creating new ideological cleavages or heresthetically bringing to the fore new issues. Krehbiel further distinguishes between policy instruments and policy effects. He asserts that legislators do not know the precise relationship between the instruments they select to get what they want, and the precise results or outcomes these instruments actually produce. He suggests that uncertainty can be modelled into a systematic and a random component. The first is the known relationship between a policy instrument and what it is likely to produce. But in any particular application of a policy instrument, the actual outcome cannot be known because of the random element. Third-generation models assume legislators know this and so need to update their information regularly. A better-informed Congress is more likely to produce legislation it prefers. The latter may be directly fed into transaction cost economics explanations of changes in legislative–executive relations, and the reshaping of bureaucratic organisational forms. Krehbiel’s model thus suggests that the committee structure of the US Congress exists because committee expertise produces productive efficiencies and increases legislative oversight of the bureaucracy. This leads to a superior view of debate, which is not simply a bargaining or logrolling process but a way of aggregating and revealing information for overall advantage. Finally, within each issue area we may assume a single dimension – this is a result of the informational modelling process Krehbiel adopts – which means that a unique equilibrium outcome is attained – i.e. the median voter result obtains. While there may be opportunities for some vote trading across issues and committee, these are constrained. Thus we attain stability. Cox and McCubbins’ (1993) third-generation variant uses the important role of parties in grouping and organising the preferences of legislators into coherent blocs. Even though US parties are weaker than those in Europe, Cox and McCubbins argue that they also provide a source of stability, providing a median party position which dominates when there is a majority party, though party outliers may at times vote with the opposition to produce a more median

24

K. Dowding

Congressional position. Versions of this model are compatible with Krehbiel’s (1993) position. In other words, these third-generation models assume that institutions have developed over time in order to increase efficiency and stability. Literature on European legislatures has been more concerned with the formation of governments, and the coalition models have been largely institutionfree, concentrating on the power of respective parties and governments as measured by one or other power index. Sometimes such models assume the equi-probability of simple power developing from Riker’s (1962) work (powerseeking models). Rival forms build in assumptions about preferences (policyseeking models) (see Laver and Schofield 1990; Dowding 1995). Laver and Shepsle (1996) introduce a European form of institutional analysis, borrowing from the second-generation literature examining the US legislature. Their spatial analysis of coalition formation introduces ministerial or departmental constraints on the possible bargaining solutions making coalition governments more stable. Each minister controls policy in any issue dimension and thus bargaining over potential policy issues is constrained by bargaining over portfolio control. Recent work on the European Union follows these same two lines of attack, with some following the power index approach and other work using spatial analysis constrained by institutional rules. The introduction of departmental constraints, like Krehbiel’s legislative oversight accounts, brings to mind the nature and role of the bureaucracy. Institutional rational choice has made important contributions to the study of bureaucracy and its relationship to the legislature. Early public choice accounts of the bureaucracy, notably Niskanen’s budget maximising model, are broadly institution-free. Niskanen (1971) sees bureaucrats as rational seekers of larger – latterly discretionary (Niskanen 1991, 1994) – budgets analogous to private-sector managers seeking to maximise profits. This very simple model, together with a rent-seeking component, leads to conclusions about the ever-expanding state, although more complex conclusions result once greater institutional detail is included. Adding institutional detail to distinguish different elements of a bureau budget, with different bureaux having different institutional forms (Dunleavy 1991), suggests that only some may seek to grow, while others reshape themselves into smaller control agencies, shifting productive activities to other public or private sector agencies. The degree of discretion politicians want to give to the bureau, and the core bureaux to agencies under their jurisdiction, depends on the nature of the transaction costs (Horn 1995).

Transaction costs for stability Institutional or transaction cost economics and transaction cost politics are more concerned with public policy than public administration. They are more concerned with considering the problems encountered in forming and implementing policies through the policy process. In these models, non-governmental actors, notably pressure groups, play an important role, as well as citizen-voters. A large part of the focus is upon the politics of the bureaucracy, both in its

Rational choice and institutional change

25

relationship to the executive and in the relationship the executive and the bureaucracy have to pressure lobbies. If everyone knew everything about everyone else, all property rights were assigned and respected, and everyone could make completely specified, fully binding agreements, then efficiency would be guaranteed, with each of the parties getting what their bargaining strength determines. Changing the initial property to another set also guarantees efficiency. This is the Coase Theorem (Coase 1960). It is also the basis of transaction cost economics (Coase 1937; Williamson 1985, 1996), which essentially argues that the main subject of economics should be that which impedes the specification, monitoring or enforcement of economic trade or transactions. Credibility is an important aspect of any transaction. Changes in the nature of rent, such as the removal of a subsidy or tariff, will hurt a given industry. An efficient solution would be to compensate the workers and employers in the industry for the capitalised value of the losses in a lump sum. However, a government would only do that in response to the promise of the people in the industry to vote for the party in government. But such a promise is not credible. Once the lump sum has been paid, there may be some residual good will, if those in the industry believe the payment was adequate compensation, but this is not enough to guarantee their vote. Of course, a lump sum paid to one group is a sum that cannot be paid to another. Promising the money over time will not help, for then there is a credibility problem from the other side. Why should the industry believe the money would be paid in the future? Can those leaving the industry be guaranteed this payment? The credibility problems, or what Williamson calls time consistency problems, are insurmountable (Dixit and Londregan 1995; Dixit 1996). North (1990) and Dixit (1996) have introduced transaction cost politics and argue that political markets are beset by even greater transaction costs than economic ones. North concentrates on what he terms failures in instrumental rationality – basically the effect of information costs on decision-making. Dixit has a broader theme, encompassing all the transaction costs identified by Williamson and arguing that systemic attempts to overcome these transactions can be seen through institutional and organisational change. The most important difference between transaction cost economics and transaction cost politics is that in economics the parties contract given the asset assignments. Whilst much trading also takes place within the asset assignment, politics is largely about asset assignment. Whereas economics is about redistribution – parties trading goods with one another for mutual benefit – politics is largely about transfers: what laws are going to be created which create the conditions under which parties will trade with each other. Moe (1990) argues that the power to govern or exercise public authority is analogous to a property right. In dictatorship we see governors governing in their own interest, albeit within certain constraints (Wintrobe 1998), but in a democracy this tendency is mitigated by regular or reasonably frequent elections. This makes the property right contingent or uncertain. Each party in power tries to attain its goals and to bind future governments to keeping them. Where

26

K. Dowding

government constantly changes, the outcome will be some sort of compromise as current policies are built on the backs of previous ones and thus embody some concessions to parties currently out of power. The US system of checks and balances, Moe argues, leads to delay, ineffectiveness and a failure of decisionmaking. In Britain, with its stronger executive and periods with one party in power, governments are more able to be effective in pursuing policies, and in transforming many aspects of state institutions. Game theorists tend to assume that institutions are the equilibrium of repeated games. Individuals continue to follow an equilibrium strategy because they have no incentive to depart from it. Of course, there may be no individual incentive to depart from some strategy, even though it would be better if all individuals behaved differently. Equilibria may be local and globally superior institutions may exist if only they can be achieved. Of course, different institutions may also be Pareto incomparable where some actors gain and some lose from a different institutional form. Thus institutions both within legislatures and across bureaucracies may form some local equilibria, though not achieve a wider efficiency for the actors concerned.

Applications to the EU The major debate among institutional rational choice writers concerns the nature of the power relationships between the major set of institutional actors. The EU has a complex set of relations. Power resides within the Council of Ministers and the European Commission and between these two and the European Parliament. The European Court of Justice is also relatively powerful, spawning some articles echoing literature on the US Supreme Court. The literature on the institutions of the EU can be split into that which uses one of the power indices to measure the power of organisational actors which concentrates mostly upon the power of coalitions of nations within the Council of Ministers (for example, Johnston and Hunt 1977; Hosli 1996; Herne and Nurmi 1993; Widgrén 1994), or the European Parliament (Lane et al. 1995; Hosli 1997) and that which uses non-cooperative game theory (the spatial approach) which concentrates upon the relative powers of Parliament, the Council and the Commission (Steunenberg 1994a, 1994b; Tsebelis 1994, 1995; Garrett and Tsebelis 1996; Tsebelis and Garrett 2000). These two different approaches not only use different techniques of analysis, they also reach different conclusions. The first models power as the capacity of actors. It looks at what actors may potentially achieve given the resources of other actors. The resources it concentrates upon are the votes that each actor has within each arena. It utilises the assumption of universal domain and analyses the set of all potential coalitions to map the power of each actor. It then examines actual votes to examine which actors seem to get what they want less often than others do. The spatial approach is less interested in the set of all potential coalitions. It restricts analysis to the coalitions predicted by spatial connectedness, arguing that only connected coalitions are likely to form. Once the spatial mapping is

Rational choice and institutional change

27

complete we may then develop a picture of the ‘true power’ of each actor – or at least see which actors will get what they want most often. The spatial approach is then able to build in other institutional features, together with the relationships between different EU institutions, trying to understand the relative powers of actors within Parliament, the Commission, the Council of Ministers and the European Court of Justice. Various non-intuitive findings have been revealed from this analysis. Which approach is the most fruitful? They are not as contradictory as is sometimes suggested. The power index approach considers what Kiser and Ostrom (1982) call the ‘constitutional choice level’, while the non-cooperative game theory approach is based at Kiser and Ostrom’s ‘collective choice level’. Both may be used to reveal important truths, though equally both can be misused. The power index approach can demonstrate the constitutional powers of actors within a given forum, such as (coalitions of) MEPs within Parliament or (coalitions of) ministers within the Council of Ministers; or it can measure the constitutional power of different bodies where the rules of voting are straightforward. What it cannot do is combine these measures, nor account for the power of different bodies where the rules of procedure are not straightforward voting games. Here we need to move to the non-cooperative game theory approach. Spatial analysis can demonstrate the way in which different actors manage to produce outcomes in their favour through coalition formation with like-minded actors. It can show how the EU may operate in ways that the constitutional designers did not intend. Importantly, it has been argued that the conditional agenda power of supranational bodies such as Parliament, the Commission and the European Court of Justice enables the selection among a set of possible equilibria, accelerates European integration and dilutes responsibility for politically unpopular measures. However, both approaches have problems. The simple way in which the power indices are summed across different constitutional bodies is problematic. One cannot simply compare the index power of one actor in one body such as Parliament with the index power of another actor, in the Council, say, even if one can compare the index power of each of these bodies. Furthermore, it is limited to actors’ voting powers, but this is only one element of the resources that actors can bring to power games (Dowding 1991, 1996). The spatial approach has been bedevilled by modelling incorrectly specifying the constitutional and conventional powers of the different bodies and has tended to be theoretical, using anecdotal rather than hard empirical evidence. With changes in the size of the Council and decision-making by qualified majority vote, calculating the voting resources of each country or sets of countries within the Council under these changes has been of great interest to political scientists. Some controversy exists over the better index to use (Garrett et al. 1995; Johnston 1995b; Felsenthal and Machover 1998, 2001), but greater, and more interesting controversy exists over the use of the indices themselves. This involves two issues: first whether the indices should take account of the preferences of each actor when calculating that actor’s power. Secondly, the index approach ignores important institutional features of the EU and therefore mis-

28

K. Dowding

specifies the true power of actors. The second criticism is robust and correct and is considered below. The first criticism is less assured and the issue somewhat messy. Power indices are constructed by assigning a score to each voter v for all winning coalitions C in which v is pivotal such that C – v is not winning. The scores are then normalised so the sum of the power of each voter is one. The power of each nation in the Council, or the power of coalitions of nations such as ‘large nations’, or the ‘poor nations’ can then be calculated given the weighted votes of each and the precise rules governing winning coalitions. This is of obvious interest, since, for example it has been argued (Johnston 1995a) that the largest member states (Germany, the UK, France and Italy) have more power when the blocking majority is twenty-seven than when it is twenty-three. As the EU is further widened the power scores with weighted voting may also be of some interest (Felsenthal and Machover 2001; Lane 2002), though it is also true that the larger a committee the less interesting such scores are in themselves. With larger committees, the distribution of preferences predicting likely coalitions becomes much more pertinent. Other writers have examined the actual and potential voting rules to see the relative powers of nations in terms of GDP, population and other factors, questioning the wisdom of the actual voting rules (Widgrén 1994; Lane and Mæland 1995; Hosli 1996). Similar applications of the indices have been given to nations and political groups within the European Parliament (Herne and Nurmi 1993; Lane et al. 1995; Hosli 1997). As far as this analysis goes, there is nothing wrong with using power indices in this manner. They can be used to demonstrate the relative voting power resources of actors. However, it is a mistake to think that measuring voting power, which is the difference an actor can make in getting what she wants, bears any relationship with the probability that some actor will actually achieve what she wants. Actors may achieve what they want in voting games without needing to vote at all. This is called their luck (Barry 1991; Dowding 1991, 1996). If actors share preferences they are more likely to vote together. This increases their power as a coalition, but does not increase their individual power; rather, the increased coalitional power of the actors is a result of their luck in sharing preferences. Where those coincident preferences are not mere happenstance, but rather depend on structurally induced preferences such as ‘being relatively poor’, ‘having large agricultural interests’, ‘northern’ versus ‘southern’ states, ‘large’ versus ‘small’ states and so on, then we may see this luck as ‘systematic’ or structured by exogenous factors (Dowding 1991, 1996). When building a constitution such systematic preference equivalencies may need to be taken into account, in order to defend what may otherwise be persistent minorities (Dowding 1999). Taking account of systematic preference equivalencies to predict how often certain coalitions form is an interesting and useful exercise, but is not the same as measuring individual powers. In criticising the power index approach, Garrett and Tsebelis (1996, 1999) argue that the ‘likely influence’ of each nation is related to their centrality in policy space and not the normalised power index score. More correctly, we should say that the probability of a nation getting what it wants is related to its centrality in policy space. This is not the same as its

Rational choice and institutional change

29

influence because it is ludicrous to suggest that countries can become more influential by discovering the preferences of other nations, then lining up their votes in the middle. All attempts to incorporate connectedness into power index calculations (Lane and Mæland 1995; Colomer and Hosli 1997) confuse the probability of coalitions forming with the normalised voting power of the constituent members of the coalitions. The non-normalised Banzhaf ‘index’ is a measure of the conditional probability of each voter turning a winning into a non-winning coalition. The Garrett and Tsebelis critique of the power index approach also makes this conceptual error. The probability of one’s preferences lining up with others in natural coalitions is separate from one’s voting power. Being central in policy space is some form of luck. It may lead to having greater power, for being pivotal in this sense may then enable a country to trade votes in one issue area for votes in another. But that is a separate issue, however, though one which directly pertains to Garrett and Tsebelis’ more important criticism of the power index approach to examining powers in the EU, notably its ignorance of other institutional features. Decision-making in the EU today involves the relationship between the Council, Parliament and the Commission. As the constitutional rules of the EU have changed over time the relative powers of these three institutions have also radically altered. The power of the Commission has drastically shifted, for example, with the change in the voting rules of the Council of Ministers from unanimity to qualified majority. Thus modelling power must take into account the internal workings of these bodies, their respective powers, and how their respective powers reflect upon their internal workings. Some have attempted to measure the respective power of these different bodies, by summing across them the normalised scores of voters within each institution (Brams 1989). This does not make a lot of sense, however, since each game within an institution is separate from those across institutions, unless the game across institutions affects the strategies of players within institutions, which cannot be captured by power indices without making assumptions about preferences. However, once we have made assumptions about preferences we can then generate strategic predictions using spatial analysis. Spatial analysis has been used both to examine strategic considerations within institutions, but to greater effect across institutions. The focus of spatial institutional analysis in the EU context has been to examine the structure of the legislative process and see how different rules governing law-making affect the relative importance of different EU institutions and the degree of agent discretion when lawmaking powers are differentiated and assigned to specialised players. Debate between protagonists largely revolves around the best modelling procedures (one-dimensional versus n-dimensional, complete versus incomplete information) and the correct interpretation of the EU Constitution and thus the powers of the different bodies. Tsebelis (1994, 1996; Tsebelis and Kreppel 1998) sees the European Parliament as a conditional agenda-setter with considerable powers to place items on the legislative agenda. Tsebelis only models the interactions between Parliament and the Council and does not include the Commission. With this simplification Parliament’s conditional agenda-setting power

30

K. Dowding

is greatest when the status quo approaches the Pareto set of the Council. Steunenberg (1994a, 1994b) and Moser (1996) see Parliament as a conditional veto player. The difference is minimal since all agree that both the Council and the Commission accept many amendments of Parliament. Steunenberg and Moser must argue that this is because the preferences of the players change during the process. Tsebelis’ analysis can explain these acceptances without the need to assume preference change. Empirical researchers may be puzzled about the heat of the exchanges between the protagonists on these issues, particularly when one considers that all these models are pared down analogies from the realities of the bargaining situation. The relationship between these institutional actors is not a single game, but a series of games over many issues, and bargaining in one game will affect later bargaining in other games. Parliament can only retain credibility if it uses its veto when confronted with a take-it-or-leave-it proposal. But the European Parliament’s ability to reflect a different set of interests from that of the Council depends on its own ability to achieve a majority on the second reading. If it can then it is unlikely to accept less than it has agreed at the second reading since second reading amendments are already a majority compromise that could be maintained if conciliation fails. Parliament has its own internal Rule 78 after Maastricht in which Parliament’s leadership will automatically propose to reject a Council text if conciliation has broken down (Hix 1999). On the single occasion that the Council tried to reconfirm the common position Parliament rejected it (Meyer-Sahling 1998: 27). The Council thus knows the likely vote in Parliament in the third reading and so is discouraged from using its agenda-setting power and must thus pursue cooperation under codecision. As Meyer-Sahling (1998: 27) reports, his interviewees ‘unanimously pointed out that once both institutions, the Council and the Parliament, have finished the first reading, the legislation is considered politically important and the two have informally committed themselves to reaching agreement’ (see also Corbett 2000). To see the type of compromises reached requires a more detailed examination of each decision than can be given by some generalised model with simple assumptions about policy positions. While each agreement requires separate study, the generalised argument that the Council will always play Hawk – citing the Tsebelis-Garrett argument obtained by backwards induction – misses the repeated nature of a game which allows rational players to reach cooperative outcomes as one of the set of possible outcomes, as predicted by the ‘folk theorems’ of repeated non-cooperative games (MeyerSahling 1998). Furthermore, the Council makes much legislation alone from the initiative of the Commission with Parliament playing no role. One of the lessons of institutional research is how sensitive models are to what appear to be innocuous assumptions. Garrett and Tsebelis (1999) make this point about the power index approach but it is equally applicable to the non-cooperative game theory approach. Where institutional or behavioural details are mistaken models may make inaccurate predictions and policy prescriptions (Crombez 2000; Steunenberg 2000; Corbett 2000). Steunenberg et al. (1999) combine institutional spatial analysis taking preferences into account, with a modified power index that recognises luck. They

Rational choice and institutional change

31

consider a player’s power to be her ability to affect the equilibrium outcome in a game defined by its decision-making procedure: the Commission can propose a policy in their game, but the Council can only veto. They distinguish power from luck by measuring the potential influence of players, independently of players’ preferences. They do this by measuring a player’s power by reference to the mean or expected distance between the equilibrium outcome and the player’s ideal point for all possible combinations of players’ preferences and all possible combinations of the status quo to give them a ‘strategic power index’. From their strategic power index they measure the status quo bias as the distance between the equilibrium outcome and the status quo in their ‘inertia index’. They report that these indices are difficult to manipulate analytically as they are based on the equilibrium outcomes of the underlying games, which have a non-linear dependence on the players’ preferences and the status quo. But they consider numerical examples, using the consultation and cooperation procedures in the EU. They argue that, contrary to earlier power index approaches, the Commission seems to be the dominant actor and, rather than Parliament gaining power under the cooperation procedure, it seems that both the Council of Ministers and the Commission gain when unanimity applies; both lose when majority is applied. Steunenberg applies this approach in Chapter 6 of this book. We can see that the institutional rational choice approach throws up numerous interesting and important debates about the key issues of the relative powers of different actors within the EU. It provides quantifiable analyses with definite predictions. What is missing in this largely theoretical literature is quantified empirical examination. By and large, empirical evidence is anecdotal rather than a systematic examination of the policy positions of different actors in different policy domains, coupled with quantitative analysis of the results as the procedures operate.

The European Parliament There is much less analysis of the party system of the European Parliament from an institutional rational choice perspective than has been done on the Council or the relationship between the major policy-making organs of the EU (where most legislative models can be placed). In part this is due to past lesser importance of the European Parliament in policy-making within the EU and a received view that there was no European party system, rather national party systems that coalesced for practical purposes within Parliament. However, the shift from the study of European integration to European politics (Hix 1996, 1998a, 1998b) in part tracks the growing proto-European party system. Academics are increasingly turning to Parliament as it takes on a more important policy-making role and Parliament is beginning to behave more like parliament with a party system. Hix (1999) has mapped out the EU political space, adapting the Manifesto Research Group of the European Consortium for Political Research (Budge et al. 1987; Klingemann et al. 1994). Hix used every major policy statement adopted by the national and Parliament’s party leaders of the Socialist, Christian Democrat, Liberal and Green party federations between 1976 and 1994 (91 in total). He argues that two dimensions were salient across the period, and explains the party

32

K. Dowding

families converging on basically integrationist positions by the early 1990s, allowing the ‘normal’ politics of left-right to emerge within a proto-party system in Parliament. This spatial analysis of Parliament can also be contrasted with the allocation of power indices to nations and political groups within Parliament, some taking into account the ideological predispositions (Colomer and Hosli 1997). Laver (1997) has also analysed the potential coalitions that might form if Parliament were to take on the executive forming function of most modern legislative bodies under both a modified Grofman (1982) dynamic model of proto-coalition formation and Laver and Shepsle’s (1996) portfolio allocation model. In fact both models produce similar results, with the liberal group holding a pivotal position and the core of the liberal group also proving pivotal if its more radical members were to split off leftwards and rightwards. The spatial analysis of Parliament will again prove the more interesting research programme, but again careful empirical analysis will provide both stronger foundations and tests of such modelling. Hix (1998b) argues however, that truly European-wide elections and a European competitive party system are unlikely to emerge no matter what powers Parliament gains, unless the EU becomes a partial presidential-interlocking system, where something like the current balance of powers remains but the Commission Presidency becomes a directly elected post. He uses a deductive framework and comparative analysis of most-similar-systems-design technique to isolate the institutional structures leading to different types of party system. Of course, this argument is somewhat speculative, but the combination of deductive and empirical analysis is suggestive. The European-wide programme to study Parliament following the June 1999 elections should provide further empirical evidence with theorised institutional rational choice underpinnings.

The Commission The European Commission is the executive of the EU. The national policy elites of each of the member states have an interest in delegating some policy functions and areas to the Commission. Such delegation produces a principal–agent problem, where the principal is multiple with some conflicting interests, and the agent must also satisfy not only the Council but also Parliament, as we have seen. Most of the analyses of the Commission (Dunleavy 1997) show that it has an interest in trying to acquire greater competencies than the Council intends to delegate. Frey (1984) suggests this is simple budget-maximising. However, most of the models of the interactions of the main institutional actors of the EU – the models considered above of the relative powers of the Commission, Parliament and the Council of Ministers – assume that the Commission has preferences that are more integrationist than the other actors. Rather than seeing this as an attempt to maximise budgets, which need not take on an integrationist form, the underlying assumption is simply that the Commission is integrationist because of its historic role. Another possibility is that its preferences are those of the personnel who are appointed. Crombez (1997) analyses the appointments of Commissioners using a

Rational choice and institutional change

33

spatial model arguing that this characterises the set of effective Commissions, viz., ones which can propose their own ideal policies, which can become EU policy. A key element in this model is the role of the Commission President. What this – and the general assumption of the preferences of the Commission vis-à-vis Parliament and especially the Council – does not take into account is the changing political map of Europe, where Commissioners are appointed, the European Parliament elected and national governments elected on different time sequences. Dunleavy (1997) examines the role of the Commission within the broader framework of the EU from the perspective of four competing public choice models. The conclusion of each is a continuing drift of competency to Brussels and the ‘democratic deficit’ becoming a permanent feature of its institutional form. Taking his two micro-level accounts of Commission behaviour – bureaushaping and transaction-costs – Dunleavy suggests that the Commission is already shaped into the configuration that senior bureaucrats are likely to desire. Virtually all directorates-general are in a regulatory or control agency form, with some transfer agencies and contract agencies. Thus they have fewer executive powers, preferring that national bureaucracies and agents carry out these functions; rather, they ensure that their own agencies are centrally located and strategic in orientation. The transaction-costs model is used to explain the delegation of responsibilities of implementation to national bureaucracies, allowing the major failures to be subject to scrutiny and litigation involving the states and the European Court of Justice. Whilst Dunleavy’s argument is plausible he produces little in the way of hard quantitative or qualitative empirical evidence, and his ideas need further clarification to produce testable predictions.

The European Court of Justice Writing on the European Court of Justice from the institutional rational choice viewpoint has generally taken a realist line suggesting governments support an activist Court in order to increase the effectiveness of the incomplete contracts which constitute the basic treaties of the EU. The Court has realised its power is limited by the acquiescence of member states and hesitates to make decisions that will not generally approve of (Cooter and Drexel 1994; Garrett 1992, 1995; Garrett and Weingast 1993). Most of the critique of this position revolves around how strategic are the decisions of the Court, and whether all national governments act as self-interestedly as each other, or whether some are more ‘proEuropean’ (Burley and Mattli 1993; Slaughter-Burley 1993, 1999; Mattli and Slaughter 1995). Garrett et al. (1998) put forward three innocuous hypotheses from a game-theoretic account of strategic reasoning of the Court and member states. The first suggests that the greater the clarity of the EU treaties, case precedent and legal norms, the greater the probability that the Court will rule against a litigant state. Second, the greater the domestic costs of an adverse Court ruling the more likely the state will contest an adverse Court decision. Third, the larger the number of states affected, and the greater the domestic costs of a number of adverse rulings in a given sphere of law, the greater the probability that EU member states will respond to Court decisions by new

34

K. Dowding

secondary legislation. They test the hypotheses against a selected set of most similar cases and find strong support for them. The literature on the European Court has fixated upon legal interpretation rather than the generation of policy. However, ‘third-generation’ institutional analysis is more interested in how institutions shape interests in the creation of policy outcomes, and how they shape other institutions. In other words it is not interested primarily in the generation of more EU law, but what type of laws. The Court should be a powerful institution for policy-making. A court’s discretion varies inversely with the probability of new legislation being introduced to repeal its decisions (Ferejohn and Weingast 1992; Cooter and Ginsburg 1997). The legislative process in the EU – separation of powers, multiple legislative bodies, and some requiring oversized majorities – suggests the Court has great discretion and so may be considered a major policy player. But there is little work examining this from the institutional perspective.

Conclusion Rational choice has revolutionised the study of the EU by bringing formal techniques to an area which previously had been subject mostly to descriptive accounts underpinned by a set of ‘theories’ consisting of classificatory schemes rather than models proposing explanatory mechanisms. Rational choice has allowed a host of new answers to be proposed to old questions, and suggest new questions to be asked. As is often the case when new techniques are proposed for studying a subject much of the debate has revolved around, first, technical issues concerning the best type of models to use to examine given issues; second, around the best models to utilise within type; and third, the applicability of these models to the actual case. More empirical evidence is required to allow us to choose between the models. As the EU continually modifies its institutions preparatory and in response to its widening, new opportunities are opened for political scientists to develop and test models. In this sense the EU has become an experiment, which allows political scientists to study changing institutions as they occur, and be able to ask the questions contemporaneously with the unfolding events. This type of institutional analysis allows us to examine the effects of institutional change, and proposed institutional change. We can see the effects on the voting resources of changing weights and voting rules in the Council. Examine the changing powers of institutional actors as rules determine the agenda-setting powers and number and nature of veto players in the legislative game. It gives a more sophisticated handle on the nature of the ‘democratic deficit’, which hangs over the EU as it expands eastwards. Thorough contemporary institutional analysis of the EU will not only better understand the EU itself as it expands but also all political institutions around the world.

Notes 1 The first part of this section depends heavily on Shepsle and Weingast (1995).

Rational choice and institutional change

35

References Arrow, K. (1951a) Social Choice and Individual Values, New Haven: Yale University Press. Arrow, K. (1951b) ‘An Extension of the Basic Theorems of Classical Welfare Economics’, in Proceedings of the Second Berkeley Symposium, Berkeley: University of California Press, 507–32. Arrow, K. (1963) Social Choice and Individual Values, New Haven: Yale University Press (2nd edition). Austen-Smith, D. and J.S. Banks (1998) ‘Social Choice Theory, Game Theory, and Positive Political Theory’, Annual Reviews of Political Science 1: 259–87. Barry, B. (1991) ‘Is it better to be powerful or lucky?’ in B. Barry (ed.) Essays in Political Theory: Democracy and Power, Oxford: Clarendon Press. Brams, S.J. (1989) ‘Are the two houses of Congress really coequal?’ in B. Grofman and D. Wittman (eds) The Federalist Papers and the New Institutionalism, New York: Agathon Press. Budge, I., D. Robertson and D.J. Hearl (eds) (1987) Ideology, Strategy and Party Change in 19 Democracies, Cambridge: Cambridge University Press. Burley, A.-M. and W. Mattli (1993) ‘Europe before the Court: a political theory of legal integration’, International Organisation 47: 41–76. Coase, R.H. (1937) ‘The nature of the firm’, Economica 4: 386–405. Coase, R.H. (1960) ‘The problem of social costs’, Journal of Law and Economics 3: 1–44. Colomer, J. and M. Hosli (1997) ‘Decision-making in the EU: the power of political parties’, Aussenwirtschaft 52: 255–80. Condorcet, M. (1785 [1994]) Foundations of Social Choice and Political Theory (translation by I. McLean and F. Hewitt), Brookfield, VT: Edward Elgar. Cooter, R. and J. Drexel (1994) ‘The Logic of power in the emerging European constitution’, International Review of Law and Economics 14: 307–26. Cooter, R. and T. Ginsburg (1997) ‘Comparative judicial discretion: an empirical test of economic models’, in D. Schmidtchen and R. Cooter (eds) Constitutional Law and Economics in the European Union, Cheltenham: Edward Elgar. Corbett, R. (2000) ‘Academic modelling of the codecision procedure: a practitioner’s puzzled reaction’, European Union Politics 1: 373–81. Cox, G. and M. McCubbins (1993) Legislative Leviathan: Party Government in the House, Berkeley: University of California Press. Crombez, C. (1997) ‘Policy making and Commission appointment in the EU’, Aussenwirtschaft, 52: 63–82. Crombez, C. (2000) ‘Codecision: towards a bicameral European Union’, European Union Politics 1: 363–8. Dixit, A.K. (1996) The Making of Economic Policy, Cambridge, MA: MIT Press. Dixit, A.K. and J. Londregan (1995) ‘Redistributive politics and economic efficiency’, American Political Science Review 89: 856–66. Dowding, K. (1991) Rational Choice and Political Power, Aldershot: Edward Elgar. Dowding, K. (1995) ‘Interpreting formal coalition theory’, in K. Dowding and D. King (eds) Preferences, Institutions and Rational Choice, Oxford: Clarendon Press, 43–59. Dowding, K. (1996) Power, Buckingham: Open University Press. Dowding, K. (1999) ‘Shaping future luck’, Journal of Conflict Processes and Change 4: 1–12. Dunleavy, P. (1991) Bureaucracy, Democracy and Public Choice, Hemel Hempstead: Harvester Wheatsheaf. Dunleavy, P. (1997) ‘Explaining the centralization of the European Union: a public choice analysis’, Aussenwirtschaft 52: 183–212.

36

K. Dowding

Felsenthal, D.S. and M. Machover (1998) The Measurement of Voting Power, Aldershot: Edward Elgar. Felsenthal, D.S. and M. Machover (2001) ‘Myths and meanings of voting power: comments on a symposium’, Journal of Theoretical Politics 13: 81–97. Felsenthal, D.S. and M. Machover (2002) ‘The Treaty of Nice and Qualified Majority Voting’, Social Choice and Welfare, forthcoming. Ferejohn, J. and M. Fiorina (1975) ‘Purposive models of legislative behavior’, American Economic Review: Papers and Proceedings 65: 407–15. Frerejohn, J. and B. Weingast (1992) ‘A positive theory of statutory interpretation’, International Journal of Law and Economics 12: 263–79. Frey, B. (1984) International Political Economics, Macmillan: Basingstoke. Garrett, G. (1992) ‘International cooperation and institutional choice: the European Community’s Internal Market’, International Organization 46: 533–60. Garrett, G. (1995) ‘The politics of legal integration in the European Union’, International Organization 49: 171–81. Garrett, G., R.D. Keleman and H. Schulz (1998) ‘The European Court of Justice, national governments, and legal integration in the European Union’, International Organization 52: 149–76. Garrett, G., I. McLean and M. Machover (1995) ‘Power, power indices and blocking power: a comment on Johnston’, British Journal of Political Science 25: 563–68. Garrett, G. and G. Tsebelis (1996) ‘An institutional critique of intergovernmentalism’, International Organization 50: 269–99. Garrett, G. and G. Tsebelis (1999) ‘Why resist the temptation to apply power indices to the EU?’, Journal of Theoretical Politics 11: 291–308. Garrett, G. and B. Weingast (1993) ‘Ideas, interests, and institutions: constructing the European Community’s Internal Market’, in J. Goldstein and R. Keohane (eds) Ideas and Foreign Policy, Ithaca, NY: Cornell University Press. Grofman, B. (1982) ‘A dynamic model of protocoalition formation in ideological N-Space’, Behavioural Science 27: 77–90. Herne, K. and H. Nurmi (1993) ‘The distribution of a priori voting power in the EC Council of Ministers and the European Parliament’, Scandinavian Political Studies 16: 269–84. Hix, S. (1996) ‘The European party federations: from transnational party cooperation to nascent European parties’, in J. Gaffney (ed.) Political Parties and the European Community, London: Routledge. Hix, S. (1998a) ‘The study of the European Union II: The ‘New Governance’ Agenda its rival’, Journal of European Public Policy 5: 38–65. Hix, S. (1998b) ‘Elections, parties and institutional design: a comparative perspective on European Union democracy’, West European Politics 21: 19–52. Hix, S. (1999) ‘Dimensions and alignments in European Union politics: cognitive constraints and partisan responses’, European Journal of Political Research 35: 69–106. Horn, M. (1995) The Political Economy of Public Administration, Cambridge: Cambridge University Press. Hosli, M.O. (1996) ‘Coalitions and power: effects of qualified majority voting in the Council of the European Union’, Journal of Common Market Studies 34: 255–73. Hosli, M.O. (1997) ‘Voting strength in the European Parliament: the influence of national and partisan actors’, European Journal of Political Research 31: 351–66. Johnston, R.J. (1995a) ‘The conflict over qualified majority voting in the European Council of Ministers: an analysis of the UK negotiating stance using power indices’, British Journal of Political Science 25: 245–54. Johnston, R.J. (1995b) ‘Can power be reduced to a quantitative index – and if so, which

Rational choice and institutional change

37

one? A response to Garrett, McLean and Machover’, British Journal of Political Science 25: 568–72. Johnston, R.J. and A.J. Hunt (1977) ‘Voting power in the EEC’s Council of Ministers: an essay on method in political geography’, Geoforum 8(1): 1–9. Kirchgässner, G. (1997) ‘Supply and demand factors of centralization: comment to P. Dunleavy’, Aussenwirtschaft 52: 213–20. Kiser, L.L. and E. Ostrom (1982) ‘The three worlds of action: a metatheoretical synthesis of institutional approaches’, in Elinor Ostrom (ed.), Strategies of Political Inquiry, Beverly Hills: Sage, 179–222. Klingemann, H.-D., R.I. Hofferbert and I. Budge (eds) (1994) Parties, Policies and Democracy, Boulder: Westview. Krehbiel, K. (1991) Information and Legislative Organization, Ann Arbor: University of Michigan Press. Krehbiel, K. (1993) ‘Where’s the Party?’, British Journal of Political Science 23: 256–66. Lane, J.-E. (2002) ‘A note on Nice’, Journal of Theoretical Politics, forthcoming. Lane, J.-E. and R. Mæland (1995) ‘Voting Power under the EU Constitution’, Journal of Theoretical Politics 7: 223–30. Lane, J.-E., R. Mæland and S. Berg (1995) ‘The EU Parliament, seats, states and political parties’, Journal of Theoretical Politics 7: 395–400. Laver, M. (1997) ‘Government formation in the European Parliament’, Aussenwirtschaft 52: 223–48. Laver, M. and N. Schofield (1990) Multiparty Government: The Politics of Coalition in Europe, Oxford: Oxford University Press. Laver, M. and K.A. Shepsle (1996) Making and Breaking Governments, Cambridge: Cambridge University Press. McKelvey, R.D. (1976) ‘Intransitivities in multidimensional voting models and some implications for majority control’, Journal of Economic Theory 12: 472–82. McKelvey, R.D. (1979) ‘General conditions for global intransitivities in formal voting models’, Econometrica 47: 1085–111. Mattli, W. and A.-M. Slaughter (1995) ‘Law and politics in the European Union: a reply to Garrett’, International Organization 49: 183–90. Meyer-Sahling, J.-H. (1998) ‘The European Parliament and the co-decision procedure of the EU-legislative process: a reassessment of Tsebelis and Garrett’, unpublished MSc. dissertation, London School of Economics. Mitchell, W.C. and M.C. Munger (1991) ‘The economic theory of interest groups: an introductory review’, American Journal of Political Science 35: 512–46. Moe, T.M. (1990) ‘The politics of structural choice: toward a theory of public bureaucracy’, in O.E. Williamson (ed.) Organization Theory, New York: Oxford University Press. Moser, Peter (1996) ‘The European Parliament as a conditional agenda setter: what are the conditions? A Critique of Tsebelis (1994)’, American Political Science Review 90: 834–8. Niskanen, W.A. (1971) Bureaucracy and Representative Government, Chicago: Aldine-Atherton. Niskanen, W.A. (1991) ‘A reflection on “Bureaucrats and representative government” ’, in A. Blais and S. Dion (eds) The Budget Maximizing Bureaucrat: Appraisals and Evidence, Pittsburgh: Pittsburgh University Press, 13–32. Niskanen, W.A. (1994) ‘A reassessment’, in: W.A. Niskanen (ed.) Bureaucracy and Public Economics, Aldershot: Edward Elgar, 269–84. North, D.C. (1990) ‘A transaction cost theory of politics’, Journal of Theoretical Politics 2: 355–67. Riker, W.H. (1962) The Theory of Political Coalitions, New Haven: Yale University Press. Riker, W.H. (1982) Liberalism Against Populism, San Francisco: Freeman.

38

K. Dowding

Schofield, N.J. (1978) ‘Instability of simple dynamic games’, Review of Economic Studies 45: 575–94. Schofield, N.J. (1983) ‘Generic instability of majority rule’, Review of Economic Studies 50: 695–705. Shepsle, K.A. and B.R. Weingast (1995) ‘Positive theories of Congressional institutions’, in K.A. Shepsle and B.R. Weingast (eds) Positive Theories of Congressional Institutions, Ann Arbor: University of Michigan Press, 5–36. Slaughter-Burley, A.-M. (1993) ‘New directions in legal research on the European Community’, Journal of Common Market Studies 31: 341–400. Slaughter-Burley, A.-M., A. Sweet Stone and J.J. Weiler (eds) (1999) The European Court of Justice and the National Courts, New York: Hart. Steunenberg, B. (1994a) ‘Decision making under different institutional arrangements: legislation by the European Community’, Journal of Institutional and Theoretical Economics 150: 642–69. Steunenberg, B. (1994b) ‘Regulatory policymaking in a parliamentary setting’, in Ph. Herder-Dorneich, K.-E. Schenk and D. Schmidtchen (eds) Jahrbuch für Neue Politischen Ökonomie, Tübingen: Mohr, 36–57. Steunenberg, B. (1996a) ‘Agenda discretion, regulatory policymaking and different institutional arrangements’, Public Choice 86: 309–19. Steunenberg, B. (2000) ‘Seeing what you want to see: the limits of current modelling on the European Union’, European Union Politics 1: 368–73. Steunenberg, B., D. Schmidtchen and C. Koboldt (1999) ‘Strategic power in the European Union: evaluating the distribution of power in policy games’, Journal of Theoretical Politics 11: 343–70. Tsebelis, G. (1994) ‘The power of the European Parliament as a conditional agenda setter’, American Political Science Review 88: 128–42. Tsebelis, G. (1995) ‘Conditional agenda-setting and decision-making inside the European Parliament’, Journal of Legislative Studies 1: 65–93. Tsebelis, G. (1996) ‘More on the European Parliament as a conditional agenda setter: response to Moser’, American Political Science Review 90: 839–44. Tsebelis, G. and G. Garrett (2000) ‘Legislative politics in the European Union’, European Union Politics 1: 9–36. Tsebelis, G. and A. Kreppel (1998) ‘The history of conditional agenda-setting in European institutions’, European Journal of Political Research 33: 41–71. Widgrén, M. (1994) ‘Voting power in the EC and the consequences of two different enlargements’, European Economic Review 38: 1153–70. Williamson, O.E. (1985) Markets and Hierarchies, New York: The Free Press. Williamson, O.E. (1996) The Mechanisms of Government, Oxford: Oxford University Press. Wintrobe, R. (1998) The Political Economy of Dictatorship, Cambridge: Cambridge University Press.

Chapter Title

Part II

Constitutional aspects of enlargement

39

40

Authors name

Chapter Title

3

41

Constitutional issues regarding European Union expansion Dennis C. Mueller1

Introduction The original intent of Jean Monnet and the other founders of the European Union was to join Germany, France and enough other European countries sufficiently closely economically to make certain that they would never go to war against one another again. Thus, the motivation behind the founding of the European Union was to provide a form of public good for all of Europe. In describing the European Union’s raison d’être today, we would expand the list of collective goods that can be provided more efficiently at the European level, and would explicitly include the gains to all Europeans from free trade. The European Union’s raison d’être is to help Europeans achieve the possible gains from a more efficient allocation of goods and services. To achieve the collective benefits from providing public goods to all Europeans, one must determine the types and amounts of public goods Europeans want, and then supply them. There are essentially two types of institutional structures for accomplishing these tasks: a federalist or a confederalist structure. Which structure is best suited to these tasks depends upon the characteristics of the distribution across Europe of citizen preferences for public goods, and thus, of course, also on the characteristics of the public goods themselves. The European Union’s institutional structure has both federalist and confederalist attributes. This mixture creates certain problems in optimally providing public goods within the EU, and in deciding whether or not to admit new members and upon what terms. These issues are the subjects of this contribution.

States as clubs: federalism Optimal club size Imagine people distributed in varying densities across a given land mass living in anarchy. They produce that which they need to survive, supplemented by that which they can obtain through exchange. Certain goods exist – roads, fire protection – that no single person or family can supply optimally. Thus, incentives exist for groups of individuals to agree to provide these goods collectively. In the tradition of contractarianism and the new constitutional political economy we can imagine various groups of individuals reaching separate agreements to

42

D.C. Mueller

provide different public goods. Let us call each union of a group of people to provide one or more public goods a ‘club’. The characteristics of a public good – joint supply and non-exclusion from benefits – often give a geographic dimension to any club formed to provide a particular good or service. The fire station benefits only those within a certain radius that covers the distance that its trucks can reach in a reasonable amount of time; the dam protects only those who live lower down along the river. This property of public goods leads us to expect that many clubs are formed to provide public goods with a geographic dimension to them.2 The natural selfinterest of individuals should thus lead them to create and join many clubs of different geographic dimensions to provide the many possible public goods that could make them better off.3 The essence of a club is that it is a voluntary association of individuals to achieve some mutually beneficial goal. As such, the formation of clubs to provide public goods is much in the spirit of Knut Wicksell’s (1896) ‘voluntary exchange’ approach to government, and of Frey and Eichenberger’s FOCJs (1996). Consistent with these approaches, one might simply define a state as a club formed to provide one or more public goods. Given the many sorts of public goods that individuals might wish to consume and the different geographic dimensions of their benefits, one might well expect that a given individual would be a citizen of several club/states, some of which would provide public services to individuals in a small geographic area, others to larger numbers of individuals spread across much larger areas. Although there is nothing that we have said so far that would imply that a series of voluntarily joined clubs would take on the hierarchical structure of a federalist state, these clubs would resemble the optimal federalist state, as described in the economics literature on federalism, in that governmental functions would be associated with different levels of government in accordance with the geographic spillovers from the various public goods, and the relationship between the benefits from membership in a club and the size of the club. The public good benefits from club membership may be of two, fundamentally different types: (1) benefits that grow without bound as club size expands, or (2) benefits that eventually decline with club size, and may even turn negative on the margin beyond some size. Under the first heading we can place the benefits from free trade to members of a customs union, or the net benefits to a club member from providing a public good when there is no crowding. In the latter case, if the benefits to a member do not fall as the club size increases, his net benefits from membership increase as the club size grows and the fixed costs of providing the public good are spread over more and more dues/tax payers.4 For many public goods, like transportation, environmental and recreation systems, adding consumers leads to congestion costs beyond some club size. When crowding occurs, an optimal club size exists at which the benefits to existing members from spreading the costs of providing the public good over one more member just equal the additional congestion costs that this new member brings with her.5

Constitutional issues of EU expansion

43

The costs of heterogeneity under federalism Although all citizens can in principle benefit from the provision of a pure public good, all do not necessarily have the same preferences for it. Institutions must be established to aggregate the diverse preferences of the citizens in a given polity to determine the optimal quantities and characteristics of the public goods offered by that polity. In all European countries except for Great Britain these institutions take the form of a proportional representation electoral system to elect members of parliament, and a simple majority rule within the parliament to aggregate the preferences represented there. This method of aggregating preferences imposes certain costs on some or even all citizens, when citizen preferences differ from one another. To see what is involved, assume that a polity provides a single public good, say roads, and its citizens must decide the amount, x, to be spent on roads. The population can be divided into five groups, each of which favours a different expenditure. Let x1 be the expenditure favoured by group 1, x2 the expenditure favoured by group 2, and so on. Assume that x1 < x2 < x3 < x4 < x5. An election is held and the parties representing groups 2 and 3 together win enough seats to form the government and decide the amount of x for the community. It is reasonable to assume that the expenditure on roads that this coalition chooses, x* falls between x2 and x3, i.e., x2 ≤ x* ≤ x3. Under this assumption, x* must differ from the most preferred quantity of x of one of these two groups, and most likely differs from the quantities most preferred by both groups. Thus, the need to form a coalition to be able to decide the quantity of the public good imposes costs on each group in comparison to what it would experience, if it could decide the quantity of the public good alone.6 Similar losses obviously also exist for members of the other three groups, and these losses are likely to be even greater than those experienced by the groups whose representatives were able to form the government. Thus, the necessity of all members of a community having to consume the same quantity of any public good it provides imposes some loss of utility on virtually all members, as opposed to what would occur if all members had the same tastes. Two consequences follow: first, we can now identify a second possible cost to the existing members of a community from expanding its membership. Even when no crowding takes place, the existing members may be made worse off from the admittance of new members, if the new members have different preferences from old members, and the increased heterogeneity of the community shifts the chosen quantities of public goods supplied away from those favoured by the old members. Second, gains may exist for all individuals from free migration across communities, if this migration can result in communities with more homogeneous preferences for public goods.7 A possible advantage of having a federalist governmental structure with numerous communities within the federation providing similar bundles of public goods is that it would allow citizens the opportunity to improve the levels of utility that they obtain from publicly provided goods and services by re-sorting themselves into communities of more homogeneous preferences.

44

D.C. Mueller

Summary Rational, self-interested individuals can be expected to form and join clubs and communities, which provide public goods. The characteristics of public goods are likely to make the formation of multiple clubs and polities optimal, so as to better match the beneficiaries of publicly provided goods to those who decide upon their characteristics and pay for them. When a club or a community provides a pure public good to its members, all members must consume the same public good. If members differ over the quantity or quality characteristics of the good that should be supplied, then these characteristics will deviate from those considered optimal by some, perhaps all, members. A second possible advantage of a federalist governmental structure is that it facilitates migration to achieve better matches between the preferences of citizens for public goods, and the characteristics of the public goods provided.

Confederations as clubs My American Heritage Dictionary defines a ‘confederation’ as a group of states united for a common purpose. One might well also define a ‘club’ as a group of individuals united for a common purpose. A confederation is thus a kind of club of nations,8 and we expect it to share the following two properties of clubs: (1) membership is voluntary, and (2) members voluntarily join because they expect to be made better off. The ‘common purpose’ that brings them together involves their obtaining mutual benefits. The benefits from membership in the club or confederation are a form of a public good to club members. When the members of a club are autonomous states, the issue arises of whether it is the interests of the citizens of these states that are being advanced by the confederation, or the interests of the members of the governments that represent these citizens. The simplest case to consider, and the one which is most compatible with the confederate form of government being the optimal way to advance the interests of the individual citizens within it, is to assume that every person within a given state has identical preferences for the public goods that the confederation provides, and that the representatives of the governments of each member state faithfully represent the preferences of their citizens. Under these assumptions, a club of states has all of the properties of a club of individuals. Namely, optimal club size is infinite in the absence of crowding or preference heterogeneity across club members, while optimal club size is likely to be finite when either of these two phenomena are present. When every person within a given state has identical preferences for the public goods provided by the confederation, an obvious and simple way to represent these preferences is to have each constituent state government send representatives to an assembly, and for these state delegations to have seats or voting rights in this assembly proportionate to their populations. Such a system of representation would produce an assembly of representatives that would resemble that which one would expect to arise under a system of proportional representation in which the citizens of the confederation directly elected representatives to the assembly.

Constitutional issues of EU expansion

45

The situation can be quite different, however, if the preferences of the citizens within each state are heterogeneous. Assume, for example, that the population of each state can be divided into five groups as in the previous section, with the five groups in each of the states having the same five most preferred quantities of the public good x, x1 < x2 < x3 < x4 < x5. Assume further that there are ten states in the confederation. In five of these, parties representing the groups favouring x1 and x2 form governments with narrow majorities, in the other five states parties favouring x3 and x4 form governments with substantial majorities. If the populations of the five states favouring x1 and x2 give them a slight majority in the confederation’s assembly, and this assembly uses the simple majority rule, then the quantity of the public good chosen using the confederate form of representation will lie between x1 and x2, where a quantity between x3 and x4 would have been chosen using the simply majority rule under a system of proportional representation in which the citizens elected their representatives directly. With heterogeneous preferences within states, the outcomes from a confederate form of representation can differ from those under proportional representation when citizens elect representatives directly, even if the unanimity rule is used in the confederate assembly. In the above example, individuals who favour the quantity x5 are not represented in any of the ten governments, and thus their preferences would presumably receive no weight in any compromise over x that representatives of the ten governments were able to reach under the unanimity rule. The benefits and costs of membership in a newly formed confederation Much of the public choice literature that deals with constitutions, and essentially all of the recent constitutional political economy literature, assumes that constitutions are written by the citizens of a country and are unanimously agreed to by these citizens.9 Although these assumptions are highly unrealistic as descriptions of the way actual constitutions of nation states are written, they do accurately describe how confederations come into being. Since the potential members of a new confederation are themselves nation states, it is possible for each of them to remain in the current ‘state of anarchy’ should they not find the conditions for joining the confederation to their advantage. Unanimity is not merely a hypothetical ideal when it comes to the formation of a confederation, it is a practical necessity. Although all countries that meet to discuss forming a confederation may not choose to join it, the agreement among those which do join must be unanimous. The contract all joiners sign must be seen as a Pareto move by all signatories.10 Both the contractarian perspective on the initial constitution and the assumption of unanimous agreement among all signatories is fully justified when the polity created by the constitution is a confederation of previously autonomous states. Let us assume now that ten nations meet to consider forming a confederation to provide some set of public goods to its members. The ten countries are identical in size and preferences for the public goods, and thus there are no obvious, natural coalition partners. Let us further assume that the properties of the public

46

D.C. Mueller

goods and the potential members’ preferences are such that the optimal size of the confederation, is seven. Obviously, an effort may be made to form a confederation of seven, and leave three countries out. But which three? If all potential members are identical no obvious way may exist to determine which countries are in the lucky seven, and which are the unlucky three. The deliberations may well result in all ten countries entering the confederation, with the benefits from membership falling below their maximum potential value. The benefits and costs of adding members to an existing confederation In joining a confederation, a nation risks losing some of its autonomy. The confederation may reach decisions that make the citizens of one of the member states worse off. This possibility can, of course, be avoided by requiring that all collective decisions of the confederation be made using the unanimity rule. The decision-making costs accompanying the unanimity rule may preclude its use for all collective decisions of the confederation, but one assumes that it will be retained for some decisions that are particularly likely to pose large potential losses for the individual members. One such collective decision would be whether or not to admit a new member. We shall assume, therefore, that the admission of new members to an existing confederation requires the unanimous approval of current members. If an existing confederation cannot compel other countries to enter by, say, threatening military action, then the unanimous agreement required to found any confederation will also be needed to expand it. Let us now assume again that the optimal size of a confederation is seven nations, but that a confederation of this size already exists. Three countries, identical to the seven already in the confederation, wish to join. The situation is now quite different from that described in the previous section. The membership of the optimal coalition of seven is obvious: the existing confederation’s members. The status quo constitutes a Schelling point for a coalition that creates one confederation of optimal size, and one of suboptimal size. If the three potential entrants are to succeed in joining the other seven countries in a confederation of ten, they are going to have to bribe the seven to let them in. The bribes must be large enough to leave the original seven members no worse off than before the confederation expands. If it is in the interest of the three potential entrants to offer such bribes, then the total benefits from a confederation of ten must exceed those arising from two clubs of sizes seven and three. The rational self-interest of both sets of countries produces the Pareto-optimal outcome, as in the previous example; only the distribution of benefits among the confederation’s members differs.11

Implications for European Union expansion The existing European confederation of fifteen nations is undoubtedly too small to supply some public goods optimally, and too large for others. The optimal membership of a confederation to address some environmental problems, like global warming, is perhaps all nations of the world. The existing confederation, however, is too large to address some more localised environmental problems, like pollution in the Rhine and Danube rivers.

Constitutional issues of EU expansion

47

If we assume for the moment that the optimal institutional structure for the European Union is a confederation, then the preferences with respect to EU policies of all citizens within each of its fifteen nations should be identical, with all preference heterogeneities existing across the fifteen members. One example of a public good that might fit this characterisation to some degree would be environmental protection, on the assumption that environmental protection is a luxury good. The richer countries in the existing confederation can then be expected to favour high environmental standards, its poorer members lower ones. European Union policies, regarding automobile and truck emissions, for example, should then be a compromise between the different policies favoured by each member country. On these assumptions we can also expect significant differences between the policies adopted by the existing members and those that the potential entrants would want to see enacted. Those heading the list of entrants into the European Union – Cyprus, the Czech Republic, Hungary, Poland and Slovenia – are much poorer on average than the existing confederation’s members. Current EU environmental standards regarding automobile and truck emissions, and the like, are undoubtedly more stringent than those that this list of entrants would like to see enforced. Thus, although expansion of the European Union is likely to provide existing Union members benefits in areas where the Union is still too small, it is also likely to impose costs on the existing members in other areas due to the increased size and heterogeneity of the confederation. If on balance the existing members of the European Union will benefit from expansion, then expansion will be a Pareto move for them, and there is no need for the entering countries to compensate or bribe the existing members. Given the heterogeneity of the existing membership, it is possible that expansion will harm some members and benefit others. If the net effect on the current fifteen members is positive, then there will again be no need for the new entrants to compensate existing members as a group to make their entry a Pareto move. Compensation can be from existing members who gain from expansion to those who lose. If, of course, the net effect on the current fifteen members of the European Union from expansion were negative, then the entrants would need to compensate the existing members to make entry a Pareto move. Such compensation would be from relatively poor countries to relatively rich ones, and thus would run counter to normal notions of equity. However, if the entrance of additional countries into the Union is to obtain unanimous support, such compensation would have to take place, and the entering countries would still be better off following entry than they were before, despite having to compensate the existing members. The biggest differences in preferences are probably between the richest current EU members – like Luxembourg and Germany – and the new entrants. Thus, to the extent that the costs of admitting Cyprus and the former Communist countries are due to the heterogeneity between existing and new members, it will be the richest current EU members who will lose the most from entry. If the existing members experience net gains from an enlargement, but these gains are

48

D.C. Mueller

unevenly distributed, it seems likely that the poorer current members would need to compensate the richer ones. Despite the obvious ethical and political difficulties with this solution, it could be necessary in order to obtain unanimous support for admitting countries that differ significantly from current members. It is a logical possibility that the potential entrants would be made worse off by joining the European Union, but that the existing members would be made better off, and thus that the EU would have to bribe candidate countries to induce them to enter. Such bribes were given to Portugal, Spain and Greece at the time of their entry. Although this is a logical possibility, the need for compensation from existing to new members to avoid the latter being made worse off seems unlikely. If the rationale behind the European Union is to create political institutions to better supply European-wide public goods, then the benefits from membership must be much greater for the potential new members than for the existing ones. No transfers from existing to new members are likely to be needed to compensate the new members for the losses they experience from membership.

Representation and citizenship in the European Union The European Union has evolved during the last fifty years through the conclusion of several treaties among the governments of European countries. Its evolution resembles that which one would expect for a confederation of autonomous states, and it seems most accurate to characterise the European Union as a confederation rather than as a federation. Consistent with this interpretation, we observe that the locus of political power in the European Union has been the Council of Ministers located in Brussels. Each member country is represented on the Council, with some collective decisions being made using the unanimity rule, and some made using a qualified majority rule, with larger countries receiving more votes than smaller ones. As discussed above, such a system of representation can accurately transmit information on individual citizen preferences only if all citizens within a country have the same preferences regarding European Union policies.12 Since each country’s representatives in the Council of Ministers are chosen by the country’s elected government, representation in the Council reflects the preferences of the dominant parties and coalitions of the member countries. Only when all citizens in a given country have identical preferences with respect to EU policies will the preferences of all citizens and the ruling majority be concurrent. If people with identical preferences on European Union policies reside in different member countries, then the mode of representation in the Council of Ministers does not adequately represent the preferences of different groups of European citizens. Those who favour a strong foreign policy for the European Union will go unrepresented in the Council, if the elected government in the country in which they have citizenship does not favour a strong foreign policy for the European Union. When individual preferences for EU policies differ within countries, but can be similar across countries, the optimal way to represent these preferences is to elect parties or persons to a European Parliament in proportion to their support across all of Europe. This is, of course, not the mode of repre-

Constitutional issues of EU expansion

49

sentation in the European Parliament today, where representatives are selected from within the individual member countries. Because all countries except the United Kingdom employ some sort of proportional representation formula, the differences in preferences among EU citizens for EU policies are likely to be better reflected in the European Parliament than they are in the Council of Ministers. Given the relative strengths of the Council and the Parliament in determining EU policies, one should not expect these policies to correspond very closely to those that would best reflect the preferences of all Europeans, if the preferences of Europeans toward these policies are not homogeneous within each country. Let us assume for the moment that the European Union has chosen the correct mode of representation given the kinds of public goods it provides, and thus that all citizens residing in a given member country have identical preferences on EU policies, but their preferences are likely to differ from those of citizens in other member countries.13 The existence of cross-national heterogeneities of preferences implies that all Europeans suffer from some of the costs of having to compromise on the characteristics of the public goods provided, as described above. It follows that these costs are likely to increase following the entry of new members that differ as widely from the existing members as, say, Germany and Poland differ. It further follows that the entry of these countries into the European Union, and the ones that will follow, must certainly impose some costs in terms of shifting the packages of public goods bundles offered away from those favoured by current members, under the assumption that the European Union’s current mode of representation has been chosen correctly. It is still possible that there are other benefits for current members from European Union expansion that outweigh these costs. These additional benefits must be greater than the costs engendered by the preference heterogeneities, however, to allow the new entrants to join the European Union without having to compensate the existing members. The existence of preference heterogeneities between residents of the new entrants and existing European Union members has an additional important implication. If the citizens of one country have preferences for public goods provided by the European Union that are relatively homogeneous as compared with the preferences for these public goods of citizens in other countries, it is also quite likely that their preferences for the public goods provided by their own country are more homogeneous than would be true of a cross-section of individuals from different European countries. Thus, membership in the European Union will entail additional costs due to individual preference heterogeneities, if the citizens of one member country have the right to take up citizenship and vote in another. This issue is explored further in the following section.

Why should there be a European Union? What form should it take? Allocative efficiency or redistribution? As Knut Wicksell (1896) first emphasised, since all citizens can benefit from a more efficient allocation of resources, the appropriate voting rule to employ when making collective decisions to improve allocative efficiency is the unanimity rule.

50

D.C. Mueller

Until fairly recently, the unanimity rule was used exclusively in the Council of Ministers to make collective decisions. Since the European Council, which contains all heads of state of EU members, must approve any expansion of membership, the enlargement process still effectively requires the unanimous agreement of all existing members’ governments. If the purpose behind the European Union is to make collective decisions that benefit all member countries, and there are collective goods that the European Union provides that do make all members better off, then existing member countries should welcome new members, and European countries outside of the Union should be eager to join. There should be no need to bribe countries outside of the Union to join. Yet such bribes were extended to Spain, Portugal, Greece and Ireland when they joined. Why? The most obvious explanation, of course, is that Spain, Portugal, Greece and Ireland were when they joined much poorer than the existing member countries. This fact in itself, however, offers neither a positive nor a normative explanation for the creation of transfer programmes to these entrants if the purpose of the Union is to provide community-wide public goods. If there are gains to being a member of a large confederation, then a single new entrant can expect to gain much more from its entry than the set of existing members. Add to this the welldocumented ‘catch-up’ phenomenon, and one now expects new members following their entry to thrive relative to the incumbent countries, as to a considerable degree they have. The presumption that the European Union’s raison d’être is to help Europeans achieve the possible gains from a more efficient allocation of goods and services does not receive much support when one examines the distribution of the Union’s budget between the provision of traditionally defined public goods and redistribution activities. Table 3.1 breaks the EU budgets in 1985 and 1995 down into various large categories. Outlays that were purely or largely redistributional made up almost 85 per cent of the EU budget in 1985 and almost 80 per cent in 1995. Activities that could be fairly clearly identified as having salient public good

Table 3.1 Distribution of European Union expenditures by budget category, 1985 and 1995 (in percentages) 1985

1995

72.9 5.9 5.7

53.6 13.6 11.9

Allocative efficiency • research, energy, transport • external policies

2.6

5.0 6.2

• •

4.6 4.4

5.1 4.5

Redistribution • agriculture and fisheries • regional policy • social policy

administrative costs miscellaneous

Source: Goodman (1996: 101, 105–6).

Constitutional issues of EU expansion

51

properties accounted for only 2.6 per cent of the EU budget in 1985. Even if we categorise the EU’s aid to non-EU countries (External Policies) as a ‘Pareto-optimal redistribution’, and thus a form of allocative efficiency activity, outlays to improve allocative efficiency accounted for only 11.2 per cent of the EU budget in 1995. Because the Council of Ministers has hitherto made the vast majority of its collective decisions using the unanimity rule, one must infer that the transfers to agriculture, poor regions, etc. have also been a form of Pareto-optimal redistribution.14 If we make the further assumption that a confederate governmental structure is the optimal way to reveal the preferences of Europeans for EU policies, then all citizens in a given country have homogeneous preferences for EU policies, and the unanimity achieved in Brussels implies that all Europeans favour large subsidies to European farmers. Every European taxpayer has been willing to fund these large redistributive programmes, because she gets positive utility from seeing her fellow Europeans working on farms, or living in rural areas, have higher incomes – indeed, in the case of European farmers, even higher than the average European taxpayer.15 If one assumes that past EU redistribution policies can be explained in this way, then one must wonder whether the average European taxpayer will get as much utility out of seeing the farmers of Poland and Hungary better off, as one must assume that she has got out of seeing French and German farmers better off? This question becomes even more salient if one pursues the logic of Paretooptimal redistribution to its conclusion. This explanation for government transfers sees the wealthier taxpayer as having the poorer transfer recipient’s welfare in the taxpayer’s utility function. Following this logic, the taxpayer gets more utility out of transfers to a person with one fourth of her income than she does to someone earning one half of her income (von Furstenberg and Mueller 1971). Since most potential entrants have much lower average incomes than even the poorest current members of the EU, the logic of Pareto-optimal redistribution would require that virtually all programmes transferring income to the poor or to poorer regions in current EU countries be abandoned in favour of similar – indeed more generous – programmes targeting individuals and regions in the new entrants. Confederation or federation? An alternative way to explain the size of redistribution programs in the European Union is to assume that the conditions that warrant a confederate governmental structure are not met. Each country’s representatives in Brussels reflect only the preferences of the voters that support that country’s government. Farm interests have not had to convince all Europeans that they deserve large subsidies, only a majority in each country. Indeed, they most likely have not even had to do that, but rather have had to join forces with other groups representing narrow interests in each country to form majority coalitions. If in fact preferences toward EU policies like agricultural subsidies are not homogeneous within each country, a federalist structure for representing these preferences should result in a better match between the policies enacted at the

52

D.C. Mueller

EU level, and those desired by its citizens. It is inconceivable that agricultural interests could exert as much influence in the European Parliament as they do in the Council of Ministers or the European Council, if the Parliament were elected under a PR formula at the European level – even assuming that the Parliament uses the simple majority rule. With redistribution a much smaller fraction of the EU budget than it now is, the potential budgetary costs to existing EU countries from admitting additional members would be greatly reduced, and the question of entry would be reduced to that of the potential for mutually beneficial outcomes, as discussed above. Thus, the mismatch between the existing structure for revealing the preferences of European citizens for EU-level goods and that which would be optimal results in both a serious distortion of existing policies and a serious hindrance to the admission of new member countries. Citizenship and mobility The issue of whether the European Union is or should be organised as a federation or a confederation also arises with respect to the rights of citizens within the EU, and in particular to their rights to employment and citizenship in other member countries. As discussed above, mobility across lower levels of government within a federation can increase the welfare of all citizens, if it results in more homogeneous preferences for the public goods provided at the local level, and there are no offsetting externalities from migration. Thus, a policy of granting the citizens of each EU country equal rights to employment, and eventually to citizenship in the other EU countries can be defended, if the conditions exist that make a federalist structure optimal for Europe. If we assume, on the other hand, that the optimal structure for the European Union is that of a confederation, and thus that Europeans’ preferences are homogeneous within countries but heterogeneous across them, then it is highly unlikely that the optimal EU policy is to provide its citizens broad rights across all Union countries. If the reason why a Dane has different preferences regarding EU policies from those of an Italian is because he is Danish, i.e., has grown up in Denmark and has absorbed the Danish culture, and not simply because he currently resides in Denmark, then there are likely to be costs imposed on the Dane by an Italian’s moving to Denmark and participating in the political process there. Free migration is likely to increase preference heterogeneity in each member country and thereby worsen the collective decisions made in each country. Conversely, if migration does not have these negative consequences – if two people living side by side in Denmark are as likely to have similar preferences with respect to the public goods provided in Denmark, when one is Danish and the other Italian, as when both are Danish – then a confederation is not the ideal form of government for Europe (Mueller 1996: Chapter 21; 1997b). This issue becomes particularly important with respect to the question of the entry of new members, because the countries in the queue to enter the EU are definitely homogeneous with respect to one important determinant of ‘tastes’ and heterogeneous with respect to existing members – namely income. If citizens in new member countries have immediate and equal claims to jobs in existing

Constitutional issues of EU expansion

53

member countries as their own citizens do, then the substantial differences in wages that exist between the new entrants and the existing members could in itself become a major cause of migration. When deciding whether to migrate to a new community, the potential migrant compares the utility level he can enjoy in her new community to the utility she experiences in her present community. If newcomers can secure the same living standards as those already present in the community, migration will continue until the average utility levels in both communities are equal. So long as migration causes no negative externalities, however, there is no reason to be concerned about it. Indeed, in the absence of negative externalities, migration leads to a rise in the utilities of the migrants without any fall in utilities to members of the recipient countries, and thus results in Pareto improvements. There are, however, reasons to expect that people living in countries experiencing inward migration will be harmed by it. We have discussed two sorts of negative externalities from migration already. First, migration can lead to population growth and reductions in welfare from crowding and congestion. Second, if the arrivees are allowed to participate in the political process, and their preferences for collective goods differ from those of the indigenous population, the outcomes from the political process may be altered to the disadvantage of the indigenous population. A third fear often expressed concerning migration from poorer countries is that the migrants compete for jobs with the indigenous population, and bring about higher unemployment and/or lower wages. The most obvious ways to avoid the second and third sources of welfare loss are to deny new arrivees the right to vote and an equal right to employment. Such measures would directly contradict what have until now been basic tenets of European integration policy, namely that all European Union citizens have equal rights to employment and, in time, to citizenship in other EU countries. But such restrictions are fully consistent with a confederate form of government, where individuals are seen as similar within countries, but different across them, and primary citizenship is defined at the level of the member country, not at the level of the confederacy. Equal rights to jobs and citizenship are difficult to defend, even within a federalist system, however, if migration has negative consequences for the indigenous population. The standard approach to any negative externality problem is to levy a Pigouvian tax on the source of the externality.16 In the present case, this would entail taxing immigrants in some way so as to discourage their migration. Subsidies are substitutes for taxes, of course, and so one could in principle achieve the same end by subsidising immigrants to induce them to remain in their homelands. An alternative to the Pareto-optimal redistribution explanation for the European Union’s programme of interregional transfers might be that these transfers are in the narrower interests of citizens in wealthier areas, because they prevent the reductions in the welfare of these citizens that would otherwise occur, if people from the poorer areas were to migrate in great numbers to the richer areas. The same arguments could be used to explain the large subsides offered to Greece, Spain, Portugal and Ireland upon their entry.

54

D.C. Mueller

Such interregional transfers from existing Union members to poorer new entrants are an extremely inefficient form of subsidy, however, if the goal is to discourage migration. The optimal tax or subsidy to achieve this objective would target potential migrants. The difficulty in administering such a tax is to identify who the potential migrants are. A tax levied on actual migrants solves this identification problem by waiting until the potential migrants identify themselves. Offering subsidies to would-be migrants before they identify themselves is obviously far less likely to hit the correct targets exactly and are thus far more expensive. Actual EU transfer policies fall even further from their mark, however. Transfers from the richer countries to regions and industries in new entrants will at best slow migration. Suppose, for example, that the average income in an existing Union member is 100, and in a new entrant is only 50. Income is growing at a 3 per cent rate in the new entrant and at a 2 per cent rate in the existing member. Transfers from the existing Union member to the new entrant raise the growth rate of the new entrant to 5 per cent, while leaving the growth of the existing member unchanged. With a 3 per cent growth rate it will take the new entrant over seventy years before its income reaches that of an existing Union member; at a 5 per cent growth rate the new entrant catches the existing member in around twenty-four years. Although people in the new entrant have less of an incentive to migrate if their country grows at 5 per cent than if it grows at 3 per cent, they still have some incentive for more than twenty years. Transfers from wealthier to poorer EU member countries to stimulate the poorer countries’ growth are likely to be a far less effective way to discourage migration than imposing negative incentives directly on the migrants themselves.

Conclusions We conclude that there is little reason why existing Union countries should offer subsidies of any kind to the countries in the queue to join it, and more generally for any other form of inter-country transfers to discourage cross-national migration. If cross-national migration has adverse effects on the countries receiving migrants, then these countries should, of course, be allowed to take actions to discourage this migration. Such actions would undoubtedly run counter to existing EU policies that provide citizens of EU countries broad rights across all EU countries. Although granting citizens of EU countries maximum freedom of movement within the European Union would obviously seem to be welfare enhancing in the absence of negative externalities from migration, in the presence of such externalities, broad rights to encourage mobility may not be optimal. The occasion of the possible entry of several countries which are much poorer than the average EU member presents an opportunity for, and indeed to some extent requires, a rethinking of the extent to which it is in the interest of all EU citizens that each one of them possess broad citizen rights across the several countries. The issue of how citizenship in the European Union should be defined and what rights should go with it is related to the more fundamental issue of whether the European Union should be organised as a confederation or as a federation of states. The current structure of the European Union is essentially that of a

Constitutional issues of EU expansion

55

confederation of autonomous states. The locus of political power in the Union has been the Council of Ministers and increasingly on major issues like enlargement the European Council where the heads of state meet. In both bodies EU citizens are only indirectly represented through their national governments. As noted above, such a structure of decision-making and modes of representation can be optimal, if all citizens within a country have the same preferences regarding European Union policies, and citizen preferences towards these policies differ across countries. Common citizenship rights for all Europeans makes eminent sense, on the other hand, if the distribution of preferences for public goods provided at the level of the European Union is such as to make the optimal governmental structure that of a federation. Indeed, if citizen preferences for public goods at all levels of government can reasonably be assumed to be randomly distributed among Europeans at birth, then the migration of Europeans may actually have positive externalities by sorting individuals into communities of more homogeneous tastes, and common citizenship rights to encourage mobility can be strongly defended. The European Union is often described, favourably, as a unique and original combination of federalist and confederalist principles. An elephant with wings might also be described as a unique and original combination, but whether one judged it favourably or not would probably depend upon how well it could fly. Up to now, the European Union has been able to fly reasonably well despite the inadequacies and logical inconsistencies in its constitutional structure. These structural deficiencies are likely to prove more harmful to the European Union’s ability to make collective decisions and to provide collective goods to its citizens as it grows in size. The occasion of the entry of several new members provides an ideal opportunity to reconsider what this structure should be.

Notes 1 The research on this contribution has received financial support from the Hochschuljubiläumsstiftung der Stadt Wien. 2 Exceptions exist, of course. One category would be that of ‘cultural clubs’, like religions. It may be possible to join a religious club and receive the benefits from membership regardless of where one lives. See Mueller (1996: Chapter 20). 3 Schmidt-Trenz and Schmidtchen (in Chapter 4 of this volume) emphasise the likelihood of different-sized protective states being optimal for different laws. 4 If B is the benefit to each member of a customs union from free trade, and n is the number of members of the union, then it is reasonable to assume that B initially grows rapidly as membership expands, but eventually its growth, although still positive, tapers off as the bulk of the gains from free trade have been realised. This relationship can be captured by positing an exponential relationship between B and n. B = α + β ln n (3.1) A similar relationship between B and n arises for a pure public good, when there is no crowding. If α is the utility a club member experiences from consuming the pure public good, and β is the public good’s total costs, then the benefits an individual obtains from club membership, net of his share of the public goods costs, are B = α-β/n2 (3.2) Both Equations (3.1) and (3.2) imply an optimal club size of infinity.

56

D.C. Mueller

5 Congestion costs can be introduced into Equation (3.2) by adding a term as given in Equation (3.3). (3.3) B = α-β/n -γ n2 Equation (3.3) implies an optimal size for club membership, n* = (β/2γ )1/3. 6 Let the members of group i have identical preferences for spending on roads, Ui = Kibi(xi-x*)2, where xi is the amount each member of this group wishes to see spent on roads, and bi is an intensity weight specific to group i. If xi = x*, a member of the group experiences a utility level Ki. The loss L to a member from having to consume x* rather than xi, increases linearly with its intensity weight, and exponentially with the distance between xi-x*, L = –bi(xi–x*)2. 7 See Tiebout (1956), Mueller (1989: Chapter 9), and Inman and Rubinfeld (1997). The latter two references also discuss the kinds of externalities from migration that negate the potential benefits from ‘voting with the feet’. 8 Indeed the European Community was often referred to in its early years as the Club of Rome. 9 This is of course the approach pioneered by Buchanan and Tullock (1962). 10 It is, of course, possible that countries that choose not to or are not allowed to join are adversely affected by the formation of a confederation. Hence, the emphasis in the text. 11 Two confederations with six and four or five members each might dominate a single confederation of ten. Here too, however, it is the three outsiders that will have to offer the bribes. 12 See also Mueller (1996: Chapter 21; 1997b). 13 The fact that the Council of Ministers has felt it necessary to adopt a qualified majority rule for some collective decisions strongly suggests that cross-national heterogeneities of preferences exist. If the preferences of citizens were homogeneous across countries, then the Council could simply use the unanimity rule. Even the threat of ‘hold-up’ under the unanimity rule would be weak in this case, since the representatives of each country would know that the other countries had the same preferences as they did, and would call the bluff of any country that tried to hold out for more favourable treatment. 14 Although the Council of Ministers has made its collective decisions using the unanimity rule, this of course does not imply that Europeans have unanimously endorsed these policies. Since the country delegates to the Council represent the governments of their countries, they do not represent the citizens who oppose their governments’ policies. 15 See Koester and Tangermann (1990). 16 A Pigouvian tax equals the marginal cost imposed on the rest of the community by the activity causing the externality.

References Buchanan, J.M. and G. Tullock (1962) The Calculus of Consent, Ann Arbor: University of Michigan Press. Frey, B. and R. Eichenberger (1996) ‘FOCJ: competitive governments for Europe’, International Review of Law and Economies 16: 315–27. Furstenberg, G.M. von, and D.C. Mueller (1971) ‘The Pareto optimal approach to income redistribution: a fiscal application’, American Economic Review 61: 628–37. Goodman, S.F. (1996) The European Union, London: Macmillan (3rd edition). Inman, R.P. and D.L. Rubinfeld (1997) ‘The political economy of federalism’, in D.C. Mueller (ed.) Perspectives on Public Choice, Cambridge: Cambridge University Press, 73–105. Koester, U. and S. Tangermann (1990) ‘The European Community’, in F.H. Sanderson (ed.) Agricultural Protectionism in the Industrial World, Washington, D.C.: Resources for the Future, 64–111.

Constitutional issues of EU expansion

57

Mueller, D.C. (1989) Public Choice II, Cambridge: Cambridge University Press. Mueller, D.C. (1996) Constitutional Democracy, New York and Oxford: Oxford University Press. Mueller, D.C. (ed.) (1997a) Perspectives on Public Choice, Cambridge: Cambridge University Press. Mueller, D.C. (1997b) ‘Federalism and the European Union: a constitutional perspective’, Public Choice 90: 255–80. Tiebout, C.M. (1956) ‘A pure theory of local expenditures’, Journal of Political Economy 64: 416–24. Wicksell, K. (1896) Ein neues Prinzip der gerechten Besteuerung: Finanztheoretische Untersuchungen, Jena; translated as ‘A new principle of just taxation’, 1958; reprinted in R.A. Musgrave and A.T. Peacock (eds) (1967) Classics in the Theory of Public Finance, London: Macmillan, 72–118.

58

4

Authors name

Enlargement of the European Union Determining the boundaries of an optimal legal area Hans-Jörg Schmidt-Trenz and Dieter Schmidtchen1

Introduction Before joining the EU, countries from Central and Eastern Europe (CEECs) are required to adopt the so-called acquis communautaire, the entire body of existing EU legislation and practices. Adoption of the acquis not only means the approximation of their legal texts, but also the establishment of administrative and organisational structures necessary to make the legislation work (European Commission 1995: 23). Expanding the Union only makes economic sense if the process enhances the wealth of both the citizens of the EU and the citizens of the CEECs. The Commission, as well as most of the observers, seems to take this for granted. However, there is no free lunch. Increasing membership of a club may be accompanied by congestion and decision-making costs. Furthermore, one cannot have approximation of law without incurring set-up costs, including the costs of restructuring the economy.2 In addition, enlargement could generate third-party effects (externalities) that should be taken into account when evaluating it from the point of view of overall welfare. Many aspects of the enlargement issue have been analysed in the Economics literature, but, somewhat surprisingly, the approximation of law has largely gone unnoticed. Studies based on established theories of economic integration focus on gains achieved from trade liberalisation. These gains are then compared with the fiscal burden of enlargement in order to identify likely winners and losers (see, for example, Baldwin et al. 1997; Kohler 2000). The techniques and models used to evaluate the costs and benefits of EU accession have typically been computable general equilibrium models, macro models, sectoral approaches, gravity models focusing on trade effects, and optimal currency area frameworks (see Mortensen and Richter 2000: Chapter 4). This contribution analyses the enlargement issue with the help of a new analytical framework provided by the economic theory of optimum legal areas. For the purpose of this contribution an ‘optimal legal area’ is defined as the group of economic agents for whom submitting to the same legal order maximises net benefits (benefits net of costs).3 Applying insights from social contract theory as developed by the Virginia School (Buchanan 1975, 1990) and the theory of clubs,4 states and state-like entities such as the European Union are viewed as law

Determining the boundaries

59

enforcement agencies, i.e. protective clubs, which are concerned with the production and enforcement of law. In order to accomplish the task of law enforcement, protective agencies are typically organised according to the territoriality principle; enforcement is linked to a certain territory following a ‘strategy to control people and things by controlling area’ (Sack 1986: 5). Additionally protective states are characterised by finite membership. Addressing the enlargement issue from this perspective, it can be identified with the boundary question. The boundary question addresses the optimal size of the European Union when considered as an area with a single law. This question gives rise to two further questions: Why are protective states territorially organised? And what are the normative implications of the territoriality of law and enforcement? This is the set of problems posed by the New Institutional Economics of International Transactions.5 In this chapter the optimal size of the Union is derived by taking account of the following economic parameters: the international allocation of human capital, the productivity of human capital, traditional trade barriers, administrative protectionism, differences in regulatory setting and legal orders, and the enforcement costs of a legal order. Whereas the costs of enlargement to the EU-15 are often or even exclusively discussed in terms of additional spending via various EU-funds, this contribution focuses on the enforcement costs to both the EU-15 and the new members, starting from the observation that the overwhelming part of EU activities relates to regulation and liberalisation, not to money (see Pelkmans et al. 2000: 22). Enforcement costs consist of: (a) the costs of running the legal system mainly reflected by tax rates and litigation fees, (b) the costs of intermediaries, such as lawyers and lobbyists, (c) the privately incurred costs of protecting private property and transactions and, most important, (d) the costs of complying with the myriad of regulatory laws imposed on firms.6 This contribution also tries to answer the question of who will belong to the winners and losers and whether enlargement improves overall welfare.

An economic theory of optimal legal areas Social contract theory The social contract theory as developed by Buchanan (1975) and others uses the traditional rational choice model in order to explain how the ‘Hobbesian problem of social order’ can be solved by a voluntary, i.e. unanimous agreement upon a system of ‘property rights’ and the simultaneous installation of an enforcement institution, the ‘protective state’. That agreement, or contract, is elaborated on the basis of an equilibrium situation without contract, the so-called natural equilibrium, prevailing in the ‘state of nature’. From the point of view of bargaining theory, the pay-offs in the natural equilibrium define the threat or disagreement points. They indicate what each player can guarantee for himself or herself. These are the pay-offs the players will receive if bargaining fails. Bargaining is concerned with dividing the rent from cooperation. With the creation and enforcement of a system of property rights, which is the content of the social contract, the players also manage to divide the gains from cooperation.

60

H.-J. Schmidt-Trenz and D. Schmidtchen

The governance of social interactions on the basis of a rule of law principle allows people to economise on transaction costs, which can broadly be defined as the costs of running a politico-economic system. In particular, a system of private law, consisting of property, tort and contract law, allows people to engage in welfare-improving exchange relations. At the same time, transaction costs may also explain why a unified global order of law and its enforcement does not come into being. In his contract theoretic explanation of the protective state, Buchanan deals with this subject only briefly, in contrast to the rather evolutionary theories of the state (see, for instance, Nozick 1974). He argues that due to the large number of economic agents, negotiations and the formation of ‘states’ will initially take place between members of subgroups ‘within the larger and more inclusive community membership’ (Buchanan 1975: 31). Thus, Buchanan presumes that a multitude of protective states formed on the basis of corresponding social contracts may coexist, at least temporarily. In this situation, anarchic conflicts continue to exist between subgroups, leading to a natural equilibrium among the different states, each one working on behalf of one of those subgroups. In Buchanan’s view, a continuing ‘process of contractual internalisation’ (Buchanan 1975: 32) will lead to larger protective units until all individuals are joined within a single constitutional structure: the ‘inclusive community’.7 In all his further analyses, Buchanan presupposes a situation in which an extensive process of contractual internalisation has reached its end. For that reason, he only speaks of the state and the constitutional contract8 that fixes the mutual behavioural limits (cease-fire treaty),9 the property rights on goods, the process of legal enforcement, and the procedural rules governing the productive state (providing other public goods than that of private law), especially the kind of majority rule. However, Buchanan does not provide a model of this ‘process of contractual internalisation’. In this contribution, we deal with the obstacles to a complete internalisation, using the concept of transaction costs. Transaction costs do not play an explicit role in the contract theory of the Virginia School. Emphasis is placed on how the costs of anarchy can be overcome by a constitutional order.10 But this school of thought does not acknowledge transaction costs that result from agreeing on this order and organising its enforcement. This neglect of the ‘transaction costs of creating and maintaining cooperation’ explains why this order is presented as being perfectly specified, fully enforced by a single inclusive protective agency, and perfectly obeyed. However, we must realistically assume that both the negotiation of a system of property rights and its enforcement do entail transaction costs. Note that they are not independent of each other. For example, the functioning of a system of detailed property rights may be threatened by excessive enforcement costs. Therefore, a comprehensive cost-benefit calculus becomes necessary. As Buchanan (1975: 132) has already realised, ‘[a] more general model must allow for the simultaneous determination of the preferred or optimal quantities of law and the quantities of enforcement’. In the following sections we ask the following question: What are the number and size of states in a contract theoretic equilibrium? We are aware of the fact

Determining the boundaries

61

that, historically, state formation depends upon a whole complex of geographical, cultural, ethnic, ideological, military, political and economic forces that can never be treated exhaustively by a single model. Nevertheless, the contract theoretic perspective may provide some insights into the process of state formation, which can usefully be applied to Union enlargement.11 In the model we show that the specification and enforcement of property rights will most probably be incomplete. Moreover, membership in a state will be exclusive. This means that limits to group size are efficient.12 Since the citizens of the European Union can be considered as forming such a group, the European Union is confronted with the problem of determining its optimum size. The model: assumptions and interpretation Our starting point is a Hobbesian state of nature (anarchy) viewed as an N-person prisoners’ dilemma game. A prisoners’ dilemma game is defined by two characteristics: all players have dominant strategies and the equilibrium of the game is inefficient. Basically, people have to decide whether to be cooperative (nice) or non-cooperative (naughty) to their friends, colleagues, roommates, spouses, coworkers, co-conspirators, allies, partners, or fellow club members (see Shepsle and Bonchek 1997: 197). Since non-cooperation is best regardless of what others decide to do, all players choose the non-cooperative strategy leading to a state of affairs in which nobody can reap gains from cooperation.13 Let us call this game, because of its absence of legal rules and governments, the natural game. Each individual can choose from a set of strategies, including the cooperative, peaceful strategy, as opposed to, the non-cooperative, aggressive strategy. In the equilibrium of anarchy, the non-cooperative strategy proves to be the individually optimal strategy for all individuals, as is well known. In that situation, every individual gets a pay-off denoted YA representing the individual equilibrium income (social product per capita) that may be gained under anarchy. This situation is Pareto inferior to a world, which results when all individuals decide to use the cooperative strategy. A Pareto optimum is defined as a state of affairs in which no one in society can be made better off without making at least one person worse off. A state of affairs is Pareto inferior if somebody can be made better off without making another person worse off. A change in the state of affairs is said to be a Pareto improvement if at least one person is made better off and no one is worse off. As the players are stuck in a prisoners’ dilemma, they cannot reach such a Pareto improvement unless some additional constraints are introduced to the game. We call these constraints a ‘legal order’. Deciding upon a specific legal order means setting up a social contract.14 All parties to the contract agree to abide by the set of legal rules. The legal order can be depicted as a matrix of sanctions for illegal, i.e. non-cooperative, behaviour.15 Adding this matrix to a matrix describing the ‘natural’ game leads to a fundamental transformation. The original non-cooperative equilibrium strategy will be replaced by a new cooperative equilibrium strategy. Thus, due to the legal order society changes from anarchy to civilisation (see Schmidtchen and Schmidt-Trenz 1990).

62

H.-J. Schmidt-Trenz and D. Schmidtchen

As has already been mentioned by Buchanan, the social contract is not selfenforcing, which means that the fundamental transformation does not work out and the non-cooperative strategy remains the dominant strategy for all individuals. Therefore, the social contract must also provide for the installation of an agency in charge of enforcing the legal order. Call this agency the ‘government’. This term stands for a bundle of administrative, judicial and economic services, which can have either characteristics of a non-rival public good or those of an impure public good (club good), with partial rivalry, or a private good. Decisions regarding the social contract (‘choice of the rules’) are taken at the constitutional stage of society. Choices within the rules are a characteristic of its post-constitutional stage.16 In the following model we consider five variables:17 1 the number of contract parties, 2 the degree of specificity of property rights in terms of its content and personal assignment, 3 the degree of enforcement of this system by the public authorities, 4 the degree of openness of the economy, and 5 the enforcement technology.18 We assume risk-neutral agents, which means that utility can be represented by the (expected) income. An individual’s decision to join a club, i.e. subscribe to the social contract, will be based on a comparison of the present value of two income streams that the representative and (for the sake of simplicity) immortal individual can expect from his resource endowment. One is the present value of the income if anarchy persists and the other is the present value of the income that results if all individuals agree to install a legal order as described above. The latter income amounts to the difference between the anarchic income (YA) and the cost of negotiation of the social contract (NC), (that is, the income during the period of negotiation or the constitutional stage), plus the present value of the net income flows at the post-constitutional stage. The net income flows at the post-constitutional stage are calculated as the difference between the income earned in the post-constitutional stage of society, which is higher than that in anarchy, and the enforcement cost of the legal order borne by the representative citizen. These costs can be considered as the individual tax charge financing the production of government services, i.e. the legal protection by the state. The difference between post-constitutional income and enforcement cost can be interpreted as the present value of individual utility from the services provided by the state. If the present value of the net income flows at the post-constitutional stage (net of negotiation costs) exceeds the present value of anarchic income, an individual will subscribe to the social contract. In the principal agent literature this condition is called the rationality or participation constraint. Let us assume that there is a domain in which this participation condition is fulfilled. Efficiency requires maximising utility, which means, given risk-neutral individuals, determining the optimal size of membership, the optimal degree of the specification of property rights, the optimal degree of enforcement of the property rights and the optimal degree of the openness of the economy.

Determining the boundaries

63

As is shown in Appendix 4, the optimum size (n*) is reached if the marginal benefit created by the adoption of one additional contract party (higher membership) is equal to the additional individual enforcement costs (see Appendix 4, Equation 4.4). These enforcement costs are made up of two components: internal enforcement costs, that may increase due to congestion caused by an additional member, and, in the case of an open economy, external transaction costs, which might also be influenced. The optimum of contract specification (s*) requires that the marginal benefit of a better contract specification must be equal to the present value of the marginal enforcement costs of better specified property rights (see Appendix 4, Equation 4.5). Regarding the optimal degree of contract enforcement (e*), the marginal benefit resulting from increased levels of contract enforcement must be equal to the marginal costs of better enforcement (see Appendix 4, Equation 4.6). The degree of openness of the economy is optimal if the marginal benefit of lowering the trade barriers is equal to the increase of the enforcement costs caused by this move (see Appendix 4, Equation 4.7). The decisions on the optimal size, the optimal specification of property rights, the optimal enforcement and optimal degree of openness are reached simultaneously. Thus, a comprehensive optimum is reached.19 We will now discuss the optimum size of the state in more detail relying on the assumption that the optimal values of the other variables are given.20 A graphical representation The post-constitutional income production curve. In Figure 4.1 the units of social product per capita (Y ) appear on the vertical axis, the number of clients of the state on the horizontal axis. The curve Y represents the present value of the individual gross benefit (that is, per capita income) in the post-constitutional stage. We assume that the marginal return of a member is constant, which means that any additional member of the club increases the per capita income by a constant amount. Given this assumption, the income production function is a linear function with positive slope (see curve labelled Y in Figure 4.1). The increase in per capita income might be due to a deeper division of labour among the members of the club, or due to positive spillovers from an additional unit in production to other units, increasing the average productivity. The latter effect can be derived from a relationship between aggregate human capital and factor productivity, as is emphasised by the endogenous growth literature (see Romer 1986). The enforcement cost curve. The curve labelled C in Figure 4.1 represents the present value of the individual cost contributions to the total costs of enforcing the legal order. The enforcement costs C are made up of two components. These are costs incurred in enforcing the legal order internally among the contract parties themselves, as well as externally against strangers to the legal order. Internal costs include costs of safeguarding domestic contracts and costs of dispute management under the procedural rules of the protective club to which

64

H.-J. Schmidt-Trenz and D. Schmidtchen

Figure 4.1 Optimal size of the state.

the legal order compatriots belong. We assume these internal enforcement cost to be an over-linearly increasing function of club size (see curve CI in Figure 4.1).21 External enforcement costs take account of the fact that the legal order needs to be defended externally, i.e. against strangers to the legal order (non-compatriots). This could be necessary in two cases. First, a state could attack members of another state in order to opportunistically appropriate their wealth and enslave the members. Second, in possible transactions between individuals belonging to different states, tort, contract or criminal conflicts may arise. Thus, private international transactions may result in conflict and ultimately in war unless there is a ‘super club’ dealing with such problems on an international level. Both factors create costs, which can be considered as the transaction costs of running a multitude of states. Given a world population of size N, increasing club size implies that a larger part of world population is comprised of compatriots. Consequently, external enforcement cost are assumed to decrease when adding additional members (see curve CE in Figure 4.1). Overall enforcement costs C may be viewed as an aggregate cost function representing the sum of the present values of the ‘internal enforcement costs’ CI, external enforcement costs CE and defence costs CD, i.e. C = C + CE + CD for each given club size. Neglecting the defence costs, the curve of the present value of the enforcement costs per capita C can be derived by vertical aggregation of

Determining the boundaries

65

the CI- and CE-curves, leading to curve C. Given the properties of CI and CE, the C function is u-shaped. The solution. Consider club size to the left of n* such as n1. This club size is suboptimal, since adding another member (that is, moving into the direction of n*) decreases per capita enforcement cost and increases per capita income. Thus, the per capita net benefit increases relative to that with club size n1. Now consider size n2 . This club size is also suboptimal. Reducing club size by a ‘small amount’ reduces per capita income as well as per capita enforcement cost. However, since the cost reduction (marginal enforcement cost) is bigger than the income reduction (marginal income) the difference between per capita income and per capita enforcement cost increases. Thus, the per capita net benefit increases. State size is optimal when marginal enforcement cost equals marginal income. Graphically, the slope of the cost curve is equal to the slope of the curve representing per capita income. This is the case with n = n* (see Figure 4.1). The optimal size n* can also be determined by calculating the maximum of G = Y – C – NC, with G denoting the net gain of being a member of the state. The graph of G is depicted in the lower part of Figure 4.1, which neglects NC, since as a constant NC has no influence on n*. That n which maximises G defines optimal club size, n*. As can be seen, it is the u shape22 of the enforcement cost function that is responsible for a finite optimal state size. The ratio k* = N/n* gives the solution regarding the efficient number of states if it is an integer. The model does not solve the problem of how to assign the world population to certain states in the case of homogeneity of the states.23 For the sake of simplicity the assumption could be made that individuals are assigned randomly. After being assigned to a particular state an individual concludes a contract with this state and pays his membership fee.

Enlargement24 The modified model We now turn to an application of the theory of optimal legal areas. If there were no transaction costs and world population were homogeneous, the number and size of the nations could be restructured in an efficient manner. This is not the world as we know it. Transaction costs are positive, the world’s population is not homogeneous, and history is relevant. As for the latter, think of the Cold War and the Iron Curtain, two factors that have been operating as constraints to a purely economic determination of the number and size of nations. This does not make the model of the preceding section useless. It must be modified in an adequate way. Our modification takes the following form. The world’s population is assigned to three groups of states: European Union (E), applicant countries (B) and the rest of the world (R). We apply a two-step optimisation procedure. In the first step we determine current Union size, denoted n*E. That is done by using a calculus as developed in the former section

66

H.-J. Schmidt-Trenz and D. Schmidtchen

under the constraint that the CEECs (representing the applicant countries) cannot become members of the Union. This constraint reflects the fact that the European Union could seek to realise its optimal size but not by expanding towards Central and Eastern Europe due to the Iron Curtain. With the fall of the Iron Curtain this constraint vanished, opening up the potential for a recalculation of the optimal size of the Union. This recalculation is captured by the second step. In the second step we start with Union size n*E as determined in the first step and use a similar procedure to calculate the optimal number of new members, denoted n*Q. Adding n*Q to n*E leads to the optimal size of the larger Union. Since we now have four groups in the world – namely old members, new members, outsiders from B-countries and R-countries – some modifications of the enforcement cost as well as the income production functions are necessary. Consequently, there are four income production functions: one for the EU-15, one for the CEECs becoming new members, one for those CEECs staying outside, and one for the rest of the world. Each production function describes how per capita income depends on the following factors: 1 2 3 4

domestic transactions, international transactions, the quality of the legal orders of the respective countries, the degree of openness of the economies.

We also have a separate enforcement cost function for each of the four country groups. Each function describes how enforcement costs depend on the following factors: 1 internal enforcement costs, 2 external enforcement costs, allowing for differences with respect to the different foreign countries. For simplicity, we assume that the enforcement cost functions of the current and the new members are identical. Knowing the income production functions as well as the enforcement cost functions we can calculate the net gains from enlargement for all groups.25 Optimal size of the Union Before approaching the enlargement issue in more detail some remarks on the optimisation procedures seem in order. Consider Figure 4.2. In the upper part of Figure 4.2 the linear curve with an initially solid and then dotted part is the graph of the income production function under the assumption nQ = 0. Note that the assumption nQ = 0 reflects the constraint that the EU in the past could not expand towards Middle and Eastern Europe due to the Iron Curtain. The cost function under the same assumption is the u-shaped curve. In the lower part of Figure 4.2 the net gain is represented by the initially solid and then dotted curve.

Determining the boundaries

67

Figure 4.2 Optimal enlargement.

In the first step, n*E is determined by the maximisation of the difference between income and costs under the assumption that the size of the B-countries is constant, which means nQ = 0. If there is an internal optimum,26 this is determined by the slope of the income production curve being equal to the slope of the cost curve, given nQ = 0. Alternatively, the optimal size can be determined with the help of the curve representing the net gain (see lower part of Figure 4.2). For the optimisation procedure regarding the enlargement issue, n*E has to be taken as given. This means that n*E is assumed to represent the current size of the Union (that is, the EU-15). Geometrically, this means taking the point n*E as origin and drawing income and cost curves defined as a function of nQ (see the curves with labels YE, CE, GE in Figure 4.2). Applying a similar procedure as in step 1, n*Q is determined by the maximisation of the difference between income YE and costs CE. If there is an internal optimum, this is given by the slope of the income production curve being equal to that of the cost curve or, alternatively, by the maximum of GE = YE – CE as shown in the lower part of Figure 4.2. Enlargement means a change in Union size from n*E to n*E + nQ. The optimal number of new members, nQ, from the point of view of the Union can be calculated by maximising Union’s net gain GE with respect to nQ (as is shown in Appendix 4). From these calculations we can derive the following five propositions:

68

H.-J. Schmidt-Trenz and D. Schmidtchen

1 The higher the internal enforcement costs, the lower the optimal number of new members, n*Q. 2 The higher the external enforcement costs, the higher the optimal number of new members, n*Q. 3 The bigger the size of former socialist countries, the higher the optimal number of new members, n*Q. 4 The higher the degree of the adoption of the acquis, the higher the optimal number of new members, n*Q. 5 The better the quality of the legal orders of the outsiders, the lower the optimal number of new members, n*Q. The winners and the losers In this section we discuss the impact of enlargement on the welfare of the current and the new members, the outsiders and the entire group (excluding the Rcountries). Knowing the respective income production and enforcement cost functions, the difference between income and enforcement costs can be easily calculated for alternative degrees of enlargement (alternative n*Q). We restrict analysis to the enlargement scenario represented by Figure 4.2. Does enlargement improve the welfare of the current members of the EU? Given the scenario shown in Figure 4.2, the answer is clearly yes. However, Figure 4.2 represents only one of nine possible scenarios. Each scenario is characterised by a specific combination of the slopes of the income production and enforcement cost functions at the current size of the union (EU-15), that is, n*E.27 If, for example, the actual enlargement scenario were to be represented by dashed curves in Figure 4.2, enlargement would not increase the welfare of the EU-15. As for the new members the following holds: their gain function, denoted GQ is parallel to that of the current members, denoted GE. The reason is that the enforcement cost functions have been assumed to be identical. New members who did not perfectly adopt the acquis receive a lower net gain. Perfect adoption would cause income production functions identical for both the new members and the EU, leading to identical net gain functions in Figure 4.3. Note further that Figure 4.3 also allows for a different interpretation. If the acquis is well suited for the EU-15 but not for the new members – as is assumed to be the case (see Baldwin et al. 1997: 128) – the lower GQ relative to GE is due to the implementation of an inefficient legal order by the new members. The impact of enlargement on the welfare of the outsiders can be analysed in a similar way. As is shown in Appendix 4, the marginal benefit at nQ = 0 remains unclear. It is possible that the net benefit is positive and increases with nQ up to a maximum from which it decreases and ends in the negative (as is shown in Figure 4.3, curve GD). However, other positions and slopes of the GD-curve cannot be ruled out. Although the position and slope of the net gain curve of the outsiders depend on the specification of the parameters (as indicated in Appendix 4), one thing can be stated with certainty (given the scenario under consideration): while the current and new members always reap gains from enlargement, the outsiders may lose.

Determining the boundaries

69

Q

Figure 4.3 Net gains of enlargement.

Adding up the gain functions of the old members using the respective group size as a weight, the new members and the outsiders deliver the net gain function of the whole group. There is a positive net gain for the whole group, since at n*Q the gains of the old and new members outweigh the losses of the outsiders. The net gain of the whole group, consisting of current and new members as well as outsiders, can be calculated by adding up the net gains of each group weighted by the respective group size (see curve labelled GW in Figure 4.3). In the scenario presented in Figure 4.3 overall welfare is at a maximum if enlargement embraces all CEECs, i.e. nQ = nB. However, it is not to the benefit of the EU-15 to increase union size beyond n*E + n*Q. Doing so would reduce the welfare of both the current and the new members. However, the loss in welfare to be borne by the outsiders outweighs the welfare reduction of the current and new members. Overall welfare increases as is indicated by the fact that the GW-curve has its maximum at nQ = nB.

Outlook Legal scientists and philosophers have been dreaming of a common European law and even a universal world law for centuries. Economists, however, have devoted relatively little attention to this issue, in particular relatively few economists have provided formal models concerning state formation and the optimal size of states (see Alesina and Spolaore 1997). This contribution is a first step towards the development of a theory of optimum legal areas and its application to the study of enlargement of the European Union. Rather than summarising our results we would like to point out avenues for further research. These avenues are connected with a modification of some of our model’s assumptions. A model with more explanatory power should include heterogeneous preferences of the population, heterogeneity of the human capital, which would lead to differences in individual incomes, and a more sophisticated treatment of the government, including the possibility of decentralisation of government.

70

H.-J. Schmidt-Trenz and D. Schmidtchen

The cost issue of setting up a new legal order in the applicant countries and the costs of institutional reform within the Union need further elaboration. If it should turn out that these costs are all but negligible, institutional alternatives to enlargement and the approximation of law must be found. What comes to mind is trade liberalisation. The theory of optimal legal areas can be applied to tackle the question whether the road to free trade is superior to the road to enlargement. However, free trade also has associated problems. Even if we have free trade as understood in International Economics, there are barriers to international transactions that are largely neglected in the literature. Whereas for domestic, internal transactions, one monopolist, the protective state, fulfils the task of law enforcement, international, external transactions make contact with a multitude of legal systems and with the monopoly of power claimed by each state within its boundaries. Furthermore, whereas the legal rules of each protective state can be judged, at least in principle, as unequivocal, in the international arena we do find a serious ‘inconsistency of rights’. Collisions of norms and gaps between different legal systems appear. An accord in court decisions is often coincidental, and the assistance of the judicial and penal institutions in foreign countries is not at all a matter of course. Thus, the territoriality of law and law enforcement results in a specific kind of attenuation of property rights and the emergence of a special kind of risk, which has been labelled constitutional uncertainty in international transactions. This kind of uncertainty gives rise to coordination problems of a special kind, reflected in corresponding transaction costs. Harmonisation of law and enlargement can be considered as a means for the reduction of these transaction costs. But one should not forget that there are also privately created means such as the law merchant and reputation (see Schmidt-Trenz and Schmidtchen 1991). In some cases these means might be cheaper and more effective than consciously cooperating governments or mergers of government. APPENDIX 4 Let n, s, e, j, o denote, respectively, the following variables: the number of contract parties, the degree of specificity of property rights in terms of content and personal assignment, the degree of enforcement of property rights, the degree of openness of the economy, and the enforcement technology. Furthermore, let YA, Y, C, NC, respectively, denote anarchic income, post-constitutional income, enforcement costs and negotiation costs. Assume Y = Y (s, e, n, j) and C = C (s, e, n, j, o). These functions have the following properties:

∂Y / ∂s > 0; ∂2 Y / ∂s2 < 0; ∂Y / ∂e > 0; ∂2 Y / ∂e2 < 0; ∂Y / ∂n > 0; ∂2 Y / ∂n2 = 0; ∂Y / ∂j < 0; ∂2 Y / ∂j2 < 0. ∂C / ∂s > 0; ∂2 C / ∂s2 > 0; ∂C / ∂e > 0; ∂2 C / ∂e2 > 0; ∂2 C / ∂n2 > 0; there exists ñ: ∂C / ∂n = 0; n < ñ: ∂C / ∂n < 0; n > ñ: ∂C / ∂n > 0. ∂C / ∂j < 0; ∂2 C / ∂j2 > 0.

Determining the boundaries

71

We suppose that n is a natural number, C > 0 for n = 0, which means that part of the enforcement cost borne by the individual is independent of the size of the club; Y = YA and C = 0 if s ∨ e = 0.

Participation constraint

[

]

YA 1 < Y ( s, e, n, j) − C ( s, e, n, j; o) − NC r r

(4.1)

with r as discount rate. Inequality (4.1) is the overall condition required for a social contract to be concluded. Present value of income, if anarchy persists (left-hand side), is smaller than present value of post-constitutional net gain minus negotiation costs (righthand side).

Optimality conditions Define Y T as

YT =

[(

) (

)]

1 Y s, e, n, j − C s, e, n, j; o − NC r

(4.2)

Then we can formulate the maximisation programme:

[ (

)]

max Y T s, e, n, j; o

(4.3)

For simplicity, we exclude cross-effects between n, s, j and e. This maximisation programme results in the following marginal conditions:28

∂Y / ∂n = ∂C / ∂n

(4.4)

∂Y / ∂s = ∂C / ∂s

(4.5)

∂Y / ∂e = ∂C / ∂e

(4.6)

∂Y / ∂j = ∂C / ∂j

(4.7)

These marginal conditions indicate the optimum size of a state (4.4), denoted n*; the optimal degree of the specification of property rights (4.5), denoted s*; the optimal degree of contract enforcement (4.6), denoted e*, and the optimal degree of openness (4.7), denoted j *.

72

H.-J. Schmidt-Trenz and D. Schmidtchen

Specification of the income production function Let h denote total human capital in the world. Divide the world into two entities, country x (which individual i belongs to) and the rest of the world. Assume that world population has mass one and each individual is endowed with the same amount of human capital. Define Hx as aggregate human capital in country x and H∼x as aggregate human capital in the rest of the world (see Alesina and Spolaore 1997: 1040, for what follows). Finally, assume that individual income, denoted Yx, depends on aggregate human capital as described in (4.8).

Yx = bo + b1 ⋅ H x + b2 ⋅ H~ x ; bo , b1 , b2 > 0

(4.8)

Equation (4.8) implies that individual income is given by a constant bo plus a linear term in aggregate human capital, both at home and abroad. Parameters b1 and b2 stand for the aggregate human capital externality. If each individual is endowed with the same amount of human capital we have: Hx =nx · h and H∼x =(1–nx) · h with nx size of country x. Setting b2 =(1–j) · b1 leads to

Yx = bo + b1 ⋅ nx ⋅ h + b1(1 − j)(1 − nx ) ⋅ h

(4.9)

The magnitude of parameter j can be interpreted as an indicator of the openness of country x. If j = 1, we have a completely closed domestic economy. Thus, individual income is only determined by the domestic aggregate human capital. If j = 0, we have the opposite case of a completely open economy. Individual income depends on world aggregate human capital. The curve labelled Y in Figure 4.1 is a graphical representation of (4.9).29

Specification of the enforcement cost function C = c1 ⋅ n2 + c2 [(1 − j)(1 − n)]2

(4.10)

represents such a function. The first term on the right-hand side describes internal enforcement costs CI. CI(0) = 0; CI(1) = c1. From ∂CI/∂n = 2c1n > 0 and ∂2CI/∂n2 = 2c1 > 0 it follows that this function is exponentially increasing in n. The second term on the right-hand side of (4.10) stands for the costs of a club member’s external transactions, CE. CE(0) = c2(1– j)2; CE(1) = 0. From ∂CE/∂n = –2c2(1– j)2(1–n) < 0 and ∂2CE/∂n2 = 2c2(1– j)2 > 0 it follows that the function is downward sloping in the way presented in Figure 4.1.

Determining the boundaries

73

Optimal state size The optimal size n* can be algebraically derived by calculating the maximum of G = Y – C – NC, with G denoting the net gain of being a member of the state. Taking the partial derivative of G with respect to n and setting it equal to zero results in the first order condition of a maximum. Solving for n leads to

n* =

c2 (1 − j)2 + b1 jh / 2 c1 + c2 (1 − j)2

,

from which it follows that n = 0 can never be an optimum and that with 2c1 < b1 · j · h optimal state size is n* = 1. In the latter case it would be optimal to have just one state in the world.

Enlargement The European Union (E-countries), the rest of the world (R-countries) and the applicant countries (B-countries) have size nE, nR and nB, respectively; with nE + nR + nB = 1. Let nQ denote the members of the B-countries qualified for entering the Union and nD the non-qualified candidates, which must remain outside the Union, i.e. nB = nQ + nD. We assume the following functions, which represent present values.

YE = bo + b1 ⋅ h ⋅ nE + α Q ⋅ b1 ⋅ h ⋅ nQ + b1 ⋅ h(1 − jR ) ⋅ (1 − nE − nB ) +

α D ⋅ b1 ⋅ h (1 − jD ) ⋅ ( nB − nQ )

(4.11)

is the income production function for the old members of the Union. This function is different from Function (4.9) in several respects: there is an additional term αQb1hnQ, capturing the positive external effect on the income of old members caused by the aggregate human capital of the new members. Since it is reasonable to assume that this externality may depend on the degree of adoption of the acquis, we introduce parameter αQ, with αQ = 1 representing a perfect adoption and αQ < 1 for an imperfect one. The term capturing the impact of foreign human capital on Union income must be split into two parts: the first part, b1h(1–jR)(1–nE –nB), relates to Union income created by aggregate human capital in the R-countries, whereas the second part, αDb1h(1–jD)(nE–nQ), shows the impact of aggregate human capital of the non-qualified candidates. The parameters jR, jD are indicators of the respective degrees of openness; αD is a parameter that allows us to represent the idea that the impact of the human capital in nonqualified countries on Union income depends on the productivity of the human capital, which is influenced by the ‘quality’ of the legal order in those countries (the lower the quality the lower αD). The structure of the income production function of the new members, denoted YQ, is similar to that of the old members:

74

H.-J. Schmidt-Trenz and D. Schmidtchen

YQ = bo + α Q ⋅ b1 ⋅ h ⋅ nQ + α Q ⋅ b1 ⋅ h ⋅ nE + α Q ⋅ b1 ⋅ h(1 − jR ) ⋅ nR + α D ⋅ b1 ⋅ h(1 − jD )( nB − nQ )

(4.12)

Note that the impact of the internal human capital as well as the human capital of the R-countries is discounted by factor αQ, indicating the degree of adaptation of the acquis. The following function represents the enforcement cost function of the Union, denoted CE:

C E = c1 ⋅ n2E + c1 ⋅ nQ2 + c2 (1 − jR )2 (1 − nE − nB )2 + c2 (1 − jD )2 ⋅ ( nB − nQ )2

(4.13)

The first and the second term on the right-hand side indicate the internal enforcement costs caused by the interactions of the old and new members, respectively. While the third term on the right-hand side of Function (4.13) reflects the external enforcement costs of transactions with R-countries, the fourth term represents enforcement costs of transactions with non-qualified Bcountries. We assume that the cost functions of the old and new members are identical. The reason is that the income functions already take account of the effect of a less than perfect adoption of the acquis. The income production and cost functions, denoted YD and CD, respectively, of the outsiders are defined in (4.14) and (4.15):

YD = ao + α D ⋅ b1 ⋅ h( nB − nQ ) + α D ⋅ b1 ⋅ h(1 − jR )(1 − nE − nB ) + α D ⋅ b1 ⋅ h(1 − jD ) ⋅ ( nE + nQ )

C D = c1( nB − nQ )2 + c2 (1 − jR )2 (1 − nE − nB )2 + c2 (1 − jD )2 ⋅ ( nE + nQ )2

(4.14)

(4.15)

The net gains of the old members, the new members, the non-qualified members and the whole group can easily be calculated using the functions (4.11)–(4.15). Note that the income production function on the domain (nE+nQ) = [0, nE+ nB] consists of two parts. On the domain (nE+nQ) = [0, n*E] the income-function holds for nQ = 0. Since on the domain (nE+nQ) = [n*E, n*E+nB] the function is steeper than up to n*E a kink occurs at n*E. The function is steeper if αQ > αD (see Equation 4.11). For a similar reason the cost function has a kink at n*E. The first part is defined on the domain (nE+nQ) = [0, nE*] with the first and second derivatives f11: = 2c1 · nE – 2c2 (1 –jD)2 · (1 – nE – nB) and f12: = 0, respectively. The second part is defined on the domain (nE+nQ) = [n*E, n*E+nB] with the first and second derivatives f21: = 2c1 · nQ – 2c2 (1 –jD) (nB – nQ) and f22: = 0, respectively. It is obvious that the gain function also has a kink at n*E.

Determining the boundaries

75

Optimal enlargement from the point of view of the current members Equations (4.11) and (4.13) allow the net gain of the current members, GE = YE + CE, to be determined. For a calculation of the optimal degree of enlargement, denoted nQ*, take the partial derivative of the GE-function

∂GE / ∂nQ = α Q ⋅ b1 ⋅ h − α D ⋅ b1 ⋅ h(1 − jD ) − 2c1 ⋅ nQ + 2c2 (1 − jD )2 ( nB − nQ )

(4.16)

Setting (4.16) equal to zero leads to the first order condition, which implies

nQ* =

2 1 α Q ⋅ b1 ⋅ h − α D ⋅ b1 ⋅ h(1 − jD ) + 2c2 (1 − jD ) ⋅ nB 2 c1 + c2 (1 − jD )2

(4.17)

Propositions 1–3 can easily be derived from (4.17). As for the other propositions take note of:

∂nQ* / ∂α Q =

1 2[ c1 + c2 (1 − jD )2 ]

∂nQ* / ∂α D = −

⋅ b1 ⋅ h > 0,

1 2[ c1 + c2 (1 − jD )2 ]

⋅ b1 ⋅ h(1 − jD ) < 0.

Optimal enlargement from the point of view of the new members Equation (4.12) and (4.13) allow the net gain of the new members, denoted GQ, to be determined. The partial derivative with respect to nQ is given in

∂GQ / ∂nQ = α Q ⋅ b1 ⋅ h − α D ⋅ b1 ⋅ h(1 − jD ) − 2c1 ⋅ nQ + 2c2 (1 − jD )2 ⋅ ( nB − nQ ).

(4.18)

Since (4.18) is identical to (4.16) the marginal benefit of enlargement on the domain [n*E, nE+nB] for the new members equals that of the old members. In other words, the slopes of the gain curves are identical for each nQ. However, new members receive a lower level of the net gain as long as αQ < 1. That means GQ is parallel to GE, and becomes identical with GE if αQ = 1.

Optimal enlargement from the point of view of the outsiders Based on (4.14) and (4.15) the partial derivative of GD with respect to nQ looks as follows:

76

H.-J. Schmidt-Trenz and D. Schmidtchen

∂GD / ∂nQ = −α D ⋅ b1 ⋅ h + α D ⋅ b1 ⋅ h(1 − jD ) + 2c1( nB − nQ ) − 2c2 (1 − jD )2 ( nE + nQ )

(4.19)

At nQ = 0, the marginal benefit is unclear. The sum of the first and second terms is always negative. Thus, it depends on [2c1 · nB – 2c2(1 – jD)2 · nE] whether the marginal benefit is positive, negative or zero. Without a specification of the parameters the net effect of enlargement cannot be determined.

Notes 1 We would like to thank P. van Rompuy, Winand Emons, Jan-Erik Lane, Karen Horn, Karen deGannes, Roland Kirstein, Christian Koboldt, Matthias Leder, Alexander Neunzig, participants in the first and fourth NEMEU research conferences, for helpful comments. 2 The costs of complying with the acquis vary from one area to another (see Mortensen and Richter 2000: 22). Mortensen and Richter (2000: 22–3) conclude that ‘the financial implications of all this may be considerable, amounting in the case of lowincome CEECs to several percent of their annual GDP: a cost that will have to be sustained over an extended period’. They quote a recent estimate, according to which the total cost of applying the EU standards and norms would, for Hungary, be in the range of 50–120 percent of one year’s GDP (see also European Commission 1997). However, these are set-up costs, which are not relevant in the medium and long run. In any case, they are fixed and sunk costs. 3 As such the basic framework is very similar to the one used in the theory of optimum currency areas, of the theory of clubs, of the modern theory of the firm, viewed as a transaction cost saving organisation, and of the theory of geographic size of markets. 4 See Buchanan (1965), Allen et al. (1974), and Sandler and Tschirhart (1980). 5 See Schmidtchen (1994), Schmidtchen and Schmidt-Trenz (1990), and Yarbrough and Yarbrough (1994). 6 According to Wallis and North (1986) who measured the transaction costs of the USeconomy in 1970, transaction costs range from 46.7 to 54.7 per cent of GDP. The lower figure does not include value added of the public sector; it comprises marketing costs and the costs of organising firms. However, marketing costs as well as the costs of organising firms are largely influenced by the law. According to a study conducted by Bier (1999) the production costs of the judicial sector (mainly the sum of a and b) for Germany amount to DM86 billion (43 billion euro). DM48 billion (24 billion euro) were financed by the state budget, 37 (18.5 billion euro) are privately incurred production costs. The opportunity costs of time and the error costs due to judicial decisions are not included. 7 See Buchanan (1975: 34). Whynes and Bowles (1981: 5) reach the same result: ‘[W]e might think of the contract theory of the state in the following terms. A group of individuals will sit around a conference table and thrash out an agreement or contract under the terms of which all potentially gain a superior outcome to that of any other action. Such a contract will specify the rules of social action necessary for the attainment of the optimal outcome and will, by virtue of the individual’s signature on the contract, be acceptable to all (...). If such conferences can be costlessly convened, we might expect these sorts of arguments to militate in the final analysis in favour of a world-wide solution.’ 8 See Buchanan (1975: 71), who uses the term ‘constitutional mix’. 9 As Buchanan (1975: 59) puts it: ‘The contract is one of bilateral behavioral exchange.’ The contract contains restrictions on behaviour, which are agreed upon voluntarily. Rights and duties are defined for the first time. Activities can now be identified as lawful or unlawful.

Determining the boundaries

77

10 These costs can be labelled ‘avoidable transaction costs’. ‘Avoidable transaction costs’ are the cooperation benefits foregone. See the classification of transaction costs based on game theory provided by Schmidt-Trenz (1990: 64). 11 If we could manage to give a rationale for the optimal size of states (that are smaller than the world), then we would deliver a ‘fundamental potential explanation’ for the division of the world into a plurality of states with territorially limited legal systems. For the notion of a ‘potential explanation’, see Nozick (1974: 32). ‘We can learn a lot if we realize how the state might have come into being, even if it didn’t really do so this way’ (Nozick 1974: 23). In Hayek’s (1979: 156) terms, this is a description of ‘conjectural history’. We shall not refer in detail to Carneiro’s (1970) discussion of how states have really come into being, even if contrary to his propositions one could find proofs for voluntary alliances. Auster and Silver (1979: 21) criticise him for these very reasons. 12 See Olson (1965: 36) on ‘exclusive’ and ‘inclusive’ groups. 13 See Shepsle and Bonchek (1997: Chapter 8), and Hardin (1982: Chapter 2), who explicitly analyses the N-person prisoner’s dilemma. 14 Binmore (1997: 4 ff.) takes a different view. He interprets a society’s social contract as a coordination device, which can be reached by direct negotiation or as a consequence of social evolution. The game of life is modelled as an infinitely repeated prisoners’ dilemma game. The social contract specifies which equilibrium path out of the infinitely large number of paths (folk theorem) is to be played in a society. Thus the social contract solves an equilibrium selection problem. Rights and duties are considered as means for sustaining the chosen equilibrium path. 15 The view is taken and elaborated in more detail in Schmidtchen and Schmidt-Trenz (1990: 17). 16 The distinction between the constitutional and the post-constitutional stage of decision-making goes back to Buchanan (1977: 287 ff.). 17 Auster and Silver (1979: 27) chose similar variables. They examine ‘the number of people’, ‘geographical size’, ‘the level of order’ and ‘the level of resources used by the state’. 18 The enforcement technology can be based either on the so-called territoriality principle or on the personality principle (see Schmidtchen and Schmidt-Trenz 1990). 19 The optimal enforcement technology is not derived in this contribution. Elsewhere it is shown that the territoriality principle Pareto-dominates the personality principle (see Schmidtchen and Schmidt-Trenz 1990). 20 Schmidt-Trenz (1990: 212) provides a very detailed discussion and analysis of the optimality calculus determining s* and e*. The marginal conditions (4.5) and (4.6) suggest that the specification and enforcement of ‘property rights’ will be incomplete. Due to (4.6), for example, preventing legal violation is inefficient. 21 In this context, compare Williamson (1967: 123), who confirms ‘the management factor is responsible for a limitation to firm size’. Auster and Silver (1979: 28) point out that opportunism becomes more important with growing membership. If the number of potential legal conflicts rises proportionally to the number of possible interactions, then – according to the formula n(n-l) – an additional member means an exponential rise in conflicts. 22 For a more thorough discussion of this point see Bean (1973: 204), Auster and Silver (1979: 29), and Moss (1980: 25). As an analogy to the theory of the firm Bean (1973: 204) assumes such a shape. Moss (1980: 25) argues that: ‘[u]nless something is said about tastes or technology of providing public services, it would seem that the optimal size of the “protective state” is the world population.’ Nozick (1974: 30), however, only seems to focus on increasing returns to scale. 23 The procedure by which the segregation may be achieved must be negotiated by all the members. 24 This part of the contribution has benefited substantially from collaboration with Alexander Neunzig.

78

H.-J. Schmidt-Trenz and D. Schmidtchen

25 26 27 28

See Appendix 4 for further details. We neglect corner solutions for the sake of simplicity. The explanation for the kinks in Figure 4.2 is given in Appendix 4. These are necessary or first order conditions for a maximum. Given the shape of the functions the second order conditions for a maximum are fulfilled. It does not make sense to derive a marginal condition with regard to the enforcement technology ‘o’ since o is a discrete variable. 29 Index x is dropped.

References Alesina, A. and E. Spolaore (1997) ‘On the number and size of nations’, Quarterly Journal of Economics 111: 1027–56. Allen, L., R.C. Amacher and R.D. Tollison (1974) ‘The economic theory of clubs: a geometric exposition’, Public Finance 19: 386–91. Auster, R.D. and M. Silver (1979) The State as a Firm. Economic Forces in Political Development, Boston: Nijhoff. Baldwin, R.E., J. Francois and R. Portes (1997) ‘The costs and benefits of Eastern enlargement: the impact on the EU and Central Europe’, Economic Policy 24: 127–76. Bean, R. (1973) ‘War and the birth of the nation state’, Journal of Economic History 33: 203–21. Bier, Ch. (1999) ‘Die volkswirtschaftlichen Kosten der Justiz’, in D. Schmidtchen and S. Weth (eds) Der Effizienz auf der Spur. Die Funktionsfähigkeit der Justiz im Lichte der ökonomischen Analyse des Rechts, Baden-Baden: Nomos Verlagsgesellschaft, 124–36. Binmore, K. (1997) Game Theory and the Social Contract II: Just Playing, Cambridge, MA: MIT Press. Buchanan, J.M. (1965) ‘An Economic Theory of Clubs’, Econometrica 32: 1–14. Buchanan, J.M. (1975) The Limits of Liberty: Between Anarchy and Leviathan, Chicago: University of Chicago Press. Buchanan, J.M. (1977) Freedom in Constitutional Contract: Perspectives of a Political Economist, College Station, TX: Texas A&M University Press. Buchanan, J.M. (1990) ‘The domain of constitutional economics’, Constitutional Political Economy 1: 1–18. Carneiro, L. (1970) ‘A theory of the origin of the state’, Science 169: 733–38. European Commission (1995) ‘Preparation of the associated countries of Central and Eastern Europe for integration into the Internal Market of the Union’, White Paper, Brussels. European Commission (1997) ‘Compliance costing for approximation of EU environmental legislation in the CEEC’, April 1997, Brussels: DG 11. Hardin, R. (1982) Collective Action, Baltimore and London: Johns Hopkins University Press. Hayek, F.A. (1979) Law, Legislation and Liberty, vol. III: The Political Order of a Free People, Chicago: University of Chicago Press. Kohler, W. (2000) ‘Die Osterweiterung der EU aus der Sicht bestehender Mitgliedsländer’, Perspektiven der Wirtschaftspolitik 1: 115–41. Mortensen, J. and S. Richter (2000) ‘Measurement of Costs and Benefits of Accession to the European Union for Selected Countries in Central and Eastern Europe’, Vienna: WIIW, Research Report 263. Moss, L.S. (1980) ‘Optimal jurisdiction and the economic theory of the state: or, anarchy and one-world government are only corner solutions’, Public Choice 35: 17–26.

Determining the boundaries

79

Nozick, R. (1974) Anarchy, State, and Utopia, New York: Basic Books. Olson, M. (1965) The Logic of Collective Action. Public Goods and the Theory of Groups, Cambridge, MA: Harvard University Press. Pelkmans, J., D. Gros and J.N. Ferrer (2000) Long-run Economics Aspects of the European Union’s Eastern Enlargement, Den Haag: WRR, Working Documents no. W109. Romer, P.M. (1986) ‘Increasing returns and long-run growth’, Journal of Political Economy 94: 1002–37. Sack, R. (1986) Human Territoriality: Its Theory and History, Cambridge: Cambridge University Press. Sandler, T. and J.T. Tschirhart (1980) ‘The economic theory of clubs: an evaluative survey’, Journal of Economic Literature 18: 1481–521. Schmidt-Trenz, H.-J. (1990) Außenhandel und Territorialität des Rechts: Grundlegung einer Neuen Institutionenökonomik des Außenhandels, Baden-Baden: Nomos Verlagsgesellschaft. Schmidt-Trenz, H.-J. and D. Schmidtchen (1991) ‘Private international trade in the shadow of the territoriality of law: why does it work?’, Southern Economic Journal 58: 329–38. Schmidtchen, D. (1994) ‘International contracting and territorial control: the boundary question’, Journal of Institutional and Theoretical Economics 150: 272–8. Schmidtchen, D. and H.-J. Schmidt-Trenz (1990) ‘The division of labor is limited by the extent of the law: a constitutional approach to international private law’, Constitutional Political Economy 1: 49–71. Shepsle, K. and M.S. Bonchek (1997) Analyzing Politics: Rationality, Behavior and Institutions, New York: Norton. Wallis, J.J. and D.C. North (1986) ‘Measuring the transaction sector in the american economy, 1870–1970’, in S.L. Engerman and R.E. Gallman (eds) Long-Term Factors in American Economic Growth, Chicago: University of Chicago Press, 95–161. Whynes, D.K. and R.A. Bowles (1981) The Economic Theory of the State, Oxford: Robertson. Williamson, O.E. (1967) ‘Hierarchical control and optimum firm size’, Journal of Political Economy 75: 123–38. Yarbrough, B. and R. Yarbrough (1994) ‘International contracting and territorial control: the boundary question’, Journal of Institutional and Theoretical Economics 150: 239–64.

80

5

Authors name

The politics of institutional change Diverging goals for reforming the European legislature Thomas König and Thomas Bräuninger

Different criteria for joining the Union’s club The dynamics of European integration can be looked at from two angles: the extension of the Union’s legislative competencies and its enlargement by the accession of additional member states. Over the past forty years competencies have been continuously extended, in particular by the Single European Act in 1987 and the Maastricht Treaty on European Union in 1993.1 Since 1958 the process of integration has witnessed expansion from a Common Market of six to a Union of fifteen, with a population today of about 372 million. As integration progresses over the next decade it may eventually include more than twenty-seven members. This process, however, will have different implications for European countries. While both present member states and applicant countries expect to profit politically and economically from further integration, the incumbents fear that a wide enlargement will change the Union’s functioning. In this contribution our concern is enlargement, and we will theoretically assess the consequences of different accession scenarios on the Union’s effectiveness and member states’ efficacy. Most accession scenarios focus on the economic performance of the applicant countries (Baldwin 1994; Hagen 1997). Although all applicants from Eastern Europe have been given the status of a formal ‘Accession partnership’,2 the economic criteria of the Commission’s ‘Agenda 2000’ proposal alone are not sufficient to determine the exact set of new members. The accession of the most promising candidates – Hungary, Poland, the Czech Republic and Estonia from Eastern Europe, and Slovenia and Cyprus from Southern Europe – will depend on their meeting exogenous and endogenous criteria: the applicants for membership not only have to be able to adhere to the aims of political, economic and monetary union, but it was acknowledged early on that the European Union itself must also prepare the institutional conditions for ensuring its proper functioning. It is feared that an extensive enlargement will dilute the Union’s capacity to act when twenty or even twenty-seven member states have to reach consensus on European legislation. With regard to the Union’s effectiveness, even British Conservatives accepted the extension of member states’ majority voting in European legislation, but the debate is about what exactly will constitute a legislative majority. When member

The politics of institutional change

81

states extended the scope of Council of Ministers’ qualified majority voting in the mid-1980s – in order to complete the internal market project – the support of outvoted member states and their citizens had already been weakened. For this reason, the Amsterdam Treaty revision also aimed to increase common support by strengthening the role of the European Parliament in European legislation. In the end, member states considered the codecision procedure to be the best way to minimise the trade-off between Council qualified majority voting and parliamentary integration. However, the participation of an additional voting body is likely to decrease the Union’s effectiveness once again. Another important aspect of the Union’s functioning concerns the balance between the efficacy of large and small member states. While decreasing the Council voting quota to simple majority may increase the Union’s effectiveness – even under the codecision procedure – it would also reduce the difference between large and small member states in terms of their individual efficacy. For this reason, France and Spain would have liked to increase large countries’ voting weights, while the reunified Germany suggested applying a quorum solution, adding a population-weighted vote to the qualified majority voting quota. By and large, the debate on re-weighting caused considerable friction in the beginning. Both proposals were rejected at the Amsterdam Intergovernmental Conference. A major task of the Nice Intergovernmental Conference in December 2000 was, therefore, to resolve the institutional issues left open in Amsterdam that needed to be settled before enlargement. In the end, the Nice Treaty established a new distribution of weighted votes, increased the majority threshold and introduced two additional hurdles for decision-making, a member state and a population quorum (Felsenthal and Machover 2001). The following analysis examines the Union’s functioning under different reform proposals and for different steps of the enlargement process. It hereby confronts both aspects of institutional reform, the Union’s effectiveness concerning functional and parliamentary integration as well as member states’ efficacy in terms of re-weighting their votes and modifying the majority threshold. Like Lane and Mæland (in Chapter 7 of this volume) we measure the Union’s effectiveness under different procedures using the Coleman index, and we also use our inclusiveness index to outline the corresponding efficacy of member states (König and Bräuninger 1998). However, while Lane and Mæland stress the number of feasible winning coalitions when applying the non-normalised Banzhaf index, we do not find evidence for introducing Council simple majority voting. The re-weighting debate and the provisions of the Nice Treaty rather indicate that large member states are interested in maintaining their ability to preserve the status quo. We accordingly argue that member states consider not only the number of feasible winning coalitions but also their ability to block decisions, which means their performance to preserve the status quo. The remainder of this contribution is divided into four sections. In the following section we outline the reasons for parliamentary integration and the transformation from unanimity to weighted qualified majority voting in the Council. We discuss the two major approaches used to explain these changes, which became apparent during the mid-1980s. Next, we present our conception of

82

T. König and T. Bräuninger

legislative actors and inter-institutional winning coalitions in order to approach the member states’ choice of voting rules. With regard to further enlargement we argue that two waves of accessions are likely. We then introduce our concept of inclusiveness, referring to changes in effectiveness and efficacy. Finally, we draw our conclusion from enlargement scenarios with respect to modifications of the standard and codecision procedure.

From unanimity to qualified majority in multi-cameral settings Despite the formal provisions of the Treaty of Rome for Council majority voting, the Luxembourg Accords of 1966 – often called an intergovernmental agreement to disagree on majority voting – granted all member states veto power up to the mid-1980s (Kapteyn and Verloren van Themaat 1990: 249). In 1986 the Council stated in the Official Journal that forty legislative proposals had been adopted by qualified majority voting – tripling the total figure for 1985 (WQ 1121/86, C306/42). By the end of 1986, the Council increased the total figure to more than one hundred qualified majority decisions, in particular on single-market issues (Nugent 1994: 147). Compared to unanimity, Council qualified majority decision-making facilitates European legislation, but the remarkable change begs the question as to why the member states allowed more majority voting. Two particular events were important for the transition to qualified majority voting in the Council: first, the member states’ intention to adopt 282 measures for completing the internal market that had been documented in the 1985 White Paper; second, the so-called ‘southern enlargement’ with the accession of Greece in 1981 and of Spain and Portugal in 1986, which increased the socio-economic variation among the member states. Similar to the contemporary discussion about the accession of Eastern and Southern European countries, most member states were convinced of the need to increase the effectiveness of European decision-making by turning away from the restrictive unanimity rule. For functional integration the qualified majority criterion was applied in the Council, setting a voting threshold of about 71.2 per cent of all votes. According to the Nice Treaty’s Declaration on the qualified majority threshold and the number of votes for a blocking minority in an enlarged Union, this will move to a maximum of 73.9 per cent when all twelve candidate countries have acceded. In addition to increasing the Union’s effectiveness by lowering the voting quota, the provision for qualified majority voting also discriminates between member states according to the allocation of voting weights: in contrast to the one-state-one-vote principle, the Council qualified majority voting rule discriminates on member states’ efficacy by distinguishing between large and small member states. Voting weights are approximately related to the size of the member states’ population, although France and Germany initially preferred a weighting according to member states’ economic power.3 In studying member states’ choice of voting rules, one must take into consideration both the Union’s effectiveness and member states’ individual efficacy. Even though effectiveness has been increased by applying the qualified majority criterion, the realisation of the internal market project has provoked much

The politics of institutional change

83

criticism of European legislation. In Britain, European legislation has been attacked for its over-regulation, manic harmonisation and state subsidies, while the French criticise its liberalism, privatisation and laissez-faire economics. Citizens increasingly dislike the European in concert with national bureaucrats coordinating intergovernmental collaboration in internal market, agricultural and trade affairs without any parliamentary control of European legislation. Moreover, treaty reforms have to be ratified by national parliaments, which led to demands for parliamentary control of the government-biased European legislation. For these reasons, the role of the European Parliament was strengthened when the Council moved to qualified majority voting in the mid-1980s and, today, about twenty procedures apply to various policy areas. In order to increase transparency the fifteen members have agreed, in principle, to reduce the number and the complexity of European legislative procedures. Basically, we find four procedures in various modifications. According to the consultation procedure established by the Treaty of Rome in 1957, European legislation originates with the Commission, and the Council adopts the proposal by simple, qualified majority or unanimity, depending on the specific treaty provision. In the consultation procedure, the European Parliament gives its opinion, Commission and Council explain their opinion to the European Parliament, but they do not operate through Parliament. The cooperation and assent procedures were introduced in 1987 and the codecision procedure was established in 1993; the names of these procedures refer to parliamentary rights in European legislation. All of these procedures establish complex multi-cameral systems of European decision-making affecting the trade-off between the Union’s effectiveness and the efficacy of participating actors. This complexity has attracted many scholars applying various approaches to power distribution and agenda setting in European multi-cameral decision-making. For the study of multi-cameral decision-making, most approaches are more concerned with the choice within rules than with the choice of voting rules. Hence they tend to concentrate on the impact of either the differences between member states or the inter-institutional interaction of the Commission, the Council and the European Parliament on European legislation. Considering the actors’ different voting weights in the case of majority voting, relative voting power analyses have focused on actors’ relative ability of being decisive in forming winning coalitions. Their main findings concern the decreasing power of larger member states due to enlargements despite weighted qualified majority voting (Brams and Affuso 1985; Lane and Mæland 1995; Hosli 1996; Sutter 2000). Others emphasise parliamentary groups’ relative ability to put forth their policy preferences in European Parliament decision-making (Lane et al. 1995). Based on extensive form games, spatial analyses call the analytical usefulness of power index analysis fundamentally into question. One argument is that recent power index applications have ignored the strategic interaction between the Commission, the Council and the European Parliament in the passage of European legislation when focusing on the relative power distribution between member states in the Council (Garrett and Tsebelis 1996: 270). Except for Article 148.2b EC, all European legislative procedures require a Commission proposal

84

T. König and T. Bräuninger

that must ultimately be adopted by the member states with unanimity, qualified or simple majority (Steunenberg 1994). Another argument is that spatial studies take into account both actors’ policy preferences and strategic interaction between European voting bodies, while voting power indices ignore the spatial location of member states’ policy preference. Studying the choice within voting rules, spatial analyses outlined that the European Parliament is only granted conditional agenda-setting power in the cooperation procedure (Steunenberg 1994; Tsebelis 1994, 1996), but blocking power in the codecision (Schneider 1995) and assent procedures (Tsebelis and Garrett 1996). Both approaches to European decision-making, however, have shortcomings when we consider the choice of European voting rules. Voting power studies usually focus on either member states’ or parliamentary groups’ relative ability to be decisive in forming winning coalitions. Since actors have the same relative ability to be decisive under unanimity and unweighted majority voting, relative power indices do not take account of the distinction between majority and unanimity voting, which define the Union’s effectiveness. By contrast, spatial models require knowledge of procedural constraints, the location of actors’ policy positions, and of the status quo. However, constitutional choice precedes policy choice, and constitutional actors do not know ex ante the dimensionality and location of their policy positions, the status quo, or the other actors involved in future decision-making processes (Buchanan and Tullock 1962: 78). Constitutional actors are not certain of the exact content of prospective proposals, their exact number or their spatial location, which may favour or hinder majority building. This is the major difference between the choice within and the choice of voting rules. In order to approach member states’ constitutional choice of voting rules, we therefore focus on two aspects of their expectations: the effectiveness of decision-making and their individual efficacy. Member states take into particular consideration the passing of a Commission proposal, the decision probability in European legislation, the formation of a majority coalition in each voting body, their expected contribution to the Union’s effectiveness, etc. To study the effects of different voting rules on the Union’s effectiveness and member states’ efficacy, we must first identify the procedural settings and second, the actors of European legislative decision-making in the present and future Union.

Actors and winning coalitions of the present and future Union According to the 1996 Turin conclusions of the European Council, the subsequent institutional reforms at the Amsterdam and Nice Intergovernmental Conferences aimed to increase the effectiveness and transparency of European legislation, with the goal of increasing parliamentary legitimacy (Steunenberg 1997: 2). In order to raise parliamentary involvement, the modified codecision procedure (Conference of Representatives 1997: 122–4) proved to be the most promising voting procedure, requiring not only the consent of the European Parliament, but also giving the agenda-setting function to parliamentary delegates in the conciliation committee. Its effectiveness is increased by lowering the

The politics of institutional change

85

Council voting quota, thereby allowing for the exclusion of some member states’ policy positions. The Amsterdam Treaty proposes to apply the standard and the modified codecision procedures extensively, while the assent procedure should be used for legislation on external relations (Conference of Representatives 1997: 119). The assent procedure does not establish an agenda-setting function for the European Parliament, which can only adopt or reject a proposal. We investigate the standard and modified codecision procedures as ideal types to study the impact of institutional politics for reforming the Union’s framework. The purpose of procedural settings is to set up formal voting rules on winning or losing, which are essential in politics. The bi- or tricameral settings of the standard and codecision procedures complicate the identification of winners and losers because they combine voting bodies as collective veto players with internal coalition building processes in the Commission, the Council and the European Parliament. We conceive these internal coalition processes as subgames, which are combined on the compound level of a bi- or multi-cameral game. The bicameral standard procedure consists of the Commission and Council subgame, while the codecision procedure adds the European Parliament subgame. Their coalition building processes differ with regard to the number of subgame actors and the corresponding voting quota.4 In order to make enlargement effects comparable, we conceive of the European Parliament as a constant voting body consisting of two major and two minor parliamentary groups. Each major group is provided with two votes and each minor group has one vote to fulfil the absolute majority criterion.5 The procedural settings define the combination of subgames with all actors necessary to adopt a proposal. However, identifying these actors is rather complicated for two reasons: first, the Council’s and the European Parliament’s voting rules vary considerably.6 Second, the role of the Commission is rather speculative.7 We investigate the compound game in a bicameral setting requiring consent between the Commission and the Council under standard conditions. We investigate it as a semi-tricameral setting under the (modified) codecision procedure, between the Council, the European Parliament, and in part the Commission, as long as the European Parliament has no decisive structure.8 Because the codecision procedure includes a non-decisive European Parliament, an additional set of winning coalitions is introduced: the first case of the original codecision procedure encompasses the Commission, a qualified majority of Council votes and at least half of the parliamentary votes; the second case consists of the unanimous member states with at least the absolute majority of parliamentary votes. The modified codecision procedure introduced by the Amsterdam Treaty has changed this second case: it decreases the Council’s voting quota to a qualified majority. Hence, the combination of two cases of winning coalitions leads to a semi-tricameral system, since the Commission can be excluded from the second case under the codecision procedure. Whilst the Amsterdam Treaty has modified the interplay of organs, it postponed the issues of institutional reform of the Council of Ministers. Two proposals for reforming its legislative procedures have been debated. According to the French proposal, large member states should be allocated twenty-five votes,

86

T. König and T. Bräuninger

the smallest member states three votes, and the current voting quota should be retained.9 The German proposal suggested retaining the pre-Nice allocation of votes but introduced an additional quorum that should provide for the representation of at least 60 per cent of the Union’s total population. Both reform options have been rejected. The European Council finally succeeded at its Nice meeting in establishing a new distribution of weighted votes and two additional prerequisites for decision-making.10 Article 3.1a(i)–(ii) of the Nice Treaty’s Protocol on the Enlargement of the European Union (Conference of Representatives 2001) states that acts of the Council shall require for their adoption a qualified majority of weighted votes in favour cast by a majority of members. . . . When a decision is to be adopted by the Council by a qualified majority, a member of the Council may request verification that the member states constituting the qualified majority represent at least 62 percent of the total population of the Union. If that condition is shown not to have been met, the decision in question shall not be adopted.

Two waves of enlargement The sets of winning coalitions represent the cornerstone of our analysis of the present Union’s effectiveness and member states’ efficacy, but we still have to identify the set(s) of potential member states in a further enlarged Union. During the Copenhagen meeting of the European Council in 1993, a French proposal specified the formal requirements for membership, as repeated by the Commission’s Agenda 2000 and laid down in Article O of the Treaty on the European Union. This admission catalogue not only includes measurements of economic development and a functioning market economy, it also requires a quantifiable level of social protection, control over public debt and inflation, an open economy, a modern fiscal system and the administrative capacity to implement European legislation (Baldwin 1994: 155). The background is that, from today’s perspective, the accession of all applicants would increase the Union’s population by around 30 per cent but its GDP by only 4 per cent. Regarding the Union’s budget financing, structural funds and Common Agricultural Policies, the risks and costs of further enlargement pose a considerable problem not only for potential, but also for actual members spending 1.27 per cent of their GDP for European budget affairs (Streit and Voigt 1996). Europe’s agricultural sector and its poorer regions currently receive about 80 per cent of all European spending and, since the Eastern applicants are populous, poor and agricultural, the unchanged Common Agricultural Policies would increase the Union’s budget by nearly ECU 40 billion. Moreover, extending the structural funds would raise the annual costs by ECU 26 billion (Baldwin 1995: 477). Tax raising, however, is not a feasible policy to cope with these deficits; enlargements will most probably be accompanied by spending cuts. The eastward enlargement is therefore less favoured by southern, poor and agricultural member states, while northern incumbents expect gains from extending their trade relations (Michalski and Wallace 1992: 54).

The politics of institutional change

87

Under these conditions, two aspects are important for delineating the set of potential member states. First, Article 2 of the Draft Treaty’s protocol on the institutions (Conference of Representatives 1997: 118) restricts the number of European member states to twenty without institutional reform. Second, the economic and geopolitical situation of applicant countries not only has to take into account their deliverance from Russian hegemony but also the ongoing crisis in the Southern hemisphere. Therefore, we expect two waves of enlargement. The first wave will be limited to five new members, and the second wave will contain the remaining seven applicants. Of the first wave, only Slovenia is meeting the economic preconditions, but it is likely that Hungary, Poland and the Czech Republic will also make up the next group to achieve accession due to their geographical proximity and their institutional embedding in the Visegrad-4 group (Hagen 1997: 375). Compared to Latvia and Lithuania, with their Russian minorities, Estonia had started to advertise itself as less of a Baltic and more of a Scandinavian state, and we expect that it will be pushed as the fifth new member of the first wave by Sweden and Finland. In contrast to the Commission’s Agenda 2000 we do not consider Cyprus as an entrant in the first wave since its participation will presuppose Greek and Turkish support. Malta has just renewed its application and is also considered an inconstant candidate.11 Besides Latvia, Lithuania and Cyprus, Slovakia has already been excluded from the first wave, because of its unwillingness to guarantee minority rights for its Hungarian and Czech population. Moreover, Bulgaria and Romania will have to wait for the second wave, despite the French support for Romania’s membership. Since both waves will change the Union’s effectiveness, present members have made these future enlargements dependent on the Union’s ability to reform its institutional framework.

Considering the consequences of enlargement For Europe, enlargement is certainly an historic opportunity to end the artificial division of Europe, and therefore it is interesting to look for strategies that may more appropriately serve the cause of institutional reform of future European legislation. In the following we argue that two aspects have to be considered: the Union’s effectiveness and current member states’ efficacy. The Union’s effectiveness in decline In conducting an analysis of the Union’s institutional reform, we are concerned with problems about how to measure the prospective change of effectiveness and efficacy by further accessions. Theoretically, the effectiveness of decision-making is determined by two components: the number of participants and the voting rule applied. Unanimity restricts the decision probability of an n-actor committee to the single favourable winning coalition of all feasible 2n coalitions, whereas simple majority permits about half of all coalitions to make binding decisions. Due to different membership size and voting quotas of the Union’s voting bodies, we measure effectiveness using decision probability. According to

88

T. König and T. Bräuninger

Coleman (1971: 278), we define the decision probability of a voting rule as the ratio between the number of winning coalitions (able to alter the status quo) and all feasible coalitions.12 Applied to the Union’s procedures, the decision probability in the standard procedure involving the Commission and a unanimous Council decreased from 0.0078 for the original six to almost 0.0001 for the twelve member states. This fundamental change indicates why member states decided to decrease the Council voting quota in order to complete the internal market up to the end of 1992. In the mid-1980s, when the twelve member states extensively started to apply qualified majority voting in the standard procedure, their decision probability was higher than ever before. In the following we use this value as the baseline value for comparison of decision probability. Table 5.1 shows the changes in comparison to the decision probability of the standard procedure under Council qualified majority voting in 1986.13 The columns show changes due to past and future enlargements as well as different reform options of redistributing member states’ voting weights, which have been discussed at the Amsterdam and the Nice Intergovernmental Conferences. Compared to 1986, with twelve members, the accession of new members represents a continuous decrease of the decision probability to about 79 per cent in 1995, 47 per cent with the first and 29 per cent with the second wave of enlargement under qualified majority voting if the old, pre-Nice weighting of member states’ votes were retained. Under unanimity decision-making becomes rather unlikely. The accession of Austria, Finland and Sweden in 1995 already limited the Union’s effectiveness, but forthcoming enlargement will almost reverse the positive effect of the transition from unanimity to Council qualified majority voting. The accession of Poland with eight votes, Hungary and the Czech Republic with five votes each, and Estonia and Slovenia with two votes each would halve present decision probability. With the accession of Romania with seven votes, Bulgaria with four votes, Lithuania and Slovakia with three votes each, Cyprus, Latvia and Malta with two votes each, decision probability would again be reduced by almost the same amount. While decision probability in the original codecision procedure, as a result of the additional parliamentary inclusion, was always lower than in the standard procedure, the modified codecision procedure increases the Union’s effectiveness by reducing the Council’s voting quota. Its decision probability is similar to that of the consultations procedure with qualified majority voting. If parliamentary participation will indeed raise common support, the modified codecision procedure may presently guarantee proper functioning because the positive effect of Council majority voting on the Union’s effectiveness will not be reversed by the inclusion of the European Parliament. However, enlargement would considerably decrease the Union’s overall effectiveness. The dramatic decline of decision probability indicates the need for institutional reform. For this purpose, the preparatory documents for the Amsterdam summit offered two proposals that were finally rejected. Applying these proposals to an enlarged Union of twenty members, the French proposal would have provided more frequent policy change, shifting decision probability to 60 per cent in the consultation and 37 per cent in the original codecision

–

Codecision (original)

Codecision (modified)

0.79

0.50

0.0003

0.79

0.47

0.30

0.00001

0.47

0.47

0.46

0.60

0.37 0.47

0.30 0.46

–

0.00001 0.00001 0.00001

0.60

Pre-Nice French German Post-Nice provisions proposal proposal provisions

First wave (20 states)

0.38

French proposal

0.29

German proposal

0.21

Post-Nice provisions

0.29

0.18 0.38

0.25

0.21

–

Weighting of votes and thresholds in the Council: • Pre-Nice provisions, 1986–95: France (10), Germany (10), Italy (10), United Kingdom (10), Spain (8), Belgium (5), Greece (5), Netherlands (5), Portugal (5), Denmark (3), Ireland (3), Luxembourg (2); 1995–: Austria (4), Sweden (4), Finland (3), threshold: 54/76 weighted votes; first wave: Poland (8), Czech Republic (5), Hungary (5), Estonia (2), Slovenia (2), threshold: 62/87 weighted votes; second wave: Romania (7), Bulgaria (4), Lithuania (3), Slovakia (3), Cyprus (2), Latvia (2), Malta (2), threshold: 94/132 weighted votes. • French proposal, first wave: France (25), Germany (25), Italy (25), United Kingdom (25), Poland (20), Spain (20), Netherlands (12), Belgium (10), Czech Republic (10), Greece (10), Hungary (10), Portugal (10), Austria (8), Sweden (8), Denmark (6), Finland (6), Ireland (6), Estonia (3), Luxembourg (3), Slovenia (3), threshold: 174/245; second wave: Romania (16), Bulgaria (8), Lithuania (6), Slovakia (6), Cyprus (3), Latvia (3), Malta (2), threshold: 206/290. • German proposal: pre-Nice provisions plus additional quorum of at least 60 per cent of total population. • Post-Nice provisions: France (29), Germany (29), Italy (29), United Kingdom (29), Poland (27), Spain (27), Netherlands (13), Belgium (12), Czech Republic (12), Greece (12), Hungary (12), Portugal (12), Austria (10), Sweden (10), Denmark (7), Finland (7), Ireland (7), Estonia (4), Luxembourg (4), Slovenia (4), threshold: 215/296 weighted votes of at least 11/20 states representing at least 62 per cent of total population; second wave: Romania (14), Bulgaria (10), Lithuania (7), Slovakia (7), Cyprus (4), Latvia (4), Malta (3), threshold: 255/345 weighted votes of at least 14/27 states representing at least 62 per cent of total population.

0.29

0.18

0.00000010 0.0000001 0.0000001 0.0000001

0.29

Pre-Nice provisions

Second wave (27 states)

Note: Decision probability of 1986 Consultation procedure (with Council qualified majority) is baseline (P=0.0491).

0.002

0.63

Consultation/unanimity

1.00

Pre-Nice provisions

Pre-Nice provisions

Consultation/qualified majority

1995 – (15 states)

1986 – 95 (12 states)

Table 5.1 Decision probability in the codecision and consultation procedure compared to 1986 baseline

Chapter Title 89

90

T. König and T. Bräuninger

procedure as compared to the 1986 baseline. With the second enlargement, however, the decision probability would again decrease to 38 per cent and 25 per cent, respectively. Re-weighting, in particular with parliamentary participation, cannot guarantee a proper functioning of European legislation. The German proposal of an additional quorum would have had the same effects on decision probability as the pre-Nice provisions and would not have provided a means to speed up the legislative processes either. The Nice Treaty introduces even higher hurdles for decision-making in the Council, namely a member state and a 62 per cent population quorum. As a result the post-Nice provisions will decrease decision probability in both consultation and the (modified) codecision procedure. Due to the increase of the threshold in the pace of enlargement, this loss in efficiency will be highest in a Union of twenty-seven member states. In sum, the French proposal would have slightly reduced the gridlock danger, but no significant effort in securing the Union’s functioning can be made as long as the 71.2 per cent Council voting quota prevails. The German proposal would have had no substantial effect, whereas the post-Nice provisions will even increase the danger of legislative gridlock. However, a lowering of the voting quota would fundamentally change the current distribution of member states’ efficacy, as established by their different voting weights. Member states’ efficacy in change In order to provide enough room for manoeuvre in case of enlargement, there seems to be only one way of reforming the procedural settings: decision probability can be raised by decreasing the Council voting quota. The question is what should constitute a majority and/or whether or not a new distribution of voting weights should be used to balance the existing differences between member states. Under unanimity all member states have the same efficacy on European decision-making, while weighted majority voting puts larger member states at an advantage. When voting weights establish differences between member states’ efficacy, some are more likely to be in a majority position than others. Decreasing the voting quota will certainly modify this distribution. In order to show this effect, we analyse the changes of individual inclusiveness compared to the 1986 baseline model. According to our index of decision probability we measure the efficacy of an actor by the inclusiveness, i.e. as the number of times an actor participates in winning coalitions in relation to the number of all feasible winning coalitions (Bräuninger 1996: 42).14 If an actor can be excluded from any feasible winning coalition, the inclusion of its policy position is determined only by luck. Thus, the actors holds a so-called dummy player position and its efficacy is 0.5, the smallest possible value. By contrast, if an actor cannot be excluded from any feasible winning coalition the actor is a veto player with the maximal efficacy of 1.0. Table 5.2 shows the changes by lowering the voting quota with respect to member states’ efficacy after the second wave of enlargement. In 1986, the efficacy of large member states was about 0.856, whereas Luxembourg was still

0.627

Denmark, Finland, Ireland, Lithuania, Slovakia

0.500

–

Austria, Bulgaria, Sweden

Parliament group with 1 vote

0.684

Belgium, Czech Republic, Greece, Hungary, Portugal

0.550

0.684

Netherlands

0.500

–

Romania

Parliament group with 2 votes

0.801

Spain, Poland

Cyprus, Estonia, Luxembourg, Latvia, Malta, Slovenia

1.000 0.856

France, Germany, Italy, United Kingdom

0.625

0.750

0.583

0.624

0.663

0.704

0.703

0.776

0.811

0.870

0.625

0.625

0.750

0.551

0.601

0.634

0.667

0.699

0.761

0.818

0.880

0.625

Codecision (modified) Pre-Nice French provisions proposal

Consultation Pre-Nice provisions

Commission

Second wave (27 states)

1986–95 (12 states)

0.625

0.750

0.583

0.624

0.663

0.704

0.703

0.776

0.811

0.870

0.625

German proposal

0.625

0.750

0.565

0.613

0.660

0.691

0.706

0.721

0.885

0.903

0.625

Post-Nice provisions

0.625

0.750

0.528

0.542

0.556

0.570

0.570

0.599

0.614

0.644

0.625

Pre-Nice provisions with simple majority

Table 5.2 Individual efficacy in the consultation and modified codecision procedure under different settings (inclusiveness indices)

Chapter Title 91

92

T. König and T. Bräuninger

provided with a value of 0.550. Both parliamentary actors had a dummy player position in the consultation procedure and were only included in majority building by chance. Compared to the baseline values we find that the two Amsterdam reform proposals and the pre-Nice weighting under qualified majority would have modestly redistributed present member states’ efficacy in the case of a Union of twenty-seven. The efficacy value of larger member states like France, Germany, Italy and the United Kingdom would increase modestly; Spain and Poland, and particularly the Netherlands would also profit from these versions, while both Amsterdam reform proposals hardly intended to decrease the efficacy of all medium-sized states. By privileging larger member states, the French proposal would have slightly increased the asymmetry between large and small members. This proposal for institutional reform can be considered a trade-off for the small (northern) members’ goal of an early Eastern enlargement, while the German proposal would have no redistributive effects. Larger medium-sized states like Belgium or Greece and small states, like Luxembourg, would have even profited from the pre-Nice settings and the German proposal. But there is no significant difference between pre-Nice settings and the two Amsterdam options, regarding either member states’ efficacy or the Union’s effectiveness. This means that none of these provisions would have provided a way out of improper functioning. This changes, however, if we move to post-Nice provisions. Figure 5.1 shows the individual efficacy in the modified codecision procedure under the pre- and

.

.

.

.

.

Figure 5.1 Consequences of reform options and the Nice reform on member states’ inclusiveness.

The politics of institutional change

93

post-Nice provisions, the French and the German proposals. The results suggest that the Nice reform benefits large member states at the expense of smaller incumbents and most (rather small) accession countries. Compared to the preNice setting, the big four as well as Spain and Poland will increase their efficacy by 15 per cent while all other member states will lose. The sharp divide is between Spain and Romania, the efficacy of the latter is quite similar to that of the Netherlands. The last column of Table 5.2 shows member states’ efficacy on simple majority voting in the Council but without any modifications to the pre-Nice voting weights. Lowering the quota would have accordingly raised the decision probability in the codecision procedure about ten times higher than the 1986 baseline value of qualified majority voting in the consultation procedure. Nevertheless, simple majority would reduce large member states’ efficacy to about 0.644, while small member states would have to bear only modest losses as compared to the 1986 distribution. Assuming that the baseline value indicates an acceptable trade-off between effectiveness and efficacy, we cannot expect member states to adopt this option, since they would become much more similar under simple majority rule.

Conclusion The accession of Eastern and Southern European countries presupposes reform of the Union’s institutional framework; originally developed by the six founding states to regulate policies as positive-sum games. New accessions will change both the effectiveness of the European decision-making system and the efficacy of member states. Today’s member states have therefore looked for procedural settings to preserve proper functioning with respect to both the Union’s capacity to act and their influence on European decision-making. With regard to the Union’s effectiveness, even British Conservatives accepted the extension of member states’ majority voting in European legislation, but the debate remained about what exactly constitutes a legislative majority. Reforming the voting rules by lowering the voting quota and re-weighting has caused considerable friction. Since all previous proposals have been rejected at the Amsterdam conference, member states had postponed the Union’s institutional reform and finally agreed on a re-weighting of votes at the Intergovernmental Conference, held at Nice in December 2000. Our results suggest that the Nice reform heavily reflects the interests of the large incumbents, in particular of France and Germany which provided themselves with the power to block about 90 per cent of all decisions. Since the efficacy of all other member states is reduced rather modestly, the large member states’ increase also comes at the expense of the Union’s effectiveness. This poses a serious threat to the proper functioning of the European Union. From our analysis it is clear that neither the French nor the German reform proposal, nor the post-Nice setting guarantee proper functioning as long as member states fail to accept a lower Council voting quota, particularly in the event of further enlargement.

94

T. König and T. Bräuninger

Notes 1 For an overview, see König (1996: 556). 2 EC 622/98, OJC L85, 20/03/98, 1–2. 3 See Garrett (1992: 546) for a discussion about the distribution of voting weights. See also Lane et al. (1995: 395), and Widgrén (1995: 78). 4 Concerning the coalition building in the Commission subgame, we argue that each Commissioner is provided with his or her own portfolio, carries the main leadership responsibility, and is independent of the Commission President in determining how to act on European legislative decisions. For this reason, we conceptualise the Commission as a unitary actor in European legislation with the responsible Commissioner as its agent (see also Spence 1994: 92; Westlake 1994: 9). In the Council, the governments of the member states are represented by delegates mediating between their own governments and those of other delegates (Johnston 1994: 27). National governments instruct their delegates, who then cast their votes homogeneously in the Council (Sabsoub 1991: 40). In the European Parliament the political group affiliation of parliamentary representatives has proved to dominate coalition formation so that the political groups can be conceived as European Parliament entities with weighted votes. 5 See König (1997: 40), for values of parliamentary groups. Furthermore, simplifying the European Parliament this way does not obstruct our findings. First, effectiveness is rarely affected by the number of actors when an absolute majority rule is applied, and second, member states’ individual efficacy is invariant when the European Parliament is added as an additional voting body. 6 The Council subgame offers two majority criteria under the standard and codecision procedures, since amendments always require unanimity among member states. Under the codecision procedure, the European Parliament may take action or no action. Preventing endorsement by no action slightly decreases the majority criterion, since the European Parliament has always been a voting body consisting of an equal number of representatives. 7 According to Article 155 TEU, the Commission holds the exclusive right to initiate legislation and the right to modify a proposal at any point of procedure (Article 189a.2 TEU), thereby making the Commission the agenda setter. Moreover, the Commission also has the right to withdraw, if the proposal’s object is felt to be emasculated by amendments (Usher 1994: 148). By contrast, Tsebelis and Garrett (1996: 13) argue that the Commission loses its agenda-setting power under the modified codecision procedure. 8 A voting body is decisive when the complement of a losing coalition is a winning one and vice versa. 9 See Consolidated Draft Treaty Texts (1997: 113) and footnotes of Table 5.1. 10 See footnotes of Table 5.1. 11 The application of Malta was frozen between October 1996 and September 1998. 12 Formally, we represent the legislative procedure as a simple game, i.e. a map v where for any coalition S of the player set N, v(S) = 1 if the coalition S is winning, and v(S) = 0 otherwise. Hence we assume that actors do not know their future preferences and, therefore, consider all Yes and No votes to be equally probable. Under these assumptions, the decision probability P(v) is

P( v) =

∑v(S)

S⊆ N

2n

In a committee of four actors, for example, 24 = 16 coalitions of supporting actors are conceivable. Applying unanimity the decision probability is 1/16 (0.06) because only the grand coalition of all actors may change the status quo. Under simple majority the decision probability increases to 5/16 (0.31), since the possible exclusion of a single actor offers a total of five winning coalitions.

The politics of institutional change

95

13 All indices are obtained using the program IOP 2.0 (Bräuninger and König 2001). 14 Formally, the inclusiveness of actor i in the game v is

ω i ( v) =

∑v(S)

S⊆ N , i∈S

∑v(S)

S⊆ N

In a four-actor-committee the inclusiveness of each actor is 1.0 under unanimity. Applying simple majority, the individual inclusiveness decreases to 4/5 (0.80) since each actor can be excluded from one out of five winning coalitions. Increasing the number of the committee to six members, inclusiveness of each actor is still 1.0 under unanimity, but it decreases to 16/22 (0.73) since each actor can be excluded from 6 out of 22 winning coalitions.

References Baldwin, R. (1994) Towards an Integrated Europe, London: Centre for Economic Policy Research. Baldwin, R.E. (1995) ‘The Eastern enlargement of the European Union’, European Economic Review 39: 474–81. Brams, S.J. and P. Affuso (1985) ‘New paradoxes of voting power on the EC Council of Ministers’, Electoral Studies 4: 135–39. Bräuninger, T. (1996) ‘Die Modellierung von Entscheidungsverfahren internationaler Organisationen am Beispiel der Meeresbodenbehörde’, Mannheim: unpublished thesis. Bräuninger, T. and T. König (2001) Indices of Power IOP 2.0 [computer program], Konstanz: University of Konstanz [http://www.uni-konstanz.de/FuF/Verwiss/koenig/IOP.html]. Buchanan, J.M. and G. Tullock (1962) The Calculus of Consent: Logical Foundations of Constitutional Democracy, Ann Arbor: University of Michigan Press. Coleman, J.S. (1971) ‘Control of collectivities and the power of a collectivity to act’, in B. Lieberman (ed.) Social Choice, New York: Gordon and Breach, 269–99. Conference of the Representatives (1997) Draft Treaty of Amsterdam. Conference of the Representatives of the Governments of the member states, 19 June 1997, Brussels (CONF/4001/97). Conference of the Representatives (2001) Draft Treaty of Nice. Conference of the Representatives of the Governments of the member states, 30 January 2001, Brussels (SN/1247/01EN). Consolidated Draft Treaty Texts (1997) 30 May 1997, Brussels (SN/600/97). Felsenthal, D. and M. Machover (2001) ‘The treaty of Nice and qualified majority voting’, Social Choice and Welfare 18: 431–64. Garrett, G. (1992) ‘International cooperation and institutional choice’, International Organization 46: 533–60. Garrett, G. and G. Tsebelis (1996) ‘An Institutional Critique of Intergovernmentalism’, International Organization 50: 269–99. Hagen, J. von (1997) ‘Wirtschaftspolitische Aspekte der Osterweiterung der EU’, in T. König, E. Rieger and H. Schmitt (eds) Europäische Institutionenpolitik, Frankfurt a.M.: Campus (Mannheimer Jahrbuch zur Europäischen Sozialforschung, vol. II), 380–96. Hosli, M.O. (1996) ‘Coalitions and power: effects of qualified majority voting in the Council of the European Union’, Journal of Common Market Studies 34: 255–73.

96

T. König and T. Bräuninger

Johnston, M. (1994) The European Council, Oxford: Westview. Kapteyn, P. and P. Verloren van Themaat (1990) Introduction to the Law of the European Communities, Deventer: Kluwer (2nd edition). König, T. (1996) ‘The constitutional development of European integration’, Journal of Theoretical Politics 8: 553–9. König, T. (1997) Europa auf dem Weg zum Mehrheitssystem. Gründe und Konsequenzen nationaler und parlamentarischer Integration, Opladen: Westdeutscher Verlag. König, T. and T. Bräuninger (1998) ‘The inclusiveness of European decision rules’, Journal of Theoretical Politics 10: 125–41. Lane, J.-E. and R. Mæland (1995) ‘Voting power under the EU constitution’, Journal of Theoretical Politics 7: 223–30. Lane, J.-E., R. Mæland and S. Berg (1995) ‘The EU Parliament: seats, states and political parties’, Journal of Theoretical Politics 7: 395–400. Ludlow, P. (1991) ‘The European Commission’, in R.O. Keohane and S. Hoffman (eds) The New European Community: Decision Making and Institutional Change, Boulder, CO: Westview, 85–132. Michalski, A. and H. Wallace (1992) The European Challenge of Enlargement, London: Royal Institute of International Affairs. Nugent, N. (1994) The Government and Politics of the European Union, Houndmills: Macmillan (3rd edition). Sabsoub, J.-P. (1991) The Council of the European Community, Luxembourg: Office for Official Publications of the European Communities. Schneider, G. (1995) ‘The limits of self-reform: institution-building in the European Union’, European Journal of International Relations 1: 59–86. Shepsle, K.A. (1986) ‘Institutional equilibrium and equilibrium institutions’, in H.F. Weisburg (ed.) Political Science: The Science of Politics, New York: Agathon, 51–82. Spence, D. (1994) ‘Structure, functions and procedures in the Commission’, in G. Edwards and D. Spence (eds) The European Commission, Essex: Longman, 92–116. Steunenberg, B. (1994) ‘Decision making under different institutional arrangements: legislation by the European community’, Journal of Institutional and Theoretical Economics 150: 642–69. Steunenberg, B. (1997) ‘Reforming legislative decision making in the European Union: an analysis of the codecision procedure and its alternatives’, Enschede: University of Twente, unpublished manuscript. Streit, M.E. and S. Voigt (1996) ‘Toward ever closer union – or ever larger? Or both?’, in D. Schmidtchen and R. Cooter (eds) Constitutional Law and Economics of the European Union, Cheltenham: Elgar, 223–47. Sutter, M. (2000) ‘Fair allocation and re-weighting of votes and voting power in the EU before and after the next enlargement’, Journal of Theoretical Politics 12: 433–49. Tsebelis, G. (1990) Nested Games, Berkeley: University of California Press. Tsebelis, G. (1994) ‘The power of the European Parliament as a conditional agenda setter’, American Political Science Review 88: 128–42. Tsebelis, G. (1996) ‘More on the European Parliament as a conditional agenda setter: response to Moser’, American Political Science Review 90: 839–44. Tsebelis, G. and G. Garrett (1996) ‘Agenda setting power, power indices and decision making in the European Union’, International Review of Law and Economics 16: 345–61. Tsebelis, G. and G. Garrett (1997) ‘Why power indices cannot explain decision making in the European Union’, in D. Schmidtchen and R. Cooter (eds) Constitutional Law and Economics of the European Union, Cheltenham: Elgar, 11–31. Usher, J. (1994) ‘The Commission and the law’, in G. Edwards and D. Spence (eds) The European Commission, Essex: Longman, 146–68. Westlake, M. (1994) The Commission and the Parliament, London: Butterworths. Widgrén, M. (1995) National Interest, EU Enlargement and Coalition Formation, Helsinki: The Research Institute of the Finnish Economy.

Chapter Title

6

97

An even wider Union The effects of enlargement on EU decision-making Bernard Steunenberg1

Introduction During the negotiations for the Amsterdam Treaty (1997) as well as the more recently adopted Treaty of Nice (2001) the main actors in the European Union expressed significant concern that the entry of new member states might reduce the effectiveness of the Union’s institutions.2 The possible institutional implications of enlargement are clearly expressed in a report submitted by Jean-Luc Dehaene, Richard von Weizsäcker, and David Simon on the invitation of Commission President Prodi. In this report, they state that ‘[a] significant increase in the number of participants automatically increases problems of decision-making and management’ (Dehaene et al. 1999: 6). More members of the Council, for instance, would increase the risk of deadlock if unanimity rule were to be used, reducing the Union’s capacity to respond to policy problems and to take decisions. The ‘three wise men’ suggested, among other things, reducing the use of unanimity rule and expand the use of majority voting prior to enlargement. However, the claim that institutional reform is a condition for enlargement can be questioned. Earlier enlargements of the Union were accompanied by debates on institutional reform, yet member states did not substantially change their decision-making rules and procedures.3 Why then, for the next enlargement, should enlargement and institutional reform be seen as connected issues? In this chapter I argue that the current member states can employ enlargement as a strategy of renegotiation in order to improve their own position. In other words, enlargement can be understood as an opportunity for the current member states to bargain for ‘new deals’ in the Union, hence the enormous importance of the internal dimension of enlargement. Thus, the requirement that applicant countries must satisfy the Copenhagen criteria is a necessary, but not a sufficient condition for their entry. The other necessary condition is a renegotiation of the Treaties that satisfies the current member states. This relates to an observation made by Christopher Preston (1997: 21), who suggests that ‘[e]xisting member states use the enlargement process to pursue their own interests and collectively to externalise internal problems’. In contrast to this claim the bulk of the scholarly literature on European integration echoes the potential ‘threats’ of enlargement as presented by the main actors in the Union. In particular, analyses using voting power indices point to the problem that more

98

B. Steunenberg

member states will lead to an increase in the number of instances in which the Union cannot take a decision and remains indecisive.4 The main concern emerging from this literature is that a larger Council will dramatically reduce the extent to which member states are able to introduce new legislation, while neglecting the underlying dynamics. My analysis is informed by the constitutional political economy perspective, in that I regard the rules and procedures on which the member states have to decide as game forms.5 These game forms define the actors that are involved, the actions or strategies available to these actors, the voting rules they have to use, and the sequence in which the actors may make a choice. The choice of a specific game form depends on how these rules and procedures are perceived to benefit the deciding actors given the fact that they will be used for a multitude of different policy issues. My comparative analysis based on game forms supports the claim that the relationship between enlargement and institutional reform is not so clear. The ‘risks of deadlock’ are not typical for an enlarged Union, while existing member states could already benefit from introducing more qualified majority voting. The analysis also shows that the changes introduced at Nice preserve the position of the larger member states after enlargement at the expense of the smaller ones and most of the applicant countries. Here enlargement clearly provided an opportunity for change, which otherwise would not have been possible. To develop my argument I start with an analysis of voting in the Council and discuss the implications of a change from unanimity to qualified majority rule within the context of the current Union. These implications concern two different aspects of the Union’s decision-making process, that is, the overall ability to respond to perceived policy problems and to take decisions in the Union as well as the ability of individual member states to affect these decisions. Second, I compare these results with the effects of enlargement on decision-making in the Union. Subsequently, I will show what the effects are of the new qualified majority rule adopted in the Treaty of Nice, which has been introduced to ‘allow’ for enlargement. Fourth, I will discuss the possible effects of different distributions of preferences for my analysis and the possible involvement of the European Parliament in the decision-making process. I conclude this contribution with an assessment of the prospects for accession.

The nature of decision-making: analytical framework In order to capture the essence of EU decision-making, I will employ two different game forms. These game forms reflect, in a rather simple and straightforward way, the dynamics of legislative decision-making in the Union. In the first basic game form, the Commission proposes a new policy to the Council, which is accepted by either unanimity or qualified majority. The second game form includes the European Parliament. In this case, the Council’s vote in favour of a new policy requires the approval by Parliament. Of course, this structure does not resemble the more complicated codecision procedure, but it reveals some of the basic properties of European Union decision-making, which are important for a

The effects on EU decision-making

99

Figure 6.1 The game trees for both legislative game forms.

discussion of the effects of enlargement. The game trees for both basic policy games are presented in Figure 6.1. The next step is to focus on the preferences of the players. I assume that players have Euclidean preferences over policies, which maintain two distinct properties. First, each player has a most preferred policy position or ideal point in a one-dimensional policy space. Second, a player’s preference changes with distance in the sense that the greater the distance between a player’s ideal point and some outcome, the smaller are the benefits or payoffs to this player, and the less preferred this outcome is. When the players know the specific preferences of the other players and the structure of the game, both sequential games can be solved using the equilibrium concept of subgame perfectness. However, a complication arises when we realise that decision-making procedures are used for a large number of different issues. Specifically, when we would like to assess the effects of enlargement on the decision-making capabilities of the Union, it seems to be more reasonable to assume that players will be uncertain about the policy issues that may arise in the future. No one will be able to indicate what may happen in various policy areas covered by the Union in the coming years. Players’ preferences will not be fixed and may vary per issue. Furthermore, the state of affairs prior to the start of a decision-making process – the status quo – may vary as well. I will call these two elements – players’ preferences and the status quo – the policy state of a decision-making process. An important insight from the formal analysis of political decision-making is that the contents of an equilibrium policy, if one can be identified, strongly depends on the policy state (Romer and Rosenthal 1978, 1979; Shepsle 1979, 1989; Riker 1980; Shepsle and Weingast 1981; Ostrom 1986). A minor change, either in players’ preferences, or the status quo point, may have a substantial impact on the position of the equilibrium policy in the outcome space. Due to this sensitivity, the characteristics of procedures cannot be analysed at the level of individual policies, but have to be assessed at a systemic or constitutional level.6 For this purpose, I will use a distribution of policy states to represent the uncertainty players have with regard to their preferences and the status quo. Based on a distribution, a game can be solved for each of the policy states described by the distribution. Using these solutions, I am able to determine various key

100

B. Steunenberg

characteristics of these game forms. First, the proportion of specific solutions that will be reached in equilibrium can be determined. These proportions provide us with a probability that some outcome may arise. Second, the role of individual players can be assessed by focusing on the mean distance between the equilibrium outcome and a player’s ideal point. The mean distance provides us with an expected value about how well a player is doing in a specific context. By comparing these expected values for different game forms – for example, for decisionmaking processes with fifteen or twenty-seven possible EU members – I am able to evaluate the consequences of enlargement to different players. The shorter the expected distance, the more beneficial a specific game form is to a player. Together with Dieter Schmidtchen and Christian Koboldt, I have suggested an alternative power index for non-cooperative games as a way to compare these expected values for different games (see Steunenberg et al. 1999). The comparison is based on the introduction of an external observer who does not play a role in the game and therefore can be regarded as a ‘powerless’ player. The resulting index, which I will use here, relates the expected values of both a player in a game and the external observer. The index – the so-called strategic power index – is restricted to the [0,1]-interval. If an actor has a value of one, it can be regarded as a dictator, since the outcome of the game will always be similar to this actor’s ideal point. If the index has a value of zero, the actor is similar to the observer and thus has no power at all.7 A second index, which is related to the strategic power index, measures the status quo bias of a game form. This bias is regarded as the extent to which players are unable to act and to pull a new policy away from the current state of affairs. The so-called inertia index relates the expected distance between the status quo and the equilibrium outcome to the value found for the external observer. A value of one for this index means that under some procedure the status quo always prevails. The smaller the value for the index, the more players are able to move the equilibrium policy away from the status quo and the more a procedure allows for flexibility in policymaking.8

Voting in the Council of the European Union The first step in the analysis is to focus on voting in the Council of the European Union (the ‘Council of Ministers’). For this purpose I start with the basic game between the Commission and the Council, in which the Council has to approve a new policy by unanimity or qualified majority. In this contribution I focus on both voting rules as rules to decide on policy issues. This is also the main focus of the reform discussion in the European Union. The major concerns as expressed in the introduction refer to decision-making in the Council of the European Union, which has to decide on the continuation or change of existing policies, or the introduction of new policies within the jurisdiction of the Union as set by the treaties. Determining the appropriate decision rule for constitutional change such as change in the rules for collective decision-making in the Union, the Union’s jurisdiction, or its power to tax, is a different issue. Here insights from constitutional economics and politics indicate that the choice of rules for collective

The effects on EU decision-making

101

decision-making has to be based on the voluntary consent of the constituting individuals (see, for instance, Buchanan and Tullock 1962; Brennan and Buchanan 1980, 1985). Based on this notion, unanimity rule seems to be the only option for constitutional choice, that is, decision-making on the rules for collective choice. Applied to the European Union, unanimity should serve as a rule for the member states’ decision to decide on whether or not to transfer some of their sovereignty to the Union. For example, unanimity rule might be applied to transfers of new policy areas to the Union by including these areas in the treaties, to changes of the Union’s decision-making rules, the Union’s power to tax if the member states agree to grant the Union this right, or the admission of new member states to the Union. However, within the context of a constitutionally grounded set of rules, decision-making need not to be restricted to unanimity voting only. Here, other decision-making rules may apply, including simple majority rule, in order to reduce costs that result from indecision.9 Key characteristics of the basic policy game in which the Council decides by unanimity or qualified majority have been computed using simulations.10 In these simulations, the policy states, that is, the preferences of the Commission and the member states and the status quo-points, were drawn from a onedimensional space following a uniform or normal distribution. Furthermore, distance is measured by simple Euclidean distance (see Enelow and Hinich 1984: 15). The results relate to two different aspects of the decision-making process. First, I present results on the current member states’ ability to take decisions, which refer to the discussion on the flexibility of policymaking in the Union.11 In order to appreciate the claim that more member states would make the Union unmanageable, we need at least to know how flexible the current Union is. Second, I will present results on how these different rules affect the individual member states’ ability to determine the outcome of the Union’s decision-making process. Policy flexibility in the current Union of fifteen member states The results on the ability to take decisions in the current Union are presented in Table 6.1. First, the table contains the proportion of solutions that will be reached in equilibrium and differ from the status quo. These proportions represent the probability that a new policy will be adopted in the Commission–Council game. Table 6.1 Comparative analysis of Council voting rules: probabilities and inertia index for the Commission–Council game Voting rule

Unanimity

QMV

Distribution of policy states

Uniform

Normal

Uniform

Normal

Expected policy choice: • current policy (status quo) • new policy

90% 10%

41% 59%

51% 49%

17% 83%

Inertia index

0.97

0.39

0.54

0.08

102

B. Steunenberg

Second, the inertia index is presented, which indicates the extent to which the status quo prevails as the outcome of the decision-making process. As the scores in the table show, the change from unanimity rule to qualified majority voting has a substantial and positive impact on the probability of choosing a new policy. This effect has been reported in various analyses using voting power indices (see, for instance, Hosli 1993; Berg and Lane 2001) and it forms the analytical basis of the suggestion to introduce more qualified majority voting in the Union. Using a uniform distribution, the inertia index drops from 0.97 to 0.54 when the member states decide to use qualified majority voting instead of unanimity voting. Similarly, the probability that the outcome of the decisionmaking process remains the status quo drops from 90 per cent to 51 per cent. These results do not depend on the specific distribution of policy states. If we move to a normal distribution for which all players have the same modus, the result is the same: decision-making in the Union allows for more change by introducing qualified majority voting. Power of the individual players The effects of more qualified majority voting to the individual players are summarised in Table 6.2. This table contains the strategic power scores for each player under both voting rules. For a uniform distribution the power scores increase when players decide to use more qualified majority voting. This effect is most substantial for the larger member states, which will gain most. They will experience a change in their power scores from 0.007 for unanimity voting to Table 6.2 Expected benefits of different voting rules in the Council: the strategic power index for the Commission–Council game Voting rule

Unanimity

QMV

Distribution of policy states

Uniform

Normal

Uniform

Normal

Value for: – Council member • Germany, France, Italy, UK (10) • Spain (8) • Belgium, Greece, Netherlands, Portugal (5) • Austria, Sweden (4) • Denmark, Finland, Ireland (3) • Luxembourg (2)

0.007 0.007

0.024 0.024

0.074 0.060

0.026 0.022

0.007 0.007 0.007 0.007

0.024 0.024 0.024 0.024

0.036 0.029 0.023 0.013

0.012 0.010 0.007 0.004

– Parliament

0

0

0

0

– Commission

0.011

0.252

0.299

0.754

Note: The power scores are based on mean distances between the equilibrium policy and the player’s ideal point. Based on a difference of means test (Blalock 1972: 226–8), the computed mean values differ significantly within and between the columns at a 1% level of significance. In addition, vote shares for the different member states are presented between parentheses.

The effects on EU decision-making

103

0.074 for qualified majority voting (an increase of about 950 per cent). A similar, although less substantial, change occurs for these large countries in the Union when we use a normal distribution of policy states. For the smaller member states, however, power scores decrease. This effect is a consequence of the Commission’s agenda-setting power in the decision-making procedure combined with the more homogeneous preferences. Since preferences will be more clustered in the centre of the distribution, the Commission is more often able to ‘pull’ the equilibrium outcome towards its own most preferred policy position (Steunenberg 1994: 651). For a non-central preference of the Commission, this will lead to a rather different policy than the ones the Council members actually prefer. Given the differences in vote shares, the smaller member states will be more affected than the larger ones, as the results reveal. This power increasing effect of introducing qualified majority voting has also been noted by König and Bräuninger (1998), who indicate that voting rules differ with regard to the extent to which they allow for policy change. Berg and Lane (2001) and Lane and Mæland (in Chapter 7 of this volume) point out that rules differ with regard to their balance between what they call the ‘power to change’ and the ‘power to block’. While the ‘power to block’ might be beneficial to a specific player aiming to avoid changes in a preferred policy, the ‘power to change’ allows a majority of players to move a policy to one it prefers more. Unanimity rule maximises a player’s ‘power to block’ since it allows any player to veto a proposal for change. However, this does not necessarily mean that players will favour this rule. Deciding on a large number of issues, players may prefer to allow for some change and thus may depart from unanimity rule since the average benefits of blocking might be lower than those of allowing for change. The simulation results indicate that this might be indeed the case. The analysis so far suggests that most of the problems associated with further enlargement already exist in the current Union. Inertia due to unanimity voting is already quite substantial. Furthermore, all current member states could on average benefit from introducing more qualified majority voting on policy issues, which is described as the power increasing effect of this change. However, the benefits of such a change are not evenly distributed over the member states. Especially the larger member states will benefit more than the smaller ones, which is a result of their higher voting weights. The existence of these benefits does not imply that member states will immediately agree to more qualified majority voting. When the member states know the policy issues at stake (such as asylum and immigration policy), and they know their preferences, they might not be inclined to change these rules in the short run.12 In the long run, however, such a move might be beneficial. Let us turn now to the potential effects of enlargement.

Enlargement of the Union Will the expected effects of deadlock indeed occur in an enlarged European Union? To analyse the effects of enlargement, I include in my analysis all countries that applied for EU membership and are included in the Declaration

104

B. Steunenberg

on enlargement, as part of the Nice Treaty (2001: 162).13 Since it is less clear when these countries will become a full member of the Union, as indicated in Chapter 1 of this volume, all applicant countries will be taken as one group. The simulations based on the Commission–Council game are now repeated for a larger European Union of twenty-seven members, which includes twelve applicant countries. In order to establish the effects of enlargement and no other changes, the current rules concerning qualified majority voting have been used (Article 205 EC). The vote shares for the applicant countries are based on an extrapolation of the current distribution of shares made by the Commission (European Commission 2000: 63). Policy flexibility in a Union of twenty-seven member states The results concerning the policy flexibility in the Union under both voting rules are presented in Table 6.3. As the table indicates, enlargement may have some effect on the possibility of choosing a new policy. On average, it will become more difficult to agree on a new policy. Under unanimity rule, this ability – that is, the probability of choosing a new policy – decreases from 10 per cent to 5 per cent. This finding is in line with the intuition expressed in the various EU policy documents and the results revealed by earlier voting power analyses. For qualified majority voting, however, this effect is hardly noticeable. Moving to an enlarged European Union with twenty-seven members, the ability to set a new policy decreases by less than 1 per cent (which is a change of about 2 per cent). This property of qualified majority voting, which suggests a substantial robustness in the face of enlargement, will be called its enlargement resistant property. Focusing on the inertia index, enlargement will barely affect the magnitude of policy change. The expected distance between the status quo and the equilibrium outcome, as expressed by this index, only slightly increases, allowing for an incremental decrease of change. In terms of policy change, the differences between the current (EU-15) and a larger European Union (EU-27) are rather modest. More substantial are the possible effects on policy flexibility by changing the voting rules in the Union. Moving away from unanimity rule and towards qualified majority voting in an enlarged Union would increase the probability of adopting a new policy from 5 per cent to 48 per cent. Also the inertia index changes dramatically. It indicates Table 6.3 Comparative analysis of Council voting rules: probabilities and inertia index for the Commission–Council game under a uniform distribution Voting rule

Unanimity

QMV

Size of the EU

EU-15

EU-27

EU-15

EU-27

Expected policy choice: • current policy (status quo) • new policy

90% 10%

95% 5%

51% 49%

52% 48%

Inertia index

0.97

0.99

0.54

0.57

The effects on EU decision-making

105

that qualified majority voting allows for much more policy flexibility than unanimity rule in both the current and an enlarged Union. Power of the individual players Different effects occur when we focus on the scores for the individual member states based on the strategic power index. These values, before and after enlargement, are presented in Table 6.4. Under unanimity rule, and assuming a uniform distribution, these scores are 0.007 for all member states in the current Union and 0.001 after enlargement. In other words, as a result of enlargement current members will experience a drop in their power of about 85 per cent. For qualified majority voting a similar effect occurs, although of a smaller magnitude. The current member states will face a decrease in their power between 23 per cent and 31 per cent. In particular, the larger member states will experience a more substantial decrease in their power than most of the smaller ones. This could be the reason why the larger member states insisted on a re-weighting of their vote shares for Council decisionmaking. Table 6.4 Expected benefits of different voting rules in the Council: the strategic power index for the Commission–Council game under a uniform distribution Voting rule

Unanimity

Size of the EU

EU-15

EU-27

EU-15

EU-27

0.007 0.007

0.001 0.001

0.074 0.060

0.052 0.042

0.007 0.007 0.007 0.007

0.001 0.001 0.001 0.001

0.036 0.029 0.023 0.013

0.026 0.021 0.015 0.010

0 0 0 0

0.001 0.001 0.001 0.001

0 0 0 0

0.042 0.031 0.026 0.021

0 0

0.001 0.001

0 0

0.015 0.010

– Parliament

0

0

0

0

– Commission

0.011

0.001

0.299

0.263

Value for: – Current Council member • Germany, France, Italy, UK (10) • Spain (8) • Netherlands, Belgium, Greece, Portugal (5) • Austria, Sweden (4) • Denmark, Finland, Ireland (3) • Luxembourg (2) – New Council member • Poland (8) • Romania (6) • Czech Republic, Hungary (5) • Bulgaria (4) • Lithuania, Slovakia, Estonia, Latvia, Slovenia (3) • Cyprus, Malta (2)

QMV

Note: The vote shares for the different member states are presented between parentheses (based on Article 205.2 EC). For the ‘new’ member states, these vote shares are based on an extrapolation made by the European Commission (2000: 63). Only countries with the same scores are grouped in the table.

106

B. Steunenberg

Enlargement as momentum The results found so far indicate that enlargement will reduce the power of the current member states in the Council. This confirms results reported by studies using voting power indices, which approach the Union’s decision-making process in a conceptually different way.14 Furthermore, we found that when decisions are taken by qualified majority, this loss will be larger to the larger member states such as Germany, France, Italy and the United Kingdom. This power-decreasing effect on the current member states forms the main link between the enlargement and institutional reform. Current member states could counteract these losses by adapting the Union’s decision-making rules, especially the rules concerning qualified majority voting. In this respect, enlargement can be regarded as momentum for institutional change. Without enlargement, reforms of this rule might not be politically feasible, since it would lead to a redistribution of power among the current member states. Those who would lose could not agree to such a change. In contrast to a reform of current decision-making rules, an institutional reform that introduces qualified majority voting in areas where unanimity rules apply may lead to a power-increasing effect to all current member states. This effect was, at least, noted for a uniform distribution of preferences. When a change of the qualified majority rule is combined with the introduction of more qualified majority voting in an enlarged Union, the net effect could be that the member states would gain in power.15 This sheds a different light on the reform proposals discussed during the 2000 Intergovernmental Conference, which led to the Treaty of Nice.

The Nice Treaty and its effects The 2000 Intergovernmental Conference and the Nice European Council aimed to introduce institutional reforms with a view to the possible enlargement of the Union. During the preparation of the new treaty several issues were discussed including the introduction of more qualified majority voting and the voting rules for Council decision-making. The introduction of more qualified majority voting in policy areas where the Council still decides by unanimity could be beneficial for all member states as the analysis in the previous section shows. This reform would increase the Union’s policy flexibility, which allows member states to reshape current policies. This could, at the individual level, lead to an increase of power to all member states, since they are, on average, more able to adapt the Union’s policies to their preferences. At Nice, the member states indeed chose to introduce qualified majority rule for several provisions. However, the number of provisions to which they agreed was smaller than initially proposed.16 The rather limited enthusiasm for the introduction of more qualified majority voting could be explained by the extent to which the issues at stake are politically sensitive. As indicated before, agreement to increase the share of qualified majority voting is affected by the policy issues that are up for decision in the near future. In as much as these issues are highly politicised so that member states have rather well-defined and divergent

The effects on EU decision-making

107

preferences, they might not be inclined to change the rules in the short run if such a change could lead to less preferred decisions. Issues at stake here are, e.g., taxation, some aspects of social policy and banking security. At Nice the member states decided to preserve unanimity in these areas. A second change introduced at Nice, which is important here, concerns the qualified majority rule. After long and intense bargaining, the member states reached agreement on a new voting rule. This includes a re-weighting of vote shares and the addition of a new criterion for the adoption of Council proposals. According to the new treaty the Council could approve a new proposal if there are at least 258 votes in favour cast by a majority of members that constitute at least 62 per cent of the total population of the Union (Nice Treaty 2001: 162). With the new population requirement, the qualified majority rule of the Union has become a ‘triple’ majority rule. If the member states ratify the treaty, the new rule will be effective from 1 January 2005. What are the effects of this new qualified majority rule on policy flexibility and the power positions of the individual member states? In the next analysis, I will compare the results found so far with those based on the agreements reached at Nice. Felsenthal and Machover (2001: 11–13) show that the new qualified majority rule is in fact a weighted rule, that is, the new rule is, in terms of outcomes, identical to another rule in which the voting weights of the member states have been transformed taking account of all majority requirements. This weighted rule, which is computationally less complex than the original formulation in the treaty, has been used in the analysis.17 Policy flexibility after Nice The results concerning the ability to take decisions after Nice based on qualified majority voting are presented in Table 6.5. The first column presents the results based on an extrapolation of the current rule made by the Commission (see also Table 6.4). The second column gives the results based on the version of qualified majority voting agreed upon at Nice. Results on unanimity voting are not included in the table since these results are not affected by Nice. As the table indicates, the new rule has a negative effect on policy flexibility. While under the current version of qualified majority voting, the probability of adopting a new policy is 48 per cent, this probability decreases to 44 per cent

Table 6.5 The effects of Nice: probabilities and inertia index for the Commission– Council game for a uniform distribution and after enlargement (EU-27) Voting rule:

Current version of QMV (Article 205 EC)

Nice-version of QMV

Expected policy choice: • current policy (status quo) • new policy

52% 48%

56% 44%

Inertia index

0.57

0.63

108

B. Steunenberg

under the Nice version of this rule. This result is confirmed by the change in the inertia index. The new rule will lead to more instances in which the status quo prevails, since the score for this index increases. In other words, despite the official effort to reduce ‘problems of decision-making’ and to make the Union more ‘manageable’, the new voting rule introduced at Nice has the opposite effect. Power of the individual players after Nice The next question is how Nice affects the individual power positions of the member states. The strategic power scores for both the current and the Nice version of qualified majority voting are presented in Table 6.6. Compared to the current version of qualified majority voting, the Nice Treaty has two important effects. First, some larger member states such as Germany, Spain and Poland will experience a slight increase in power. The effect for Germany is due to its larger population size (see Felsenthal and Machover 2001: 11–12). The scores for France, Italy and the United Kingdom remain the same. Second, the smaller member states will experience a decrease in their power by the

Table 6.6 The effects of Nice: the strategic power index for individual players in an enlarged Union for the Commission–Council game and a uniform distribution Voting rule:

Current version of QMV

Nice-version of QMV

Value for: – Council member • Germany (29) • France, Italy, UK (29) • Spain (27) • Netherlands (13) • Belgium, Greece, Portugal (12) • Austria, Sweden (10) • Denmark, Finland, Ireland (7) • Luxembourg (4)

0.052 0.052 0.042 0.026 0.026 0.021 0.015 0.010

0.053 0.052 0.049 0.023 0.021 0.017 0.012 0.007

– Extended Council member • Poland (27) • Romania (14) • Czech Republic, Hungary (12) • Bulgaria (10) • Lithuania, Slovakia (7) • Estonia, Latvia, Slovenia (4) • Cyprus (4) • Malta (3)

0.042 0.031 0.026 0.021 0.015 0.015 0.010 0.010

0.049 0.025 0.021 0.017 0.012 0.007 0.007 0.005

– Parliament

0

0

– Commission

0.263

0.211

Note: Vote shares for the different member states and based on the Nice Treaty are presented between parentheses. The differences between the scores for Germany and France, Italy and UK are due to their differences in population size, which appear to be relevant. Only countries with the same scores are grouped in the table.

The effects on EU decision-making

109

introduction of the new rule. This decrease in power varies between 12 per cent for the Netherlands (a change from 0.026 to 0.023) to 30 per cent for Luxembourg (a change from 0.010 to 0.007). Also almost all applicant countries will be worse off under the new rules. Compared to the current version of qualified majority voting, the smaller applicant countries will experience a decrease in power. The only exception is Poland, which will experience an increase in power.

The effects of preferences and other players Preferences The results presented so far are based on the uniformity assumption, which assigns an equal probability to the occurrence of any possible ordering of preferences and a status quo point. This means that preference configurations in which the political actors hold rather extreme and divergent preferences are regarded as equally likely as configurations in which actors have preferences that are more similar and ‘centrally’ located. However, a uniform distribution may not represent preferences correctly. The current member states (EU-15) may have preferences that are more similar than those of the applicant countries. Based on economic indicators, future member states will be quite different from the present members (see König and Bräuninger 2000). For this reason one would expect that preferences in an enlarged European Union (EU-27) will be more heterogeneous. Such an increase in heterogeneity can be associated with a change from normally distributed preferences in the current Union to uniformly distributed preferences in a Union of twentyseven. Under a normal distribution, the preferences of the member states do vary, but the differences between them are smaller than for a uniform distribution. After enlargement, however, the preferences of the current as well as the new member states may become more alike. This may be the result of adaptations on the part of the new member states that are currently adjusting national regulations and policies to the acquis communautaire, and of current member states that discover new ways of doing things through working with the new members.18 This possible change in preferences after enlargement can be represented by replacing a uniform with a normal distribution for an enlarged Union. The consequences of these different preference distributions for the decisionmaking process are analysed by moving from a normal to a uniform distribution, and then back to a normal distribution for the current and an enlarged Union. The results are presented in Table 6.7 for both unanimity voting (Table 6.7(a)) and qualified majority voting (Table 6.7(b)). In the first column, a normal distribution of policy states has been used for the current Union (EU-15) in which all players have the same modus. The second column presents the results for an enlarged Union (EU-27) in which all players are assumed to have uniformly distributed preferences. Finally, and to reflect a possible convergence of preferences in the years after enlargement, the third column presents results for normally distributed preferences in an enlarged Union. Compared to today’s Union, in which the current member states are assumed to have homogeneous preferences associated with a normal distribution, enlargement

110

B. Steunenberg

Table 6.7 Comparative analysis of Council voting rules after enlargement: probabilities for the Commission–Council game under different distributions 1 EU-15 Normal

2 EU-27 Uniform

3 EU-27 Normal

(a) Unanimity voting Expected policy choice: • current policy (status quo) • new policy

41% 59%

95% 5%

47% 53%

(b) Qualified majority votinga Expected policy choice: • current policy (status quo) • new policy

17% 83%

56% 44%

18% 82%

Size of the EU: Distribution:

a

For EU-27 based on the voting rules as embedded in the Nice Treaty (2001: 162).

will lead to a decrease of policy flexibility. The probability of setting a new policy drops from 59 per cent for the EU-15 to 5 per cent for an EU-27 with heterogeneous preferences based on a uniform distribution. However, if preferences in an enlarged European Union become more homogeneous in the years after enlargement, this effect will be mitigated. As indicated in the third column, the probability of selecting a new policy could increase to 53 per cent, which is slightly less than in the current Union (59 per cent). These results illustrate that preference homogeneity is a critical factor for the Union’s decision-making capacity. For qualified majority voting a similar pattern occurs as for unanimity rule. Starting with a current Union (EU-15) with homogeneous preferences, enlargement will reduce policy flexibility as the probability of changing policy drops from 83 per cent for today’s Union to 44 per cent for an enlarged Union (EU-27) with heterogeneous preferences.19 However, if the preferences of the Union’s future members become more alike, this decrease in flexibility would completely disappear. The probability of changing policy in an enlarged Union is similar to that in today’s Union, that is, about 82 per cent, assuming that preferences are normally distributed (compare columns 1 and 3). This is again an expression of the enlargement resistant property of qualified majority voting. Involvement of the European Parliament The game form analysed so far has been the Commission–Council game without involvement of the European Parliament. In many areas the European Parliament plays a role in the making of EU policies as part of the assent and the codecision procedure.20 These procedures grant the Parliament formal decision-making power, so that no new policy can be made without its explicit approval. The basic dynamics of decision-making with a role of Parliament is reflected in the Commission–Council–Parliament game, in which Parliament has to approve a policy selected by the Council upon a proposal of the Commission. Here I

The effects on EU decision-making

111

assume that Council decisions have to be taken by a qualified majority, which is the case for most instances in which Parliament plays a role. The question is whether the involvement of Parliament affects my findings based on the simple Commission–Council game. Table 6.8 presents the results of this game with Parliament both for the current Union as well as a Union after enlargement (and for both versions of qualified majority voting in the Council). These results lead to three conclusions. The first conclusion concerns the impact of parliamentary involvement. Compared to the third and fourth column in Table 6.1, which present the results for the Commission–Council game under qualified majority voting, parliamentary involvement reduces the probability of selecting a new policy (49 per cent in the Commission–Council game to 41 per cent in the extended game under a uniform distribution and for the current Union). This reduction is a consequence of the fact that Parliament, as an additional player, may also decide to block policy change leading to more cases of deadlock. The second conclusion concerns the effects of enlargement. For both a uniform and a normal distribution of policy states enlargement does not seem to have a substantial impact on a decision-making process in which Parliament plays a role. Using the current qualified majority rule, the probabilities of selecting a new policy remain the same (41 per cent for a uniform distribution and 80 per cent for a normal one), independent of the size of the Union. The remarkable robustness of these results regarding an increase of the number of Council members is again an expression of the enlargement resistant property of qualified majority voting. Since Council decisions in the current game are based on this rule, it affects the overall result. The third and last conclusion relates to the Nice version of qualified majority voting. Moving from the current version of qualified majority voting to this new rule, the probability of selecting a new policy decreases from 41 per cent to 38 per cent (a change of about 7 per cent). For a normal distribution this change is much less, from 80 percent to 79 per cent (a change of about 1 per cent). The reason for this is that when the preferences of the players are more homogeneous, this would lead to less instances of deadlock. These results again show that the Nice version of qualified majority voting will reduce the Union’s policy Table 6.8 Comparative analysis of procedures with Parliament’s involvement: probabilities for different distributions Distribution of policy states:

Uniform distribution

Version of QMV: Size of the EU:

EU-15

Current Nice EU-27 EU-27

EU-15

Current Nice EU-27 EU-27

Expected policy choice: • current policy (status quo) • new policy

59% 41%

59% 41%

20% 80%

20% 80%

62% 38%

Normal distribution

21% 79%

Note: ‘Current’ stands for an extrapolation of the current qualified majority rule based on Article 205 EC; ‘Nice’ stands for the Nice-version of qualified majority voting (see Nice Treaty, 2001: 162).

112

B. Steunenberg

flexibility, especially when the member states have heterogeneous preferences. However, as for all effects of institutions, the extent to which these differences occur depends on whether political actors have divergent views and therefore need to apply decision-making rules in order to make collective decisions.

Conclusions In this chapter I have analysed the relationship between enlargement and institutional reform in the European Union. Although the ‘official’ belief is that enlargement requires reform, the nature of the relationship between these processes is not so clear. The Union’s decision-making capacities depend critically on the voting rules that are used and on the heterogeneity of the preferences of the member states. This is true for the current Union as well as for an enlarged Union, and it is not necessarily related to enlargement as such. A first result based on the analysis in this chapter is that problems of policy inflexibility are not typical for an enlarged European Union. Unanimity voting already leads to substantial deadlock in the current Union. Moving to an enlarged Union of twenty-seven member states, the probability that the Union will not be able to act is expected to increase only slightly (from 90 to 95 per cent). Under qualified majority voting, enlargement will not affect the Union’s ability to take new decisions. This has been labelled the enlargement resistant effect of qualified majority voting. In this respect, enlargement and the problems of inflexibility, which would make the Union ‘unmanageable’, seem to be hardly related. A second result is that enlargement is expected to decrease the power of the individual member states in the Union’s decision-making process. In addition, when qualified majority voting is used, the larger member states will lose more power than the smaller ones. This could be the reason why the larger member states insist on a re-weighting of vote shares for Council decision-making as part of the institutional reforms. Here, enlargement and the institutional reforms within the Union are closely related to the position of individual member states in the decision-making process. Analysing the Nice reform package, we find that the larger member states managed to introduce a new rule concerning qualified majority voting that will preserve their power positions in the Council after enlargement. The smaller member states will experience a decrease of their power due to this reform. This redistribution of power between the member states was achieved by linking the reform with enlargement. Why did the smaller member states agree to these reforms if they stand to lose power? The first explanation is the link with the reform of the composition of the European Commission. In the intergovernmental conference leading to the Nice summit the member states already reached agreement on the redistribution of the number of Commissioners. The solution was that the larger member states would give up their right to fill two posts and have only one Commissioner per country. It implies that larger member states lose power in the Union’s decisionmaking process, while the smaller countries maintain their position, assuming that Commissioners would act as their ‘representatives.’ By casting the composition of the Commission in terms of ‘national’ representation the larger

The effects on EU decision-making

113

member states could call for compensation for this loss. The proposal on the number of Commissioners made the smaller member states well aware of the fact that some ‘compensation’ to the larger member states might be needed in order to make the whole package of reforms acceptable. However, this reform does not provide a sufficient reason to accept a less beneficial re-weighting of vote shares for Council decision-making. The second and most important reason for reaching agreement was that the reform of the Council’s voting rule was linked to the enlargement as such. During the 2000 Intergovernmental Conference this reform was presented as one of the key reforms that would make enlargement possible. This made all member states – large and small – well aware of the political importance of seeking agreement on this issue and the possible risk of being blamed for ‘blocking’ enlargement by withholding the ‘necessary’ internal reforms of the Union. The confidential minutes of the summit indicate that the proposal on the re-weighting of the votes was heavily contested by some of the smaller member states, especially Portugal and Belgium. Finally, these countries, and especially Belgium, gave in at the very last moment. As the Belgium premier Verhofstadt remarked in the early morning of Monday 11 December: ‘I will accept for the sake of enlargement.’21 Here, the larger member states, and especially the French Presidency, used enlargement as a renegotiation strategy to improve their position in the Union. The new qualified majority rule introduced by Nice also reduces the potential influence of most of the applicant countries in Council decision-making. Compared to the current rules, the mostly smaller applicant countries will be worse off under the new rules agreed upon in Nice. This is another consequence of the ‘shadow of enlargement.’ The larger member states in particular aimed to secure their interests in view of enlargement, in this case by improving their power position vis-à-vis the members of an enlarged Union. The new rule agreed upon in Nice has the possibly unintended effect that it will lead to more instances of deadlock, which reduces the Union’s decision-making capacity. Instead of increasing policy flexibility, the changes adopted at Nice lead to less flexibility, which would ‘weaken’ the Union in the near future, as initially feared by the Reflection Group (1995: 3). The process of enlargement appears to be caught between external and internal pressures, both affecting the timing of entry of new member states. The external pressures consist of a successful transformation on the side of the applicant countries of their political, economic and legal system in view of the acquis communautaire. This process requires substantial political efforts in the applicant countries, which need to be recognised and rewarded by the Union. The speed of these domestic reforms and the demands imposed by the Commission on the applicant countries as part of the accession negotiations determine the prospects of entry. The internal pressures consist of the reforms linked to the process of enlargement, which take place within the current Union. This implies that the time path of enlargement is also based on these pressures. Furthermore, if some of the desired changes are not adopted, one or more member states may block further enlargement. In other words, the willingness to accept new members could be made fully dependent on issues that are not related to enlargement.

B. Steunenberg

114

APPENDIX 6

Indices on strategic power and inertia The strategic power index, as proposed by Steunenberg et al. (1999: 347–50), is based on the average distance between the outcomes from the application of a given game form and a player’s ideal point in each application. Let n ∈ N be the number of players. For an m-dimensional and finite outcome space X ∈ Rm, these players have Euclidean preferences, which can be characterised by player i’s ideal point xi = (x1i, x2i, ... xmi). Let q ∈ X denote the status quo. I call a combination of a particular ideal point for each player and the status quo a ‘policy state’, which will be denoted as s = (x1, x2, ... xn, q). Finally, let xπ(s) be the unique equilibrium outcome of the game based on game form π and policy state s. Assume that each particular policy state is the instance of a random variable s = (x1, x2, ... xn, q). The expected (or mean) distance between the equilibrium outcomes for some game form, π, and player i’s ideal point is then given by

()()

∫

∆πi = δ iπ f s d s

(6.1)

where

δ iπ =

m

 π x s 

∑ () k =1

k

 − x ik  

2

(6.2)

is the Euclidean distance between the equilibrium outcome of the game and the ideal point of player i in any particular state of the world, and f(s) is the density function if s is a continuous random variable.22 To construct an index that measures power in absolute terms, Steunenberg et al. (1999: 348–9) introduce an external observer, that is, an actor that does not have any decision-making power and thus can be regarded as ‘powerless’. Let this observer be e, which is an actor that is not a player according to the game form π, then the power of a player i in this game form can be defined as:

Ψiπ =

∆πe − ∆πi ∆πe

=1−

∆πi ∆πe

(6.3)

which is called the strategic power index. This index lies in the interval [0,1] and increases with the power of a player. The expected distance for a player that is ‘powerful’ enough to dictate the outcome of a game under any preference configuration would be zero leading to a corresponding value for the index of one. By contrast, if a player has a similar effect on the outcome of a game as the external observer (which, by definition, is ‘powerless’), the expected distance for this player is the same as for the external observer leading to a corresponding value for the index of zero.

The effects on EU decision-making

115

Based on this approach, there is a natural way to express the extent to which players are unable to act and to pull a new policy away from the current state of affairs or status quo. For a specific game form, that status quo bias can be measured by the expected distance between the equilibrium outcome and the status quo, which is defined as ∆πq. Substituting this value for the expected distance found for a player in the strategic power index, we get

Ψqπ =

∆πe − ∆πq ∆πe

=1−

∆πq ∆πe

(6.4)

which is called the inertia index. A value of one for this index means that under some procedure the status quo always prevails. The smaller the value for the index, the more players are able to move the equilibrium policy away from the status quo.

Notes 1 Some parts of this contribution are based on my paper ‘Enlargement and institutional reform in the European Union: Separate or connected issues?’, published in Constitutional Political Economy 12: 349–68. 2 See, for instance, the European Commission (1996: 2), which fears that ‘[a]s the number of Union members increases, it creates a risk of the Union being watered down’. See Chapter 1 for a further discussion. 3 See, for instance, Preston’s (1997: 20–1) observation that ‘[n]ew members are integrated into the EC’s institutional structure on the basis of limited incremental adaptation facilitated by the promise of a more fundamental review after enlargement’. Such reviews have been postponed in the past or achieved limited results. 4 See Hosli (1993), Widgrén (1994), Turnovec (1998), and Paterson and Silárszky (1999). See also Lane and Mæland in Chapter 7 of this volume. 5 See also Pattanaik and Suzumura (1996) and Miller (1999: 225–6), who relate political institutions, defined as rules or procedures, to game forms. The term ‘game form’ goes back to Gibbard (1973: 589). 6 At the same time, and in view of this reason, the analysis does not focus on ways in which political actors may reach sufficient support in order to pass legislation, such as coalition formation or vote trading. I aim to focus on the properties of specific sets of rules, which are represented by a game form, and not on the specific preferences of political actors. 7 The strategic power index differs from voting power indices, since it is based on a game-theoretical modelling of a decision-making process in which institutions as rules or procedures are included. Voting power indices approach decision-making as a process of coalition formation and are less suited to include institutional features (see Garrett and Tsebelis 1996, 1999; Steunenberg et al. 1999). 8 See Appendix 6 for a more detailed discussion of both indices. 9 See Mueller’s analysis of the choice of an optimal majority, in which he identifies a simple majority as the optimal size (Mueller 1989: 55–7). His analysis is an extension of Buchanan and Tullock’s analysis in which the choice of an optimal majority is based on a trade-off between decision costs and the expected loss of utility from making a decision in which one or more individuals are opposed by a majority (Buchanan and Tullock, 1962: 111–16). 10 The sample size used in these simulations is based on the decision to estimate the expected distance for each player within an accuracy of at least ±0.1 distance-units

116

11

12 13 14 15 16 17

18

19

20

21 22

B. Steunenberg

and to make use of a 99 per cent confidence interval. In addition, if both the outcome and the ideal point of a player were randomly determined, the expected standard deviation would be 25 (with an expected mean of 50). Using the formula for a confidence interval (see Blalock, 1972: 213–15), the minimum sample size in that case is √N = (2.58 × 25)/0.1 = 645, or N = 416,025. Smaller standard deviations, which can be associated with players that have some impact on the outcome of a game, imply a smaller minimum sample size. A sample of 10,000,000 observations, which is above the required minimum, is sufficiently large to get an accurate impression about the mean values. These numbers have been used to construct the index. Based on the precision used and the values for the expected distance, the index can be interpreted with an accuracy of at least three figures after the digit. In discussions on the treaties of the Union the term ‘flexibility’ is used to refer to arrangements that allow member states to opt out of certain decisions, or to increase cooperation between some of the member states (such as the Schengen agreement), which is more intensive than what other members prefer. In Chapter 7, Lane and Mæland, for instance, refer to this flexibility concept. See Steunenberg and Dimitrova (1999), who point at this problem with regard to the negotiations on the Amsterdam Treaty. Turkey is not included in this declaration and has not been assigned a voting weight, so I will not include it in my analysis. See the studies mentioned in the Introduction of this chapter. See Chapter 1 of this volume for a discussion of the different conceptual approaches on decision-making in the Union. Compare the first and the last column of Table 6.4. As Ludlow (2001: 3) indicates: ‘Overall, with approximately 40 new provisions for QMV out of a possible list of around 75, the result was not bad, particularly as everybody shared the Commission’s view that upwards of 20 items were off limits.’ The adjusted voting weights based on this weighted rule can be found in Table 8 of the paper of Felsenthal and Machover (2001). Another issue concerns the size of a ‘blocking’ majority in a future Union, which is part of yet another declaration of the Nice Treaty (2001: 165). As noted by Felsenthal and Machover, application of this declaration is not consistent with the earlier declaration and its application would lead to a slightly different voting rule. In my analysis, I restrict myself to the voting rule as expressed by the first declaration (Nice Treaty 2001: 162). Also note that in the current Union, relatively recent members, such as Spain and Portugal, do not vote differently from most of the other member states (see Mattila and Lane 2001). In contrast, UK, Sweden, the Netherlands, Germany and Denmark often appear to vote against a majority in the Council. The results for the EU-27 are based on the new rules agreed at Nice. For a normal distribution, as we will see, the effect of the Nice version of qualified majority voting is less substantial than for a uniform distribution. This is a consequence of the fact that under a normal distribution preferences are less dispersed and therefore lead to fewer cases of deadlock. A third well-known legislative procedure with parliamentary involvement was the cooperation procedure, which applied to several treaty articles until the Amsterdam summit (1997). After the coming into force of the Amsterdam Treaty (1999), the significance of the cooperation procedure is reduced to only a few articles (Articles 99.5, 102.2, 103.2 and 106.2 EC) dealing with EMU and the introduction of the Euro. See the minutes of the European Council in Nice as initially published by El País. This quote is taken from NRC Handelsblad, 18 December 2000, page 5. If the game does not have a unique equilibrium, but multiple equilibria, the simple Euclidean distance can be replaced by the average Euclidean distance, i.e. the sum of the Euclidean distances between each equilibrium outcome and the player’s ideal point for all equilibria in a particular state of the world, divided by the number of equilibria.

The effects on EU decision-making

117

References Berg, S. and J.-E Lane (2001) ‘Measurements of voting power: individual and collective properties with applications to the EU’, in M. Holler and G. Owen (eds) Power indices and coalition formation, Dordrecht: Kluwer, forthcoming. Blalock, H.M. (1972) Social Statistics, Tokyo: McGraw-Hill (2nd edition). Brennan, G. and J.M. Buchanan (1980) The Power to Tax: Analytical Foundations of a Fiscal Constitution, Cambridge: Cambridge University Press. Brennan, G. and J.M. Buchanan (1985) The Reason of Rules: Constitutional Political Economy, Cambridge: Cambridge University Press. Buchanan, J.M. and G. Tullock (1962) The Calculus of Consent, Ann Arbor: University of Michigan Press. Dehaene, J.-L., R. von Weizsäcker and D. Simon (1999) ‘The institutional implications of enlargement: report to the European Commission’, Brussels, 18 October 1999. Dimitrova, A.L. (2000) ‘Enlargement from a historical institutional perspective: an irreversible process?’, Leiden: Leiden University, unpublished manuscript. Enelow, J.M. and M.J. Hinich (1984) The Spatial Theory of Voting: An Introduction, Cambridge: Cambridge University Press. European Commission (1996) ‘Reinforcing political union and preparing for enlargement’, Commission Opinion, Brussels, February 1996. European Commission (2000) ‘Adapting the institutions to make a success of enlargement’, Commission Opinion in accordance with Article 48 of the Treaty on European Union on the calling of a Conference of Representatives of the Governments of the Member states to amend the Treaties, Brussels, 26 January (COM(2000)34). Felsenthal, D.S. and M. Machover (2001) The Treaty of Nice and Qualified Majority Voting, London: LSE Voting Power Report VVP 02/01. Garrett, G. and G. Tsebelis (1996) ‘An institutional critique of intergovernmentalism’, International Organization 50: 269–99. Garrett, G. and G. Tsebelis (1999) ‘Why resist the temptation of power indices in the European Union’, Journal of Theoretical Politics 11: 291–308. Gibbard, A. (1973) ‘Manipulation of voting schemes: a general result’, Econometrica 41: 587–601. Hosli, M.O. (1993) ‘Admission of European Free Trade Association states to the European Community: effects on voting power in the European Community Council of Ministers’, International Organization 47: 629–43. König, T. and T. Bräuninger (1998) ‘The inclusiveness of European decision rules’, Journal of Theoretical Politics 10: 125–42. König, T. and T. Bräuninger (2000) ‘Governing the enlarged European Union: accession scenarios and institutional reform’, Central European Political Science Review 1: 42–62. Ludlow, P. (2001) ‘The Treaty of Nice: neither triumph nor disaster’, ECSA Review 14(2): 1–4. Mattila, M. and J.-E. Lane (2001) ‘Why unanimity in the Council? A roll call analysis of Council voting’, European Union Politics 2: 31–52. McKelvey, R.D. (1976) ‘Intransitivities in multidimensional voting models’, Journal of Economic Theory 12: 472–82. Miller, N.R. (1999) ‘Power in game forms’, Homo Oeconomicus 16: 219–43. Moe, T.M. (1990) ‘Political institutions: the neglected side of the story’, Journal of Law, Economics, and Organization 6 (Special Issue): S213–53. Mueller, D.C. (1989) Public Choice II: A Revised Edition of Public Choice, Cambridge: Cambridge University Press.

118

B. Steunenberg

Nice Treaty (2001) ‘ Treaty of Nice amending the Treaty on European Union, the Treaties establishing the European Communities and certain related Acts. Conference of the representatives of the governments of the member states’, Brussels, 14 February 2001 (SN1247/1/01 REV1). Ostrom, E. (1986) ‘An agenda for the study of institutions’, Public Choice 48: 3–25. Paterson, I. and P. Silárszky (1999) ‘Redesigning the institution of the Council of Ministers in advance of EU enlargement: issues and options’, Journal for Institutional Innovation, Development and Transition 3: 43–54. Pattanaik, P.K. and K. Suzumura (1996) ‘Individual rights and social evaluation: a conceptual framework’, Oxford Economic Papers 48: 194–212. Preston (1997) Enlargement and Integration in the European Union, London and New York: Routledge. Reflection Group (1995) Reflection Group’s report, Brussels, 5 December 1995. Riker, W.H. (1980) ‘Implications from the disequilibrium of majority rule for the study of institutions’, American Political Science Review 74: 432–47. Romer, T. and H. Rosenthal (1978) ‘Political resource allocation, controlled agendas, and the status quo’, Public Choice 33: 27–43. Romer, T. and H. Rosenthal (1979) ‘Bureaucrats versus voters: on the political economy of resource allocation by direct democracy’, Quarterly Journal of Economics 93: 561–87. Schofield, N. (1978) ‘Instability in simple dynamic games’, Review of Economic Studies 45: 575–94. Shepsle, K.A. (1979) ‘Institutional arrangements and equilibrium in multidimensional voting models’, American Journal of Political Science 23: 27–60. Shepsle, K.A. (1989) ‘Studying institutions: some lessons from the rational choice approach’, Journal of Theoretical Politics 1: 131–47. Shepsle, K.A. and B.R. Weingast (1981) ‘Structure-induced equilibrium and legislative choice’, Public Choice 37: 503–19. Steunenberg, B. (1994) ‘Decision making under different institutional arrangements: Legislation by the European Community’, Journal of Theoretical and Institutional Economics 150: 642–69. Steunenberg, B. and A. Dimitrova (1999) ‘Interests, legitimacy, and constitutional choice: the extension of the codecision procedure in Amsterdam’, Enschede: University of Twente, NEMEU Working Paper 99–2. Steunenberg, B., D. Schmidtchen and C. Koboldt (1999) ‘Strategic power in the European Union: evaluating the distribution of power in policy games’, Journal of Theoretical Politics 11: 339–66. Turnovec, F. (1997) The Double Majority Principle and Decision Making Games in Extending European Union, Vienna: Institute for Advanced Studies, East European Series no. 48. Turnovec, F. (1998) ‘Decision making games in the EU: Commission, Council and European Parliament’, Prague: CERGE of Charles University, unpublished manuscript. Widgrén, M. (1994) ‘Voting power in the EC decision making and the consequences of two different enlargements’, European Economic Review 38: 1153–70.

Chapter Title

Part III

Reforming European Union decision-making

119

120

Authors name

Chapter Title

7

121

Enlargement and the Council of Ministers Jan-Erik Lane and Reinert Mæland

Introduction The enlargement of the European Union into former Eastern Europe has been discussed for some time. An enlargement into Eastern Europe is a more difficult project to carry through than the earlier enlargements made between 1972 and 1995, given the huge differences in both income and institutions between Western and Eastern Europe. The economic aspects of EU enlargement eastward weigh heavily, but here we focus upon the political aspects of making the club of states larger, especially constitutional matters concerning decision-making in the key EU organ, the Council of Ministers. Taking on new members in a club questions the power relationships within the club. One may wish to consider not only how many votes (Council) or seats (Parliament) the new member states are going to be entitled to, but also by which voting rule decisions are going to be made in an enlarged club. We use the voting power index approach in order to model the implications of alternative arrangements for an enlarged EU. Today, the EU constitution as it has been laid down in the key treaties covers fifteen member states to which has been allocated a number of votes in the Council as well as a number of seats in the Parliament, held by political parties. As the basic institutions have developed with each enlargement, the number of votes and seats has been changed in order to accommodate more members. At the same time, the decision rule in order to arrive at a winner has slowly been transformed, from unanimity in the Council to qualified majority voting, whereas the European Parliament has always voted by means of simple majority. In this chapter we will discuss the logic of enlargement in relation to the Council of Ministers. It remains the chief decision-making body of the EU and it is the organ representing the member states, and the member states remaining the owners of the acquis communautaire.

Ambiguity of enlargements As emphasised, the enlargement of the Union is based upon both economic and political considerations. Actually, several requirements have from time to time been considered in the deliberations about enlargement, and they may contradict each other.

122 1

2

3

4

5

6

J.-E. Lane and R. Mæland State influence: It cannot be stressed enough that the members of the EU club are states – according to any legal interpretation. It is governments or ministers who act in the Council of Ministers in the EU as the key decisionmaking body of that club, but they always do it on behalf of their states. Thus, EU law must emanate from the will of the member states, as they are the formal owners of the acquis communautaire. And law-making in the EU must present the Council with a major role, as long as it sticks to the fiction that the EU results from treaties between states. In principle, governments exercise influence over the Council through voting power in that body. Representation of the people of Europe: The much-debated issue of a democratic deficit in the Union has added a dimension to the first requirement about state influence. It has more and more been considered essential that the people in the various member states have a direct say on Union law-making and similar important matters. Correspondingly, the European Parliament has been given more powers in Union decision-making, from consultation only, to amendment, to veto powers. Need for coordination mechanism: The Union has come to emphasise more and more coherence in its decision-making, relying upon the Commission as setting the agenda and being involved through the entire process. The Commissioners work for the Union, although they are nominated by the member states and confirmed by the European Parliament. To coordinate decision-making, the Commission has been given the agenda competency. Recognition of size: Since member states are very different in terms of size – population and area – and states are the key stakeholders in the Union, a principle about a relationship between votes in the Council on the one hand and size of country on the other hand has been used since 1958. For the European Parliament, a somewhat similar principle has been at the foundation of the allocation of seats. Thus, the EU is replete with weighted voting or quantitative voting schemes, which never recognise fully the huge differences in size of member states but do take them into account to some extent. Decisiveness: Although the norm about state influence is deeply entrenched in the Union, the developments since 1985 have placed more emphasis upon group decisiveness than the individual member state’s capacity to veto what it may not like. The Council only employs simple majority in relation to procedural matters, but on many important issues it does use qualified or double qualified majorities and not unanimity. The combination of weighted voting and qualified majority raises intriguing questions about the voting power of various coalitions. Sovereignty: The Union is not a federal state. Thus, it is a federation under which member states retain full sovereignty. Union legislation is basically done on the basis of delegation of certain competencies from the member states to the federation, specified in the treaties. Future developments of the Union do not include the possibility that sovereignty may be once and for all transferred from the member states to the Union. The Union is delegation from sovereign states, not a dual division of competencies between two governments of the same unit, central and provincial.

Enlargement and the Council of Ministers 7

8

9

123

Vital interests’ protection: The member states, owning the Union, retain a veto power on all matters relating to the Union where they believe their vital interests are concerned. Thus, unanimity is still the decision rule that applies in matters such as raising taxes, accepting new members, changing the constitution, etc. Deciding about such vital concerns is not part of the delegation made in the treaties. Efficiency: The EU is not a confederation or an international organisation. In order to make the Union important in the member states, its legislation has two very distinctive features: direct effect and immediate impact. Thus, EU laws do not require ratification and they hold not for states but for the people in these states. EU law is not simply recommendations, although the Union bodies also make lots of recommendations. Transparency: The European Court of Justice has been given a broad mandate to interpret EU law and its constitution, whenever ambiguities arise. The European Court of Justice employs on the one hand the treaties and on the other hand right reason as the sources of interpretation. The national courts in the member states must implement the case law produced by the European Court of Justice, or a member state faces fines imposed by the Commission for non-compliance with the Union regulations.

The various requirements listed above include a number of contradictions, which have to be resolved by trade-offs. Thus, the governments of these countries wish to maintain control over the development of the Union and its law-making, according to the criteria of state influence, sovereignty, vital interests’ protection. At the same time, the people of these countries wish to make their voice heard, but the powers of the states in the Council and the powers of the political parties in the European Parliament can only be maximised at the same time to a limited extent. Influence of the people and control by states constitute considerations that loom large, but another important aspect is the power of the Union as a body, meaning its own capacity to act in the sense of taking decisions and implementing them. In order for EU decision-making to be decisive, it is necessary to have a strong Commission that sets the agenda of EU law-making and to empower the Council to be able to take decisions. Veto powers for member states or more power for the European Parliament may weaken the capacity of EU to act as a body. In order to make trade-offs between popular influence, state control and regulative efficacy, the EU constitution has constructed a most complicated institutional web. When organisational complexity increases, then institutional transparency may decrease.

The EU as a club of states The European Union is from a legal point of view an ad hoc organisation, the owners of which are the member states of the Union. Governments pay for the Union organisations and they recruit people who decide in the name of the Union. The organisations of the Union are regulated by public law instruments

124

J.-E. Lane and R. Mæland

such as treaties between the governments, own law and regulations as well as case law emerging from the rulings of the European Court of Justice. The basic problem of the European Union is that of understanding what kind of club the governments of Europe wish to have, as membership is after all voluntary. One may look upon the creation and enlargement of the Union as similar to the economic approach to the provision of goods and services involving economies of scale, or so-called local public goods (Cornes and Sandler 1986). Yet creating or enlarging a club of states has certain specific characteristics. The members are not individuals but states, i.e. highly complex organisations with a diverse set of interests. When governments create a club, then it is bound to be a multi-purpose unit. Second, states have claims to sovereignty, meaning that they will have great difficulty in accepting that the club takes action against their vital interests. Third, states possess enormous resources, which they can use for a variety of purposes, as e.g. instructing a club to take some action on the basis of resources delegated to it by the owners of the club, i.e. the states. These three basic characteristics pull a club of states in different directions. The sovereignty principle is conducive towards a kind of confederal institutions for a club of states, whereas the access to resources makes it possible to empower the club for certain specific purposes, as in federal arrangements. The multi-purpose nature of states makes it highly unlikely that a new club would be delegated one single purpose only. When sovereign states come together and create new institutions for cooperation among them, then they will consider two things. First, they will have to consider the economics of any kind of the club they will set up. A prediction of some sort of the benefits and costs must be done. Second, they will have to reflect upon the organisations of the new club, or more specifically upon the competencies of these as well as the decision-making procedures to be employed in the new organisations. The economics and the politics of clubs constitute distinct perspectives, but at the same time there are links between the two. When a group is set up voluntarily by choice participants, then certain conditions will be satisfied. The European Union has been a mechanism for cooperation between the states in Western Europe and it may become a club comprising almost all of Europe in a not so distant future. No government is forced to participate in this group, as the governments of relevant states apply for membership and the group is constituted by means of an agreement (an international treaty). The fabric of the Union underlines the voluntary nature of the entire enterprise as states may accept some elements of the treaties and stay outside of other elements (variable geometry). There exist different clubs of states with overlapping membership, such as the EU, the EMU and the Schengen group. The prevailing philosophy about group membership in the EU emphasises flexibility, meaning that governments have a certain amount of discretion when choosing which groups to adhere to. Thus, there is a strong element of choice and rationality in opting in or out group membership. Governments act on behalf of states. From a legal point of view group membership is held by states but it is exercised by governments. The following considerations are as minimal when a government decides to apply for and accept

Enlargement and the Council of Ministers

125

membership: the new group or the enlarged club must be able to do something taking decisions that benefit the group members in general. No government would want to be member of a group that is idle or whose operations result in more costs than benefits, at least not in the long-run. This is the requirement of collective rationality (first principle). Groups or clubs typically have two sides: benefits and costs. If the group accomplishes nothing, then there will presumably only exist costs to be picked up somehow by the choice participants. Why give assent to such a club? Groups may, of course, allocate various kinds of goods or services in such a manner that the total benefits are larger than the total costs. The member states will at the end of the day have to pay for the corresponding costs of the club by alternative schemes, as with the European Union, the EMU, etc. The requirement of collective rationality entails that the average benefits of the group are bigger than the average costs of the group, no matter how the benefits and the costs are ultimately distributed onto the members. A stronger version would require that the group maximises the difference between benefits and costs. Groups may calculate the average benefit and cost curves and derive an optimal supply of programmes or decide about an optimal size of group members. One would be inclined to extend the group rationality requirement to the expansion of group. Thus, EU enlargement makes sense collectively if the total benefits of the enlarged group increase more than the total costs of the enlarged group, at least in the long run. Group requirements must be supplemented by individual requirements. A choice participant will not be satisfied knowing that the group as a whole confers more benefits than costs or that it maximises the difference between benefits and costs. Each and every choice participant would want to know whether group membership pays individually for him or her. It could be the case that a member state gains nothing from group or club membership, although the collective requirement is satisfied. It just happens to be the case that under a certain distribution of benefits and costs one member state receives an unfavourable deal. A member state may refuse membership for another reason, however. It may find the alternative of free riding attractive. Even if a group may be predicted to bring about more benefits and costs collectively, it is not automatically forthcoming. The same applies to the enlargement of the group. Thus, we have another condition that is different from collective rationality. A government would want to do better being a group member than staying entirely outside of the group. This is the requirement of individual rationality (second principle). The requirement of collective rationality does not entail the requirement of individual rationality, as one country may stand to lose considerably from group membership, although all the other group members gain. Using principles like the first and second principle above, one may arrive at variable geometry or a flexible structure of clubs. However, a member state may wish to insert a decision rule, which will guarantee that it could never lose. Then we move to the political science perspective on the forming of clubs. Thus, if a government wishes to make sure that it never loses out, then it would want to have a veto. This would be the case when it has a general capacity to block any decision it considers non-

126

J.-E. Lane and R. Mæland

advantageous. However, if one member calls for a veto over certain issues, then how can it reject a veto for other members on issues they believe salient? And if the group endorses the veto principle of group decision-making, unanimity, then it will face staggering transaction costs threatening the requirement of collective rationality. The veto rule is not the only one that is conceivable in relation to the requirement of individual rationality. A government may accept to lose out in certain programmes if it can compensate itself in other programmes. Each government would wish to be a member of the group if membership results in more individual benefits than costs on the whole or over the long run. This requirement is more plausible than the veto rule, as it recognises that the group may have a variety of programmes, which add up differently for various member states in terms of benefits and costs. Governments would wish to maximise their gains from membership on an average or in the long run. Thus, governments would wish to know not only whether they can block programmes that are not advantageous but also whether they can introduce programmes that are advantageous. As suggested by Wicksell, a member of a group may be highly sensitive to losing upon certain issues while remaining lukewarm on gaining on other ones (Wicksell 1967). If this is the case, then it may wish to have a plethora of decision rules, e.g. choice rules which provide for some degree of veto power over certain issues or programme changes. When a government enters a group, then ideally it would want to have a group, which can take decisions involving more benefits than costs, but it would also wish to be able to block any decision that goes against its interests. Let us look at the political perspective on clubs, underlining decision-making as a group capacity. When groups have been constituted, they have a set of institutions, which deliver the rules of the game, namely voting. This is the focus of politics. Group decision-making is basically about winning or losing. Members look for opportunities to prevail. Thus, what is at stake is decidability, or scoring a winner and consequently a loser. Political efficiency is the capacity of a group to deliver a collective choice. Political efficiency avoids stalemate. It results when a group approves of one proposal and defeats another. A decisive group is one that can deliver decisions. A group using unanimity is not a decisive group in general. Political efficiency entails that you defeat your opponent. Thus, you do not negotiate with your opponents. The vote is the typical instrument of deciding within a group. One debates until one is convinced that only the vote can decide the issue or the outcomes. Voting is a zero sum game, where one coalition defeats another. It is not about economic efficiency, i.e. a variable sum game. The decisiveness of a group depends upon institutions first and foremost. What matters is both the number of votes cast and the aggregation rule employed to score a winner. Decisiveness means that in relation to a set of preferences a choice rule delivers an outcome. Preferences and coalitions are basically the same thing, as members will make the coalitions that correspond to their preferences. Focusing upon voting and group decidability the conditions for rationality must be reformulated. Thus, we have the requirement of group power (third principle),

Enlargement and the Council of Ministers

127

that is, the group must be capable to taking positive decisions. Given any configuration of preferences among the members the group will arrive at a decision to act through the voting procedure. This is group decisiveness, and it precludes the employment of the unanimity rule. Under unanimity there is only one coalition that can take action – the grand coalition. Thus, economic efficiency is different from political efficiency. Furthermore, we have the requirement of individual power (fourth principle), or no member will accept that it becomes a dummy. It will ask for as much individual voting power as it can get the others to accept. When quantitative voting schemes are used, then differences in the number of votes cast will be accepted, if they can be motivated on the basis of criteria that are systematically related to the accomplishment of the goals of the group, such as population size. This is individual decisiveness and any variation in the individual capacity to impact upon outcomes presupposes that one does not use the veto rule. Under unanimity, all choice participants have the same voting power. The constitution of a group needs not satisfy requirements of equity. Groups may be composed of players with different weights thus violating anonymity. And deviations from neutrality are conceivable in more ways than simply unanimity. The European Union has for a long time deviated from both neutrality and anonymity and the Nice framework constitutes no exception. It is regarded as essential that the bigger member states dispose of more votes (Council) or more representatives (Parliament) than small member states, although the exact formula for translating population differences into differences in terms of seats and votes is a contested matter due to the immense scale of the differences – compare Luxembourg with Germany. The Union employs qualified majority. When simple majority is employed, then it is equally likely that the status quo prevails or not. Deviations from simple majority entails that more votes are needed to change the status quo. One may go from 51 per cent all the way up to unanimity. The European Union has positioned itself in between qualified majority at 71 per cent and unanimity during its forty years’ existence. The Nice constitution does not change anything in principle but has added another rule that deviates from equity, namely the requirement of a population backing of 62 percent of the total Union population. Moreover, the Union employs weighted voting. One may give up the requirement of undifferentiatedness in order to single out the choice participants with different voting weights. When it is a matter of a group constituted by subgroups, then such a procedure appears appealing from the side of both self-interests and moral norms. When members carry different burdens in relation to a group, then they may raise a claim for having more votes than others. Quantitative voting results from the allocation of different votes to choice participants. The key question is what are the implications of differences in votes for the capacity to influence decisions? One cannot draw an immediate conclusion from differences in votes to a variation in power. How votes translate into power depends upon the aggregation rule employed in scoring a winner, given the preferences of the choice participants. We suggest one employs the power index approach to analyse this question (Lane and Mæland 2000).

128

J.-E. Lane and R. Mæland

Relevance of voting power This capacity to influence decision-making so that the acquis communautaire reflects the preferences of member states is closely connected with the institutions that govern collective choice in the EU. We will use the power index method to estimate a state’s capacity to influence decisions under alternative constitutional arrangements. N-person game theory forms the basis of models that study the interaction in collective decision-making processes such as those found in legislative bodies, all characterised by voting over different proposals. Thus, the main objective of these models must be to obtain a measure of the voting power of the individual decision-makers where committee decision-making is described in cooperative terms. Voting at any level involves discussion, debate, communications and coordinated action, which are all characteristics of cooperative games. Cooperative game theory models those situations where communication is a central feature of human interaction, admitting the formation of coalitions or a binding agreement among two or more players to coordinate their actions. Thus, the focus is on the coalition, on what gain a coalition may obtain (Shubik 1995; see Appendix 7). The Banzhaf index can be interpreted as the expected likelihood of being decisive in a political assembly. The capacity to be decisive consists of two kinds of power: the power to change, represented by the probability of decisiveness, and the power to block, represented by the individual probability of blocking. The probability of decisiveness δ is a monotonic decreasing (discrete step) function of the decision rule. An increasing decision rule lowers the capacity of power to change, represented by the probability of decisiveness, while the effect on the capacity of power to block is the opposite: the individual probability of blocking, is an increasing (discrete step) function of the decision rule, or, more precisely, an oscillating increasing trend. If we focus on decisive games, the final result on the capacity of a player to exercise decisive power, represented by the Banzhaf index, is negative: decreases with a stricter decision rule. Thus, the Banzhaf power indices are maximised at simple majority and decreases towards its minimum value 1/2 n–1, which obtains with the strictest decision rule: unanimity. The trade-off between increasing and decreasing is negative, that is, a rise in the capability of blocking power is offset by an even greater reduction in the capacity of power to change. A player’s voting power is always twice the individual power to block, multiplied by the power to change of a group. The basic relationships between these parameters may now be used to state some fundamental generalisations highly relevant to the analysis of constitutions. The core of any constitution is the rules about the votes the political players can cast in decision-making. The power index approach to identify and measure voting power in a group of choice participants such as a legislative assembly or committee focuses upon a few key parameters. These parameters have been derived axiomatically in Appendix 7 and we now proceed to apply them for the purposes of constitutional analysis. An increasing (stricter) decision rule, c, results in an increasing trend in the individual probability of blocking. Note that only the non-normalised Banzhaf index captures the drastic impact of the unanimity institution upon the power of the players, reducing it

Enlargement and the Council of Ministers

129

almost to nothing. The other indices only take the relative distribution of power into account, and not the absolute capacity of the group to take action.

The present situation The constitution of the Union contains the provision that enlargement decisions must be taken by unanimity. Thus, each member state has a veto on the entrance of new club members. A member state shall not find itself in a position that it will have to live with members it cannot put up with. In that respect, unanimity protects member states from adverse consequences of enlargement. However, should unanimity also be the decision-making method in the Council when it decides ordinary matters? If unanimity is employed in the normal operations of the Council, then each member state becomes a veto player. It can always block decisions that it considers adverse to its interests. However, unanimity implies that the system of different votes that states can cast becomes meaningless. A more profound reason though leads us to question the unanimity option. Consider Figure 7.1, which depicts the individual member state voting power. From Figure 7.1 appears the decisive argument against the use of unanimity as a constitutional provision for law-making in the Union. Although each member

Figure 7.1 Unanimity power: individual Banzhaf power indices.

130

J.-E. Lane and R. Mæland

state would have the highest possible power to block, the capacity of member states to take positive decisions would be small. And each enlargement would make it even more infinitesimally small. In a large club with thirty member states, the use of unanimity would make each member state, whatever its size, completely powerless. Figure 7.1 shows that each and every enlargement is a Paretoinferior move for each member state, from the point of view of voting power. Although it must be important for a member state to be able to block the entrance of new members that it does not like, each member state would have strong interests in being able to getting things done within the Union. When the Union makes policy, then a member state is not only interested in being able to stop what it does not want. A member state would also be interested in wanting to promote the policies it likes and see them enacted into law. This interest in making policy implies that a member state would not want each and every one to have a veto against its policies. If a member state would only be interested in blocking negative decisions, then it would do better by staying outside of the Union. If it wants a mighty Union, capable of making policy, then the unanimity option has to be abandoned, because it makes the Union completely powerless. The constitutional development of the EU reflects this basic contrast between the power to change and the power to block. In order to increase the first the Union has moved away from unanimity, but that involves reducing the power to block. This trade-off is unavoidable. When unanimity is no longer going to be the choice rule, then quantitative voting makes a difference. Using the power index method, we can analyse the present distribution of power within the Union and compare it with alternative enlargement scenarios. A number of enlargements seem possible for the EU, as the countries that have expressed an interest in becoming members include not only many East European countries but also Muslim countries around the Mediterranean as well as Turkey and Israel. We will not speculate about whether the EU could or would or even should accept this or that country. What we wish to pin down are the power implications of alternative extensions. Let us first consider the situation according to the Treaty of Amsterdam. The Council of Ministers is still the chief law-making body of the Union, as has always been the case. Since the Maastricht Treaty, the European Parliament can exercise a definitive veto in certain questions. Most matters are now decided by means of qualified majority voting, or 70 per cent of the votes in favour to arrive at a decision. The member states cast different amounts of votes. The voting power of a member state appears from Table 7.1. From following first the columns and then the rows in Table 7.1, we may observe some of the logic of decision-making in the Council. The first feature derives from the use of alternative aggregation rules, whereas the second characteristic is entailed in the quantitative voting schemes that the Council applies. Simple majority would entail much voting power. Banzhaf voting power, i.e. the combined capacity of a group to change and of an individual member state to block, is maximised by a simple majority scheme. Whereas the Council used to adhere to the so-called Luxembourg Agreement, consisting basically of unanimity, it has been almost completely abandoned in the treaty revisions since the 1980s,

Enlargement and the Council of Ministers

131

Table 7.1 Member state voting power in the Council: Banzhaf power Members

Votes

Simple majority

Qualified majority

Double qualified majority

• • • • • • • • • • • • • • •

10 10 10 10 8 5 5 5 5 4 4 3 3 3 2

0.326 0.326 0.326 0.326 0.254 0.156 0.156 0.156 0.156 0.130 0.130 0.092 0.092 0.092 0.061

0.113 0.113 0.113 0.113 0.093 0.059 0.059 0.059 0.059 0.048 0.048 0.036 0.036 0.036 0.023

0.098 0.098 0.098 0.098 0.083 0.059 0.059 0.059 0.059 0.051 0.051 0.042 0.042 0.042 0.030

87

2.782

1.011

0.968

Germany France Italy United Kingdom Spain Netherlands Belgium Greece Portugal Austria Sweden Finland Denmark Ireland Luxembourg

Sum

because unanimity implies almost nil Banzhaf voting power. Compare Table 7.1 with Figure 7.1 in so far as a fifteen-member state Union is concerned. Qualified majority, or double qualified majority has, however, one virtue in relation to simple majority, viz.: it raises the individual capacity to block (Table 7.2). It should be pointed out that the EU version of qualified majority voting (about 70 per cent) does give good protection against adverse decisions. The version with qualified-qualified majority presents the small states with a better capacity to block whereas the opposite is true of the large states. Table 7.2 The capacity to block in the EU Council Members

Votes

Simple majority

Qualified majority

Double qualified majority

• • • • • • • • • • • • • • •

10 10 10 10 8 5 5 5 5 4 4 3 3 3 2

0.326 0.326 0.326 0.326 0.254 0.156 0.156 0.156 0.156 0.130 0.130 0.092 0.092 0.092 0.061

0.725 0.725 0.725 0.725 0.601 0.382 0.382 0.382 0.382 0.311 0.311 0.233 0.233 0.233 0.147

0.698 0.698 0.698 0.698 0.588 0.420 0.420 0.420 0.420 0.360 0.360 0.297 0.297 0.297 0.214

0.500

0.078

0.070

Germany France Italy United Kingdom Spain Netherlands Belgium Greece Portugal Austria Sweden Finland Denmark Ireland Luxembourg

Probability of decisiveness

132

J.-E. Lane and R. Mæland

We observe from Table 7.2 that the blocking power of any member state rises rapidly as one moves from simple majority over qualified majority towards unanimity. Comparing Table 7.1 with Table 7.2, we observe that member states have to trade-off voting power against only veto power. The United Kingdom under Thatcher had one such trade-off that was very much different from the one that Germany has had under Kohl. At the same time, most countries do not seem prepared to move to simple majority, which maximises their voting power but minimises their veto power. Yet, such a possible trade-off has been mentioned in the deliberations about the necessary changes in the EU constitution before enlargement can take place. Member states can exercise different amounts of influence over EU policies. Thus, the Big Four have much more influence than the other eleven, as long as unanimity is not used. The differences between the member states in both voting power and the power to block may appear striking. However, the differences are in no way proportional to differences in population. Dividing the voting power by population, one arrives at an entirely different picture, namely that the small countries have more influence than the big countries, relatively speaking (see Table 7.3). With such huge differences in citizen influence as revealed in Table 7.3, one may speculate about future constitutional changes in the EU. For how long, one may ask, will the citizens in the populous countries in the Union accept that they have far less voting power than the citizens in the small countries? Either one can increase the votes of the huge states, or one can supplement decisionmaking in the Council with other bodies where population differences are more recognised.

Table 7.3 Banzhaf voting power related to population Members • • • • • • • • • • • • • • •

Germany France Italy United Kingdom Spain Netherlands Belgium Greece Portugal Austria Sweden Finland Denmark Ireland Luxembourg

Sum

Population (million)

Votes

Per capita vote (10–6)

Per capita power (10–6)

81.0 57.5 57.5 56.5 39.0 15.0 10.5 10.0 10.0 8.5 7.5 5.0 5.0 4.0 0.5

10 10 10 10 8 5 5 5 5 4 4 3 3 3 2

0.1235 0.1739 0.1739 0.1770 0.2051 0.3333 0.4762 0.5000 0.5000 0.4706 0.5333 0.6000 0.6000 0.7500 4.0000

0.001395 0.001965 0.001965 0.002000 0.002385 0.003933 0.005619 0.005900 0.005900 0.005647 0.006400 0.007200 0.007200 0.009000 0.046000

367.5

87

Enlargement and the Council of Ministers

133

Now, the key question is whether the Union can take on new members without changing either the principle for the allocation of votes or the principle for aggregating the votes into a final decision, or both. Thus, if the voting power of member states is not to be reduced, then the enlargement presupposes both a move towards simple majority and a redistribution of the votes in favour of the large states. Let us follow up to see what is involved in enlargement in principle and compare it with the Nice rules of 2000. The constant drive towards changing the aggregation rule, from unanimity towards various majoritarian formulas, is explainable simple as a Pareto improvement. In a fifteen-member state Union, each member state would have 0.000061 voting power under the unanimity institution. However, the qualified majority scheme presents the large states with 0.113 voting power and the small states with between 0.059 and 0.036 voting power. When the Union has moved from unanimity to qualified majority, then it has made a Pareto improvement.

Enlargement: Pareto-inferior and Pareto-superior changes The EU has been considering enlargement for quite some time. First, enlargement always implies less individual voting power as long as one does not change the aggregation rule. Although total voting power increases when new member states are added, this collective or group effect is not enough to cancel out the distributional effect of having to share voting power with new members. Second, enlargement implies more group power, only as long as one employs simple majority. The other side of the coin is the capacity to block in an enlarged EU. Again, we have the same tendency: the individual member state power to block decreases as power has to be shared with new incoming members. In order to salvage the principle of Pareto optimality with regard to voting power weights, one cannot simply add new member states to the existing institutions. Enlargement implies that more members share voting power, which almost always means that member states end up having less voting power. Thus, an enlargement is a Pareto-inferior movement as long as one maintains the same aggregation rule. The Nice meeting of the European Union in late 2000 resulted in constitutional revisions opening up for a much larger union from 2005. However, it did not accomplish a Pareto improvement. Let us analyse the situation in 2005 assuming that an enlarged Union is formed on the basis of the Nice constitution. Table 7.4 has the voting power scores. The outcome is actually a Pareto deterioration and it reflects two forces. On the one hand, as more members enter the voting power available has to be split upon more actors. On the other hand, the orientation of the Union towards qualified majority including the new requirement of 62 per cent of the population increases the majority requirement to 73.5 per cent and drives down group decisiveness. The Nice constitution makes the Union almost incapable of taking action, involving a drastic reduction in group decisiveness, from a probability of a decision at 0.078 to 0.017. There is one clear set of winners in the Nice framework, namely the big states which search for a mechanism to prevent changes to the existing Union from the enlargement of the Union. Thus, the blocking

134

J.-E. Lane and R. Mæland

Table 7.4 Voting power 2005 Members

Pop.

Votes

Banzhaf

Banzhaf norm.

Banzhaf/ population

Capacity to block

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

82.2 58.9 58.7 57.3 39.6 39.0 22.5 15.7 10.6 10.3 10.2 10.1 10.0 8.9 8.2 8.2 5.4 5.3 5.2 3.7 3.7 2.4 2.0 1.4 0.7 0.4 0.4

29 29 29 29 27 27 14 13 12 12 12 12 12 10 10 10 7 7 7 7 7 4 4 4 4 4 3

0.033 0.033 0.033 0.033 0.031 0.031 0.018 0.017 0.015 0.015 0.015 0.015 0.015 0.013 0.013 0.013 0.009 0.009 0.009 0.009 0.009 0.005 0.005 0.005 0.005 0.005 0.004

0.078 0.078 0.078 0.078 0.074 0.074 0.043 0.040 0.037 0.037 0.037 0.037 0.037 0.031 0.031 0.031 0.022 0.022 0.022 0.022 0.022 0.013 0.013 0.013 0.013 0.013 0.009

0.040 0.055 0.056 0.057 0.079 0.080 0.080 0.106 0.146 0.150 0.152 0.153 0.155 0.146 0.158 0.158 0.170 0.173 0.176 0.248 0.248 0.219 0.263 0.375 0.750 1.313 0.989

0.807 0.807 0.807 0.807 0.769 0.769 0.442 0.412 0.382 0.382 0.382 0.382 0.382 0.321 0.321 0.321 0.226 0.226 0.226 0.226 0.226 0.130 0.130 0.130 0.130 0.130 0.098

481.0

345

0.420

1.000

Germany United Kingdom France Italy Spain Poland Romania Netherlands Greece Czech Republic Belgium Hungary Portugal Sweden Austria Bulgaria Slovakia Denmark Finland Ireland Lithuania Latvia Slovenia Estonia Cyprus Luxembourg Malta

Sum

power of the big states is sharply up, making them almost veto players. For the small countries the new allocation of voting makes them almost dummies. The probability that Denmark or Finland casts a decisive vote is 0.008 – a drastic reduction in comparison with the present situation. However, their power to block is not reduced. In the rhetoric of the Nice meeting it was stated that the new rules will empower the Union, increasing its capacity to act in a quasi-federal manner. In reality, things are entirely different. Despite the emphasis upon qualified majority instead of unanimity, the Union has chosen a constitution, which is close to the unanimity outcome. In reality, all member states will lose voting power in 2005. The Nice rules entail a Pareto-inferior change. The power to block has been considered more important than the power to change, at least by the dominant member states. Yet the members of the EU are driven by an inexorable logic to make institutional changes as it extends the club to new members. Although the acceptance of a new state as member falls under the unanimity prerogatives, having unanimity or almost unanimity as decision rule for day-to-day policymaking in the

Enlargement and the Council of Ministers

135

Council of Ministers will not work. Adding new members empties the club of all group power. The only remedy is to accept more and more other decision rules, closer to simple majority. However, any majoritarian scheme involves that a member state can lose in a vote in the Council, which goes against the requirement that states can be sure that their vital interests will not be hurt in the Union. Thus, decisiveness conflicts with vital interests protection. The reform of the institutions of the Union allowing for majority decisionmaking implies that decisiveness will be traded off against state sovereignty. The closer the Union moves towards simple majority voting, the more the group is capable of taking decisions. At the same time, the risk for vital interests protection and state sovereignty increases. There is no way around this conflict between group decisiveness and individual member state veto. Table 7.5 shows how an enlarged Union could be based upon Pareto improvements in voting power. As long as one reforms the institutions towards accepting simple majority at the same time as one includes more members one may be able to make Pareto improvements. However, there is a limit to how far one can go. Once one reaches simple majority, then any enlargement entails a Pareto-inferior change. Table 7.5 The attractiveness of simple majority Members

Pop.

Votes

Banzhaf

Banzhaf norm.

Banzhaf/ population

Capacity to block

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

82.2 58.9 58.7 57.3 39.6 39.0 22.5 15.7 10.6 10.3 10.2 10.1 10.0 8.9 8.2 8.2 5.4 5.3 5.2 3.7 3.7 2.4 2.0 1.4 0.7 0.4 0.4

29 29 29 29 27 27 14 13 12 12 12 12 12 10 10 10 7 7 7 7 7 4 4 4 4 4 3

0.292 0.292 0.292 0.292 0.270 0.270 0.136 0.126 0.116 0.116 0.116 0.116 0.116 0.097 0.097 0.097 0.068 0.068 0.068 0.068 0.068 0.038 0.038 0.038 0.038 0.038 0.029

0.086 0.086 0.086 0.086 0.079 0.079 0.040 0.037 0.034 0.034 0.034 0.034 0.034 0.028 0.028 0.028 0.020 0.020 0.020 0.020 0.020 0.011 0.011 0.011 0.011 0.011 0.009

0.355 0.495 0.497 0.509 0.681 0.692 0.604 0.804 1.098 1.130 1.141 1.152 1.163 1.087 1.180 1.180 1.250 1.274 1.298 1.825 1.825 1.602 1.923 2.746 5.493 9.613 7.235

0.292 0.292 0.292 0.292 0.270 0.270 0.136 0.126 0.116 0.116 0.116 0.116 0.116 0.097 0.097 0.097 0.068 0.068 0.068 0.068 0.068 0.038 0.038 0.038 0.038 0.038 0.029

481.0

345

3.400

1.000

Germany United Kingdom France Italy Spain Poland Romania Netherlands Greece Czech Republic Belgium Hungary Portugal Sweden Austria Bulgaria Slovakia Denmark Finland Ireland Lithuania Latvia Slovenia Estonia Cyprus Luxembourg Malta

Sum

136

J.-E. Lane and R. Mæland

Under quantitative voting institutions, simple majority makes the most difference in enhancing the voting power of the player with the most votes to cast. Thus, moving away from unanimity is in the interest of the big states at the same time as it is in the interest of all states, as long as no one of them has a particular interest in being able to block some changes of the status quo. Yet the big member states may require that future enlargements not only change the aggregation rule but also redistributes the votes cast by each state in order to reflect population size better or reflect the money contributed to the EU. Any redistribution of votes between member states, existing or new, cannot be motivated on the basis of Paretian criteria. Such a redistribution would involve preferring one attribution of votes ahead of another on the basis of an argument about justice. Since there is no agreement about what justice implies in terms of voting weights in a union, there is bound to be considerable distributional conflict when the EU begins to discuss alternatives to the existing allocation of votes in the Council.

Conclusion Enlargement of a union has institutional implications. The existing member states need not only come together and accept unanimously a new member, but they also have to be aware of the institutional consequences of enlarging the number of members. The EU is a club of states in Western Europe that has decided it wishes to go towards Eastern Europe to find new willing members. Enlarging the Union has all sorts of implications: economic, social and political. But here we have focused upon the institutional consequences from the point of view of voting power. At the beginning, we laid down nine institutional requirements that are relevant for evaluating alternative frameworks for the enlargements of the Union. Emphasising state influence, decisiveness and efficiency we have examined various reforms of the voting procedure in the Ministry of Council. Cooperative game theory allows us to state the implications of alternative voting institutions, also in an enlargement perspective. We argue that the requirement of vital interests’ protection must be abandoned. Any enlargement of the present fifteen-member Union should be accompanied by institutional reform, accepting more of simple majority voting in the Council of Ministers. Unanimity can only be used for unique decisions about the constitution of the Union. The larger the Union becomes, the greater will be the pressure to accept simple majority voting. This follows from Paretian considerations about making enlargement a Pareto-superior change in voting power. The logic of the rules arrived at the Nice meeting is different as they reduce voting power in general. To make the Union more viable the Nice rules may have to be revised sooner or later as they emphasise the power to block too much at the expense of the power to change or group decisiveness. Enlargement and institutional change are linked not only in terms of Pareto efficiency but also through the relevance of distributional considerations. When the Council of Ministers employs a majoritarian scheme, then quantitative voting

Enlargement and the Council of Ministers

137

makes a huge difference for a member state’s voting power. Enlargement presents the huge states with an opportunity to call for a redistribution of votes, increasing their number and decreasing that of the small states. When the present voting power of the various states in the EU are related to the population size of the member states, then it is crystal clear that voting power is not proportional to country size. When states come together and collaborate in institutional forms, then one tends to think about federal or unitary solutions. However, this dichotomy is too simplistic. Creating or enlarging a club of states involves the resolution of the fundamental problems about economic efficiency as well as political efficiency. Different forms of clubs of governments are conceivable as a function of and alternative rules about group decisiveness and member state blocking power in decision-making. The choice of decision institutions in a club reflects or expresses what the members wish the club to be. The Nice rules favour the capacity to block ahead of club power. If a federal solution for Europe is ever enacted, then the preference for group decisiveness must go ahead of the individual member’s capacity to block.

APPENDIX 7 Given a set of players, N = {1,…, n}, collective choice in a political assembly can be modelled as a simple game, (a voting game), where subsets, S ⊆ N , of players form winning coalitions (or win sets). These winning coalitions are assigned the total value of the game, normalised to one, while the losing coalitions get nothing. The possible gain of a simple game is given by a characteristic function υ : 2N → {0, 1},

 1 if S is winning, υ (S) =   0 otherwise,

(7.1)

with υ(Ø) = 0 and υ (N) = 1. The characteristic function combined with the set of players, the pair (N, υ), will then be a proper definition of a simple game. Furthermore, we observe that simple games are constant sum games where υ (N) = υ (S) + υ (N \ S),

(7.2)

which implies υ (S) = 1 – υ (N \ S).

(7.3)

A general form of simple games is the weighted majority game which is the basic model used to study power distributions among players (voters) when individual amounts of votes may differ. A player i ∈ N thus casts wi votes, while a decision rule c decides the quota of votes needed to pass a bill. In other words, in a weighted majority game (c, w) = [c; w1,…, wn], the characteristic function has the form

138

J.-E. Lane and R. Mæland

∑

 1 wi ≥ c i∈S υ (S)  ,  0 otherwise.

(7.4)

A solution concept to these simple game models is the power index. A power index measures individual voting power in a decision-making system where voting power is the capability of a player (voter) to be decisive in a voting session. It is an a priori measure that summarises the results of all possible plays of a game, that is, all 2 N possible coalitions are checked for decisive players. The Banzhaf index considers a set of equally probable coalitions, while the Shapley-Shubik index considers equally likely permutations of the same coalitions. Thus each permutation is given weights, which are functions of coalition size (the cardinality |S|). Every permutation has only one pivotal (decisive) player making the Shapley-Shubik index a relative power measure, which always sum to one. On the other hand, the coalitions considered by the Banzhaf index may include several decisive players making the index an absolute power measure which only in exceptional cases sums to one. The preferred measures voting power have been relative power indices such as the Shapley-Shubik index or a normalised version of the Banzhaf index. However, the normalisation of the Banzhaf index distorts important aspects of this index, removing information on collective decisiveness as well as information on individual blocking capabilities. A formal presentation of the Banzhaf index is obtained by considering the power set 2N of all possible and by assumption equally likely coalitions S ⊆ N, the empty set {Ø} included. From these 2n coalitions we find the Banzhaf score of player i as the number of coalitions in which i is critical, that is, the number of coalitions that player i is able to swing. There are two types of swings present, on the one hand type (a) swings where player i turns a winning coalition into losing by leaving it, and, on the other hand, type (b) swings where player i turns a losing coalition into winning by joining it. Mathematically, given the set of winning coalitions W(N,υ ) = { S ⊆ N | υ(S) = 1},

(7.5)

type (a) and type (b) swings are represented by (S ∈ W(N, υ) , S \ {i} ∉ W(N, υ) ) respectively (S ∉ W(N,υ), S ∪ {i} ∈ W(N, υ) ). However, because of symmetry, the number of the two types of swings will always be equal and we may therefore focus on just the count of type (a) swings. A simple proof of this symmetry, is obtained by considering the so called dual game, (N, υ∗) where the dual characteristic function is defined as υ* (S) = υ(N) - υ(N \ S).

(7.6)

By (7.2) and (7.6) we find that a constant sum game (N, υ) is self dual: (N, υ*) = (N, υ).

(7.7)

Enlargement and the Council of Ministers

139

From the original game, (N, υ), the number of type (a) swings of player i is given by the cardinality ηi = |{S ⊆ N | S ∈ W(N, υ), S \ {i} ∉ W(N, υ)}|,

(7.8)

and is readily obtained by summing over all possible coalitions

ηi =

∑ [υ (S) − υ (S \ {i})]

S⊆ N i ∈S

(7.9)

Comparing the original game, (N, υ), with its dual (N, υ*) it is evident that the type (b) swings of the original game are identical with the type (a) swings of the dual game, say ηi*, ηi* = |{S ⊆ N | S ∈ W(N, υ∗), S \ {i} ∉ W(N, υ∗)}| = |{S ⊆ N | S ∈ W(N, υ), S \ {i} ∉ W(N, υ)}| = ηi,

(7.10)

where the second equality is obtained by plug-in of the self dual result, (7.7). Hence, the symmetry between type (a) and type (b) swings. Now, assume equally likely coalitions S ⊆ N then the Banzhaf power index of player i is defined by the ratio

βi =

ηi 2n −1

,

(7.11)

or, inserting (7.9), by the commonly used characteristic function form

βi =

1 2

n −1

∑ [υ (S) − υ (S \ {i})]

S⊆ N i ∈S

(7.12)

An interesting implication is that the sum of the individual Banzhaf indices diminishes with increasing decision rule, c. Moving from simple majority rule via qualified majority rules to the strictest unanimity rule has a substantial negative effect on this sum. To see this, we note that a unanimity game, UN, has only one winning coalition, the grand coalition N, which all players are able to swing, that is, ηi equals one for all i. Note, the counterpart type (b) swing, player i is able to turn a losing coalition N \ {i} into winning by joining. Thus, the (minimised) unanimity Banzhaf index simplifies to

β i |U N =

1 2

n −1

.

(7.13)

The individual unanimity power will be small for even a modest number of players. A question one may now ask is if it is possible to find some general (collective) effect that the decision rule has on the Banzhaf index. We want a

140

J.-E. Lane and R. Mæland

number that, given a certain decision rule, c, measures the capability of an assembly to form winning coalitions. A simple indicator of this capability is just a simple count of the winning coalitions of a game, d = |W(N, υ)|, which is obtained by

∑

d=

υ(S)

S⊆ N

(7.14)

Relating this number to the total number of possible coalitions of a game we obtain what we the call the probability of decisiveness of a game,

δ=

d 2n

.

(7.15)

With no restrictions on c, absolute decisiveness, δ = 1, is obtained by the improper game of a zero decision rule. In this uninteresting case all proposals will be automatically accepted. The class of improper games are simple games where at least two sets may simultaneously form winning coalitions thus implying a decision rule less than 1/2. A decisive game, on the other hand, can only have one winning coalition at the time. In this class of games, δ has a maximum value of 1/2, which obtains by simple majority decision rule, or in cases where one player has dictatorial power. Not surprisingly, minimum decisiveness obtains with unanimity decision rule. With only one winning coalition minimum δ is equal to 1/2n, and, as the number of player increases, the probability of decisiveness approaches nil quite rapidly, lim {δ | U N } = 0. n→∞

(7.16)

The apparently high relative power connected with unanimity decision rule actually leaves players quite powerless when probability of decisiveness is considered. Our next step is to find the individual factor, which combined with δ results in the Banzhaf index. The number of individual swings, ηi, is a crucial factor in this index and relating this number with the total number of winning coalitions, d, we obtain what we define as the individual probability of blocking

θi =

ηi d

(7.17)

Using (7.15) and (7.17) in (7.11) we see that the Banzhaf index can be factorised into the product of the probability of decisiveness, δ, and twice the individual probability of blocking, θi, βi = 2δθi.

(7.18)

The doubling of the individual probability of blocking reflects the fact that the Banzhaf index is based on the two symmetric capabilities to swing a game,

Enlargement and the Council of Ministers

141

Banzhaf (1965). A slightly different interpretation of (7.18) is given by the dual game approach to the type (b) swings where we interpret a dual to the probability of decisiveness as a collective probability to block, δ*, d* then denotes the total number blocking coalitions. Following this line of thought, the dual to the individual probability of blocking will then become an individual probability to break a blocking coalition, θi*. Hence, we may define the Banzhaf index by the sum of these two products βi = δθi + δ* θi*.

(7.19)

Or, following the theory of simple games, we would say that the Banzhaf equals twice the product of the probability of decisiveness, δ, and the individual probability of blocking, θi. A similar measure to our blocking probability is the inclusiveness index by König and Bräuninger (1998). This index is a simple measure of a player’s participation in winning coalitions and defined by

wi =

di d

(7.20)

where d is the number of winning coalitions (7.14) and di is the number of winning coalitions where i is a member, di = |{S ∈ W(N, υ), i ∈ S}|.

(7.21)

Formally, the size of wi is limited by the boundary conditions

1 ≤ wi ≤ 1. 2

(7.22)

To see this, consider an arbitrary player, player k, say. In any voting situation, this player votes yes and no with equal probability. Thus, half of all 2n possible coalitions will have player k as a member. If player k has dictatorial power in a game, all coalitions where he/she is a member will be winning coalitions, while all the other will be losing coalitions, thus dk = d, which results in maximal inclusiveness: wk = 1. When player k has dictatorial power, all other players i ∈ N \ {k} of the game above will be dummy players, accordingly, exactly half of the winning coalitions will have the dummy player i as a member. Hence, for all dummy players i ∈ N \ {k} the value of di will be equal to d/2 and the minimum inclusiveness value, wi = 1/2, follows. To study how the probability of decisiveness enters the Banzhaf index we consider the alternative definition of this index where we let player i turn a losing coalition S ⊆ N \ {i} into winning by joining it (type (b) swings). By symmetry ηi is given by ηi = |{S ⊆ N \ {i} | S ∉ W(N, υ), S ∪ {i} ∈ W(N, υ)}|,

(7.23)

142

J.-E. Lane and R. Mæland

which is obtained by the summation

ηi =

∑ [υ (S ∪{i}) − υ (S)],

S ⊆ N \{ i}

(7.24)

where the first term is the number of winning coalitions where i is a member, cf. the cardinality (7.21),

∑

υ (S ∪ {i}) = di ,

S ⊆ N \{ i}

(7.25)

while the second term of the summation is the number of winning coalitions where i is not a member. By (7.14) we have

∑

υ (S) = d − di .

S ⊆ N \{ i}

(7.26)

Inserting (7.25) and (7.26) back into (7.24) we obtain identity ηi ≡ 2di – d,

(7.27)

This is a central element in the so-called Chow’s theorem. This theorem tells us that ηi and d will distinguish any weighted game from all other simple games (Dubey and Shapley 1979). Dividing through this identity by d, using (7.17) and (7.20), we find a transformation between the inclusiveness index and the individual probability of blocking given by θi = 2wi –1.

(7.28)

Finally, by inserting (7.28) into (7.18), we define the relationship between the inclusiveness and the Banzhaf indices by the following relationship: βI = 2δ (2wi – 1).

(7.29)

By Equation (7.29), we observe how the behaviour of the inclusiveness index above applies to the Banzhaf index and, by Equation (7.28), we furthermore observe how this behaviour pattern transforms directly to our individual probability of blocking, θi.

References Banzhaf, J.F. III (1965) ‘Weighted voting doesn’t work: a mathematical analysis’, Rutgers Law Review, 19: 317–43. Cornes, R. and T. Sandler (1986) The Theory of Externalities, Public Goods and Club Goods, Cambridge: Cambridge University Press. Deegan Jr., J. and E.W. Packel (1979) ‘A new index of power for simple n-person games’, International Journal of Game Theory 7: 113–23.

Enlargement and the Council of Ministers

143

Dubey, P. and L.S. Shapley (1979) ‘Mathematical properties of the Banzhaf power index’, Mathematics of Operations Research 4: 99–131. König, T. and T. Bräuninger (1998) ‘The inclusiveness of European decision rules’, Journal of Theoretical Politics 10: 125–42. Lane, J.E. and R. Mæland (2000) ‘Constitutional analysis: the power index approach’, European Journal of Political Research 37: 31–56. Shapley, L.S. (1953) ‘A value for n-person games’, Annals of Mathematics Studies 28: 307–17. Shapley, L.S. and M. Shubik (1954) ‘A method for evaluating the distribution of power in a committee system’, American Political Science Review 48: 787–92. Shubik, M. (1995) Game Theory in the Social Sciences, Cambridge, MA: MIT Press. Wicksell, K. (1967) ‘A new principle of just taxation’, in: R.A. Musgrave and A.T. Peacock (eds) Classics in the Theory of Public Finance, New York: St. Martin’s Press, 72–118.

144

Authors name

8

Enlargement and the European budget Budgetary decision-making and fiscal constraints Lars P. Feld, Gebhard Kirchgässner and Hannelore Weck-Hannemann

Introduction Although some national referenda, e.g. in Denmark, France and Ireland, were held for a ratification of treaty changes in the European Union (EU), direct legislation is highly neglected in most European Union countries as a viable supplement to purely parliamentary decision-making. Moreover, whenever the democratic deficit in the EU is addressed, an increase of the power of the European Parliament is regularly proposed without considering referenda and initiatives as institutions that would not only reduce the democratic deficit, but also start the creation of a ‘European Public’ for EU decisions (Hauser and Müller 1995; Grimm 1997).1 Only a few authors propose the introduction of referenda and initiatives at the EU level in order to enhance the legitimacy and the efficiency of political decisions in the EU (Kriesi 1991; Schneider 1993; Feld and Kirchgässner 1996). There is some discussion among public choice scholars as to the conditions under which direct democracy leads to policy outcomes that are closer to citizens’ preferences (Romer and Rosenthal 1978, 1979; Ingberman 1985; Steunenberg 1992; Moser 1996) or increases the information of representatives about constituents’ preferences (Matsusaka 1992; Matsusaka and McCarty 2001). The former is also systematically analysed in two empirical studies. First, Pommerehne (1978) shows with aggregate data on Swiss cities in 1970 that the median voter model performs better in jurisdictions in which political decisions are made by voters directly in local assemblies or in referenda and initiatives. Second, Gerber (1999) and Gerber and Hug (1999) present evidence for the US strongly supporting Pommerehne’s result. In the case of three political issues, parental consent laws, the death penalty and gay rights, they show that initiatives correct policy outcomes towards the preferences of the median voter. These results imply that citizens’ preferences are enforced in political systems with direct legislation to a larger extent than in those without. On the basis of these results, it could be expected that policy outcomes in representative democracies with and without direct legislation differ from each other. Systematic comparative studies2 for Switzerland and the US find that, first, public expenditure is, ceteris paribus, lower in direct than in representative democracies (Matsusaka 1995; Feld and Matsusaka 2001; Feld and Kirchgässner

Enlargement and the European budget

145

2001a). Second, public revenue of initiative states is composed to a higher extent of charges than of broad-based taxes (Matsusaka 1995). Third, public debt is lower in direct democratic than in representative democratic jurisdictions (Kiewiet and Szakaly 1996; Feld and Kirchgässner 1999, 2001b). The question remains whether the lower level of public spending also leads to a more efficient public sector. Studies that attempt to empirically investigate efficiency of political institutions, even in the restricted sense of cost efficiency, are rare.3 Pommerehne (1983) analyses whether costs and prices of local garbage collection differ in 103 Swiss cities in 1970 according to its different organisation. He finds that average refuse collection costs (per household) are, ceteris paribus, the lowest in cities with direct legislation and private garbage collection. These costs are about 10 per cent higher if garbage collection is organised publicly instead of privately. In cities with representative democracy the costs of private garbage collection is about 20 per cent higher than in direct democracies. Average costs of garbage collection are the highest in cities with representative democracies and public organisation (30 per cent higher costs than in the first case). Moreover, Pommerehne and Weck-Hannemann (1996) show that tax evasion is lower in Swiss cantons in the years 1965, 1970 and 1978 in which citizens have an impact on budgetary policy in direct legislation. In those cantons, tax evasion is, ceteris paribus, about SFr 1,500 lower as compared to the average of those cantons per year without such a direct influence of citizens on fiscal policy.4 If, however, the willingness to pay taxes is higher, the more satisfied citizens are with public services supplied, then these results are evidence for a higher satisfaction of citizens and thus for a higher efficiency of the provision of public services in direct democratic systems. Frey and Stutzer (2000) find indeed that citizens’ satisfaction with their lives is, ceteris paribus, higher in direct democracies. These latter studies lend support to the hypothesis that direct democratic systems are more efficient than representative democratic ones. Does this also hold in a more generalised perspective? Is aggregate economic performance higher in jurisdictions with direct legislation? Feld and Savioz (1997) study the relationship between budgetary referenda and economic performance of Swiss cantons measured by GDP per employee. In a panel with annual data from 1984 to 1993 for the twenty-six Swiss cantons, they obtain the result that GDP per employee is, ceteris paribus, by about 5 per cent higher in those cantons with budgetary referenda as compared to the cantons without those referenda. This result is tested in various ways on its robustness. In particular, reversed causality is tested hypothesising that the richer cantons could afford more direct legislation. On the basis of the empirical results, this hypothesis could be rejected. Moreover, the impact of direct democracy is hardly decreased if additional explanatory variables are included in the empirical model. These results imply that political competition between political parties in a representative democracy is not sufficiently perfect to enforce median voters’ preferences. The restriction of competition may result from a collusion between representatives, such that representatives have leeway to follow their self-interest between elections. The referendum results in an effective restraint on the collusion ability of the government and the relevant parties in the parliament. It prevents

146

L.P. Feld, G. Kirchgässner and H. Weck-Hannemann

political decisions that result from such a collusion. The initiative creates new policy outcomes and prevents a collusion of representatives by additionally weakening their agenda-setting power. In presidential and parliamentary systems, representatives in the government are controlled by representatives in the parliament. In direct democratic systems, citizens control and sanction their representatives directly. Therefore, these studies also indicate that an introduction of referenda and initiatives at the EU level could have some merits. In this chapter, such a constitutional reform is proposed for the EU budgetary process. We consider budgetary referenda alone instead of a general introduction of referenda for two reasons. First, the power to tax, spend and issue debt has always been one of the most crucial powers of policymakers. Without financial resources political agents cannot realise their policy goals. Consequently, kings and queens in history did not want to transfer budgetary powers to parliaments. Second, the empirical studies conducted at Swiss sub-federal levels mainly consider budgetary referenda. The favourable impact of direct legislation on economic policy outcomes can therefore be mainly attributed to budgetary referenda. While the impact of balanced budget rules and of a strong position of the finance minister in the budgetary process on budgetary outcomes are intensively discussed (Von Hagen and Harden 1995; Poterba 1997) and policy recommendations are deducted from this discussion (Inman 1997), direct legislation is not considered as an alternative to tighten the budgetary process. This contribution is the first attempt to specifically propose budgetary referenda as a constitutional alternative to balanced budget rules and analyse their impact on the decisionmaking power of players in the budgetary game. This impact is analysed in a spatial voting model in the tradition of Shepsle (1979), Weingast (1989) and Moser (2000). Spatial voting theory has been applied in the analysis of budgetary referenda5 and of EU legislative procedures,6 but in neither that of EU budgetary decision-making nor of the impact of an introduction of referenda in this procedure. The chapter is organised as follows: after describing the budgetary procedures (for compulsory and non-compulsory expenditures) in the EU and discussing how a referendum might enter it, we analyse both procedures in a spatial model and discuss the impact of the referendum on the position of the different EU bodies involved in the process. Subsequently, we discuss the effects of the introduction of direct legislation in the budgetary process on the enlargement of the EU. We conclude with some final remarks.

The EU budgetary process The EU has no power to tax or to issue bonds. It receives its revenue mainly from four sources. First, levies on agricultural and sugar trade, which are transferred directly to the EU. Second, customs duties on trade with third countries. Third, resources accruing from a proportion of VAT on the basis of a uniform VAT base throughout the Union. Those three revenue sources are the traditional so called ‘own resources’. A new fourth resource is based on a topping-up of the revenue

Enlargement and the European budget

147

available from the other sources and related to GNP. The rate of this resource is determined each year during the budgetary procedure in the light of the total amount available from all other sources of revenue related to the total GNP of all member states. In 1999, revenue from levies and customs duties amounted to 16 per cent, revenue from the VAT resource to 35.9 per cent and revenue from the GNP resource to 43.2 per cent of EU revenue (Commission of the European Union 2000). Although the EU cannot incur public debt officially, it uses several external financial instruments to get resources on financial markets. The most notable offbudget activities are those of the European Investment Bank, which amounted to !31.9 Billion and thus nearly 38 per cent of the official EU budget in 1999 (Commission of the European Union 2000). Simplified, the European Investment Bank borrows money for the Commission on financial markets to grant loans to member countries, to specific industries in member countries, but also to nonmember countries like the Eastern and Central European countries. Total external EU borrowing has increased from !3.9 Billion in 1980 to !31.9 Billion in 1999 (Commission of the European Union 2000). All in all, EU budgetary decision-making thus mainly consists of decisions about the structure of expenditure while the level of spending is more or less predetermined by financial resources granted by member countries. Nevertheless, the increasing external borrowing activities lend support to the view that EU authorities will manage to expand their power to issue bonds and perhaps also to tax. Thus, the EU budgetary process will become increasingly important in the future of European politics. The budgetary procedure is established by Article 203 of the EU Treaty, which sets out a precise timetable and procedure for drafting the annual budget (Laffan 1997: Chapter 4). Decision-making power on budgetary matters is divided between the Council and the European Parliament. The Commission works out a draft budget and proceeds it to the Council, which opens the formal budgetary procedure by reaching a first decision on the budget. It is then sent to the Parliament for a first reading. The Council gets the budget for a second reading and sends it back to the Parliament. The Parliament has the final say in the budgetary process by adopting or rejecting the budget. The Council and the Parliament have different powers over different parts of the budget depending on the categorisation of expenditure. The Council determines so-called ‘compulsory’ spending – expenditure necessarily arising from the treaties. Compulsory spending consists mainly of guaranteed agricultural expenditure, i.e. direct subsidies to the agricultural sectors of the member states mounting to 46.5 per cent of EU expenditure in 1999 (Commission of the European Union 2000). The Parliament determines ‘non-compulsory’ spending subject to a ‘maximum rate of increase’ decided upon each year on the basis of a statistical calculation. The ‘maximum rate of increase’ has lost importance due to the establishment of expenditure ceilings in the financial perspective. The distribution of budgetary power between the Council and the Parliament is a permanent source of conflict between the two institutions (Laffan 1997: 70).

148

L.P. Feld, G. Kirchgässner and H. Weck-Hannemann

The sequential structure of the budgetary process for compulsory expenditure is illustrated in Figure 8.1.7 The budget cycle starts with conciliations among the three EU institutions concerned with budget-making, the so-called interinstitutional trialogue on the priorities for the budget, which is followed by two ad hoc conciliation meetings on compulsory expenditures. The Commission prepares a Preliminary Draft Budget (PDB) for submission to the budgetary

Key: C = Commission; P = a vote in the Parliament; 1 = county 1; p = the EU policy; V = a vote in the Council; q = the status quo; BCOM(.) = the Commission’s proposal strategy; Ecl(.) = country one’s amendment stragegy for expenditure; Vci(.) = country one’s voting strategy on the amendment; VP(.) = the Parliament’s voting strategy on the common position; VPci(.) = country one’s voting strategy on the amendments of the Parliament.

Figure 8.1 Decision structure of EU compulsory expenditures.

Enlargement and the European budget

149

authority (BCOM). Before sending the PDB to the Council, it calculates the maximum rate of increase and collects expenditure estimates from EU institutions. After the submission of the budget, but before the first reading of the budget by the Council, another conciliation between the Council and the Parliament takes place. During the first reading in the Council, each EU member state either proposes amendments or does not propose any.8 The Council accepts or rejects the proposed amendments by its qualified majority procedure (at the moment 62 weighted votes) and forwards the draft budget to the Parliament. Consider first the budgeting of compulsory expenditures and the case of the Council accepting amendments by one or more member countries. The Parliament either proposes an amendment of the draft budget or does not propose any. If it does not propose an amendment, country i’s voting strategy (Eci) will succeed: the Council and the Parliament will accept this budget in their second readings. If the Parliament proposes an amendment by simple majority of the votes, which is usually the case, and after another conciliation meeting between Council and Parliament, the Council decides upon the budget in its second reading. If the Council accepts the proposal(s) by the parliament, the budget is adopted because the final vote by the Parliament will only be a formal agreement. The voting strategy of the Parliament on the common position (VP) succeeds. If the Council rejects the amendments of compulsory spending by the Parliament with its qualified majority, the Parliament can only reject the whole budget by a two-thirds qualified majority or accept the Council’s rejection of its amendments and adopt the budget. In the first case, the outcome of the budget cycle is the status quo q, which is the budget of the year before.9 In the second case, country i’s voting strategy (VPci) will succeed. The case if the Council rejects amendments proposed by member states in its second reading follows in an analogous fashion. The policy outcomes are the same with the exception that instead of a success of country i’s voting strategy (VPci) the Commission’s proposal strategy (BCOM) will succeed. Consider now the budgetary process for non-compulsory expenditures in Figure 8.2. The budgetary procedure is the same as in the case of compulsory expenditures. The difference between both expenditure categories lies in the rule relating to which EU institution has the final say over expenditures. In the case of compulsory expenditures, the Council has the final say and the Parliament can only reject the whole budget. In the case of non-compulsory expenditures, the Parliament has the final say. If the Parliament rejects the rejection of its first reading amendments of non-compulsory expenditures by the Council and accepts the budget, the budget is finally adopted by a simple majority of the members of the Parliament and a three-fifths majority of votes. Then, the voting strategy of the Parliament on the common position (VP) succeeds instead of country i’s voting strategy (VPci) or the proposal strategy of the Commission (BCOM) in the case the Council has accepted or rejected the amendments by single member states in the Council’s first reading. Basically, a fiscal referendum may enter a budgetary procedure threefold.10 First, citizens can vote on the whole budget after the legislature has adopted it. The budget will then finally be adopted by voters or it has to be revised in the

150

L.P. Feld, G. Kirchgässner and H. Weck-Hannemann

Figure 8.2 Decision structure of EU non-compulsory expenditures.

legislature in order to get voter approval. Second, a constitution can impose the obligation on representatives to ask voters in a referendum for tax increases or more generally changes in tax laws. Third, voters can vote in a referendum ex post after the fiscal year in which the actual revenue and expenditure occurred in order to finally approve the budget deficit. Since the EU does not have the power to tax and issue bonds, the introduction of tax referenda and ex post fiscal referenda is not very interesting. The easiest way to introduce a fiscal referendum in the EU budgetary procedure therefore is the referendum approval of the whole budget ex ante. After the European Parliament has adopted the budget in

Enlargement and the European budget

151

its second reading, European voters will decide on it in a referendum. If the Parliament rejects the budget, voters can approve it without the Parliament’s adoption. A third reading in the EU budgetary process will only occur if voters reject the budget draft. Further qualifications of such a referendum are necessary. First, as in the case of the three kinds of fiscal referenda mentioned above, the EU budgetary referendum is assumed to be binding and it is obligatory. Second, only a simple majority of votes is considered. There is no requirement for a majority of votes of the member countries in the referendum. In Switzerland, for example, in order to succeed, changes of the federal constitution require that voters of a majority of cantons (states) adopt the changes. This institutional arrangement would make it possible to keep a Council-like weighting of votes. Third, no weighting of votes will occur consequently. As it would be the case in a United States of Europe, the principle ‘one person, one vote’ holds. Differences between the voters in one country are the usual case in politics and they may be expressed at the ballots. In contrast to the Council, it thus cannot necessarily be assumed that voters in a particular country prefer higher or lower spending. Fourth, no minimum of voting turnover or participation rate is required for the referendum to be binding. A simple majority of voters from the member countries of the EU who actually vote on the ballots is sufficient to reject and adopt the EU budget. If voters abstain from the ballots, such a behaviour is assumed to be a kind of preference revelation of its own that has to be accepted as such. A result of such a rule may be that voters from Southern European member countries that are assumed to have more intense preferences for higher EU spending may get a majority of votes at the ballots because of higher participation rates as compared to the Northern European countries. Fifth, we assume that, if the voters adopt a budget which induces revenue for the EU exceeding the official revenue ceilings, then the EU gets higher revenue from the fourth resource.11 On the other hand, if voters adopt a budget that requires lower revenue, this revenue has to be refunded to the member countries according to their share of revenue granted to the EU. All in all, this referendum design is a simple and basic reform proposal.

A basic model of referendum approval in EU budgetary procedures In the following we analyse the impact of a fiscal referendum in the EU budgetary procedure on policy outcomes and the position of the different players in that process. How is the power of the Council or the Parliament changed if a fiscal referendum is introduced? Is there any agenda-setting power of the Commission? If it exists, is it changed by the referendum? Our analysis is as follows: first, we analyse the current EU budgetary process in a simple unidimensional spatial model. Second, we consider the changes that occur in this model if a fiscal referendum is introduced. A basic model of EU budgetary procedures To show how different institutional arrangements may affect budgetary outcomes in the EU, assume that public expenditures are solely financed by taxes. A drafted

152

L.P. Feld, G. Kirchgässner and H. Weck-Hannemann

increase of EU expenditures thus induces higher financial contributions of the member countries, in our case only based on the GDP resource. The policy space is assumed to be unidimensional although budgetary decisions involve more than one dimension in order to keep the model as simple as possible.12 All political actors are assumed to have Euclidean preferences with respect to the level of expenditures E.13 We assume a quadratic utility function indicating that the utility of political actors is the more reduced, the farther away actual spending outcomes are from their most preferred spending level (ideal point). The ideal points of EU policymakers are shown in Figure 8.3. The Commission and the Parliament use simple majority rule, and there are no restrictions on amendments of the budget.14 Assume that the preferences of Commissioners and the Members of the European Parliament are Euclidian as well. Since the policy space is assumed to be unidimensional, the Commission and the Parliament can thus be treated as unitary actors with ideal policies equal to their median voters’ ideal policies, ÊC and ÊP, respectively. The Council can however not be treated as a unitary actor because it does not use simple majority rule but qualified majority with weighted votes. A qualified majority in the Council consists of 62 out of a total of 87 votes.15 The analysis of budgetary policymaking can be simplified by focusing on the countries that are pivotal under the qualified majority rule. The countries with the 26th vote from the left that are pivotal for a move to the right, i.e. higher EU spending, thus have an ideal policy to the left of the countries with the median vote. These pivotal countries and the countries to its right then have 62 votes and have the majority to decide higher spending levels. The countries to the right of these pivotal countries do not achieve a qualified majority without them. The countries with the 62nd vote are pivotal for a move to the left. For further simplification, EU member countries with presumably similar preferences are represented as unitary actors with an ideal point. We distinguish four groups of EU countries with similar preferences: the Southern European states Italy, Spain, Portugal and Greece together with Ireland with ideal policies ÊSE; the French-speaking countries France, Belgium and Luxembourg with ideal policies ÊF; the Northern European countries Germany, the Netherlands, Austria and Finland with ideal policies ÊNE; and the countries usually sceptical of centralisation of competencies at the EU level, the United Kingdom, Denmark and Sweden with ideal policies ÊOO. Although these coalitions are not totally stable in EU politics, we assume them to be stable with respect to spending levels. We also assume that Southern European EU countries would prefer the highest spending level, the French-speaking countries the second highest level,

Figure 8.3 EU budgetary policymaking without a fiscal referendum.

Enlargement and the European budget

153

the Northern European countries the third highest level, while the opting-out countries are assumed to prefer the lowest spending level. There are several reasons why we assume the distribution of spending preferences in that particular way, among them the positions of countries as contributors and beneficiaries from the EU budget, scepticism towards the EU and, not the least, cultural homogeneity. Southern European countries together with Ireland are assumed to have the highest spending preferences. Spain, Portugal, Greece and Ireland obtain the highest receipts from the EU budget. Italy is supposed to belong to that group because it receives a large amount from structural funds for the Mezzogiorno, money that would have to be paid directly by Northern Italy otherwise. Urging other European countries to share that burden with Italy makes these money flows to Southern Italy politically acceptable for Italian Northerners. The French-speaking countries have the second highest spending preference. France benefits considerably from EU agricultural policy. Belgium and Luxembourg directly gain from the fact that EU institutions are located there. The Northern European countries are net contributors to the EU budget, like Germany, the Netherlands or Austria. Finland has developed favourably in recent years such that it can be expected to have a reduced amount of receipts soon. Finally, the Euro-sceptics are grouped together. From each additional Euro spent by the EU, they can expect to finance a larger share, the UK by a reduction of its refund, Sweden and Denmark in general. Considering the number of votes of each group, the Northern European countries are pivotal for a move to the right and the French-speaking countries for a move to the left.16 Finally, the Commission is assumed to prefer a spending level between those of the Southern European and French-speaking countries because it obtains a higher power and prestige in administering higher levels of spending. Parliament prefers spending levels between the Northern European and the French-speaking EU countries since the population is considerably well represented in Parliament. Let q denote the status quo, i.e. last year’s budget. Preferences for compulsory and non-compulsory spending levels are assumed not to differ from each other for single political actors, in order to simplify the analysis. We only consider spending levels that are higher than last year’s budget to capture the EU expenditure growth effect. Recall now the sequence of decision-making with respect to budgetary decisions in the EU: 1 2 3 4

5

the Commission proposes a preliminary draft budget; the Council has the authority to amend it and decides on it in its first reading; the Parliament has the authority to amend it and decides on it in its first reading; the Council has the authority to amend it and to decide on it in its second reading (non-compulsory expenditure) or to determine the budget (compulsory expenditure); and the Parliament has the authority to amend non-compulsory expenditures, to determine the non-compulsory expenditures in the budget and to adopt or reject the whole budget draft in its second reading.

154

L.P. Feld, G. Kirchgässner and H. Weck-Hannemann

Consider first the easier case of non-compulsory expenditures in Figure 8.3. In this case, the Parliament can always overrule the budget amendments by the Council by a simple majority of its members. Since the Council knows the preferred expenditure level of the Parliament, it will only propose amendments that are both close to the preferences of the Parliament’s median legislator and will gain a qualified majority in the Council. Consequently, the Council will propose expenditure level Ê´NCE which is located in the middle between ÊNE and ÊF. Since the Parliament has the final say over non-compulsory expenditures, it will change these amendments to its ideal point Êp, which equals ÊNCE. With respect to non-compulsory expenditures the Commission thus has no agendasetting power. The outcome changes in the case of compulsory expenditures. Then, the Commission has agenda-setting power. Suppose it proposes Ê*CE, because it knows the preferences of the pivotal groups of countries in the Council. This expenditure level is not changed by the Council since the Northern European and the opting-out EU countries need the group of the French-speaking countries for a qualified majority for changes of EU expenditure levels. The French-speaking countries prefer however Ê*CE which is closer to their ideal point, ÊF. In the configuration of Figure 8.3, the equilibrium level of EU compulsory expenditure is thus Ê*CE. Although the two-thirds qualified majority in the Parliament may give the pivotal vote to the member of Parliament with higher spending preferences, the status quo will be sufficiently far away from the Parliament’s ideal point. Therefore, the Parliament will not reject the budget with Ê*CE, in order to get an outcome that is closer to its preferences. Comparing both equilibrium outcomes in this preference configuration, Ê*CE>Ê*NCE =Êp. The outcome changes, however, the closer the status quo is located to the preferences of the Parliament’s median legislator. Thus, the location of the status quo q is important because the Parliament can always reject the whole budget by a qualified majority.17 The introduction of referendum approval in EU budgetary procedures Consider now the introduction of a referendum in the EU budgetary process. Neither the ideal point of the median legislator in the European parliament nor the pivotal group of countries in the Council is necessarily equal to the median preference of all voters. One reason for such a deviation of preferences may be imperfect competition in the political market enabling political parties to follow their ideological positions. Another explanation is to assume that legislators are elected by districts. If a legislator represents the median voter of his district, then the preferences of the median of all voters is represented by the median legislator if and only if the median voter in the legislator’s district is equivalent to the median of all voters. This is a very restrictive assumption that is usually not satisfied in reality. The preferences of the median voter are represented by its ideal point, ÊV, in Figure 8.4. Recall again the sequence of decision-making with respect to budgetary decisions in the EU with a fiscal referendum:

Enlargement and the European budget

155

Figure 8.4 EU budgetary policymaking with a fiscal referendum.

1 2 3 4

5

6

the Commission proposes a preliminary draft budget; the Council has the authority to amend it and decides on it in its first reading; the Parliament has the authority to amend it and decides on it in its first reading; the Council has the authority to amend it and to decide on it in its second reading (non-compulsory expenditure) or to determine compulsory expenditure; the Parliament has the authority to amend non-compulsory expenditures and to determine the non-compulsory expenditures in the budget in its second reading; and voters have to adopt or reject the whole budget draft finally.

Whether the fiscal referendum becomes really binding in the EU budgetary process strongly depends on the location of the status quo, q. In the preference configuration depicted in Figure 8.4, with the median voter’s ideal point, ÊV, located between ÊNE and ÊP, the status quo is located farther away from the median voter’s ideal point than both the Parliament’s ideal point, ÊP, determining the equilibrium level of non-compulsory expenditures, Ê*NCE, and the equilibrium level of compulsory expenditures, Ê*CE, decided in the Council by qualified majority and determined by agenda-setting power of the Commission. Although voters would prefer lower EU spending that is pretty close to the ideal point of the Northern European countries, ÊNE, they would not reject the budget proposed because the status quo is worse for them. The policy outcome changes, however, the closer the status quo is located to the ideal point of the median voter. In Ê´V, the median voter is indifferent with respect to a comparison of the status quo and the expenditure level of non-compulsory expenditures proposed by the Parliament. In Ê˝V, the median voter is indifferent with respect to the comparison of the status quo and compulsory expenditures. Then, the probability that he rejects the whole EU budget is pretty high. It will definitely be rejected in this preference configuration of the other players, if the ideal point of the median voter is located at the left of q, i.e. the median voter prefers spending cuts. Moreover, the impact of the Commission’s agenda-setting power is weakened. The Commission’s proposal of the level of compulsory expenditures leads to higher EU spending and thus increases the probability that the median voter will reject the budget because of a spending level that is too high. Although the budgetary referendum is not immune against the impact of agenda-setting power, expenditure levels that are too far away from citizens’ preferences have no

156

L.P. Feld, G. Kirchgässner and H. Weck-Hannemann

chance to be adopted in a referendum. This is all the more true considering the current EU practice of specifying expenditure increases in advance. In this case, the status quo will be located relatively close to the preferences of the median voter. Deviations can thus only be relatively small.

Referendum approval of the budget and EU enlargement Eastern enlargement of the EU is supposed to lead to a need for major reform of EU institutions. In particular the weighting of votes in the Council, the number of Commissioners and the size of the Parliament are matters of concern. The introduction of fiscal referenda before enlargement has taken place may be regarded as a further possibility of postponing enlargement in the far future. We do not consider the question whether Eastern enlargement will take place and what direct reforms will need to be made. With respect to the introduction of referenda, we do not consider the impact of an EU-wide referendum on Eastern enlargement and its impact on the speed of enlargement. We assume that enlargement is a matter of fact and analyse how the fiscal referendum discussed above works with additional twelve EU member countries.18 The question addressed in this section thus is whether the power position of one of the political actors in the EU budget game and the political outcome is changed.19 Cyprus and Malta are supposed to vote in accordance with the group of Southern European countries (plus Ireland). The Eastern European countries are supposed to like even higher spending levels than the Southern European EU countries. Their ideal point, ÊEE, is located on the extreme right of the unidimensional policy space in Figure 8.5. This assumption is not too unrealistic given the fact that these countries will gain the most from the EU budget either concerning agricultural subsidies or payments from structural funds. Furthermore the change in the share of Commissioners and members of Parliament liking higher spending levels will move the position of the ideal points of the Commission and the European Parliament to the right as well. Finally, the ideal point of the median voter in an EU-wide referendum will move also to the right since Eastern European voters presumably like higher spending as well. The preference configuration of the political actors depicted in Figure 8.5 is similar to the ones in the previous section. The underlying budgetary process corresponds to the process outlined earlier. Not surprisingly, Eastern enlargement does not change the outcome of the budgetary referendum procedure

Figure 8.5 EU budgetary policymaking with a fiscal referendum after enlargement by twelve new member states.

Enlargement and the European budget

157

basically. By assumption, EU spending levels are higher in the case of enlargement than without it. But this is the result of the assumptions with respect to positions of ideal points after enlargement. Generally, the equilibrium outcome for non-compulsory expenditures, Ê*NCE, corresponds again to the ideal point of the Parliament, Ê P, if the status quo q is not sufficiently closer to the ideal point of the median voter, Ê V, than Ê *NCE. If this condition does not hold, the median voter in the EU will again reject the budget and prefers to stick to last year’s budget.20 The equilibrium outcome for compulsory expenditures, Ê *CE, is again determined by the Commission’s agenda-setting power because it proposes a level of compulsory expenditures between the ideal points of the Northern European countries, Ê NE, and the French-speaking countries, Ê F. Note that the new allocation of votes in the Council does not change the pivotal group of countries. In order to get a qualified majority in the Council 258 votes are necessary. The Eastern European, Southern European and French-speaking countries have 240 votes together. They thus need the agreement of the Northern European countries to pass compulsory expenditures. The fiscal referendum restricts only compulsory expenditure levels that are located farther away from the median voter’s ideal point, Ê V, than the status quo q. As in the case without enlargement the referendum requirement only prevents representatives from deviating too much from voters’ preferences.

Concluding remarks In this chapter we have investigated the impact of an introduction of a fiscal referendum in the EU budgetary procedure on policy outcomes and the power of the different political actors. The analysis of the current EU budgetary process yields the following results. First, the level of non-compulsory expenditures corresponds to the position of the European Parliament’s median legislator. This outcome follows from the power of the Parliament over and its unrestricted amendment capacity of non-compulsory expenditures. Second, the Commission has no agenda-setting power with respect to non-compulsory expenditures. Third, it has agenda-setting power with respect to compulsory expenditures because it can use the qualified majority requirement in the Council to get a policy outcome that is closer to its preferences than a vote in the Council without the influence of the Commission would yield. Thus, expenditures in the EU tend to be higher than a non-negligible part of member countries prefers. A creeping centralisation of spending to the EU level is compatible with this observation. The introduction of a referendum in the EU budgetary process as a reform proposal is based on the favourable economic performance of sub-federal jurisdictions in Switzerland and the US with direct legislation in their constitutions. The proposed referendum is designed as simple as possible and basically requires a simple majority of votes at the ballots in the end of the budgetary process for a budget draft to be adopted by voters. The changes that occur by such an institutional reform are not dramatic. The referendum will only prevent budget outcomes that deviate strongly or even extremely from the median voter’s ideal point. Since we assume a full information budgetary game, all political

158

L.P. Feld, G. Kirchgässner and H. Weck-Hannemann

actors know the ideal points of others. Thus, the Parliament can propose a noncompulsory expenditure level as close as possible to its ideal point that is located less far away from the ideal point of the median voter than the status quo. In a similar fashion, the Commission will consider the spending preferences of the pivotal groups of countries in the Council but also the ideal point of the median voter and proposes compulsory expenditures that exploit the position of the status quo as far as possible to come closest to the Commission’s ideal point. Nevertheless, these considerations indicate that citizens’ preferences cannot be neglected in budgetary decision-making if a referendum requirement exists. Budgetary outcomes are corrected towards voters’ preferences. Finally, we show that a simple increase in the number of member countries does not change the way budgets are made if they have to be adopted by voters. Thus, the considerations of the Parliament and the Commission with respect to referendum approval regarding non-compulsory and compulsory expenditures hold whether the EU has fifteen or twenty-seven member countries. The analysis of budgetary decision-making in the EU and the impact of an introduction of referenda in this chapter suffers from two shortcomings in particular. First, we only analyse changes in the level of EU expenditures without investigating the structure of the EU budget, i.e. the distinction between compulsory and non-compulsory spending. In the analysis of these structural aspects, the additional complication has to be considered that a relative increase in non-compulsory spending increases the relative power of the Parliament in budgetary decision-making. Second, we have not investigated to what extent the two-thirds majority, which is necessary in the Parliament to reject the whole budget, restricts the powers of the Council and the Commission in the budgetary procedure. Given the assumed spending preferences in this chapter, the agendasetting power of the Commission in the determination of compulsory spending would not be reduced because a qualified majority in the Parliament moves the ideal point of the Parliament closer to the ideal position of the Commission. In the event that the distribution of spending preferences differs from the one made above, the qualified majority may successfully restrict the Commission’s agenda-setting power. If a budgetary referendum is introduced with a simple majority requirement, agenda-setting power of the Commission may then be increased by a replacement of the qualified majority in the Parliament by the simple majority referendum. This problem can be solved by introducing a double majority referendum not only requiring a majority of all voters in the EU, but also a majority of the majority of voters of each country without weighting the votes. Such a double majority referendum is used in Switzerland for constitutional changes. However, both questions are beyond the scope of this chapter and must be left for future research. Nonetheless, the discussion in this contribution indicates that budgetary referenda are feasible without prohibitively high implementation costs. Furthermore, they lead to a correction of budgetary outcomes towards citizens’ preferences. Third, referenda do not lead to dramatic changes unless policy proposals do not strongly deviate from the median voter’s ideal point and, in addition, have no major repercussions on Eastern enlargement. Given the fact that legislation and

Enlargement and the European budget

159

budgetary decision-making in the EU will need more democratic legitimacy in the future in particular if Eastern enlargement reduces the representation of voters of EU member countries in the Commission, the Council and the Parliament, the (single or double majority) referendum proposal will be a way to tie EU decisions to the preferences of the European people. It would be a major step towards a ‘European public’.

Notes 1 Also, proposals exist that suggest a strengthening of national parliaments in the EU decision-making process in order to prevent the centralisation of competencies in the EU. See particularly Vaubel (1995). 2 For a detailed description of these studies see Pommerehne (1990), Kirchgässner et al. (1999, Chapters 4 and 5) as well as Feld and Kirchgässner (2000). 3 An exception is Noam (1980). 4 For theoretical arguments why tax evasion in direct democracies is lower than in representative democracies see Pommerehne and Weck-Hannemann (1996) and Pommerehne et al. (1997). 5 See Romer and Rosenthal (1978, 1979), Ingberman (1985), Steunenberg (1992), Moser (1996) and Feld (1997). 6 See Steunenberg (1994), Tsebelis (1994), Crombez (1996, 1998) and Moser (1997). 7 Figures 8.1 and 8.2 adapt the illustration of EU legislative procedures by Crombez (1996) in order to enable the reader to compare EU budgetary procedures with EU legislative procedures. 8 In the following, we only consider the case of such a proposal. The other case follows in an analogous fashion. 9 The Parliament has rejected the budget three times in 1979, in 1982 and in 1984. Usually, the Commission will propose some amendments of the budget to the Council and the Parliament for a third reading. 10 There are actually more possibilities. These three alternatives are usually found at Swiss sub-federal levels and do not exclude each other. They can also be combined. 11 We choose the GDP resource as the easiest solution since it is used to fill up EU revenue to the ceiling. 12 This assumption is simplifying because the legal distinction in compulsory and noncompulsory spending at the EU level seems to make a two-dimensional analysis natural. Extending the model to the two-dimensional case doesn’t change the results and only complicates the representation. 13 These assumptions are standard in spatial voting models, see, e.g., Romer and Rosenthal (1979), Crombez (1996), and in particular Steunenberg (1992: 504). The preference of a political actor i can be represented by the utility function Vi(.) defined over E: Vi = –(Êi – E)2. The most preferred point by a political actor is called his ideal point, Êi, which has the following property: Vi (Êi) > Vi (E) ∀ E ≠ Ê(i). All political actors are assumed to be fully informed on the preferences of others and the structure of the game, and to prefer their decisions not to be revised by others. Decisions are made sequentially without offering actors the chance of commiting to future courses of action. 14 With respect to decision-making in the parliament, this assumption corresponds to the assumption that all members of the European Parliament are present during budgetary decisions. Thus, we do not consider the three-fifths majority for budgetary decisions in the Parliament. 15 France, Germany, Italy and the UK have 10 votes each; Spain 8; Belgium, Greece, Portugal and the Netherlands 5 each; Austria and Sweden 4 each; Denmark, Finland and Ireland 3 each; and Luxembourg 2.

160

L.P. Feld, G. Kirchgässner and H. Weck-Hannemann

16 The Southern European countries together with Ireland have 31 votes; the Frenchspeaking countries have 17 votes; the Northern European countries have 22 votes; and the opting-out countries have also 17 votes. 17 We abstract from an analysis of the interactions between both EU expenditure categories. The higher compulsory expenditure, the lower non-compulsory expenditure must be, given a certain revenue ceiling for the EU. If, however, the status quo as the reversion point is sufficiently far away from Parliament’s desired level of noncompulsory expenditures, it will accept a reduction in this spending category for higher compulsory spending since it loses by a reversion to last year’s budget. 18 Whether a EU-wide fiscal referendum is technically feasible if another twelve countries join the EU does not concern us here. With new technological developments the geographic and demographic size of a polity becomes less important in voting. Increasing information processing capacities will render each referendum basically to a local one when votes at the ballots are collected in a community and the result is immediately processed to higher government levels. 19 We assume that the twelve additional countries that will join the EU are allocated the following number of votes in the Council (http://ue.eu.int/CIG/default.asp?lang=en, p. 80): Poland 27 votes; Romania 14 votes; the Czech Republic and Hungary 12 votes each; Bulgaria 10 votes; Slovakia and Lithuania 7 votes each; Slovenia, Latvia, Estonia and Cyprus 4 votes each; and Malta 3 votes. In total, EU member states will thus have 345 votes in the Council. We assume for simplicity that a qualified majority in the Council then will require 258 votes (74.8 per cent of the votes). The blocking minority is 88 votes. 20 The reversion to the status quo q causes unexpected problems if the budget to be decided in the referendum is the first budget after enlargement. Last year’s budget would then necessarily be insufficient to cover the expenditures occurring due to Eastern enlargement. We do not consider this special, but very interesting case here, but assume that voter interests are already considered in the basic EU enlargement decision.

References Commission of the European Union (2000) Haushaltsvademecum 2000, Luxembourg: European Union. Crombez, C. (1996) ‘Legislative procedures in the European Community’, British Journal of Political Science 26: 199–228. Crombez, C. (1998) The Treaty of Amsterdam and the Codecision Procedure, Leuven: Catholic University of Leuven, Department of Economics Working Paper No. 9827. Feld, L.P. (1997) ‘Exit, voice and income taxes: the loyalty of voters’, European Journal of Political Economy 13: 455–78. Feld, L.P. and G. Kirchgässner (1996) ‘Omne agens agendo perficitur: the economic meaning of subsidiarity’, in R. Holzmann (ed.) Maastricht: Monetary Constitution Without a Fiscal Constitution?, Baden-Baden: Nomos, 195–226. Feld, L.P. and G. Kirchgässner (1999) ‘Public debt and budgetary procedures: top down or bottom up? Some evidence from Swiss municipalities’, in J.M. Poterba and J. von Hagen (eds) Fiscal Institutions and Fiscal Performance, Chicago: Chicago University Press and NBER: 151–79. Feld, L.P. and G. Kirchgässner (2000) ‘Direct democracy, political culture, and the outcome of economic policy: a report on the swiss experience’, European Journal of Political Economy 16: 287–306. Feld, L.P. and G. Kirchgässner (2001a) ‘The political economy of direct legislation: direct democracy and local decision-making’, Economic Policy 33: 329–67.

Enlargement and the European budget

161

Feld, L.P. and G. Kirchgässner (2001b) ‘Does direct democracy reduce public debt?’, Evidence from Swiss municipalities, Public Choice 109: 347–70. Feld, L.P. and J.G. Matsusaka (2001) ‘Budget referendums and government spending: evidence from Swiss cantons’, University of St Gallen, unpublished manuscript. Feld, L.P. and M.R. Savioz (1997) ‘Direct democracy matters for economic performance: an empirical investigation’, Kyklos 50: 507–38. Frey, B.S. and A. Stutzer (2000) ‘Happiness, Economy and Institutions’, Economic Journal 110: 918–38. Gerber, E.R. (1999) The Populist Paradox: Interest Group Influence and the Promise of Direct Legislation, Princeton: Princeton University Press. Gerber, E.R. and S. Hug (1999) Minority Rights and Direct Legislation: Theory and Methods, La Jolla: University of California at San Diego, Department of Political Science, unpublished manuscript. Grimm, D. (1997) ‘Which constitution for Europe?’, in H. Siebert (ed.) Quo Vadis Europe?, Tübingen: Mohr Siebeck, 3–9. Hagen, J. von and I.J. Harden (1995) ‘Budget processes and commitment to fiscal discipline’, European Economic Review 39: 771–9. Hauser, H. and A. Müller (1995) ‘Legitimacy: the missing link for explaining EU institution building’, Aussenwirtschaft 50: 17–42. Ingberman, D.E. (1985) ‘Running against the status quo: institutions for direct democracy, referenda and allocations over time’, Public Choice 46: 19–43. Inman, R.P. (1997) ‘Do balanced budget rules work? U.S. experience and possible lessons for EMU’, in H. Siebert (ed.) Quo Vadis Europe?, Tübingen: Mohr Siebeck, 307–32. Kiewiet, D.R. and K. Szakaly (1996) ‘Constitutional limitations on borrowing: an analysis of state bonded indebtedness’, Journal of Law, Economics and Organization 12: 62–97. Kirchgässner, G., L.P. Feld and M.R. Savioz (1999) Die direkte Demokratie: Modern, erfolgreich, entwicklungs- und exportfähig, Basel: Helbing & Lichtenhahn/Vahlen/Beck. Kriesi, H. (1991) ‘Direkte Demokratie in der Schweiz’, Aus Politik und Zeitgeschichte 23: 44–54. Laffan, B. (1997) The Finances of the European Union, New York: St Martin’s Press. Matsusaka, J.G. (1992) ‘Economics of direct legislation’, Quarterly Journal of Economics 107: 541–71. Matsusaka, J.G. (1995) ‘Fiscal effects of the voter initiative: evidence from the last 30 years’, Journal of Political Economy 103: 587–623. Matsusaka, J.G. and N.M. McCarty (2001) ‘Political resource allocations: benefits and costs of voter initiatives’, Journal of Law, Economics, and Organization 17: 413–48. Moser, P. (1996) ‘Why is Swiss politics so stable?’, Schweizerische Zeitschrift für Volkswirtschaft und Statistik 132: 31–60. Moser, P. (1997) ‘A theory of the conditional influence of the European Parliament in the cooperative procedure’, Public Choice 9: 333–50. Moser, P. (2000) The Political Economy of Democratic Institutions, Cheltenham: Edward Elgar. Noam, E.M. (1980) ‘The efficiency of direct democracy’, Journal of Political Economy 88: 803–10. Pommerehne, W.W. (1978) ‘Institutional approaches to public expenditure: empirical evidence from Swiss municipalities’, Journal of Public Economics 9: 255–80. Pommerehne, W.W. (1983) ‘Private versus öffentliche Müllabfuhr: nochmals betrachtet’, Finanzarchiv 41: 466–75. Pommerehne, W.W. (1990) ‘The empirical relevance of comparative institutional analysis’, European Economic Review 34: 458–69. Pommerehne, W.W. and H. Weck-Hannemann (1996) ‘Tax rates, tax administration and income tax evasion in Switzerland’, Public Choice 88: 161–70.

162

L.P. Feld, G. Kirchgässner and H. Weck-Hannemann

Pommerehne, W.W., A. Hart and L.P. Feld (1997) ‘Steuerhinterziehung und ihre Kontrolle in unterschiedlichen politischen Systemen’, Homo Oeconomicus 14: 469–87. Poterba, James M. (1997) ‘Do budget rules work?’, in A.J. Auerbach (ed.) Fiscal Policy: Lessons from Economic Research, Cambridge: MIT Press, 53–86. Romer, Th. and H. Rosenthal (1978) ‘Political resource allocation, controlled agendas, and the status quo’, Public Choice 33: 27–43. Romer, Th. and H. Rosenthal (1979) ‘Bureaucrats versus voters: on the political economy of resource allocation by direct democracies’, Quarterly Journal of Economics 93: 563–87. Schneider, F. (1993) ‘The federal and fiscal structures of representative and direct democracies as models for a European Federal Union: some ideas using the public choice approach’, European Economy 5: 191–212. Shepsle, K.A. (1979) ‘Institutional arrangements and equilibrium in multidimensional voting models’, American Journal of Political Science 23: 27–59. Steunenberg, B. (1992) ‘Referendum, initiative, and veto power: budgetary decision making in local government’, Kyklos 45: 501–29. Steunenberg, B. (1994) ‘Decision making under different institutional arrangements: legislation by the European Community’, Journal of Institutional and Theoretical Economics 150: 642–69. Tsebelis, G. (1994) ‘The power of the European Parliament as a conditional agenda setter’, American Political Science Review 88: 128–42. Vaubel, R. (1995) The Centralization of Western Europe, London: Institute of Economic Affairs. Weingast, B.R. (1989) ‘The political institutions of representative government: legislatures’, Journal of Institutional and Theoretical Economics 145: 693–703.

Chapter Title

9

163

Enlargement and the European Central Bank Jordi Bacaria, Georgios E. Chortareas and Andreas P. Kyriacou1

Introduction The European Monetary Union (EMU) has been delegated to the European System of Central Banks (ESCBs), made up of the European Central Bank (ECB) and the National Central Banks (NCBs) of the participating countries. The primary objective of the ECB is ‘price stability’ and towards this aim the ECB and participating National Central Banks are prohibited from granting credit facilities to, or purchasing debt from, Community institutions or bodies, central governments, regional, local or other public authorities, other bodies governed by public law or public undertakings of member states. Both the ECB and participating National Central Banks are moreover prohibited from taking instructions from these institutions. The requirement that the participating National Central Banks are legally independent facilitates the setting of a common monetary policy within EMU. In this chapter we consider whether the presence of an ever-increasing number of independent National Central Banks in the system reduces the degree of policy cohesion in monetary policy decision-making, and the implications for policy credibility and the central bank’s ability to achieve its policy objectives. Having advanced this argument, we recognise that the participation of National Central Banks introduces a measure of regional representation when setting monetary policy in EMU. We examine whether some form of regional representation when setting a common monetary policy is a desirable feature. We evaluate suggestions that seek to preserve the benefits of regional representation while avoiding the corresponding costs that could possibly result from the inclusion of an ever-greater number of participants in the monetary policy-setting process. Such institutional arrangements include the denationalisation of regional representation through the creation of supra-national monetary districts and the institution of rotating membership in the monetary policy-setting body. The chapter is structured as follows. First, we discuss remedies to the time inconsistency problem which plagues the setting of monetary policy and points out that in the case of the EMU one of them – making the ECB legally independent – is comparatively more conducive to the setting of a common monetary policy. Subsequently, we examine the contention that the participation of an evergreater number of National Central Banks in the Governing Council of the ECB

164

J. Bacaria, G.E. Chortareas and A.P. Kyriacou

may lead to a decline in policy cohesion when this body sets the Union-wide monetary policy, and we focus on the issue of regional representation in the system. Finally, we closely examine alternative institutional arrangements of regional representation, drawing on the relevant experience of the German and US monetary unions. Appendix 9 presents a more formal theoretical framework for the analysis of the costs that individual countries may suffer from the setting of a central common monetary policy.

The case for independent central banks: the preference versus the incentives approach The theoretical case for central bank independence has stressed the time inconsistency problem (Kydland and Prescott 1977; Barro and Gordon 1983a, 1983b). In a nutshell, assuming that economic agents respond optimally to the policies of the authorities, the ex post optimal move of the government is to renege on a past promise to keep its inflation rate at a given level and try instead to engineer a surprise inflation in the short run hoping to achieve a reduction in real wages and temporary output gains in the economy. Implicit in this contention is the assumption that electoral politics is inefficient since, on the one hand, it may force politicians to attach a strong weight to the short-term gains of inflationary policies (De Grauwe 1994), and on the other, voters are myopic or suffer from some informational disadvantage (Rogoff and Sibert 1988). Many studies empirically support the idea that in developed countries,2 more independent central banks are associated with lower levels of inflation (see Eijffinger and De Haan 1996). Not only is central bank independence negatively related to inflation, it is also costless in terms of growth and inflation and output variances (Eijffinger and Schaling 1993). Similarly, Alesina and Summers (1993) find that while central bank independence promotes price stability, it has no measurable impact on real economic performance such as growth, unemployment and real interest rates. Alesina and Gatti (1995) show that this may be due to the elimination of output variability, which comes from alternating governments with different preferences over inflation and unemployment. This assumes that this type of output variability is larger than that which comes from exogenous shocks (which would tend to be ignored by conservative and independent central bankers). On the other hand, if exogenous shocks are expected to be important, then the possibility emerges that a legally independent and conservative central bank may be willing to pay a high price in terms of output stabilisation possibilities foregone, in order to defend price stability. This problem, incidentally, is less likely to occur under the principal-agent approach. In this vein, Lohmann (1992) has shown that it may be optimal to allow the central bank to be overruled, albeit at a large but finite cost.3 Moreover, some authors reverse the implied causality of the relationship between legally independent central banks and price stability by suggesting that countries which are inflation-averse develop the institutions to support their aversion. For Posen (1993) this inflation aversion may be due to a coalition of

Enlargement and the European Central Bank

165

interests in society and, in particular, a politically powerful financial sector. Goodman (1991) has similarly argued that although such coalitions are a necessary condition for the creation of an independent central bank, they are not sufficient. He points to an additional requirement, namely, the expectation by politicians that their party will be in office for only a short period of time. De Haan and Van ‘t Hag (1995) attribute the aversion to inflation to the general public. In this vein, central bank independence is endogenous to public’s preference. It has often been argued that the Bundesbank’s preferences are representative of those of the German people and so the Bank can follow antiinflationary policies because the public wants it to, given a bitter historical experience with hyperinflation (see, for instance, Bofinger 1992).4 Persson and Tabellini (1993) and Walsh (1995) derived an optimal contract for central bankers. Based on this contract, the bankers are faced with a linear tax or subsidy for obtaining an inflation rate respectively above or below a targeted level. As previously stated, New Zealand represents a real-world example of the incentives approach. Brash (1996) explains how section 9 of the Reserve Bank of New Zealand Act 1989 requires a contract to be signed between the Minister of Finance and Governor such that the central bank will maintain the rate of inflation within the margin of 0–2 per cent without interference from Government, Treasury, or anybody else. According to Fratianni et al. (1997), a major problem with the principal-agent approach is that the contract must be negotiated with elected leaders who are responsible for creating the problem in the first place. This, and the need for flexibility in the face of a large supply shock, points to the possibility that the government may try to reset the contract in a non-optimal way. Partly as a result of this possibility, the above authors come out in favour of legal independence as a better guarantee that central bankers will pursue policies in society’s best interests. In the case of EMU, a Rogoff-type approach has been taken by establishing a legally independent European System of Central Banks, or in other words, by making both the ECB and the National Central Banks legally independent of governments. Bacaria and Kyriacou (1996) have elsewhere argued that this approach makes possible the setting of a common monetary policy within EMU. While the principal-agent approach may have similarly been able to generate the desired anti-inflationary credibility across National Central Banks, it would have made it exceedingly difficult to coordinate monetary policies within EMU. In particular, since presumably each National Central Bank would have a contract with its government to obtain an inflation rate within a given margin, this contractual relation would hamper efforts towards coordination with the other central banks in the system. Compared to the principal-agent approach, then, the Rogoff approach would facilitate the setting of a common monetary policy within EMU.

Enlargement of EMU and the problem of reduced policy cohesion The more cohesive an institution is regarding policy matters, the stronger it will be in the face of external pressures, and the greater the commitment towards

166

J. Bacaria, G.E. Chortareas and A.P. Kyriacou

policy implementation (see, for instance, Krause 1996). In the specific area of monetary policy decision-making, the search for consensus basically stems from the need to maintain the power and credibility of the monetary policy setting body (Borins 1972). Under the actual institutional structure of the ECB, the successive enlargement of EMU is likely to increase the likelihood of dissent voting emerging or, in other words, to reduce the degree of policy cohesion when monetary policy decisions are taken in the Governing Council. To see why, and to facilitate our discussion in other parts of this chapter, it is worth considering the institutional structure of the ECB in greater detail. The ECB has a Governing Council (which is to set monetary policy) and an Executive Board (which is to implement it). The Executive Board is made up of the President and Vice President of the ECB and four other members. The members of the Executive Board are appointed for eight-year, non-renewable terms.5 Decisions in both the Council and the Board are to be taken by majority, with the President being able to cast a tie-breaking vote. The Governing Council is made up of the Executive Board and the governors of the National Central Banks of those member states that participate in the EMU. The members of the National Central Banks are variously appointed either directly or indirectly by national governments and will be so for at least five years. Their reappointment is not ruled out by the European System of Central Banks’ statute (Kenen 1995). For our purposes here it is important to note that monetary policy is to be set by a Governing Council, which is made up of the six members of the Executive Board and the National Central Banks governors of participating member states. In the EMU, that is made up of twelve countries, this means that monetary policy will be set within a Governing Council of eighteen members. In the event that Denmark, Sweden and the UK eventually join the EMU, the Council would increase to twenty-one members. And if the EU is eventually enlarged to admit up to eleven new member states (ten Central and Eastern European countries plus Cyprus), then monetary policy would potentially have to be set by a Council of thirty-two members. This situation, of course, is not unique to the Governing Council of the ECB. It also affects other EU institutions, such as the European Commission and the Council of Ministers. In an EU of twenty-six members the number of Commissioners could have increased from the current twenty to thirty-two (two for Poland and one for the remaining ten candidates) and the Council of Ministers would increase to twenty-six members. A priori, the fact that decisions in the Governing Council of the ECB are to be taken by simple majority, while those in the Council of Ministers are typically taken under qualified majority and those in the Commission by consensus (given its collegial nature), means that the ability to take decisions would be relatively easier in the Governing Council of the ECB; relative, that is, to the ability to do so in the Council of Ministers and, even more so, in the Commission. This would be the case in an EU of fifteen members and increasingly so in one of twenty-six. This is probably behind the provision in the Amsterdam Treaty that before the next enlargement the number of Commissioners is reduced to one per country.

Enlargement and the European Central Bank

167

Table 9.1 Regional representation in three monetary unions

Central Bank • • • • • • •

ECB (EU-15)a ECB (EU-15) b ECB (EU-15)c ECB (EU-21)d ECB (EU-26)e Bundesbank Federal Reserve

Notes:

Size of monetary policy setting body

Regions represented in monetary policy setting body

Proportion of regional representation in monetary policy setting body

17 18 21 27 32 17 12

11 12 15 21 26 9 5

0.65 0.66 0.71 0.78 0.81 0.53 0.42

a

Those member states which will take part in EMU from the beginning. a + Greece (in EMU since January 2001). b + Denmark, Sweden and the UK. d c + Cyprus, Czech Republic, Estonia, Hungary, Poland, Slovenia. e d + Bulgaria, Latvia, Lithuania, Romania, Slovak Republic. b c

Consistent with this, Alesina and Grilli (1992: 262) have argued that given that decisions in the Council are to be taken under majority rule, ‘the policy that will be implemented is one that is (ex post) preferred by the median Council member’. Implicit in this result is the assumption that each member of the Governing Council may disagree, for example, about the level of short-term interest rates that should be set by the ECB, but they all agree that a movement away from one’s ideal level makes him/her worse off.6 Under this assumption, the level of short-term interest rates that would be chosen by the majority would correspond to the ideal point of the median (Black 1958). This point is an equilibrium, in the sense that a majority prefers the median’s ideal to any other possible alternative. On the face of it, the likelihood that, under majority rule, the monetary policy set by the Governing Council will be that preferred by its median member, points to the potential ease of decision-making within the Council. The median voter decides and that is the end of the matter. To the extent that the monetary policy actually adopted by the majority of members generates important and/or persistent output costs for one or more Euro-member states, however, it becomes increasingly likely that the national central bankers appointed by the member states will increasingly voice their disagreement with the majority. This in itself would not be of concern except for the possibility, previously raised, that the emergence of such dissent voting may strike at the credibility of the ECB and thereby undermine its ability to pursue the objective of price stability. The decision not to publish the minutes of policy meetings or the details of how members vote until, possibly, sixteen years later,7 must surely be motivated by the expectation of such policy dissent emerging and the fear that if it were to be made public, it would strike at the credibility of the system and reduce the ability of the ECB to attain its primary objective. One relevant question that emerges here is under what circumstances a Euromember state or group of states is likely to suffer large and/or persistent output

168

J. Bacaria, G.E. Chortareas and A.P. Kyriacou

costs from that monetary policy which is decided by a majority of member states. While here we limit ourselves to a description of the most important costs, in the Appendix at the end of this chapter we develop a theoretical framework to examine these and other costs. Such costs are more likely the smaller the correlation between domestic and European output due for example to non-synchronised business cycles or the strength and incidence of asymmetric external shocks (Alesina and Grilli 1992). Thus, a country or group of countries whose business cycle is out of phase with that of the majority of Euro-member states or which is relatively prone to suffer from asymmetric external shocks is likely to bear substantial costs from the setting of a Union-wide monetary policy by the Governing Council, at least in the early years of the monetary union.8 Second, large costs are more likely the greater the domestic output variance of a country relative to the variability of European output (Alesina and Grilli 1992). Thus, even if the members of a monetary union have synchronised business cycles and are hit by symmetric external shocks, those countries with a greater output sensitivity to such shocks are likely to incur greater costs and would thus tend to prefer a more flexible monetary policy than the majority. Structural differences across countries may be one factor underlying international differentials in domestic output variance. For example, countries with more flexible labour markets would tend to have a smaller output variability (Waller 1992). A third reason why countries may incur costs from the conduct of a common European-wide monetary policy is international differences in the way monetary policy is transmitted to the economy. In particular, recent evidence suggests that, because of structural differences in financial markets and varying degrees of openness of economies, the full effect of a contractionary monetary shock on activity may vary across countries (see Ramaswamy and Slok 1998; Dornbusch et al. 1998). While the evidence is not unanimous on the actual numbers, the overriding presumption is that such differences do exist. For our purposes here we would suggest that such differences may contribute to the emergence of costs for some countries from a European-wide monetary policy – again, at least during the initial years of the monetary union.9 Unsurprisingly then, given the large volume of literature on optimum currency areas since Mundell (1961), the likelihood of large and/or persistent costs emerging for participating countries when setting a common monetary policy within the Governing Council of the ECB is positively related to the differences in the participating member states’ economies (structural and/or cyclical). To these we should add historical and cultural factors that may lead to international differences in the tolerance to any given output loss in the pursuit of price stability. To the extent that such differences increase as EMU is enlarged beyond the current twelve members, this would tend to increase the likelihood of policy dissent in the conduct of monetary policy by the ECB. This is especially so given the fact that even an EMU with only the current twelve participants represents a relatively decentralised monetary policy setting; relative, that is, to the case of the Bundesbank and even more so, to the Federal Reserve (see Table 9.1). Under the current institutional set-up, each successive

Enlargement and the European Central Bank

169

enlargement would increase the decentralization of the ECB’s Governing Council, thereby making it increasingly difficult for the centrally appointed Executive Board to exercise a stabilising or cohesive influence.

The need for regional representation in the ECB Thus, the enlargement of the EMU to include new member states could increase the likelihood of dissentient voting emerging when these decisions are taken and – the secrecy regarding the minutes of the monetary policy decision-making process notwithstanding – it could increase the possibility that the credibility of the ECB may be undermined. These problems were seen to be a direct result of the fact that under the current institutional set-up of the ECB, enlargement will necessarily increase the number of National Central Banks that form an integral part of the Governing Council. It is important to point out, however, that the direct participation of National Central Banks itself introduces a measure of regional representation into the system. The emerging question is under what conditions is some degree of regional representation in monetary policy setting desirable. In as much as the viability of the system is concerned, regional representation has been seen to address the fact that the consequences of central bank decisions are often likely to vary across regions (Bowles and White 1994). In this vein, Eichengreen (1990) points to the presence in the US of representatives of Federal Reserve districts in an attempt to reconcile regional monetary preferences and in particular to allay the fears of a priori pro-inflation Westerners and anti-inflation Easterners alike. Similarly, Leonard Gleske, former directorate member of the Bundesbank states: You need people on the Council who reflect different developments in different parts of Germany. They come from different backgrounds, and are chosen by different procedures corresponding to different political currents. Such a Council makes it possible for the central bank to be independent. Otherwise I would have my doubts about giving such a body so much independence. (Quoted in Marsh 1993) An additional point that can be made here stems from Hirschman’s (1970) idea of the substitutability between voice and exit options. In the context of our discussion, the absence of a voice option on the part of economically heterogeneous regions when setting a common monetary policy may make the exercise of an exit option increasingly likely and thus jeopardise the viability of the monetary union.10 This points to the need to institute some kind of regional representation in the Governing Council so that the interests of the different economic areas are represented. In the case of EMU this need is particularly urgent for at least two reasons. It is so, first, in view of the fact that the ECB’s formal democratic accountability to elected politicians, and thus its democratic legitimacy, is more limited than that of other central banks (see, for instance, De Haan 1997). This

170

J. Bacaria, G.E. Chortareas and A.P. Kyriacou

is especially pertinent given the intention to throw a shroud of secrecy over the monetary policy-setting deliberations of the ECB. Second, the need for regional representation increases given the absence of effective adjustment mechanisms such as labour mobility, wage flexibility and centralised or, in view of the stability pact, even decentralised transfer schemes. Conversely, the existence of alternative adjustment mechanisms in both Germany and the US means that, compared to the future EMU, the urgency of some form of regional representation within each country’s central banking system is reduced.11 To the extent that regional representation can increase the democratic legitimacy of the ECB, it can also, of course, make it easier for it to defend its primary objective of price stability, against possible political pressures. Regional representation may also make it easier for a central bank to pursue its objectives in another, less obvious way. Goodman (1991) argues that the presence of a majority of Land Central Bank presidents in the Bundesbank’s Central Bank Council, where German monetary policy is set, maintains the central bank’s independence vis-à-vis the federal government.12 Along these lines, the Federal Reserve constitutes the Districts as a permanent minority in the Federal Open Market Committee and to this effect may reduce the independence of the Federal Reserve from the federal government. This is in line with studies that argue that the Fed is a creature of Congress (see, for instance, Beck 1990). From the above discussion it emerges that the proportion of regional representation in the monetary policy-setting body of the ECB will be greater in the EMU as compared to the German or the US monetary unions. Moreover, it will successively increase as more members are added to the EU and ultimately the EMU itself. Given that the Governing Council of the ECB constitutes the National Central Banks as a permanent and ever-increasing majority, it arguably promotes the independence of the ECB from any centralised political authority such as the euro-zone finance ministers group.13 The possibility that the degree of decentralization of the ECB may be positively related to its independence from any centralised political authority points to the desirability of constituting regional representatives as a permanent majority in the Governing Council. Recall, however, that decentralisation has its costs, namely, a lower degree of policy cohesion when setting monetary policy. This points to the need for institutional solutions that can put a limit on the degree of decentralisation of the system in the face of enlargement.

Possible institutional solutions: insights from Germany and the US Thus, the successive enlargement of the EMU may, under the given institutional framework, reduce the degree of policy cohesion when setting the common monetary policy. On the other hand, some degree of regional representation could improve the workings of the monetary union by enhancing its cohesion. In searching for possible institutional structures that can address this apparent trade-off, we turn to the experience of two existing monetary unions namely, those of Germany and the US.

Enlargement and the European Central Bank

171

The institutional structure within which monetary policy is conducted in these countries is not only of interest because of the longevity of these monetary unions but also because both have survived despite being composed, like the EMU, of numerous heterogeneous regions. Therefore, a closer examination of the German and US examples should be able to bring to light some of the key elements of a resolution of the above trade-off. Consider first how monetary policy has been conducted in Germany. German monetary policy has been the responsibility of the Bundesbank, which is made up of a Central Bank Council, a Directorate and the Boards of nine Land (regional) Central Banks. Monetary policy is set by the Central Bank Council, which is made up of the Directorate (the President and Deputy President of the Bundesbank and a maximum of six more members) and the nine presidents of the Land (regional) Central Banks (European Monetary Institute 1996).14 Decisions in the Central Bank Council are taken by simple majority vote and implemented by the Directorate. The Directorate is also responsible for open market operations, foreign exchange transactions and transactions with nonresidents and the Federal Government. The members of the Directorate are appointed by the President of the Federal Republic on proposal by the Federal Government, after consultation with the Central Bank Council. The Presidents of the Land Central Banks are appointed by the President of the Federal Republic on proposal by the Bundesrat (the Upper Chamber of Parliament), following the submission of a proposal from the authority designated under the laws of the Land or Länder concerned and after consultation with the Central Bank Council. All the members of the Bundesbank are appointed for eight-year terms that are renewable. In the US, monetary policy is in the hands of the US Federal Reserve System, which is a federal system composed of the Board of Governors and twelve regional Federal Reserve Banks (FED 1994). A major organ of the Federal Reserve System is the Federal Open Market Committee, which supervises open market operations; it is the main tool used by the Federal Reserve to influence money market conditions and the growth of money and credit. It is made up of the Board of Governors and the presidents of five Federal Reserve Banks, four of whom serve on a rotating basis. More specifically, the president of the New York Federal Bank has a permanent seat while the Presidents of the Cleveland and Chicago Federal Reserve Banks become members every other year and the Presidents of the remaining nine Federal Reserve Banks become a member every third year. Decisions in the Federal Open Market Committee are taken by simple majority. The Board of Governors is composed of seven members, including the Chairman. The Board is appointed for a period of fourteen years by the President of the US and confirmed by the Senate. The Chairman and Vice-Chairman are appointed for four years. The regional Federal Reserve Banks are, in a legal sense, private corporations with presidents, and boards of directors and even shareholders. Each has its own board of nine directors of whom six – designated class A and B (three each) – are elected by those commercial banks in each District which are members of the Federal Reserve System. The three remaining

172

J. Bacaria, G.E. Chortareas and A.P. Kyriacou

– designated class C – are appointed by the Board of Governors in Washington DC and from these the Board selects one person as Chairman and another as Deputy Chairman. The important thing to note here is how the number of members in the monetary policy-setting bodies is reduced in either system. Common to both systems is the creation of monetary districts, which do not correspond to the regional or state borders. This is the case in Germany where there are nine Land Central Banks and sixteen Länder, and it is even more so the case in the US where the fifty states of the Union far outnumber the twelve Federal Reserve Banks which make up the Federal Reserve System. An additional and relevant feature, specific to the US system, is the institution of rotating membership in the Federal Open Market Committee where monetary policy is decided. Specifically, the twelve Federal Reserve Banks which make up the system alternate in such a way that only five of them can make up the Federal Open Market Committee at any one time, with the additional proviso that the New York Federal Reserve has a permanent seat. The adoption of monetary districts or that of alternating membership has been proposed as one way to cope with the ever-larger committee size that would result from the enlargement of EMU. In particular, the Centre for Economic Policy Research (1992) has proposed either the creation of supra-national monetary districts or the institution of rotating membership. By way of a theoretical exercise meant to illustrate the application of these proposals, the CEPR suggests that the districts could be made up of one or more countries based on relative population size and geographical proximity (small and neighbouring countries being grouped together or with a large neighbour) while rotating membership could mean permanent seats for the largest countries, with the others taking seats every two, three, or four years depending on their size (the smaller they are, the less frequent their participation).15 The setting up of monetary districts of similar population size has the advantage of enhancing the democratic legitimacy of the ECB. Under the current setting each National Central Bank has one vote, despite huge population differences across countries (ranging from as little as 0.1 per cent of the euroarea for Luxembourg to as much as 28.2 per cent for Germany). In addition, our analysis points to the desirability that each such monetary district be, as far as possible, defined over homogeneous economic regions. This would give different economic regions a separate voice in the monetary policy-setting process and reduce the relative attractiveness of exit. If the creation of economically homogeneous supra-national monetary districts is not possible by grouping together existing member states, then monetary districts could be drawn across them, following the US example. In either case, such institutional arrangements correspond to the denationalisation of regional representation within the ECB, to paraphrase Hayek (1990). In order to increase the degree of policy consensus in the system, there may be also a case for consulting with the Governing Council before appointing those individuals who are proposed by each monetary district to act as their representative (e.g., Bundesbank).

Enlargement and the European Central Bank

173

The desire to give different economic regions a separate voice in the monetary policy-setting process would tend to favour the creation of supra-national monetary districts appointed to the rotating membership of member states, given that the latter would, at times, leave some states without a voice. Having said this, such a voice can be instituted under a system of rotating membership by requiring that participating countries take into consideration the interests of those countries, which – at times – may not form part of the decision-making process. One way to do this would be to require that all the different economies which may form part of the monetary union are at all times represented by those member states that happen to form part of the Governing Council at any point in time.

Conclusions We began this chapter by reviewing remedies to monetary policy problems that arise from the strategic and/or political nature of the policymaking process and how they pertain to the design of monetary institutions in the euro-area. We considered the possibility that nominally independent national central bankers may disagree with that monetary policy decided under a simple majority rule. The greater the differences in the participating member states’ economies (both cyclical and structural) as well as international differences in the tolerance to any given output loss in the pursuit of price stability (due to historical and cultural factors), the more likely this possibility is. The greater these differences the greater is the likelihood of a dissenting voting emerging, which, if it were to be somehow made public, could threaten the credibility of the ECB. The possibility of policy dissent can be expected to become even more acute as each successive enlargement of the EMU includes an increasingly diverse range of economies and inevitably leads to an increasingly decentralised ECB. Having said this, however, we have emphasised the importance of some sort of regional representation in EMU as a way of giving economically diverse regions a voice when setting monetary policies and as a check against centralised political power. The former may be particularly necessary given that the ECB has been viewed as suffering from a democratic deficit and, compared to other monetary unions, the EMU is short on alternative adjustment mechanisms. In response to this trade-off between the benefits of regional representation and the potential costs from the inclusion of more and more National Central Banks in the monetary policy-setting process, we have considered two alternative institutional arrangements. First, we have examined the denationalisation of regional representation through the establishment of supra-national monetary districts. Such an arrangement may enhance the democratic legitimacy of the ECB by assigning votes equally across regions on the basis of population and by giving different economic regions a separate voice when setting the common monetary policy. Second, we examined the institution of rotating membership of participant countries in the monetary policy-setting body. Rotating membership implies that, at times, member states may be left without a voice, which is not so in the case of supra-national monetary districts.

174

J. Bacaria, G.E. Chortareas and A.P. Kyriacou

One way of avoiding this problem is to have the rotation of membership based on the requirement that all the different economies that may form part of the monetary union, be at all times represented by those member states that happen to form part of the Governing Council. Regardless of its form, regional representation designed to constitute a permanent majority in the Governing Council (e.g. as in Bundesbank) would enhance the independence of the ECB from any centralized political authority.

APPENDIX 9 In this appendix we propose one way to examine the costs, which may flow for individual countries from a centrally set common monetary policy. Consider a typical Phillips curve that relates inflation (π) and unemployment (u) in the European Monetary Union (EMU) and the ith member country as follows:

uE = uEn − λ E (π E − π Ee ) + ν E ,

(9.1)

ui = uin − λ i (π i − π ie ) + ν i ,

(9.2)

and

where vE and vi are a union-specific and a country-specific shock and the subscripts denote EMU and country i variables respectively. The European Central Bank’s (ECB) and the ith country’s policymakers’ preferences about the unemployment and inflation rates in their constituencies are given by the following loss functions:

E( LECB E )=

() ( 1 2

 E  uE − u*E 

)

2

(

)

2 + bE π E − π *E , 

(9.3)

and

E( Lii ) =

( ) E(u − u ) + b (π − π ) , 1 2

i

* i

2

i

i

* i

2

(9.4)

where b reflects the degree of policymakers’ conservatism. We assume that π*E = π*i n n n n = 0, u*E = uE (1/θE), θE ≥ 1, u*i = u i (1/θ i), θ i > 1, θ i > θE and uE , u i is the natural rate of unemployment in the EMU and country i respectively. In other words, although both the ECB and the national policymaker dislike inflation, the last is more willing to trade off inflation for lower unemployment rates. If the policymakers of country i could set monetary policy autonomously they would choose the inflation rate that minimises the following expected loss function:

Enlargement and the European Central Bank

min E( Lii ) = πi

( ) E(u − u (1 θ )) + b (π ) . 1 2

2

n i

i

2

i

i

i

175 (9.5)

The optimising behaviour of the national policymakers, subject to the constraints implied by the structure of the economy, delivers the following inflation and unemployment rates:

(

)(

)

π i = λ i bi qi uin ν i ,

(9.6)

ui = uin + ν i ,

(9.7)

and

where qi = (θ i – 1/θ i) and for θ i >1, 1 > qi > 0. Note that (∂qi/∂θ i)>0, which implies that the lower the unemployment rate desired by national policymakers is, the higher the resulting inflation rate is (see Equation 9.6). When country i participates in the EMU and monetary policy is set by the ECB, then its expected loss is

(

 E( LECB ) =  uEn qE i 

)

2

((

 + σ ν2,E  bE + λ2E 

) 2b ), E

(9.8)

which results from minimisation of Equation 9.3. The common monetary policy implies a uniform unemployment rate for the EMU. The expected loss of the national policymaker, however, when monetary policy is set autonomously, is

( )

 E( Lii ) =  uin qi 

2

((

 + σ ν2,i  bi + λ2i 

) 2b ). i

(9.9)

The expected costs for the nationally motivated policymakers in country i from the common monetary policy (i.e., from losing monetary policy autonomy) are given by E(Ci), that is, the difference between the expected losses for country i when monetary policy is set by the ECB and the expected losses for country i when monetary policy is set autonomously,

E(C i ) = E( LECB ) − E( Lii ). i

(9.10)

We now examine how country i’s costs of having a single monetary policy decided by the ECB depend on the differences regarding the tolerable (desired/ targeted) unemployment rate between the national policymakers and the ECB, and the differences regarding their degrees of inflation aversion (b). The same procedure can be applied to examine the cost for country i from a common monetary policy due to differences in external shocks between country i and the EMU or differences in the way that monetary policy affects the rate of unemployment in country i compared to EMU.

176

J. Bacaria, G.E. Chortareas and A.P. Kyriacou

Different b’s Assume that the economic structure of country i is representative of that of EMU. That is, this country is subject to the same shocks as the EMU (σ 2v,i = σ 2v,E = σ 2v), monetary policy affects the rate of unemployment in a similar way in both country i and the EMU (λE = λ i = λ), the natural rate of unemployment in n n n country i is the same as for the EMU (u i = u E = u ), and the preferences of national policymakers about the unemployment rate are the same as the ECB (θE = θ i = θ). The only difference is that the national policymakers prefer more flexibility in monetary policy. In other words, they consider the ECB as ‘too conservative’ and willing to incur high stabilisation costs in order to achieve low inflation rates. This difference between the ECB and the national policymakers regarding their degree of conservatism is (bE – bi). Then,

∂E(C i ) ∂( bE − bi )

(

)

= - λ (u n q + σ ν2 ) / 2bE bi < 0.

(9.11)

That is, the higher the perceived degree of ECB’s ‘excessive’ conservatism, the higher are the perceived costs of common monetary policy for country i.

Different θ ’s Consider the possibility that although the ECB and the national policymakers have the same preferences regarding the inflation-stabilization trade off, they disagree about what the targeted unemployment rate should be. Therefore, now bE = bi = b, but θ E < θ i. This situation may emerge either because the EMU is not structurally homogeneous and the average natural unemployment rate targeted by the ECB is higher than the natural unemployment rate in country i, or because the natural unemployment rate may be considered intolerable (politically unacceptable) by the policymakers of this country. This difference in preferences is given by θ E  θ i, and we assume that the ECB is free from any expansion bias, i.e., θ E  1. Then we have

∂E(C i ) ∂(θ E − θ i )

(

)

= (u n )2 ( b + λ2 ) / 2bθ i > 0.

(9.12)

That is, as θ i increases reflecting the national policymakers’ preference for a lower than the natural unemployment rate (assuming θ E=1), the perceived costs of country i’s policymakers from losing monetary policy autonomy increase as well.

Notes 1 Jordi Bacaria and Andreas Kyriacou received support from CIRIT SCR97–333 and CICYT SEC96–2300. Additional support has been received from the European Science Foundation in the context of the Network on Enlargement and new Membership of the European Union (NEMEU). When this contribution was written the second author was at the University of Connecticut. We thank the participants of

Enlargement and the European Central Bank

2

3 4

5 6 7 8 9 10 11 12 13 14 15

177

the third and fourth meetings of the Network on Enlargement and New Membership of the European Union (and especially our discussant Dieter Schmidtchen) as well as two anonymous referees for their comments. Any errors are the authors’ responsibility. The views are those of the authors alone and not of the institutions they are affiliated with. Cukierman (1992) reports a positive relationship between legal independence and average inflation for a group of seventy-two developing countries. Fischer (1995) suggests that this may be due to either the non-observance of the law and/or the absence of key aspects of independence (such as prohibition of debt financing). An example of this is New Zealand where, according to the Reserve Bank of New Zealand Act 1990, the government, after parliamentary approval, can demand that the bank adopt a particular monetary policy for six months (Garcia de Paso 1993). This is supported by ‘the fact that the Bank’s statute does not unambiguously reflect the primacy of maintaining price stability’ (EMI 1996: Annex 1). For recent tests of the political independence of the Bundesbank see – for evidence that supports the idea of independence – Berger (1997) and Berger and Woitek (1997), and, for the contrary view, see Vaubel (1997a, 1997b). Treaty Establishing the European Community, Art. 112. In more technical terms, the issue of setting short-term interest rates can be defined along one dimension along which members’ preferences are single peaked. See The Economist, 31 October 1998: 91–2. Recent empirical evidence suggests that closer economic integration (like that which arguably will be experienced by those countries participating in the Euro) tends to lead to more highly synchronised business cycles (Frankel and Rose 1997). These authors suggest that the creation of a single financial market and the operation of the common monetary policy itself will probably narrow, over time, the differences in the monetary policy transmission mechanism across EU countries. See, ‘Will the euro work? The ins and outs of EMU?’, by David Currie of the Economist Intelligence Unit, 1998. This may partly explain why there is conflicting empirical evidence on whether or not Reserve Bank Presidents in the Federal Reserve vote in defence of regional interests. See, for example, Gildea (1992) and Tootell (1991) for both views. More generally, Moser (1999) concludes that OECD countries, whose legislative decision rules are characterised by extensive checks and balances and based on effective bicameral systems, are associated with significantly more independent central banks. Notwithstanding this, Fratianni and von Hagen (1992) have suggested that the ECB will enjoy more independence than the Federal Reserve System because, rather than facing one unified government, it will face twelve (at the time) separate national ones. The Land Central Banks are located in Berlin, Düsseldorf, Frankfurt-am-Main, Hamburg, Hanover, Leipzig, Mainz, Munich and Stuttgart. Rotating membership in the Governing Council has also been proposed by the European Constitutional Group (1993).

References Alesina, A. and R. Gatti (1995) ‘How independent should a central bank be? Independent central banks: Low inflation at no cost?’, American Economic Review 85: 196–200. Alesina, A. and V. Grilli (1992) ‘The European Central Bank: Reshaping monetary politics’, in M. Canzoneri, V. Grilli and P. Masson (eds) Establishing a Central Bank: Issues in Europe and Lessons from the US, Cambridge: Cambridge University Press, 49–85. Alesina, A. and L. Summers (1993) ‘Central bank independence and macroeconomic performance: some comparative evidence’, Journal of Money, Credit and Banking 25: 151–62.

178

J. Bacaria, G.E. Chortareas and A.P. Kyriacou

Bacaria, J. (1993) ‘Supra-national or national monetary rules as constitutions: reflections on the European Monetary Union’, Economia delle Scelte Pubbliche 2: 157–71. Bacaria, J. and A. Kyriacou (1996) ‘Districts and the coordination of monetary policies in the European System of Central Banks’, Enschede: University of Twente, NEMEU Working Paper 96–5. Barro, G. and D. Gordon (1983a) ‘A positive theory of monetary policy in a natural rate model’, Journal of Political Economy 91: 589–610. Barro, G. and D. Gordon (1983b) ‘Rules, discretion and reputation in a model of monetary policy’, Journal of Monetary Economics 12: 101–21. Beck, N. (1990) ‘Congress and the Fed: why the dog does not bark in the night’, in T. Mayer (ed.) The political economy of American monetary policy, Cambridge: Cambridge University Press, 131–150. Berger, H. (1997) ‘The Bundesbank’s path to independence: evidence from the 1950s’, Public Choice 93: 427–53. Berger, H. and U. Woitek (1997) ‘How opportunistic are partisan German central bankers: evidence on the Vaubel hypothesis’, European Journal of Political Economy 13: 807–21. Bernholtz, P. (1987) ‘The implementation and maintenance of a monetary constitution’, in J. Dorn and A. Schwartz (eds) The Search for Stable Money. Essays in Monetary Reform, Chicago: University of Chicago Press, 83–117. Black, D. (1958) Theory of Committees and Elections, Cambridge: Cambridge University Press. Bofinger, P. (1992) ‘Discussion’, in M. Canzoneri, V. Grilli and P. Masson (eds) Establishing a Central Bank: Issues in Europe and Lessons from the US, Cambridge: Cambridge University Press, 77–80 Borins, S. (1972) ‘The political economy of the Fed’, Public Policy 20: 175–98. Bowles, P. and G. White (1994) ‘Central bank independence: a political economy approach’, Journal of Development Studies 31: 235–64. Brash, D.T. (1996) ‘New Zealand’s remarkable reforms’, London: Institute of Economic Affairs, Occasional Paper No. 100. Brennan, G. and J. Buchanan (1981) ‘Monopoly in money and inflation. The case for a constitution to discipline government’, London: Institute of Economic Affairs, Hobart Paper No. 88. Buchanan, J. and G. Tullock (1962) The Calculus of Consent: Logical Foundations in Constitutional Democracy, Ann Arbor: University of Michigan Press. CEPR (1992) Is Bigger Better? The Economics of EC Enlargement. Monitoring European Integration, Annual report, London: Center for Economic Policy Research. Cukierman, A. (1992) Central Bank Strategy, Credibility and Independence: Theory and Evidence, Cambridge, MA: MIT Press. De Grauwe, P. (1994) The Economics of Monetary Integration, Oxford: Oxford University Press. De Haan, J. (1997) ‘The European Central Bank: independence, accountability and strategy. A review’, Public Choice 93: 395–426. De Haan, J. and J. van ‘t Hag (1995) ‘Variation in central bank independence across countries: some provisional empirical evidence’, Public Choice 85: 335–51. Dornbusch, R., C. Favero and F. Giavazzi (1998) ‘Immediate challenges for the European Central Bank’, Economic Policy 26: 17–63. Eichengreen, B. (1990) ‘One money for Europe? Lesson from the US currency union’, Economic Policy 9: 118–87. Eijffinger, S. and J. de Haan (1996) The Political Economy of Central Bank Independence, Princeton: Princeton University, Princeton Special Papers in International Economics No. 19.

Enlargement and the European Central Bank

179

Eijffinger, S. and E. Schaling (1993) ‘Central bank independence in twelve industrial countries’, Banca Nazionale del Lavoro Quarterly Review 184: 1–41. European Constitutional Group (1993) A Proposal for a European Constitution, London: European Policy Forum. European Monetary Institute (1996) ‘Progress towards convergence 1996’, Frankfurt am Main. FED (1994) The Federal Reserve System: Purposes and Functions, Washington, DC: Board of Governors of the Federal Reserve System. Fischer, S. (1995) ‘Central bank independence revisited’, American Economic Review 85: 201–6. Frankel, J. and A. Rose (1997) ‘The endogenity of optimum currency criteria’, Swedish Economic Policy Review 4: 487–512. Fratianni, M. and J. von Hagen (1992) The European Monetary System and European Monetary Union, Boulder: Westview. Fratianni, M., J. von Hagen and C. Waller (1997) ‘Central banking as a political principalagent problem’, Economic Inquiry 35: 378–93. Garcia de Paso, J. (1993) ‘Sobre la independencia de los bancos centrales’, Papeles de Economía Española 57: 99–102. Gildea, J. (1992) ‘The regional representation of Federal Reserve Bank Presidents’, Journal of Money, Credit, and Banking 24: 215–25. Goodman, J. (1991) ‘The politics of central bank independence’, Comparative Politics 23: 329–49. Hayek, F.A. (1990) ‘Denationalization of money’, London: Institute of Economic Affairs, Hobart Paper No. 70. Hirschman, A. (1970) Exit, Voice and Loyalty: Responses to Decline in Firms, Organizations and States, Cambridge, MA: Harvard University Press. Kenen, P. (1995) Economic and Monetary Union in Europe, Cambridge: Cambridge University Press. Krause, G. (1996) ‘Agent heterogeneity and consensual decision making on the Federal Open Market Committee’, Public Choice 88: 83–101. Kydland, F. and E. Prescott (1977) ‘Rules rather than discretion: the inconsistency of optimal plans’, Journal of Political Economy 85: 473–91. Lohmann, S. (1992) ‘Optimal commitment in monetary policy: credibility versus flexibility’, American Economic Review 82: 273–86. Marquis, M. (1996) Monetary Theory and Policy, Minneapolis St Paul: West Publishing Company. Marsh, D. (1993) The Most Powerful Bank: Inside Germany’s Bundesbank, New York: Times Books. Mitchell, W. (1989) ‘The calculus of consent: enduring contributions to public choice and political science’, Public Choice 60: 201–10. Moser, P. (1999) ‘Checks and balances, and the supply of central bank independence’, European Economic Review 43:1569–93. Mundell, R. (1961) ‘A theory of optimum currency areas’, American Economic Review 51: 657–65. Persson, T. and G. Tabellini (1993) ‘Designing institutions for monetary stability’, CarnegieRochester Conference Series in Public Policy 39: 53–84. Posen, A. (1993) ‘Why central bank independence does not cause low inflation: there is no institutional fix for politics’, in R. O’Brien (ed.) Finance and the International Economy: Volume 7, Oxford: Oxford University Press, 40–65.

180

J. Bacaria, G.E. Chortareas and A.P. Kyriacou

Ramaswamy, R. and T. Slok (1998) ‘The real effects of monetary policy in the European Union: what are the differences?’, IMF Staff Papers 45: 374–93. Rogoff, K. (1985) ‘The optimal degree of commitment to an intermediate monetary target’, Quarterly Journal of Economics 100: 1169–90. Rogoff, K. and A. Sibert (1988) ‘Equilibrium political business cycles’, Review of Economic Studies 55: 1–16. Tootell, G. (1991) ‘Regional economic conditions and the FOMC votes of district presidents’, New England Economic Review 2: 3–16. Vaubel, R. (1997a) ‘The bureaucratic and partisan behavior of independent central banks: German and international evidence’, European Journal of Political Economy 23: 201–24. Vaubel, R. (1997b) ‘Reply to Berger and Woitek’, European Journal of Political Economy 13: 823–7. Waller, C. (1995) ‘The choice of a conservative central banker in a multisectoral economy’, American Economic Review 82: 1006–12. Walsh, C. (1995) ‘Optimal contracts for central bankers’, American Economic Review 85: 150–67.

Chapter Title

Part IV

Dynamics of widening

181

182

Authors name

Chapter Title

183

10 A never-ending success story? The dynamics of widening and deepening European integration Gerald Schneider1

Introduction One of the persistent puzzles of regional integration is the connection between enlargement and internal reform. The academic literature has not yet come to any conclusive judgement on how the two distinct forms of institutional evolution – the so-called processes of widening and deepening – are interrelated. This neglect is surprising since the European Union has often faced these challenges simultaneously. One example is the second southern enlargement, which more or less coincided with the conclusion of the Internal Market Programme and the Single European Act. The perplexing parallelism between the internal consolidation and the geographic expansion of the collaborative framework has not yet provoked any substantive evaluation, however. At the moment we do not know whether widening and deepening hinder or help each other or, as a third option, are not interrelated in any meaningful way at all. To complicate matters even more, all of these three hypotheses on the linkage between the depth and scope of regional integration have found their proponents in the current enlargement debate. The main reason for our deplorable lack of understanding of this interrelationship between the two most distinctive features of the integration process was, at least until the conclusion of the Maastricht Treaty, the tendency to perceive the development of the European Union as a never-ending success story. Before the organisation reached its mid-life crisis in the mid-1990s, the pace of integration seemingly followed a deterministic logic of a stepwise development, with a new institutional success always demanding a further enlargement round. The widening in return, at least in the perspective of the integration optimists, inevitably invited an attempt to advance the collaborative network further towards a situation of full integration. The more sceptical voices only became louder before the current Eastern enlargement. They contend that it will become increasingly difficult to succeed on both fronts at the same time. In their view, there is almost no possibility of gaining recalcitrant states through the conclusion of package deals that improve the situation of all current and future members of the organisation. This pessimistic view, however, contrasts considerably with a few important analytical articles. While some authors even stress that the two processes neces-

184

G. Schneider

sitate each other (Baldwin 1994; Pahre 1995), a paper by Downs et al. (1998) at least cautions against the current wave of integration scepticism. Their trade model shows that the level of integration can deepen and widen simultaneously as long as the admission process proceeds in a piecemeal manner and progressive states are among the front-line applicant countries. Although this model demonstrates the possible conditionality of the widening-deepening nexus, it does not explicitly examine how different institutional and geographical evolutions of a multilateral organisation affect the likelihood of an ever-expanding and ever-deepening level of cooperation in a region of states. Understanding the eventual relationship between widening and deepening requires an analysis of the simultaneous choices that a multilateral organisation like the European Union can make. I model these interrelated decisions in a simple formal model that tries to capture the essentials of integration negotiations. I rely on a modelling strategy because more empirical alternatives do not promise valid results. Although econometric tests might show the beneficiality of certain accessions, they are not able to capture how states use their institutional and economic structure to influence bargaining outcomes. Schelling’s paradox of weakness and the ensuing literature on two-level games suggest that the relationship between power and bargaining outcomes is much more complex (Schelling 1960; Putnam 1988; Schneider and Cederman 1994). This contribution is structured as follows: After a brief introduction to the empirical and theoretical literature, the chapter will first present a trade model that analyses the dynamics of integration from the perspective of a national government. I then move on to a game that shows how the level of integration helps to enforce future bargains. Finally, I discuss how uncertainty affects the dual process of widening and deepening within the collaborative network of the European Union.

Three views on the interrelationship between widening and deepening One of the most distinctive developments in the global political economy has been the resurgence of regional economic agreements during the late 1980s and the 1990s. The last boom of the concept dates back to the 1950s and 1960s when a great number of states tried to follow the example of the European Community. However, in most regimes time was not yet ripe for formalised economic cooperation. The trend ended in the recommendation that only less advanced forms of formalised interstate cooperation are feasible in the near future and that the study of regional integration is doomed (Haas 1975). Against this pessimistic backdrop, most member states of the World Trade Organization (WTO) have currently signed general trade treaties with other nations. To date, the WTO counts 214 agreements and enlargements, 134 of which are still in force.2 The renaissance of regionalism as both economic and political concept has invited much speculation on the repercussions that this process has on the political capacity of the nation state and these emerging polities. One theoretical

A never-ending success story?

185

strand argues that the trend towards closer cooperation at the supranational level gradually undermines the governance capacity of nation states, while statecentric authors contend that the transfer of sovereignty to Brussels strengthens rather than weakens the nation state (Vaubel 1994; Moravcsik 1994). In a series of provocative papers, Alberto Alesina and his colleagues (Alesina and Spolare 1997; Alesina and Wacziarg 1998; Alesina et al. 2000) show that the interrelationship might be much more complex. Their formal argument leads us to suspect that increased levels of economic interdependence further the disintegrative tendencies in nation states and political unions (Alesina et al. 2000). In this chapter I explore another salient feature of the surging trend towards political integration and analyse formally the relationship between the geographic scope and political depth of regional agreements. Obviously, interstate agreements differ widely along these two dimensions. Figure 10.1 accordingly presents four ideal types of formalised cooperation, distinguishing between the level of integration and the number of countries that have joined a specific framework. Although a clear benchmark for the distinction between ‘minilateralism’ and ‘multilateralism’ does not exist, bilateral and trilateral treaties are significant examples of the former type of arrangements; multilateralism involves a larger number of participating states. Political integration, broadly speaking, measures whether or not a treaty allows decisions to be made against the opposition of individual member states. A supranational level of integration indicates that some form of majority voting is possible. ‘Intergovernmentalism’, conversely, stands for an institutional setting in which each member state preserves its veto rights.

Figure 10.1 The depth and scope of regional integration.

186

G. Schneider

To begin, ‘intergovernmental minilateralism’ stands for a situation where a few nations can maintain much national decision-making power and only agree on some modest form of trade liberalisation. Examples fitting into this category are some of the Latin American arrangements that were founded during the past decade. Historically, they have their predecessors in the German customs unions of the nineteenth century. A second sort of formalised interstate cooperation, ‘intergovernmental multilateralism’, involves a greater number of countries, but is still characterised by a modest and mainly sectoral level of cooperation. Since the collaborative framework does not yet reach far beyond government declarations, the cooperation between the industrialised states around the Pacific in the form of the Asia-Pacific Economic Cooperation (APEC) currently still resembles a mainstream multilateral model and consequently falls into this category. Conversely, the cooperation intensity of the more club-like Northern American Free Trade Agreement (NAFTA) reaches deeper; yet the treaty is limited to only three member states, with discussion, however, underway to widen the geographical scope considerably. The European Union, even in its early years of existence and in considerable contrast to its Northern American counterpart, had already moved towards a considerable delegation of sovereignty from the national to the international level. ‘Supranational minilateralism’ stands for the junction of a modest number of member states and a relatively deep level of integration, which was typical of the European Union up to the 1970s. The last enlargements as well as some crucial constitutional reforms have transformed the organisational structure into a unique form of ‘supranational multilateralism’. Although the current reform debate is not complete after the Nice summit and the envisioned Eastern enlargement far from being settled, the organisation therefore represents an extreme case along both dimensions under consideration. Unsurprisingly, such an advanced stage of integration also provokes the question of whether or not the organisation can prevail on both fronts simultaneously. The erratic evolution of the European Union suggests on the one hand that disintegration is an unlikely event. The only withdrawal from EU membership occurred in 1985 when Greenland – or, in politically correct parlance, Kalaadtlit-Nunaat – left the organisation. However, although the organisation has never experienced the formal withdrawal or exclusion of a member state, its institutional structure is extremely vulnerable to any possible blackmailing strategies on which the least integrationist member might rely to advance its own interests. The foot-dragging by some well-known mavericks such as President De Gaulle and Prime Minister Thatcher has largely contributed to the stop-and-go nature of European integration, where one actor alone can still single-handedly block the accession of some applicant countries or the conclusion of both fundamental or partial reform processes (Schneider and Cederman 1994). Yet the formal analysis of individual negotiations cannot tell us anything about the long-term evolution of the process. This is only possible through a model that captures the interrelationship between enlargement and internal reform. To move in this direction, I will first look at the characteristic features of the two processes individually. Table 10.1 summarises the geographical expansion of the European Union, suggesting that accession tends to become longer over

A never-ending success story?

187

Table 10.1 The length between the application and accession

Country

Formal application

Accession (real/forecasted)a

Denmark Ireland United Kingdom Greece Portugal Spain Austria Cyprus Malta Sweden Finland Hungary Poland Romania Slovakia Latvia Estonia Lithuania Bulgaria Czech Republic Slovenia

May 1967 May 1967 May 1967 June 1975 March 1977 July 1977 July 1989 July 1990 July 1990 July 1991 March 1992 April 1994 April 1994 June 1995 June 1995 October 1995 November 1995 December 1995 December 1995 January 1996 June 1996

January 1973 January 1973 January 1973 January 1981 January 1986 January 1986 January 1995 (2005) (>2006) January 1995 January 1995 (2005) (2005) (>2006) (>2006) (>2006) (2005) (>2006) (>2006) (2005) (2005)

Years between application and accession 7 7 7 6 9 9 6 (15) (>16) 4 3 (11) (11) (>11) (>11) (>11) (10) (>11) (>11) (9) (10)

Source: Adopted from Nicolaides et al. (1999). Norway, Switzerland and Turkey are excluded because the accession process has been stopped or stalled. Note: a There is still no firm decision on when the first Eastern European applicant state will enter the organisation. Judgments are based on the yearly Commission reports on enlargements and a study published by Deutsche Bank.

time and that the inclusion of nations with a GDP below the average of the current EU members provokes more extensive debates. Table 10.1 distinguishes between real and forecast accessions. If we believe that the original frontline applicant countries will be able to join the European Union some time between 2003 and 2006, then the average time between the application and the accession of these groups is around nine to ten years. For the other applicants, in most cases less wealthy and less democratic than the firstwave applicant countries, the average time between the application and the accession will exceed ten years even if we assume that the negotiations with the European Union will be very short following the eventually successful entry of the 5+1 countries that have lined up to the expectation of the European Commission. Obviously, the wealth of a country is an important intervening variable. This becomes clear if one takes a look at the rather short entry negotiations that led to the membership of Austria, Finland and Sweden. Be this as it may at the moment, a similar trend can be observed if we look at the fate of the official plans that tried to alter the level of cooperation within the European Union. With the notable exception of the reform attempts in the

188

G. Schneider

1960s, all official blueprints tried to increase the scope of integration.3 This can happen either through granting supranational institutions agenda-setting power or through the introduction of qualified majority voting. The first intergovernmental treaty, the ECSC Treaty, was announced in May 1950 and came into force two years later. The EDC Treaty was equally launched in the year 1950, but ultimately failed to be ratified four years later. The debates on the EEC/Euratom treaties also lasted for four years while the formal negotiations about the next significant and successful reform attempt, the Single European Act, only took two years. However, long internal reform discussions with a host of ultimately unsuccessful reform attempts preceded the launching of the Single European Act. The Tindemans report, the Genscher–Colombo plan, the Solemn Declaration of Stuttgart and the draft treaty of the European Parliament are some of the more memorable endeavours that belong to this period of institutional Eurosclerosis. Ratification problems in Denmark, Germany and Great Britain delayed the entry into force of the Maastricht Treaty for some time, increasing the discussion and negotiation period to about eight years. The debate on the Amsterdam and Nice Treaties did not last so long, although the plan to engage into some sort of ‘permanent reform’ was officially announced through the Maastricht Treaty in the early 1990s. This is not so surprising because the last intergovernmental conferences were interrelated events with modest aspirations. In particular, the Nice summit with its much deplored results showed that the organisation has moved to a piecemeal strategy. The comparison of the institutional and geographic evolution of the European Union suggests that negotiation processes seem generally to become more cumbersome the more advanced and the larger the organisation is. The early failures with reform attempts also reveal that the likelihood of a stalemate is greater in the organisation’s early years of existence, reflecting that the costs of a breakdown have grown in line with the institutional development. Although the organisational progress has occasionally stagnated in the past, most political commentators currently share the intuition that the European Union has reached a dangerous situation of institutional inertia. They believe that the organisation can advance on one but certainly not on both fronts simultaneously. Some even go as far as describing the current enlargement negotiations as a suicide attempt in which the organisation signs its own ‘death warrant’ (Dehousse 1995). The conventional wisdom maintains that the costs that the new membership creates are simply too large if the organisation wants to keep its current level of internal cooperation. The decision of the Prodi Commission to extend the negotiations to other applicant countries has not rendered the task easier. The gloomy forecasts of the empirical studies partly contradict the theoretical literature on this issue. While a first group of authors expects – in line with the enlargement pessimists – that the European Union finally approaches the institutional deadlock of a ‘joint decision trap’ (Scharpf 1988; Schneider 1997), some theorists conclude just the opposite and predict a positive relationship. Against the backdrop of these conflicting explanations, the third analytically viable option is the null hypothesis that no systematic relationship exists between

A never-ending success story?

189

the two processes. Although there is no space to review the more empirical literature on the current expansion of the European Union, I shall briefly outline these competing theoretical positions.4

The theoretical debate on enlargement The literature on the provision of collective goods offers a theoretical foundation for the optimist’s viewpoint. These models show that expanding the depth and scope of an agreement necessitate each other, at least up to a certain point. Proponents of the collective goods approach traditionally perceive the coincidence of widening and deepening as the result of overlapping jurisdictions. In this perspective, the European Union as a political entity provides a public good which simultaneously produces positive spillovers within this territory and a geographically larger economic jurisdiction (Olson 1969). Standard textbook versions of the approach demonstrate that reducing costs in the production of the joint public good leads to a simultaneous increase in the scope and size of integration (for instance, Cornes and Sandler 1986: 168–70). Unsurprisingly, therefore, Pahre’s application of the standard collective goods model (1995) to the European Union shows that the two processes do not inevitably contradict each other. He even argues that the demand for widening grows with every enlargement round since ‘adding members worsens the nonintegration outcome and raises the optimal level of integration’ (Pahre 1995: 112). Baldwin (1994) follows a similar line of inquiry and suggests that nonmember states will fall like dominoes into the arms of the European Union. Downs et al. (1998) challenge the view that there is a consistent and unconditional relationship between the evolution of depth and breadth of a multilateral agreement. They show that the long-term fate of a multilateral agreement might depend on the initial construction strategy. In their view, a sequential geographical expansion might lead to deeper integration than an inclusive strategy, provided that the liberal states are admitted before the more conservative ones. Their model further implies that a slow admission process might be beneficial since this form of timing forces conservative states to liberalise considerably prior to their admission to an international treaty. Most negative predictions assume that the increased preference diversity, which is most likely, but not necessarily, the result of an expansion of a multilateral treaty, poses a severe reform problem. Hence, expansion should heighten the tensions between member states, increase the transaction costs and enhance the bargaining complexity. This will allegedly happen in the event that the organisation favours the admission of countries with extreme preferences, which will constantly be in the minority in the Council of Ministers. In other words, the coupling of a naive expansion and an unconvincing reform strategy might easily further the disintegrative tendency within a multilateral treaty. As I have shown elsewhere, moving to qualified majority voting increases the tendency of the laggard member states of an organisation to block the further unanimous constitutional development of the organisation. The entry of preference outliers furthers the tendency (Schneider 1995a, 1995b).

190

G. Schneider

A similar argument is made by Alesina et al. (2000, see also Alesina and Spolare 1997) who predict a negative relationship between trade liberalisation and political integration. This formally deduced result follows from the insight that the size of a country’s market does not depend on the size of the political jurisdiction in a world of free trade. To bolster their rather counterintuitive claim, they note that the number of countries increased rapidly after World War II while the tariff rates were gradually eroding. Although these authors do not explicitly check the impact of systemic changes and do not distinguish between the possibly offsetting effects of internationalisation and regionalisation, the analysis suggests that the growing success of the European Union could ultimately incite political disintegration. Unsurprisingly, Alesina and Spolare (1999) contend in a more normative paper that European integration has gone ‘too far’. Even though such a development does not seem very likely for the European Union, the optimism of collective goods models is still in considerable opposition to the history of the organisation. If we take collective goods theory at its face value, we would have to expect a more or less smooth development towards increased levels of integration and a growing number of members. Further, collective goods models do not seem to be adequate because becoming a member of the European Union is a costly endeavour. As the authorities of the European organisation make it plainly clear to any self-declared applicant, new members have to pay an entrance fee and to accept the rules of the club they wish to join. In the language of public goods theory, the European Union constitutes a club that is characterised by a considerable heterogeneity among its members. One modelling strategy that one could pursue would accordingly be the development of a club good model.5 One drawback of this literature, however, is that it is not easily linked to the bargaining literature and often lacks any concern with institutional issues. Since incoming members have to accept the new status quo, the rules of the organisation are largely determined by the least integrationist member of the organisation. The unchallenged role of unanimity in the enlargement and the constitutional negotiations of the European Union suggests that the framework used by Downs et al. (1998) is highly inadequate for the purposes of this study. In their model, states rely on qualified majority voting to found and strengthen multilateral arrangements. This invites the question of why members or applicants might ratify an agreement that has been concluded against their own will. Hence, the model by Downs et al. assumes that a set of nations is able to strike a deal in the centre of the core. This optimality perspective runs counter to most bargaining models that stress the power of extremist states in a multilateral setting (Putnam 1988; Schneider and Cederman 1994). Hence, this alternative model neglects the possibility that most negotiators try to employ outside options to influence the course of action. The history of the European Union is full of evidence supporting the claim that the process of institution building frequently follows the pace of the slowest member. To overcome the limitations of the competing models, I will analyse the challenge of reconciling the widening and deepening of regional integration with an alternative design. The model which will be presented in the next section

A never-ending success story?

191

and which is fully fleshed out in Appendix 10 offers a framework of analysis that will be extended later. The machinery for this model derives largely from Schneider (1994), Schneider and Weitsman (1994) and Downs et al. (1998).

Model of widening and deepening Both widening and deepening are bargaining processes in which actors are concerned about the distribution of the anticipated gains. In a path-breaking paper, Fearon (1998; see also Hovi 1998) has shown that negotiations will become increasingly harder the longer the time horizon of the actors. However, enforcing an agreement might become more difficult if actors become impatient and if the ‘shadow of the future’ shortens. This makes it plausible that the relationship between widening and deepening crucially depends on the initial conditions. I assume with Schneider (1994) that multilateral decision-making on the level and size of integration can be depicted as a unidimensional bargaining space x (0 ≤ x ≤ 1), with 0 standing for a situation of complete integration and 1 depicting a situation of unbroken sovereignty. Further, some informal arrangement exists prior to the possible integration negotiations, with the status quo xs standing for the old international equilibrium. The model is based on the assumption that starting conditions are identical for all nations, with xsi ≡ xsj. The government of the first country i will be referred to as the home government, government j as the foreign government. A regional integration treaty involves a subset K ⊂ N{1,2,3, . . . n} of all nation states that potentially gather in a regional organisation. In a multilateral treaty, the pivot player is the government that is able to take the decisive position, xp, during the integration negotiations. Given the prevalence of unanimity in important multilateral treaties, this is obviously the least integration-minded member of the organisation. This assumption might seem unrealistic, given the increasing prevalence of majority voting in the European Union. Yet member states still resort to unanimity when they negotiate on the scope and depth of the integration project. While the organisation increasingly makes policy decisions by qualified majority, laggard member states continue to veto constitutional choices. Before they enter a bargaining round, governments establish distinct unilateral negotiation stances ¯xi. Since these positions reflect differences in the preferences of interest groups and the decisive foreign government, both domestic and international factors enter the integration game at an early stage. I assume that a simple political economy framework inspired by the conference version of Downs et al. (1998) is able to capture the essence of domestic politics. In this model, xij is the level of independence that the home government maintains. By contrast, xji represents the discretionary power of the foreign government towards the home government. Assuming a quadratic loss function, discrimination dij of i is minimal in the case of complete integration and maximal in the case of unshared sovereignty. Further, the more nations participate in a regional agreement, the smaller the costs of discrimination. This linearity follows from the insight that an expansion of the organisation changes the actors’ opportunity costs. Most partic-

192

G. Schneider

ularly, an increasing scope of the European Union allows member states to trade more within than outside the regional bloc. Moving towards further integration, in return, generates costs since the economy and society have to adapt to the new rules. The vulnerability parameter vij reflects that states depend to different degrees on the influence of other states. The more nations that are becoming members of an organisation, the larger this factor looms in the calculations of the governments. The third element of the objective function are the bargaining costs, which are minimal if the home government is in the pivot position. These bargaining costs increase non-linearly with the size of the organisation; hence, negotiations over an enlargement or an internal reform package create the costs k2(xp-xij)2, reflecting the assumption often made in the literature that a very large and heterogeneous organisation faces explosively increasing transaction costs. We can use these assumptions to derive unilateral integration stances expressing the tension between widening and deepening. Mathematically, these optimal negotiation positions that are derived fully in Appendix 10 take the following form:

xi =

viK i xs + knx p n − k Ki di + viK i + kn k

The most important and counterintuitive result that follows from these unilateral integration stances is that the tension between widening and deepening is nonlinear and conditional upon the interplay between the level of integration and the number of possible and actual applicants. This means that adding members will not be problematic as long as the size of the region expands through a system change, like the end of the cold war, or through the more gradual development of decreasing transportation costs. Hence, if the transaction costs start to weigh heavily, the average member state of an integration treaty will ask for a looser net of cooperation. In the actual situation, however, this trend is partly offset by the countervailing influence that the increasing number of potential members has on the integration process itself. Figure 10.2 illustrates this crucial dilemma. The simulation shows the difficulty in evaluating the beneficiality of the Eastern enlargement. The impact of the other influences is straightforward, with an increasing vulnerability for instance rendering governments more prudent on the market for international collaboration. The relative impact of the different independent variables on the integration stances is summarised in Table 10.2.

Enforcement, bargaining and uncertainty: adding some additional complexities The likelihood of reaching enduring agreements is higher in an open-ended process than in a single-shot interaction. Hence, nation states form integration policies by considering the benefits that might arise from a long-term collaboration with other states. Governments can agree to relinquish sovereignty and

A never-ending success story?

193

Figure 10.2 The impact of the number of member states (y) and the size of the regional system (x) on the depth of integration (vertical axis).

Table 10.2 Effects of independent variables on the depth of a unilateral integration stance Increase in:

Effect

Size of region (n) Size of member states (k) Domestic vulnerability (v) Foreign discrimination (d) Status quo (xs) Position of pivot government (xp)

– ± + + + +

Note: + (–) signifies that a government wishes to move towards a more (less) reluctant unilateral integration stance.

enforce integration by punishing a defecting partner state. This is done through the adoption of trigger strategies: a state’s unilateral stance beyond a certain threshold of acceptable behaviour sets off punishment by other member states for a predetermined time period. The range of such trigger strategies goes from ‘Tit-for-Tat’ to ‘Grim Trigger’ (Downs and Rocke 1990). In the latter strategy, a single defection sets off perpetual punishment, while in the former strategy punishment is limited to one iteration. The implicit threat of exclusion elicits cooperation from member states by reducing the range of acceptable behaviour. The threat of exclusion is equivalent to a Grim Trigger strategy. The presence of this threat is particularly important if the objective function of a government for a single-shot interaction changes and disintegration becomes more profitable than integration. I assume that the enforcement costs c of the treaty grow with the number of participating states.

194

G. Schneider

This means that it becomes more and more costly to defect the larger the cooperative network is. Since I am not interested in the optimal period of punishment, I study Grim Trigger enforcement. States maintain their integration agreement under Grim Trigger if the discounted rewards from cooperation exceed the gains from a one-time shift to defection and the subsequent punishment. The condition for the application of a Grim Trigger strategy is that the discounted benefits of cooperation exceed the utility of one defection plus the discounted value of the ensuing punishment. Ri + wiRi / (1 – wi) ≥ Ti + Pi / (1 – wi) where R is according to the standard notation the reward for cooperation and T stands for the temptation to cheat on the other member states. Further, P represents the punishment payoff and w stands for an individual discount parameter. Enforcement problems might possibly arise if states join the organisation, but later on refuse to live up to the obligations of the accession treaty. A further and often neglected eventuality is that the organisation a nation is about to join does not fulfil the financial promises that it made during the enlargement negotiations. Quite typically, some kind of financial concession is made by the organisation to entice an eventual member state to swallow the pill of unshared sovereignty in some policy domains. I model the enforcement problem in line with Hovi (1998) as a two-actor game in which the applicant can enforce the treaty obligations or not and the EU is able to compensate the new member or to break its financial promises. Since states anticipate the possible enforcement problems, compliance considerations enter their bargaining calculations before starting the actual negotiations. If the European Union compensates a state, it has to pay the costs xEU. The compensation takes the form of xA for the applicant state which, in return, has to pay the price x if it implements the treaty. The enforcement costs that the organisation has to pay if a new member does not live up to the treaty obligations increase with the size of the organisation and accordingly amount to kEEU. I also assume, however, that the size of the bargain increases with the number of states that have joined the organisation. The applicant state, finally, has to discriminate against the other states if they do not live up to the treaty obligations. These costs are depicted as eA. The basic enforcement problem that results from the confrontation of an applicant state with an organisation like the European Union is depicted in Figure 10.3.6 Comparative statistics show how a member state of the European Union approaches the prospect of enlargement if the possible enforcement problems have to be taken into account. First of all, the larger k is, the more integrationist the bargaining position of the organisation. Most interestingly, the bargaining space narrows, the larger the discount factor. This basically means, in line with Fearon (1998) and Hovi (1998), that states are more concerned about a new member if the organisation is successful and the opportunity costs of destroying the collaborative framework are high. Hence, bargaining becomes increasingly difficult if the organisation has reached a certain momentum.

A never-ending success story?

195

Figure 10.3 Enforcement and compensation in enlargement negotiations.

This result nicely illustrates why the internal reform debates within the European Union have become increasingly cumbersome.7 It also suggests that the bargaining space is wider if the European Union is about to accept a preference extremist as its new member. In such an instance, states will not negotiate as hard as with an applicant state whose preference is much closer to those of the member states. The counterintuitive insight that dealing with preference outliers might be beneficial to some extent probably accounts for much of the naivety with which the European Union has embarked on the current enlargement negotiations. This influence of a short ‘shadow of the future’, however, is partially offset by the impact that uncertainty has over the future course of action within a possible member state. Most obviously, an organisation might start to mistrust a negotiation partner, which is not able to commit itself towards a predictable course of action. In the view of the European Union, a clear commitment towards the principles of the market economy and democracy are the major prerequisites for long-term stability. Uncertainty over the domestic situation in an applicant state thus accounts for much of the discriminatory strategy of the European Union, which distinguishes between the front-line applicant states and a second row of membership candidates. In general, establishing the reservation price for the international bargain constitutes the creation of what Schneider and Weitsman (1994) call the ‘entry threshold’ of regional collaboration. The threat of exclusion that the reliance on a Grim Trigger strategy entails automatically reduces the range of acceptable behaviour. Applicant countries have to make it clear that they must take this first step irrevocable. If the European Union expects a shift towards a disintegrative policy, it automatically alters its integration stance and subsequently its negoti-

196

G. Schneider

ation position. This explains why states with highly volatile domestic politics are unattractive partners in an integration agreement, although the international treaty might help them to overcome anti-integrationist tendencies. The same is true for authoritarian partners where governments have the capacity to suddenly alter integration stances. Uncertainty about the other side’s behaviour and subsequent severe enforcement costs might generally be the main reason why noncoercive integration is largely limited to democracies. By the same token, while some institutional convergence eases the way towards long-term agreements, significantly different states can attain enduring cooperation as long as some zone of agreement exists. The enforcement issue also allows us to speculate about the possible evolution of an organisation. If future interactions within the cooperative framework loom large enough, a regional agreement is most likely to survive for quite some time. While such an organisation might be rather inflexible in its negotiations with an applicant state, it has also a great interest in continuing and eventually enlarging the regional arrangement. The final extension of the model that we will discuss, but not formally prove here for space limitations, is the question of how the possibility of enforcement problems affect the negotiation.8 In enlargement negotiations between the organisation and the applicant, the first move is typically made by the European Union (or the European Commission, to be more precise). The next move is then up to the applicant state. Ståhl (1972) and Rubinstein (1982) have shown that the solution to such a bargaining game with changing offers and discounting is equivalent to

x=

1−δ A 1 − δ EUδ A

where δEU and δA are the bargaining discount parameters of the European Union and the applicant country, respectively.9 In contrast to the enforcement situation, patience is an asset in international negotiations. If we have two actors with similar and high discount parameters, they can expect similar shares of the immediate benefits of regional integration. In the event that the power differential between two states is large, however, the more patient and thus probably the more powerful negotiator will largely prevail. This explains on the one hand the almost anticipated resentment of some negotiators who think that the European Union is dictating the negotiation agenda too much in the current enlargement negotiations. The power differential is also one major reason why the Eastern enlargement round is taking so long. While the applicant states would rather like to join the organisation today, the European Union plays with time to ensure that the entry threshold of the organisation is really taken before new members join the organisation.

Conclusion The analysis presented in this contribution supports the view of Downs et al. (1998) that the relationship between the depth and scope of regional integration

A never-ending success story?

197

is more complex than the sombre and, in a few cases, rosy speculations of some policy analysts indicate. First, the relationship is basically curvilinear since moving from a status of exclusive sovereignty to some sort of formalised cooperation is easier for a large rather than a small number of states. Yet, in the event that a regional organisation gains momentum, adding members becomes less desirable because of the dramatically increasing transaction costs. This non-linear, but plausible relationship between widening and deepening has been derived from a trade model that synthesises the assumptions of both enlargement pessimists and optimists. Second, more successful organisations and states with a long time horizon might negotiate harder since an enlarged shadow of the future aggravates rather than alleviates the enforcement problem. Taking into account the overall success of the integration process, this explains why both enlargement and internal reforms have become more complicated over time. Third and finally, it pays off to be patient in integration negotiations. Since rich states do not discount the future as heavily as poor applicant states, they are more likely to be on an equal footing with the European Union. Neither side has, in such a situation, an interest in playing with time. Short bargaining rounds will be the most likely consequence of such a constellation. With regard to the current negotiations, this contribution tentatively argues that enlarging the European Union will most likely increase the wish of some member states to loosen the collaborative network. To move towards a clearer picture, however, we need to compare the costs of integration and disintegration more carefully. The analysis suggests that the current negotiations are hampered by the serious enforcement and renegotiations problems that the European Union will face if it includes more that twenty members. Implementation rather than treaty reform will be the main problem in the decades following the next enlargement rounds. While this contribution addressed the logic of the eventual nexus between widening and deepening, it has not moved beyond empirical illustrations to hone the basic insights. The analysis only suggests that the current negotiations, provided that the organisation really extends beyond its nadir, could be one of the underestimated disintegrative forces that will shape the integration discourse of the coming decades. The chapter offers, however, the analytical tools that make us understand why the old members accept new members and do not fear that their achievements will be diluted. The future will show whether they were right.

APPENDIX 10 This appendix contains the derivation of the unilateral negotiation stances and the negotiation positions of both the EU and the applicant state. The first step in establishing the unilateral negotiation stances is the deduction of a cost function. Adding these elements together leads to the following

198

G. Schneider

cost function:

C=

n− k n

∑d x ij

ji

+

j≠i

k n

∑v (x ij

s

− xij )2 + k2 ( x p − xi )2

j≠i

(10.1)

which can be transformed into a more general cost function that differentiates between the costs imposed on a probable member by non-participating and participating states in a region.

C=

n− k n

∑d x

2 ij i

+

j ∈K , j≠i

2 n− k k dij xs + n j ∉K , n

∑ j≠i

∑ v (x ij

s

− xi )2 + k2 ( x p − xi )2

j ∈K , j≠i

(10.2)

Equation 10.2 can be rewritten as

C=

n − k Ki 2 k ( di xi + di( N − K ) i x2s ) + viK i ( xs − xi )2 + k2 ( x p − xi )2 n n

(10.3)

in which Ki equals K-{1} and

diKi =

∑d

ij

i ∈K

(10.4)

is an example of how the summation over a set is symbolised in the following. Equation 10.3 can then be used to derive the following first order condition:

0=2

n − k ki k di xi − 2 viK i ( xs − xi ) − 2k2 ( xp − xi ) n n

(10.5)

After some manipulation, we derive the one-shot unilateral integration stance, which amounts to

xi =

viK i xs + knx p n − k Ki di + viK i + kn k

(10.6)

If we only look at the further expansion of a treaty, we can further derive the positions with which the two negotiation partners, the European Union and the applicant state, enter the treaty negotiations from this framework of analysis. If the organisation lives up to the daunting task of Grim Trigger enforcement, enlargement negotiations will become feasible under the following conditions:

wEU ≥

xEU k −1 x+ xEU − eEU k

(10.7)

A never-ending success story?

199

and

wA ≥

x eA

(10.8)

which we can use to derive the minimal negotiation positions:

X EU ≥

(

)

WEU eEU − x

 k −1  WEU  − 1  k 

(10.9)

and

x A = w Ae A

(10.10)

Notes 1 Previous versions of this contribution have been presented at the Joint Meeting of the DVPW-sections ‘Staatslehre und politische Verwaltung’ and ‘Integrationsforschung’, Munich, 29–31 October 1998, the fourth NEMEU research seminar, Prague, 7–10 January 1999, and on an internal research seminar at the University of Konstanz. I would like to thank participants for their comments, especially Steffen Baumann, Thomas Plümper, Bernard Steunenberg and Dieter Wolf. My gratitude goes also to Simon Hug for his detailed comments. 2 See the information on www.wto.org/english/tratop_e/region_e/regfac_e.htm (last consulted, 9 May 2001). 3 The exceptions were ‘Fouchet I’, ‘Fouchet II’ and the proposal that became ultimately known as the ‘Luxembourg compromise’. 4 Useful and systematic information on these issues can, for instance, be found in Baldwin (1994) or von Hagen (1996). Lippert (1998), Prange (1998) and Rode (1998) are some of the more descriptive accounts of the enlargement process. 5 Sandler and Tschirhart (1997) offer a recent review of this literature. 6 Note that the resulting game resembles a Chicken Game rather than the Prisoners’ Dilemma situation that Hovi (1998) and Fearon (1998) suggest. 7 In the set-up of spatial models, it can equally be shown that bargaining becomes more cumbersome if the status quo in the intergovernmental bargains moves closer to the ideal point and thus the Pareto set of the organisation’s members. See Schneider (1997) for an illustration of this point. 8 A similar modelling strategy is pursued by Hovi (1998), while Fearon (1998) uses a war of attrition model to study the bargaining sequence in the creation of international regimes. 9 I thus distinguish between the expectation actors build vis-à-vis the length of the negotiations and vis-à-vis the durability of the agreement that they might eventually conclude.

References Alesina, A. and E. Spolare (1997) ‘On the number and size of nations’, Quarterly Journal of Economics 112:1027–56. Alesina, A. and E. Spolare (1999) ‘Has european integration gone too far?’, NBER Working Paper No. 6883.

200

G. Schneider

Alesina, A., E. Spolare and R. Wacziarg (2000) ‘Economic integration and political disintegration’, American Economic Review 90: 1276–96. Alesina, A. and R. Wacziarg (1998) ‘Openness, country size and government’, Journal of Public Economics 69: 305–21. Baldwin, R.E. (1994) Towards an Integrated Europe, London: Centre for Economic Policy Research. Barnes, J. and M.P. Barnes (1995) The Enlarged European Union, London: Longman. Colomer, J.M. (1996) ‘A dynamical model of decisions by unanimity rule’, Paper presented at the Joint Sessions of Workshops of the European Consortium for Political Research. Cornes, R. and T. Sandler (1986) The Theory of Externalities, Public Goods and Club Goods, Cambridge, MA: Cambridge University Press. Dehousse, F. (1995) ‘In deciding to push east, the EU has signed its death warrant’, International Herald Tribune, 29 December 1995. Downs, G. and D.M. Rocke (1990) Tacit Bargaining, Ann Arbor: University of Michigan Press. Downs, G., D.M. Rocke and P.N. Barsoom (1998) ‘Managing the evolution of multilateralism’, International Organization 52: 397–419. Fearon, J.D. (1998) ‘Bargaining, enforcement, and international cooperation’, International Organization 52: 269–305. Haas, E.B. (1975) The Obsolescence of Regional Integration Theory, Berkeley, CA: Institute of International Studies (Research Series No. 25). Hagen, J. von (1996) ‘The political economy of Eastern Enlargement of the EU’, in L. Ambrus-Lakatos and M.E. Schaffer (eds) Coming to Terms with Accession, London: CEPR, Institute for East-West Studies (Forum Report of the Economic Policy Initiative 2), 1–41. Hovi, J. (1998) ‘Reducing global carbon emissions: joint implementation, sequential bargaining, and enforcement’, Paper presented at the Joint Meeting of the ECPR Standing Group on International Relations and the International Studies Association, Vienna, 16–19 September 1998. Inotai, A. (1998) ‘The CEECs: from the Association Agreements to full membership’, in J. Redmond and G.G. Rosenthal (eds) The Expanding European Union: Past, Present, Future, Boulder, CO: Lynne Rienner, 157–76. Kawecka-Wyrzykowska, E. (1996) ‘On the benefits of the Accession for Western and Eastern Europe’, in L. Ambrus-Lakatos and M.E. Schaffer (eds) Coming to Terms with Accession, London: CEPR, Institute for East–West Studies (Forum Report of the Economic Policy Initiative, 2), 85–107. Lippert, B. (1998) ‘Der Gipfel von Luxemburg: Startschuß für das Abenteuer Erweiterung’, Integration 21: 12–31. Moravcsik, A. (1994) ‘Why the European Community strengthens the state: domestic politics and international cooperation’, Paper presented at the Annual Meeting of the American Political Science Association, New York, NY, 1–4 September 1994. Nicolaides, P., S.R. Boean, F. Bollen and P. Pezaros (1999) A Guide to the Enlargement of the European Union (II): A Review of the Process, Negotiations, Policy Reforms and Enforcement Capacity, Maastricht: European Institute of Public Administration (revised edition). Olson, M. (1969) ‘The principle of ‘fiscal equivalence’: the division of responsibilities among different levels of government’, American Economic Review 59: 479–87. Pahre, R. (1995) ‘Wider and deeper: the links between expansion and integration in the European Union’, in G. Schneider, P.A. Weitsman and T. Bernauer (eds) Towards a New Europe: Stops and Starts in Regional Integration, Westport, CT: Praeger, 111–36. Prange, H. (1998) Die Ostintegrationspolitik der Europäischen Union, Marburg: Tectum.

A never-ending success story?

201

Preston, C. (1997) Enlargement and Integration in the European Union. London and New York: Routledge. Putnam, R.D. (1988) ‘Diplomacy and domestic politics: the logic of two-level games’, International Organization 42: 427–60. Rode, R. (1998) Die Integration Mittelosteuropas in die Weltwirtschaft, Amsterdam: Fakultas. Rubinstein, A. (1982) ‘Perfect equilibrium in a bargaining model’, Econometrica 50: 97–109. Sandler, T. and J. Tschirhart (1997) ‘Club theory: thirty years later’, Public Choice 93: 335–55. Scharpf, F.W. (1988) ‘The joint decision trap: lessons from German federalism and European integration’, Public Administration 66: 239–78. Schelling, T. (1960) Strategy of Conflict, Cambridge, MA: Harvard University Press. Schneider, G. (1994) ‘Getting closer at different speed: strategic interaction in widening European integration’, in P. Allan and C. Schmidt (eds) Game Theory and International Relations, Cheltenham: Edward Elgar, 125–55. Schneider, G. (1995a) ‘The limits of self-reform: institution building in the European Community’, European Journal of International Relations 1: 59–86. Schneider, G. (1995b) ‘Ein Opfer des eigenen Erfolges: Mehrheitsentscheide als Sprengsatz für die Selbsterneuerung der Europäischen Union’, Homo Oeconomicus 12: 437–53. Schneider, G. (1997) ‘Auswege aus der Verflechtungsfalle? Eine Analyse der wichtigsten Reformoptionen’, in T. König, E. Rieger and H. Schmitt (eds) Mannheimer Jahrbuch für Europäische Sozialforschung: Band 2, Frankfurt: Campus, 163–77. Schneider, G. and L.-E. Cederman (1994) ‘The change of tide in political cooperation: a limited information model of European integration’, International Organization 48: 633–62. Schneider, G. and P.A. Weitsman (1994) ‘Cooperation among equals: a theory of regional integration’, Paper presented at the 1994 Annual Meeting of the American Political Science Association, New York, USA, September 1–4. Sedelmeier, U. and H. Wallace (1996) ‘Policies towards Central and Eastern Europe’, in H. Wallace and W. Wallace (eds) Policy-Making in the European Union, Oxford: Oxford University Press (3rd edition), 353–87. Ståhl, I. (1972) Bargaining Theory, Stockholm: Economic Research Institute, Stockholm School of Economics. Vaubel, R. (1994) ‘The public choice analysis of European integration: a survey’, European Journal of Political Economy 10: 227–49.

202

Authors name

11 How political parties, rather than member states, are building the European Union Josep M. Colomer1

Introduction The evolution of the European Community (EC) and the European Union (EU) demonstrates that not only formal institutions but also organisations, such as political parties, can play an aggregating role, rendering institutional decisionmaking feasible in a pluralistic community. In the long run, political parties tend to adapt the institutional framework to their own ‘play’, in other words, to transform patterns of behavior into formal rules. The basic model that can cast light on the EU’s development in this respect runs as follows. When the initially established regulatory institutional framework is relatively loose, parties choose certain strategies and informal rules of behavior that shape their relations and produce collective outcomes. These patterns of behavior tend to become more formal and constraining institutional rules through incremental changes of the existing institutions. In this process, the aggregating role of political parties is extended to the institutional framework, which makes decision-making more regular and more predictable.2 The institutional development of the European Union can be approached with three different models, involving different roles of actors and parties, which can be called, respectively, ‘intergovernmental’, ‘imperial’, and ‘federal’. The intergovernmental model was dominant from the foundation of the European Economic Community in 1957 until mid 1980s. The institutional relations corresponding to this model are those of an international or diplomatic organisation. The main actors are the governments of member states. The Council of the European Union (formerly called the ‘Council of Ministers’) appears to be the central institution, as representatives of the governments of the member states form it. As they consider themselves to be sovereign, the member states have veto power on collective decisions, which means that most decisions are made by unanimity. Similarly, the European Council, consisting of the EU’s heads of state or government, generally acts on the basis of unanimity. After an initial period of innovative decisions in favour of a common market, during the 1960s and 1970s, unanimity rule proved to be highly ineffective in making decisions, especially after successive enlargements of the EC from six to twelve member states. A single member’s veto (as was frequently the case of French governments led by General De Gaulle, British governments led by

How political parties are building the EU

203

Margaret Thatcher, and others) was able to paralyse any initiative. Unanimity rule in the Council led to a period of stagnation, ‘Eurosclerosis’ and ‘Europessimism’, as it was called in the early 1980s. In reaction to increasingly pervading sclerosis, the EC accepted that certain decisions could be made by qualified majorities and subsequently enforced only by a subset of member states. This model has been called ‘reinforced cooperation’ or ‘closer cooperation’, as well as ‘variable geometry’, ‘concentric circles’, ‘strong core’, ‘two speeds’, ‘à la carte’, ‘flexible’, and other interesting metaphors. I suggest that it should be called ‘imperial’ because it corresponds to formulas used in classical empires, including, for instance, the German-Roman Empire or Austria-Hungary.3 The multilevel institutional structure produces multiple but different overlaps between institutions at different levels regarding the territory and the issues on which they make final decisions. In fact, each territorial unit may implement a particular subset of the set of enforcing decisions made at higher institutional levels. For example, the Social Chapter of the European Community was approved without the United Kingdom, the euro was initially accepted by only twelve out of the fifteen member states, the Schengen agreement on border controls affects a different subset of countries, etc. The main weakness of this model is that it always creates the opportunity for every member state to undertake its own decisions on common matters and withdraw from some Europe-wide developments. The third, federal model includes the principle of territorial representation, as the other two models previously sketched, but it is more effective in decisionmaking because it is not based upon unanimity rule (in contrast to the intergovernmental model) and makes enforceable decisions for all countries (in contrast to the imperial model). The deployment of the federal model would require certain changes in inter-institutional relations. This should include in particular an increase in the legal powers of the European Parliament in order to cover the so-called ‘democratic deficit’, in other words, the relative limited role traditionally played by the only directly elected European institution. Political parties are the main actors able to make decision-making in the European Parliament feasible and, by this way, foster federalising inter-institutional relations. If the typical formula of a federal state were to be adopted, representatives of the people of the EU at large would share decision power with representatives of the territorial units. The European Parliament, which is elected directly by universal suffrage in a more or less simultaneous election and is internally organised in large European Political Groups, should be considered to be the lower chamber, the representative institution of the European citizens. The Council, which represents the citizens of each country through the institutions of each member state and tends to work on the basis of country’s interests rather than on strict ideological positions, should be considered an upper chamber of territorial representation. A further improvement could give the Council as many representatives per member state as votes. Both chambers, the Parliament and the Council, should have significant legislative and mutual veto powers. The Parliament and the Council are elected separately and with different rules. This usually produces the possibility that different political party majorities can be formed in each chamber. As a result,

204

J.M. Colomer

EU inter-institutional decisions have to be based on very broad agreements encompassing two different majorities, which is also a common formal feature of several federal states with symmetrical bicameralism (such as Germany). In this framework, the European Commission should be considered to constitute the EU’s executive.4 Accordingly, the Commission would be appointed by, and made accountable to, the two parliamentary chambers. It would then rely on a large consensual supermajority, which would not be very different from the one which was obtained with the intergovernmental procedures. Some fears of ‘ungovernability’ of the EU have been expressed in the face of the possibility to move towards more innovative federal formulas. Yet, as suggested here, the federal model could actually maintain some of the consensual features that have traditionally characterised the working of the Union’s institutions, but with higher effectiveness. As a matter of fact, the decision powers of the European Parliament have increased as a result of a series of incremental institutional reforms. The European Parliament was required only to be ‘consulted’ by the Council according to the Rome Treaty (1957). Significant changes began after the first direct elections in 1979. Further stages include the introduction of new rules forcing the Council to share certain legislation powers with the Parliament or cooperation procedure, as embodied in the Single European Act (1987), the codecision procedure for legislative acts, as well as the formal requirement of the Parliament’s support for the appointment of the Commission President, as introduced by the Maastricht Treaty on the European Union (1992), and the virtual replacement of the cooperation procedure with codecision rules for most policy areas, as well as the requirement of a formal investiture of the Commission and its President by the Parliament prior to the Commissioners taking office, as introduced by the Amsterdam Treaty (1997) and enforced from 1999 on. This chapter maintains that these changes would not have been able to give effective capability of decisionmaking to the highly pluralistic EU in the late 1980s and the 1990s if they had not been responded to with aggregating initiatives by political parties. Political parties in the European Parliament benefit from the opportunities and incentives given by the existing institutional rules at any given moment. Accordingly, they have developed more cooperative and aggregative strategies in order to gain influence in the inter-institutional process. But, at the same time, political parties create these opportunities by promoting treaties and constitutional-like regulations that replace the dominant role of the Council with more complex formulas. Political party formation and coalition building in the European Parliament is a driving force that makes the governance of the highly pluralistic European Union relatively effective and consensual. Electoral parties in the European Parliament have been involved in a long process of forming large political groups of increasingly Europe-wide scope. They have also developed continuous activities of majority coalition formation, in both processes of voting and the appointment of office holders. In spite of the fact that some national parties have been hesitant to participate in these processes and the developments have tended to be slow, gradual changes have rendered inter-institutional decisionmaking in the EU feasible and increasingly effective.

How political parties are building the EU

205

Multinational European Political Groups and multiparty parliamentary coalitions create intermediate aggregation mechanisms between highly pluralistic political and territorial representation and European-level decision-making. By this, the Parliament has been able to become an active partner in the inter-institutional process and to create federalising links with the other institutions. In other words, political party strategies fill the gap that is left by loose institutional regulations and constraints and make EU inter-institutional relations relatively tight and effective. The plan of this chapter is as follows. First, I analyse the formation of European Political Groups in the European Parliament from the time of the first direct election in 1979 to the fifth election in 1999. Then, I apply certain models of coalition formation and distribution of voting power to measuring the power of European Political Groups in forming multiparty coalitions in the European Parliament. I discuss the dilemma faced by certain parties between trying to lead a small group within the Parliament or to join a larger, more influential group. The analysis of voting power both of European Political Groups in the Parliament and of national parties within the European Political Groups shows the incentives for national parties to favour aggregation in decision-making. I provide some empirical evidence on the increasing role of the European Parliament in EU inter-institutional decision-making. I conclude with an evaluation and some positive expectations with respect to the governability of the European Union and its further evolution in the direction of a more federal system.

Party formation in the European Parliament Successive enlargements of the EU have led not only to an increasing number of member states, but also to more political parties obtaining representation in the EU institutions. The members of the Council and the European Parliament (as well as most Commissioners) are members of political parties, which have run separately in national and in European elections respectively. No Europe-wide significant aggregation is developed at the stage of forming electoral candidacies for the European Parliament. Electoral rules in the EU’s member states vary considerably. All of them are now based on the principle of proportional representation, but electoral formulas and ballot procedures differ for each member. Candidacies, the selection of issues and campaign strategies are primarily conducted in a national framework. The European Parliament’s elections, which have been termed ‘second-order national elections’, appear to have more of a character of national ‘midterms’ than of true European events (Reif 1985; Marsh 1998). Over one hundred different political parties have obtained representation at once in the European Parliament. Such a high level of party fragmentation might make inter-institutional decision procedures very ineffective and contribute to preserving the basic features of the EU’s ‘intergovernmental’ model. Yet, interestingly, enlargements and the corresponding increase in the number of parties obtaining representation in the European Parliament have not caused a

206

J.M. Colomer

growth in the total number of European Political Groups. On the contrary, the number of European Party Groups decreased after the two most recent elections, in 1994 and 1999. Remarkably, the effective number of European Party Groups has decreased steadily since the first election to the European Parliament in 1979. The ‘effective number’ takes into account the relative size of each party, by this way measuring the degree of fragmentation of the party system.5 In spite of representing fifteen member states with their own political regimes and party systems, the European Parliament has a lower effective number of parties than the parliaments of several European countries, such as Belgium, Denmark, Finland, Italy and the Netherlands. The relatively small number of effective European Party Groups and the decrease of this number over time provide support for the interpretation that political parties have developed some effective strategies to reduce fragmentation and to form European Party Groups of an aggregating character. This trend has favoured decision-making within the European Parliament and interinstitutional relations in the EU. Table 11.1 provides an overview of the number of member states, national political parties, and European Political Groups, as well as the corresponding effective numbers, by taking every unit’s relative proportion of seats into account. As can be seen, both the absolute and the effective number of member states have considerably increased between 1979 and 1999. The absolute number of member states has passed from nine to fifteen. The effective number of member states, weighting for their size, has increased in accordance, from six to 9.5.6 The number of political parties obtaining representation in the European Parliament has also increased, from forty-one in 1979 to 110 in the election of 1999. Yet the absolute number of European Political Groups, which increased twice during the first period, from eight to eleven, has decreased twice afterwards, from eleven to nine.7 The effective number of European Political Groups, as weighted by their size and thus reflecting the degree of aggregation from an increasingly pluralistic setting of elected parties, has decreased from a value of 5.2 in 1979 (with a slight increase by 1984) to 4.0 in 1999.

Table 11.1 Fragmentation of the European Union (1979–1999) Election year Member states – effective number Political parties European political groups – effective number

1979

1984

1989

1994

1999

9 6.0

10 6.6

12 8.2

12 8.1

15 9.5

46

70

95

11 5.0

10 4.6

41 8 5.2

9 5.3

110 9 4.0

Note: Data correspond to the initial composition of the European Parliament after every election. The ‘effective number’ (N) reflects the number and the relative size of the units represented; it is calculated according to the formula N = 1 / pi2 , where p is the proportion of seats of member state, party or group i.

How political parties are building the EU

207

‘Moderate’ parties have tended to prevail in the European Parliament. In all elections since 1979, the European Socialist Party and the Christian Democrats or European People’s Party together have gathered more than 50 per cent of total number of seats. The Socialists and the Christians together have encompassed increasing proportions of seats with every election, from 54 per cent in 1979 to 66 per cent in 1999. The smaller group of the Liberals, always the third in number of seats, is usually placed in an intermediate position between the Socialist and the Christians when measured on a left–right ideological scale. Hence, a ‘connected’ winning coalition could always be formed with the Socialists, the Liberals and the People’s Party. This scheme based on these three Europeans Political Groups – the Socialists on the centre-left, the smaller, intermediate Liberals in the centre, and the Christians on the centre-right – closely resembles the domestic party structure in Germany, and it is not dissimilar to the distribution of parties in Belgium, the Netherlands and Luxembourg. Yet significant parties from most of the other EU member states made the initial choice of forming their own European Political Groups, which produced a number of small groups dominated by a single large national party (i.e., one that holds more than half of the seat total in the group). This choice reflected the resistance of certain domestic-based parties formed in specific party-system configurations to adapt to the basic configuration of the party-system in the European Parliament previously described. As will be shown below, however, smaller groups have moved toward aggregation into the two largest European Political Groups, the Socialist Party and the People’s Party.8 The formation of aggregative European Political Groups has followed different paces among the centre-left and the centre-right parties. In only two out of five elections, the set of centre-left parties (those included in the Groups of Socialists, Rainbow, Greens and Radicals) collectively obtained more seats than the set of centre-right parties (those included in the groups of Christian Democrats, Conservatives, Democratic Alliance, Forza Europa and Europe of Nations). Yet, the Socialists were the largest European Political Group in the Parliament after the first four elections. The Socialist Party integrated national parties from all the member states since the first election in 1979.9 In contrast, the most integrative party on the centre-right side of the spectrum, the People’s Party, included national parties from only seven out of nine member states in 1979, while representatives from all member states were included only since 1989. In other words, while, for instance, the British Labour, the Central European and Nordic Social-democrats, and the Southern Socialists were integrated in the same European Political Group since the first direct election of the European Parliament, the British Conservatives, the Central European Christian Democrats and other major centre-right parties, such as the French Gaullists and Forza Italia, were in different European Political Groups for quite a long time. The Liberal Group and all other smaller Groups have never been based on representation from all member states. These trends are illustrated with figures regarding proportion of seats in Table 11.2.

208

J.M. Colomer

Table 11.2 Seats and voting power of European Political Groups (1979–1999) Election year s

1979 vp

1984 s vp

1989 s vp

s

1994 vp

s

1999 vp

Left • C/UL • EUL

11 –

6 –

9 –

1 –

3 5

1 1

5 –

1 –

7 –

3 –

Centre-left • V • Rb/Ra • PS

– – 28

– – 28

– 5 30

– 3 24

6 2 35

8 3 34

4 3 35

3 3 28

8

6

29

6

Centre • LDR

10

47

7

39

10

36

8

38

8

46

Centre-right • PP • DG/FE • DA • EN

26 15 5 –

10 9 0 –

25 12 7 –

14 11 8 –

23 7 4 –

7 5 4 –

28 5 4 3

10 6 6 4

37 – – 5

34 – – 3

Right • ER/NA/DD

–

–

4

0

3

1

5

1

2

2

Others • NA • CDI

2 3

– –

1 –

– –

2 –

– –

– –

– –

4 –

–

100

100

100 100

100

100

100

100

100

100

Total

Note: s = proportion of seats; vp = voting power based on connected winning coalitions. Numbers are rounded by the method of greatest remainders, according to rules by Balinski and Rachev (1997). Left: C/LU/UL: Communists (1979, 1984)/United and Nordic Left (1989, 1994, 1999); EUL: European United Left (1989). Centre-left: Rb/Ra: Rainbow (1984, 1989)/Radicals (1994); V: Greens (1989, 1994, 1999); PS: Socialists (1979, 1984, 1989, 1994, 1999). Centre: LDR: Liberals (1979, 1984, 1989, 1994, 1999). Centre-right: PP: Christians (1979, 1984, 1989, 1994, 1999); DG: Conservatives (1979, 1984, 1989); FE: Forza Europa (1994); DA: Democratic Alliance (1979, 1984, 1989, 1994); EN: Europe of Nations (1994, 1999). Right: ER/NA: European Right (1984, 1989)/Non-Attached (1994); DD: Democracies and Diversities (1999). Others: NA: Non-Attached (1979, 1984, 1989, 1999); CDI: Coordination Defense of Independents (1979).

Parties, coalitions and power In order to discuss in the following section the incentives for national parties to contribute to form aggregating European Party Groups, in this section I introduce some concepts on coalition formation and party bargaining and voting power. The type of coalition that parties tend to form depends on the institutional setting and the motives of the participating members. Basically, party motives are distinguished as office-seeking and policy- or ideology-seeking.

How political parties are building the EU

209

These motives introduce restrictions on the size of the coalition and on the policy-ideology position of the preferred partners, respectively. Four basic types of ‘viable coalitions’ produced by different party motives can be distinguished: winning coalitions (no specific restrictions); minimal winning coalitions (restriction on size); connected winning coalitions (restriction on ideology); and minimal connected winning coalitions (restrictions on ideology and on size). It can be expected that different institutional incentives will promote the formation of different types of coalition. Winning coalitions without restrictions can be found, for example, in international organisations, as well as in the EU Council, where the multidimensional issues submitted to collective decisions and the absence of offices to distribute among its members make rigid restrictions on the choice of partners irrelevant. Minimal winning coalitions are likely to be found in office elections in corporate bodies with no significant ideological cleavages. Connected winning coalitions can be expected to correspond to ideologically oriented parties in settings in which no offices are elected, such as multiparty assemblies in presidential regimes, as well as the European Parliament. Minimal connected winning coalitions correspond to ideologically oriented parties in parliamentary regimes. When applied to a committee or a parliament, every model produces a different set of ‘viable coalitions’. In order to measure the distribution of power for each party within every viable coalition, a general assumption is that only non-superfluous partners of the coalition can be considered to be ‘decisive’ or powerful.10 In the following sections, I use two models of coalition formation and the corresponding power indices on the basis of their appropriateness to measure political party power within the EU institutions. First, I analyse coalition formation among European Political Groups in the European Parliament from the assumption that the European Political Groups care about their relative ideological positions and form ideologically ‘connected’ winning coalitions between neighbour partners on the left–right spectrum, in a similar way as has widely been assumed for the analysis of multiparty coalition formation in domestic parliaments, but they do not care about the size of the coalitions because, in contrast to domestic parliaments, they do not elect executive officers. Second, the power of national parties within every European Political Group in the Parliament is measured on the assumption that, on the basis of sharing the same basic ideological allegiance (as happens, for example, among the members of the European Socialist Party, or among the members of the European People’s Party, and so on), they form winning coalitions without restrictions. When political parties form coalitions in the national Parliaments of most EU member states, their relative policy or ideology positions matter. A compelling, albeit informal rule of behaviour is that parties tend to prefer close neighbours in the ideological space to more distant potential partners even if the latter could be sufficient in terms of their number of seats to form a majority coalition. This criterion of behaviour reduces the set of viable winning coalitions and, as a result, public policy decisions tend to be more consistent.11 The basic ideological affinities between the European Political Groups, as mentioned above, tend to be situated on the left–right axis, as is the case in most

210

J.M. Colomer

national parliaments. The relevance of the left–right dimension in the European Parliament and the location of the European Political Groups has been analysed with empirical data, finding that ‘coalitions form on the basis of ideology, not nationality’ (Kreppel and Tsebelis 1996: 2).12 Based on the relative positions of European Political Groups, as presented by the European Parliament itself, five major party positions can be distinguished from left to right:13 1

2

3 4

5

Left: the Communists in 1979 and 1984, and the Communist-dominated United and Nordic Left in 1989/United Left in 1994, as well as the European United Left (1989); Centre-Left: the Groups formed by the Socialists in the five elections from 1979 to 1999, the Rainbow in 1984 and 1989 the Greens since 1989, and the Radicals in 1994 Centre: the Liberals in the five elections; Centre-Right: the Christian Democrats in the five elections, the Conservative Democrats in 1979, 1984 and 1989, the Gaullist Democratic Alliance in 1979, 1984, 1989 and 1994, Forza Europa in 1994, and Europe of Nations in 1994 and 1999; Right: the Right in 1984 and 1989, Democracies and Diversities in 1999, as well as the Non-Attached in 1994 due to it rightist and nationalist composition.

In order to calculate the set of viable coalitions corresponding to the assumption that the European Political Groups maintain their ideological connection when they chose coalition partners, I assume that a coalition can be formed by groups placed on contiguous positions in the policy space, for example, between the Socialist, the Liberal and the People’s Groups. We assume that groups cannot form a coalition with a group placed in a non-contiguous position without including all the groups in the intermediate positions. For example, in the 1999 Parliament a coalition encompassing the United Left and the Liberals should also include the Socialists and the Greens. Similarly, a coalition including the Liberals and the rightist Democracies and Diversities should also have the People’s and the Europe of Nations Groups as partners. As mentioned, I assume that European Political Groups keep their ideological connection between the partners of the viable winning coalitions, but they do not necessarily care about the size of these coalitions. The latter assumption is realistic in this institutional context because the European Parliament does not elect or control a government and hence, the groups cannot be office-seekers. Like ideological political parties acting in legislatures of non-parliamentary regimes, and in contrast to typical parliamentary regimes more frequently analysed in the coalition formation literature, the European Political Groups are considered to be policy-seekers, restricting partnership in coalitions to their neighbours, but accepting the addition of numerically superfluous coalition partners. On the basis of these criteria, I assume that European Political Groups in the European Parliament form connected winning coalitions, that is, coalitions of ideologically neighbour parties regardless of the size of the coalition. Within the

How political parties are building the EU

211

set of viable connected winning coalitions, there is some variation depending on the subject matter.14 Unpublished data for all 611 votes from July 1999 to March 2000 show that a centre-left majority, including the Socialists and the Liberals as its core, is formed in 84 per cent of votes on civil liberties, foreign affairs and environment (the Greens join the majority on these issues in 78 per cent of votes). The majority built around the centre by the Socialist, the Liberal and the People’s groups, forms in 72 per cent of votes on constitutional, institutional, budgetary affairs and fishery. Centre-right coalitions based on the Liberal and the People’s groups are also formed in 88 per cent of votes on economic affairs, industry, budget control and employment.15 More generally, in the European Parliament elected in 1999, it is possible to form sixteen connected winning coalitions, of which five are minimal connected winning coalitions. These coalitions are formed by: 1 2 3 4 5

United Left–Greens–Socialists–Liberals, with 328 seats; Socialists–Liberals–People, with 463 seats; Greens–Liberals–People, with 331 seats; Liberals–People–Nations, with 313 seats; and Liberals–People–Democracies and Diversities, with 299 seats.

The non-attached are not taken into account, on the assumption that their members may be added to any of these coalitions. The corresponding power index is calculated by dividing the number of decisive partnerships of a European Political Group in the set of connected winning coalitions by the total of decisive partnerships for all parties. This calculation shows that ‘centre’ players tend to gain in influence. When parties respect their ideological connectedness, the intermediate parties on the policy-ideology scale are advantaged. In particular, the party containing the median seat is a sure partner of any viable winning coalition. As a result, its relative bargaining power exceeds its proportion of seats (as well as its power if it is measured on the basis of an approach that abstracts from actors’ policy or ideology motives). Table 11.2 provides the values for the ‘legislative’ power index in the European Parliament based on connected winning coalitions. The comparison of these values with the proportions of seats held by the respective European Political Groups illustrates that the relative influence of parties in forming coalitions by bargaining and voting, as assessed on the basis of this power index, is highly favourable for the Liberals (which always contain the median seat). It is more or less fair for the largest party, the Socialists in the first four elections and the People’s Group in 1999, and to the disadvantage of all the other Groups.

The dilemma of political parties Generally, political parties face a dilemma with respect to their mode of representation in the European Parliament. They can choose either to try to lead a small European Political Group, and hence be highly influential within this

212

J.M. Colomer

Group, or to join a larger European Political Group that is more influential in the Parliament, within which they may have, however, little influence. The rules for forming European Political Groups are rather permissive. According to the 14th edition of the Rules of Procedure for the European Parliament enforced from May 1999 on, a European Political Group can be formed by 23 deputies if they come from two member states, or by 18 deputies if they come from three member states, or by 14 deputies if they come from four or more member states. Once the European Political Groups have formed, every group makes decisions on the basis of a majority of its members. National party members are submitted to the internal voting discipline of the European Political Group to which they belong.16 Under the rather flexible requirements for forming European Political Groups in the five elections since 1979, only an average of 3.9 per cent of the deputies has not been able to become a member of a European Political Group and thus be forced to join the group of the ‘Non-Attached’ or remain as independent. Incentives for national parties to either lead a small group or join a larger group can be assessed by comparing the relative voting power corresponding to either choice. For this purpose, I again resort to power index measurements. The power of a national party within the European Parliament can be approached by multiplying the ‘legislative’ power index previously derived for a European Political Group in the European Parliament by the appropriate power index of the national party within the European Political Group. This will give us an estimation of the relative power of a national party in the European Parliament as it results from its belonging to a specific European Political Group. In order to measure the power of the national parties within a European Political Group, I assume that any majority coalition of national parties can be formed within a European Political Group without significant restrictions. I assume that there are no significant ideological barriers between the political parties within a European Political Group because they share a common basic ideological position. Similarly, size should not be a relevant restriction, since the purpose of coalition-formation in this context is not the allocation of offices, but the establishment of a policy position of the respective European Political Group in the European Parliament. Hence, the internal bargaining power of the political parties within the European Political Groups is here analysed on the basis of the normalised Banzhaf power index. This is calculated by summing the decisive partnerships of a party by the total of decisive partnerships of all parties in the set of winning coalitions. By weighting the (Banzhaf) power index of a political party within its European Political Group with the ‘legislative’ power index of the European Political Group in the European Parliament, we obtain a measure of the power of the national party in the European Parliament.17 The global power of party i in the European Parliament will be referred to as P(i). The values of this measurement appear to have been significantly different for the national parties on the centre-left and for those on the centre-right at different moments. This suggests an explanation for the different paces in the process of successive aggregation of national parties in powerful European Political Groups.

How political parties are building the EU

213

In summary, while the very large European Socialist Party was still enlarged with the Italian post-Communists and reunited up to twenty parties, the European People’s Party was gradually enlarged with the members of previous small Groups led by the British Conservatives, the French Gaullists and Forza Italia, among others. Within the centre-left, a rather strong degree of concentration into the European Socialist Party can be observed from the first election in 1979. After the election of 1989, the Italian Communists (PCI), the largest party on the left side of the policy spectrum in Italy at the time, left the European Communist Group and formed the European United Left, together with a few minor partners. Yet this turned out to have been only an intermediate step, since at the following election, in 1994, the Italian Communists, already transformed into the Party of Democratic Left (PDS), entered the European Socialist Party. Domestic political strategies can largely account for the ideological evolution of the PCI, which led to the corresponding realignment within the European Parliament. But the approach developed here can cast light on the advantages of the corresponding moves in the European Parliament, which, given its specific party-system configuration, might not necessarily be as beneficial for the party as those in domestic politics. The PCI was the dominant party within the European Communist Group in 1979 and 1984, as it had more than 50 per cent of the seats of the group. Yet, the group became decreasingly powerful within the European Parliament, as illustrated in Table 11.2. The relative power of the PCI in the Parliament, as estimated according to the procedure previously discussed (i.e., its power index within the European Communist Group multiplied by the power index of the Communist Group in the European Parliament), can be estimated with the following values: P(PCI): 6.2 in 1979; P(PCI): 1.5 in 1984; P(PCI): 0.7 in 1989; P(PDS): 2.5 in 1994. These values suggest that the PCI made an initial somewhat costly move by creating its own group, the European United Left, in 1989, since it made it even less powerful than in its previous membership of the Communist Group. The corresponding loss, however, was minor in absolute terms, given the low relative power already attained by the Communist Group in the Parliament. Once in the European Socialist Party in 1994, the PCI, now PDS, was scarcely influential within its group (the corresponding Banzhaf value is about 9 per cent). But by this way it obtained higher power, that is, more global influence within the Parliament, than in its previous grouping. In parallel, the power of the remaining European Communist Group was reduced to almost zero. The evolution of political parties in forming European Political Groups has been more intricate within the centre-right of the European Parliament. A longterm tendency to an increasing concentration in the People’s Party can be identified. However, the incentives for every political party to enter the PP were not always immediately positive. Using the approach previously presented, we analyse the relative positions of three small European Political Groups in the centre-right space: the Conservatives or Democratic Group, the Gaullists of Democratic Alliance and Forza Europa. These groups were formed at the initiative of major centre-right parties in countries not fitting the German and Benelux configur-

214

J.M. Colomer

ation of party system, together with minor parties or fractions. Most of them have held nationalistic positions and rather reluctant stands in the face of further advances toward a closer union in Europe. The British Conservatives dominated the Democratic Group, as they always had more than 50 per cent of the seats in the group. The corresponding power values of the British Conservatives (Con) in the European Parliament were: P(Con): 9.4 in 1979; P(Con): 10.6 in 1984. The Spanish Popular Party provisionally joined the Democratic Group in 1987, after Spain entered the EC. The Democratic Group became then quite powerful, even obtaining the Presidency of the European Parliament in 1986–9. Yet at the following election in 1989 (the first one for the Spaniards), the Spanish Popular Party, faced with what still would be an absolute majority of Britons in the group that would have nullified its power, moved to the European People’s Party, where it obtained some positive influence. The British Conservatives, remaining in their own dominated group, attained then lower power than in the previous two legislatures: P(Con): 5.5 in 1989. In 1992 the British Conservatives, followed by their previous Danish minor partner in the Democratic Group, moved into the European People’s Party, in spite of significant policy and ideological differences in their traditional positions, including in regard to the EU further union and enlargement. Yet, the corresponding power value was P(Con): 1.2 in 1994, that is, a less powerful position than in its previous grouping. Although this loss was partly produced by the lower representation in seats of the British Conservatives in the Parliament of 1994 in comparison with the previous elections, the move to join the People’s Party in itself did not produce clear advantages for them. This may have rendered the appeal of the European People’s Party for further potential new partners rather uncertain. The French Gaullists (RPR) dominated the Group of Democratic Alliance, as they always had more than 50 per cent of the seats in the group. The corresponding power values of the French Gaullists in the European Parliament were: P(RPR): 0 in 1979; P(RPR): 7.6 in 1984. Some French Liberal and Republican candidates who were included in the UDF coalition in domestic politics and had run separately from the Gaullists in previous elections to the European Parliament, ran together with them in 1989. Yet they split into two groups once in the Parliament. The new electoral partners returned to the European Liberal Group, while the Gaullists formed again their own-dominated group, Democratic Alliance. The corresponding power value was P(RPR): 4.1 in 1989, that is, lower than in the previous occasion. At the following election in 1994, all the Christian Democrats, Republicans and Liberals forming the UDF and the Gaullists of the RPR ran together in a single candidacy with the commitment to remain in the same European Political Group after the election. Yet they split again, although this time a larger fraction of the previously coalesced candidates joined the European People’s Party: thirteen European deputies went to the People’s Party, only one returned to the Liberal group, and fourteen remained in the Democratic Alliance. In addition, an anti-EU candidacy was formed by ex-members of the Gaullist party, obtaining thirteen seats and the possibility of forming their own European Group, Europe of Nations. This group obtained a notable power in the European Parliament, P(EN): 4.4 in 1994; P(EN): 2.9 in 1999. In spite of all

How political parties are building the EU

215

of this, the Gaullists still retained some significant global power in the European Parliament, giving them incentives to maintain their own group. The corresponding power value was P(RPR): 6.1 in 1994. The European Political Group Forza Europa was dominated by Forza Italia (FI), the new centre-right party formed in the early 1990s after the crisis of the Italian Christian Democrats in domestic politics. FI became a national government party in 1994 and, on the basis of its relative success in the European Parliament election of that year, it formed its own European Group. The corresponding power value of Forza Italia in the European Parliament was P(FI): 5.7 in 1994. During the 1994 legislature FI merged with the French Gaullists in the Democratic Alliance Group, forming the new European Political Group Union for Europe. While the RPR had dominated the Democratic Alliance Group, as illustrated before, FI obtained a very influential position within the new group at the expense of the RPR. The corresponding power values in the Parliament were: P(FI): 12.7 in 1995; P(RPR): 0.8 in 1995. The European People’s Party has, thus, not only gathered together Christian Democratic parties from nine countries (Belgium, France, Germany, Greece, Ireland, Italy, Luxembourg, the Netherlands, as well as the Catalan and Basque nationalists). It has gradually integrated other national parties, including the Conservatives from Britain and Denmark having previously formed a separate group, and the Spanish Popular Party who had provisionally joined the latter. The number of French seats in the European People’s Party has also increased. Initially, the fragmentation of the French centre-right worked against the group, whose French membership went from nine in 1979 to eight in 1984 and six in 1991 (to the advantage of the Liberal Group and the Democratic Alliance). Yet during the 1989 legislature the People’s Party welcomed four French members of the Liberal Group (led by the former president of the group) and one from the Democratic Alliance, gathering together eleven at the end of the period, and then increased its number to thirteen after the 1994 election. A minor Portuguese party, the CDS, left the People’s Party after the 1994 election on the basis of its anti-EU stance. Yet the major centre-right Portuguese party, the PSD, previously in the Liberal Group, moved to the People’s Party during the 1994 legislature, thus globally increasing also the Portuguese membership in the Group. Shortly before the end of the 1994 legislature, Forza Italia joined the People’s Party and tried to persuade its Gaullist former partner in the Union for Europe Group to do the same. The French finally did so after the election of 1999, obtaining the symbolic concession of renaming the group as ‘European People’s Party (Christian Democrats)/European Democrats’. The corresponding power value of the new French component of the People’s Party is P(RPR): 2.9 in 1999, thus making their move into the more aggregative People’s Party positively rewarding. These moves suggest that the process of forming an encompassing European Political Group within the centre-right space in the European Parliament by the gradual enlargement of the People’s Party has been a relatively costly process. Some national parties having previously formed their own dominated, small

216

J.M. Colomer

groups had to accept temporary sacrifices in the expectation of obtaining similar moves from other parties to enhance the global power of the People’s Party. Observers have noted that the degree of internal ideological homogeneity of the group has somewhat diminished with successive integration of new parties in it, with some tensions developing between Christian Democrats and more Conservative partners. But, in the long-term, aggregation within the centre-right cluster of the European Parliament has been as effective as the aggregation in the Socialist Party, not only in terms of relative proportion of seats, but, according to our analysis, also with respect to bargaining and voting power in order to form majority coalitions in the European Parliament. Actually, some degree of initial heterogeneity is essential in any aggregation process. The formation of increasingly large, aggregative European Political Groups in the European Union during the late twentieth century can be compared to certain processes of forming electoral and parliamentary nationwide political parties in West European countries during the late nineteenth and early twentieth centuries (including, among others, Britain, France, Germany and Spain). Political representatives of the time were strongly rooted in local constituencies, they promoted local interests with no clear priorities on national issues, and many of them held vague ideological positions. The parallel can be extended to the fact that some domestic parliaments at the time were relatively weak in making the executive accountable (especially in certain monarchies). But the aggregation of local representatives in national parties created higher policy and ideology homogeneity among political representatives while contributing to reinforcing the role of parliaments, in a comparable way as the aggregation of national representatives into European Political Groups is now promoting more effective decision-making and contributes to giving the European Parliament more institutional strength. To the extent that national parties are submitted to majority decisions within large European Political Groups and to their group’s voting discipline in the Parliament, nationalistic positions against an ever-closer union obtain lower prominence in the institutional process.

The increasing role of the European Parliament According to the assumptions and expectations previously developed, empirical evidence shows remarkably high and increasing degrees of parliamentarians’ participation and internal cohesion of the European Political Groups in the European Parliament. Participation of individual members of the European Parliament in voting sessions has increased from about 48 per cent in 1989–94 to 61 per cent in 1994–9 and to 72 per cent in 1999–2000. In the Parliament elected in 1999, the highest levels of attendance correspond to the largest groups, the People’s, the Socialists, the Liberals and the Greens, all of them above the average, with values between 73 and 77 per cent.18 Internal cohesion of the European Political Groups has also increased steadily at successive legislative periods. On the basis of analysis of Roll Call votes, Table 11.3 shows the average proportions of individual members of the Parliament

How political parties are building the EU

217

Table 11.3 Voting cohesion of European Political Groups (1984–2004) election year

1984

1989

1994

1999

Average

74

84

88

90

Left • C/UL • EUL

71 –

93 94

84 –

84 –

Centre-left • V • Rb/Ra • PS

88 70 79

85 100 89

94

68 62

92

Centre • LDR

70

86

80

95

Centre-right • PP • DG/FE • DA • EN

84 83 76 –

88 92 65 –

90 81 93 71

90

Right • ER/NA/DD

96

100

–

68

83

Sources: For the periods starting in 1984, 1989, and 1994, adapted from calculations on samples of votes in Attinà (1990), Raumio (1996), Hix and Lord (1997) and Hix (1999). For 1999–2000, author’s calculations based on all 611 votes from July 1999 to March 2000 from Secretariat General European Liberal, Democratic and Reformist Group (ELDR) (2000). Note: Internal cohesion of European Political Groups is measured on the basis of roll call votes. The averages at the top are percentages of all individual members of the European Parliament (excluding Non-Attached) having voted according to their political group (these numbers are different from the Group averages given by the sources mentioned below). The other numbers are percentages of individual members of each political group voting as a block. See Table 11.2 for full party names and the method for rounding numbers.

voting in accordance to their European Political Group. As can be seen, internal cohesion rises over time: from 74 per cent in 1984–9 to 84 per cent in 1989–94, 88 per cent in 1994–9, and 90 per cent in 1999–2000. Specifically, the largest groups, the People’s, the Socialists, the Liberals and the Greens, have reached the highest degrees of discipline in voting, between 90 and 95 per cent in the Parliament elected in 1999. Internal cohesion of European Political Groups in the European Parliament is certainly higher than that of political parties in the United States Congress. These striking results contradict conventional assumptions that the larger the European Political Groups and the higher the number of their national components, the less cohesive they would tend to be. In contrast, as larger groups have higher chances of being decisive in the Parliament, as discussed previously, decisiveness increases the incentives for individual members to participate and to vote as a block.

218

J.M. Colomer

As previously suggested, the aggregation of national parties in large aggregative European Political Groups, their internal cohesion and their choice of ideologically close partners in forming majority coalitions, make decision-making in the European Parliament feasible. By becoming more effective in decisionmaking, the European Parliament is able to play a more significant role in interinstitutional relations. There is ample empirical evidence showing an increasing degree of influence of the European Parliament in EU inter-institutional processes, as can be observed for the Parliament’s role in legislative activity, in budget approvals and in its control on the European Commission. First, the introduction of the codecision procedure, which permits a decision to be made with the support of only a qualified majority of the Council together with the majority of the Parliament, has increased the volume of legislation. An annual average of fifty-five legislative acts were approved by the cooperation procedure, in an overwhelming proportion by the Council unanimity, during the period 1987–1993. In contrast, an annual average of forty-nine legislative acts by cooperation and twenty-six by codecision – so, a total of seventy-five legislative acts per year – were approved during the period 1993–8. In other words, the codecision procedure has been used in about 35 per cent of the total legislative activity of the EU from its introduction in 1993 to 1998.19 This is consistent with the claim that under the Amsterdam Treaty’s version of the codecision procedure, the Parliament is a co-equal legislator with the Council, whereas the Commission’s influence is likely to be more informal than formal (Tsebelis and Garrett 1999: 3). Empirical analysis also based upon the data published by the European Parliament shows that ‘on the aggregate there is a higher success rate of parliamentary amendments under codecision than under cooperation’ (Tsebelis et al. 1999: 3). Second, the role of the European Parliament in the elaboration and approval of the EU budget has become increasingly relevant. During the early years of the Community, from 1958 to 1974, the Parliament had no real budgetary powers. From 1975 on, the Council and the Parliament developed frequent conflicts regarding the classification of expenditures between ‘compulsory’ (as such expenditure necessarily derived from the EU treaties, according to the Council interpretation) and ‘noncompulsory’ (which initially concerned only administrative expenditure). By this way, the Council reserved for itself exclusive ‘competences’ on agriculture and certain international agreements, which amounted to the bulk of the EU budget. At the requirement of the Parliament, the European Court of Justice judged the 1986 budget, calling upon the two institutions to reach an agreement. More cooperative decisions were fostered by the Interinstitutional Agreement on budgetary discipline and improvement of the procedure signed in 1988, which led to the approval of the financial framework for the period 1994–9. In 1995, on the basis of the lack of agreement with the Council, the President of the European Parliament declared the final adoption of the EU budget unilaterally, although a new decision of the Court of Justice forced resumption of the previous procedure. Inter-institutional cooperation between the Council and the Parliament led to the approval of the EU Financial Agenda 2000–2006 in May 1999.20

How political parties are building the EU

219

Finally, the European Parliament provoked the resignation of the whole European Commission, including its President, after voting a motion of censure and threatening with further initiatives of control in March 1999. Interestingly, successive motions of censure on the Commission by the Parliament had obtained increasing number of favourable votes: 16 in June 1990, 8 in July 1991, 96 in December 1992, 118 in February 1997, and 232 in January 1999. The latter was still short of a majority, but the credible threat that a new vote could obtain larger support precipitated the resignation of the Commissioners a few weeks later. Similarly, votes of confidence on incoming Commissions have obtained increasing numbers of negative votes: 31 in February 1981 (Thorn Commission), 34 in January 1985, not recorded in January 1989, 84 in January 1993 (on the three Delors Commissions), 104 in January 1995 (Santer Commission), and 138 and 153 in September 1999 (Prodi Commission) (see Corbett et al. 2000: 236, 245).

Conclusion The evolution of the institutions of the European Union shows that the role of political parties in rendering institutional decision-making feasible in a pluralistic community can be crucial. European political parties have played an aggregating role within a loose institutional framework that initially allowed mere ‘intergovernmental’ relations to develop. Aggregating initiatives from a highly fragmented set of national political parties in order to form larger European Political Groups, as well as majority coalitions among these Groups in the European Parliament, fill the gap left by relatively low constraining institutional rules. The corresponding success in decision-making helps the institutions themselves to become more consistent and effective. To the extent that the European Parliament was highly fragmented, the Council, making decisions by unanimity, could remain the dominant institution of the EU. The basic features of the intergovernmental model were maintained, but this led the EC to a long period of paralysis or ‘Eurosclerosis’, as it was called in the 1970s and early 1980s, provoked by the opportunities given to some individual national governments, as represented in the Council, of vetoing further measures of European integration and union. To the extent that the formation of European Political Groups and multiparty coalitions makes decisions in the European Parliament relatively easier, interinstitutional relations corresponding to the federal model for the European Union are fostered. Decision-making becomes more effective not only regarding public policies, but also regarding treaties and constitutional-like regulations, so reinforcing the tightness of the institutional framework and the role of the Parliament in it. As nationalistic groupings are replaced by larger European Political Groups in which every national party is submitted to the group discipline, further decisions toward a closer union can be made with less nationalistic resistance. In contrast to previous periods of ‘Europessimism’, new treaties and major decisions have been approved in the 1990s with increasing frequency, including a new enlargement up to fifteen member states and the creation of a new common currency, the euro.

220

J.M. Colomer

From the perspective developed here, further enlargements of the EU may be challenged not only by factors traditionally considered as being relevant, such as major differences between countries in economic structures or in public policy, but also by a low degree of fitness of new members’ party systems. New candidates to become member states of the EU could facilitate the governance of a highly pluralistic Europe by adapting their political representation to the patterns already developed within the EU institutions, especially in regard to the configuration of European Political Groups. The analysis here presented shows that Europe-wide political party building, ideologically consistent coalition formation and an ever-closer Union advance together and mutually reinforce each other.

Notes 1 I am very grateful to the Patronat Català Pro-Europa and MEP Carles Gasòliba for inviting me to visit the European Parliament in Brussels, as well as to Bo Manderup Jensen, Secretary General of the ELDR Group, Frank A. Wall, Council’s Director of Parliamentary and Institutional Affairs, and Jaume Duch, from the Directorate General for the Press, for information and unpublished data. Previous drafts of this contribution were presented at conferences of the NEMEU project in Prague and the American Political Science Association in Atlanta, both in 1999; I acknowledge help and useful suggestions received from the organisers and participants, especially Madeleine Hosli, Bernard Steunenberg, and John Smith. 2 See North (1990), for the basic concepts here applied. 3 See Axtmann (2001) for a comparison between early modern empires and the EU. 4 Including, for instance, the ‘minister’ in charge of the Common Foreign and Security Policy, now depending on the Council. 5 According to the formula of Laakso and Taagepera (1979) as given in Table 11.1. 6 Except for a small adjustment at the time of the German reunification, since it created a larger member state. 7 The numbers are from six to ten and then to seven if the non-attached and independents’ groups are not counted. 8 For the European Parliament and the European party systems, see Gaffney (1996), Morgan and Tame (1996), Hix and Lord (1997), Raunio (1997), Corbett et al. (2000). For a comparative analysis of different national party-system ‘configurations’, see Laver and Hunt (1992). In contrast to the ‘German’ or ‘Benelux’ party-system configuration alluded here, both the ‘Anglo’ and the ‘Mediterranean’ configurations distinguished by the authors involve higher degrees of bipolarisation and are characterised by more specific party ideological positions. 9 With the exception of Ireland in 1984 when the Irish Labour Party failed to gain representation. 10 For more details on the approach followed in this article, see Colomer (1996a, 1996b, 2000), Colomer and Hosli (1997). Power indices and their appropriateness for the analysis of the EU institutions are discussed in Brams (1975, 1985: 101–7; 1990), Hosli (1995), Bruckner and Peters (1996), Garrett and Tsebelis (1996), Nurmi (1997), and Felsenthal and Machover (1998). 11 For the analysis of coalition formation in European national parliaments with parliamentary regimes, see Laver and Schofield (1990), Laver and Budge (1992), and Laver and Shepsle (1996). See Deheza (1997, 1998) for the analysis of coalition formation in Latin American assemblies with presidential regimes. 12 See also Laver (1997), Kreppel (1999), Hix (1999: 79–84), and Corbett et al. (2000: 90–3). Well-established party positions on the left–right dimension for domestic

How political parties are building the EU

13

14 15 16 17 18 19 20

221

parliaments can be found in Castles and Mair (1984), Huber and Inglehart (1995), Knutsen (1998). A compilation is found in the Appendix to Laver and Schofield (1990). For the relative positions of parties in the European Parliament after the 1994 election, see Guide to the European Parliament 1994–99, Brussels: EU Committee of the American Chamber of Commerce; for the 1990 election, see The European Parliament, Luxemburg: Publications of the European Communities. See Kreppel and Tsebelis (1996), Hix and Lord (1997: 137–9), and Kreppel (1999). Author’s calculations with data from Secretariat General ELDR (2000). See Attinà (1990), Bowler et al. (1999) and Corbett et al. (2000: 89–90). For more details on calculation procedures and an illustration with the 1994 European Parliament, see Colomer and Hosli (1997). Raunio (1997), and author’s recalculations for individual members average from Hix and Lord (1997: 135), and from unpublished data in Secretariat General ELDR (2000). Author’s own calculation with data from European Parliament (1994, 1997, 1998). From data in European Parliament (1996) and press.

References Attinà, F. (1990) ‘The voting behavior of the European Parliament Members and the problem of Europarties’, European Journal of Political Research 8: 557–79. Axtmann, R. (2001) ‘State formation, multi-level governance and supranationalism in Europe: the case of the Holy Roman Empire of the German Nation’, Paper given at the First General Conference of the European Consortium for Political Research, Canterbury, 6–10 September. Balinski, M.L. and S.T. Rachev (1997) ‘Rounding proportions: methods of rounding’, Mathematical Scientist 22: 1–26. Bowler, S., D.M. Farrell and R. Katz (1999) Party Discipline and Parliamentary Government, Columbus: Ohio State University Press. Brams, S.J. (1975) Game Theory and Politics, New York: Free Press. Brams, S.J. (1985) Rational Politics: Decisions, Games, and Strategy, Washington: Congressional Quarterly Press. Brams, S.J. (1990) Negotiation Games: Applying Game Theory to Bargaining and Arbitration, New York: Routledge. Bruckner, M. and T. Peters (1996) ‘Research note: further evidence on EU voting power’, Journal of Theoretical Politics 8: 317–21. Castles, F. and P. Mair (1984) ‘Left-Right political scales’, European Journal of Political Research 12: 83–8. Colomer, J.M. (1996a) ‘Measuring parliamentary deviation’, European Journal of Political Research 30: 87–101. Colomer, J.M. (1996b) Political Institutions in Europe, London and New York: Routledge. Colomer, J.M. (2000) Political Institutions: Social Choice and Democracy, Oxford: Oxford University Press. Colomer, J.M. and M.O. Hosli (1997) ‘Decision-making in the European Union: the power of political parties’, Aussenwirtschaft 52: 255–80. Corbett, R., F. Jacobs and M. Shackleton (2000) The European Parliament, London: John Harper. Deheza, G.I. (1997) ‘Coalition cabinets in presidential systems: South America’, Doctoral dissertation, Florence: European University Institute. Deheza, G.I. (1998) ‘Gobiernos de coalición en el sistema presidencial: América del Sur’,

222

J.M. Colomer

in D. Nohlen and M. Fernández (eds) El presidencialismo renovado: Institucionalismo y cambio político en América Latina, Caracas: Nueva Sociedad, 151–69. European Parliament (1994) ‘Les avis législatifs du Parlément Européen et leur impact’, Luxembourg: European Parliament, Direction General des Études European Parliament. European Parliament (1996) ‘Report on the budgetary procedure’, Luxembourg: European Parliament. European Parliament (1997) ‘Progress Report 1 August 1996 to 31 July 1997’, Luxembourg: European Parliament, Conciliation Committee. European Parliament (1998) ‘Progress Report 1 August 1997 to 31 July 1998’, Luxembourg: European Parliament, Conciliation Committee. Felsenthal, D.S. and M. Machover (1998) The Measurement of Voting Power: Theory and Practice, Problems and Paradoxes, Cheltenham: Edward Elgar. Gaffney, J. (ed) (1996) Political Parties and the European Union, London and New York: Routledge. Garrett, G. and G. Tsebelis (1996) ‘An institutional critique of intergovernmentalism’, International Organization 50: 269–99. Hix, S. (1999) The Political System of the European Union, London: Macmillan. Hix, S. and C. Lord (1997) Political Parties in the European Parliament, London: Macmillan. Hosli, M.O. (1995) ‘The balance between small and large: effects of double-majority system on voting power in the European Union’, International Studies Quarterly 39: 351–70. Huber, J.D. and R. Inglehart (1995) ‘Expert interpretations of party space and party locations in 42 societies’, Party Politics 1: 73–111. Knutsen, O. (1998) ‘Expert judgements of the left-right location of political parties: a comparative longitudinal study’, West European Politics 21: 63–94. Kreppel, A. (1999) ‘Rules, ideology and coalition formation in the European Parliament: past, present and future’, Paper to the Annual Meeting of the American Political Science Association, Atlanta, Georgia. Kreppel, A. and G. Tsebelis (1996) ‘Coalition formation in the European Parliament’, Paper to the Annual Meeting of the American Political Science Association, San Francisco, California. Laakso, M. and R. Taagepera (1979) ‘Effective number of parties: a measure with application to West Europe’, Comparative Political Studies 12: 3–21. Laver, M. (1997) ‘Government Formation in the European Parliament’, Aussenwirtschaft 52: 223–48. Laver, M. and I. Budge (eds) (1992) Party Policy and Government Coalition, London: Macmillan. Laver, M. and W.B. Hunt (1992) Policy and Party Competition, London and New York: Routledge. Laver, M. and N. Schofield (1990) Multiparty Government: The Politics of Coalition in Europe, Oxford: Oxford University Press. Laver, M. and K.A. Shepsle (1996) Making and Breaking Governments: Cabinets and Legislatures in Parliamentary Democracies, Cambridge: Cambridge University Press. Marsh, M. (1998) ‘Testing the second-order election model after four European elections’, British Journal of Political Science 28: 591–607. Morgan, R. and C. Tame (eds) (1996) Parliaments and Parties: The European Parliament and the Political Life in Europe, London: Macmillan. North, D.C. (1990) Institutions, Institutional Change, and Economic Performance, Cambridge: Cambridge University Press. Nurmi, H. (1997) ‘The representation of votes groups in the European Parliament: a Penrose-Banzhaf Index analysis’, Electoral Studies 16: 317–27.

How political parties are building the EU

223

Raunio, T. (1996) Party Group Behaviour in the European Parliament, Tampere: Acta Universitatis Tamperensis, A 521. Raunio, T. (1997) The European Perspective: Transnational Party Groups in the 1989–1994 European Parliament, Ashgate: Brookfield. Reif, K. (ed.) (1985) Ten European Elections, Aldershot: Gower. Secretariat General European Liberal, Democratic and Reformist Group (ELDR) (2000) Internal Report, Brussels: European Parliament, Manuscript. Tsebelis, G. and G. Garrett (1999) ‘Legislative politics in the European Union’, Los Angeles: UCLA, unpublished manuscript. Tsebelis, G., C.B. Jensen, A. Kalandrakis and A. Kreppel (1999) ‘Legislative procedures in the European Union: an empirical analysis’, Los Angeles: UCLA, unpublished manuscript.

224

Authors name

12 A federal European constitution for an enlarged European Union Insights from constitutional economics Friedrich Schneider

Introduction After successfully completing the European Economic and Monetary Union we are currently observing two processes: a process of reforming the financing of the European Union budget (Agenda 2000) in order to secure future financial position of the EU and a first process of negotiations between the EU and some transition countries (Poland, Czech Republic, Hungary, Slovenia) about EU membership of these countries. Imagine a European Union with twenty or more members without a sound financial system and without some basic constitutional reforms (ending in a European federal constitution) – this would lead to a situation where the advantages of the EU are fewer than the disadvantages, with the consequence of destroying the EU. In order to avoid a major crisis of the functioning of the EU and in order to give the transition countries a real chance to become EU members in a foreseeable time, some (minimal) European federal union will be necessary. In this chapter some basic elements of a federal European constitution, like subsidiarity, federalism and direct democratic institution, are elaborated with the help of constitutional economics. In the first section six basic elements of a European constitution are introduced and in subsequent sections an attempt is made to scientifically justify these propositions, like the design of European legislation, the subsidiarity and federalism principle, and direct democratic institutions. The final section presents my conclusions.

Six basic elements of a future European federal constitution The completion of European Economic and Monetary Union provides the opportunity to achieve a number of efficiency gains, and it also provides the opportunity to stimulate the growth rates of western European economies. Furthermore, competition is promoted between member states by having a single currency and by weakening state-owned monopolies (like power plants and telecommunication systems). However, there is also the danger that these positive influences would be weakened if national regulations were replaced by EU regulations, and even more important if a ‘new federal’ government at the EU level is ‘created’ without

A federal European constitution

225

operating in a carefully designed institutional framework and with the consequence that it might grow and take over responsibilities from EU member states, which they do not want.1 But then there should be careful considerations when creating such a federal European constitution in order to prevent or limit a growing share of European federal government expenditures and the influence of interest groups. Buchanan (1990: 1) has drawn attention to the fact that Europe is now presented with a historically unique opportunity... The (constitutional) contract must be such as to ensure mutual gains from trade... The only constitutional structure that is consistent with the historically constrained setting of the 1990s is that of a federal union. Buchanan (1990: 10) also stresses that ‘a central political authority must come into being with some sovereignty over citizens in all of the nation-states’. Following Buchanan, it seems advisable to support the introduction of a federal European constitution. It has already been stressed in Schneider (1993, 1996) that democratic systems with market economies, if unchecked, show a strong tendency towards increasing state activity and interest group influence.2 As a consequence, the motivation of individuals to work efficiently, to engage in risky productive activities and to innovate is weakening. Whereas the removal of intra-European barriers to the movement of people, goods, capital, and services might weaken the influence of special interest groups and bureaucracies in EU member states, a growth at the federal European level has to be expected as soon as Europe-wide interest groups and parties have been fully established. A European constitution thus has to contain provisions to counterbalance such tendencies. Also such a constitution is necessary in order to guarantee a proper and efficient enlargement of the European Union, which will draw social attention to the Union’s institutional framework. In particular, it is feared that a further enlargement will dilute the Union’s legislative activities and in the current political debate it is clearly argued that present EU members have in fact made further enlargements dependent on the European Union’s ability to reform its institutional framework in order to maintain efficiency as well as the support of the European voters.3 It is obvious that a European Federal constitution is needed, so that the widening of the EU can proceed. The following elements should be an essential part of such a constitution:4 1

The European Commission should be turned into a European government with strictly limited tasks (for instance, the ones set up in point (2)), the Council of Prime Ministers into a second chamber (a newly formed ‘European Council’) where each country has the same weight of voting power. The simple majority approval of both chambers (the European Parliament and the ‘European Council’) is necessary for any legislation passed. Obviously, the European Parliament and the second chamber should solely have the full authority and responsibility for all European budgetary and federal items. If

226

2

3

4

5

6

F. Schneider

the two chambers cannot agree on a legislative or budgetary item, the Parliament can overrule the decision of the second chamber by a qualified (for instance, 2/3) majority. The jurisdiction of the European federal government should consist of defence, foreign policy, foreign trade policy, the enforcement of free intracommunity movement (of people, goods, services and capital), anti-cartel and anti-monopoly policy and environmental policy concerning communitywide environmental problems. All these policy issues should only be taken over by the European government if there is consensus between member states that the highest federal unit should do it, and if a referendum over these issues is approved by simple majority of the European voters and by simple majority of the member states. For the federal European government it should not be possible to run or accumulate deficits on its (current) budget over a legislative period. If a budget deficit still occurs at the end of a legislative period, either expenditures should be cut or revenues should be increased, given that the political conditions for a tax rate increase are fulfilled (compare elements (4) and (5)) so that the budget will be balanced again. Longer (than a legislative period) lasting public debt at the European federal level should only be allowed for financing investment expenditures and only if the federal government has the financial capacity to pay the interest and amortization of the debt out of its current budget.5 The activities of the Community should be financed by one specifically labelled tax, like a proportional (indirect) tax. Increases of the rate of this tax should be subject to 2/3 majority of the European Parliament and of the second chamber, and to the approval of a popular referendum. The institution of a popular referendum should be introduced for major policy issues (like a change of the European constitution, change of tax rate, etc.). Furthermore, a popular referendum should be held if a certain number of voters ask for it and if at least a certain percentage of all people entitled to vote participate.6 The issue over which the referendum is held is only accepted if it is approved by simple majority of the European voters and by simple majority of the member states. EU member states should have the right to secede from the European federal union. A member state should, by qualified (2/3 or 3/4) majority vote of their population, be able to leave the European federal constitution to become an independent state once again. However, the political and all other procedures for the secession should be precisely fixed in advance and a transition period of a considerable length of time (e.g. five or ten years) should be allowed. If a EU member state fails to reach such a qualified majority the next attempt should be possible after twenty years.

The design of the European legislation As the first elements in the previous section propose the formation of a European government with a two-chamber system, the first chamber being the European

A federal European constitution

227

Parliament and the second chamber being the Council of Prime Ministers, and obviously the control of both chambers over all federal items, which are delegated to the European Union, some general remarks will be made to justify this proposition. It is difficult to find studies analysing the separation of powers from the perspective of positive constitutional economics. In a review article Posner (1987) argues that the separation of powers increases the transaction costs of governing. This would hold for welfare-enhancing as well as for redistributive or even exploitative measures. The concept of separation of powers can be classified into the horizontal separation (legislature, executive, judiciary) and vertical separation (federalism). The structure of isolated powers can vary to a considerable extent. Some progress has been made in analysing the effects of separation of powers and political accountability. For example, Persson et al. (1997) show in a formal principal-agent model that separation of powers improves the accountability of elected officials, and thereby the utility of voters, but only under appropriate checks and balances. Two central provisions are needed: there must be a conflict of interest between the executive and the legislative. Moreover it must be impossible to implement any policy unilaterally, i.e. without the consent of both bodies. The basis for these results lies in the modelling of real-world political constitutions as ‘incomplete contracts’: Elected politicians are not offered an explicit incentive scheme associating well defined payoffs with actions in all states of the world. Political constitutions only specify who has the right to make decisions, and according to which procedures for which circumstances. This makes it hard to tie specific rewards or sanctions to the contents of those decisions. (Persson et al. 1997: 5) The application of these results to the European Union makes it necessary that institutions are created which leave only limited leeway to European politicians for selfish actions. The various effects of unicameral and bicameral legislators were first analysed from a public choice perspective by Buchanan and Tullock in their famous book Calculus of Consent in 1962. One of their major conclusions in their analytical framework is that the optimal decision rule is the one that leads to a minimum of the sum of external and decision costs (interdependence costs). Buchanan and Tullock (1962: 235) conclude that: . . . in comparison with unicameral systems, bicameral systems have higher decision costs . . . On the other hand, if the basis of representation can be made significantly different in the two houses, the institutions of bicameral legislature may prove to be an effective means of securing a substantial reduction in expected external cost of collective action without incurring as much added decision-making cost as a more inclusive rule would involve in a single house. The larger the majority required to reach a certain decision, the lower the external costs connected with that decision, because the number of opponents to a decision

228

F. Schneider

is negatively correlated with the required majority. On the other hand, it will become increasingly difficult to reach a decision at all, because the decision costs are positively correlated with a required majority. In a more recent study Levmore (1992) investigated the advantages and disadvantages of a bicameral versus a unicameral system. He concludes that a bicameral system might be better suited than a corresponding qualified majority in a unicameral system, to reduce the power of the agenda setter (mostly the government). Bicameral systems are often interpreted as a ‘brake’ against overly active legislators. Summarising the effects of bicameral systems, one could conclude that the legislative activities in a bicameral system are indeed lower than in a unicameral one and this should be reflected in a lower government consumption of economic output and in higher growth rates.7 Some papers in constitutional economics (see, for instance, Moser and Schneider 1997) try to give an analysis of the consequences of a change in the procedure on the power of the European government organs.8 Within the European Union, the strengthening of the European Parliament can be attributed as an introduction of a second force in addition to the second chamber. A bicameral system is also demanded in the first element, since it reduces the capability of rent seeking, because it is much more difficult to get a majority in both chambers than in only one.9 This is especially important if the European Union is widened, because the more member states the European Union has, the more likely rent seeking might occur. The draft report of the European Constitutional Group (1993) stresses the importance of competition as a mechanism to best fulfil consumer preferences. Competition, however, is not only crucial for the working of economic markets; it is also needed in political markets. This concept of institutional competition has a long tradition (starting with Tiebout’s ‘voting with the feet’). A better implementation of the distribution of powers, turning the Commission into a European government, the Council of Prime Ministers into a second chamber and strengthening the European Parliament (which means giving these two chambers full legislative power), can be seen as a first step in applying the democratic principle to a European Union. This can be one way to reduce the political inefficiencies which are normally discussed when investigating the democratic deficit of the European Union.

The subsidiarity and federalism principles as safeguards against government growth The subsidiarity principle One key element of a European federal union is the tasks of the European federal government, which have to be carefully defined. This basic proposition comes from the idea of using the subsidiarity principle, which is in substance a constitutional norm. Vanberg (1994) argues that this norm is meant to provide a criterion for what can be considered as a desirable constitutional order, a criterion that concerns an allocation of political authority in a multilayer system of government. To put it simply, it requires that in a multilevel policy the distribution of

A federal European constitution

229

power should be in favour of lower level governments, and hence smaller jurisdictions. In other words, it demands that the political authority be always located at the lowest possible level that is as close as possible to the citizens, the ultimate sovereign. Again, the consistent use of this principle is a necessary prerequisite for widening the European Union in a productive way for all member states. In the Commission report on the adaptation of existing legislation to the subsidiarity principle (European Commission 1993: 545) one reads ‘. . . the aim of the subsidiarity principle is, to see to it that decisions are taken as close as possible to the citizen, a constant watch being kept to ensure that action taking at community level is justified in the light of the means available to national, regional or local authorities’. Of course, the phrase ‘as close as possible’ is in urgent need of interpretation if the subsidiarity principle is to have specific normative content. Of course the constitutional norm to allocate political authority in favour of more local levels of governments is in itself not a very operational instruction for the design of constitutional frameworks, and the question of how the general principle is to translate it into more specific constitutional provisions is by no means a trivial issue. Judgements on the preferability of particular constitutional arrangements (for instance, using the subsidiarity principle in a strict way) over others always refer to some party to whom these alternatives are claimed to be preferable. In other words, all such judgements are directed to some addressee to whose interests they appeal. In democratic systems the ultimate addressees of constitutional proposals are, of course, the citizens who constitute this union. If the subsidiarity principle is claimed to be a desirable constitutional norm for the European Union, this means that such claims must be supported by arguments that can convince its citizens that it would be in their interest if efforts in the constitutional construction were guided by this principle. More precisely, these citizens would have to be convinced that adopting this principle would be in their constitutional interest, i.e. the interests that would form their choice if it were up to them to select the rules for the polity in which they live. What kind of arguments could one put forward in support of the subsidiarity principle as a constitutional norm? In other words, what kind of arguments could be made in favour of this principle, when designing a federal European constitution? One major argument for this principle is the central concern on part of the members of any democratic organisation about the principal-agent problem, that is the issue of how one can ensure that power delegated to agents can, on the one hand, be used to the benefit of the principals and, on the other, be prevented from being used against the principals’ interests. As far as democratic politics is concerned there is a long tradition of inquiry, in political economy as well as in other social sciences, into the advantages of decentralisation in political organisations. The results of this inquiry are of direct relevance to the subsidiarity principle. However, it is obvious that the subsidiarity principle alone neither constitutes a basis to regulate the intergovernmental relations in a future widened European Union, nor does it protect the collectivities at the grass roots (Feld and Kirchgässner 1996: 4).10 Moreover, the authors argue that the introduction of the subsidiarity

230

F. Schneider

principle in the Maastricht Treaty has in fact shifted the burden of proof at least somewhat more toward the centralists; the notion of subsidiarity nevertheless remains very general and open to many interpretations. Hence, the use of the subsidiarity principle does not solve the dynamic organisational problem, under which conditions, competencies or ‘rights’ should be given to lower governmental units. From a public choice point of view, there is a need for constitutional rules, which might prevent the ‘misuse’ of instruments by politicians, bureaucrats and interest groups. Therefore the subsidiarity principle must be ‘filled with life’ and the theory of federalism may represent an operational means to regulate the horizontal and vertical relationship between governmental units in the light of a potential Leviathan. Fiscal federalism in a European constitution Federalism is a crucial institution that serves to establish competition within the political arena. Costs rise for the voters as taxpayers if certain groups are able to appropriate the benefits of a publicly supplied good, but do not have to pay the price for it. This group can be the politicians and/or the bureaucrats, who are self-interested rent seekers, or special interest groups, who try to attain their selfish goals. Although it is not argued here that politicians and bureaucrats always seek to maximise their own utility up to the extent of actively exploiting the citizens and taxpayers, politicians and bureaucrats will do it from time to time, if they have the opportunity. Thus federal competition helps to save as a ‘safeguard’ against decision-makers taking unfair advantage of their discretionary power. Federal competition and federal institutions might also be a very crucial argument in a future European constitution. As has been discussed, the highest federal unit in the European Union should only be given those tasks that bring additional benefits (for instance, due to EU-wide spillovers) to voters/citizens, if they are fulfilled by the highest federal unit, such as foreign defence policy and environmental policy. The restriction of these tasks is necessary so that a more or less automatic centralisation of tasks (especially in the area of redistribution) at the highest federal level will be avoided. All other tasks should be provided by the EU member states (at what level within the EU member states is not discussed here).11 The operating principles of a lively federalism can be summarised as follows:12 as already argued, the European federal government will be constitutionally restricted in its domain of action – severely so. Within its assigned sphere, however, the central government would be strong; sufficiently so to allow it to enforce economic freedom or openness over the whole of the EU territory. The EU member states would be prevented by the federal European authority from placing restrictions on the free flow of resources and goods across their borders. In order to guarantee, additional to the constitutional arrangements already suggested, that the central power does not take over either fiscal or other items from the EU member states, Buchanan (1995) suggests the exit option in the following way: the EU member states must be constitutionally empowered to secede from the federal European Union. The secession, or the threat thereof, represents the only means through which the

A federal European constitution

231

ultimate powers of the European federal government might be held in balance. In the absence of the secession prospect, the federal European government may, by overstepping its constitutionally assigned limits, extract surplus value from the citizenry almost at will, because there would exist no effective means of escape.13 With an operative secession threat on the part of the EU member states, the European federal government could be held roughly to its assigned constitutional limits, while the EU member states could be left to compete among themselves in their capacities to meet the demand of citizens for collectively provided goods and services.14 Considering the arguments about federalism and subsidiarity, which policies should now be allocated at the European federal level? As proposed in reference to the jurisdiction of the federal government, the European federal government should be responsible for competition policy, defence and environmental policy, all those areas where one can expect EU-wide spillovers, so that the single EU citizen profits from them. The rationale for EU-wide environmental policy is first of all given by the global nature of some of today’s environmental problems: for instance, the thinning of the ozone layer and the danger of global warming by increased CO2 pollution.

The tax base of the European government In Kirchgässner (1994) and Schneider (1993, 1996) it is argued that the activities of the European government should only be financed by proportional (indirect) taxes. The rationale behind this idea lies in the different control possibilities (coming from the public choice literature) which exist on different governmental levels. First, any government will act more in accordance with the preference of the individuals/voters, the more the citizens are able to control it. At the lower levels, with smaller communities, the citizens have better opportunities to force the government to act according to their preferences; hence, this is just another argument to assign government activities to the lowest possible level, again using the subsidiarity principle in a very strict way. Second, as Feld and Kirchgässner (1996: 23) argue, it implies that tasks as well as financial means that are easier to control are more suited to a higher governmental level than those that are difficult to control. The proposed indirect tax can only be changed by changing a law, which means that any relative increase of the government share has to be decided by means of due parliamentary process and also via referendum, as suggested here. This ensures a public discussion, and at least as long as the European government seeks re-election it will hesitate to increase this tax, and it might face difficulties in getting an approval by the parliament and the voters. Such proportional taxes leave relatively little room for Leviathan behaviour by a European federal government, especially if an increase in the tax rate has to be subject to a two-thirds majority of both chambers in the European Parliament and the approval of a popular referendum. Therefore, at the European top federal level only the revenues from this indirect tax should be available.

232

F. Schneider

Institutions of direct democracy in a future European constitution Beside the important issues of federalism and subsidiarity, institutions of direct democracy like popular initiatives and (obligatory) referenda could also be a crucial factor in a future European constitution. They should be seen as a necessary supplement for the institutions of the representative democracy such as parliament (here the proposed two chamber system) and the government. There is a second crucial institutional feature, when introducing institutions of direct democracy. Referenda do not simply consist of a choice between given alternatives, but should also be seen as a quite important ‘political education’ process over time. According to Frey (1994) and Frey and Bohnet (1994a, 1994b) three stages can be differentiated. The first stage is the pre-referendum one, in which the possibility of undertaking a referendum encourages discussion both among citizens and between politicians and voters. Pre-referendum discussion produces a number of important effects. Preferences are articulated, enabling mutually beneficial bargaining and exchange to occur. Moreover, the agenda of alternatives is to a great extent determined by the citizens, thus constituting the relevant decision space. The pre-referendum stage screens the alternatives to be voted upon, reduces the number of relevant alternatives (quite often to only two) and makes the preferences somewhat more homogeneous, thereby lowering the chance that the preference aggregation paradox will occur.15 The second stage is the formal decision situation, in which it can be seen that voters clearly express their content or discontent with a proposed referendum and quite often give a government a clear task to do. At the third, the post-referendum stage on the one hand, as just argued, the government has a clear task to do, and on the other hand, quite often, initiators of a referendum force the government to change their policy merely by threatening to bring an issue into a popular referendum. But in some cases the government can also undertake unpopular measures (like tax increases), if they are supported in a popular referendum.16 The institutions of direct democracies also have other important means, such as their possible use by the voters to break politicians’ cartels directed against them. As Frey and Bohnet (1994: 151) proposed, rent-seeking theory argues that representatives have a common interest in forming a cartel to protect and possibly extend political rents.17 Referenda and initiatives can be means to break the politicians’ coalition against voters. Initiatives require a certain number of signatures and if the initiators obtain these signatures they can force the government to undertake a referendum on a given (mostly disputed) issue. They are a particularly important institution, because they take the agenda-setting monopoly away from the politicians and enable outsiders to propose issues for democratic decision, including those that many elected officials might have preferred to exclude from the agenda. As has been demonstrated in public choice theory the group determining which propositions are voted on, and in what order, has a considerable advantage, because it decides to a large extent the issues that will be discussed and which ones will be left out.18 Referenda, whether obligatory or optional, enable the voters to state their preferences to the politicians more

A federal European constitution

233

effectively than in a representative democracy. In a representative system, deviating preferences with respect to specific issues can only be expressed by informal protests, which are difficult to organise and to make politically relevant. Accumulating research on the properties of a popular referendum has revealed two major aspects on which institutional economics has to focus. The first element is the importance of discussion in the pre-referendum state (Frey 1994: 341). This implies that the number of propositions and the frequency of ballots must be low enough that the voters have an incentive and the opportunity to collect and digest the respective information in order to participate actively in the decision. Second, direct democratic institutions enable voters to break politicians’ and parties’ coalitions directed against them. Direct participation serves to keep the ultimate agenda-setting power with the voters. Initiatives and referenda are effective means by which the voters might regain some control over the politicians. Hence, introduction of direct democratic institutions like the referendum at the highest constitutional level in European constitution is an absolute necessity, especially if the European federal government wants to change the tax structure or introduce new taxes or wants to run a deficit or wants to take over new policy fields.19 This can only be implemented if it is approved by the legislation of the two chambers and by a popular referendum and if it is approved by a majority of the states. Hence, the introduction of direct democratic elements would be crucial for a future European constitution so that the European government keeps strictly to its given tasks.

Conclusions In this contribution an attempt has been made to provide some basic elements of a future European federal constitution in order to make possible a successful enlargement of the European Union. Six basic elements have been put forward, namely that the European Commission should be turned into the European government, the Council of Ministers into a second chamber, and that the European Parliament should get full control and responsibility over all federal items together with the chamber. The jurisdiction of the European federal government should consist of a few specific items, which are best suited for the highest federal level, like foreign defence and environmental policy. The activities of the federal European government should be financed by one proportional (indirect) tax, and direct democratic institutions should be introduced in a European federal constitution, like the possibility of forcing the European government to set up a referendum. These elements are then justified by arguments dealing with the subsidiarity principle, the idea of federalism and with the effects of direct democratic institutions. As has been demonstrated, these elements are best suited to limit the domain of central European authority in the long run, even in the face of a strong tendency to centralisation in nearly all federal states that has been observed during recent decades. The idea is also proposed that the constitution should be structured in such a way that any attempt at a future concentration of government activities at the European federal level will be prevented by explicitly assigning specific govern-

234

F. Schneider

mental functions to each level of government and to put in additional safeguards, like the subsidiarity principle and the direct democratic institutions, so that the federal European government cannot take over tasks which are not approved by the majority of voters and European member states. From another perspective it is also very difficult for the European government to take over additional tasks, because the necessary widening of the tax base can only be done if a majority of European voters/taxpayers approve it and also a majority of the member states. Neither can the European government accumulate large deficits, which might hold as another safeguard to limit the domain of a future central authority in Europe. Two central features – direct democracy and fiscal federalism – should be key principles in a future European constitution, as these have proved in other federal units (like Switzerland and the United States) to be strong safeguards against policies that are not in line with a majority of voters’ preferences. Moreover, direct democratic elements provide European citizens in a federal state with the possibility of participating actively in political decision-making, to break politicians’ and interest groups’ cartels and prevent a shifting of responsibilities from EU member state levels to the EU federal government level. A proper assignment of the tax competencies may also help to restrain centralisation. If, as suggested here, the tax competencies lie within the EU member states – with the expectation of one specific indirect tax for the European government – such an element should also work. With respect to the actual political discussion of an enlargement of the European Union, such constitutional perspectives (for instance, the introduction of a two-chamber system, both chambers being fully responsible for all federal items, and the possibility of a referendum) would be very fruitful and important and might help to overcome the fears of a number of citizens. Since, according to Downs (1957), rational voters are rational ignorant, it is necessary to build the European Union in a way that at least some advantages of this Union can be perceived by the ordinary citizens as well, even if they are not well informed. Up to now, however, the large advantages of the European process for the citizens are very indirect and often not at all obvious, while the public discussion focuses on the interests of producer interest groups and the influence of bureaucracy in Brussels. Thus, today the political opinion of ordinary citizens as to the enlargement of the European Union varies between apathy and refusal; hence the public support for the widening of the EU is rather small in most EU member countries. If such a constitutional perspective could be provided in the actual political process, this could lead to a greater acceptance of the widening of the European Union by the ordinary voter and taxpayer.

Notes 1 Compare Alesina and Perotti (1994), Alesina and Rodrick (1992) and Kirchgässner and Pommerehne (1995). 2 Compare the studies by Olson (1982), Bernholz (1990a, 1990b), Schneider and Frey (1988), Peden and Bradley (1989), Weede (1986, 1990) and De Haan and Siermann (1995, 1996).

A federal European constitution

235

3 Necessary reforms of the EU are, for instance, discussed by König and Bräuninger in Chapter 5 of this volume. They concentrate on two aspects of institutional reform and functional as well as parliamentary integration. 4 For a recent discussion of constitutional issues, see Gwartney and Wagner (1988), Vanberg and Buchanan (1989), Schneider (1993), Kirchgässner (1994) and Holzmann (1996); for fiscal federalism, Oates (1985), Blöchliger and Frey (1992), and Eichenberger (1994). 5 Whether the European economic and monetary integration provides incentives to increase public deficits is investigated in Horstmann and Schneider (1994). Compare also Feld and Kirchgässner (1997). 6 The precise figures determining the conditions under which a referendum has to be held and is accepted have to be specified; the only important point is that the option exists to force the European government to hold a referendum over a certain issue. 7 A similar conclusion is reached and preliminary empirical evidence is given by the study of Feld and Kirchgässner (1996). For a more general view of this aspect, see Weingast (1993). 8 Compare Peters (1996a and 1996b) and Steunenberg (1994). 9 The introduction of a bicameral system not only reduces attempts at rent seeking but also is important to strengthen the federal structure of the European Union. The second chamber, in which the Prime Ministers of every EU member state have equal voting power, can be seen as an institution which solely represents the interests of the EU member countries, like the German Bundesrat or the Schweizer Ständerat. It might especially help to take care of the interests of the smaller EU member countries. 10 In this context it is not surprising that Delors (1991: 12) argues that ‘subsidiarity does not enact any restriction for the Commission to take political action’ and he continues to argue ‘solely on the basis of the Maastricht Treaty subsidiarity is not judiciable.’ See also Sachverständigenrat zur Begutachtung der gesamtwirtschaftlichen Entwicklung (1992), Vaubel (1993, 1995) and Möschel (1993, 1995). 11 One could go a step further and put forward the idea of federal competition between and within EU member states when providing goods and services, but also financing them. As has been shown in the extensive research for Switzerland by Kirchgässner and Pommerehne (1995) and Pommerehne et al. (1995) there is extensive tax competition in Switzerland, for instance, between the cantons at very small local distances. This tax competition did not result in a breakdown of public good supply in Switzerland and there is no indication of an underprovision of public goods. According to Feld and Kirchgässner (1996: 8), the income taxes in Switzerland vary quite a lot between the different cantons. Taking the value of the (weighted) average for Switzerland as 100, in 1990 the index of burden in the form of personal income and property taxes ranged from 56.1 in the canton of Zug to 154.1 in Valais. For instance, a family with two children that earns a gross income of SFr 175,000.00 had to pay SFr 16,083.00 in cantonal and local income taxes in Zug, but SFr 34,475.00 in Bern, two cities only about 100 km apart! Genser (1992) and Sinn (1990) present a critical view on the issue of tax competition. 12 Compare here the path-breaking study by Brennan and Buchanan (1980). 13 Compare here again the pioneering work of Buchanan and Faith (1987) and Buchanan (1990, 1995). 14 The (threat of) secession should be seen here as an ultimate ‘weapon’ for every EU member country in order to avoid a development that is not wanted by EU member states and their citizens. In principle it gives the single countries a power over (for them) critical issues and if it is used one might end up in a prisoners’ dilemma situation. But as the execution of the secession for a single EU country is not so easy (for instance, a qualified majority of voters in that country is necessary) it is unlikely that this instrument will be used frequently only for tactical reasons. In terms of game theory, the threat is unlikely to be credible. 15 See Frey and Bohnet (1994a, 1994b), and Frey (1994: 339).

236

F. Schneider

16 For instance, a referendum might help the government to undertake unpopular measures in environment policy, compare Frey and Schneider (1997). 17 The literature on rent seeking was developed by Tullock (1967); one of the latest surveys is by Tollison (1982). 18 Compare Denzau (1985) and Mueller (1987, 1989). 19 Such a conclusion is also reached by Bernholz (1990b), Feld and Kirchgässner (1996) and Vaubel (1993 and 1995).

References Alesina, A. and R. Perotti (1994) ‘The political economy of growth: a critical survey of the recent literature’, World Bank Economic Review, 8: 351–71. Alesina, A. and D. Rodrick (1992) ‘Distribution, political conflict, and economic growth: a simple theory and some empirical evidence’, in A. Cukierman, Z. Hercowitz and L. Leiderman (eds) Political Economy, Growth, and Business Cycles, Cambridge, MA: MIT Press, 23–50. Bernholz, P. (1990a) ‘The completion of the European market: opportunities and dangers from an institutional perspective’, in The Macroeconomics of 1992, Bruxelles: CEPS Paper No. 42. Bernholz, P. (1990b) ‘Grundzüge einer europäischen Verfassung: Ein Bundesstaat mit begrenzter Zentralgewalt?’, in Argumente zur Europapolitik, Frankfurt: Frankfurter Institut, 2–12. Bernholz, P. (1993) ‘Institutional aspects of European integration’, in S. Borner and H. Grubel (eds) EC after 1992: Perspective From the Outside, London: Macmillan Publishing Company, 241–66. Blöchliger, H.-J. and R.L. Frey (1992) ‘Der Schweizerische Föderalismus: Ein Modell für den institutionellen Aufbau der europäischen Union?’, Aussenwirtschaft 47: 515–48. Brennan, G. and J. Buchanan (1980) The Power to Tax: Analytical Foundations of a Fiscal Constitution, New York: Cambridge University Press. Buchanan, J.M. (1990) Europe’s Constitutional Opportunity, Fairfax: Virginia Center for Study of Public Choice, George Mason University. Buchanan, J.M. (1995) ‘Federalism as an ideal political order and an objective for constitutional reform’, Publius: The Journal of Federalism 25: 19–27. Buchanan, J. and R. Faith (1987) ‘Secession and the limits of taxation: toward a theory of internal exit’, American Economic Review 91: 1023–31. Buchanan, J.M. and G. Tullock (1962) The Calculus of Consent: Logical Foundations of Constitutional Democracy, Ann Arbor: University of Michigan Press. Delors, J. (1991) ‘The principle of subsidiarity: contribution to the debate’, Maastricht: European Institute of Public Administration, 7–18. Denzau, A.T. (1985) ‘Constitutional change and agenda control’, Carnegie Papers on Political Economy 47: 183–217. Downs, A. (1957) An Economic Theory of Democracy, London: Harper & Row. Eichenberger, R. (1994) ‘The benefits of federalism and the risk of overcentralization’, Kyklos 47: 403–20. European Commission (1993) Commission report to the European Council on the adaptation of existing legislation to the subsidiarity principle, Bruxelles. European Constitutional Group (1993) ‘A proposal for a European Constitution, report by the European Constitutional Group’, London: European Policy Forum. Feld, L.P. and G. Kirchgässner (1996) ‘Omne agens agendo perficitur: the economic meaning of subsidiarity’, in R. Holzmann (ed.) Maastricht: Monetary Constitution Without a Fiscal Constitution?, Baden-Baden: Nomos Publishing Company, 195–226.

A federal European constitution

237

Feld, L.P. and G. Kirchgässner (1997) ‘Public debt and budgetary procedures: top down or bottom up? Some evidence from Swiss municipalities’, St Gallen: Universität St Gallen, Volkswirtschaftliche Abteilung, Discussion Paper 97–17. Frey, B.S. (1994) ‘Direct democracy: politico-economic lessons from Swiss experience’, American Economic Association Papers and Proceedings 84: 338–42. Frey, B.S. and I. Bohnet (1994a) ‘The Swiss experience with referenda and federalism’, IDIOMA, Revue de linguistique et des traductology 10: 147–60. Frey, B.S. and I. Bohnet (1994b) ‘Direct democratic rules: the role of discussion’, Kyklos 47: 341–54. Frey, B.S. and F. Schneider (1997) ‘Warum wird die Umweltökonomik kaum angewendet?’, Zeitschrift für Umweltpolitik und Umweltrecht 20: 153–70. Genser, B. (1992) ‘Tax competition and harmonization in federal economies’, in H.-J. Vosgerau (ed.) European Integration in the World Economy, Heidelberg: Springer Publishing Company, 184–205. Gwartney, J.D. and R.E. Wagner (1988) Public Choice and Constitutional Economics, Greenwich, CT: J.A.I. Press. Haan, J. de, and C.L.J. Siermann (1995) ‘A sensitivity analysis of the impact of democracy on economic growth’, Empirical Economics 20: 197–215. Haan, J. de, and C.L.J. Siermann (1996) ‘New evidence on the relationship between democracy and economic growth’, Public Choice 86: 175–98. Holzmann, R. (ed.) (1996) Maastricht: Monetary Constitution Without a Fiscal Constitution?, Baden-Baden: Nomos Publishing Company. Horstmann, W. and F. Schneider (1994) ‘Deficits, bailout and free riders: fiscal elements of a European constitution’, Kyklos 47: 355–83. Kirchgässner, G. (1994) ‘Constitutional economics and its relevance for the evolution of rules’, Kyklos 47: 321–39. Kirchgässner, G. and W.W. Pommerehne (1995) ‘Public spending in federal states: a comparative econometric analysis’, in D. Meulders (ed.) Public Expenditure and Taxation: Recent Econometric Research, London: Routledge, 112–36. Levmore, S. (1992) ‘Bicameralism: when are two decisions better than one?’, International Review of Law and Economics 12: 145–62. Möschl, W. (1993) ‘Eine Verfassungskonzeption über die Europäische Union’, in H. Gröner and A. Schüller (eds) Die Europäische Integration als ordnungspolitische Aufgabe, Stuttgart: Verlag Gustav Fischer, 21–39. Möschl, W. (1995) ‘Subsidiaritätsprinzip im Zwielicht’, WiSt 5: 232–6. Moser, P. and G. Schneider (1997) ‘Rational choice and the governance structure of the European Union: an introduction’, Aussenwirtschaft 52: 64–82. Mueller, D.C. (1987) ‘The growth of government: a public choice perspective’, IMF-Staff Papers 34: 115–49. Mueller, D.C. (1989) Public Choice II, Cambridge: Cambridge University Press. Oates, W.E. (ed.) (1977) The Political Economy of Fiscal Federalism, Lexington, MA: Lexington Books. Oates, W.E. (1985) ‘Searching for Leviathan: an empirical study’, American Economic Review 75: 578–83. Olson, M. (1982) The Rise and Decline of Nations: Economic Growth, Stagnation and Rigidities, New Haven: Yale University Press. Olson, M. (1983) ‘The political economy of comparative growth rate’, in D.C. Mueller (ed.) The Political Economy of Growth, New Haven: Yale University Press, 222–47. Peden, E.A. and M. Bradley (1989) ‘Government size, productivity and economic growth: the post war experience’, Public Choice 61: 229–45.

238

F. Schneider

Persson, T., G. Roland and G. Tabellini (1997) ‘Separation of powers and political accountability’, Munich: Center for Economic Studies at the University of Munich, Working Paper No. 136. Peters, T. (1996a) ‘Decision-making after the EU-Intergovernmental Conference’, European Law Journal 2: 251–66. Peters, T. (1996b) ‘Voting power after the enlargement and options for decision making in the European Union’, Aussenwirtschaft 51: 223–43. Pommerehne, W.W., G. Kirchgässner and L.P. Feld (1995) ‘Tax harmonization and tax competition at the state-local levels: lessons from Switzerland’, in G. Pola, R. Levaggi and G. Francke (eds) New Issues in Local Government Finance: Theory and Policy, Aldershot: Edward Elgar, 201–55. Posner, R. (1987) ‘The constitution as an economic document’, George Washington Law Review 56: 4–38. Sachverständigenrat zur Begutachtung der gesamtwirtschaftlichen Entwicklung (1992) Für Wachstumsorientierung gegen lähmenden Verteilungsstreit, Jahresgutachten 1992/93, Stuttgart: Verlag Poeschl. Schneider, F. (1993) ‘The federal and fiscal structures of representative and direct democracies as models for a European federal union: some ideas using the public choice approach’, in European Economy Report and Studies: Volume 5, Brussels: European Commission, 191–212. Schneider, F. (1996) ‘The design of a minimal European federal union: some ideas using the public choice approach’, in J.C. Pardo and F. Schneider (eds) Current Issues in Public Choice, Cheltenham: Edward Elgar, 203–22. Schneider, F. and B.S. Frey (1988) ‘Politico-economic models of macro economic policy: the political economy of money, inflation and employment’, in W. Thomas (ed.) The Political Business Cycle, Durham and London: Duke University Press, 239–75. Sinn, H.-W. (1990) ‘Tax harmonization and tax competition in Europe’, European Economic Review 34: 489–504. Steunenberg, B. (1994) ‘Decision-making under different institutional arrangements: legislation by the European Community’, Journal of Institutional and Theoretical Economics 150: 642–63. Tollison, R.D. (1982) ‘Rent-seeking: a survey’, Kyklos 25: 575–602. Tullock, G. (1967) ‘The welfare costs of tariff, monopolies and theft’, Western Economic Journal 5: 224–32. Vanberg, V. (1994) ‘Subsidiarity, responsive government and individual liberty’, in K.W. Nörr and T. Oppernmann (eds) Subsidiarität: Idee und Wirklichkeit zur Reichweite eines Prinzips in Deutschland und Europa, Tübingen: Mohr, 253–69. Vanberg, V. and J.M. Buchanan (1989) ‘Interests and theories in constitutional choice’, Journal of Theoretical Politics 1: 49–62. Vaubel, R. (1993) ‘Perspektiven der europäischen Integration: die politische Ökonomie der Vertiefung und Erweiterung’, in H. Siebert (ed.) Die zweifache Integration: Deutschland und Europa, Tübingen: Mohr, 3–31. Vaubel, R. (1995) ‘Constitutional safeguards against centralization in federal states: an international cross-section analysis’, Mannheim: Universität Mannheim, Discussion Paper 532/95. Weede, E. (1986) ‘Catch-up distributional coalitions and government growth or decline in industrialized democracies’, British Journal of Sociology 37: 194–220. Weede, E. (1990) Wirtschaft, Staat und Gesellschaft, Tübingen: Mohr. Weingast, B. (1993) ‘Constitutions as governance structures: the political foundations of secure markets’, Journal of Institutional and Theoretical Economics 149: 286–311.

Chapter Title

Part V

Conclusion

239

240

Authors name

Chapter Title

241

13 Changing times The future of enlargement of the European Union and the theoretical challenges ahead Bernard Steunenberg and Antoaneta Dimitrova

Reforms and enlargement: perspectives and problems This book has addressed the relationship between internal institutional reform in the European Union and enlargement. Such reforms are important, as Dennis Mueller concludes, because as the Union increases its size, structural deficiencies are likely to prove more harmful to its ability to make decisions and to provide collective goods to its citizens. His concern is supported by the analysis by HansJörg Schmidt-Trenz and Dieter Schmidtchen, who show that the possible benefits of further extending a legal order, such as the acquis communautaire, depend on various factors and the extension is not necessarily a ‘better’ arrangement for existing and new members. Gerald Schneider shows how widening and deepening of the Union are connected in a non-linear way. He suggests that enlargement will most likely lead to the wish of some member states to loosen the collaboration, as some of the benefits of cooperation can no longer be obtained. Focusing on Council decision-making, Thomas König and Thomas Bräuninger, as well as Jan-Erik Lane and Reinert Mæland, urge reducing the threshold for majority voting to avoid the Union’s ability to take decisions being substantially limited by the member states’ capacity to block. Lars Feld, Gebhard Kirchgässner and Hannelore Weck-Hannemann suggest introducing referenda at the European Union level, especially on the Union’s budget. Jordi Bacaria, Georgios Chortareas and Andreas Kyriacou propose alternative arrangements for the European System of Central Banks, which would improve monetary policymaking in the euro-zone in view of further enlargement. Finally, Friedrich Schneider argues that a European constitution is necessary in order to guarantee ‘a proper and efficient enlargement’. These findings make it clear that even after Nice a further debate on the Union’s institutional structure is required. The link between institutional reform and enlargement is getting stronger as enlargement prospects are becoming a reality. As discussed in some of the chapters, the difficulties in reaching agreement on the necessary changes have also increased, threatening to delay enlargement further. In a speech to the European Parliament in November 2001, Commission President Prodi warned against postponing enlargement by making it conditional on further institutional and policy reforms:

242

B. Steunenberg and A. Dimitrova Further reforms cannot be made a pre-condition for enlargement, although the responsibilities which we are taking on with enlargement must be seen as providing the opportunity to begin preparing – with considerable momentum and political vision – the new institutional process due to be launched at the Laeken European Council. Similarly, the current negotiations should not be linked to the issue of the Union’s finances after 2006: these are separate issues. Any attempt to link them would mean changing the rules and would inevitably disrupt our carefully balanced negotiation process. (Prodi 2001)

Prodi’s words remind us that, having just completed one round of institutional negotiations at Nice, the European Union is trying to launch the next round of institutional reforms and prepare the 2004 Intergovernmental Conference. In contrast to the Nice negotiations, the process starting with the European Council held at 14–15 December 2001 in Laeken, Belgium, aims to address the broader questions of the ‘future of Europe’ and the nature of governance in Europe. The ‘future of Europe’ debate is related to the proposals for sweeping institutional changes leading to the establishment of a constitution for Europe, which received the greatest prominence in the speech made by the German Foreign Minister, Joschka Fisher, in May 2000 (Fischer 2000). The need for such changes is linked very strongly to enlargement by Fischer, who forcefully argued that with the pending enlargement, the usefulness of the EU institutional system in its present form is coming to an end. Similarly, the debate on governance in Europe, launched by the European Commission’s White Paper on Governance in October 2001, discusses changes in the way the institutions of the Union relate to each other and to the citizens in preparation for an enlarged Union (European Commission 2001). The process of change in the Union reflected in these debates is larger than the adaptations that may be needed for enlargement. Following Bernard Steunenberg’s argument in Chapter 6, it becomes more and more evident that enlargement has provided an opportunity to initiate reforms.

Enlargement and the institutional balance in the Union There are several ways in which enlargement provides an opening for actors to attempt to change the institutional balance of power in their favour. As Chapters 5, 6 and 7 in this book show, the Nice Treaty reform resulted in changes in the balance of power between large and small member states, changes negotiated in the shadow of the future and accepted as a means to open the way for Europe’s historic unification, as this enlargement is often called. Not only member states but also institutional actors are involved in attempts to increase their power or defend their position, in particular in the light of the constitutional reforms mentioned above. The Commission lobbied for the management of assistance for Central and Eastern Europe at the start of enlargement.1 It obviously saw some advantage in its extensive responsibility of managing the enlargement process on a day-to-day

Changing times

243

basis for over a decade. Giandomenico Majone (1996: 65) has suggested that the utility function of the Commission is positively related to the scope of its competences – a hypothesis which could account for the Commission’s exceptionally active role in this enlargement. Despite claims that preparing enlargement is a heavy responsibility, the Commission has had the chance to make gains in the institutional balance and vis-à-vis the member states in two ways. First, the Commission has the opportunity to exploit its information advantage in the relationship with the member states – a feature that is common to principalagent relationships (Pollack 1997). In the current enlargement process, the Commission’s role as a negotiator with the candidates from Central and Eastern Europe has allowed it to have access to a great amount of information regarding the post-communist states. This information was not available before 1989 and has not been always made readily available to the member states. Evidence of this is provided by the fact that the Commissioner responsible for enlargement in the previous Commission, Hans van den Broek, did not send to the member states the position papers of the candidate states in the first half of the 1990s.2 Later, as enlargement negotiations proper started in 1998, Commission officials described them as a purely technical, a frame that increased considerably the Commission’s discretion. Second, the Commission’s position in the enlargement process has allowed it to expand the scope of its competences without much scrutiny from the member states, as this expansion was mostly realised vis-à-vis the candidate states. It possesses certain competences in the candidate states, for instance in the context of the yearly progress report evaluations, which it does not have in the Union. Examples are the evaluation of administrative capacity, regional and administrative reform, monitoring the rights of minorities and other competences (Dimitrova 2001). An interesting question for further research is to what extent the Commission will be able to preserve these gains in the scope of its competences after the current candidates join. There is a possibility that it will be reined in by the member states as the period of technical decisions ends and the time for the final bargains of this enlargement approaches. Yet it is possible that the final stages of the accession negotiations may allow the Commission an opening to extend the scope of its powers by helping overcome potential vetoes on enlargement, similar to the increase of competences in regional policy accompanying the Greek accession described by Marc Smyrl (1998). The link between enlargement and internal reform, and the reform of governance in particular, provides another opening. In the 2001 White Paper on Governance, the Commission (2001: 25) has suggested that currently existing mechanisms used to monitor and enhance administrative capacity in the candidate states should be used after enlargement. This would constitute a transfer of some of the competences it has acquired in the enlargement process into the Union’s permanent institutional arrangements. Such potential gains for the Commission vis-à-vis the member states have to be evaluated in the light of the trend in the Union towards decision-making which does not use the Community method in which the Commission has

244

B. Steunenberg and A. Dimitrova

traditionally established its place. The Commission’s recent plea in its White Paper on Governance for refocusing the institutions and revitalising the Community method can be seen as a struggle for survival in the context of the increased dominance of intergovernmental decision-making procedures in CFSP and Justice and Home Affairs areas where the Commission does not have an exclusive right of initiative (European Commission 2001). The increased importance of areas where the EU member states take decisions without using the Community method, such as visas, asylum and judicial cooperation in criminal matters, presents a challenge to the European Parliament as well. The Parliament has been somewhat excluded from the enlargement negotiations and has even taken to producing its own reports on the candidates to compensate for the lack of information. Even with the Parliament’s pro-active and positive attitude to enlargement it is possible that it would use its assent powers to protest its exclusions from decision-making on the future of the Union, for example in the 2004 Intergovernmental Conference. The European Council has had a unique position during the enlargement process. The European Council has driven and directed the process through a series of landmark decisions, such as the decision to invite the candidates to join, taken in Copenhagen in 1993, and the decisions to start negotiations (Luxembourg 1997 and Helsinki 1999). In fact the European Council has had a particular institutional role in enlargement (see also Smyrl 1998). The Council of Ministers has been somewhat sidelined as a result, although it has a central role in the accession negotiations in which the Council has to give common positions, which the Commission negotiates, and later to give assent to the Accession treaties. On the whole, the enlargement negotiations seem to have been an interplay between the Council and the Commission. Even with the power of vetoing enlargement, the European Parliament may not be able to compensate for its long exclusion from the formal part of the process. However, it must be remembered that the institutional balance of power will be affected by the debate on the future of Europe, the constitutional convention and the debate around the Union’s Charter of Human Rights in which the Parliament plays a more active role than in enlargement.

Institutional reforms in the Union and the candidate states The importance of the debate on the future of Europe is recognised by the candidate states, which have argued that the legitimacy of the reforms the Union intends to introduce would be greatly diminished if they were made without taking into account the preferences of soon-to-be-members. A contribution from a group of experts from the current and future member states points out that ‘[u]ntil recently, politicians, experts and the general public in the EU-15 have focused on the consequences of enlargement without involving the views of the applicant states seriously. A next effort to reform the Union cannot be an exclusive exercise of the EU-15, but must involve also the accession countries’ (Villa Faber Group 2001: 1).

Changing times

245

The importance of taking into account the preferences of the future members is threefold. First, as they point out, any constitutional arrangements negotiated by the current EU member states would be more just if they are not specifically designed to suit current members but as if behind a veil of ignorance about future configurations (Villa Faber Group 2001). As König and Bräuninger show in Chapter 5, the institutional reforms reached at Nice reflect the interests of the current member states, especially the larger ones. Using a different method, Steunenberg reaches a similar conclusion in Chapter 6. Second, the potential destabilising effects which may occur as a result of the wholesale adoption of the acquis based on self-interested bargains between the existing EU member states would mostly affect the economies of the current candidates. Based on the East German post-unification experience, Fritz Scharpf (2001) has issued a similar warning about the potential impact of enlargement on the domestic political arena of future member states. The potential domestic implications of enlargement-driven reforms should be considered by the Union alongside with the effects on the current institutions and the future member states. Third, as Josep Colomer argues in Chapter 11, accounting for the future members’ political preferences and especially their ideological fit to the current system of political party groupings would facilitate the governance of a highly pluralistic Europe. The current member states would avoid the occurrence of sharp divisions between current and new members, or between ‘East’ and ‘West’, which could create new cleavages in European politics. Such cleavages are not very helpful in formulating European policy and could reduce the Union’s effectiveness. As the reform perspectives reviewed in this section show, the enormous challenges presented by the next rounds of institutional reform and enlargement are sufficient to generate many questions for social research. Some have been addressed in this book, such as the need for differentiation in the Union, the decision-making rules in the Council, the budget procedure, the European System of Central Banks and the need for a European constitution with elements of subsidiarity, federalism and direct democracy. Others, however, require further theoretical advances to enable us to explain institutional and above all constitutional change.

Analysing institutional change: the way ahead? Reforming the institutions of the European Union has been so far the most difficult part of the process of enlargement. Institutions have turned out to be far more stable than anticipated by most of the rational choice literature in which the main question has been why institutions are stable at all. This requires reformulating the basic research question: from asking how institutions affect decision-making to asking how they can be changed. This means that institutions should no longer be taken as a factor explaining the process and outcome of decision-making, but as the factor that needs to be explained. Analysing institutional change implies that we have to concentrate on a more extensive ‘game’ in which political actors not only decide on public policies, but

246

B. Steunenberg and A. Dimitrova

also select institutions that will be used to select public policies. At this point, it is helpful to distinguish between two different levels of choice. The first level is constitutional choice, which concerns decisions on the rules and procedures that regulate the process in which public policies have to be selected. The second level is called policy choice, which involves the choice of a public policy within the context of a predetermined set of institutions. Thus, constitutional choice precedes policy choice, which is made at a lower level (see also Mueller 1996). In this perspective, policy choice is part of a larger ‘game’, in which players not only decide on public policies, but also on the institutional arrangements under which public policies will be selected. Constitutional choice is mostly perceived as a ‘logical’ extension of the institutionalist program, in which institutions, like policies, are the result of equilibrium in the behavioural options to political actors.3 In other words, in view of their preferences, institutions are the best possible choice for each actor given the choices made by the others. The question arises whether institutions which are necessary in a process of collective choice that leads to policy, can be distinguished from this policy within an analysis that starts from preferences. If both policies and institutions are a result of preferences, how could institutions provide stability in the choice of policy? William Riker (1980), who first noted this problem, defines institutions as ‘relative’ constants in the analysis. He suggests that the costs of changing institutions might be higher than for preferences, so they can be treated as constants in the short run. ‘Thus the only difference between values and institutions is that the revelation of institutional disequilibrium is probably a longer process than the revelation of disequilibria of taste’ (Riker 1980: 445). The relative stability of institutions compared to preferences, which Riker explains in terms of costs, is often related to how well institutions perform certain functions. This view is widespread in institutional rational choice, as discussed by Keith Dowding in Chapter 2 of this book. Terry Moe (1990), for instance, regards institutions as primarily reducing uncertainty about future policy benefits, which allow politicians to secure the support of their constituencies. Albert Breton (1996) points out that when institutions facilitate an efficient exchange of interests in the ‘political’ market, political actors do not have an incentive to change them. Finally, Keith Krehbiel (1991) perceives institutions as means to collect and use information efficiently. Administrative law, political appointments and administrative procedures are, according to Krehbiel, ways to induce ‘bureaucratic compliance’, which allows politicians to utilise bureaucratic expertise. An important drawback of these functionalist ‘explanations’ of the relative stability of institutions is that they do not provide a clear reason why political actors have chosen some institutions in the first place (Thelen 1999). The fact that a herd of cattle allows a farmer to produce milk and cheese does not provide an answer to the question why this farmer would keep cattle and not sheep, or would not just grow corn or maize in the fields. The possible reasons for a higher stability of institutions are not necessarily the same as the reasons for the choice of institutions. If other considerations than an efficient use of information, an efficient reduction of uncertainty, or an efficient exchange of interests play a role,

Changing times

247

political actors will select other institutions than the ones optimally performing these functions. To advance insights in institutional change further, institutional rational choice has moved towards non-institutional factors. Discussing ways to approach institutional change, Kenneth Shepsle (2001: 325) suggests that ‘[i]n the end it comes down to information’, information about ‘how political games are played and what can be expected.’ In his view institutions are grounded in the beliefs of political actors, which are updated on the basis of experience. A change from one set of institutions to another could be triggered by a change of these beliefs so that all political actors that are involved would continue to make decisions using another set of rules or procedures. At the same time, the main emphasis of the analysis shifts away from institutions. Michael Gorges (2001: 141) indicates in a recent review that in order to address institutional change, institutionalist accounts heavily rely on exogenous, non-institutional factors, such as information (Shepsle 2001), general social-economic change (North 1990), leadership (Ostrom 1998), or ideas (Garrett and Weingast 1993). These non-institutional factors are introduced to explain the use of institutions and their potential change. Based on these recent developments, institutional rational choice seems to have reached the limits of institutional explanations. This opens up a new area of research in which the role of institutions in the process of political decisionmaking is analysed using non-institutional factors, including ‘behavioural’ or ‘normative’ ones.4 At the same time, if new elements are included in the analysis, they should, as in any attempt to formulate new theoretical insights, be handled in the same systematic fashion as before. One should avoid, as Gorges (2001: 141–2) warns, adding new explanatory factors on an ad hoc basis, so that they become a ‘deus ex machina on a “per need” basis’. Understanding institutional change is too important, as the developments in the European Union show, to be glossed over by casual explanations.

Notes 1 As Ross (1995: 48) reports: ‘[f]ollowing skillful maneuvering by Delors and Pascal Lamy at the top, the Community was given responsibility for coordinating G-24 (the OECD countries) and Community aid to Poland and Hungary at the July 14, 1989 Paris G-7 summit, a task later expanded to include the rest of the CEECs.’ 2 This statement is based on information obtained during an informal meeting with a Dutch politician. 3 In the case of ‘constitutional economics’ an equilibrium is often derived by extreme state of uncertainty in which political actors do not know their ‘future’ positions – or roles – in society (see, for instance, Brennan and Buchanan 1985). 4 The term ‘normative’ often leads to confusion, since it may imply a shift away from developing a positive theory of political decision-making. Here the term is used to address the shared understandings or common beliefs of political actors, which are properties of a group of individual actors, and not of the individual itself.

References Brennan, G. and J.M. Buchanan (1985) The Reason of Rules: Constitutional Political Economy, Cambridge: Cambridge University Press.

248

B. Steunenberg and A. Dimitrova

Breton, A. (1996) Competitive Government: An Economic Theory of Politics and Public Finance, Cambridge: Cambridge University Press. Dimitrova, A.L. (2001) ‘Governance by enlargement? The case of the administrative capacity requirement in the EU’s Eastern enlargement’, Paper presented at the ECPR General Conference, 6–8 September 2001, University of Kent, Canterbury. European Commission (2001) ‘European governance: a White Paper’, Brussels, 25 July 2001 (COM(2001)428) (also at http://www.europa.eu.int/comm/governance/index_en.htm). Fischer, J. (2000) ‘From confederacy to federation: thoughts on the finality of European integration’, Speech at the Humboldt University in Berlin, 12 May 2000 (http://www. germanembassy.org.au/eu-fisch.htm). Garrett, G. and B. Weingast (1993) ‘Ideas, interests, and institutions: constructing the European Community’s internal market’, in J. Goldstein and R. Keohane (eds) Ideas and Foreign Policy, Ithaca, NY: Cornell University Press, 173–206. Gorges, M.J. (2001) ‘New institutionalist explanations for institutional change: a note of caution’, Politics 21:137–45. Krehbiel, K. (1991) Information and Legislative Organization, Ann Arbor: University of Michigan Press. Majone, G. (1996) Regulating Europe, London: Routledge. Moe, T.M. (1990) ‘Political institutions: the neglected side of the story’, Journal of Law, Economics, and Organization 6 (Special Issue): S213–53. Mueller, D.C. (1996) Constitutional Democracy, Cambridge: Cambridge University Press. North, D.C. (1990) Institutions, Institutional Change and Economic Performance, Cambridge: Cambridge University Press. Ostrom, E. (1998) ‘A behavioral approach to the rational choice theory of collective action’, American Political Science Review 92: 1–22. Pollack, M. (1997) ‘Delegation, agency and agenda setting power in the European Community’, International Organization 51: 128–42. Prodi, R. (2001) The enlargement speech to the European Parliament Strasbourg by President of the European Commission, 13 November 2001 (DN:SPEECH/01/531). Riker, W.H. (1980) ‘Implications from the disequilibrium of majority rule for the study of institutions’, American Political Science Review 74: 432–47. Ross, G. (1995) Jacques Delors and European integration, Cambridge: Polity Press. Scharpf, F.W. (2001) ‘European governance: common concerns versus the challenge of diversity’, Jean Monnet Working Paper no. 6/01 (http://www.jeanmonnetprogram.org/ papers/01/010701.html). Shepsle, K.A. (2001) ‘A comment on institutional change’, Journal of Theoretical Politics 13: 321–5. Smyrl, M. (1998). ‘When (and how) do the Commission’s preferences matter?’, Journal of Common Market Studies 36: 79–101. Thelen, K. (1999) ‘Historical institutionalism in comparative politics’, in Annual Reviews of Political Science, Volume 2, Palo Alto: Annual Reviews, 369–404. Villa Faber Group (2001) ‘Thinking enlarged: the accession countries and the future of the European Union. A strategy for reform’, Munich: Bertelsmann Foundation and Center for Applied Policy Research.

Chapter Title

249

Index

acquis communautaire, adoption of 58–9, 109, 113, 241 agricultural levies, EU budget source 146–7 allocative efficiency, resources 49–51 ambiguity, enlargement 121–3 Amsterdam Treaty 84–6, 97 effects 106–9 application-accession times 187 Bank, European Central see European Central Bank Bank, European Investment 147 Banks, Federal Reserve 170–3 Banks, National Central see National Central Banks Banzhaf index 128–9, 212 Banzhaf voting power 128–36, 212 bargaining, enlargement negotiations 192–6 benefits confederations 45–6 public goods 42–4 voting rules 105 benefits/costs of membership, EU 45–8 blocking the Council of Ministers see voting power boundary issues 58–79 Buchanan, J. 59–61 budgets budgetary procedures 151–9 budgetary process 146–51, 220 compulsory/non-compulsory spending 147–51 decision-making 14, 144–62 enlargement and 144–62 PDB 148–9

political parties and 220 referendum approval 151–9 Centre for Economic Policy Research 172 change, efficacy in 90–3 change, institutional see institutional change changing times 241–8 chaos theorems 21–2 choice, rational 21–38 citizenship EU 48–9 and mobility 52–4 clubs confederations as 44–6 states as 41–4, 123–7 coalitions, political parties 208–11 collective rationality 125–6 Commission see European Commission compulsory/non-compulsory spending 147–51 confederations adding members 46 benefits 45–6 as clubs 44–6 costs of membership 45–6 defined 44 cf. federations 51–2 consequences, enlargement 87–93 Conservative Party, UK 213, 214 constitutional issues 12–13, 39–118 basic elements, federal constitution 224–6 constitutional economics, insights 224–38 direct democracy institutions 232–3, 234

250

Index

EU expansion 41–57 federal European constitution 224–38 future European constitution 232–3 implications, EU expansion 46–8 coordination mechanism, enlargement 122 costs/benefits of membership, EU 45–8 costs, heterogeneity 43 costs of membership, confederations 45–6 Council of Ministers blocking see voting power and enlargement 121–43 voting 100–3, 110 Cox, G., rational choice modelling 23–4 customs duties, EU budget source 146–7 decision-making 119–80 analytical framework 98–100 budgetary 14, 144–62 effects, enlargement 97–118 European Parliament 110–12 group 126–7 inertia index 100, 101, 104 legal issues 82–4 nature of 98–100 power increasing effect 103 reforming 14–15 strategic power index 100, 102, 105, 114–15 decision probabilities, comparative 87–90 decisiveness, enlargement 122 deepening integration 183–201 model 191–2, 193 and widening 184–9 direct democracy institutions 232–3, 234 dynamics of widening 15–16, 181–238 ECB see European Central Bank economic theory, optimal legal areas 59–65 effectiveness decline, EU 87–90 efficiency enlargement 123 resource allocation 49–51 EMU see European Monetary Union enforcement cost function, optimal size determination 72 enforcement, enlargement negotiations 192–6 enlargement ambiguity 121–3 consequences 87–93

constitutional aspects see constitutional issues and Council of Ministers 121–43 decision-making effects 97–118 and ECB 163–80 EU 58–79 and European budget 144–62 federal European constitution 224–38 future 241–8 and institutional balance 242–4 as institutional reform process 8–12 modified model 65–6 as momentum 106 negotiations 192–6 Pareto-inferior changes 133–6 Pareto-superior changes 133–6 process 5–6 reform interdependency 6–8 reforms and 3–20, 241–2 state influence 122 theoretical debate 189–91 two waves 86–7 winners/losers 68–9 see also widening enlargement resistant effect 112 entry subsidies 53–5 EU see European Union European Central Bank (ECB) enlargement and 163–80 institutional structure 166–9 regional representation 169–70 European Commission 32–3 enlargement and institutional balance 242–4 role 33 European Court of Justice 33–4 European government, tax base 231 European Investment Bank 147 European Monetary Union (EMU) and banking systems 163–80 reduced policy cohesion 165–9 European Parliament 31–2 decision-making 110–12 increasing role 216–19 political party formation 205–8 spatial analysis 31–2 European Union (EU) benefits/costs of membership 45–8 budgets see budgets citizenship 48–9 as club of states 41–4, 123–7 effectiveness decline 87–90 enlargement 58–79 external borrowing 147 fragmentation 206–7

Index joining criteria 80–2 policy flexibility 101–2, 104–5, 107–8 raison d’être 41, 49–54 representation 48–9 external borrowing 147

251

integration deepening 183–201 regional, depth/scope 184–9 interdependency, enlargement 6–8 joining criteria, EU 80–2

Federal Reserve Banks 170–3 federalism 41–4 federal European constitution 224–38 fiscal federalism 230–1, 234 government growth safeguard 230–1, 234 optimal club size 41–2 federations, cf. confederations 51–2 FI see Forza Italia fiscal constraints 144–62 see also budgets fiscal federalism 230–1, 234 Forza Italia (FI) 213, 215 French Gaullists (RPR) 213, 215 future enlargement 241–8 European constitution 232–3 theoretical challenges 241–8 game theory 26–31, 61–3, 128–9 Germany, monetary union insights 170–3 GNP resource, EU budget source 147 government growth safeguards 228–33, 234 government tax base 231 Grim Trigger strategy 192–6, 198–9 group decision-making 126–7 Helsinki European Council 6 income production, optimal size determination 72 independence, NCBs 164–5 individual players, power 105, 108–9 inertia index, decision-making 100, 101, 104 inflation, and NCBs 164 institutional balance, and enlargement 242–4 institutional change analysing 245–7 institutional reforms 8–12, 244–5 politics 80–96 rational choice and 9, 21–38 institutional reform process, enlargement as 8–12 institutional rules, and political battles 21–2 institutions for stability 22–4

Krehbiel, K., rational choice modelling 23–4 legal issues decision-making 82–4 diverging goals 80–96 legislation design, federal constitution 226–8 optimal legal area boundaries 58–79 unanimity 82–4 losers/winners, enlargement 68–9 McCubbins, M., rational choice modelling 23–4 migration 52–4 mobility, and citizenship 52–4 momentum, enlargement as 106 monetary policy 14–15 see also budgets monetary union insights, Germany/US 170–3 National Central Banks (NCBs) 163–4 independence, case for 164–5 and inflation 164 negotiation stances/positions 197–9 Nice Council 4 Nice Intergovernmental Conference 81, 84–6 Nice reform consequences 90–3 Nice Treaty 4, 13, 82, 97 policy flexibility 107–8 power, individual players 108–9 non-compulsory/compulsory spending 147–51 optimal enlargement federalism 41–2 legal area boundaries 58–79 points of view 75–6 optimal legal area boundaries 58–79 economic theory 59–65 optimal size issues 41–2, 63–70 optimality conditions, optimal size determination 71 Pareto-inferior changes 133–6 Pareto-superior changes 133–6

252

Index

Parliament see European Parliament participation constraint, optimal size determination 71 Party of Democratic Left (PDS) 213 PDB see Preliminary Draft Budget PDS see Party of Democratic Left policy cohesion, EMU and 165–9 policy flexibility after Nice 107–8 EU 101–2, 104–5 political battles, and institutional rules 21–2 political parties 202–23 budgetary process and 219 coalitions 208–11 dilemma 211–16 evolution 213–14 political party formation, European Parliament 205–8 relative positions 210–11 voting cohesion 216–19 voting power 207–11 politics EU 10–12 institutional change 80–96 power decreasing effect 106 power increasing effect 106 decision-making 103 power index approach 27–31, 127, 211–16 power, individual players 105, 108–9 Preliminary Draft Budget (PDB) 148–9 prisoners’ dilemma game 61–3 see also game theory procedural studies 11–12 Prodi, R. 241–2 public goods, benefits 42–4 rational choice, and institutional change 21–38 redistribution, resources 49–51 referendum approval, budgets 151–9 reform decision-making 14–15 and enlargement 3–20, 8–12, 241–2 enlargement interdependency 6–8 regional integration, depth/scope 184–9 regional representation, ECB 169–70 representation enlargement 122 EU 48–9 requirement of collective rationality 125–6 requirement of group power 126–7 requirement of individual power 127

requirement of individual rationality 125–6 resources allocative efficiency 49–51 redistribution 49–51 RPR see French Gaullists size issues enlargement 122 EU 46–7, 63–70 federalism 41–2 optimal size issues 41–2, 63–70 social contract theory 59–63 sovereignty 122 spatial analysis, European Parliament 31–2 spatial approach 26–31 spending, compulsory/non-compulsory 147–51 stability institutions for 22–4 transaction cost 24–6 state influence, enlargement 122 states, as clubs 41–4, 123–7 strategic power index, decision-making 100, 102, 105, 114–15 subsidiarity principle, government growth safeguard 228–30 subsidies, entry 53–5 tax base, European government 231 theoretical challenges, future 241–8 Tit-for-Tat strategy 193–6 transaction cost, stability 24–6 transparency, enlargement 123 trigger strategy, Grim Trigger 192–6, 198–9 unanimity argument against 128–30 legal issues 82–4 uncertainty, enlargement negotiations 192–6 United States, monetary union insights 170–3 VAT, EU budget source 146–7 vital interests protection 123 voting cohesion, political groups 216–19 voting power 2005; 134 Banzhaf 128–36, 212 EU Council 100–3, 110 political parties 207–11

Index relevance 128–9 studies 10–12, 14, 28–31 weighted 127 voting rules 10–12, 110 benefits 105 weighted voting 127

253

widening and deepening 184–9 dynamics of 15–16, 181–238 model 191–2, 193 see also enlargement winners/losers, enlargement 68–9 World Trade Organization (WTO) 184