Welfare State Change in Leading OECD Countries: The Influence of Post-Industrial and Global Economic Developments (ebs-Forschung, Schriftenreihe der ... BUSINESS SCHOOL Schloß Reichartshausen, 74) 3834919012, 9783834919014

Citation preview

Ingmar J. Schustereder Welfare State Change in Leading OECD Countries

GABLER RESEARCH EBS Forschung Schriftenreihe der European Business School (EBS) International University · Schloss Reichartshausen Herausgegeben von Univ.-Prof. Ansgar Richter, PhD

Band 74

Die European Business School (EBS) – gegründet im Jahr 1971 – ist Deutschlands älteste private Wissenschaftliche Hochschule für Betriebswirtschaftslehre im Universitätsrang. Dieser Vorreiterrolle fühlen sich ihre Professoren und Doktoranden in Forschung und Lehre verpflichtet. Mit der Schriftenreihe präsentiert die European Business School (EBS) ausgewählte Ergebnisse ihrer betriebs- und volkswirtschaftlichen Forschung.

Ingmar J. Schustereder

Welfare State Change in Leading OECD Countries The Influence of Post-Industrial and Global Economic Developments With a foreword by Prof. Dr. Joachim Ahrens

RESEARCH

Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de.

Dissertation European Business School, International University Schloss Reichartshausen, Oestrich-Winkel, 2009 D 1540

1st Edition 2010 All rights reserved © Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2010 Editorial Office: Ute Wrasmann | Britta Göhrisch-Radmacher Gabler Verlag is a brand of Springer Fachmedien. Springer Fachmedien is part of Springer Science+Business Media. www.gabler.de No part of this publication may be reproduced, stored in a retrieval system or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the copyright holder. Registered and/or industrial names, trade names, trade descriptions etc. cited in this publication are part of the law for trade-mark protection and may not be used free in any form or by any means even if this is not specifically marked. Umschlaggestaltung: KünkelLopka Medienentwicklung, Heidelberg Printed on acid-free paper Printed in Germany ISBN 978-3-8349-1901-4

)RUHZRUG For many years, academics and practitioners in economics as well as political science have been engaged in a controversial discussion about the driving forces behind the emergence, change, and reform of modern welfare states. This study aims at enhancing our understanding of institutional, structural, and political adjustment processes in advanced capitalist welfare states. This is being investigated against the background of globalization processes of labor, capital, and goods markets, on the one hand, and post-industrial developments within nation states as outlined by Pierson (2001a) (such as population ageing, de-industrialization, changing household structures, rising unemployment) on the other hand. The key research objective is to analyze the processes of the recent past between 1975 and 2000. In particular, the author addresses two research questions: (1) what is the relative influence of post-industrial and globally economic forces, respectively, with respect to recent changes of welfare states?, and (2) to what extent could different national systems of social security maintain their institutional core characteristics over time? Addressing these questions is not only of great relevance from an economic point of view, but also highlights the enormous political importance. Eventually, the question takes center stage in how far specific national regulatory systems are viable in times of globalization and to what extent national governments are still capable of acting independently regarding the creation and reform of institutions. Ingmar Schustereder argues that countries adjust in different ways to these challenges. Although all welfare states have been under adjustment pressure, some nations appear to be better able than others to confront these contemporary challenges. Based on an in-depth discussion of different theoretical approaches, the author applies both sophisticated quantitative and qualitative techniques to shed light on these different adjustment strategies. This investigation at the interface of economics, political science, and sociology represents, without doubt, an outstanding piece of research. From a theoretical perspective, the author provides a concise and yet differentiated survey of all relevant approaches, which are applied in an intellectually stimulating way. The subsequent empirical analysis is based on a broad range of variables, applies pooled cross-section time series models, and is complemented by illustrative case studies. This balanced, theoretically well-grounded, and empirically convincing analysis will benefit a wide range of readers in the academic as well as in the policy-making community. It significantly contributes to a better understanding of changes in contemporary welfares states. Most important new insights relate to the realms of comparative empirical welfare state research, the research of globalization processes, and the so-called varieties-of-capitalism approach. Göttingen, November 2009

Prof. Dr. Joachim Ahrens

Acknowledgement Conceptualizing and writing the dissertation at hand has been both a challenging and interesting experience for me. Along this academic journey, many people have accompanied me to whom I am very grateful. In particular, I would like to thank my Doktorvater Prof. Dr. Joachim Ahrens for valuable advice and guidance. I am also indebted to Prof. Dr. André Schmidt for taking on the role as second supervisor upon the unfortunate death of Prof. Dr. Rolf Caspers. Without them, the completion of this dissertation would not have been possible. During the course of my research, I have gained important insights from my classmates in the Advanced Studies Program 2005/ 2006 at the Kiel Institute for the World Economy, researchers at the Department of Political and Social Sciences at Pompeu Fabra University, and participants of the Oslo Summer School in Comparative Social Science Studies. To them I give my deepest appreciation. Furthermore, I would like to thank Dipl.-Kfm. Robert Esser, Dipl.-Vw. Christian Gäckle, Dipl.-Ing. Stephan Salinger, Mag. Dr. Sandra Stötzer, and Dipl.-Ök. Dr. Dennis Stremmel for their moral support. Last, but not least I am very grateful to my family and Melody Garcia for their endless patience, love, and support throughout all these years. I dedicate this dissertation to them. Ingmar Josef Schustereder

Contents )RUHZRUG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII0 List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XIII0 List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XV0 List of Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XVII0 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 1.2 1.3 1.4

1

Research Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theoretical Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Course of Investigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 2 3 6

Part I The Welfare State and Welfare Capitalism . . . . . . . . . . . . . . . . . . . . .

9

2 Theoretical Foundations of the Welfare State . . . . . . . . . . . . . . . . . . . . . . .

11

2.1 The Welfare State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1 Definition of the Terms Welfare and Welfare State . . . . . . . . . . . 2.1.2 The Core Functions of the Welfare State . . . . . . . . . . . . . . . . . . . 2.2 Political Economic Theories on the Rise of the Welfare State . . . . . . . . . 2.3 Welfare State Regimes and the Classification of Advanced Capitalist Welfare States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.1 Esping-Andersen’s Welfare State Regime Approach . . . . . . . . . 2.3.1.1 The Welfare State Regimes and De-Commodification . 2.3.1.2 The Welfare State Regimes and Social Stratification . . 2.3.1.3 The Welfare State Regimes and De-Familialization . . . 2.3.2 Classification of Advanced Capitalist Welfare States . . . . . . . . .

11 11 12 15

3 Worlds of Welfare Capitalism: Interlinking Welfare State and Production Regimes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

18 18 19 20 20 21

25

3.1 Varieties of Production Regimes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 3.2 The Welfare State-Economy Nexus . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 3.3 Classification of Welfare Capitalist Systems . . . . . . . . . . . . . . . . . . . . . 33

X

Contents

Part II Historical Development and Contemporary Challenges of the Advanced Capitalist Welfare State . . . . . . . . . . . . . . . . . . . . . .

39

4 Historical Phases of Welfare State Development . . . . . . . . . . . . . . . . . . . . .

41

4.1 4.2 4.3 4.4 4.5

Social Protection in the Pre-Welfare State Era . . . . . . . . . . . . . . . . . . . . The Genesis of the Welfare State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Interwar Period of Consolidation . . . . . . . . . . . . . . . . . . . . . . . . . . . The “Golden Age” of Major Welfare State Expansion . . . . . . . . . . . . . . The Welfare State in the Post-“Golden Age” Era . . . . . . . . . . . . . . . . . .

41 42 44 45 48

5 Post-Industrial Challenges of the Contemporary Capitalist Welfare State

49

5.1 5.2 5.3 5.4 5.5

The Phenomenon of Deindustrialization . . . . . . . . . . . . . . . . . . . . . . . . . 49 The Demographic Shift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Changing Household Structures and the Rise of New Social Risks . . . . 59 Welfare State Maturation and Labor Market Imbalances . . . . . . . . . . . . 63 Post-Industrial Pressures and Agents of Welfare State Change: Partisan Politics versus New Politics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

6 Economic Globalization and the Contemporary Capitalist Welfare State 6.1 The Underlying Causal Forces of Economic Globalization . . . . . . . . . . 6.2 The Three Dimensions of Economic Globalization . . . . . . . . . . . . . . . . 6.2.1 The Globalization of Product Markets . . . . . . . . . . . . . . . . . . . . 6.2.1.1 Trade Globalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2.1.2 Foreign Direct Investment and the Globalization of Corporate Activities . . . . . . . . . . . . . . . . . . . . . . . . . 6.2.2 The Globalization of Capital Markets . . . . . . . . . . . . . . . . . . . . . 6.2.3 The Globalization of Labor Markets . . . . . . . . . . . . . . . . . . . . . . 6.3 The Effects of Global Economic Forces on the Welfare State: Four Different Hypotheses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.1 The Openness Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.1.1 The Competitiveness View . . . . . . . . . . . . . . . . . . . . . . 6.3.1.2 The Compensation View . . . . . . . . . . . . . . . . . . . . . . . . 6.3.1.3 The Curvilinear View . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.2 The Context-Contingent View – Globalization and Varieties of Welfare Capitalism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

71 71 72 73 73 77 80 87 94 95 95 97 99 99

Part III Empirical Analyses of the Driving Forces Behind the Contemporary Capitalist Welfare State . . . . . . . . . . . . . . . . . . . 105 7 Methodological Issues in Comparative Welfare Capitalism Research . . . . 107

Contents

XI

8 A Quantitative Study on the Determinants of Welfare Spending . . . . . . . . 111 8.1 Patterns of Public Social Spending . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2 The Econometric Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2.1 Contribution to Comparative Welfare State Research . . . . . . . . . 8.2.2 Modeling Potential Determinants of Welfare Spending . . . . . . . 8.2.2.1 Explanatory Variables as a Proxy for Post-Industrial Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2.2.2 Explanatory Variables as a Proxy for the Openness Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2.2.3 Explanatory Variables as a Proxy for the Context-Contingent View . . . . . . . . . . . . . . . . . . . . . . . 8.2.2.4 Control Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4 Empirical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4.1 Baseline Model Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4.2 Interaction Model Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4.3 Structural Break Model Results . . . . . . . . . . . . . . . . . . . . . . . . . 8.4.4 Shortcomings of the Quantitative Analysis and Further Research Needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

111 116 116 118 119 120 120 121 124 128 129 135 145 148

9 A Qualitative Study on the Recent Development Process of Advanced Capitalist Welfare States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 9.1 The Development of Welfare State Revenues in the Contemporary Era . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1.1 The Competitiveness View and the Hypothesis of International Tax Competition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1.2 The Hypothesis of Tax Competition in the Empirical Literature . 9.1.3 The Evolution of Welfare State Revenues: A Cross-Country Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2 The Development of Welfare Expenditure Patterns in the Contemporary Era . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.1 A Brief Summary Review of Welfare State Regime Analysis . . . 9.2.2 The Recent Development Process of the Three Welfare State Clusters: A Comparative Analysis . . . . . . . . . . . . . . . . . . . . . . . . 9.2.2.1 The Three Welfare State Clusters in Broad Perspective: Size and Principal Categories of Public Social Spending . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.2.2 The Three Welfare State Clusters and the Welfare State-Labor Market Nexus . . . . . . . . . . . . . . . . . . . . . .

151 151 153 154 164 165 165

166 170

XII

Contents

10 Summarized Findings and Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203

List of Figures Figure 1: Welfare State and Social Protection . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Figure 2: Developed Countries and the Process of Deindustrialization, 1960-2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

50

Figure 3: Developed Countries and the Evolution of Female Labor Force (as % of Total Labor Force), 1975-2000 . . . . . . . . . . . . . . . . . . . . . .

60

Figure 4: Total Government Disbursements (as % of GDP), 1960-2005 . . . . .

64

Figure 5: Developed Countries and Current Account Liberalization, 1975-1999 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

77

Figure 6: Developed Countries and Capital Account Liberalization, 1975-1999 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

86

Figure 7: The Average Development of Four Welfare State Indicators in 17 OECD Countries, 1975-2000 . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Figure 8: The Average Development of Public Social Expenditures in the Three Welfare State Clusters, 1980-2000 . . . . . . . . . . . . . . . . . . . 114 Figure 9: The Average Development of Social Security Transfers in the Three Welfare State Clusters, 1975-2000 . . . . . . . . . . . . . . . . . . . 115 Figure 10: The Average Development of the Benefit Generosity Index in the Three Welfare State Clusters, 1975-2000 . . . . . . . . . . . . . . . . . . . 115 Figure 11: The Average Development of the Pension Generosity Index in the Three Welfare State Clusters, 1975-2000 . . . . . . . . . . . . . . . . . . . 116

List of Tables Table Table Table Table

1: 2: 3: 4:

Table 5: Table Table Table Table

6: 7: 8: 9:

Table 10: Table 11: Table 12: Table 13: Table 14: Table 15: Table 16: Table 17: Table 18: Table 19: Table 20: Table 21: Table 22:

A Summary Overview of Regime Characteristics . . . . . . . . . . . . . . . . Classification of 17 Advanced Capitalist Welfare States . . . . . . . . . . . Selected Institutional Domains in Capitalist Economies . . . . . . . . . . . Two Varieties of Capitalism: Uncoordinated vs Coordinated Market Economies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Continental European and the Anglo-Saxon Pension System in Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Classification of 17 Advanced Market Economies . . . . . . . . . . . . . . . The Introduction of Social Insurance Legislation in OECD Countries The Expansion of Citizenship in OECD Countries . . . . . . . . . . . . . . . Welfare State Expansion and Unionization in OECD Countries, 1950-1975 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ratio of Population Aged 65 and Over to the Total Population in 17 OECD Countries, 1950-2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . Total Fertility Rates in 17 OECD Countries, 1970-2005 . . . . . . . . . . . The Continental European and the Anglo-Saxon Pension System in Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Poverty Rates by Employment Status/ Household Type and the Employment/ Population Ratio of Sole Parent Households . . . . . . Trends in Average Applied Tariff Rates in Industrialized and Developing Countries, 1988-2007 (unweighted in %) . . . . . . . . . . . . Degree of Trade Openness in OECD Countries, 1975-2000 . . . . . . . . Degree of Foreign Direct Investment Openness in OECD Countries, 1965-1998 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gross Private Capital Flows, 1977-2004 (in % of GDP) . . . . . . . . . . . The Five Largest SWFs by Assets Under Management (as of January 2009) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Net Migration Flows to OECD Countries, 1970-2003 . . . . . . . . . . . . The Educational Attainment of the Native- and Foreign-Born Populations (circa 2000) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Definitions of Dependent, Explanatory and Control Variables . . . . . . Overview of the Different Hypotheses on the Determinants of Public Social Spending in OECD Countries . . . . . . . . . . . . . . . . . .

22 23 27 28 32 35 43 44 47 54 56 58 62 74 76 78 85 87 89 91 122 123

XVI

List of Tables

Table 23: Baseline Model Results – Dependent Variable: Public Social Expenditures (PSEX), 1980-2000 . . . . . . . . . . . . . . . . . Table 24: Baseline Model Results – Dependent Variable: Social Security Transfer (SSTRAN), 1975-2000 . . . . . . . . . . . . . . . . Table 25: Baseline Model Results – Dependent Variable: Benefit Generosity Index, 1975-2000 . . . . . . . . . . . . . . . . . . . . . . . . . Table 26: Baseline Model Results – Dependent Variable: Pension Generosity Index, 1975-2000 . . . . . . . . . . . . . . . . . . . . . . . . . Table 27: Interaction Model Results – Dependent Variable: PSEX, 1980-2000 . Table 28: Interaction Model Results – Dependent Variable: SSTRAN, 1975-2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Table 29: Interaction Model Results – Dependent Variable: Benefit Generosity Index, 1975-2000 . . . . . . . . . . . . . . . . . . . . . . . . . Table 30: Interaction Model Results – Dependent Variable: Pension Generosity Index, 1975-2000 . . . . . . . . . . . . . . . . . . . . . . . . . Table 31: Structural Break Model Results – Dependent Variable: PSEX, 1980-2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Table 32: Structural Break Model Results – Dependent Variable: SSTRAN, 1975-2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Table 33: Total Tax Revenues in Comparative Analysis, 1975-2005 . . . . . . . . . . Table 34: Corporate Income Tax Revenues in Comparative Analysis, 1975-2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Table 35: Personal Income Tax Revenues in Comparative Analysis, 1975-2005 Table 36: Payroll Tax Revenues in Comparative Analysis, 1975-2005 . . . . . . . . Table 37: Social Security Contributions in Comparative Analysis, 1975-2005 . . Table 38: General Consumption Tax Revenues in Comparative Analysis, 1975-2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Table 39: The Recent Development of the Three Welfare State Clusters: Size of Public Social Spending and Income Equality, 1980-2000 . . . . Table 40: The Recent Development of the Three Welfare State Clusters: Principal Welfare State Spending Categories, 1980-2000 . . . . . . . . . . Table 41: The Recent Development of the Three Welfare State Clusters: The Welfare State-Labor Market Nexus (I) . . . . . . . . . . . . . . . . . . . . . Table 42: The Recent Development of the Three Welfare State Clusters: The Welfare State-Labor Market Nexus (II) . . . . . . . . . . . . . . . . . . . .

130 131 132 134 136 139 141 143 147 148 156 158 159 160 162 163 167 169 171 173

List of Abbreviations AR (1) BIS EPL FDI FX system GATS GATT GDP GNP IMF ITO MNCs MR NGOs OECD OLS PCSE PSEX SSTRAN SWFs TSCS UK US/ USA TNCs TRIMs TRIPS UN UNCTAD VoC vs WDI WTO WWI WWII

First-Order Autocorrelation Bank for International Settlements Employment Protection Legislation Foreign Direct Investment Fixed Exchange-Rate System General Agreement on Trade in Services General Agreement on Tariffs and Trade Gross Domestic Product Gross National Product International Monetary Fund International Trade Organization Multinational Corporations Multiple Regression Non-Governmental Organizations Organisation for Economic Co-operation and Development Ordinary Least Squares Panel Corrected Standard Errors Public Social Expenditures as percentage of GDP Social Security Transfers as percentage of GDP Sovereign Wealth Funds Time-Series Cross-Section United Kingdom United States Transnational Corporations Agreement on Trade-Related Investment Measures Agreement on Trade-Related Aspects of Intellectual Property Rights United Nations United Nations Conference on Trade and Development Varieties of Capitalism versus World Bank Development Indicators Database World Trade Organisation First World War Second World War

1

Introduction

1.1

Research Scope

Ever since the introduction of social insurance legislation beginning in the last decades of the 19th century, social scientists have been interested in analyzing the causal forces behind the development of the modern welfare state. Precisely, for most of the period the academic debate has hereby mainly centered on the following questions: Why was the welfare state founded in the first place? And which factors have contributed to its rapid expansion in the aftermath of World War Two (WWII)?1 With the advanced capitalist welfare states beginning to face increasing challenges from the mid-1970s onwards, academic research has recently shifted from providing explanations on the rise of the welfare state to understanding processes of welfare state adaptation. In general, advanced capitalist welfare states confront today two types of challenges: on the one hand, pressures from post-industrial developments, and on the other hand, pressures from global economic forces (see Buti, Franco, and Pench, 1998, pp. 17-26). Following Pierson (2001a), the most severe “post-industrial pressures” (p. 80) involve ongoing processes of deindustrialization, population ageing, changing household structures, and welfare state maturation (pp. 82-83). Apart from these domestic challenges, advanced capitalist welfare states have recently also been affected by developments occurring in the global sphere. While processes of international economic integration had already resumed in the immediate aftermath of WWII, they have significantly intensified during the last four decades. Besides the gradual lifting of impediments to international trade, the emergence of multinational corporations (MNCs) and the deregulation of financial markets (following the liberalization of the exchange rate system in the early 1970s) have in particular contributed to the growing interdependence of nation states (see OECD, 2005a, p. 16). Although global economic forces appear to shape social policy, their intensity and direction are still under debate (for an overview of theories on the globalization-welfare state nexus, see Chapter 1.2). In light of rising challenges both in the domestic and global realm, the aim of this dissertation is to analyze the recent development process of advanced capitalist welfare states. Precisely, the author will focus on the following two research questions: Firstly, what has been the relative influence of post-industrial and global economic developments on recent welfare state change? And secondly, to what extent have different national systems of social protection preserved their core institutional features? 1

For a thorough review of political economic theories on the rise of the welfare state, see Esping-Andersen (1990/ 1998, Chapter 1) and Chapter 2.2 in this dissertation.

2

1.2

1 Introduction

Theoretical Framework

Taking a holistic approach, the author tests various theories on the driving forces behind the contemporary advanced capitalist welfare state. Partly in line with recent categorizations by a number of scholars (for example, Brady, Beckfield, and SeeleibKaiser, 2005, pp. 922-925; Brady, Beckfield, and Zhao, 2007, pp. 318-320; Ellison, 2006, pp. 48-59; Genschel, 2004; Hicks, 1999, pp. 157-168, Chapter 7; Koster, 2008, pp. 2-3; Swank, 2002/ 2004, 20-58), the author hereby classifies the theories into three broad schools of thought: Proponents of the first school of thought (for example, Castles, 2001; Iversen, 2001; Iversen & Cusack, 2000; Pierson, 2001a; Pierson, 1996; Pierson, 1994/ 1997) downplay the role of global economic forces, suggesting instead that domestic factors are by far more important drivers of contemporary welfare state development. Precisely, deindustrialization, population ageing, household structure transformation, and welfare state maturation are considered to be the key challenges of social policy (see Pierson, 2001a). As awareness of these domestic pressures has significantly grown in recent decades, an intense debate has erupted about the potential implications these post-industrial developments are likely to have for the process of political decisionmaking. Social scientists in comparative welfare state research, affiliated with the power resources approach and the class mobilization thesis2 (such as Allan & Scruggs, 2004; Castles, 1982; Esping-Andersen, 1985/ 1988; Garrett, 1998a; Huber & Stephens, 2000; Korpi, 1989; Korpi, 1980; Korpi & Palme, 2003), are of the opinion “that it is fruitful to view welfare states as outcomes of, and arenas for, conflicts between class-related, socioeconomic interest groups and that in these distributive conflicts partisan politics is likely to matter” (Korpi & Palme, 2003, p. 425). Recently, this view has been challenged by the emergence of the new politics approach which states that political parties and labor organizations have ceded power to pro-welfare state oriented interest groups (mainly welfare state employees and recipients) within the social policy context (Pierson, 1996, p. 147; see also Pierson, 1994/ 1997, 2001b, 2001c).3 In contrast to the prior academic camp, the second school of thought – the socalled Openness Literature – places global economic forces at the core of recent processes of welfare state change. Precisely, this strand of literature revolves around the following three perspectives: the competitiveness view, the compensation view, and the curvilinear view (see Hicks, 1999, pp. 204-208; Brady et al., 2005, pp. 922-924). According to the competitiveness view, economic globalization – fostering 2 3

See Chapter 2.2 for a detailed explanation of these theoretical concepts. A detailed analysis of the post-industrial challenges and their potential implications for the political decision-making process is provided in Chapter 5.

1.3 Methodology

3

locational competition between nation states to attract mobile production factors (Siebert, 1999/ 2000, p. 241) – entails a decline in state capacity (Cerny, 1994, p. 334) and a simultaneous erosion of the post-WWII welfare state (Sinn, 2003, pp. 64-65; see Razin & Sadka, 2005, p. 4, 35-36; Sinn, 2001, pp. 13-15). By contrast, proponents of the compensation view assume that economic globalization induces a rise in public social spending since citizens need to be protected from the fluctuations of global markets (see Cameron, 1978, pp. 1255-1258, 1260; Hicks & Swank, 1992, pp. 666-667; Katzenstein, 1985, p. 9, 24; Ruggie, 1982, pp. 392-393, 399). Finally, the curvilinear view combines the compensation and the competitiveness view in a sequential order. Accordingly, the process of international economic integration is at first accompanied by a gradual augmentation of public transfer programs; however, beyond a certain level of international capital openness, welfare spending is cut in order to reduce the tax burden on mobile production factors since the latter would otherwise escape from the country thereby undermining the welfare state’s fiscal stance (Rodrik, 1997, pp. 89-90; Hicks, 1999, pp. 206-211).4 The third school of thought – the so-called context-contingent view5 or “‘weak globalization’ thesis” (Ellison, 2006, p. 48) – acknowledges that global economic processes have been augmenting in intensity; it nevertheless claims that the direction and force of these globalization effects on national systems of social protection are shaped by respective countries’ institutional context. Consequently, adjustment strategies – being influenced by path dependency – are assumed to vary leaving the general diversity of existing welfare state clusters more or less unchanged (Bonoli & Palier, 2001, p. 58; Ellison, 2006, pp. 48-49; Leibfried & Obinger, 2001, p. 5; Scharpf, 2000, p. 224; Swank, 2003, pp. 60-61; Swank, 2002/ 2004, p. 5).6

1.3

Methodology

Throughout the dissertation, a three-step methodological approach is applied. In the first step, the author conducts a comprehensive classification of 17 leading welfare economies. Relying to a great extent on Esping-Andersen’s (1990/ 1998, 2000) welfare state regime indicators, each welfare state is assessed according to the level of de-commodification, the modes of social stratification, and the degree of de-familialization. Subsequently, in line with Hall & Soskice’s (2001/ 2004a, 2001/ 2004b) Varieties of Capitalism (VoC) approach, the production side of the 17 welfare economies 4

A detailed presentation of the Openness Literature is provided in Chapter 6.3.1. The author coined this term to best describe the essence of the literature on globalization and varieties of welfare capitalism. 6 A detailed presentation of the context-contingent view is provided in Chapter 6.3.2. 5

4

1 Introduction

is evaluated. Following these comparative analyses, the author examines the recent development of post-industrial challenges and global economic forces. The second and third methodological step comprise in-depth empirical analyses to test all prevailing theories on the driving forces behind the contemporary advanced capitalist welfare state. Similar to prior studies in this field of research (for example, Huber & Stephens, 2001a; Swank, 2002/ 2004), the author pursues a hybrid approach consisting of quantitative and qualitative analyses (“triangulation”7). With respect to the quantitative analysis (second methodological step), a set of various pooled crosssection time series models is applied. The dataset hereby includes the 17 welfare economies for the time period 1975-2000. Brady et al.’s (2005) model provides a basis for the econometric analysis because it takes a truly holistic approach to testing all of the major theoretical hypotheses on the determinants of public social spending in OECD countries.8 In order to make the results of this quantitative study comparable with their earlier findings, the author has opted to use to a certain extent a similar dataset for the explanatory variables as well as the welfare state measures. Moreover, with respect to model operationalizations, the author follows Brady et al. (2005) in relying to a great extent on Huber & Stephens (2001a, 2000) prior works. Although the econometric analysis is built on previous academic studies, it also addresses a number of limitations of past research. Specifically, the following contributions are made to contemporary comparative welfare state research: Firstly, to the best of the author’s knowledge, this quantitative study is one of the first of its kind to link the VoC and the Three Worlds of Welfare Capitalism approach in a comprehensive econometric analysis for the recent period. While Brady et al. (2005) conduct a series of sensitivity analyses for liberal/ non-liberal as well as European/ nonEuropean welfare states (pp. 940-943), the author constructs a series of interaction models that simultaneously control for liberal, conservative, and social-democratic welfare state regime clusters. In this context, it should be noted that Swank (2002/ 2004) has already run regression analyses for each of the three welfare state clusters. The dataset of Swank (2002/ 2004) however includes only the time period 1965-1993 (pp. 103-121). Recently, Kim & Zurlo (2008) have conducted a similar comparative study using a dataset from 1980-2001. Secondly, this quantitative study sheds light on the rapid decline of public social spending in the Nordic countries in the early 1990s. Using a structural break model, the author analyzes the hump-shaped pattern of public social spending in the socialdemocratic welfare regime cluster in detail.

7

According to Shalev (2007), “triangulation […] means combining MR [Multiple Regression] with other types of analysis – quantitative, qualitative or both” (p. 296). 8 It should be noted that Brady et al. (2005) have heavily relied on Huber & Stephens (2001a, 2000) when constructing their own baseline model (p. 927).

1.3 Methodology

5

Thirdly, with respect to the dependent variables, Scruggs’ (2004a) Pension Generosity Index as well as the 2006 updated version of the de-commodification index, Scruggs’ (2004a) Benefit Generosity Index, is applied. The use of both welfare state indicators not only circumvents the problems related to aggregate data, but should also clarify how welfare policy changes affect the individual life chances (see Allan & Scruggs, 2004, p. 498).9 Fourthly, the author uses both annual and cumulative political partisanship variables, but runs each model with only one of the four political partisanship variables (annual left, cumulative left, annual right, and cumulative right) at a time. By comparing annual and cumulative political partisanship models, the paper attempts to make a contribution to the ongoing debate about which measure is more appropriate for comparative welfare state research (see Hicks, 1999, pp. 189-193). Fifthly and more generally, this quantitative study takes a slightly different modeling approach to Brady et al. (2005). Building an econometric model that controls for a multitude of different theoretical approaches, is a very difficult task since the author is compelled to construct a parsimonious while at the same time comprehensive model. While Brady et al. (2005) placed more emphasis on comprehensiveness this study aims to strike a balance between comprehensiveness and parsimoniousness. Starting with ten principal independent variables, the baseline model is then reduced to include only seven variables. Unlike in Brady et al.’s (2005) study, global economic forces are not categorized according to globalization advantage, openness, or threat (pp. 928-929); rather, they are related to the product (Trade Openness for international trade in goods and services) and factor (FDI Openness for international capital mobility and Net Migration for international migration) markets across which they tend to affect public social spending.10 Sixthly, in contrast to many previous studies in comparative political economy, this quantitative study takes possible nonstationarity issues in time-series cross-section (TSCS) analysis into account. Nonstationarity, if left untreated, can pose serious problems when interpreting the findings in pooled models (Kittel & Winner, 2005, p. 288). Due to recent developments in econometrics, this study tested for stationarity using the Levin-Lin-Chu panel unit root test (Levin, Lin, and Chu, 2002, pp. 2-3), a 9

For this reason, several scholars (for example, Allan & Scruggs, 2004; Brady et al., 2005; Fernandez, 2008; Korpi & Palme, 2003) have recently begun to use disaggregated welfare state measures. 10 Apart from differences in how to model globalization, the author has also not considered a number of domestic political and economic proxy variables which have been incorporated in Brady et al.’s (2005) earlier model. These variables include the following: Christian Democratic Cabinet, Union Density, Constitutional Structure, Female Labor Force Participation-Left Party Interaction Variable, Voter Turnout, Strikes, Authoritarian Legacy, GDP per Capita, Year, Military Spending, and Wage Coordination.

6

1 Introduction

commonly used test for stationarity, available as a special coded program in STATA 9. According to Bornhorst & Baum (2001), “the test may be viewed as a pooled Dickey-Fuller test, or an Augmented Dickey-Fuller (ADF) test when lags are included, with the null hypothesis that of nonstationarity (I(1) behavior)” (n. p.). Building on the insights of the quantitative study, the author conducts two types of qualitative analyses in the third methodological step. Partly following Ganghof (2000/ 2004), the focus of the first qualitative analysis is laid on the evolution of welfare state revenue flows in the 17 welfare economies during the period 1975-2005. Subsequently, in keeping partly with Swank (2002/ 2004, Chapter 4-6), the objective of the second qualitative analysis is to examine if the expenditure pattern and the underlying institutional configuration of the three welfare capitalist systems have changed in recent decades.

1.4

Course of Investigation

The dissertation is organized into three parts. Part I lays the theoretical foundations by making the reader familiar with the welfare state and the concept of welfare capitalism. Precisely, Chapter 2 defines the welfare state, gives an overview of its core functions and presents the principal political economic theories on the rise of the welfare state. Furthermore, adhering to Esping-Andersen’s (1990/ 1998, 2000) welfare state regime approach, existing variations between national systems of social protection are highlighted. In this context, a classification of 17 advanced capitalist welfare states is conducted. Chapter 3 looks at the broader context of welfare capitalism by laying the focus on the welfare state-economy nexus. At first, the author gives a brief introduction into existing varieties of production regimes. A special focus is hereby laid on Hall & Soskice’s (2001/ 2004a, 2001/ 2004b) VoC approach. Subsequently, in line with Ebbinghaus & Manow (2001a, 2001b), linkages between the welfare state and three institutional domains of the economy (the system of industrial relations, the production system and employment regime, and the corporate governance and financial system) are analyzed. Finally, the chapter is rounded off with a classification of the production regimes of the 17 welfare economies. Building on the theoretical foundations laid in Chapter 2 and 3, Part II of the dissertation brings into focus the contemporary challenges of advanced capitalist welfare states. Before analyzing these challenges, Chapter 4 reviews the historical process of welfare state development. Subsequently, adhering to Pierson’s (2001a) classification of “post-industrial pressures” (p. 80), Chapter 5 sketches the nature of welfare state challenges which are domestically provoked in the political and socioeconomic realm. Precisely, the following post-industrial developments are hereby at

1.4 Course of Investigation

7

the core of the focus: deindustrialization, population ageing, household structure transformation, and welfare state maturation (see Pierson, 2001a). Finally, the author also analyzes the potential implications of these challenges for the political decisionmaking process. Chapter 6 shifts the focus to welfare state challenges arising from economic globalization. For this reason, the author analyzes at first the causal forces behind the process of economic globalization. In this context, the development of international trade, financial and migration flows for the post-WWII era is examined. Subsequently, an overview of the principal theories on the welfare state-globalization nexus is presented. Finally, Part III of the dissertation conducts in-depth empirical analyses in order to test all prevailing theories on the driving forces behind the contemporary advanced capitalist welfare state. Similar to prior studies (for example, Huber & Stephens, 2001a; Swank, 2002/ 2004), the author applies a hybrid approach comprising of quantitative and qualitative analyses. Given the complexity of methodological issues in comparative welfare capitalism research, Chapter 7 explains in detail why this empirical approach has been chosen. Subsequently, Chapter 8 presents various econometric models and their respective findings while Chapter 9 conducts a set of qualitative analyses shedding light on both the revenue and expenditure side of national systems of social protection in all three welfare state clusters. Finally, Chapter 10 summarizes the empirical findings and concludes the dissertation.

Part I The Welfare State and Welfare Capitalism Laying the theoretical foundations, Part I makes the reader familiar with the welfare state and the concept of welfare capitalism. Precisely, Chapter 2 defines the welfare state and gives an overview of its core functions. Furthermore, the author presents the principal political economic theories on the rise of the welfare state and classifies advanced capitalist welfare states according to Esping-Andersen’s (1990/ 1998, 2000) welfare state regime approach. Subsequently, Chapter 3 looks at the broader context of welfare capitalism by focusing on the interrelationships between the welfare state and the economy.

2

Theoretical Foundations of the Welfare State

2.1

The Welfare State

2.1.1

Definition of the Terms Welfare and Welfare State

The term welfare defies clear-cut definition. Abstractly speaking, welfare may be defined as “the state of doing well especially in respect to good fortune, happiness, wellbeing, or prosperity” (Merriam-Webster, n.d.). Barr (2004) distinguishes between four different sources of individual welfare: Firstly, the labor market provides welfare in the form of wage income and occupational insurance schemes. Secondly, individual savings and private insurance contracts represent another source of welfare. Thirdly, individual welfare can also be derived through free or below-market-price services offered voluntarily by family members or others. Finally, the state also contributes to individual welfare in the form of cash benefits, benefits in kind, and subsidies (p. 6; Barr, 1992, pp. 742-743). Following Pierson’s (1998) schema, the first two categories are subsumed under economic welfare, the third category is defined as social welfare, and the fourth category falls under the rubric of state welfare (p. 6). State welfare, commonly dubbed the welfare state, can be defined in various ways. According to a narrow definition, it involves two types of public welfare provision: cash benefits and benefits-in-kind (Barr, 2004, p. 7; Lindbeck, 2006, p. 2). According to Barr (2004), cash benefits consist of two principal components: “1. Social insurance is awarded without an income or wealth test, generally on the basis of (a) previous contributions and (b) the occurrence of a specified contingency, such as becoming unemployed or reaching a specified age” (p. 7). “2. Non-contributory benefits are of two sorts. So-called universal benefits are awarded on the basis of a specified contingency, without either contributions or an income test [(for example, child allowances)]. […] social assistance is awarded on the basis of an income test. It is generally a benefit of last resort, designed to help individuals and families who are in poverty, whether as an exceptional emergency, or because they are not covered by social insurance, or as a supplement to social insurance” (p. 7). Besides cash benefits, the second category of public welfare provision – benefits-in-kind – comprises the following services: education, health care, and various forms of care for children, handicapped, and senior citizens (Barr, 2004, p. 7; Lindbeck, 2006, p. 2). To summarize, Figure 1 (see p. 12) depicts the structure of the welfare state outlined above. In a broader sense, the welfare state also involves, besides the two previously explained types of public welfare provision, “price regulation (such as rent control and agricultural price support), housing policies, regulation of the work environment,

12

2 Theoretical Foundations of the Welfare State

"(D>K %>I F$ACIA6

E$>D$ F(D($

" ID> "($ (I"A

$A$I(F IAKIACF C D(I"A

" ID> DFFIF(DA $

B" FIA6

" ID> IAF DA $

?$(I$$A( #A "$KF "(

1A$>"$A(

B$D>( D$

/DI> D>>"DA $F

Figure 1: Welfare State and Social Protection Source: From Pestieau (2006), p. 4.

job-security legislation, and environmental policies” (Lindbeck, 2006, p. 2). Finally, in a general sense, the welfare state also refers to “(1) a particular form of state, (2) a distinctive form of polity or (3) a specific type of society” (Pierson, 1998, p. 7). Throughout this book all three definitions of the welfare state will be taken into account in order to study the variations of existing welfare states and their distinctive influence on formal economies and societies.

2.1.2

The Core Functions of the Welfare State

Although the efficiency-equity trade-off 11 has remained a recurrent theme in public policy debate (Pestieau, 2006, pp. 5-7), the welfare state has been intended to achieve 11

The efficiency-equity trade-off was first mentioned by Okun (1975) who stated that “the contrasts among American families in living standards and in material wealth reflect a system of rewards and penalties that is intended to encourage effort and channel it into socially productive activity. To the extent that the system succeeds, it generates an efficient economy. But that pursuit of efficiency necessarily creates inequalities. And hence society faces a tradeoff between equality and efficiency” (p. 1). Recently, however, several scholars (for example, Kenworthy, 1995; Pressman, 2005) could not find any empirical evidence for such a trade-off. For an elaboration on the difficulty of grasping the notion of the efficiency-equity tradeoff, see Le Grand (1990).

2.1 The Welfare State

13

both objectives simultaneously in the context of a capitalist system (Barr, 2004, p. 10).12 Following Lindbeck (2006), social institutions contribute to economic efficiency in three ways: Firstly, according to the microeconomic literature, government intervention is needed to correct for the following market failures with respect to private income insurance: “advantageous selection (‘cream skimming’) of insurance applicants, when insurance providers can differentiate between low-risk and high-risk individuals; adverse selection, when insurance providers cannot do so; myopia, when individuals underestimate their future income needs; and free riding on the altruism of others, when individuals expect others to assist them in case of economic distress” (p. 3). Social security legislation could surmount all of the above problems. Introducing mandatory income insurance would prevent problems of information asymmetry (between insurance providers and consumers) and undermine free-riding behavior by individual citizens. Furthermore, it would foster consumption smoothing over the lifecycle (p. 3; see Barr, 2001). Secondly, mandatory income insurance would also create a risk-sharing pool between different generations (p. 4). The latter can hardly be achieved “by voluntary contracts alone since the potential parties of such contracts may not live simultaneously – both when the contract is signed and when it is supposed to be fulfilled” (p. 4). Thirdly, a broad-based consensus exists in academia “that investment in human capital […] tends to be suboptimal without government intervention (in the form of subsidies or direct government provision) – either because of the difficulties in borrowing with expected future human capital as collateral, or because of unexploited (positive) externalities in connection with such investment” (p. 4). In this sense, public social spending could lead to higher rates of labor productivity and economic growth (p. 4). Besides promoting efficiency, the second objective of the welfare state has been to reduce insecurity on citizens’ behalf and create an equitable society (Barr, 2004, p. 10; Briggs, 2006, p. 16). However, perceptions of equity differ largely among political groups and market economies. Adhering to prior research (Bell, 1973), Flora & Heidenheimer (1981) identify two concepts of equality in the Western cultural tradition: The socialist ethic propagates equality of result, implying “an equalization in the disposal of resources, commodities, and services, a redistribution according to needs” (pp. 24-25). In contrast, the liberal ethic understands equality as equality of 12

For a comprehensive analysis of the core functions of the welfare state, see Barr (2004) and Barr (1992).

14

2 Theoretical Foundations of the Welfare State

opportunity. According to this view, equal opportunities (granted for example through the mode of public education) are more important than income equality among citizens (p. 25). With respect to equity issues, Barr (2004) contends that a central task of the welfare state should be to support living standards. In this sense, poverty relief and income insurance programs are a means to alleviate poverty and guarantee a minimum living standard, which in turn leads to social inclusion (p. 10, 12). In other words, the welfare state should function as a social safety net that protects people against “an unexpected and unacceptably large drop in […] [their] living standard” (Barr, 2004, p. 10). Furthermore, besides providing minimum insurance standards, the welfare state should also aim at achieving equity both vertically and horizontally. Vertical equity entails income redistribution geared towards citizens on the lower rungs of the income ladder. Progressive income taxation is one way to fulfill this goal. Horizontal equity refers to redistribution of income according to horizontal characteristics, such as age or family size (Barr, 2004, pp. 11-12). Although the objectives of the welfare state are commonly divided into efficiency and equity issues, it is worth noting that these categories often overlap. Indeed, by promoting solidarity income redistribution can lead to higher economic growth rates.13 Equity issues can therefore also be justified by efficiency considerations (Lindbeck, 2006, pp. 4-5). Yet as Siebert (2005) remarks, the nature of the association between efficiency and equity is dependent on the size of the welfare state. Precisely, he notes that the relationship between the two welfare state objectives follows a hump-shaped pattern. Accordingly, at lower levels of public social spending an expanding welfare state can raise economic efficiency; by contrast, in mature welfare states, an increase in equity normally entails a decline in efficiency (pp. 34-35). In this context, Buti et al. (1998), for instance, have highlighted the negative effects of overly generous public social institutions on labor market efficiency (pp. 19-22; see also Lindbeck, 2006, pp. 14-15). Against this backdrop, adequate welfare state institutions are a fundamental precondition to ensure the efficient working of the economy. In this context, the author concurs with Barr (2004) in considering a welfare state efficient if the following three criteria are fulfilled: Firstly, social “policy should avoid distortions leading to cost explosions” (macro-efficiency); secondly, social “policy should ensure the efficient division of total welfare-state resources between the different cash benefits” and benefits-in-kind (micro-efficiency); and finally, with respect to publicly funded institutions, “their finance and the structure of benefits should minimize adverse effects on labour supply, employment, and saving” (incentives) (p. 10). 13

In this context, Lindbeck (2006, p. 5) refers to Alesina & Rodrik (1994) who have found empirical support for this hypothesis.

2.2 Political Economic Theories on the Rise of the Welfare State

2.2

15

Political Economic Theories on the Rise of the Welfare State

Due to the complex nature of the topic as well as the vast academic literature that has piled up over recent decades, it is a truly challenging task to classify contemporary political economic theories with regards to the question why the welfare state was founded in the first place. Reviewing recent classifications by Esping-Andersen (1990/ 1998, pp. 13-18), Huber & Stephens (2001a, pp. 14-32, 41-46, 48-50), and Bonoli, George, and Taylor-Gooby (2000, pp. 9-12), four explanatory approaches appear to have dominated the academic debate: the logic of industrialism approach, the power resources approach, the state-centric approach, and the institutional approach.14 The first theoretical school, the logic of industrialism approach, considers economic progress, more precisely the process of industrialization, to be the main causal force behind the development of the welfare state (Wilensky, 1975, p. 47). Two central arguments are advanced in support of this view: Firstly, advocates emphasize the dramatic socio-economic changes that were induced by the transformation from a former agrarian into an urban-industrial society. With people moving from rural to urban areas, traditional structures of social protection (for example, extended tribal family, guild system) dissolved (see Esping-Andersen, 1990/ 1998, p. 13; Kerr, Dunlop, Harbison, and Myers, 1960/ 1966, p. 46); as a result, the state was called upon to assume a principal role as provider of social welfare. The process of industrialization, in turn, made rising public welfare expenditures possible since state income increased rapidly as a result of economic progress (Wilensky & Lebeaux, 1965/ 1968, p. 14). Secondly, with the emergence of industrial production work accidents no longer automatically conferred guilt on the employer, but were to be seen as an unavoidable side effect of the new economic era. In this sense, it became natural for employers to compensate the work force in case of industrial accident. Consequently, it does not appear surprising that social insurance for industrial accidents was the first welfare scheme to be established, followed subsequently by social insurance legislation for sickness, invalidity, old age and unemployment (Flora & Alber, 1981, pp. 50-51). Secondly, the power resources approach and the class mobilization thesis consider social classes as principal agents of change and social policy as a function of the balance of class power. This balance is primarily determined by the number of electoral votes, the share of seats in parliament, and the degree of collective bargaining (Esping-Andersen, 1990/ 1998, p. 16; Korpi, 2006, pp. 80-83). Advocates of the power resources approach therefore consider “welfare states as outcomes of, and are14

See Pierson (1998, Chapter 1-3) for a comprehensive overview of all existing welfare state theories.

16

2 Theoretical Foundations of the Welfare State

nas for, conflicts between class-related, socioeconomic interest groups and that in these distributive conflicts partisan politics is likely to matter” (Korpi & Palme, 2003, p. 425). Accordingly, they contend that welfare state expansion has largely been driven by social democratic governments whereas policies to contain public social spending have primarily been initiated by parties of the secular right political spectrum (Castles, 1982, p. 85; Korpi, 1989, p. 323). In this context, it is worthwhile noting that welfare-enhancing effects have also lately been attributed to Christian-Democratic government parties (Huber, Ragin, and Stephens, 1993, pp. 734-737; Huber & Stephens, 2000, p. 334). In line with this view, Esping-Andersen (1990/ 1998) concludes that the welfare state cannot be merely perceived as an institutional mechanism to compensate for market risks, but must also be viewed as a power resource itself. In social democratic political economies, left or labor parties – engaging in political alliances with other classes such as the peasants (Esping-Andersen, 1985/ 1988, pp. 37-38) – have pursued the development of the welfare state in order to make the labor movement less captive to the capital owner side, thereby strengthening their own position within the political arena (p. 16). Securing a critical power resource has therefore mainly depended “on the resources of contending forces, on the historical durability of its mobilization, and on patterns of power alliances” (EspingAndersen, 1990/ 1998, p. 16). The third theoretical school, the state-centric approach15, considers civil servants to be important actors in the development process of the welfare state. Generally, the following line of argument is stressed why modern bureaucracies – having emerged as a by-product of industrialization in order to administer the expanding state (Wilensky & Lebeaux, 1965/ 1968, p. 15) – have had a decisive impact on social-policy making: According to Heclo (1974/ 1975), “a great deal of policy development – its creation, alteration, or redirection – has been settled prior to or outside of substantial exercises of power” (p. 306). This is explained by the fact that political actors often lack knowledge on complex issues and therefore turn to civil servants – latter are assigned to “the task of gathering, coding, storing, and interpreting policy experience” (Heclo, 1974/ 1975, p. 303; see Kuhnle, 1996, pp. 251-253) – for policy advise (Heclo, 1974/ 1975, pp. 303-304).16 Being aware that state officials pursue their own career goals, directions for welfare state reform will therefore be promoted that are conducible to their interests (Skocpol, 1992, p. 42).17 15

See Skocpol (1985/ 1987, pp. 9-20) for a concise overview of the literature on state autonomy and capacity. 16 Against this backdrop, Heclo (1974/ 1975) concludes that only in very rare cases, “when analysis, deliberation, and persuasion have failed to achieve agreement […] [,] political power has been resorted to and sometimes proven decisive” (p. 306). 17 See Dahlström’s (2006) case study on Sweden for a thorough analysis of civil servants’ impact on the social policy making process.

2.2 Political Economic Theories on the Rise of the Welfare State

17

Apart from the three prior theoretical schools, various perspectives have emerged that stress the importance of a country’s politico-institutional setting for welfare state development. Adhering to Esping-Andersen’s (1990/ 1998, pp. 14-16) overview of the institutional approach, two variants – deemed critical for the understanding of subsequent analyses – are explained in more detail: The first perspective views the cultivation of the welfare state as a direct response to international economic pressures. This hypothesis has primarily originated in the works of Cameron (1978), Katzenstein (1985), and Ruggie (1982). Ruggie (1982) – adhering to Polanyi’s (1944) ground-breaking work The Great Transformation – has noted that the post-World War Two (WWII) international economic order had been built on the concept of embedded liberalism (pp. 392-393), the latter implying “a grand domestic bargain: societies were asked to embrace the change and dislocation attending international liberalization, but the state promised to cushion those effects by means of its newly acquired domestic economic and social policy roles. Unlike the economic nationalism of the thirties, then, the postwar international economic order was designed to be multilateral in character. But unlike the laissez-faire liberalism of the gold standard and free trade, its multilateralism was predicated on the interventionist character of the modern capitalist state” (Ruggie, 1998, pp. 89-90; see Ruggie, 1982, pp. 392-393, 399). In this context, compensating citizens for increasing risks related to international economic integration was perceived as a means to remove their resistance against further liberalization (see Ruggie, 2003, pp. 93-94). Looking from a different angle, Cameron (1978) has found that the positive correlation between the degree of trade openness and the size of the public economy has been most pronounced in small open economies. Latter has been explained in the following way: Given their small size and high dependence on international markets, these countries could not enforce neo-mercantilist or protectionist measures to shelter their economy from international competitive pressures. As a consequence, they have developed corporatist arrangements that fostered the creation of a large public sector in order to cushion the negative effects of volatile international markets (pp. 1255-1258, 1260; see also Katzenstein, 1985, p. 9, 24). Taking a similar view, Katzenstein (2003) concludes that the “perceived vulnerability generated an ideology of social partnership that had acted like a glue for the corporatist politics of the small European states” (p. 11). The second perspective, the democracy thesis, zeroes in on the relationship between democracy and the development of the welfare state. As noted by EspingAndersen (1990/ 1998), classical forerunners were of the opinion that democratic majorities would always support the welfare state. Following this line of argumentation, a modern variant of this thesis takes the view that the rapid growth of the welfare state is the logical outcome of democracy and its inherent electoral competition among parties to attract the median voter (p. 15). Against this backdrop, Persson & Tabellini (2000), for instance, have developed a median voter model of pensions,

18

2 Theoretical Foundations of the Welfare State

which “predicts that pensions per retiree will be higher, the higher the weight on old voters […], as this shifts the median-voter equilibrium toward a more generous pension system” (p. 130; see Pampel & Williamson, 1992, pp. 165-166).

2.3

Welfare State Regimes and the Classification of Advanced Capitalist Welfare States

Parallel to the development of political economic theories on the rise of the welfare state, several scholars began to lay the focus on existing variations between national systems of social protection. Amongst all, Esping-Andersen’s (1990/ 1998, 2000) Three Worlds of Welfare Capitalism approach has had the greatest influence on comparative welfare state research.18 Before analyzing this approach, two classifications are worth mentioning as they provide the basis: Wilensky & Lebeaux (1965/ 1968) – being one of the first to shed light on the institutional configuration of welfare states – distinguished two concepts of social welfare: “the residual and the institutional. The first holds that social welfare institutions should come into play only when the normal structures of supply, the family and the market, break down. The second, in contrast, sees the welfare services as normal, ‘first line’ functions of modern industrial society” (p. 138). Building on Wilensky & Lebeaux’s (1965/ 1968) prior work, Titmuss (1974) extended their classificatory framework by a third conception of social welfare termed the industrial achievement-performance model of social policy. According to his view, the latter, being situated in the middle between the residual welfare and the institutional redistributive model of social policy, “incorporates a significant role for social welfare institutions as adjuncts of the economy. It holds that social needs should be met on the basis of merit, work performance and productivity” (p. 31).19 Having briefly elaborated on the theoretical predecessors of Esping-Andersen’s welfare state regime approach, the latter will be now presented in detail below.

2.3.1

Esping-Andersen’s Welfare State Regime Approach

Central to Esping-Andersen’s (1990/ 1998, 2000) classification approach is the concept of welfare state regimes, the latter referring “to the ways in which welfare pro18

For a similar, but more comprehensive analysis of Esping-Andersen’s (1990/ 1998, 2000) welfare state regime approach and an overview of alternative welfare state typologies, see Bonoli et al. (2000, pp. 12-22), Ebbinghaus & Manow (2001b, pp. 7-10), Ellison (2006, pp. 10-22), and Arts & Gelissen (2002). 19 See also Furniss & Tilton (1977) who have developed an alternative classificatory approach which distinguishes between the following three types of welfare states: the positive state, the social security state, and the social welfare state (pp.14-20).

2.3 Welfare State Regimes and the Classification of Advanced Capitalist Welfare States

19

duction is allocated between state, market, and households” (Esping-Andersen, 2000, p. 73). In light of the sheer complexity of the welfare state, Esping-Andersen (1990/ 1998, 2000) focuses on three key defining dimensions when operationalizing the concept of welfare state regimes. Latter involve the level of de-commodification, the modes of social stratification, and the degree of de-familialization.20 2.3.1.1 The Welfare State Regimes and De-Commodification Building on Marshall’s (1950) concept of social citizenship21, Esping-Andersen (1990/ 1998) considers the process of “de-commodification of the status of individuals vis-à-vis the market” as a fundamental precondition for the development of the modern welfare state (p. 21). Originally, the rise of capitalism at the beginning of the Industrial Age was accompanied by a process of commodification of the working population. Human welfare, in other words, was at the mercy of the raw cash-nexus of the market. With the implementation of social insurance legislation starting in the last decades of the 19th century, the process of commodification was gradually reversed. Individual welfare ceased to depend exclusively upon the market since the state began in part to provide insurance programs by virtue of citizenship (EspingAndersen, 1990/ 1998, pp. 21-22; see Pierson, 1998, p. 103). While this development ultimately occurred in all industrialized countries, the degree of de-commodification varies today considerably between advanced capitalist welfare states. Using a decommodification index22 that takes into account the range of and conditions for social benefit entitlements as well as the levels of income replacement, Esping-Andersen (1990/ 1998) identifies three types of welfare systems: Firstly, a liberal system in which the market mechanism is strengthened by only granting social assistance on a needs-tested basis (p. 22, 53-54); secondly, a conservative system in which social benefits are almost entirely related to employment-based contributions (p. 22, 53); finally, a social democratic system in which social rights are granted on a universalistic basis by virtue of citizenship. This system comes closely to full de-commodification since social benefits nearly reach income levels (p. 23, 53). 20

It is worthwhile noting that in his groundbreaking book The Three Worlds of Welfare Capitalism Esping-Andersen’s (1990/ 1998) focus was primarily on the first two dimensions of the welfare state regime (pp. 21-26, Chapter 2 and 3). The third dimension, de-familialization, was introduced by Esping-Andersen (2000, Chapter 4). 21 Marshall (2006) defines citizenship as “a status bestowed on those who are full members of a community. All who possess the status are equal with respect to the rights and duties with which the status is endowed” (p. 34). Unlike in the Poor Law period, when people in need of social protection had to give up their citizenship in exchange for poverty relief (p. 33; Pierson, 1998, p. 103; see Chapter 4.1 in this book), Marshall’s concept of social citizenship envisages the inclusion of social rights in the status of citizenship (Marshall, 2006, p. 37). 22 See Esping-Andersen (1990/ 1998, pp. 47-54) for a detailed presentation of the components as well as the scoring procedure of the de-commodification index.

20

2 Theoretical Foundations of the Welfare State

2.3.1.2 The Welfare State Regimes and Social Stratification Although the welfare state has been founded to correct adverse market outcomes and nurture equality through redistributive power, “it is, in its own right, a system of stratification. It is an active force in the ordering of social relations” (Esping-Andersen, 1990/ 1998, p. 23). Due to different path dependences, advanced capitalist welfare states vary today considerably with respect to the degree of status differentiation within their respective societies. Using various indices (corporatism, etatism, meanstested poor relief, average benefit equality, and so forth)23, Esping-Andersen (1990/ 1998) identifies three dominant systems of social stratification across advanced capitalist welfare states: Firstly, the two-tiered liberal system in which minimum social insurance is provided on a universalistic basis by the state, whereas comprehensive welfare is furnished by the market (p. 26); secondly, the conservative system which is characterized by a high differentiation of program benefits across different classes. The objective of this corporatist, earnings-related social insurance structure has been to guarantee the perpetuation of power in the hands of the monarchy or the ruling class by means of consolidating class differences and endowing loyal groups (such as the civil service) with privileged welfare provision (pp. 24-25); finally, the social democratic system in which on top of the minimum social insurance scheme (as explained in the case of the liberal system) an earnings-related, universally inclusive public welfare program is provided (p. 26). 2.3.1.3 The Welfare State Regimes and De-Familialization Moving beyond the analysis of the state – market mix of welfare provision, EspingAndersen (2000) has recently introduced the degree of de-familialization as the third key-defining dimension of his welfare state regime concept in order to take into account the role of the family in the wider welfare context. In his eyes, “a familialistic welfare regime is […] one that assigns a maximum of welfare obligations to the household” (p. 45). Accordingly, the term “de-familialization” is applied “to capture policies that lessen individuals’ reliance on the family; that maximize individuals’ command of economic resources independently of familial or conjugal reciprocities. It is, like the concept of de-commodification, empirically more a matter of degree than of an ‘either-or’” (p. 45). Having this said, de-familialization is a process that can be initiated via two channels according to Esping-Andersen (2000): Firstly, the welfare state can provide services (for example, child and old age care, and public education) and subsidies to families (p. 61). Secondly, the market can supply services such as child care centers or nursing homes. The market channel, however, is only efficient if a low-wage service sector is existent; otherwise, most families will not opt 23

See Esping-Andersen (1990/ 1998, pp.70-71, 77-78) for a detailed presentation of the different indices of social stratification.

2.3 Welfare State Regimes and the Classification of Advanced Capitalist Welfare States

21

to demand these services due to budget constraints (pp. 63-64). Against this backdrop, Esping-Andersen (2000) applies various de-familialization indicators (for example, public spending on family services, public daycare coverage, and so forth), familialism indicators (for example, average weekly unpaid hours of women, percentage of aged people living with their children, and so forth), and market indicators (for example, rates of private daycare for the under-3-year olds) and identifies once again three types of welfare systems: Firstly, the social democratic system is characterized by high levels of de-familialization. Traditional household services are provided to a great extent by the state (pp. 61-64, 66). Consequently, female labor force participation rates are high (p. 45). Secondly, the conservative system is dominated by the notion of the male bread-winner model. With family-oriented services provided only to a minor degree by the state and the market, the family remains the central source of social protection (pp. 61-66). As a result, female labor force participation tends to be low (pp. 45-46). Finally, in the liberal system the market rather than the state or the family is considered to be the main provider of social care (pp. 61-65). However, given the fact that family subsidies are often low, a great number of households cannot afford private market care. Consequently, mothers in poorer households are forced to stay at home instead of entering the labor market (pp. 66-67). De-familialization is therefore much less advanced in the liberal compared to the social democratic system.

2.3.2

Classification of Advanced Capitalist Welfare States

Analyzing advanced capitalist welfare states along the three prior key defining dimensions, Esping-Andersen (1990/ 1998, 2000) has identified three types of welfare state regimes: a liberal, a conservative, and a social democratic regime. Since these regimes are ideal types of social protection systems, existing welfare states do not entirely mirror one specific welfare state regime. Esping-Andersen (1990/ 1998) therefore clusters welfare states loosely around, rather than rigidly assigning them to a specific regime type (p. 26; Esping-Andersen, 2000, p. 86). In the liberal cluster (the category of welfare states grouped around the liberal welfare state regime) Anglo-Saxon countries are predominant. Continental European countries such as Austria, France, and Germany, by contrast, revolve around the conservative regime type. Finally, the Nordic countries are considered to adhere closest to the social democratic welfare state regime type (Esping-Andersen, 1990/ 1998, pp. 26-29; Esping-Andersen, 2000, pp. 74-86). Table 1 (see p. 22) summarizes the main characteristics of each welfare state regime. Based on these theoretical groundings, Table 2 (see p. 23) provides a classification of 17 advanced capitalist welfare states for the year 1980. Relying to a great extent on Esping-Andersen’s (1990/ 1998, 2000) welfare state regime indicators, the

22

2 Theoretical Foundations of the Welfare State

Table 1: A Summary Overview of Regime Characteristics Liberal

Social democratic

Conservative

Marginal Central Marginal

Marginal Marginal Central

Central Marginal Subsidiary

Dominant mode of solidarity

Individual

Universal

Kinship Corporatism Etatism

Dominant locus of solidarity

Market

State

Family

Degree of decommodification

Minimal

Maximum

High (for breadwinner)

USA

Sweden

Germany Italy

Role of: Family Market State Welfare state:

Modal Examples Source: From Esping-Andersen (2000, p. 85).

author assesses each welfare state according to the level of de-commodification, the modes of social stratification, and the degree of de-familialization.24 Largely in line with Esping-Andersen’s (1990/ 1998, pp. 26-29; 2000, pp. 74-86) prior findings, three welfare state clusters emerge: Firstly, the Nordic countries (Denmark, Finland, Norway, and Sweden) form the social democratic group of welfare states which is shaped by a universally inclusive system of social protection. As mentioned above, the central characteristics involve high levels of de-commodification and de-familialization coupled with relatively low degrees of status differentiation. Secondly, the Continental European countries (Austria, Belgium, France, Germany, Italy, the Netherlands, and Switzerland25) and Japan cluster around the conservative welfare state regime. While they feature lower de-commodification and de-familialization scores than reported for Northern Europe, corporatism and etatism is much more pronounced than in the rest of the OECD world. Thirdly, the Anglo-Saxon countries 24

For similar analyses in the area of comparative welfare state research, see Huber & Stephens (2001a, pp. 85-105), Huber & Stephens (2001b), Kitschelt, Lange, Marks, and Stephens (1999/ 2005, pp. 429-438), and Swank (2003, pp. 60-72). 25 Although Switzerland does not appear to be the typical case of a conservative welfare state it is assigned to this group due to its relatively high level of de-commodification.

De-commodification

31.1 32.4 27.5 27.7 24.1 27.1 32.4 29.8 29.0

13 22 23.3 23.4 13.8 19.1

Austria Belgium France Germany Italy Japan Netherlands Switzerland Conservative average

Australia Canada Ireland UK USA Liberal average

5 7.7 6.7 8.5 7 7.0

11.9 15 12 8.5 9.6 10.5 10.8 9 10.9

15 14 14.9 17 15.2

(1980)

4 6.3 8.3 7.7 0a 5.3

12.5 8.8 9.2 11.3 9.4 6.8 10.5 12 10.1

15 10 14 15 13.5

(1980)

4 8 8.3 7.2 7.2 6.9

6.7 8.6 6.3 7.9 5.1 5 11.1 8.8 7.4

8.1 5.2 9.4 7.1 7.5

(1980)

Unemployment4

1 2 1 2 2 1.6

7 5 10 6 12 7 3 2 6.5

2 4 4 2 3.0

(1980)

0.7 0.2 2.2 2 1.5 1.3

3.8 3 3.1 2.2 2.2 0.9 1.8 1 2.3

1.1 2.5 0.9 1 1.4

(1980)

Etatism6

1.00 0.48 0.77 0.64 0.22 0.62

0.52 0.79 0.55 0.56 0.52 0.32 0.57 0.48 0.54

0.99 0.72 0.69 0.82 0.81

(1980)

Ave. Benefit Equality7

Social Stratification Corporatism5

42.4 56.9 30.6 42.5 54 45.3

1.5 0.6 1.6 1.9 0.5 1.2

2.7 2.3 2.8 1.7 1 0.4 1.7 1 1.7

3.3 3.2 2.8 4.5 3.5

..b 64.3 ..b 85 74.7 50.8 64.4 61.3 41.4 45.3 37.2 37 33.5 46.4

Total Public Family Expenditures (1990)9

De-familialization Employment Rates for Mothers (1990)8

/x[6h__xg6d 0--0goo h?o0exoeg6]/x#6x_0)6x?#6xl 06x?#6xl-h__6xolx?#6xg6S66x0xg6d 0--0goo h?o0e/x.0 6IxUgh?6gx0-xloeSdUeg6l6exT

Kx

+x/xAP$/x P/xF6d 0--0goo h?o0exoeg6]x)o?#x6Shglx?0x0_gxhS6x6elo0ex0Sh-l/x.0 6IxUgh?6gx0-xloeSdUeg6l6exT

Kx

+x/ AK$/ /xF6d 0--0goo h?o0exoeg6]x)o?#x6Shglx?0xlo e6llx76e6o?x0Sh-lxThIx0Sh-xe0ed6]ol?6e?xhegx?#6606xl 06gxK$/x.0 6IxUgh?6gx0-xloeSdUeg6l6exT

Kx

+x/xAK$/ m/xF6d 0--0goo h?o0exoeg6]x)o?#x6Shglx?0xe6-_0,-6e?x76e6o?x0Sh-l/x.0 6IxUgh?6gx0-xloeSdUeg6l6exT

Kx

+x/xAK$/ A/x#6xg6S66x0 00h?ol-xol -6hl6gxhlxe-76x0x-hC0x0 h?o0eh__,xgol?oe ?x7_o x6elo0exl #6-6l/x.0 6IxUgh?6gx0-xloeSdUeg6l6exT

Kx

+x/x\Kd\$/x /x?h?ol-xol -6hl6gxhlx6]6ego?6x0ex6elo0elx?0 S0[6e-6e?x6-_0,66lxhlx6 6e?hS6x0xFD/x.0 6IxUgh?6gx0-xloeSdUeg6l6exT

Kx

+x/x\Kd\$/ \/xU[6hS6x76e6o?x6Lh_o?,xolxg6oe6gxhlxh[6hS6xgo66e?oh_x76?)66ex7hlo xhegx-h]o--xl0 oh_x76e6o?lx0xlo e6llxe6-_0,-6e?xhegx6elo0elxT7hl6gx0e e6?xh?6d?h]x76e6o?l$/x%6e6o?xgo66e?oh_lxh6x7hl6gx0ex?#6x h?o0x0 Shhe?66gx7hlo xl0 oh_x76e6o?x?0x?#6x_6Sh_x-h]o--x76e6o?x0llo7_6xoex?#6xl,l?6-/x.0 6IxUgh?6gx0-xloeSdUeg6l6exT

Kx

+x/x\Kd\$/ +/x-_0,-6e?xh?6lx0x-0?#6lx)o?#x,0eS6l?x #o_gxhS6gxeg6x T7IxFh?hxe0?xh[ho_h7_6$/x.0 6IxFh?hx 0-o_6gx0-xtF TPKKA7xe/ /$/
/x0?h_x7_o xh-o_,x6]6ego?6lx-6hl6gxhlxh 6 6e?hS6x0xFD/x.0 6IxFh?hx 0-o_6gx0-xtF TPKKmhxe/ /$/

38.1 29.2 38.3 39.1 36.2

(1980)

Overall1 Pensions2 Sickness3

Denmark Finland Norway Sweden Social democratic ave.

Countries

Table 2: Classification of 17 Advanced Capitalist Welfare States

2.3 Welfare State Regimes and the Classification of Advanced Capitalist Welfare States

23

24

2 Theoretical Foundations of the Welfare State

(Australia, Canada, Ireland, UK, and the USA) constitute the liberal welfare state model. Given the modest universal provision of social protection, all indicators achieve much lower scores on average than in the previous two welfare state categories (compare Huber & Stephens, 2001a, pp. 87-93; Huber & Stephens, 2001b; Kitschelt et al., 1999/ 2005, p. 435; Swank, 2003, pp. 61-65). Following Esping-Andersen (1990/ 1998), the above outlined allocation of industrialized countries into three welfare state clusters also corresponds with the broader political-economic view underlying the power resources approach and the class mobilization thesis26: Accordingly, Anglo-Saxon countries, which most closely mirror the liberal welfare state regime, developed a residual welfare state model because the middle class could not be successfully captured by pro-welfarist political groups and their notion of an encompassing welfare state. As a consequence, the middle class in these countries has traditionally leaned politically toward the market rather than the state. The two-tiered welfare system (see above), in other words, is a reflection of these political processes (p. 31). Continental European countries, which primarily embody attributes of the conservative welfare state regime, have mainly been influenced by conservative political-economic forces (p. 31). Precisely, these countries have “institutionalized a middle-class loyalty to the preservation of both occupationally segregated social-insurance programs and, ultimately, to the political forces that brought them into being” (p. 32). Finally, the Nordic countries have been shaped by social democracy which pursued “to synthesize working-class and white-collar demands without sacrificing the commitment to solidarity” (p. 31). The newly formed working-class-white-collar middle class backed the construction of a universally inclusive welfare system, which reinforced social democratic power in the political landscape of the Nordic countries (p. 18, 31).

26

See Chapter 2.2 for a presentation of the power resources approach and the class mobilization thesis.

3

Worlds of Welfare Capitalism: Interlinking Welfare State and Production Regimes

The concept of welfare capitalism refers to the notion that the welfare state is inextricably linked to central pillars of the economy (Esping-Andersen, 1990/ 1998, p. 2).27 In this context, Esping-Andersen (1990/ 1998) has addressed the interrelationships that bind the welfare state with distinctive labor market institutions. Precisely, he has advanced the view that “social policy and the labor market have become interwoven and mutually interdependent institutions. To an extent, the welfare state has become a major agent of labor-market clearing. It eases the exit of women, with family programs; of older people, with early retirement. It upholds labor demand by employing people in health, education, and welfare. It facilitates female labor-supply by providing necessary social services. It helps people reconcile their role as economic producers, social citizens, and family members by granting workers paid vacations and temporary absence from work” (p. 149).28 Recently, several researchers – most prominently Ebbinghaus & Manow (2001a), Huber & Stephens (2001a, pp. 85-105), Huber & Stephens (2001b), Kitschelt et al. (1999/ 2005, pp. 429-438), and Swank (2003, pp. 60-72) – have begun to move beyond the welfare state – labor market nexus to take a broader set of institutional complementarities between welfare state and production regimes into account. Before analyzing these linkages the author will give a brief introduction into existing varieties of production regimes. A special focus will hereby be laid on Hall & Soskice’s (2001/ 2004a, 2001/ 2004b) Varieties of Capitalism (VoC) approach.

3.1

Varieties of Production Regimes

Parallel to Esping-Andersen’s (1990/ 1998, 2000) welfare state regime analysis, researchers have recently started to classify capitalist economies in a comprehensive, systemic way. This so-called comparative capitalism (CC) literature centers hereby on the notion of production regimes.29 Following Soskice (1999/ 2005), 27

This chapter draws heavily on Ebbinghaus & Manow (2001b). To illustrate an exemplary dimension of the welfare state – labor market nexus, see Chapter 2.3.1.3 for an overview of how different welfare state systems enhance or hamper women’s participation in the labor market. 29 According to Soskice (1999/ 2005), the term production regime was originally coined by Hollingsworth, Schmitter, and Streeck (1994). 28

26

3 Worlds of Welfare Capitalism

“by a production regime is meant the organization of production through markets and market-related institutions. It analyzes the ways in which the microagents of capitalist systems – companies, customers, employees, owners of capital – organize and structure their interrelationships, within a framework of incentives and constraints […] set by a range of market-related institutions within which the microagents are embedded. These framework incentives and constraints are sometimes summarized as the ‘institutional framework’ of the production side of the economy” (pp. 101-102). Although the CC literature has rapidly expanded in recent years, all classificatory approaches adhere to the following theoretical cornerstones (Jackson & Deeg, 2006): The institutional subsystems constituting each national production regime are strategic complements that give rise to country-specific institutional configurations (p. 6).30 Against this backdrop, each political economy is considered to contain a number of distinct comparative institutional advantages and disadvantages for specific types of economic transactions (p. 6). The existence of these institutional strengths and weaknesses, in turn, suggests a path-dependent development of the overall capitalist economy. Consequently, each production regime is expected to accommodate and respond to common pressures in distinct ways (p. 6). Table 3 gives an overview of the principle institutional domains of a national production system and depicts representative typologies for each of these subsystems.31 Of all theoretical concepts prevalent in the CC literature, Hall & Soskice’s (2001/ 2004b) firm-centered VoC approach has been most influential in recent academic debates. Placing the firm and its relationships with internal (for example, employees) and external actors (for example, suppliers, clients, stakeholders, trade unions, business associations, governments) at the core of the analysis (p. 6), the VoC approach identifies two ideal types of production regime: Firstly, liberal market economies shaped by firms, which “coordinate their activities primarily via hierarchies and competitive market arrangements” (p. 8); secondly, coordinated market economies characterized by firms, which engage mainly in “non-market relationships to coordinate their endeavors with other actors and to construct their core competencies” (p. 8). In both production regimes, firms are hereby considered to coordinate their activities across five interdependent institutional sub-spheres of a political economy: vocational training and education, industrial relations, corporate governance, inter-firm 30

In this context, Jackson & Deeg (2006) note that “unlike in neoclassical economics, the emerging literature on institutional economics and economic sociology was coming to understand efficiency in terms of strategic complementarities among organizational elements, which led to the argument that successful capitalism came in more than one variety (Milgrom/ Roberts 1994; Sorge/ Warner 1986; Aoki 1984)” (p. 5). 31 See also Ebbinghaus & Manow (2001b, pp. 3-7) for an overview of the literature on varieties of capitalism.

27

3.1 Varieties of Production Regimes

Table 3: Selected Institutional Domains in Capitalist Economies 4el?o??o0eh_xg0-hoe

c66l6e?h?o[6x?,0_0S,

.6_6 ?6gx6]h-_6lx0x60-he 6x66 ?l

Goehe oh_xl,l?6-l

%hedx-h6?d7hl6g

4e[6l?-6e?xh??6ex 00h?6xS0[6ehe 6

t00h?6xS0[6ehe 6

4elog6x0?log6x0 l#h6#0_gd 6xl?h6#0_g6

Go-xl?h?6S,xoe 0-6xgol?o7?o0ex lo__lxoe[6l?-6e?xoee0[h?o0e

4e?6do-x6_h?o0el

U-"lx_6eS?#x6_h?o0ex 07_oSh?o0eh_x6_h?o0e

t006h?o0exhegx 0-6?o?o0ex 00h?6x S0[6ehe 6xoee0[h?o0e

4egl?oh_x6_h?o0el

t0e_o ?hohex_h_ol?x 00h?ol?

4e?6eh_x[l/x6]?6eh_x_h70x-h6?x _6]o7o_o?,x)hS6x_6[6_lxe6-_0,-6e?

.o__x 6h?o0e

4e 0-6xgol?o7?o0ex)0x0Sheoh?o0ex .?h?6xhll0 oh?o0elx-h6?xo-l oee0[h?o0exoegl?oh_x6_h?o0elxo-x l?h?6S,

k0x0Sheoh?o0e

G0gol-x_6]o7_6xl6 oh_oh?o0ex %loe6llxl?h?6S,xoegl?oh_x6_h?o0el go[6loo6gxLh_o?,x0g ?o0e

k6_h6xl?h?6l

!o76h_x 0el6[h?o[6xl0 oh_x g6-0 h?o

!h70x-h6?xh?o oh?o0exh??6elx0 lh[oeSlxhegxoe[6l?-6e?x0Sheoh?o0ex0 _h70xeo0el

4ee0[h?o0e

chgo h_xoe 6-6e?h_

k0x0Sheoh?o0ex7loe6llxl?h?6S,x 6-_0,-6e?

Source: From Jackson & Deeg (2006, p. 13).

relations, and intra-firm relations (pp. 6-7).32 In line with the CC literature, Hall & Soskice (2001/ 2004b) hold the view that the mechanism of coordination is not randomly applied, but follows a capitalist system inherent logic that can be subsumed under the concept of institutional complementarities (p. 18). Adhering to Aoki (1994), the latter term is defined by Hall & Soskice (2001/ 2004b) in the following way: “two institutions [in different spheres of a political economy] can be said to be complementary if the presence (or efficiency) of one increases the returns from (or efficiency of) the other” (p. 17).33 An often stated example in this respect is the interrelation of vocational training with measures of employment protection: EstevezAbe, Iversen, and Soskice (2001/ 2004), for instance, demonstrate that employers pursuing product market strategies based on firm- or industry-specific skills are sup32

See Hall & Soskice (2001/ 2004b) for a detailed analysis of the five institutional sub-spheres constituting the coordinated production regime (pp. 21-27) and the liberal production regime (pp. 27-33). 33 See Hall & Gingerich (2004) for an overview of methods how to measure institutional complementarities empirically.

28

3 Worlds of Welfare Capitalism

portive of strong employment and/ or unemployment protection through legislation, collective agreements or company-based criteria (pp. 152-153; 165, 168, 170). In light of these institutional complementarities, Hall & Soskice (2001/ 2004b) note “that the institutional structure of a particular political economy provides firms with advantages for engaging in specific types of activities there. Firms can perform some types of activities, which allow them to produce some kinds of goods, more efficiently than others because of the institutional support they receive for those activities in the political economy, and the institutions relevant to these activities are not distributed evenly across nations” (p. 37). Consequently, they believe in the persistence of these two diverse patterns of production regime (Hall & Soskice, 2001/ 2004b, pp. 56-60).34 Table 4 summarizes the main characteristics of both varieties of capitalism. Table 4: Two Varieties of Capitalism: Uncoordinated vs Coordinated Market Economies

Do-6x6]h-_6l

Goehe oh_xhegx6 0e0-o x S0[6ehe 6

D0g ?o0exl,l?6-

>Q++SZQT5-S_(T-5_-+Q+('

[++SZQT5-S_(T-5_-+Q+('

J.UxJQ

6-he,x hhe

.#0?d?6-xoehe oh_x-h6?lx 6Lo?,xoehe oeSxTl#h6#0_g6x [h_6$x_o-o?6gx7loe6llx 00goeh?o0exhe?o?l?x_h)l

!0eSd?6-xh?o6e?x ho?h_xg67?x oehe oeSxTl?h6#0_g6x[h_6$x l?0eSx7loe6llxhll0 oh?o0elx oe?6d 0-he,xe6?)0l

!0)dlo__x0g ?o0ex-hllx 0g ?lxe-6o x_6]o7o_olh?o0e

foS#dlo__x0g ?o0ex#oS#d Lh_o?,x0g ?lx_6]o7_6x l6 oh_olh?o0e

nhehS6-6e?d_h70x6_h?o0el F6 6e?h_ol6gx7hShoeoeSx 0e?6e?o0lx)0_h 6x6_h?o0el hoeoeSxhegx6-_0,-6e?

6e6h_x6g h?o0exl#0?x?6e6x #oS#x?e0[6xhegxoe?6do-x -07o_o?,

t00goeh?6gx7hShoeoeSx l?h??0,x)06x 66l6e?h?o0e 0 h?o0eh_x?hoeoeSx_0eSx ?6e6x_0)x?e0[6xhegxoe?hd o-x-07o_o?,

Source: From Ebbinghaus & Manow (2001b, p. 6).

34

For a brief summary of the VoC approach, see also Casey (2004, pp. 6-11), Ebbinghaus & Manow (2001b, pp. 3-7), Jackson & Deeg (2006, pp. 21-24), and Streeck & Thelen (2005, p. 5). Streeck & Thelen (2005) have criticized the VoC approach for “providing a compelling account of observed institutional resiliency” (p. 5). According to their view, “the prevailing emphasis on institutional stability even in the face of indisputable and important change points to a general problem in contemporary institutional analysis, which has always emphasized structural constraints and continuity” (p. 6). Recently, Hall & Thelen (2009) have elaborated on processes of institutional change within the VoC perspective.

3.2 The Welfare State-Economy Nexus

3.2

29

The Welfare State-Economy Nexus

Following Esping-Andersen’s research on the relationship between welfare state regimes and labor markets, several scholars have begun to look in more detail at the overall welfare state – economy nexus. Attempting to bridge Esping-Andersen’s Three Worlds of Welfare Capitalism approach with Hall & Soskice’s VoC approach, Ebbinghaus & Manow (2001b) lay the focus on linkages between the welfare state regime and the following three institutional domains of the production regime: the system of industrial relations, the production system and employment regime, and the corporate governance and financial system (pp. 12-17). In the following, the author will briefly shed light on each of these three interrelationships.35 According to Brandl & Traxler (2005), the system of industrial relations and the welfare state are closely intertwined since both institutions “share the purpose of containing the risks arising from the ‘commodification’ of labour” (p. 635). Against this backdrop, it does not come at a surprise that the ascendance of the modern welfare state and the emergence of neocorporatist institutional arrangements – the latter referring to the interplay between the state, employer organizations, and unions – have been mutually reinforcing processes (Hicks, 1999, pp. 127-128). As noted by EspingAndersen (1990/ 1998), left or labor parties in social democratic political economies pursued the development of the welfare state in order to make the labor movement less captive to the capital owner side, thereby strengthening their own position within the political arena (p. 16). Simultaneously, Hicks (1999) remarks that the neocorporatist institutional setting became “a bridge from the working class and social insurance reformism of the first half of the twentieth century to the welfare state in the second half ” (p. 128). Nowadays, the interdependence of industrial relations and welfare state institutions has not lost relevance as Brandl & Traxler (2005) specify in the following: “The unemployment effects of collective bargaining have an impact on the welfare state insofar as unemployment benefits are provided by the state. Collective wage agreements even more directly affect the public welfare system, when the level of and change in social benefits is linked formally to wage movements. Conversely, the welfare state essentially influences industrial relations, when establishing social protection against the risks resulting from the labour market. For example, unemployment benefits enable unemployed persons to ‘wait’ and thus enhance the bargaining power of labour, because there is less incentive for the unemployed to undercut the given labour standards in these circumstances” (pp. 635-636). 35

A more detailed analysis of the welfare state – economy linkages is provided in the book Comparing Welfare Capitalism: Social policy and political economy in Europe, Japan and the USA edited by Ebbinghaus & Manow (2001a).

30

3 Worlds of Welfare Capitalism

In light of the institutional complementarity of the two institutional spheres, market economies with relatively weak (strong) systems of collective bargaining tend to be paralleled by residual (comprehensive, universalistic) welfare states and vice-versa (Esping-Andersen, 2000, p. 17; see Swank, 2003, pp. 61-65). Besides the above mentioned interrelationship, another dimension of the welfare state – economy nexus has recently featured prominently in academic literature. Precisely, several studies (for example, Estevez-Abe et al., 2001/ 2004; Mares, 2001/ 2004) have focused on the linkage of the welfare state with employees’ skill formation and firms’ product market strategies. Following Estevez-Abe et al. (2001/ 2004), individual workers decide to invest in firm-/ industry-specific or general skills depending on the degree of social protection in the labor market. This can be explained in the following way: Where protection of employment, unemployment or wages is low and therefore future job insecurity is high, individual workers will opt to invest in general, portable skills since these provide them with better chances to find a job in case of unemployment (pp. 149-150, 153). By contrast, where social protection in the labor market is high, individual workers will be more willing to invest in asset-specific skills (p. 150).36 Against this backdrop, the welfare state induced mode of skill formation can provide institutional advantages to certain firms (Mares, 2001/ 2004, p. 186). In this context, Estevez-Abe et al. (2001/ 2004) note the following: “For firms pursuing product market strategies which depend heavily on firmand industry-specific skills, promise of employment and unemployment security can thus provide a cost-effective path to improving the firms’ competitive position in international markets (cf. Ohashi and Tachibanaki 1998; Koike 1994)” (p. 152). Consequently, “the more a welfare production system emphasizes the creation of specific skills, the more likely it is that the median voter will be someone with considerable investments in specific skills, and the more likely it is that employers’ interest organizations will be dominated by firms pursuing specific skills strategies. Both will contribute to perpetuating institutions and policies that advantage firms and workers with heavy investments in specific skills” (p. 182).37 36

Precisely, Estevez-Abe et al. (2001/ 2004) note that high levels of unemployment protection (high replacement ratios and long benefit duration) enhance individual workers’ willingness to invest in industry-specific skills (p. 152); by contrast, adhering to Aoki (1988), the same authors remark that high degrees of employment protection (long job tenure) will foster individual workers’ acquisition of firm-specific skills (p. 150). 37 Although Mares (2001/ 2004) agrees that the skill profile of a firm has a considerable impact on how the stance of a firm towards social policy is, other corporate factors, in conjunction with the skill composition, need to be taken into account. These include, namely, the size of a particular firm and the incidence of labor market risk (for example, workplace accidents, (Continued at p. 31)

3.2 The Welfare State-Economy Nexus

31

Taking a different angle on the interrelationship between the welfare state and the production system as well as employment regime, several studies (for example, Iversen & Wren, 1998; Scharpf, 2001; Wren, 2001) have recently focused on the effects of welfare state institutions on employment ratios. As they demonstrate, adjustment strategies in response to challenges such as deindustrialization largely vary across welfare state clusters. The author will not further elaborate on this interdependence now since it will be dealt with in close detail in Chapter 5.1. Finally, the last major dimension of the welfare state – economy nexus centers on the interrelationship between welfare state institutions, particularly the pension regime, and the financial as well as corporate governance system. Following Jackson & Vitols (2001), two types of linkages can be distinguished: “Policy choices balancing the three pillars of pension regimes (public, occupational and individual) impact the supply side of national savings. The regulation of private pensions (second and third pillars) shape how savings are channelled into capital markets through investments in stocks, bonds, loans or internal company reserves. Conversely, pension regimes impact the demand side for different financial assets by shaping the personal sector’s portfolio distribution between bank deposits and securities” (p. 171). Castles (2004) distinguishes between three types of public pension regimes, which largely cluster developed countries in accordance with Esping-Andersen’s Three Worlds of Welfare Capitalism approach: Firstly, the majority of Continental European countries have institutionalized a Bismarckian-type pension regime, which provides income-related, contributory benefits (p. 130).38 Secondly, the Anglo-Saxon countries have implemented a Beveridgean-type pension regime, which grants flatrate, public pensions on a universal or means-tested basis (p. 130). Finally, the Nordic countries have developed a hybrid pension regime, which incorporates both Bismarckian and Beveridgean elements (pp. 130-131). Looking now in more detail at a representative country for the Bismarckian (Germany) as well as for the Beveridgean group (USA), Table 5 presents regime-specific characteristics and depicts to what extent income is derived from each of the three pension pillars (public, occupational, and individual). As indicated Germany has relied in the past heavily on the first pillar; the latter institutionalizing a pay-as-you-go 37

(Continued from p. 30) unemployment due to demand volatility) employees of this firm are confronted with (pp. 186-187, 193). Depending on the degree of all three factors, firms are expected to prefer either firm-level, contributory or universalistic social policy. See Mares (2001/ 2004, pp. 187-203) for the theoretical firm-centered model on the trade-off between risk redistribution and control. 38 Notable exemptions to this system within Continental Europe can only be found in the Netherlands and Switzerland, which largely provide flat, public pensions and rely to a great extent on funded systems (Börsch-Supan, 2001a, p. 1).

32

3 Worlds of Welfare Capitalism

pension system, in which both employers and employees finance current retirees’ pensions via contributions (Börsch-Supan, 2001b, p. 16). Firm pensions and private savings, by contrast, have only accounted for a small amount of total retirement income.39 Turning now to the United States, a different picture emerges: Although the Anglo-Saxon country has also institutionalized a pay-as-you-go pension scheme, its relevance for the overall system has been by far less critical since the vast amount of retirement income is derived from funded pensions constituting the second and third pillar.40 Table 5: The Continental European and the Anglo-Saxon Pension System in Comparison

t0e?oe6e?h_x06hex

UeS_0d.h]0ex

)T-_+

-(TQ'_



,Q+(-_ +(_ -T_:ZKKT_

)F_



P

,Q+(-_ +(_ -T_:ZKKT_P

.0 oh_x.6 o?,

mA

Go-x6elo0elx A

T-0l?_,xoe[6l?6gxoex



700x6l6[6l$

-?h_xegl$

ZS_

)exlh[oeSlx

)exlh[oeSlx

mP T0x

:ZKKT

T-hoe_,x)#0_6x_o6x

T-hoe_,xoego[ogh_x

)#o #xPA xolx

6?o6-6e?x

6heoeSlx

h 0e?$x0?#6x

0- 0_gdhS6x

oe 0-6/

6-_0,-6e?$

oelhe 6$x0?#6x oe 0-6/

K

Note: The figures are related to the combined income of a two-person retiree household. Source: (1) Adapted from Börsch-Supan (2001a, p. 4); (2) Adapted from Börsch-Supan (2001a, p. 5).

39

It is worthwhile noting however that a series of reforms have recently been implemented in Germany which have a considerable impact on the German pension system. For a detailed analysis of the German pension reform process see Börsch-Supan & Wilke (2005, pp. 5-16). 40 Although private savings account for a vast share of retirement income in the US (see Table 5), it is worthwhile noting that these funds tend to be largely concentrated within a small portion of the population. As remarked by Wise (2001), “many retirees have virtually no personal saving other than home equity and only about half receive employer-provided pension benefits” (p. 111).

3.3 Classification of Welfare Capitalist Systems

33

Besides these general differences in the public-private mix of pensions, two institutional distinctions between the German and the US pension system need to be further highlighted since they have significant implications for the organization of the financial and corporate governance system of the respective countries: Firstly, private savings (third pillar) in Germany have been mainly directed at banks and life insurance companies.41 This situation has stood in sharp contrast to the US, where private savings have mostly been channeled via pension funds into equity markets. As a consequence, Germany has developed a rather bank-centered financial system whereas the US epitomizes a capital market-based financial system (Jackson & Vitols, 2001, pp. 172-173, 185). Secondly, firm pensions (second pillar) in Germany tend to be “accumulated as internal reserves on the balance sheets of large firms and used as a means of internal company finance. […] thus, pension reserves increase the long-term financial autonomy of firms, reducing their dependence on external equity finance” (Jackson & Vitols, 2001, p. 184). In the UK and the US, by contrast, firm pensions “are largely accumulated and controlled by external pension funds. […] given their institutional trustee relation, pension funds pursue maximum shareholder returns in the interests of their members, subject to prudent rules concerning the diversification of assets” (Jackson & Vitols, 2001, p. 185). Consequently, as opposed to Germany, firm ownership tends to be largely dispersed and long-term voice-based relationships between investors and firms are rare (Vitols, 2001/ 2004, p. 345).

3.3

Classification of Welfare Capitalist Systems

Although the author has already classified advanced capitalist welfare states in Chapter 2.3.2 (see Table 2), a comprehensive analysis of welfare capitalist systems also necessitates a close look at the production side of the 17 leading welfare economies. Similar to prior analyses conducted by Huber & Stephens (2001a, pp. 85-105), Huber & Stephens (2001b), Kitschelt et al. (1999/ 2005, pp. 429-438), and Swank (2003, pp. 60-72), the author assesses these countries based on principal production regime indicators. Precisely, in line with Hall & Soskice’s (2001/ 2004a, 2001/ 2004b) VoC approach, the following five institutional domains of the political economy are the central focus of Table 6: industrial relations, corporate governance, inter-firm relations, intra-firm relations, and vocational training. At first sight, the nine production regime 41

According to Jackson & Vitols (2001), “pension funds and life insurers [in Germany] invest in equities below the legal maximums. However, German equities have not underperformed. But the small differences in returns between German bonds and stocks may not be large enough to justify the extra risk involved” (p. 187).

34

3 Worlds of Welfare Capitalism

indicators42 reflect the distinction proposed by the VoC approach, namely two types of market economies: Firstly, the Anglo-Saxon countries cluster around the liberal production regime which is primarily characterized by a strong market mechanism. Accordingly, institutional features of non-market coordination, such as the system of industrial relations, are underdeveloped. Furthermore, in light of a capital markets oriented financial and corporate governance system, long-term voice-based relationships between firms and their investors are rare. The same holds true for intra-firm relationships as demonstrated by the indicators on long-term employment security and median length of enterprise tenure. Consequently, firms pursue production strategies that seldom require vocational training on behalf of their employees (see Hall & Soskice, 2001/ 2004b, pp. 27-33; Huber & Stephens, 2001a, pp. 93-105; Kitschelt et al. 1999/ 2005, p. 435; Swank, 2003, p. 68). Secondly, Japan and the Continental European as well as the Nordic countries cluster around the coordinated production regime which relies less on the pure market as a mechanism of coordination. Instead, firms engage mainly in “non-market relationships to coordinate their endeavors with other actors and to construct their core competencies” (Hall & Soskice, 2001/ 2004b, p. 8; see Huber & Stephens, 2001a, pp. 93-105; Kitschelt et al., 1999/ 2005, p. 435; Swank, 2003, pp. 66-68). Consequently, all indicators on non-market oriented coordination achieve much higher scores on average than in the Anglo-Saxon category. By contrast, given the fact that the corporate governance system tends to be bank-centered, stock-market capitalization relative to GDP is fairly low (see Hall & Soskice, 2001/ 2004b, pp. 21-27). Finally, another striking difference is observable with regards to vocational training. As depicted in Table 6, the level of this indicator is close to five times as high as the level reported for Anglo-Saxon countries. This empirical evidence substantiates Estevez-Abe et al.’s (2001/ 2004) argument that firms in coordinated market economies pursue production strategies that require a firm- or industry-specific skill profile on behalf of their employees (pp. 169-172; see Hall & Soskice, 2001/ 2004b, pp. 24-25).43 Although the Nordic and Continental European countries are both assigned to the coordinated market economy cluster two distinctions in the area of industrial rela42

The majority of these production regime indicators have previously been used by Swank (2003) in order to conduct a principal components analysis testing in how far these indicators load on the following two dimensions: national coordination-corporatism and sector-group coordination (pp. 65-66). He finds that Nordic market economies tend to be nationally coordinated whereas their Continental European counterparts appear to be sector-coordinated (pp. 62-63, 66-68). Anglo-Saxon countries, by contrast, seem to be generally uncoordinated (p. 68). See also the analysis by Huber & Stephens (2001a, p. 100, 102-103) who have relied on a coordination index that consists of similar production regime indicators. 43 Following Estevez-Abe et al. (2001/ 2004), “the figure for Japan is only 16 per cent, but much training in this country goes on in large companies and is not recorded in the data” (p. 171).

75.3 72.5 58.5 80.8 71.8 46.9 53.9 10.1 31.2 38.8 25.4 24.3 24.3 31.9 40.5 32.9 51.1 39.3 15.5 35.9

0 1 1 1 0.8 1 0.5 0 1 0 1 0.5 0.5 0.6 1 0 0.5 0 0 0.3

0.30 0.20 0.22 0.46 0.30 0.11 0.36 0.28 0.22 0.15 1.22 0.48 0.71 0.44 0.41 0.47 …a 0.85 0.58 0.58

Stock market capitaliz. as % GDP (1990)3 1 1 1 1 1.0 1 1 1 1 1 1 0 1 0.9 0 0 0 0 0 0.0

Long-term voice-based relationship (1990)4

Corporate Governance

0.5 1 0.5 0.5 0.6 0.5 0.5 0.5 1 1 1 0.5 0.5 0.7 0.5 0.5 0.5 0.5 0.5 0.5

Alliances among competing firms (1990)5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 1 1 0.5 0.5 0.6 0.5 0.5 0.5 0.5 0.5 0.5

Long-term voice-based relationship (1990)6

Intra-Firm Relationships

0.5 1 0.5 1 0.8 1 0.5 0 0.5 0.5 1 0.5 0.5 0.6 0 0 0 0 0 0.0

Long-term employment security (1990)7 4.4 7.8 6.5b 7.8 6.6 6.9 8.4 7.7 10.7 8.9 8.3 5.5 6 7.8 3.4 5.9 5.3 5 4.2 4.8

31 32 37 36 34.0 22 53 28 34 35 16 43 23 31.8 9 5 6 11 3 6.8

Vocational Training Median Vocational length of train tenure share (1995)8 (1990s)9

Industrial Relations

.0 6Ix.66xU6ego]x/


/xG0__0)oeSxl?6[6dU76x6?xh_/xTPKKxPKKm$x?#6x[0 h?o0eh_x?hoeoeSxl#h6xolxg6oe6gxhlx?#6xl#h6x0xhe hS6x 0#0?xoe 6o?#6xl6 0egh,x0x0l?dl6 0egh,xT4.tFA$x[0 h?o0eh_x?hoeoeSxT/x\$/

+/xn6gohex_6eS?#x0x6e?6ol6x?6e6xoe ,6hlxT7IxFh?hx0x

$/

m/x!0eSd?6-x[0o 6d7hl6gx6_h?o0el#olx76?)66exo-lxhegx?#6oxoe[6l?0lx _hlloo6gxhlx0__0)oeSIxx _hS6xoe[6l?0lx#0_gxloSeoo he?x0)e6l#oxl#h6lx0x_0eSx6o0glx/A  6_h?o[6_,xg6 6e?h_o6gx0)e6l#ox7?x)o?#x0e_,x-0g6h?6x oe[6l?0x?e0[6xK  g6 6e?h_o6gx0)e6l#ox)o?#xh #oS#x?e0[6xh?6xTf76x6?xh_/ PKKm7x/x+$/ A/xU__ohe 6lxh-0eSx 0-6?oeSxo-lx0xcRFxhegx0?#6x 00h?6xh ?o[o?o6lx _hlloo6gxhlx0__0)oeSIx  6]?6elo[6xl6x0xh__ohe 6lx0?6exoe[0_[oeSx-06x?#hex?)0xo-lx/A  -0g6h?6xl6x0xh__ohe 6lx7,x-he,xo-lx0x6]?6elo[6xl6x 7,xl0-6xo-lxK  oe6L6e?xl6x0 h__ohe 6lxTf76x6?xh_/xPKKm7x/x+$/ /x!0eSd?6-x[0o 6d7hl6gx6_h?o0el#olx76?)66ex #hl6xhegxl_o6xo-lx _hlloo6gxhlx0__0)oeSIxx 6]?6elo[6xl6x0 #oS#_,x 006h?o[6xl_o6xh?e6l#olx7,x-he,x_hS6xo-lx/A  l6x0x-0g6h?6_,x 006h?o[6xl_o6x h?e6l#olx7,x-he,xo-lx0x0x#oS#_,x 006h?o[6xh?e6l#olx7, l0-6xo-lxK  oe6L6e?xl6x0 l_o6xh?e6l#olxTf76x6? h_/ PKKm7x/ +$/ \/x!0eSd?6-x6-_0,-6e?xl6 o?,xShhe?66gx7, o-lx _hlloo6gxhlx0__0)oeSIx  _0eSd?6-xToexl0-6x hl6lx_o6?o-6$x6-_0,-6e?xl6 o?,x 0--0exoe _hS6xo-lx/A  l0-6xo-lx0[og6x-6go-dx0x_0eSd?6-xl6 o?,xTh o_o?h?6gx 7, h 6_h?o[6_,x_0)xe6-_0,-6e?xh?6$xK  6-_0,-6e?xl6 o?,x6_h?o[6_, e 0--0exTf76x6?xh_/xPKKm7x/ +$/

/xG0__0)oeSxoll6xTPKK$xeo0exg6elo?,xh?6lxh6xg6oe6gxhlxeo0ex-6-76l#oxhlxh 00?o0ex0 )hS6xhegxlh_h,x6he6lxoe 6-_0,-6e?xT/x +$/ P/xt00goeh?6gx)hS6x7hShoeoeSx _hlloo6gxhlx0__0)oeSIx  )hS6xe6S0?oh?o0elx 00goeh?6gxheg0x 0eg ?6gx7, 6e?h_o6gx0x 0e 6e?h?6gx_h70xhegx6-_0,6x 0e6g6h?o0elxoe l0-6x hl6lx)o?#xS0[6e-6e?xoe[0_[6-6e?x/A  -0g6h?6x 00goeh?o0exh?x?#6x 6e?h_x0 oegl?,x_6[6_xK  hS-6e?6gx7hShoeoeSx 0eoe6gx_hS6_,x?0 oego[ogh_xo-lx0x_he?lxTf76xchSoex.?6#6elx%hg,xhegx%6 o6_gxPKKm7x/x
$/ /x.?0 x-h6?x ho?h_oh?o0ex6_h?o[6x?0 FDxT 0el6[h?o[6xh[6hS6x[h_6xl?0eS_,x)hgx7ohl6gxg6x?0 0?_o6x hhex)o?#0?x hhex[h_6xh-0e?lx?0 0e_,xK/ xhIxFh?hxe0?xh[ho_h7_6$/

Denmark Finland Norway Sweden Social-democratic ave. Austria Belgium France Germany Italy Japan Netherlands Switzerland Conservative average Australia Canada Ireland UK USA Liberal average

Countries

Inter-Firm Relationships Union Coordinated density wage bargaining (1990)1 (1990)2

Table 6: Classification of 17 Advanced Market Economies

3.3 Classification of Welfare Capitalist Systems

35

36

3 Worlds of Welfare Capitalism

tions become apparent when looking at Table 6: Firstly, the levels of union density for Northern Europe are more than twice as high as the levels reported for the remaining European countries (see Huber & Stephens, 2001a, pp. 93-94; Huber & Stephens, 2001b, pp. 116-117; Kitschelt et al. 1999/ 2005, p. 435). Latter might be reasoned by the traditionally influential role of social democracy in the Nordic countries. As partly mentioned in Chapter 2.2, social democrats have foreseen comprehensive social security legislation as well as neocorporatist institution building (power mobilization) as programmatic cornerstones when developing the Nordic welfare capitalist system (Esping-Andersen, 1990/ 1998, p. 12; see Hicks, 1999, p. 127). Secondly, in three of the four Nordic countries wage bargaining is coordinated on a national level; by contrast, several Continental European countries exist in which wage negotiations occur on an industry (sector) level (Kitschelt et al., 1999/ 2005, pp. 432-434; see Swank, 2003, pp. 65-67). In light of these institutional distinctions, the author advances the view that Nordic countries form a social democratic production regime cluster whereas Continental European countries and Japan constitute a conservative production regime cluster.44 Overall, following the comparative assessment of leading welfare economies based on principal welfare state regime indicators (see Table 2 in Chapter 2.3.2) and production regime indicators (Table 6), the author identifies three distinct varieties of welfare capitalism: Firstly, Australia, Canada, Ireland, the UK, and the USA make up the liberal welfare capitalism cluster. This group is characterized by a residual welfare state and a market-coordinated production regime. Secondly, Austria, Belgium, France, Germany, Italy, Japan, the Netherlands, and Switzerland form the conservative welfare capitalism cluster. The main features of this category are a class-differentiated, transfer-based welfare system coupled with a corporatist (sector-coordinated), bank-centered production regime. Thirdly, Denmark, Finland, Norway, and Sweden constitute the social democratic welfare capitalism cluster. This Nordic welfare stateeconomy nexus consists of a comprehensive, universally inclusive system of social

44

Although Continental Europe and Japan are assigned to the same production regime cluster, it is worthwhile noting that both slightly differ with regards to the mechanism of coordination. As Hall & Soskice (2001/ 2004b) point out, Continental European countries center on industry-based coordination whereas Japan relies on group-based coordination (p. 34; see Kitschelt et al., 1999/ 2005, p. 434). Precisely, as they demonstrate, “in Germany, coordination depends on business associations and trade unions that are organized primarily along sectoral lines, giving rise to vocational training schemes that cultivate industry-specific skills, a system of wage coordination that negotiates wages by sector, and corporate collaboration that is often industry-specific. By contrast, the business networks of most importance in Japan are built on keiretsu, families of companies with dense interconnections cutting across sectors, the most important of which is nowadays the vertical keiretsu with one major company at its center” (Hall & Soskice, 2001/ 2004b, p. 34).

3.3 Classification of Welfare Capitalist Systems

37

protection and a corporatist (nationally coordinated), bank-centered production regime that is characterized by high levels of female employment and wage equality (see Ebbinghaus & Manow, 2001b, pp. 11-12). Given the fact that the 17 welfare economies have been increasingly challenged by domestic pressures and global economic forces in recent decades, it is worthwhile analyzing in how far the three patterns of welfare capitalism outlined above are still prevalent today. Thus, Parts II and III will look in close detail at the driving forces behind the contemporary welfare state in affluent democracies and assess their impact on the institutional configuration of these countries’ welfare state-economy nexus.

Part II Historical Development and Contemporary Challenges of the Advanced Capitalist Welfare State Building on the theoretical foundations laid in Part I, Chapter 4 reviews the historical process of welfare state development while Chapter 5 and 6 confront the reader with the contemporary challenges of the advanced capitalist welfare state. In general, the academic literature distinguishes two broad categories of welfare state challenges: on the one hand, domestic political and economic pressures; on the other hand, global economic forces commonly known as “economic globalization” (Buti et al., 1998, pp. 17-26). Adhering to Pierson’s (2001a) classification of “post-industrial pressures”, Chapter 5 sketches the nature of welfare state challenges which are domestically provoked in the political and socio-economic realm. Subsequently, Chapter 6 lays the central focus on the process of economic globalization. Besides conducting a thorough analysis of all three dimensions of global economic integration (product, capital, and labor market globalization), the author – with the aid of recent classifications by Andersen (2003), Brady et al. (2005, pp. 922-925), Brady et al. (2007, pp. 318-320), Ellison (2006, pp. 48-59), Garrett & Mitchell (2001, pp. 149-153), Genschel (2004), Hicks (1999, pp. 157-168, Chapter 7), Koster (2008, pp. 2-3), Schulze & Ursprung (1999), Swank (2003, pp. 59-61), and Swank (2002/ 2004, pp. 32-58) – presents an overview of theories, which assign a decisive role to global economic forces in shaping the recent development process of advanced capitalist welfare states.

4

Historical Phases of Welfare State Development

Pierson (1998) measures the development of the advanced capitalist welfare state according to the following three criteria: the stages of social insurance legislation, the expansion of citizenship to grant male universal suffrage and eventually universal adult suffrage, and the growth in public social expenditures relative to GDP (pp. 103-104). Following this analytical framework and reviewing prior research (Alber, 1982, pp. 56-69; Barr, 2004, Chapter 2; Flora & Alber, 1981, pp. 48-57; Heclo, 1981, pp. 386-387; Pierson, 1998, Chapters 4 and 5), the author identifies four discrete time periods of modern welfare state development: the genesis of the welfare state (1870-1918), the interwar period of consolidation (1918-1945), the “golden age” of expansion (1945-1975), and the period of structural adjustment in the era of globalization (from 1975 onwards). Before analyzing in detail each of these four stages, the author briefly sheds some light on the prehistory of the modern welfare state.

4.1

Social Protection in the Pre-Welfare State Era

Although the birth of the welfare state is generally dated at around the last quarter of the 19th century, it is worthwhile noting that scarce features of national social protection – the so-called Poor Law policies – were already prevalent throughout the Western World since the 16th century. These poverty relief programs, however, differed largely from the modern welfare state legislation (Rimlinger, 1971, p. 13; Flora & Alber, 1981, p. 48). Initially perceived “as the aggressive champion of the social rights of citizenship […] [due to its original] claim to trespass on the territory of the wages system, or to interfere with the forces of the free market” (Marshall, 2006, p. 32), the poor law eventually turned into an effective instrument to defend the traditional authoritarian social order (Rimlinger, 1971, p. 13; Flora & Alber, 1981, p. 48). Precisely, it “treated the claims of the poor, not as an integral part of the rights of the citizens, but as an alternative to them – as claims which could be met only if the claimants ceased to be citizens in any true sense of the word” (Marshall, 2006, p. 33). With the occurrence of major events in the political (American and French Revolution) and economic sphere (Industrial Revolution) at the end of the 18th century (Rimlinger, 1971, p. 2), the principles of the Poor Law period were on the wane. Liberal ideas spread proclaiming that “every man was to be free to pursue his fortune and was to be responsible for his success and failure” (Rimlinger, 1971, p. 35; see Flora & Alber, 1981, p. 48). Against this backdrop, “the old poor laws were [considered to be] a hindrance to labor mobility and to the workers’ commitment to an urban, industrial life” (Rimlinger, 1971, p. 36).

42

4 Historical Phases of Welfare State Development

4.2

The Genesis of the Welfare State

Building on the political-economic foundations laid in Chapter 2.2, the rise of capitalism was accompanied by the formation of the industrial working class which openly engaged in power class struggles to transform traditional authoritarian systems into mass democracies and establish a system of social protection against the new risks related to industrialization (Flora, 1986, pp. XII-XVI; Flora, 1985, p. 12; see Flora & Alber, 1981, p. 48). In this sense, the introduction of social insurance legislation starting in the last decades of the 19th century marked the beginning of the modern welfare state. By the outbreak of the First World War (WWI), several countries had developed four main social insurance systems in order to address the following risks: industrial accident, sickness and invalidity, old age and invalidity, and unemployment (Flora & Alber, 1981, p. 50). Although no common sequencing of policy reforms existed, social insurance for industrial accidents was the first to be established according to Flora & Alber (1981) since it represented the least radical break from the pre-existing mindset; employers were already liable for work accidents by that time. Yet this new legislation nonetheless constituted a break with the past in two ways: Firstly, work accidents no longer automatically conferred guilt on the employer, but were now considered an unavoidable side-effect of industrial production. Secondly, the introduction of the social insurance institution induced employers within a given industrial branch to pool the risk of compensation payments in case of work accidents (pp. 50-51). In contrast to these relatively minor tweaks to the social insurance system, the institutionalization of the remaining three insurance systems constituted a major breakthrough (Flora & Alber, 1981, pp. 51-52), the latter involving “the recognition that the incapacity to earn a living through contingencies such as old age, sickness or unemployment […] [was] a normal condition in industrialized market societies and that it […] [was] legitimately the business of the state to organize for collective provision against the loss of income arising from these contingencies (Flora and Heidenheimer, 1981; Flora, 1986; see also the reservations of Jones, 1985)” (Pierson, 1998, p. 103). As Table 7 indicates, Germany and Denmark stood out as pioneers; other Continental European and Nordic countries (with the exception of Finland) followed swiftly. In contrast, Anglo-Saxon countries (except for Ireland and the UK) were characterized by a pronounced distrust of state interventions and thus lagged behind. In a broader context, the implementation of the social insurance legislation also implied “a change in the relationship of the state to the citizen and of both to the provision of public welfare” (Pierson, 1998, p. 103). Unlike in the Poor Law period,

43

4.2 The Genesis of the Welfare State

Table 7: The Introduction of Social Insurance Legislation in OECD Countries

,QSW5ZTK_ TZS-Q5

)ZQ-WW

:-QWZ+Q

>Q-(K+'P

B 

-T_:ZKKT P "$A(K" >D(I"AK D(I"K7KD" ACK422*4

Source: (1) Adapted from OECD (2007c, p. 65); (2) Adapted from OECD (2007c, p. 16).

Note: a: The child poverty rate is defined as the share of children with equivalized incomes less than 50 percent of the median for the entire population. b: Data are for 2005 except for Denmark (1999), Belgium, Canada, Germany, Italy, Japan (2001), Finland (2002), Norway (2003), the Netherlands (2004), and 2006 second quarter for Switzerland.

P/m /m / / *

t#o_g6eh

F6e-h Goe_heg B0)h, .)6g6e 8+SZ_FG-TD-

J" A(

-":$(KD($FK7KD" ACK4222

Table 13: Poverty Rates by Employment Status/ Household Type and the Employment/ Population Ratio of Sole Parent Households

62 5 Post-Industrial Challenges of the Contemporary Capitalist Welfare State

5.4 Welfare State Maturation and Labor Market Imbalances

63

In light of these findings, it seems evident that the Nordic countries appear to be much better adapted to the ongoing transformation of gender roles and household structures than their Anglo-Saxon and Continental-European counterparts. As mentioned above, they have institutionalized a welfare state – economy nexus that allows for and encourages the reconciliation of work and family life, latter being key to resolving the challenges (decline in fertility rates, rise in child poverty) accompanied by ongoing changes in family structures (Esping-Andersen, 2002b, pp. 20-21; OECD, 2007c, pp. 15-17). Although the Nordic countries’ policy to reduce poverty in single-parent households – expanding public sector employment, active labor market programs and welfare services in order to augment female labor force participation (Esping-Andersen, 2000, pp. 78-81) – might appear very costly at first sight, Esping-Andersen’s (2000) empirical analysis clearly demonstrates that the long-term benefits of these measures outweigh the costs (pp. 164-165). In line with these findings, he states that “Sweden urges lone mothers to work; America ‘taxes away’ their incentive to do so. The price of ‘urging’ may be quite high to the Swedish welfare state, but the hidden cost of not doing so may be even higher. If, like in America (and Britain), lone parents leave the labour market and have little incentive or capacity to return, they are more likely to become long-term welfare clients (as we have seen) and the net loss due to forgone lifetime earnings (and inferior life chances of their offspring) may become extremely high” (Esping-Andersen, 2000, p. 163). Against this backdrop, the OECD (2007c) has called upon governments in advanced market economies to pursue family-oriented policies. Among others, latter involve “improved access to affordable and quality childcare, financial support for children, arrangements that allow working parents to take leave to care for children, and flexible workplace practices that allow a better reconciliation of work and care commitments. They also include financial incentives to work for families with children and employment support for jobless parents” (p. 13).62 With pressures increasing to pursue this shift towards a more family-friendly policy agenda, contemporary welfare states are challenged to take on new tasks previously accomplished within the boundaries of the family household (Mahon, 2002, p. 1).

5.4

Welfare State Maturation and Labor Market Imbalances

Although research on post-industrial pressures has largely centered on exogenously driven processes such as deindustrialization, population ageing as well as household structure transformation, “the gradual expansion, maturation, and ‘growth to limits’ of governmental commitments” (Pierson, 2001a, p. 83) has increasingly come to be 62

See Esping-Andersen (2002a, 2002c) for a detailed analysis on how these family-friendly policies should be crafted.

64

5 Post-Industrial Challenges of the Contemporary Capitalist Welfare State

seen as a fundamental challenge for advanced market economies in recent years.63 According to Pierson (2001a), “the wide scope of established governmental commitments is a defining feature of affluent, post-industrial democracies. By themselves, these expanded governmental commitments generate persistent budgetary pressure and a pronounced loss of policy flexibility in the contemporary period” (p. 88). Taking the same line, Buti et al. (1998) specify two challenges related to welfare state maturation: Firstly, given the fact that the welfare state largely provides labor-intensive human services it is increasingly prone to Baumol’s cost disease (p. 19). Secondly, overly generous public social institutions (for example, high unemployment benefits, long benefit duration) appear to exert detrimental effects on labor market efficiency, the latter in turn backfiring on the fiscal sustainability of the overall welfare state system (pp. 19-22). Since the end of WWII, government spending – in particular spending on welfare state activities – has rapidly increased. As indicated in Figure 4, total government disbursements (as a percentage of GDP) of five selected OECD countries have risen on average from a level of 29 percent in 1960 to approximately 50 percent in the mid-

–䉱–

×

– –

Figure 4: Total Government Disbursements (as % of GDP), 1960-2005 Source: Data compiled from OECD (2005c). 63

For a comprehensive analysis of the causes and consequences of welfare state maturation, see Pierson (2001a, pp. 88-92).

5.4 Welfare State Maturation and Labor Market Imbalances

65

1990s. Recently, however, one can witness stagnation or even decline in public spending across these high income countries. In 2005 total government disbursements accounted for approximately 45 percent of GDP on average. Flora (1986) has been one of the first researchers to consider the phenomenon of welfare state maturation. According to his view, with welfare states maturing “there will be a ‘natural’ tendency in their expansion to slow down or stagnate […] – similar to the slowdown of population growth in the course of the demographic transition and to the deceleration of the sectoral change of employment from an industrial towards a service economy” (p. XXIII).64 Furthermore, given the comprehensiveness of existing welfare assistance, a continuous expansion of the welfare state (similar to the postWWII era) appears to be unnecessary (Flora, 1986, p. XXIII). Recently, adhering to Flora’s (1986) “Growth to limits” view (p. XXI) and contemporary research (for example, Castles, 2004, pp. 168-174), Obinger & Starke (2007) have found evidence that advanced capitalist welfare states have been moving towards a country-specific steadystate ratio of public social expenditures relative to GDP. The level of this steady state seems to be determined by national power resource constellations of the past (p. 477). Looking from a different angle at the phenomenon of welfare state maturation, Pierson (2001a) advances three reasons to explain why welfare state commitments are stagnating or even declining in the contemporary era: Firstly, given the high level of public social provision, “the opportunity cost associated with further expansions has risen (Klein and O’Higgins 1988). If social expenditure grows at twice the rate of national income, the cost in foregone resources is much greater when social spending is 30 per cent of GDP than it was when that share was 15 per cent of GDP” (Pierson, 2001a, p. 90).65 Secondly, an increase in welfare state spending would necessitate a rise in taxation. The latter is particularly in Continental Europe hardly economically feasible, because an augmentation of non-wage labor costs would further aggravate the unemployment problem (Scharpf, 2000, p. 222; see Buti et al., 1998, p. 21).66 Moreover, political tensions would likely ensue as an increasing part of the electorate could be expected to block efforts to impose a higher tax burden (Pierson, 2001a, p. 90). Thirdly, given the fact that numerous contemporary welfare states have evolved into generosity systems that exert major “disincentive effects via tax distortions and moral hazard” (Lindbeck, 2006, p. 14) on labor supply, further expansion in public social spending is expected to arise public controversy (Pierson, 2001a, p. 90). 64

See also on this topic Kohl (1981). In this context, Tanzi & Schuknecht (1997) have empirically shown “that by the time countries reach the level of public spending shown by the small governments, namely, between 30 and 40 percent of GDP, much of the potential social gain from public spending has been obtained. Spending beyond that level does not contribute much” (p. 168). 66 See also Chen & Funke (2003) who have analyzed the development of non-wage labor costs and their impact on labor demand in Germany. 65

66

5 Post-Industrial Challenges of the Contemporary Capitalist Welfare State

In light of the pressures associated with the increasing maturation of governmental commitments, national policymakers find themselves forced to limit public social provision and raise the efficiency of existing welfare state institutions. In this context, several scholars have recently proposed various reform paths: Buti et al. (1998) advocate a tightening of unemployment benefit schemes (level of compensation and eligibility rules) and a relaxation of labor market regulations throughout Europe (pp. 20-22). Apart from these suggestions for policy reform, Hemerijck (2002) argues for the introduction of activation programs and employment subsidies (pp. 195-197). Proposing a more radical step, Orszag & Snower (1998) envisage welfare accounts (retirement, unemployment, education and training, and health) for each citizen replacing the current non-transparent system of social protection (pp. 110-113). On a more general level, revamping efficiency also involves policies for tackling “the growing discrepancy between existing programme design and social demands. The contemporary welfare state addresses a past social order; its ideals of universalism and equality emerged with reference to a relatively homogenous industrial working class” (Esping-Andersen, 1996, pp. 8-9). Yet, as explained in the previous sub-chapters, the post-industrial era is characterized by challenges arising from new social risks such as deindustrialization, demographic shifts, and changing household structures. Thus, welfare states in OECD countries are under pressure to implement major adaptations in order to respond adequately to the new needs of heterogeneous groups within society.

5.5

Post-Industrial Pressures and Agents of Welfare State Change: Partisan Politics versus New Politics

As awareness of the above described domestic pressures has significantly grown in recent decades, an intense academic debate has erupted about the potential implications these post-industrial developments are likely to have for the process of political decision-making. Specifically, the following question has been put up for discussion: Who are the agents of change determining public social spending activities in the contemporary era? Social scientists in comparative welfare state literature, affiliated with the power resources approach and the class mobilization thesis67 (such as Allan & Scruggs, 2004; Castles, 1982; Esping-Andersen, 1985/ 1988; Garrett, 1998a; Huber & Stephens, 2000; Korpi, 1989; Korpi, 1980; Korpi & Palme, 2003), are of the opinion “that it is fruitful to view welfare states as outcomes of, and arenas for, conflicts between class-related, socioeconomic interest groups and that in these distributive conflicts partisan politics 67

See Chapter 2.2 for a detailed explanation of these theoretical concepts.

5.5 Post-Industrial Pressures and Agents of Welfare State Change

67

is likely to matter” (Korpi & Palme, 2003, p. 425). Accordingly, they contend that welfare state expansion has largely been driven by social-democratic governments whereas policies to contain public social spending have primarily been initiated by parties of the secular right political spectrum (Castles, 1982, p. 85; Korpi, 1989, p. 323). Similar partisan effects have also been found for the contemporary era of welfare state maturation and retrenchment: While left government parties tend to block attempts to cut public spending or even expand the generosity of social programs as a result of economic shocks, right-wing political parties openly pursue the rolling back of the welfare state (Amable, Gatti, and Schumacher, 2006, p. 441; Korpi & Palme, 2003, p. 441; see also Allan & Scruggs, 2004, pp. 505-50768; Brady et al., 2005, pp. 937-940). Besides social-democratic governance, welfare-enhancing effects have also lately been attributed to Christian-Democratic government parties (Brady et al., 2005, pp. 937-940, Huber et al., 1993, pp. 734-737; Huber & Stephens, 2000, p. 334). Recently, the above view has been challenged by the emergence of the new politics approach which has been propagated most prominently by Pierson (1994/ 1997, 1996, 2001b, 2001c). With governmental commitments reaching maturation and policymakers pressured to roll back the welfare state, Pierson (1996) envisages a radical transformation of the political landscape within industrialized economies (p. 144) since “retrenchment is a distinctive and difficult political enterprise. It is in no sense a simple mirror image of welfare state expansion, in which actors translate (in the view of competing theories) a favorable balance of class ‘power resources’ or institutional advantages into political success. Retrenchment advocates must operate on a terrain that the welfare state itself has fundamentally transformed” (Pierson, 1994/ 1997, pp. 1-2). In this context, Pierson (1996) refers to the group of welfare state employees and recipients who constitute a powerful political force in favor of social programs in each advanced capitalist welfare state. Given the fact that these interest groups revolving around the welfare state are well organized, policymakers aiming to retrench the welfare state are assumed to encounter fierce political resistance from large parts of the electorate (pp. 174-175; Pierson, 2001b, p. 8). Furthermore, Pierson (1996) also emphasizes the role of path dependence with respect to welfare state reform. As he notes, “organizations and individuals adapt to particular arrangements, making commitments that may render the costs of change […] far higher than the costs of continuity. Existing commitments lock in policymakers” (p. 175). Against this backdrop, he argues that processes of welfare state cuts are very rare. According to his view, they will only take place if one of the following four political preconditions is preva68

Allan & Scruggs (2004) provide evidence for a strong partisan effect of right government parties since the recession of the early 1980s. By contrast, they do not find any significant influence of left government parties on welfare effort for the same period (pp. 505-507).

68

5 Post-Industrial Challenges of the Contemporary Capitalist Welfare State

lent (Pierson, 1996): Firstly, retrenchment may be feasible “when governments believe that they are in a strong enough position to absorb the electoral consequences of unpopular decisions” (p. 176). Secondly, the enforcement of policies to roll back the welfare state may be possible in case of fiscal imbalances. In such a scenario, policymakers could justify retrenchment measures as an endeavor to preserve rather than destroy the welfare state (p. 177). Thirdly, welfare state restructuring may be facilitated if policymakers are capable to minimize “the visibility of reforms” (p. 177). Fourthly, welfare state cuts might be politically feasible if policymakers succeed in changing “the ways in which trade-offs between taxes, spending, and deficits are presented, evaluated, and decided” (p. 177). Overall, given the fact that the advanced capitalist welfare states have stayed fairly resilient during recent decades, Pierson (1996) notes that governments have rarely been capable to apply one of the strategies explained above in order to avoid the growing resistance from interest groups surrounding the welfare state (p. 179). As a consequence, he concludes in line with the new politics approach that political parties and labor organizations have ceded power to these pro-welfare state oriented interest groups within the social policy context (Pierson, 1996, p. 147). The influence of special interest group activity on welfare effort has been analyzed particularly in the domain of national pension systems. In this context, Pampel & Williamson (1992) have already noted close to two decades ago that “something more than demographic accounting […] is involved in welfare spending. […] we find that pension spending per aged person grows with percent aged in advanced industrial democracies, even though the spending measure controls for the number of potential aged recipients. We find that political participation and party competition interact with the size of the aged population to jointly raise welfare spending. We find that the effect of percent aged on welfare spending increases over time, thus showing that the aged more effectively translate their numbers into greater spending in more recent time periods” (pp. 165-166). Following these and similar results by Strömberg (1996), Persson & Tabellini (2000) have developed a theoretical model of pensions that “predicts that pensions per retiree will be higher, the higher the weight on old voters […], as this shifts the medianvoter equilibrium toward a more generous pension system” (p. 130). Interestingly, Razin, Sadka, and Swagel’s (2002) empirical results oppose the above findings. Analyzing the relationship between the dependency ratio (“defined as usual as one minus the labor force as a share of the population” (Razin et al., 2002, p. 912)) on the one hand and two welfare state measures (labor tax rate and the generosity of per capita social transfers) on the other hand, they show that in a representative democracy the dependency ratio can be negatively correlated with the size of the welfare state after controlling for other factors (p. 901). Razin et al. (2002) explained this phenomenon in terms of

5.5 Post-Industrial Pressures and Agents of Welfare State Change

69

“a fiscal ‘leakage’ from the median voter to the net beneficiaries of the welfare state. The mechanism for the determination of the tax burden and generosity of social transfers emphasizes the demand for redistribution by the median voter. A crucial factor determining the political economy tax-transfer policy is whether this decisive voter is a net contributor or a net beneficiary of the pay-as-you-go social security system” (pp. 915-916). As long as the median voter is still working, he or she will stand up for a reduction of labor tax rates and social transfer generosity (Razin et al., 2002, p. 916). Recently, however, several researchers (for example, Bryant, 2003; Disney, 2007; Simonovits, 2007; Simonovits, 2003) have expressed their doubts about Razin et al.’s (2002) results. Bryant (2003) considers their dependent variable as highly misleading because it assumes that only elderly people constitute the dependent population (p. 1; Simonovits, 2003, pp. 1-3). When disaggregating the share of dependent people into youth and old-age groups and repeating Razin et al.’s (2002) model, he finds that “youth dependency seems to be negatively correlated with tax and benefit levels, while old-age dependency seems to be positively correlated with taxes and (possibly) benefits” (Bryant, 2003, p. 8). Furthermore, Disney (2007) observes that Razin et al.’s (2002) “set-up explicitly assumes no link between current tax payment and future entitlements” (p. 549; see Simonovits, 2007, p. 536; Simonovits, 2003, p. 3). While Anglo-Saxon countries are characterized by a rather loose link between a citizen’s social security contributions and benefits, the opposite is true for most Continental European and Nordic countries. Thus, Razin et al.’s (2002) model assumptions are too simple to mirror the diversity of existing welfare systems in a comprehensive way (Disney, 2007, p. 549; see Simonovits, 2003, p. 3).

6

Economic Globalization and the Contemporary Capitalist Welfare State

Apart from domestic political and economic pressures, advanced capitalist welfare states are increasingly confronted with processes of international economic integration (see Buti et al., 1998, pp. 17-26). The aim of this chapter is to identify the underlying causal forces and analyze the intensity of economic globalization on product, capital, and labor markets. Subsequently, the author will present the principal hypotheses on the welfare state-globalization nexus.69

6.1

The Underlying Causal Forces of Economic Globalization

Economic globalization describes the process of “increasing internationalisation of markets for goods and services, the financial system, corporations and industries, technology and competition” (OECD, 2005a, p. 16). Three causal forces are considered to be at the root of these global economic processes: firstly, the liberalization of the exchange rate system and the deregulation of financial services starting in the mid-1970s; secondly, the gradual reduction of impediments to international trade and investment within the General Agreement on Tariffs and Trade (GATT)/ World Trade Organization (WTO) system; and finally, the emergence of modern information and communication as well as transportation technologies, which have significantly shrunk time-space distances on a global scale (OECD, 2005a, p. 16). According to this logic, economic globalization is often attributed mainly to liberal public policy and technological advancement. Yet the subsequent intensification of economic globalization during the last three decades has also to be considered “as a microeconomic phenomenon, driven by the strategies and behaviour of firms

69

Throughout this chapter the author will focus on the economic dimension of globalization. For a thorough introduction into the multiple dimensions of globalization, see for example Giddens, Duneier, and Appelbaum (2007, Chapter 20). Furthermore, Held, McGrew, Goldblatt, and Perraton (1999) provide a comprehensive compendium on globalization which has been of great use to the author in the course of writing this chapter. For a general definition of globalization, see Cohn (2003), who describes it as “a process that involves both the broadening and deepening of interdependence among societies and states throughout the world. Broadening refers to the geographic extension of linkages to encompass virtually all major societies and states, and deepening refers to an increase in the frequency and intensity of interactions” (p. 417).

72

6 Economic Globalization and the Contemporary Capitalist Welfare State

that have responded to these [regulatory and technological] changes” (OECD, 2005a, p. 16). The emergence of transnational corporations (TNCs) has given rise to global production networks and significantly increased the volume of foreign direct investment flows around the globe. Consequently, global trade patterns have increasingly shifted from inter-national to intra-firm trade during recent decades (Dicken, 1986/ 1987, pp. 54-56, 60-65). As mentioned above, the process of economic globalization has not only been constrained to product markets, but has also radically shaped financial markets. Unlike in the post-WWII era, when capital was largely bound to the national territory, international investment has become a central feature of national economies (see Scharpf, 2000, pp. 195-196). Besides the rise in longterm financial flows (foreign direct investment, hereafter referred to as FDI), the globalization of financial markets has been in particular accelerated by the strong upswing of short-term financial flows (portfolio capital) on behalf of financial investors (OECD, 2005a, p. 16). Finally, these economic developments have also been accompanied by rising tides of migrants from developing countries as a result of falling transportation costs and periods of liberal immigration policies (Held et al., 1999, p. 299).70

6.2

The Three Dimensions of Economic Globalization

When talking of globalization economists generally refer to processes of international economic integration occurring on product, capital, and labor markets (see Bordo, Taylor, and Williamson, 2005, pp. 1-2). Given the fact that the world had already witnessed a major “broadening and deepening” (Cohn, 2003, p. 417) of international economic relations in previous periods (most prominently during the classical gold standard era at the turn of the 19th century), some commentators do not regard the contemporary wave of globalization as a radically new phenomenon (Bordo et al., 2005, p. 3).71 Although this perspective might not be wrong, previous era of globalization are not of central research interest to the author since they occurred in absence of the advanced capitalist welfare state (see Tanzi, 2002, p. 116). Studying effects of global economic forces on developed welfare state institutions requires foremost an analysis focusing on the development of international trade, financial and migration flows in the post-WWII-era.

70 71

See also Glyn (2007) for an introduction into the topic of economic globalization. For a detailed overview of the principal schools of thought on globalization, see Held et al. (1999, pp. 2-10) and Genschel (2004, pp. 615-623).

6.2 The Three Dimensions of Economic Globalization

6.2.1

73

The Globalization of Product Markets

6.2.1.1 Trade Globalization72 Provoking a sharp fall in international trade volumes and severely aggravating nationalistic tendencies among industrialized countries, the rise of protectionism during the interwar period has often been cited as a principal cause of WWII. With this in mind, leaders of the allied nations, spearheaded by the United States and the United Kingdom, began to establish a multilateral world order at the end of WWII. Known as the Bretton Woods system, it aimed at uniting former conflict parties for the pursuit of common prosperity (Balaam & Veseth, 2001, p. 116, 147-148; see Bagwell & Staiger, 2004, pp. 43-45; Jackson & Sykes, 1997, pp. 2-3). The year 1948 saw the implementation of GATT, which strived to gradually liberalize international trade through a series of multilateral negotiations.73 Central to GATT were the principles of reciprocity and non-discrimination: The former ensured that all member countries would agree to reduce trade barriers in a mutual way while the latter was intended to prevent attempts to give preference to the imports of one nation over another (Most Favored Nation trading status) (Balaam & Veseth, 2001, p. 117; see Hoekman & Kostecki, 1995, pp. 26-30). Although international trade remains a contentious issue, the eight GATT rounds (from Geneva in 1947 to Uruguay in 1994) can be considered a success story in review.74 Average tariff rates were whittled down from 40 percent in 1945 to under 4 percent in the developed world. Tariff bindings, in turn, were increased to approximately 97 percent of all tariffs in industrialized countries and 65 percent in developing countries (Grimwade, 2000, p. 330, 332; see WTO, 2001, p. 11, 16-17). Table 14 (see p. 74) depicts the development of average applied tariff rates throughout the developed and developing world for the last two decades. Simultaneously, international trade volume ascended rapidly during the same period. According to WTO statistics, the total value of worldwide merchandise exports/ 72

See Ellison (2006, pp. 29-33) and Yeates (2001, pp. 33-37) for a similar analysis in the field of economic sociology. Findlay & O’Rourke (2005) provide a detailed analysis of commodity market integration in the field of economic history while Balaam & Veseth (2001, pp. 110-132) look at the development of international trade from a political economic angle. 73 In fact, the Bretton Woods Agreement did not explicitly lay the foundations of GATT. Rather, proposals were made that envisaged the establishment of an International Trade Organization (ITO) with extensive competences. Yet several national governments did not want to cede too much power to an international organization. As a consequence, the ITO was never founded. Nation states instead launched GATT as a provisional solution in order to promote free trade (Balaam & Veseth, 2001, pp. 116-117; Krueger, 1998, pp. 2-3; Senti & Conlan, 1998, pp. 11-13). 74 For a historical overview of the eight GATT rounds, see Hoekman & Kostecki (1995, Chapter 1).

74

6 Economic Globalization and the Contemporary Capitalist Welfare State

Table 14: Trends in Average Applied Tariff Rates in Industrialized and Developing Countries, 1988-2007 (unweighted in %) 
++ PP/K



+/



+ m/\

PKK K/A

PKK\ +/+

foS#doe 0-6x 0e?o6l High-income countries Non-OECDs B0edtFl tFl OECDs

/
/A

P/
/+

A/
m/

m/A /A

A/ P/


&Rmiddle income countries Low !0)x -ogg_6x oe 0-6x 0e?o6l !0)x oe 0-6 Low income nogg_6x oe 0-6 Middle income

P/ P/ PP/P

P/ P
/ /+

/ 
/
m/A

P/K m/K K/\


/+ P/K +/\

World k0_g

Note: The World Bank (n.d.) mainly compiled his data from UNCTAD TRAINS database. Source: Adapted from World Bank (n.d.).

imports increased from 58 billion/ 62 billion US dollar (at current prices) in 1948 to approximately 14 trillion/ 14.2 trillion US dollar (at current prices) in 2007 (WTO, n.d.). As a result, Irwin (1995a) notes that the GATT’s long-term influence has been by far more important for the world economy than the impact exerted by the World Bank or the International Monetary Fund (IMF) (p. 327; see Irwin, 1995b). With the closing agreement of the Uruguay round in 1994, a new epoch in international trade dawned. Although the previous two rounds (Kennedy and Tokyo round) had already considerably widened the GATT, the Uruguay round marked a significant break from past developments since it enlarged the world trading system by introducing three agreements (the General Agreement on Trade in Services (GATS), the Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS), and the Agreement on Trade-Related Investment Measures (TRIMs)) complementary to GATT. In light of this increasingly broad array of trade-related issues, the post-WWII trading system needed a major institutional overhaul. Thus, legislators decided to establish the WTO within which all four general agreements were institutionally embedded (Balaam & Veseth, 2001, pp. 119-120; Senti & Conlan, 1998, pp. 15-17). Created in 1995 with a status similar to the IMF and the World Bank, the WTO can be seen as the third central pillar of the international economic order (Hoekman & Kostecki, 1995, pp. 37-40, 51-52). Its principal objective has been “to help trade flow smoothly, freely, fairly and predictably. It does this by: […] Administering trade agreements [,] […] Acting as a forum for trade negotiations [,] […] Settling trade disputes [,] […] Reviewing national trade policies [,] […] Assisting developing countries in trade policy issues, through technical assistance and training programmes [,] […] Cooperating with other international organizations” (WTO, 2008, p. 7). During the last decade the WTO has evolved from an international body mainly occupied with negotiating tariff reductions into an institution that increasingly understands itself as a forum for setting rules for global competition (Balaam & Veseth, 2001, p. 121). Its prominent role within the system of global governance has

6.2 The Three Dimensions of Economic Globalization

75

recently caused considerable public controversy. While industrialized countries as well as economic liberals consider the WTO as a catalyst for world-wide growth and prosperity, developing countries as well as numerous Non-Governmental Organizations (NGOs) feel that the institution is an instrument of the “North” that preaches trade liberalizations that are advantageous to developed countries. As an example, critics of the WTO and the existing world trade system often point to the industrialized countries’ unwillingness to open up their agricultural markets (Balaam & Veseth, 2001, pp. 122-124; see Schirm, 2004, pp. 266-274). In light of these opposing standpoints between “North” and “South”, recent ministerial conferences (Cancún 2003 and Hong Kong 2005) have not achieved much progress in their efforts to further liberalize international trade. Now that the theoretical underpinnings and critical issues underlying the institutional framework of the current world trade system have been discussed, an empirical analysis examining the development of international trade flows and assessing the current degree of global product market integration will be presented. As Findlay & O’Rourke (2005) point out, market integration can best be measured by comparing the prices of identical products in different markets. If market integration is taking place, these prices should converge across countries and considerably boost trade volumes (pp. 13-14).75 However, given the lack of available price data that is comparable for a wide range of tradable goods and services in numerous countries, the author is compelled to apply, in line with Ellison (2006, p. 30), Findlay & O’Rourke (2005, p. 41), and Yeates (2001, p. 35), the conventional method – using historical data on the ratio of total trade to output in each country – when measuring aspects of trade globalization. Table 15 (see p. 76) presents the degree of trade openness – specified as annual sum of exports and imports of goods and services as a percentage of GDP – for 17 OECD countries in four different time periods. With the exception of Japan and Norway, all countries – confirming prior sources – indicate a dramatic upswing of trade volume relative to GDP. This finding is a clear indication that national markets are becoming increasingly integrated. With respect to trade volume levels, evidence is provided in favor of Katzenstein’s (1985) finding that small states have in the past been more eager to open up their economy in order to find growth opportunities beyond their small home markets (p. 22, 24; see Cameron, 1978). For example, trade volume levels exceeded respective national output levels in Austria, Belgium, Ireland, and the Netherlands in 2000. As mentioned above, the dramatic rise in trade flows can be mainly attributed to the liberal trade policy pursued by national governments during 75

It should be noted that Findlay & O’Rourke (2005) assume other parameters (such as tax policy) to be equal between different countries in the scenario of price convergence. Furthermore, they admit that changes in trade volume are not necessarily the result of globalization. Domestic issues (such as shifts in supply or demand behavior) can also have an influence on trade flows (p. 13).

76

6 Economic Globalization and the Contemporary Capitalist Welfare State

Table 15: Degree of Trade Openness in OECD Countries, 1975-2000

t0e?,

0)?#xch?6


\A


+A



A

PKKK

Ul?h_oh

P+/P

m/\


/+

mA/+

P/m

Ul?oh

/A

\
/m

\m/m

K/P

m/

%6_So-

KA/

mP/A



+/

A
/

thehgh

m\/

Am/\

\P/

+/m

+ /m

F6e-h

A
/A

\P/A

/+

+P/P

+/P

Goe_heg

A/\

A/


A/

\/\

m+/m

Ghe 6

/

m/+

m /

AA/


Am/


6-he,


/

AP/A

m+/

\/P

\/K

46_heg

+/m

P/P

m/m

+/+

K/m

4?h_,

mK/

mA/m

AK/K

AA/

\/K

hhe

PA/m

PA/

/+

PK/P

dPK/A

B6?#6_hegl


m/\

PK/K

K
/K

P
/\

\/K

B0)h,

\+/

\+/m


/\

\/

d /P

.)6g6e

Am/


+/K

\/+

+/m

A\/m

.)o?6_heg

AA/

\P/\

A/\

+A/A

Am/

JQ

AP/A

A/

A\/

A+/P

K/


J.U

/

\/P

P /m

P/

 /m

/+


\AdPKKK

Source: Data compiled from World Bank (2005).

the post-WWII era. Figure 5 depicts the degree of current account liberalization across four groups of developed market economies for the time period 1975-1999. Beyond these mere facts, two features of the contemporary wave of trade globalization are worth mentioning: First of all, international trade patterns have significantly changed with respect to the composition of trade. While inter-industry trade dominated international trade flows during the first half of the 20th century intra-industry trade is nowadays the leitmotif for international trade. Furthermore, given the rapid expansion of multinational companies’ activities beyond nation state borders, intra-firm trade plays an indisputably prominent role (Held et al., 1999, p. 176; OECD, 2005a, p. 17; Yarbrough & Yarbrough, 1997, pp. 87-88; Yeates, 2001, pp. 35-37). Secondly, international trade patterns have also radically changed with respect to the countries involved.

77

Degree of liberalization

6.2 The Three Dimensions of Economic Globalization

–䉱–

Figure 5: Developed Countries and Current Account Liberalization, 1975-1999 Note: Huber et al. (2004a) take the data from Quinn’s work on international financial regulation. The degree of liberalization ranges from zero to eight. Source: Data compiled from Huber et al. (2004a).

Many developing countries were formerly colonies of the European empires; as such, they were marginalized into a position of raw material suppliers. An increasing number of countries have since emancipated themselves from the historical core-periphery dependency in recent decades. As an example, exports of light manufactured goods, which amounted to only 5 percent of all exports from the developing to the developed world in 1955, rose to 58 percent by 1992 (Hillman, 2003, p. 14; see Findlay & O’Rourke, 2005, p. 55). Combined, developing countries managed to account for 31 percent of global merchandise trade volume in 2004 (WTO, 2005, p. 1). This indicates that the traditional North-South pattern no longer reflects macroeconomic realities; an increasing number of former developing countries have transformed into fast-growing, newly-industrializing economies (Held et al., 1999, pp. 176-177). 6.2.1.2

Foreign Direct Investment and the Globalization of Corporate Activities76

Apart from the strong upswing of trade in goods and services, the globalization of product markets is also reflected in the rapid development of FDI flows.77 Table 16 (see p. 78) presents the degree of FDI openness – specified as annual sum of FDI in76

See Ellison (2006, pp. 33-37), Swank (2002/ 2004, pp. 15-20), and Yeates (2001, pp. 37-38, 45-58) for a similar analysis in the field of economic sociology. Grunberg (2001) examines the development of FDI flows and the emergence of MNCs from a political-economic angle. 77 Held et al. (1999) define FDI “as the acquisition and management of overseas productive assets” (p. 203).

78

6 Economic Globalization and the Contemporary Capitalist Welfare State

and outflows as a percentage of GDP – for 17 OECD countries in five time periods. All countries indicate a dramatic rise in investment openness. The latter trend is a clear indication that national markets are becoming increasingly integrated (see Ellison, 2006, pp. 34-37; Obstfeld & Taylor, 2005, pp. 139-144; Swank, 2002/ 2004, pp. 15-18). The author assigns the rapid growth of cross-border direct investment flows to an increasing internationalization of corporate activities throughout the world.78 Table 16: Degree of Foreign Direct Investment Openness in OECD Countries, 1965-1998

t0e?,


A


\A


+A



A



+

Ul?h_oh

K/

K/\

P/

A/

P/

Ul?oh

K/P

K/

K/m

/

/A

%6_So-

/

/\

/A

+/

PK/

thehgh

/P

K/


K/P

/K

+/+

F6e-h

/

K/


K/\

m/

/

Goe_heg

K/

K/

K/


/


P /+

Ghe 6

K/

K/+

K/+

P/+

A/P

6-he,

K/


K/\

K/


/


A/P

46_heg

/m

/


/K

m/


+/

4?h_,

K/\

K/A

K/

/

/

hhe

K/P

K/m

K/A

K/m

K/\

B6?#6_hegl

P/m

/

P/\

A/


/A

B0)h,

K/

/

/m

m/K

m/P

.)6g6e

K/


K/\

/

K/+

\/

//

//

/K

m/\


/

JQ

/m

P/\

/A

A/A

 /K

J.U

K/+

/

K/+

P/

/+

.)o?6_heg

Note: Huber et al. (2004a) take the data from Swank’s Data Set on the Comparative Political Economy of Advanced Industrial Democracies. Source: Data compiled from Huber et al. (2004a). 78

Kreikebaum (1997) provides a comprehensive overview of different internationalization strategies (pp. 256-258). Schmid (2007), in turn, assesses the degree of internationalization of the 30 largest German joint stock companies.

6.2 The Three Dimensions of Economic Globalization

79

According to Navaretti, Haaland, and Venables (2002), firms invest abroad for two distinct reasons: Firstly, they want to enter a foreign market in order to acquire a new customer base. This so-called horizontal FDI is a means to avoid transportation costs and circumvent trade barriers, which constitute a significant obstacle for firms that export their products from their country of origin. In this sense, horizontal FDI acts as a substitute for trade (p. 14). Secondly, firms go abroad in order to minimize procurement and production costs. This form of investment is termed vertical FDI “since it involves slicing the vertical chain of production and relocating part of this chain in a low-cost location” (p. 14). Vertical FDI, in other words, generates intra-firm trade, which is today the dominant component of international trade flows (p. 14). Among all companies, MNCs play a prominent role in direct investment flows (Dicken, 1986/ 1987, pp. 60-65; Ellison, 2006, pp. 34-37; Grunberg, 2001, p. 346; OECD, 2005a, p. 16). MNCs are generally defined as “economic organizations engaged in productive activities in two or more countries. Typically, they have their headquarters in their country of origin (their home country) and expand overseas by building or acquiring affiliates or subsidiaries in other countries (the host)” (Grunberg, 2001, p. 347). The strong growth of MNCs in the aftermath of WWII was mainly driven by the goal to capture economies of scale and economies of scope (hierarchical capitalism) (Dunning, 1995, pp. 462-466, 475-476). Recently, however, MNCs have begun to engage increasingly in cooperative alliances (alliance capitalism) (Dunning, 1995, pp. 466-473, 475-476). In the following both determinants of MNC activity will be explained in more detail. Economies of scale arise in a firm when an increase in the scale of production induces a decrease in average production costs of a single good or service (Roach, 2005, p. 28). This effect can either be accomplished by “division of labor through specialization” (Roach, 2005, p. 28) or by the combination and/ or replacement of “human labor with mechanized production. Investment in large-scale capital equipment and the latest production technologies often requires significant up-front costs. These investments may be affordable only for firms with substantial retained earnings or access to credit” (Roach, 2005, p. 29). By providing a competitive advantage, economies of scale, in other words, nurture the development of MNCs (Roach, 2005, p. 29). Economies of scope, by contrast, refer to efficiency gains related to an expansion of the range of goods and services produced. By increasing the scope of production, MNCs raise synergies along common distribution and marketing networks, which in turn reduce average production costs (Roach, 2005, p. 29). Finally, cooperative alliances refer to collaborative arrangements (for example, joint venture) that MNCs establish with other firms in different markets (Dunning, 1995, p. 467). Reviewing the empirical literature on this topic, Dunning (1995) notes that “the principal incentives for alliance formation have been to lower transaction costs, develop new skills and to overcome or create barriers to entry in national or international markets” (p. 467).

80

6 Economic Globalization and the Contemporary Capitalist Welfare State

Although the lion’s share of FDI flows is still concentrated on the industrialized world, developing and emerging countries in Africa, Asia, Eastern Europe, and Latin America are quickly catching up (see Siebert, 1999/ 2000, p. 63). By 2005, these countries have already accounted for more than 40 percent of global FDI inflows (United Nations Conference on Trade and Development (UNCTAD), 2006, p. 299). In particular, China has been very successful in luring foreign direct investment. In 2005, FDI inflows amounted to 72 billion US dollars (China excluding Hong Kong), making the country the world’s third most attractive business location behind the United Kingdom and the United States (UNCTAD, 2006, pp. 299-301). Yet it would be a mistake to understand cross-border financial flows involving developing countries as unidirectional. While it is true that the developing world still reports huge net capital inflows79, it is worth noting that an increasing number of countries have recently emerged as outward investors (UNCTAD, 2006, p. xxiii, 6-7). In contrast to previous periods of internationalization, developing countries are today host to an augmenting number of MNCs and TNCs80 that invest in developing as well as developed countries in order to serve new markets and acquire know-how (UNCTAD, 2006, p. xxiii; Grunberg, 2001, p. 350). In 2004, 5 out of the top 100 TNCs worldwide were headquartered in emerging economies (China, Hong Kong (China), Malaysia, the Republic of Korea, and Singapore) (UNCTAD, 2006, p. 31).

6.2.2

The Globalization of Capital Markets81

The multilateral world order that was planned for the post-WWII era by the Allied nations at the Bretton Woods economic conference not only foresaw the creation of international organizations (the GATT, the IMF, and the World Bank) acting as central governing bodies of the reemerging world economy, but also envisaged the installment of a world-wide fixed exchange rate (FX) system (Balaam & Veseth, 2001, pp. 147-149). Stable exchange rates were considered key to a successful reconstruction 79

In 2005, developing countries’ share in global FDI outflows amounted to more than 17 percent (UNCTAD, 2006, p. 299). 80 UNCTAD (n.d.) defines a TNC “as an enterprise comprising entities in more than one country which operate under a system of decision-making that permits coherent policies and a common strategy. The entities are so linked, by ownership or otherwise, that one or more of them may be able to exercise a significant influence over the others and, in particular, to share knowledge, resources and responsibilities with the others” (no page). 81 See Ellison (2006, pp. 37-40), Swank (2002/ 2004, pp. 15-20) and Yeates (2001, pp. 38-43) for a similar analysis in the field of economic sociology. Obstfeld & Taylor (2005) provide a detailed analysis of capital market integration in the field of economic history while Balaam & Veseth (2001, pp. 133-181) look at the development of international capital flows from a political economic angle.

6.2 The Three Dimensions of Economic Globalization

81

of international trade within the framework of the GATT rounds (Eichengreen & Sussman, 2000, p. 29). The so-called Bretton Woods FX system introduced in 1946 was based on a gold-dollar standard. Although all countries set a fixed exchange rate between their respective currencies and gold, the hegemon’s currency, the US dollar, constituted the medium of exchange for international transactions. Thus, the US dollar was convertible to gold at a fixed rate (35 US dollars per ounce), while other countries’ currencies remained convertible to US dollars at a fixed rate to ensure current account transactions (Balaam & Veseth, 2001, p. 149). In comparison to prior FX systems (such as the 19th century classical gold standard system), the Bretton Woods FX system contained a number of key innovations: “First, exchange rates, although pegged, could now be adjusted in the event of fundamental disequilibrium. Second, governments that wished to gain room for maneuver to address domestic problems and to avoid a replay of the interwar experience with ‘destabilizing speculation’ were expressly permitted to resort to capital controls. And third, the International Monetary Fund was created as a way of placing international monetary cooperation at one remove from domestic politics” (Eichengreen & Sussman, 2000, p. 29). The United States, acting as the world’s economic powerhouse – the US accounted for “more than 60 percent of the world’s output of manufactures in the late 1940s” (Branson, Giersch, and Peterson, 1980, p. 183) – assumed principal responsibility for the stability of the international monetary system (Balaam & Veseth, 2001, p. 149). During the inception of the Bretton Woods regime, the virtue of a fixed exchange rate could not be undermined. It became synonymous to stable business, stable growth, and less political repercussions. Despite its potential, the regime also had its shortcomings and eventually became infeasible. Eichengreen & Sussman (2000) describe three key factors that potentially led to its collapse, namely “the U.S. balance of payments deficit, rising capital mobility, and the inadequacy of international reserves” (p. 32). Below is a summary explanation of the factors following closely the story line of Eichengreen & Sussman (2000, pp. 32-34; see also Balaam & Veseth, 2001, pp. 149-150; Garber, 1993; Genberg & Swoboda, 1993; James, 1996, pp. 148-227). Until the early 1960s, the balance of payments position of the United States was favorable and the gold reserves were more than sufficient to ensure dollar convertibility. Starting in 1960, however, annual trade surpluses ceased to exceed capital account deficits, because foreign direct investment by US corporations and portfolio capital lending by US banks soared. At that point, US gold reserves were no longer sufficient to cover US foreign dollar liabilities. The future of the international monetary system was thus at stake. The growing instability was further aggravated by the fact that the United States and other industrial countries did not coordinate, but rather set diverging political and economic priorities. The United States, for instance, carried out monetary expansion to finance the Viet Nam war. In light of these growing

82

6 Economic Globalization and the Contemporary Capitalist Welfare State

discrepancies, a devaluation of the US dollar as a means of restoring the international competitiveness of the US economy and stabilizing the US balance of payments was debated. Fearing a loss of political credibility at the height of the Cold War and balking at their allies’ determination to reestablish competitiveness by simultaneously devaluing their own currencies, the US government eventually opted against a devaluation of the US dollar. As a consequence, the US balance of payments significantly deteriorated while the inadequacy of US dollar gold reserve coverage further aggravated matters. These pressures became even more intense as capital controls gradually lost their effectiveness, which was due to the restoration of current account convertibility and the establishment of the Eurocurrency market. Against the background of these mounting macroeconomic challenges, the United States was no longer willing to forego national interest and assume principal responsibility for the international monetary system. In 1971, President Nixon finally declared the end of the US dollar-gold peg (Balaam & Veseth, 2001, p. 149). The subsequent break-up of the Bretton Woods FX system in 1973 cleared the way for a system of flexible exchange rates (Balaam & Veseth, 2001, p. 149). Unlike in the past, domestic politics were no longer shielded from market activity. The market, in other words, became a sanctioning mechanism for public policy (Siebert, 1998, pp. 44-45; see Caspers & Kreis-Hoyer, 2004, p. 69; Tanzi, 1998, p. 13). In this context, policymakers began to confront a key challenge that has been summarized under the term “fundamental macroeconomic policy trilemma” (Obstfeld & Taylor (1998, p. 354) or “unholy trinity” (Cohen, 1995, p. 256). Precisely, the latter refers to “the intrinsic incompatibility of three key desiderata of governments: exchange-rate stability, capital mobility, and national policy autonomy” (Cohen, 1995, p. 256; see also Balaam & Veseth, 2001, p. 152; Obstfeld & Taylor, 2005, pp. 127-129). Monetary autonomy and stable exchange rates were accomplished in the Bretton Woods FX system through the use of capital controls. With the break-up of the post-WWII international monetary system, these controls were lifted and capital became internationally mobile again, just as in the 19th century classical gold standard system. In the context of international capital mobility, national policymakers have been confronted with the choice between monetary autonomy and stable exchange rates. Small states characterized by high degrees of trade openness, such as Hong Kong, tend to favor exchange rate stability over monetary autonomy. Large states with less open economies, such as the United States, are more eager to retain domestic autonomy at the expense of exchange rate stability (Balaam & Veseth, 2001, p. 152).82

82

When the 19th century classical gold standard system was in place, countries were also confronted with the “macroeconomic policy trilemma” (Obstfeld & Taylor, 1998, p. 354) and opted for the following policy mix: fixed exchange rate and capital mobility at the expense of monetary autonomy (see Obstfeld & Taylor, 2005, p. 135).

6.2 The Three Dimensions of Economic Globalization

83

There is much controversy surrounding capital market liberalization in comparison with trade liberalization.83 The contention is caused – despite the potential benefits that economists have hypothesized – by the lack of empirical evidence supporting the view that capital market liberalization has a positive effect on growth (Rodrik, 1998a, p. 61; Stiglitz, 2000, p. 1078). The gain from capital market liberalization depends on whether capital comes from a speculative short-term capital flow or foreign direct investment (FDI). The positive effect of FDI on growth is much more straightforward. FDI promotes knowledge transfer and increases technological adoption across border aiding countries in their ‘catch up’ process (Borensztein, De Gregorio & Lee, 1998, p. 117; De Mello, 1999, p. 148). Increased short-term capital flows could also enhance output and efficiency. Stiglitz (2000) argues five cases in support of the aforementioned theory (p. 1077): Firstly, if citizens are allowed to invest outside their country where investment return is higher then income will increase. Secondly, capital market openness promotes stability through ‘diversification’ of portfolios. Thirdly, these flows act as an alternative funding source. Fourthly, if capital is viewed as a good that can be traded, then capital openness is also beneficial in the same way trade is. And lastly, international financial flows create competition resulting in a better financial infrastructure. The latter has been empirically tested by Levine & Zervos (1998a, pp. 1178-1179). They showed, among other scholars, that stock markets improve after countries eliminate capital controls and that in turn promote accelerated or faster growth (Beck & Levine, 2004, p. 427; King & Levine, 1993, p. 719; Levine & Zervos, 1998b, p. 538). The theory and evidence against capital market liberalization is also substantial. Firstly, the free movement of capital produces instability since policymakers would either have to give up exchange rate control or monetary autonomy, popularly known as the “unholy trinity” (Cohen, 1995, p. 256).84 Secondly, countries appear to be more susceptible to banking crises under financial liberalization, especially those which do not have a strong regulatory infrastructure (Demirgüc-Kunt & Detragiache, 1998 p. 5). Thirdly, contagion and herd behavior is a potential consequence of capital liberalization. Since international flows appear to be volatile, countries are more vulnerable contracting cross-border financial downturns and negative spill-over (Dornbusch, Park, and Claessens, 2000 p. 179-181; see Banerjee, 1992). And lastly, based from the experiences of the developing countries, Calvo, Leiderman, and Reinhart (1996) stress that “large capital inflows can also have less desirable macroeconomic effects, including rapid monetary expansion, inflationary pressures, real exchange rate appreciation and widening current account deficits” (p. 124). 83 84

See Caspers & Kreis-Hoyer (2004, pp. 60-67) for a more detailed overview of the debate. Contrary to this view, Rose (1996) does not find any empirical evidence for an existing incompatibility between capital mobility, exchange-rate stability and monetary autonomy (p. 942).

84

6 Economic Globalization and the Contemporary Capitalist Welfare State

Following these theoretical groundings an empirical analysis is conducted in order to examine the development of international financial flows and to assess the current degree of global capital market integration. Upon reviewing the historical development of the international monetary system in the post-WWII era, one can conclude that the so-called globalization of capital markets appears to be a rather recent phenomenon dating back to the transition from fixed to floating exchange rate regimes in the mid-1970s. Characteristic of this global financial structure is the fact that “finance is less and less a structure that links nations, often through official lending by governments, and is more and more a concern of the global market, dominated by private capital flows, where national borders and regulations are relatively unimportant” (Balaam & Veseth, 2001, p. 166). In general, the following types of private cross-border financial flows can be distinguished: “FDI, international bank lending, international bonds, equities, new financial instruments (notably derivatives) and foreign exchange transactions” (Held et al., 1999, p. 203). According to Obstfeld & Taylor (2005), capital market integration can be measured by analyzing historical volume data (for example, the stock of foreign asset holdings relative to a country’s GDP) and price-based criteria such as the nominal interest arbitrage, real interest rate convergence, and equity returns (p. 138). If market integration has been taking place, the degree of investment openness (the gross stock of foreign capital relative to GDP) should have significantly augmented during the last three decades (Obstfeld & Taylor, 2005, pp. 138-145). Furthermore, real interest rates should have converged and stock markets should be increasingly correlated throughout the world by now (Obstfeld & Taylor, 2005, pp. 145-146, 157-161). Table 17 presents the development of gross private capital flows as a percentage of GDP for the developed and developing world from the late 1970s onwards. In line with prior studies’ findings (Ellison, 2006, pp. 37-38; Obstfeld & Taylor, 2005, pp. 139-144; Swank, 2002/ 2004, pp. 15-20), developed countries have experienced a dramatic increase in direct, portfolio, and other investment in- and outflows. The latter has in particular been pronounced starting in the mid-1990s after capital account liberalization had largely been implemented (see Figure 6, p. 86). Developing countries have also witnessed an upswing in gross private capital flows; yet the development was much less intense than in the high income country group. Among capital movements, international monetary flows (foreign exchange transactions) as well as international transactions in new financial instruments, subsumed under the term “derivatives”85, have accounted by far for the largest share of 85

Held et al. (1999) define derivatives as “instruments derived from other financial products where the basic payment is only a fraction of the total notional value of the product; derivatives essentially are futures, options or swaps […]. […] Futures are simply an agreement to trade a quantity of a product – currency, commodity, bonds, shares, etc. – at an agreed future (Continued at p. 85)

85

6.2 The Three Dimensions of Economic Globalization

Table 17: Gross Private Capital Flows, 1977-2004 (in % of GDP) k0_g foS#xoe 0-6 !0)xRx-ogg_6xoe 0-6 nogg_6xoe 0-6 !0)xoe 0-6 hl?xUlohxRxDh oo 06xR t6e?h_xUloh !h?oexU-6o hxRxtho776he nogg_6xhl?xR B0?#xUo h .0?#xUloh .7d.h#hhexUo h

k0_g foS#xoe 0-6 !0)xRx-ogg_6xoe 0-6 nogg_6xoe 0-6 !0)xoe 0-6 hl?xUlohxRxDh oo 06xR t6e?h_xUloh !h?oexU-6o hxRxtho776he nogg_6xhl?xR B0?#xUo h .0?#xUloh .7d.h#hhexUo h


\\d\
/P /m A/ /A /\ // m/K \/ \/m K/m P/+


+Kd+m \/\ +/ A/A /m P/ * P/A
/A /P K/ m/


+Ad+

/P K/ m/\ A/P P/\ /+ *A \/P P/
/P A/



Kd
m /P /+ \/ +/P m/P +/K //
/P \/P P/
/




Ad

+/A PK/m K/m /P A/m K/A /
/+ A/A /+  /A

PKKK P+/m P/ /m /
\/+ P/+ m/A K/+ A/
/A  /m

PKK PP/K Pm/ /P P/K / K/  / P/ // A/K /

PKKP P/ P /m K/+ / /P
/ m/  / // m/P  /

PKK Pm/A P\/A / P/ /
/+ /m
/+ // A/m /A

Note: van der Hoeven & Lübker (2006) compile the data from World Bank World Development Indicators Database 2006. In accordance with the latter, they define gross private capital flows as “the sum of the absolute values of direct, portfolio, and other investment inflows and outflows recorded in the balance of payments financial account, excluding changes in the assets and liabilities of monetary authorities and general government” (World Bank (2006), quoted according to van der Hoeven & Lübker (2006, p. 34)). Figures for the periods from 1997 to 1999 are arithmetic averages; figures in italics refer to 1982-1984 (East Asia & Pacific) and 1985-1988 (Europe & Central Asia), respectively. Source: From van der Hoeven & Lübker (2006, p. 34).

cross-border financial activity (Held et al., 1999, pp. 207-209; see Ellison, 2006, p. 37; Yeates, 2001, pp. 38-40). Although these instruments were originally provided for hedging purposes related to international trade and FDI, “financial activity […] [has become] largely divorced from the requirements of trade, that is, the exchange of 85

(Continued from p. 84) date for an agreed price. An option gives the holder the right to buy an agreed quantity of a product at an agreed price on an agreed date. A swap is where agents swap the payments associated with two assets – for example, two borrowers may swap payments between a fixed and a floating interest rate loan” (p. 207).

86

Degree of liberalization

6 Economic Globalization and the Contemporary Capitalist Welfare State

–䉱–

Figure 6: Developed Countries and Capital Account Liberalization, 1975-1999 Note: Huber et al. (2004a) take the data from Quinn’s work on international financial regulation. The degree of liberalization ranges from zero to four. Source: Data compiled from Huber et al. (2004a).

goods and services” (Held et al., 1999, p. 221). The annual global turnover of foreign exchange transactions was estimated at approximately 300 trillion US dollars in 1995, while the notional value of derivatives contracts outstanding in 1996 was evaluated at approximately 34 trillion US dollars. These figures dwarf the value of world exports, which were estimated at only five trillion US dollars in 1995 (Bank for International Settlements (BIS), 1997, quoted according to: Held et al., 1999, pp. 208-209). Thus, Held et al. (1999) conclude that the current pattern of “almost instantaneous, real-time, round the clock, global financial trading” has given rise to “an autonomous global financial market” that is unparalleled in history (p. 221). Driving agents behind these developments have mainly been institutional investors (most prominently pension funds) concentrated in the high-income countries (Ellison, 2006, p. 37, 40; Yeates, 2001, pp. 40-41). Recently, however, as a result of the commodity boom and large trade surpluses, a number of emerging countries (for example, China, Russia, Singapore, United Arab Emirates) have set up sovereign wealth

86

Kern (2008) defines SWFs in the following way: “SWFs are government-owned investment funds which are commonly funded by the transfer of foreign exchange assets, and which are set up to serve the objectives of a stabilisation fund, a savings fund for future generations, a reserve investment corporation, a development fund, or a contingent pension reserve fund by investing the funds on a long-term basis, often overseas” (p. 2).

87

6.2 The Three Dimensions of Economic Globalization

Table 18: The Five Largest SWFs by Assets Under Management (as of January 2009) t0e?, Jeo?6gxUh7x-oh?6lxd U7xF#h7o .hgoxUh7oh t#oeh .oeSh06 B0)h,

GegxBh-6

Ull6?lxJ.x5 %o__o0e

4e 6?o0e

oSoe

U7xF#h7ox4e[6l?-6e?xU?#0o?,

+\A


\

o_

eh

o_

.UnUxG06oSexf0_goeSl

m

B0ed t0--0go?,

.UGx4e[6l?-6e?xt0-he,

m\/h

0[6e-6e?x0x.oeSh06x4e[6l?-6e?x t00h?o0e

K


+

B0ed t0--0go?,

0[6e-6e?xD6elo0exGegxd _07h_

K



K

o_

Note: a: This number is a best guess estimation. Source: Adapted from Sovereign Wealth Fund (SWF) Institute (2009).

funds (SWFs)86 which are today central actors in the global financial markets (Reisen, 2008, p. 1). Table 18 gives an overview of the five largest SWFs. Apart from the above findings, evidence for ongoing capital market integration is also found when price-based criteria are examined. Obstfeld & Taylor (2005) analyze long-term real interest rate differentials between the United States and other industrialized countries and find clear indication of statistical stationarity87 suggesting an integrated financial market (pp. 145-150). Fujii & Chinn (2000) – lending support to the results above – have empirically found that real interest rate equalization among the G-7 economies holds stronger if tested over long horizons than in short (p. 2). Another price-based measure is equity return. Obstfeld & Taylor (2005) find evidence for an increasing correlation between the stock markets of the industrialized economies. However, this latter observation has to be treated with caution since the convergence pattern could be the result of a common shock within a group of countries rather than the consequence of economic globalization (pp. 157-161).

6.2.3

The Globalization of Labor Markets88

The contemporary phenomenon of international economic integration is not merely constrained to commodity and capital markets; it also factors in the labor market. The globalization of the labor market involves a new level of complexity, however. Inter87

Obstfeld & Taylor (2005) excluded data from the German hyperinflation period (1919-1923) as well as the wartime periods (1914-1918, 1939-1945) in order to avoid noisy data (p. 148). 88 See Chiswick & Hatton (2005) for a detailed analysis of labor market integration in the field of economic history.

88

6 Economic Globalization and the Contemporary Capitalist Welfare State

national migration entails that “the factor of production (labor services) crosses national boundaries embodied in individuals. […] In Adam Smith’s words, ‘man is of all sorts of luggage the most difficult to be transported’” (Chiswick & Hatton, 2005, p. 65). Even though immigration issues have figured prominently in national and international politics for decades, the labor market today ranks as “the least integrated of global markets” (Glyn, 2007, p. 102). International migration – legal or not – is a complex phenomenon driven by political, religious, environmental, and economic motives (Demeny, 2002, p. 67). In a narrow economic context, it “alters the labor supply and the demographic characteristics of both the sending and the receiving countries. Moreover, it influences economic growth, patterns of trade, income distribution, and the distribution of political power within and between countries” (Chiswick & Hatton, 2005, p. 65). This chapter will focus on the impact of international migration on the receiving countries of the developed world during the second half of the 20th century. The historical pattern of large-scale migration into advanced market economies during the second half of the 20th century can be broadly categorized into four distinct time periods according to Held et al. (1999): Starting in the immediate aftermath of WWII, migration flows were bound to the regional level as people displaced by the events of the war returned to their former homes or were forced to establish new homes elsewhere (p. 299, 303). The second phase was ushered in by the rapidly mounting need for labor to reconstruct the devastated European economies; countries like Belgium, France, and Germany began to recruit workers, initially from Southern Europe in the 1950s and subsequently from Turkey and North Africa in the 1960s. These so-called “guest worker” programs targeted a mostly low-skilled labor force that was supposed to work on a temporary basis. As economies boomed throughout the Western world, migration streams intensified, peaking on the eve of the first oil price shock in 1973 (p. 299, 303). This crisis was accompanied by a fundamental shift in immigration policies, leading to the third period. As advanced market economies faced increasing unemployment and low economic growth rates, the demand for overseas workers came to a sudden end. Immigration then began to be largely restricted to family members of former “guest workers” who had not returned to their countries of origin (p. 299, 303-304). With inflation finally brought under control, economic growth resumed and migration into the industrialized world augmented again in the second half of the 1980s; this marked the start of the fourth phase. The implosion of Soviet-style communism as well as the outbreak of the Yugoslavian civil war accelerated this trend (p. 304).89 Table 19 summarizes the evolution of net migration streams as a percentage of total population for 17 OECD countries in six time peri89

See Zimmermann (1995, pp. 46-47) for a similar historical analysis of European migration patterns.

89

6.2 The Three Dimensions of Economic Globalization

Table 19: Net Migration Flows to OECD Countries, 1970-2003

t0e?,


\K


\A


+K



K



A

PKK

Ul?h_oh

K/


K/

K/\

K/\

K/

K/

Ul?oh

K/P

dK/

K/

K/+

K/K

K/m

%6_So-

K/K

K/P

K/K

K/P

K/

d

thehgh

K/

K/

K/A

K/\

K/A

K/

F6e-h

K/P

dK/P

K/K

K/P

K/

K/

Goe_heg

dK/+

dK/

K/K

K/

K/

K/

Ghe 6

K/m

K/K

K/

K/

K/

K/P

6-he,

K/


dK/

K/A

/

K/A

K/P

46_heg

dK/

K/A

K/K

dK/P

K/P

K/+

4?h_,

dK/

K/K

K/K

K/K

K/P

/

hhe

K/K

K/K

K/K

K/K

K/K

d

B6?#6_hegl

K/

K/A

K/m

K/m

K/P

K/K

B0)h,

K/K

K/

K/

K/

K/

K/P

.)6g6e

K/

K/P

K/

K/m

K/

K/

.)o?6_heg

dK/

dK/


K/

K/+

K/P

K/

JQ

K/K

dK/

dK/

K/

K/

d

J.U

K/P

K/P

K/

K/

K/m

K/m

Source: Data compiled from OECD (2006a, n. p.).

ods. The stock of immigrants in high-income OECD countries climbed to 90.9 million or approximately 9.8 percent of total population in 2005 (World Bank, 2008, p. 38).90 Numerous studies have been conducted to assess the impact of international migration on national wage levels of the native population in host economies. The literature has two opposing views. The first and the most popular belief is that, if both immigrants and natives are close substitutes, an increase in the number of immigrants reduces the natives’ wage due to higher competition and thus providing firms with an option to reduce wages. The second is that, if immigrants and natives are comple-

90

Refugees accounted for only 2.6 percent of all immigrants in high-income OECD countries (World Bank, 2008, p. 38). Immigration is therefore primarily driven by economic motives.

90

6 Economic Globalization and the Contemporary Capitalist Welfare State

ments, an increase in the number of immigrants increases natives’ wage by giving them an opportunity to move to a higher skilled job level thereby increasing economic productivity (Borjas, 1995, p. 34).91 The popular belief on the adverse effect of migration to the general wage and employment levels of the natives has proved empirically weak (Friedberg & Hunt, 1995, p. 42; Hunt, 1992, pp. 568-570; Simon, Moore, and Sullivan, 1993, p. 308; Simon, 1999, p. 154, 170). Following Card (1990), this finding even holds in the case of close substitutability of immigrants and natives (p. 256). The latter stands in sharp contrast to Borjas, Freeman, and Katz (1997) who find that unskilled immigration has exerted a significantly negative effect on the income of low-skilled workers (for example, high-school dropouts) (p. 67). Why do empirical results differ so much from the popular belief of an adverse effect of immigration? Borjas (1995) discusses two potential reasons: The first explanation is that native-born citizens could move to a different region as a response to wage depression caused by immigration inflows. The second is that capital investments could also increase in areas where there is influx of immigrants since firms can view this as an opportunity and thus reducing the potential negative effects of immigration (p. 36). Although the academic debate on the impact of unskilled immigration on national wage and employment levels is still ongoing, all OECD countries have recently adjusted their immigration agenda to attract higher skilled workers and staunch the flow of unskilled migration (see Chiswick & Hatton, 2005, pp. 109-112). The success of this policy shift has varied widely among OECD countries. As indicated in Table 20, Anglo-Saxon countries tend to have foreign-born populations that are better qualified than the native-born citizens, while for the majority of the Continental European and Nordic countries, the opposite is true. Indeed, countries such as “Canada, Australia, and New Zealand [have] placed far greater emphasis on the applicants’ skills in the rationing of immigration visas” (Chiswick & Hatton, 2005, p. 96). As a result, it is not surprising that “figures released recently by the European Commission showed that 85 percent of unskilled labor from developing countries had gone to the European Union and only 5 percent to the United States, whereas 55 percent of skilled labor had gone to the United States and only 5 percent to the Union” (Bilefsky, 2007, p. 1). Now that several EU member countries have already launched separate initiatives (for example, the German “green card” initiative to attract foreign information technology specialists in 2000), European Union Commis91

For a detailed overview and analysis of the immigration impact on national wage levels, see also Chiswick & Hatton (2005, Chapter 2.4) and Greenwood & McDowell (1994, pp. 169-189). For a more general study, see Nannestad (2007) who provides a comprehensive survey of the last 15 years of research on the welfare state – immigration nexus.

m\ P m mK m Pm m+ m PA m P PA P A PP

m

+D(I:$@%"A

 A K \ mK  \ A\ P
P+ m\ mK A AP A P
A


 P  
P \ 
P + P\ 
P P + PK P\

J6xl6 0egh,xhegx 0l?dl6 0egh,xe0ed 6?oh, ?6?oh, KK KK KK KK KK KK KK KK KK KK KK KK KK KK KK KK KK

U__x6l0el + m
Am K m
A AA mm K Am P A + K mP m mK

!6llx?#hex 6x l6 0egh,

/"$I6A@%"A

mm P
A m A PA m



Pm P P P+ P\ m P


J6xl6 0egh,xhegx 0l?dl6 0egh,xe0ed ?6?oh,

m  PP + 

+ A m P K +  Pm Pm A P

6?oh,

KK KK KK KK KK KK KK KK KK KK KK KK KK KK KK KK KK

U__x6l0el

Note: The data on the educational attainment of the native- and foreign-born populations is presented as a percentage of the population aged 15 and above (within each group, circa 2000). Given that the figures have been rounded up the total number sometimes surpasses 100. Source: Adapted from OECD (2007b, n. p.).

Ul?h_oh Ul?oh %6_Sothehgh F6e-h Goe_heg Ghe 6 6-he, 46_heg 4?h_, hhe B6?#6_hegl B0)h, .)6g6e .)o?6_heg JQ J.U

t0e?o6l

!6llx?#hex 6x l6 0egh,

Table 20: The Educational Attainment of the Native- and Foreign-Born Populations (circa 2000)

6.2 The Three Dimensions of Economic Globalization

91

92

6 Economic Globalization and the Contemporary Capitalist Welfare State

sioner for Freedom, Security and Justice Franco Frattini has recently proposed the introduction of an EU-wide “blue card” system that would grant work permits based on a points system to non-EU citizens in order to combat rising skill shortages (Dixon, 2007).92 Although Anglo-Saxon countries have enforced immigration policies that place more emphasis on the migrants’ skill levels, some academic researchers feel that this alone does not explain why the educational levels of foreign-born populations vary so widely between Anglo-Saxon countries on the one hand and Continental European and Nordic countries on the other. Following Razin & Sadka’s (2001) line of argument, the extent of low-skilled migration is also determined by the generosity degree of a national welfare system since “higher taxes or benefits […] [are] expected to lead to more immigration of low-skilled workers (with higher-education immigrants moving for reasons other than benefits)” (p. 79). This view broadly concurs with Borjas’ (1999) welfare magnet hypothesis. Studying welfare participation rates of immigrants across US states, he finds “that less-skilled immigrants […] are much more heavily clustered in high-benefit states than immigrants who do not receive welfare, or than natives” (p. 609). Against this backdrop, migratory pressures may likely induce national policymakers to tighten social spending programs (Glyn, 2007, p. 102; Sinn, 1997, p. 3; see also Razin & Sadka, 2004, pp. 709-711). Recently, an academic debate has launched about the effects of migration on the fiscal sustainability of welfare programs – mainly pension systems – in OECD countries. Razin & Sadka (2000) demonstrate in a theoretical model involving a dynamic set-up that inflows of unskilled migrants can ease public finance problems of payas-you-go pension systems. The following two assumptions are hereby made: Firstly, the time horizon of the economy is infinite whereas a citizen’s life span is limited to two generations (working-age and retirement). Secondly, the economy has sufficient access to international capital markets such that factor prices do not change due to migration. Under these circumstances, inflows of working-age immigrants – even if the latter are net welfare beneficiaries over the life cycle – exert a positive effect on the fiscal sustainability of the pension system since the children of these immigrants contribute to the welfare system in the second period thus easing pension payments of the current working age population (pp. 464-473). To sum up, in a dynamic scenario with open capital markets “migration is a Pareto-improving measure. That is, all the existing income (low and high) and all age (young and old) groups living at the time of the migrants’ arrival would be better-off ” (Razin & Sadka, 2000, p. 477). 92

The EU “blue card” system implies that “workers would be able to come to an EU member state to work for two years. After that, they could apply for a one-year extension to stay within the same country or move to another European country” (Dixon, 2007).

6.2 The Three Dimensions of Economic Globalization

93

In contrast to Razin & Sadka (2000), Schou (2006) applies a computable general equilibrium model (DREAM93) assessing the overall effect of immigration on the economy. More concrete than Razin & Sadka’s (2000) theoretical approach, Schou’s (2006) empirical method takes a comprehensive look at the impact of immigration on Denmark’s fiscal sustainability: “The role of immigrants is an important part of the picture to understand the projection of the Danish economy, as the share of immigrants and their descendants in the population rises in the future, and their economic and demographic behaviour, concerning age distribution, participation rates, wage/ productivity levels, unemployment levels, take-up rates of public consumption and (re-) emigration and fertility rates, differs in various ways from that of the RP [remaining population]. These various differences may have opposite signs influencing the total effect of immigration upon sustainability of public finances” (p. 688). Taking all these characteristics into account, Schou (2006) detects a slightly negative effect of immigration on fiscal sustainability in the Danish case (p. 688). A positive effect is retrieved only under the hypothetical, very unrealistic assumption of “full, costless and immediate economic integration of the immigrants and descendants in the base-line population forecast” (Schou, 2006, p. 688). In other words, unless restricted “to immigrants from relatively developed countries” (Schou, 2006, p. 688), immigration is a burden rather than a benefit for the sustainability of public finances in hosting countries. This finding is broadly in line with Nannestad (2007) who has conducted a survey of 15 years of research on the welfare state – immigration nexus and has come to the following conclusion: “Everything considered, there is not much empirical evidence to support a conclusion that immigrants arriving during the last 15 years have been economic assets for Western welfare states. The main reason is relatively low labor market participation rates and relatively high unemployment levels. It also appears that the negative impacts of immigration are strongest in the most generous welfare states. Thus one may possibly consider it a stylized fact that immigration into Western welfare states of predominately low-skilled immigrants is advantageous to the immigrants, but not for natives, in the short run, and that the gains from migration are most asymmetrically distributed between immigrants and natives in the most generous welfare states” (pp. 528-529). To summarize, the contemporary period of large-scale migration into OECD countries is somewhat similar to the pre-WWI era with respect to annual immigration 93

According to Schou (2006), “the computable general equilibrium model DREAM [Danish rational economic agents model] features an overlapping generation structure with 17 generations of representative households. The model represents a small open economy with a fixed exchange rate regime and perfect international mobility of financial capital” (p. 674). For further details about the model, see Schou (2006, pp. 674-680).

94

6 Economic Globalization and the Contemporary Capitalist Welfare State

rates. Yet it differs in two qualitative ways: Firstly, while the migration regime of the classical gold standard era was characterized by large inflows of unskilled labor, which found extensive employment opportunities in the agricultural and industrial sector of the host economy, the demand for low-skilled labor from overseas has significantly declined during the last decades. Particularly since the 1970s, there has been a shift in immigration policies to target more highly skilled migrants (Chiswick & Hatton, 2005, pp. 110-111). Secondly, the current period has been shaped by the increasing role of public welfare provision. In contrast to the pre-WWI era, when social policy was still in its infancy, measures of income redistribution have since been institutionalized in all advanced market economies in order to contain the rising income inequality resulting from globalizing product and labor markets as well as technological change. With low-skilled migrants continuing to flock to OECD countries in the hopes of a better economic future, the share of government transfer programs in national income will probably keep on rising. As a consequence of the increased fiscal burden imposed by unskilled migration, advanced market economies will likely maintain an immigration policy targeting high-skilled immigrants (Chiswick & Hatton, 2005, p. 111).

6.3

The Effects of Global Economic Forces on the Welfare State: Four Different Hypotheses

Numerous studies have found a link between economic globalization and the welfare state. Yet debates continue to roil with respect to the force (including its direction) that global economic processes exert on national systems of social protection. Reviewing recent classifications by Andersen (2003), Brady et al. (2005, pp. 922-925), Brady et al. (2007, pp. 318-320), Ellison (2006, pp. 48-59), Garrett & Mitchell (2001, pp. 149-153), Genschel (2004), Hicks (1999, pp. 157-168, Chapter 7), Koster (2008, pp. 2-3), Schulze & Ursprung (1999), Swank (2003, pp. 59-61), and Swank (2002/ 2004, pp. 20-58), the author distinguishes between two broad schools of thought asserting that economic globalization exerts a fundamental influence on the contemporary welfare state. Firstly, globalization theorists consider global economic processes in product (trade in goods and services) and factor markets (international capital and migration flows) to have had a major impact on public social spending activities in all affluent democracies. This so-called Openness Literature, which summarizes the spectrum of globalization theories with respect to recent welfare state development, revolves around three perspectives: the competitiveness view, the compensation view, and the curvilinear view (Hicks, 1999, pp. 204-208).

6.3 The Effects of Global Economic Forces on the Welfare State

95

Secondly, proponents of the so-called context-contingent view94 or “‘weak globalization’ thesis”, as Ellison (2006, p. 48) coins it, acknowledge that global economic processes have been augmenting in intensity; they nevertheless claim that the direction and force of these globalization effects on national systems of social protection are shaped by respective countries’ institutional context. Consequently, adjustment strategies – being influenced by path dependency – are assumed to vary leaving the general diversity of existing welfare state clusters more or less unchanged (Bonoli & Palier, 2001, p. 58; Ellison, 2006, pp. 48-49; Leibfried & Obinger, 2001, p. 5; Scharpf, 2000, p. 224; Swank, 2003, pp. 60-61; Swank, 2002/ 2004, p. 5).95 In the following, the underlying hypotheses of both schools of thought will be presented in close detail.

6.3.1

The Openness Literature

6.3.1.1 The Competitiveness View According to proponents of the competitiveness view, national politics are considered today as subordinate to the borderless global economy (Ohmae, 1995; Ohmae, 1991; Strange, 1996). Following Reinicke (1998), the contemporary process of economic globalization poses a significant challenge to the internal sovereignty of the nation state: “Specifically, since the organizational logic of globalization induces corporations to seek the fusion of multiple, formerly segmented national markets into a single whole, it generates an economic geography that subsumes multiple political geographies. As a result, a government no longer has a monopoly of legitimate power over the territory within which corporations organize themselves. The greater the mismatch between political and economic geography, the more difficult it will be for national governments to act in an inclusive manner, allowing individuals to coexist and interact in a relatively predictable environment” (p. 65). With national borders continually blurring, world-wide competition has intensified, resulting in locational competition between nation states to attract mobile production factors (Siebert, 1999/ 2000, p. 241).96 In this context, Siebert (1999/ 2000) specifies that competition takes place today on three different levels: 94

The author coined this term to best describe the essence of the literature on globalization and varieties of welfare capitalism. 95 Although Ellison (2006) concurs with the view that globalization forces do not exert a uniform impact on welfare state systems, he nevertheless observes a “drift” of social policy towards greater liberalization in most advanced capitalist welfare states (pp. 48-49, 178-179). 96 For a detailed overview of the concept, the channels and the instruments of locational competition, see Siebert (1999/ 2000, Chapter 12) and Siebert (2000).

96

6 Economic Globalization and the Contemporary Capitalist Welfare State

“First, firms compete in the world product markets. Second, countries compete in the world market for capital, for technological knowledge, for high-skilled mobile labor, and, to some extent […], for residents. Third, there is a complex and intricate system of interrelations between the immobile production factors of countries, especially labor, by which these immobile factors also compete on a worldwide scale” (p. 241). Thus, the competition for locational advantage, fuelled mainly by what Hirschman (1970) has termed the exit option97 (p. 4), has given rise to a new systems competition98, which fundamentally alters the role and capability of the state (Sinn, 2001, p. 3; see Tanzi, 2002, pp. 122-127; Tanzi, 1998). While in the past national governments were fully able to determine “the conditions under which transnational economic transactions would take place” (Scharpf, 2000, p. 191), nowadays, “locational competition changes the cost-benefit calculus of national politics […] [and] reduces governmental power” (Siebert, 1999/ 2000, p. 250). More precisely, Scharpf (2000) explains the status quo of national politics in the following way: “Since governments and unions are no longer dealing with captive capital owners and captive consumers, national systems of taxation, regulation and industrial relations have now become vulnerable to the extent that they reduce the attractiveness of the national economy to mobile capital and the competitiveness of nationally produced goods and services in international product markets” (p. 196). Alesina & Perotti (1997) have conducted an empirical study using a panel of 14 OECD countries in order to analyze the welfare state’s potential burden on the competitiveness degree of the overall economy. They have found the following result: If the welfare state aims to finance the old and unemployed by redistributing income through labor taxation higher labor costs and substantial loss in competitiveness will be a consequence since labor unions will demand higher wages. This, in turn, will eventually result in lower export demand and rising unemployment (pp. 921-922). It is worthwhile noting that the above described negative effect on the competitiveness of the economy is hump-shaped: “a given increase in taxation induces a larger increase in relative unit labor costs in countries with intermediate levels of centralization, and smaller increases in countries with highly centralized and highly decentralized labor markets” (Alesina & Perotti, 1997, p. 923). 97

In Hirschman’s (1970) terminology, the exit option refers to the consumer’s option to “stop buying the firm’s products” or to members’ option to “leave the organization” (p. 4). Applying the concept to the contemporary era, the exit option refers to the fact that production factors – being globally mobile – can leave or exit a country in case of unfavorable conditions. 98 According to Sinn (2001), the old systems competition was the competition between capitalism and communism (p. 3).

6.3 The Effects of Global Economic Forces on the Welfare State

97

In light of the increasing international economic pressures99, proponents of the competitiveness view therefore envisage a decline in state capacity (Cerny, 1994, p. 334) and a simultaneous erosion of the post-WWII welfare state (Sinn, 2003, pp. 64-65; see Sinn, 2001, pp. 13-15). Precisely, they contend that with contemporary welfare states facing increasing pressure to curtail taxation on mobile production factors (Razin & Sadka, 2005, p. 107), “it seems inevitable that the high-tax countries will have to succumb to the forces of tax competition and sharply cut their corporate tax rates” (Razin & Sadka, 2005, p. 4). As tax revenues fall, “the state […] [will] have to reduce its role as a direct provider of social protection” (Tanzi, 2002, p. 127).100 Empirical evidence for this hypothesis has lately been provided by a number of scholars: Rodrik (1997), for instance, finds that social spending cuts due to trade openness are particularly severe in countries which have lifted all capital account restrictions (p. 62). Supporting Rodrik’s (1997) results, Garrett & Mitchell’s (2001) panel data analysis of 18 OECD countries for the time period 1961-1993 reports a significant negative relationship between trade and various welfare state spending measures (total government spending, government consumption, and social security transfers) (pp. 168-169).101 6.3.1.2 The Compensation View The compensation view follows a different line of argument with respect to the relationship between globalization and the welfare state. In contrast to the competitiveness view, it assumes that processes of international economic integration will induce a rise in public welfare spending. This hypothesis has primarily originated in the works of Cameron (1978), Katzenstein (1985) and Ruggie (1982). Ruggie (1982) – adhering to Polanyi’s (1944) ground-breaking work The Great Transformation – has noted that the post-WWII international economic order had been built on the concept of embedded liberalism (pp. 392-393), the latter implying “a grand domestic bargain: societies were asked to embrace the change and dislocation attending international liberalization, but the state promised to cushion those effects by means of its newly acquired domestic economic and social policy 99

It is hereby worthwhile noting that pressures are not only arising from the product and capital market side, but also from international migration. As Sinn (1997) notes “each single welfare state will have an incentive to scale down its activities since this will deter the net recipients of public funds and attract the net contributors” (p. 3). 100 For a more detailed analysis of the hypothesis of tax competition, see Chapter 9.1. 101 In this context, however, it is worthwhile noting that the same authors do not find any significant empirical evidence for a negative effect of international economic integration on effective rates of taxation. Quite on the contrary, foreign direct investment appears to be positively associated with capital taxation while an increase in trade volume apparently induces the capital/ (labor + consumption) tax ratio to rise (Garrett & Mitchell, 2001, pp. 171-173).

98

6 Economic Globalization and the Contemporary Capitalist Welfare State

roles. Unlike the economic nationalism of the thirties, then, the postwar international economic order was designed to be multilateral in character. But unlike the laissez-faire liberalism of the gold standard and free trade, its multilateralism was predicated on the interventionist character of the modern capitalist state” (Ruggie, 1998, pp. 89-90; see Ruggie, 1982, pp. 392-393, 399). In this context, compensating citizens for increasing risks related to international economic integration was perceived as a means to remove their resistance against further liberalization (see Ruggie, 2003, pp. 93-94).102 Looking from a different angle, Cameron (1978) has found that the positive correlation between the degree of trade openness and the size of the public economy has been most pronounced in small open economies. Latter has been explained in the following way: Given their small size and high dependence on international markets, these countries could not enforce neo-mercantilist or protectionist measures to shelter their economy from international competitive pressures. As a consequence, they have developed corporatist arrangements that fostered the creation of a large public sector in order to cushion the negative effects of volatile international markets (pp. 1255-1258, 1260; see also Katzenstein, 1985, p. 9, 24). Taking a similar view, Katzenstein (2003) concludes that the “perceived vulnerability generated an ideology of social partnership that had acted like a glue for the corporatist politics of the small European states” (p. 11). Recently, several authors have confirmed the compensation view. Hicks & Swank (1992), for instance, show empirically that trade openness has had a significant positive impact on total social welfare expenditures (as a percentage of GDP) in 18 OECD countries for the period 1960-82 (pp. 666-667). Furthermore, Rodrik (1998b) studies various groups of countries and finds a significant positive correlation between exposure to external risk (defined as trade openness in conjunction with terms-of-trade variability) and social security and welfare spending in the group comprising of 19 OECD countries (pp. 1019-1021). Alesina & Wacziarg (1998) demonstrate that trade openness tends to have a significantly positive effect on the size of government transfer programs (pp. 319-320). Hicks (1999) confirms this result (p. 188). Brady et al. (2005) provide evidence that various facets of globalization appear to exert a positive impact on welfare spending.103 Finally, Hicks & Zorn (2005) analyze determinants of welfare state retrenchment and show empirically that trade openness as well as finan102

One should however note that nowadays Ruggie (1998) contends that the compromise is more and more “surpassed and enveloped externally by forces it cannot easily grasp” and that “it finds itself being hollowed out from the inside by political postures it was intended to replace” (p. 90). 103 For an overview of the empirical results strengthening the compensation view, see Brady et al. (2005, pp. 941-943).

6.3 The Effects of Global Economic Forces on the Welfare State

99

cial liberalization work against welfare state cuts. FDI outflows, by contrast, tend to operate for welfare state cuts, but lack statistical significance (p. 633, 648-649). 6.3.1.3 The Curvilinear View According to Hicks’ (1999) classification, the curvilinear view is the third perspective of the Openness Literature (pp. 204-208). It was originally developed by Rodrik (1997, pp. 89-95). Unlike the convergence process following a race to the bottom, this new convergence thesis implies an inverted U-shaped relationship between globalization and the welfare state. In this sense, it combines the compensation and the competitiveness view in a sequential order. Accordingly, the process of international economic integration is at first accompanied by a gradual augmentation of public income transfer programs in order to compensate workers for increasing risks associated with the opening up of the economy (for example, real wage fluctuations). Governments are expected to finance these welfare programs by raising the tax on capital. This strategy is successful until international capital mobility reaches a certain level. Beyond this threshold, public social spending is assumed to be cut in order to reduce the tax burden on mobile production factors which would otherwise escape from the country thereby undermining the welfare state’s fiscal stance even further (Rodrik, 1997, pp. 89-90; Hicks, 1999, pp. 206-211). Reviewing prior research, Brady et al. (2005) note that proponents of the curvilinear view expect welfare states to converge towards one common welfare state model which should be situated in the middle between the residualist welfare state model of the less globalized Anglo-Saxon countries and the universalist welfare state model of the highly globalized Nordic countries (p. 924). Several quantitative studies have recently confirmed the curvilinear relationship between globalization and the welfare state. Hicks (1999), for instance, provides empirical evidence that investment openness exerts a hump-shaped influence on governmental benefit payments to households (pp. 208-209, 213). Furthermore, Brady et al. (2005) find that net migration affects social security transfers in European and coordinated market economies in the curvilinear way explained above (p. 943).104

6.3.2

The Context-Contingent View – Globalization and Varieties of Welfare Capitalism

In contrast to the previous theoretical approaches, a new strand in academic literature (for example, Bonoli et al., 2000; Bonoli & Palier, 2001; Burgoon, 2001; Ebbinghaus & Manow, 2001a, 2001b; Ellison, 2006; Hall & Soskice, 2001/ 2004a, 2001/ 2004b; Kitschelt et al., 1999/ 2005; Leibfried & Obinger, 2001; Swank, 2003; Swank, 2002/ 104

Interestingly, Brady et al. (2005) also find evidence for a U-shaped curvilinear effect (p. 944).

100

6 Economic Globalization and the Contemporary Capitalist Welfare State

2004; Swank, 2001) has recently emerged asserting that the impact of global economic forces on the welfare state is determined by the respective country’s institutional context (Bonoli & Palier, 2001, p. 58; Ellison, 2006, p. 178; Leibfried & Obinger, 2001, p. 5; Swank, 2003, pp. 60-61; Swank, 2002/ 2004, p. 5). This so-called context-contingent view mainly builds on the insights from the Varieties of Welfare Capitalism Literature. As mentioned in Chapter 3, two strands of research form the core of this literature according to Ebbinghaus & Manow (2001b, p. 1): Hall & Soskice’s (2001/ 2004a, 2001/ 2004b) VoC approach in the field of comparative capitalism literature and Esping-Andersen’s (1990/ 1998, 2000) Three Worlds of Welfare Capitalism approach in the area of comparative welfare state research. Placing the firm and its relationships with actors in five different spheres of the political economy (vocational training and education, industrial relations, corporate governance, inter-firm relations, and intra-firm relations) at the core of the analysis (pp. 6-7), Hall & Soskice (2001/ 2004b) identify two ideal types of production regime: Firstly, liberal market economies shaped by firms, which “coordinate their activities primarily via hierarchies and competitive market arrangements” (p. 8); secondly, coordinated market economies characterized by firms, which engage mainly in “non-market relationships to coordinate their endeavors with other actors and to construct their core competencies” (p. 8). In line with the CC literature, the VoC approach advances the view that firms’ coordination practices follow a capitalist system inherent logic that can be subsumed under the concept of institutional complementarities (Hall & Soskice, 2001/ 2004b, pp. 17-18). Adhering to Aoki (1994), the latter term is defined by Hall & Soskice (2001/ 2004b) in the following way: “two institutions [in different spheres of a political economy] can be said to be complementary if the presence (or efficiency) of one increases the returns from (or efficiency of) the other” (p. 17). An often stated example in this respect is the interrelation of vocational training with measures of employment protection: Estevez-Abe, Iversen, and Soskice (2001/ 2004), for instance, demonstrate that employers pursuing product market strategies based on firm- or industry-specific skills are supportive of strong employment and/ or unemployment protection through legislation, collective agreements or company-based criteria (pp. 152-153; 165, 168, 170). Thus, given the presence of these institutional complementarities, Hall & Soskice (2001/ 2004b) conclude that each production regime provides firms with a particular set of comparative institutional advantages to engage in certain economic transactions (pp. 36-37). Complementary to the production regime analysis of the VoC approach, the second strand of literature – in particular Esping-Andersen’s (1990/ 1998, 2000) Three Worlds of Welfare Capitalism approach – focuses on the significance of institutional differences for national protection systems (Ebbinghaus & Manow, 2001b, p. 1). As mentioned in Chapter 2.3, Esping-Andersen (1990/ 1998) – classifying welfare states along three dimensions (the level of de-commodification, the system of stratification,

6.3 The Effects of Global Economic Forces on the Welfare State

101

and the degree of de-familialization) – identifies three welfare state regimes which cluster the following countries: Firstly, the liberal welfare state regime clusters Anglo-Saxon countries (pp. 26-27; Esping-Andersen, 2000, pp. 74-77). Secondly, the conservative welfare state regime clusters Continental European countries (p. 27; Esping-Andersen, 2000, pp. 81-83). Thirdly, the social democratic welfare state regime clusters Nordic countries (pp. 27-28; Esping-Andersen, 2000, pp. 78-81). Bringing together the vast amount of literature on different welfare state – economy inter-relationships, Ebbinghaus & Manow (2001b) primarily consider institutional complementarities between the social insurance system and the following three institutional domains of the political economy: “The first area is the system of industrial relations, comprising corporate actors, such as the state, labour and employers, and the laws and rules governing the conditions of employment relations, most importantly via collective bargaining. The second area entails the production system and employment regime, which involves the social organisation of the production process and the employment strategies of firms. Finally, the financial and corporate finance system is the third area, which defines the specificity of a capitalist system: the way financial markets are organised, companies are controlled and investments are financed” (pp. 12-13). Comparing advanced capitalist welfare states and their respective political economies along the above stated institutional complementarities, the author has identified – largely in line with prior research (for example, Huber & Stephens, 2001a, pp. 85-105; Huber & Stephens, 2001b; Kitschelt et al., 1999/ 2005, pp. 429-438; Swank, 2003, pp. 60-72) – the following three varieties of welfare capitalism: Liberal (Anglo-Saxon countries), conservative (Continental European countries and Japan), and social-democratic (Nordic countries) welfare economies (see Table 2 in Chapter 2.3.2 and Table 6 in Chapter 3.3). Given the fact that institutional configurations vary significantly between each variety of welfare capitalism, reform paths will do so too.105 As a result, advocates of the context-contingent view assume that global economic forces do not have a uniform impact on all advanced capitalist democracies (Kitschelt et al., 1999/ 2005, pp. 449-457) and their respective welfare states (Bonoli & Palier, 2001, p. 58; Ellison, 2006, p. 178; Leibfried & Obinger, 2001, p. 5; Scharpf, 2000, p. 224; Scharpf & Schmidt, 2000, p. 335; Swank, 2003, pp. 60-61; Swank, 2002/ 2004, p. 5). While a common consensus exists in favor of persistence of diverse patterns of welfare capitalism, disagreement prevails regarding the question which welfare capitalist systems will be rather affected by global economic pressures. Hall & Soskice 105

In this context, Chapter 5 has already provided extensive evidence that adjustment strategies as a response to challenges arising in the domestic realm differ quite dramatically between the three welfare state clusters. Now, the author will focus exclusively on the potential impact of global economic forces on the three clusters.

102

6 Economic Globalization and the Contemporary Capitalist Welfare State

(2001/ 2004b) consider Anglo-Saxon welfare economies – being market-driven and therefore rather accommodative to liberalization tendencies – as more prone to globalization forces in comparison to their Continental European (including Japan) and Nordic counterparts: “In the face of more intense international competition, business interests in LMEs [liberal market economies] are likely to pressure governments for deregulation, since firms that coordinate their endeavors primarily through the market can improve their competencies by sharpening its edges. The government is likely to be sympathetic because the comparative advantage of the economy as a whole rests on the effectiveness of market mechanisms. […] in coordinated market economies, however, the political dynamic inspired by globalization should be quite different. Here, governments should be less sympathetic to deregulation because it threatens the nation’s comparative institutional advantages. Although there will be some calls for deregulation even in such settings, the business community is likely to provide less support for it, because many firms draw competitive advantages from systems of relational contracting that depend on the presence of supportive regulatory regimes” (pp. 57-58; see Swank, 2003, pp. 74-75, 79). Swank (2002/ 2004) draws a similar conclusion; however, by doing so, he does not primarily adopt a firm-centered perspective, but rather looks at it from a broader political-institutional angle. Precisely, he considers the welfare state and political regime specific context106 as the structuring force that determines how globalization affects the respective systems of social protection: “Where universalistic and corporatist conservative welfare state structures are strong and in political institutional contexts of moderate-to-strong social corporatism, inclusive electoral institutions, and centralized policy-making authority, the conventionally hypothesized globalization dynamics are absent. Internationalization has no direct and systematic impact on social welfare policy change, or it is related to small positive increments in social protection in these institutional contexts” (pp. 279-280). Naturally, where the opposite holds (as in Anglo-Saxon countries), global economic pressures are expected to induce major welfare state cuts (Swank, 2002/ 2004, p. 279). In contrast to this view, Scharpf & Schmidt (2000) contend that adjustment to contemporary challenges in a globalizing economy is most severe in Continental European welfare economies (p. 335) since the latter have institutionalized “pay-as-yougo, contribution-based, and earnings-related pension systems which […] create a 106

Apart from the central dimensions constituting each welfare state regime (see Chapter 2.3 and 3), Swank (2001) regards three political-institutional features as key to the welfare state and political regime specific context: the mode of interest representation, the embodiment of the electoral system, and the level of centralization within the political decision-making process (pp. 207-212; see Bonoli et al., 2000, pp. 144-151).

6.3 The Effects of Global Economic Forces on the Welfare State

103

large tax wedge, discourage female labor participation, exacerbate intergenerational conflict, and are highly vulnerable to international regulatory competition” (pp. 333; see Ellison, 2006, p. 59). Finally, focusing on the global economic pressures arising out of international migration, Nannestad (2007) holds the view that their fiscal impact will be most severe in the generous Nordic welfare states since latter attract in particular low-skilled immigrants (pp. 528-529).107 As a result, he envisages two adjustment strategies: Firstly, policy-makers can restrict immigration to skilled workers. Secondly, policymakers can down-scale the welfare state in order to dissuade migration (p. 529).

107

For a more detailed elaboration on this issue, see Chapter 6.2.3.

Part III Empirical Analyses of the Driving Forces Behind the Contemporary Capitalist Welfare State Reviewing the two prior chapters, advanced capitalist welfare states face challenges both in the domestic as well as global sphere. As mentioned, academic debates center on the question of force that global economic processes relative to domestic political and economic pressures exert on national systems of social protection. Against this backdrop, Part III conducts in-depth empirical analyses in order to test all prevailing theories on the driving forces behind the contemporary capitalist welfare state. Similar to prior studies in this field of research (for example, Huber & Stephens, 2001a; Swank, 2002/ 2004), the author applies a hybrid approach comprising of quantitative and qualitative analyses (“triangulation”108). Given the complexity of methodological issues in comparative welfare capitalism research, Chapter 7 explains in detail why this empirical approach has been chosen. Subsequently, Chapter 8 presents various econometric models and their respective findings. Finally, Chapter 9 conducts a set of qualitative analyses shedding light on both the revenue and expenditure side of national systems of social protection in all three welfare state clusters.

108

According to Shalev (2007), “triangulation […] means combining MR [Multiple Regression] with other types of analysis – quantitative, qualitative or both” (p. 296).

7

Methodological Issues in Comparative Welfare Capitalism Research

According to Hall (2003), one of the major problems in the field of comparative social sciences is the growing discrepancy between methodologies and ontologies109: In his eyes, comparative research has recently turned to standard multivariate regression models (MR) as the predominant mode of statistical analysis (p. 374). At the same time, ontologies have shifted “toward theories, such as those based on path dependence or strategic interaction, whose conceptions of the causal structures underlying outcomes are at odds with the assumptions required for standard regression techniques and conventional comparative method to provide valid causal inferences (cf. Bates, Greif, Levi, Rosenthal, and Weingast 1998; Pierson 2000a)” (p. 375). Thus, given the causal complexity of prevailing theories, he contends that “the ontologies of comparative politics have substantially outrun its methodologies” (Hall, 2003, p. 375). Taking the same line, Shalev (2007) gives five reasons why MR is not an appropriate method for conducting research in the field of comparative political economy: First of all, he holds that the MR method – considered suitable for marginal effect estimations under the ceteris paribus conditions widely used in economic research – is less appropriate for comparative political economy because the latter discipline does not primarily estimate marginal effects (pp. 265-266), but rather attempts to assess which impact “the presence or absence of certain conditions” has on the macro-level of a national political economy (p. 266).110 Secondly, he doubts that researchers using the MR method can effectively model the causal complexity underlying prevailing

109

The term ontology refers “to the fundamental assumptions scholars make about the nature of the social and political world and especially about the nature of causal relationships within that world” (Hall, 2003, p. 374). 110 To illustrate the existing difference between economic and comparative political economic research, Shalev (2007) mentions two distinct cases: On the one hand, economists estimate the marginal effect of a certain variable on another, holding all other variables’ impact constant (p. 265). In this sense, a potential research question could be the following: “If prices rise, what will be the likely effect on economic growth, net of other known influences like the rate of investment and the terms of trade?” (p. 265). On the other hand, comparative political economists are interested in determining which impact the presence of a certain institution (for example, corporatism) has on the national political economy (p. 266). In this vein, a potential research question could be the following: “Will economic growth be higher in the presence of corporatist trade unions (or a hegemonic social-democratic party, or an independent central bank)?” (p. 266).

108

7 Methodological Issues in Comparative Welfare Capitalism Research

ontologies (p. 266).111 Thirdly, from a more pragmatic point of view, he notes that given the small number of observation cases in comparative political economy (small N problem), a profound empirical analysis is rarely feasible with MR; he therefore doubts the reliability of empirical findings that have been derived by these methods (Shalev, 2007, p. 267). Fourthly, Shalev (2007) criticizes the general practice of pooling datasets (combining cross-country and time-series data) in order to tackle “the problem of ‘too many variables and not enough cases’” (p. 269). According to his view, “most pooled designs utilize multiple cross-sections in order to fortify comparative generalizations, or multiple time series to fortify dynamic generalizations, on the implicit assumption that there is no difference in causality between the two dimensions” (p. 280). Yet, as he points out, prior works (for example, Korpi & Shalev, 1980) have clearly demonstrated that the effect of an explanatory variable may differ across time-series and cross-country variation (pp. 279-280).112 Finally, from a more technical point of view, Shalev (2007) notes that the MR technique can lead to biased results since “it estimates partial parameter effects as if all (linearly-fitting) configurations were possible” (p. 267).113 This notion has recently also been shared by Scruggs (2007a) who recognizes that “many MR studies ‘span’ many empty cells and convey an impression of linear effects that is not really justified” (p. 316). 111

In this context, Swank (2007) notes that ”Shalev makes the distinct point that the theories we seek to test in comparative political research entail complex and often non-linear causal sequences […] [and] that the linear and additive logic of general MR analysis, as well as the more sophisticated versions with non-linear specifications and interaction terms, cannot adequately test our complex theories” (pp. 362-363). This is much in line with Abbott’s (1988) earlier criticism that sociologists all too often consider the social and political world “in terms of a ‘general linear reality‘“, with the latter assuming “(1) that the social world consists of fixed entities with variable attributes, (2) that cause cannot flow from ‘small’ to ‘large’ attributes/ events, (3) that causal attributes have only one causal pattern at once, (4) that the sequence of events does not influence their outcome, (5) that the ‘careers’ of entities are largely independent, and (6) that causal attributes are generally independent of each other” (p. 169). 112 Kenworthy (2007) illustrates the three different types of variation (variation across countries, variation over time, variation across countries and over time) in a graphical analysis (p. 347). 113 To illustrate, Shalev (2007) constructs a simple MR with social security expenditures (as percentage of GDP) as the dependent variable and left party power, trade exposure, and the elderly population (as a percentage of total population) as independent variables. He then assumes that all variables are measured on a 5-point scale, opening up a multiway table of 625 (5*5*5*5) cells. Given the fact that most comparative studies on OECD countries include no more than twenty countries, at maximum only twenty cells are actually filled with data (p. 267). As a consequence, Shalev (2007) concludes that the regression method “in effect places imaginary countries in some of these empty cells when it seeks out the best linear fit that can be generated for the data at hand [.]. Because it estimates partial parameter effects as if all (linearly-fitting) configurations were possible, MR can easily yield problematic results” (p. 267).

7 Methodological Issues in Comparative Welfare Capitalism Research

109

Several leading scholars (Esping-Andersen, 2007; Kenworthy, 2007; Pontusson, 2007; Rothstein, 2007; Rubinson & Ragin, 2007; Scruggs, 2007a; Swank, 2007) have recently responded to Shalev’s (2007) critical comments about the use of MR in comparative political economy. While acknowledging the fact that comparativists “too often […] oversimplify tests of complex theories” (Swank, 2007, pp. 369-370), some hold the view that MR, if specified properly, can sidestep the caveats mentioned by Shalev. Pontusson (2007), for instance, makes the point that regression analysis does not assume a sole “linear-additive conception of causality” (p. 326); rather, controlling for conditional causal complexity has become a central characteristic of the MR models applied in recent papers (p. 326). In this context, Pontusson (2007) refers to an earlier work by Rueda & Pontusson (2000) who have examined if causal relationships vary between social, liberal and mixed market economies. For this purpose they have “dummy variables for each of these political-economy types [interact] with all the independent variables included in their model” (Pontusson, 2007, p. 327) and find clear evidence that several variables (for example, government partisanship) tend to have different effects on wage inequality depending on the type of market economy (Pontusson, 2007, p. 327; see Rueda & Pontusson, 2000, pp. 370-374). Just as causal heterogeneity is analyzed across country clusters by Rueda & Pontusson (2000) it can also be tested over time by interacting dummy variables for specific time periods with all the independent variables comprising the model (Pontusson, 2007, p. 328).114 MR, in other words, can be applied to both linear and non-linear causal sequences. Under the assumption that the regression model is specified properly, pooling datasets across countries and time is therefore not perceived as problematic (Pontusson, 2007, pp. 331-332). On the contrary, proponents of the regression analysis are convinced that it provides “more accurate estimates of the ‘statistical associations’ to be explained” (Pontusson, 2007, p. 330) than Shalev’s (2007, pp. 296-297) alternative methods (for example, factor analysis, tree diagrams). Precisely, they contend that these alternative “techniques only work well with two or three explanatory variables and, with three explanatory variables, the interaction effects among these variables become quite opaque” (Pontusson, 2007, p. 329); MR, by contrast, allows controlling for more effects and therefore provides the appropriate framework to model both linear and non-linear causal sequences (Pontusson, 2007, pp. 331-332). With respect to Shalev’s (2007) final critique that the MR technique can lead to biased results since “it estimates partial parameter effects as if all (linearly-fitting) configurations were 114

In this context, Scruggs (2007a) notes that “conjunctural causation can essentially be accounted for by some type of ‘interaction term’ in a regression model. This would test whether the effect of two things together is greater (or less) than the sum of the parts. In a simple case, one can simply take the interaction as the intersection of two variables. If, for example, having A or B alone is jointly bad for you, but having A and B (together) is good for you, this can be incorporated into an MR model” (p. 313).

110

7 Methodological Issues in Comparative Welfare Capitalism Research

possible” (p. 267), Scruggs (2007a) feels that this problem is overstated. According to his view, the econometric literature provides various techniques (for example, residual diagnostics) to check for the robustness of regression estimations (p. 317). While disagreeing with Shalev on numerous accounts, proponents of the MR approach, however, concur with him on the following ground: They are convinced that “the operationalization and estimation of MR models should be the last […] stage in the development and testing of hypotheses. The earlier stages involve developing a theoretical model, mapping the theoretical model to observables, [and] developing a statistical model of the process generating the observables” (Scruggs, 2007a, p. 310). Theoretical causation, in other words, is not determined by empirical correlation but vice versa (Scruggs, 2007a, p. 310). Furthermore, several researchers have recently recognized that a comprehensive analysis of the dynamic process between causal effects and ultimate outcomes necessitates a combined methodological approach consisting of MR and other quantitative as well as qualitative techniques (Pontusson, 2007, p. 332; Swank, 2007, p. 370). Precisely, Swank (2002/ 2004) elaborates on this issue in the following way: “While one can derive estimates from the quantitative models of the effects of causal factors (e.g., international capital mobility) on the phenomenon of interest (social welfare protection) in specific cases (i.e., nations), these estimates and the important inferences we derive from them are only as good as our conceptualization, measurement, data sources, and statistical estimation techniques. Moreover, despite the power of ‘large-N’ quantitative tests of hypotheses to contribute to theoretical knowledge (e.g., by assessing the generalizability of propositions), the analysis often leaves many questions concerning causal sequences, actors’ motivations and the interpretation of events by political agents, and how nation-specific contexts modify causal effects” (p. 10). In light of these methodological constraints, several researchers in the field of comparative social sciences (for example, Huber & Stephens, 2001a; Swank, 2002/ 2004) have begun to complement MR analysis with different types of qualitative analyses.115 Accordingly, the author also pursues a hybrid approach comprising of MR (Chapter 8) and qualitative analyses (Chapter 9).

115

Combining quantitative and qualitative techniques (a method referred to as “triangulation”) has also been proposed by Shalev (2007) as a means to overcome the shortcomings associated with MR analysis (p. 296). In this context, Pontusson (2007) has suggested conceiving “the ‘low-tech’ quantitative methods advocated by Shalev as a bridge between MR and theoretically informed, process-oriented case studies” (p. 332).

8

A Quantitative Study on the Determinants of Welfare Spending

Applying a set of various pooled cross-section time series models, Chapter 8 will conduct a quantitative analysis in order to test prevailing theories on the driving forces behind the contemporary welfare state. The dataset will include 17 OECD countries for the time period 1975-2000. Before presenting the econometric approach and the respective empirical findings in close detail, the author will analyze current patterns of public social spending across advanced welfare economies.

8.1

Patterns of Public Social Spending

The ongoing debate on the “dependent variable problem” discusses which variable would be most appropriate to conceptualize and measure welfare state development (Green-Pedersen, 2004, pp. 3-4; Allan & Scruggs, 2004, p. 497; Korpi & Palme, 2003, pp. 426-427, 432-433). So far numerous studies in the area of comparative welfare state research (for example, Burgoon, 2001; De Grauwe & Polan, 2005; Hicks & Swank, 1992; Rodrik, 1998b; Rodrik, 1997) have been applying aggregate social spending data (for example, government expenditures, public social expenditures, social security transfers as a ratio of GDP) in order to capture welfare state commitments. However, objections against the use of these aggregate indicators are wellfounded: From a conceptual point of view, Esping-Andersen (1990/ 1998) remarks that welfare effort should be perceived in levels of de-commodification. In particular, he notes that “if our aim is to test causal theories that involve actors, we should begin with the demands that were actually promoted by those actors that we deem critical in the history of welfare-state development. It is difficult to imagine that anyone struggled for spending per se” (p. 21). Supporting this view, Allan & Scruggs (2004) note that aggregate data is a bad measure since it “cannot tell us very much about how, or on whom, the money is spent” (p. 498). Moreover, Allan & Scruggs (2004) point out three caveats with respect to the use of aggregate social spending data: First of all, an increase in the welfare-benefitreceiving population can obscure actual spending cuts on the individual citizen level (p. 498). Secondly, given the fact that services provided by the state exhibit inherently lower productivity growth relative to other sectors of the economy (Baumol’s cost disease), social security expenditures are assumed to augment relative to GDP (p. 498). Thirdly, “differences in the tax treatment of transfers […] distort the degree to which social spending, as measured in national accounts, translates into disposable

112

8 A Quantitative Study on the Determinants of Welfare Spending

income for program recipients” (p. 498), further complicating comparative welfare state research. In light of these objections against the use of aggregate spending data, Scruggs (2004a) – building on Esping-Andersen’s (1990/ 1998) de-commodification indices – has constructed the Overall Benefit Generosity Index (hereafter referred to as Benefit Generosity Index), which looks at welfare effort on an individual basis. More precisely, Scruggs’ 2006 updated index is composed of three major social insurance programs (public pension, unemployment insurance, and sick pay insurance), which are evaluated according to the following three criteria: net replacement rates, qualifying conditions, and coverage rates (Scruggs, 2004b, pp. 1-5).116 The use of the Benefit Generosity Index not only circumvents the problems related to aggregate data, but should also give a clear picture of how welfare policy changes affect the individual life chances of individual citizens (see Allan & Scruggs, 2004, p. 498). Against the background of the ongoing “dependent variable” debate, the author chooses to use both aggregated and disaggregated variables in order to adequately measure welfare effort. Similar to Brady et al.’s (2005, p. 927) prior study, the first three dependent variables involve the following welfare state indicators: the Benefit Generosity Index117, public social expenditures as a ratio of GDP (hereafter referred to as PSEX)118, and social security transfers as a ratio of GDP (hereafter referred to as SSTRAN)119. The fourth dependent variable is Scruggs’ (2004a) Pension Generosity Index (the third sub-component of the Benefit Generosity Index besides Unemployment Generosity and Sickness Generosity).120 The latter has been chosen for the following reasons: Firstly, public pensions represent the most important welfare state 116

For more detailed information on the construction of the index, see also Scruggs & Allan (2006, pp. 62-67) as well as Scruggs (2007b, pp. 139-161). 117 Brady et al. (2005, p. 927) use Scruggs’ de-commodification index, which is an older version of the Benefit Generosity Index. For more information on the Benefit Generosity Index, see Scruggs’ Comparative Welfare Entitlements Dataset (http://www.sp.uconn.edu/ ~scruggs/wp.htm). 118 In line with Lindbeck’s (2006) definition, PSEX refers to the sum of “(i) cash benefits to households (transfers, including mandatory income insurance) and (ii) subsidies or direct government provision of human services (such as child care, pre-schooling, education, health care, and old-age care)” (p. 2). 119 SSTRAN refers to the sum of all cash benefits paid to households as a ratio of GDP (Lindbeck, 2006, p. 2). See also Chapter 2.1 for an explanation of the different types of welfare state benefits. 120 The Pension Generosity Index is evaluated according to the following three criteria: pension net replacement rate, qualifying conditions, and coverage rates (Scruggs, 2004b, pp. 1-5). For more detailed information on the construction of the index, see also Scruggs & Allan (2006, pp. 62-67) as well as Scruggs (2007b, pp. 139-161).

113

8.1 Patterns of Public Social Spending

program in the vast majority of affluent democracies.121 Secondly, the Pension Generosity Index as a dependent variable offers an appropriate measure to test the new politics approach, which considers retirees to constitute a powerful political force capable to block politicians’ retrenchment efforts (Pierson, 1996, pp. 174-175). Thirdly, from a methodological point of view, it is worthwhile examining in how far results differ between the regressions using the Benefit Generosity Index and the Pension Generosity Index as dependent variable. Similar to Brady et al.’s (2005, pp. 933-935) prior study, Figure 7 depicts the average development of the four welfare state indicators in 17 OECD countries for the period 1975-2000. In the case of PSEX, average welfare spending rose sharply until 1993 and then fell by approximately eight percent of the 1993 peak value until 2000. SSTRAN experienced similar growth; however, the ongoing welfare retrenchment process starting in the early 1990s has so far been more pronounced (15 percent of the 1993 peak value until 2000). Looking at the Benefit Generosity Index and the

–䉱–

Figure 7: The Average Development of Four Welfare State Indicators in 17 OECD Countries, 1975-2000 Note: Data on PSEX is only available starting in 1980. Source: Data compiled from Huber et al. (2004a), OECD (2007e), Scruggs (2004a).

121

With respect to the 17 welfare states under empirical investigation, only Australia, Canada, Ireland, and the United States spend more on public health than on public pensions. For a detailed analysis, see Table 40 in Chapter 9.2.2.1 and Appendix 5 and 6.

114

8 A Quantitative Study on the Determinants of Welfare Spending

Pension Generosity Index, a different picture emerges: Firstly, welfare spending peaked by as early as 1986. Secondly, welfare retrenchment in the 1990s has been less dramatic (only five and seven percent respectively of the 1986 peak value until 2000). In summary, all four social spending indicators show clear signs of welfare expansion until the mid 1980s or early 1990s and a subsequent process of welfare retrenchment throughout the 1990s. Yet it has to be noted that all welfare state measures display higher values in 2000 than at the beginning of the financial globalization era in 1975. These findings are in line with Brady et al.’s (2005, pp. 933-935) earlier graphical analyses. Based on this general graphical analysis, the 17 OECD countries are now grouped according to Esping-Andersen’s (1990/ 1998, 2000) welfare state regime typology into a social-democratic (Nordic countries), a conservative (Continental European countries and Japan), and a liberal (Anglo-Saxon countries) welfare state regime cluster. Figure 8 to 11 depict the average development of the four welfare state measures for each country group. While all three welfare state clusters exhibit the above-described pattern of public social spending, both the expansion and the retrenchment process are particularly pronounced in the Nordic countries.122

–䉱–

Figure 8: The Average Development of Public Social Expenditures in the Three Welfare State Clusters, 1980-2000 Source: Data compiled from OECD (2007e).

122

See also Scruggs (2006) who conducted graphical analyses of unemployment, sickness and pension net replacement rate trends across liberal, conservative and social-democratic welfare state clusters (pp. 352-358).

115

8.1 Patterns of Public Social Spending

–䉱–

Figure 9: The Average Development of Social Security Transfers in the Three Welfare State Clusters, 1975-2000 Source: Data compiled from Huber et al. (2004a).

–䉱–

Figure 10: The Average Development of the Benefit Generosity Index in the Three Welfare State Clusters, 1975-2000 Source: Data compiled from Scruggs (2004a).

116

8 A Quantitative Study on the Determinants of Welfare Spending

–䉱–

Figure 11: The Average Development of the Pension Generosity Index in the Three Welfare State Clusters, 1975-2000 Source: Data compiled from Scruggs (2004a).

8.2

The Econometric Analysis

8.2.1

Contribution to Comparative Welfare State Research

Brady et al.’s (2005) model provides a basis for the econometric analysis because it takes a truly holistic approach to testing all of the major theoretical hypotheses on the determinants of public social spending in OECD countries.123 In order to make the results of this quantitative study comparable with their earlier findings, the author has opted to use to a certain extent a similar dataset for the explanatory variables as well as the welfare state measures. Moreover, with respect to model operationalizations, the author follows Brady et al. (2005) in relying to a great extent on Huber & Stephens (2001a, 2000) prior works. Although the econometric analysis is built on previous academic studies, it also addresses a number of limitations of past research. Specifically, the following contributions are made to contemporary comparative welfare state research: Firstly, to the 123

It should be noted that Brady et al. (2005) have heavily relied on Huber & Stephens (2001a, 2000) when constructing their own baseline model (p. 927).

8.2 The Econometric Analysis

117

best of the author’s knowledge, this quantitative study is one of the first of its kind to link the VoC and the Three Worlds of Welfare Capitalism approach in a comprehensive econometric analysis for the recent period. While Brady et al. (2005) conduct a series of sensitivity analyses for liberal/ non-liberal as well as European/ non-European welfare states (pp. 940-943), the author constructs a series of interaction models that simultaneously control for liberal, conservative, and social-democratic welfare state regime clusters. In this context, it should be noted that Swank (2002/ 2004) has already run regression analyses for each of the three welfare state clusters. The dataset of Swank (2002/ 2004) however includes only the time period 1965-1993 (pp. 103-121). Recently, Kim & Zurlo (2008) have conducted a similar comparative study using a dataset from 1980-2001. Secondly, this quantitative study sheds light on the rapid decline of public social spending in the Nordic countries in the early 1990s. Using a structural break model, the author analyzes the hump-shaped pattern of public social spending in the socialdemocratic welfare regime cluster in detail. Thirdly, with respect to the dependent variables, Scruggs’ (2004a) Pension Generosity Index as well as the 2006 updated version of the de-commodification index, Scruggs’ (2004a) Benefit Generosity Index, is applied. The use of both welfare state indicators not only circumvents the problems related to aggregate data, but should also clarify how welfare policy changes affect the individual life chances (see Allan & Scruggs, 2004, p. 498).124 Fourthly, the author uses both annual and cumulative political partisanship variables, but runs each model with only one of the four political partisanship variables (annual left, cumulative left, annual right, and cumulative right) at a time. By comparing annual and cumulative political partisanship models, the paper attempts to make a contribution to the ongoing debate about which measure is more appropriate for comparative welfare state research (see Hicks, 1999, pp. 189-193). Fifthly and more generally, this quantitative study takes a slightly different modeling approach to Brady et al. (2005). Building an econometric model that controls for a multitude of different theoretical approaches, is a very difficult task since the author is compelled to construct a parsimonious while at the same time comprehensive model. While Brady et al. (2005) placed more emphasis on comprehensiveness this study aims to strike a balance between comprehensiveness and parsimoniousness. Starting with ten principal independent variables, the baseline model is then reduced to include only seven variables. Unlike in Brady et al.’s (2005) study, global economic forces are not categorized according to globalization advantage, openness, or threat (pp. 928-929); rather, they are related to the product (Trade Openness for in124

For this reason, several scholars (for example, Allan & Scruggs, 2004; Brady et al., 2005; Fernandez, 2008; Korpi & Palme, 2003) have recently begun to use disaggregated welfare state measures.

118

8 A Quantitative Study on the Determinants of Welfare Spending

ternational trade in goods and services) and factor (FDI Openness for international capital mobility and Net Migration for international migration) markets across which they tend to affect public social spending.125 Sixthly, in contrast to many previous studies in comparative political economy, this quantitative study takes possible nonstationarity issues in time-series cross-section (TSCS) analysis into account. Nonstationarity, if left untreated, can pose serious problems when interpreting the findings in pooled models (Kittel & Winner, 2005, p. 288). Due to recent developments in econometrics, this study tested for stationarity using the Levin-Lin-Chu panel unit root test (Levin et al., 2002, pp. 2-3), a commonly used test for stationarity, available as a special coded program in STATA 9. According to Bornhorst & Baum (2001), “the test may be viewed as a pooled Dickey-Fuller test, or an Augmented Dickey-Fuller (ADF) test when lags are included, with the null hypothesis that of nonstationarity (I(1) behavior)” (n. p.).

8.2.2

Modeling Potential Determinants of Welfare Spending

As mentioned in Chapter 8.1, welfare effort will be modeled with four dependent variables (PSEX, SSTRAN, the Benefit Generosity Index, and the Pension Generosity Index).126 Since these have already been explained in detail, the author will focus on the independent variables in the following.127 125

Apart from differences in how to model globalization, the author has also not considered a number of domestic political and economic proxy variables which have been incorporated in Brady et al.’s (2005) earlier model. These variables include the following: Christian Democratic Cabinet, Union Density, Constitutional Structure, Female Labor Force ParticipationLeft Party Interaction Variable, Voter Turnout, Strikes, Authoritarian Legacy, GDP per Capita, Year, Military Spending, and Wage Coordination. 126 Some scholars might criticize the author for using level instead of change measures. However, according to Huber & Stephens (2001a), several reasons speak in favor of the chosen approach: First of all, most pooled time series analyses in comparative welfare state research have used level data since the development of the welfare state follows a slow, path-dependent process that can only be reasonably captured with level data. Annual change data, by contrast, would be incapable of mirroring the historical legacy of the welfare state (pp. 57-58). Secondly, annual change data is overwhelmingly influenced by economic cycles (p. 57). Thirdly, data errors might have a much greater impact on the results if annual change data are used than if level data are used. The latter is explained by the fact that the coefficients for the independent variables are generally much smaller for change data than for level data analysis (p. 58). Fourthly, “some important social policy programs, most notably pensions but also most other programs in which entitlements are based on the duration and /or level of contributions, mature over a long period of time. Thus, expenditure will gradually climb without any additional legislation by the sitting government” (p. 58). 127 With respect to independent variables, the model follows closely Brady et al. (2005, pp. 922-925, 926-929) in controlling for post-industrial and globalization effects.

8.2 The Econometric Analysis

119

8.2.2.1 Explanatory Variables as a Proxy for Post-Industrial Challenges The following three of Pierson’s (2001a) post-industrial pressures are modeled: the processes of deindustrialization, population ageing, and changing household structures (pp. 82-83; see Chapter 5.1-5.3). Furthermore, the author also tests the hypotheses underlying the power resources and the new politics approach (see Chapter 5.5). Employment in Agriculture and Industry (hereafter referred to as Employment in A & I) is used as a proxy variable for deindustrialization. According to Iversen & Cusack (2000), falling employment in agriculture and industry should have a positive effect on welfare spending (pp. 324-325). Population ageing (hereafter referred to as Population over 65) is also expected to have a welfare-expanding influence (see Chapter 5.2). Furthermore, the variable Female Employment is applied as a proxy in order to assess the impact of ongoing changes in gender roles and household structures on social policy. As mentioned in Chapter 5.3, the effects of an increase in female employment are ambivalent: Although the rise in female employment has amplified the base of taxcontributing citizens thereby softening the fiscal burden on the welfare state (EspingAndersen, 2002b, p. 20; Pierson, 2001a, p. 95), a number of new social risks in conjunction with this development have emerged that need to be addressed by contemporary welfare states (Esping-Andersen, 2002b, p. 20; Pierson, 2001a, pp. 95-97). In the following the author turns to the political factors. In order to analyze if political partisanship remains a significant determinant of welfare state development, the following variables are incorporated in the model from Huber et al.’s (2004a) Comparative Welfare States Data Set: left seats as a percentage of seats held by all government parties (hereafter referred to as LEFTCAB), cumulative LEFTCAB score from 1946 to the year of observation (hereafter referred to as LTCABCUM), right seats (including right secular parties, right Christian parties, and right Catholic parties) as a percentage of seats held by all government parties (RIGHTCAB), and the cumulative RIGHTCAB score from 1946 to the year of observation (hereafter referred to as RIGHTCABCUM) (Huber et al., 2004b, pp. 19-21). Due to potential problems of multicollinearity, only one political variable at a time is included in the model. As mentioned above, annual and cumulative data are included since researchers are still debating which measure is more appropriate. According to Hicks (1999), “cumulative measures are used to operationalize the idea that because policy outputs accumulate over time from past policies they are functions of cumulative histories of political variables” (p. 189). However, as Hicks (1999) further elaborates, this notion is problematic since “static measures of legacies of partisan rule from origin (1946) until the first time periods in a data set” are combined with “more dynamic, […] temporally variable measures of partisan rule” (p. 189). Overall, following the hypothesis of the power resources approach, the author expects right government parties to have a welfare-regressing and left as well as Christian Democratic government parties to have a rather welfare-expanding (or -stabilizing) influence (see Chap-

120

8 A Quantitative Study on the Determinants of Welfare Spending

ter 5.5).128 These assumptions are challenged by the new politics approach, which envisages a radical transformation of the political landscape within industrialized economies (Pierson, 1996, p. 144). Precisely, Pierson (1996) contends that political parties and labor organizations have ceded power to pro-welfare state oriented interest groups (such as welfare state employees and recipients) within the social policy context (p. 147). The influence of special interest group activity on behalf of retired and unemployed citizens has been covered extensively in Chapter 5.5. Proxies for both groups have been incorporated with the variables Population over 65 and Unemployment. 8.2.2.2 Explanatory Variables as a Proxy for the Openness Literature As mentioned above, global economic forces exert influence via product and factor markets. Thus, the following three proxy variables are chosen to model globalization effects: Trade Openness, FDI Openness, and Net Migration.129 Following the Openness Literature, three different hypotheses can be distinguished: The competitiveness view envisages a negative association between global economic forces and the welfare state. The compensation view, by contrast, argues in favor of a positive relationship. Finally, the curvilinear view implies an inverted U-shaped relationship between globalization and public social spending (Hicks, 1999, pp. 204-207; see Chapter 6.3.1 in this book). 8.2.2.3 Explanatory Variables as a Proxy for the Context-Contingent View Following the explanation in Chapter 6.3.2, the context-contingent view mainly builds on the insights from the Varieties of Welfare Capitalism Literature. As mentioned in Chapter 3, two strands of research form the core of this literature according to Ebbinghaus & Manow (2001b, p. 1): Hall & Soskice’s (2001/ 2004a, 2001/ 2004b) VoC approach in the field of comparative capitalism literature and Esping-Andersen’s (1990/ 1998, 2000) Three Worlds of Welfare Capitalism approach in the area of 128

Critics of the econometric model set-up might hint at potential multicollinearity problems with regards to a possible transmission effect of global economic forces via political partisanship on public social spending. Without discounting this fact, the author tested for various measures of association, such as the Pearson and the Spearman correlation coefficient, and found very little correlation between the three globalization variables and the political partisanship variables. Furthermore, the author tested for multicollinearity by checking for the robustness of the signs of all the independent variables while omitting/ adding the respective political partisanship and globalization variables. The results proved to be robust with the ones reported in Chapter 8.4. 129 Critics of the econometric model set-up might hint at potential problems of endogeneity with respect to the relationship between Net Migration and the respective dependent variables since recent studies (Borjas, 1999, p. 609; Razin & Sadka, 2001, p. 79) have found empirical evidence that Net Migration is to a certain extent determined by the generosity of the migrant receiving country. To circumvent any potential endogeneity bias, the variable Net Migration was lagged by one year. Furthermore, the author checked for the robustness of the signs of all the independent variables while omitting/ adding Net Migration to the model. Results proved to be robust with the ones reported in Chapter 8.4.

8.2 The Econometric Analysis

121

comparative welfare state research. Attempting to bridge Esping-Andersen’s Three Worlds of Welfare Capitalism approach with Hall & Soskice’s VoC approach, Ebbinghaus & Manow (2001b) lay the focus on linkages between the welfare state regime and the following three institutional domains of the production regime: the system of industrial relations, the production system and employment regime, and the corporate governance and financial system (pp. 12-17). Comparing advanced capitalist welfare states and their respective political economies along the above stated institutional complementarities, the author has identified – largely in line with prior research (for example, Huber & Stephens, 2001a, pp. 85105; Huber & Stephens, 2001b; Kitschelt et al., 1999/ 2005, pp. 429-438; Swank, 2003, pp. 60-72) – the following three varieties of welfare capitalism: Liberal (Anglo-Saxon countries), conservative (Continental European countries and Japan), and social democratic (Nordic countries) welfare economies (see Table 2 in Chapter 2.3.2 and Table 6 in Chapter 3.3). Given the fact that institutional configurations vary significantly between each variety of welfare capitalism, reform paths will do so too. As a result, advocates of the context-contingent view assume that global economic forces do not have a uniform impact on all advanced capitalist democracies (Kitschelt et al., 1999/ 2005, pp. 449-457; Hall & Soskice, 2001/ 2004b, pp. 57-58) and their respective welfare states (Bonoli & Palier, 2001, p. 58; Ellison, 2006, p. 178; Leibfried & Obinger, 2001, p. 5; Scharpf & Schmidt, 2000, p. 335; Swank, 2003, pp. 60-61; Swank, 2002/ 2004, p. 5).130 Against this backdrop, the author – adhering to Rueda & Pontusson’s (2000) methodological approach – builds a series of interaction models that simultaneously control for the liberal, the conservative, and the social-democratic welfare state cluster. 8.2.2.4 Control Variables Besides the explanatory variables, two macroeconomic control variables – GDP Growth and Inflation – are incorporated into the model. GDP Growth, the growth rate of real GDP, controls for business cycle fluctuations. The author hypothesizes that welfare expenditures should increase during phases of recession and decrease during boom phases (Swank, 2002/ 2004, p. 81). Inflation, defined as the annual percentage change of the consumer price index, can be a major determinant of welfare spending since “high rates of inflation generate demands for increments to social benefits, activate cost-of-living indexation, and, when and where non-indexed progressive taxation forms an important source of welfare state funding, engender rises in the revenue base of the welfare state through bracket creep (e.g., Hicks and Swank [1992]; Pampel and Williamson [1989])” (Swank, 2002/ 2004, p. 81). To summarize, Table 21 (see p. 122) presents all dependent, explanatory and control variables (including their definitions) comprising the econometric model while 130

See Chapter 6.3.2 for a detailed overview of the principal hypotheses prevalent within the context-contingent view.

122

8 A Quantitative Study on the Determinants of Welfare Spending

Table 21: Definitions of Dependent, Explanatory and Control Variables DID%>$F .$$AC$A(KDID%>$F D7_o x.0 oh_x]6ego?6lxhlx 0 FD .0 oh_x.6 o?,xhel6lx hlx 0xxFD %6e6o?x6e60lo?,x4eg6]

D6elo0ex6e60lo?,x4eg6]

DAD("KDID%>$F D0_h?o0ex0[6xAx Je6-_0,-6e?xch?6 -_0,-6e?xoexUSo _?6xR 4egl?, Ueeh_x!6?xth7oe6? t-_h?o[6 !6?xth7oe6? Ueeh_xcoS#?xth7oe6? t-_h?o[6 coS#?xth7oe6? G6-h_6x-_0,-6e? GF4x6ee6ll hg6x6ee6ll B6?xnoSh?o0e J"A(">KDID%>$F FDx0)?# 4e_h?o0e

.$IAI(I"A Ueeh_x?0?h_ 7_o xl0 oh_x6]6ego?6lxhl h 6 6e?hS6x0 FD Ueeh_xl0 oh_xl6 o?,x?hel6lxhlxh 6 6e?hS6x0 FDx Ueeh_xoeg6]x 0elol?oeSx0 ?#66x-hC0xl0 oh_x oelhe 6x0Sh-lxT6elo0exe6-_0,-6e? hegxlo e6llxh,-6e?l$x)#o #xh6x6[h_h?6gx h 0goeSx?0x?#6x0__0)oeSx?#66x o?6ohIxe6?x 6_h 6-6e?xh?6lxLh_o,oeSx 0ego?o0elxhegx 0[6hS6xh?6l Ueeh_xoeg6]x 0elol?oeSx0 6elo0exoelhe 6x 0Sh-lx)#o #xh6x6[h_h?6g h 0goeSx?0x ?#6x0__0)oeSx?#66x o?6ohIxe6?x6_h 6-6e?x h?6lxLh_o,oeSx 0ego?o0elxhegx 0[6hS6x h?6l Ueeh_xl#h6 0x0_h?o0exhS6gxh70[6xAx ,6hlxoe ?#6x?0?h_x0_h?o0e Ueeh_x6 6e?hS6x0 ?#6x?0?h_x_h70x0 6x e6-_0,6g Ueeh_x6-_0,-6e? oe hSo _?6xhegx oegl?, hlxh l#h6x0 ?#6x?0?h_x_h70x0 6 Ueeh_x_6?xl6h?lxhlxh 6 6e?hS6x0 l6h?lx#6_gx 7, h__xS0[6e-6e?xh?o6l t-_h?o[6xheeh_x_6?xl6h?lxhl 6 6e?hS6x Cumulative annual left seats as ahpercentage of by all government parties from 0 seats held #6_gx7,xh__xS0[6e-6e?xh?o6lx01946 
m onwards 0e)hgl Ueeh_xoS#?xl6h?lxhl h 6 6e?hS6x0 l6h?lx #6_gx7, h__xS0[6e-6e?xh?o6l t-_h?o[6xheeh_xoS#?xl6h?lxhl 6 6e?hS6x Cumulative annual right seats as a hpercentage 0 seats held #6_gx7,xh__xS0[6e-6e?xh?o6lx0of by all government parties from 1946 
m onwards 0e)hgl 0?h_x6-h_6x o[o_ohex6-_0,-6e?xhl hx 6 6e?hS6x0 ?0?h_x6-h_6x0_h?o0e Ueeh_xl-x0 go6 ?xoe[6l?-6e?xoe_0)lxheg 0?_0)lxhl hx6 6e?hS6x0 FD Ueeh_xl-x0 o-0?lxhegx6]0?lxhlxh 6 6e?hS6x0 FD B6?x-oSh?o0exhl h 6 6e?hS6x0x?0?h_x 0_h?o0e Ueeh_xS0)?#xh?6x0 6h_xFD Ueeh_ 6 6e?hS6x #heS6x0 ?#6x 0el-6x o 6xoeg6]

+***

-***

+***

New politics approach

-/+*

-/+*

-/+*

Liberal

-/+*

-/+*

-/+*

Conservative

-/+*

-/+*

-/+*

Social democratic

Context-Contingent View

Note: Independent variables characterized with *** are considered to be the most significant determinants of welfare state spending according to each respective theory. Source: Own depiction.

-***

Net migration2

+***

Trade openness2

+***

+*** -***

-***

Net migration

+***

FDI openness2

-***

-***

-***

Cum. right cabinet

Trade openness

-***

Annual right cabinet

FDI openness

+***

+***/'-***

+***

Power resources approach

+***

-***

Deindus- Household trializatranstion formation

Post-Industrial Challenges

Cum. left cabinet

+***

Competi- Compen- Convertiveness sation gence view view view

Openness Literature

Annual left cabinet

Female employment

Employment in A & I

Unemployment

Population over 65

Independent Variables

Expected associations with welfare effort, taking principal theoretical approaches into account

Table 22: Overview of the Different Hypotheses on the Determinants of Public Social Spending in OECD Countries

8.2 The Econometric Analysis

123

124

8 A Quantitative Study on the Determinants of Welfare Spending

Table 22 (see p. 123) gives an overview of all the above stated hypotheses on what determines public social spending. The data sources and descriptive statistics of all variables are depicted in Appendix 2.

8.3

Methodology

Similar to Brady et al. (2005), the study uses time-series cross-section (TSCS) analysis of 17 OECD countries, namely: Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, Netherlands, Norway, Sweden, Switzerland, the UK, and the US, from 1975 to 2000, yielding a sample size of 442 “country-years”.131 1975 is set as the starting year of this study since this date – shortly after the outbreak of the first oil crisis and the implosion of the Bretton Woods system – marks the approximate beginning of the era of intense globalization (Brady et al., 2005, p. 926). In line with previous studies in the field of comparative welfare state research (for example, Allan & Scruggs, 2004; Brady et al., 2005, Hicks, 1999; Huber & Stephens, 2001a; Swank, 2002/ 2004) the methodology used in this paper is an ordinary least squares (OLS) regression with a panel corrected standard error (PCSE). Beck & Katz (1995a) performed a Monte Carlo study on the accuracy of estimating TSCS data using panel corrected standard errors (PCSE) in the presence of heteroscedasticity and contemporaneous correlation of the error terms in comparison to standard OLS (pp. 634-635). According to them, “ordinary least squares is optimal […] for TSCS models if the errors are assumed to be generated in an uncomplicated (‘spherical’) manner. In particular, for OLS to be optimal it is necessary to assume that all the error processes have the same variance (homoscedasticity) and that all of the error processes are independent of each other” (Beck & Katz, 1995a, p. 636). Moreover, to deal with serial correlation of errors, Beck & Katz (1995a) have suggested either using a lagged dependent variable or “transform[ing the data] based on an estimate of the common serial correlation” (p. 645; see Beck & Katz, 1995b, pp. 7-8). Due to the temporal and spatial nature of the data presented in this study, standard OLS would produce biased estimates since errors tend to be autocorrelated and correlated across units. Following the approach suggested by Beck & Katz (1995a) and other researchers in comparative political economy, an OLS-PCSE technique corrected for contemporaneous correlation, heteroscedasticity, and first-order autocorrelation (AR(1)) of errors is employed. Whether to include or exclude a lagged dependent variable depends 131

The author has a sample size of 442 “country-years” for the dependent variables SSTRAN, Benefit Generosity Index and Pension Generosity Index. Since data for PSEX was only available starting in 1980, the sample size for this dependent variable is 357 “country-years”.

8.3 Methodology

125

on the nature of the data. Plümper, Troeger, and Manow (2005) explain that including a lagged dependent variable for this type of study would be inappropriate and could pose additional problems, resulting in biased estimates. They suggest the use of the AR(1) correction approach to deal with serial correlation, consistent with the methodology mentioned above (pp. 334-343). Furthermore, the inclusion of a lagged dependent variable in fixed effects models when panels are short would also lead to biased coefficients (Nickell, 1981, pp. 1418-1423). This implies the use of instrumental variable estimation. However, improper instrumentation would only lead to spurious results. Therefore, to circumvent all problems associated with including a lagged dependent variable, the author has decided to follow the approach suggested by Plümper et al. (2005, pp. 334-343) and similarly applied by Brady et al. (2005, pp. 926-927). To reduce the problem of simultaneity bias, the author uses the lagged values of all independent variables as suggested by Allan & Scruggs (2004, p. 505) and also performed by Brady et al. (2005, p. 927). Plümper et al. (2005) have also stressed the importance of identifying the optimal lag length of each variable (pp. 343-345). However, formal testing for this does not exist in panel analysis. In this case, to reduce simultaneity bias and to address the complex issue of optimal lag, the author uses a uniform lag length of one year for simplicity without losing sight of the fact that political processes and adjustments may vary in different countries. To control for the omitted variable bias due to country-specific differences, country dummies are always included in each model.132 Furthermore, to address the problem of missing data, the author interpolates whenever there are missing values for less than three years found in the series.133 Previous studies in comparative political economy rarely take into account possible nonstationarity issues in TSCS; literature and formal tests in this area of research are still at a nascent stage. Nonstationarity, if left untreated, can pose serious problems when interpreting the findings in pooled models (Kittel & Winner, 2005, p. 288). Due to recent developments in econometrics, this study tested for stationarity using the Levin-Lin-Chu panel unit root test. This test stems from a paper by Levin et al. (2002, pp. 2-3) which is specifically designed for application in cross-section and 132

In this context, some scholars might argue that one would also have to incorporate time fixed effects to control for the gradual process of decline occurring during two specific periods (aftermath of second oil price crisis in the early 1980s, retrenchment process starting in the early 1990s). Brady et al. (2005) have tested for both periods and did not find any significant difference in comparison to the overall model results (p. 926). However, it has to be noted that given the Nordic countries’ strong upswing and subsequent rapid decline in public social spending, the author tested for a structural break with regards to the Nordic welfare regime cluster. The output tables of the structural break model are presented in Chapter 8.4.3. 133 Exceptions to this rule were only made in the following cases: Austria (variable PSEX, 1981-1984, 1986-1989), Norway (variable PSEX, 1981-1984), and the US (variable SSTRAN, 1998-2000).

126

8 A Quantitative Study on the Determinants of Welfare Spending

time-series data. The test can be performed using an available special coded program in STATA 9 developed by Bornhorst & Baum (2001). According to these authors, “the test may be viewed as a pooled Dickey-Fuller test, or an Augmented DickeyFuller (ADF) test when lags are included, with the null hypothesis that of nonstationarity (I(1) behavior)” (Bornhorst & Baum, 2001, n. p.). Performing the Levin-LinChu panel unit root test under the null hypothesis of a unit root or nonstationarity, the result shows that all variables are stationary except for the variable Trade Openness, which reveals weak stationarity. Aside from formally testing for stationarity, the author individually inspects the graph of each time series. To double-check the consistency of the PCSE findings under stationarity, another set of regression is performed in which each variable is assumed nonstationary and the first-differencing of each is performed. The regression results are essentially the same with a few minor exceptions. Overall, against the backdrop of the principal hypotheses depicted in Table 22, the author estimates the following four pooled OLS-PCSE models: a linear model, a linear model with curvilinear effects, an interaction model, and an interaction model with curvilinear effects. Moreover, in order to analyze in detail the Nordic welfare regime cluster’s strong upswing and subsequent rapid decline in public social spending, a structural break OLS-PCSE model is applied for the Nordic welfare states.134 To verify the robustness of the results, all equations are also estimated using fixed effect models (see Appendix 3). The fixed effects model takes into account the panel characteristic of the data and has the power to control for unobserved heterogeneity (Greene, 2003, pp. 284-285). Naturally, all non-time-varying variables will be dropped when using fixed effects. Overall, the regression results for the fixed effect model have, with minor exceptions, confirmed the empirical results presented in Chapter 8.4. For the linear baseline model the following equation is estimated:

Z5 


 T

T4

T Z T 5 $




 ( 

(   ; T (   $ Z T 5   $



Z5

(1)

Yit refers to the dependent variable, X denotes the a independent variables, D’s are m country dummies, the subscript i refers to cross section units where i = 1, 2, … N, and the subscript t to time units where t = 1, 2, … T. _ is the common intercept. `a refers to the slopes of the explanatory variables and ^m–1denotes the intercept of each coun134

Brady et al.’s (2005) model serves as a basis for the construction of the first two models: the linear model and the linear model with curvilinear effects. The methodology of the interaction model is largely based on Rueda & Pontusson’s (2000) model, which researches the determinants of wage inequality in three country clusters (pp. 370-371). Finally, the structural break model equation is to a great extent borrowed from Allan & Scruggs (2004), who have tested whether a structural break occurred within advanced capitalist welfare states in the early 1980s (pp. 504-505).

127

8.3 Methodology

try dummy, whereas ¡it is a random error term normally distributed around a mean of 0 with a variance of m2. Notice that all X are lagged one year to take into account their delayed adjustment before having an effect on Y. The set-up of the baseline model with curvilinear effects is the same with the exception that the three X representing the globalization variables are complemented by the squared term X2 in order to model possible curvilinear effects. Again, all independent variables are lagged by one year and country dummies are incorporated to control for country-specific differences. The a and b subscripts denote the number of independent variables. The equation takes the following form:

Z5 




T

4 T  Z T 5 $


T

 4 .

P .  Z T 5 $




.



( 

;T ( $ Z T 5 $
 Z5

(2)

( 

The set-up of the first interaction model is very similar to equation (1) except that three additional dummy variables – LWC for the liberal, CWC for the conservative, and SWC for the social-democratic welfare economy cluster (see Table 2 in Chapter 2.3.2 and Table 6 in Chapter 3.3 for a classification of advanced capitalist welfare states and advanced market economies) – are introduced, which all interact with the control variables X. Again, all independent variables are lagged by one year and country dummies are incorporated to control for country-specific differences. Similarly, the subscripts a and b represent the number of independent variables. The equation takes the following form:

Z5 








T

T

TT4

T4

T Z T 5  $

T Z T 5   $

 V[

 )V[

$


T

$








T

;

T (   $ Z T 5   $

T



( 

T4

T Z T 5   $

 [V[




$
(3)

Z5

( 

The set-up of the second interaction model is very similar to equation (2) except that the three additional dummy variables (LWC, CWC, and SWC) interact not only with the linear control variables X, but also with the curvilinear control variables X2. Again, all independent variables are lagged by one year and country dummies are incorporated to control for country-specific differences. And the subscripts in each independent variable denote its total number. The equation takes the following form:

Z5 




TT4



TT4





T

T

( 

 T



T

T Z T 5  $

T4

T Z T 5   $

 [V[

 V[ $


$




.T4

P . Z T 5   $

 V[ $


.



.T4



.T4

P . Z T 5   $

 [V[

$


.

T Z T 5  $

 )V[ $


(   ; T (  $ Z T 5  $

.




Z5

P . Z T 5  $

 )V[ $


(4)

128

8 A Quantitative Study on the Determinants of Welfare Spending

The structural break model for the Nordic countries is very similar to regression equation (1) except that one dummy variable, SB (for structural break), is incorporated and interacts with all linear control variables X. Again, all independent variables are lagged by one year and country dummies for the Nordic countries (u) are incorporated to control for country-specific differences. Like above, the subscript a denotes the number of independent variables. The equation takes the following form:

Z5 


8.4

 T

T

4

T Z T 5  $






T

T )!4

T  Z T 5  $






 

   ; T    $ Z T 5   $


Z5

(5)

Empirical Results

Before presenting the results of the regression models for all four dependent variables, five features have to be noted: Firstly, the depicted models do not include the variable Employment in A&I. The proxy variable for Iversen & Cusack’s (2000) deindustrialization hypothesis has been dropped since it appears to bias the coefficient of the variable Unemployment (multicollinearity).135 Given the fact that a primary cause for unemployment has been the phenomenon of deindustrialization, the results associated with the variable Unemployment will be interpreted in such a way. Secondly, the proxy variable Female Employment has been dropped due to endogeneity problems – the variable both drives the welfare state and is driven by it (see Chapter 5.3 for a detailed explanation of the relationship).136 Thirdly, the control variable Inflation was dropped due to evidence for multicollinearity with other proxy variables. Fourthly, all proxy variables for globalization are simultaneously controlled for in the model to grasp the full effect of global economic forces on welfare effort. Yet the author has also run models controlling only for one globalization variable at a time; the results proved to be robust with the ones reported in this chapter. Finally, Table 23 to 25 only report the results of the baseline model without curvilinear effects since curvilinear effects were not found for any of the three models with the dependent variables PSEX, SSTRAN, and the Pension Generosity Index. Table 26, by contrast, presents the baseline model with curvilinear effects for the fourth dependent variable, the 135

The author has also run models substituting the variable Unemployment with Employment in A&I. The explanatory variable for Iversen & Cusack’s (2000) deindustrialization hypothesis has produced a significant negative coefficient for the majority of the models. 136 Further research is needed how to quantify the effects of changing gender roles and household structures in a proper way. This could involve the use of refined independent variable measures (for example, the amount of single-parent households as a percentage of the total number of households), which might help to elucidate the effect of household transformation on public welfare spending.

8.4 Empirical Results

129

Benefit Generosity Index. The same procedure will be applied with respect to the Output Table 27 to 30 of the interaction models.137

8.4.1

Baseline Model Results

Tables 23 and 24 (see p. 130 and p. 131) present the baseline model results for PSEX and SSTRAN. While all proxy variables for post-industrial challenges (Population over 65, Unemployment, Political Partisanship Variables) have a statistically significant influence on welfare effort, this is true for only one of the three globalization proxy variables (Trade Openness). As expected, rising unemployment and population ageing have a significantly positive influence on welfare effort, whereas high economic growth rates are significantly negatively associated with government spending. The cumulative political partisanship measures indicate that both left and right parties have been participating in expanding the welfare state during the second half of the 20th century. In contrast, the two annual political partisanship measures show a mixed and insignificant association with welfare spending (compare Brady et al., 2005, p. 933).138 The discrepancy between the cumulative and annual variable coefficients reflects the debate over the appropriate political partisanship measure in an empirical analysis (see Chapter 8.2.2.1). While all political and economic variables of the domestic realm report a clearly significant impact on welfare spending, global economic forces seem to have only a minor effect on PSEX and SSTRAN. Trade Openness has a welfare regressing influence, which strengthens the hypothesis of the competitiveness view (see Chapter 6.3.1.1). The above results are broadly in line with Brady et al.’s (2005) earlier findings, with the exception of the proxy variables Cumulative Right Cabinet, Net Migration, and FDI Openness.139 With respect to Cumulative Right Cabinet, Brady et al. (2005) report a significantly negative coefficient in both welfare state models (pp. 938-939). Turning to the two globalization variables, Brady et al. (2005) detect a positive significant influence of Net Migration on the dependent variable PSEX and a curvilinear significant association (a U-shaped relationship) of FDI Openness with the dependent variable SSTRAN (pp. 938-939). 137

Given the fact that Brady et al.’s (2005) model has served as the basis for the construction of the baseline models, the results of the latter will be primarily compared with Brady et al.’s (2005) earlier findings. For a comparison with other studies’ findings, the reader is urged to review the literature presented in Chapter 5 and 6.3. 138 These findings could serve as proof that Pierson (1996) is right when arguing that “the emergence of powerful groups surrounding social programs may make the welfare state less dependent on the political parties […] that expanded social programs in the first place” (p. 147). 139 It is worthwhile noting that Brady et al. (2005) do not find a significant relationship between Trade Openness and PSEX or SSTRAN; however, their proxy variable Net Trade reports a significantly negative coefficient in the model with PSEX as dependent variable (p. 938).

Dh?olhel#ox-6hl6gxhl Ueeh_x!6?xth7oe6?x t06o o6e? ?dl?h?ol?o K/\sss A/A K/P sss P/\A dK/Psss dP/
P K/K K/Km dK/KK dK/ dK/K s d/\ dK/
dK/A d/P+ d/P \/\\sss A/\
/
Ksss /A P/Ass P/KP +/m\sss  /\K /A
sss m/P \/\sss \/AA /Asss
/AA P/+
s /
P/ss P/  d /+ sss d / \ +/
msss \/K m/K\sss /KA /

sss  /\+ K/K K/m dP/+ss dP/m+ +/Kmsss /P mK K/+\ P

/ K/KK K/\K

Dh?olhel#ox-6hl6gxhl t-/x!6?xth7oe6? t06o o6e? ?dl?h?ol?o K/m+sss A/P K/
ss P/AA dK/ sss d /P K/P+sss A/P
dK/Km d/ \ dK/Kmss dP/
dK/P dK/ d /KPsss d / /mAsss /A
K/P sss \/P\ m/Asss m/\ m/mmsss m/\ m/\sss /+\ \/sss
/K / sss K/K m/
Ksss /K\ /+sss m/P dP/KPs d/\\
/Asss +/\ d/\+ d/ \ A/+\sss /
K dK/

d/m+ dK/\m dK/K
/ sss m/A+ mK K/
K mA
m/m K/KK K/K

Dh?olhel#ox-6hl6gxhl Ueeh_xcoS#?xth7oe6? t06o o6e? ?dl?h?ol?o A/K K/\sss P/+P K/P sss dP/
P dK/Psss d/m
dK/ K dK/ dK/K d/+K dK/Kms dK/m dK/ d/ d/ m A/
\/\msss /m

/+Asss /
 P/K\s  /\ +/mmsss m/m
/m
sss \/mK \/K
sss
/A /A\sss / P/+ P/P /
ss d /P d /\sss /
+ +/
sss /KP m/Ksss  /m /
Psss K/Pm K/+ dP/AP dP/
Pss /+ +/PAsss mK K/+\ P +/+P K/KK K/\K

Dh?olhel#ox-6hl6gxhl t-/xcoS#?xth7oe6? t06o o6e? ?dl?h?ol?o K/ sss /A K/PAsss / K dK/msss d / \ K/
sss m/A dK/K dK/ dK/Kmss dP/m\ dK/\ dK/m dm/\sss d /

+/
sss PK/\K  /+sss
/m P/+sss P/
P /P\sss m/\
/\Psss \/Pm /KPsss /++ / Psss K/PP m/K sss P/m A/+Ksss A/ d
/PAsss d/ K/++sss
/m \/\sss +/m /\\sss m/ d/K d/  dm/\Ksss dm/mP
/\\sss m/\+ mK K/
K \
/  K/KK K/A


Note: Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively.

D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e FUl?h_oh FUl?oh F%6_SoFthehgh FF6e-h FGoe_heg FGhe 6 F6-he, F46_heg F4?h_, F hhe FB6?#6_hegl FB0)h, F.)6g6e FJQ FJ.U t0el?he? B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0x

4eg66eg6e?xhoh7_6l

Table 23: Baseline Model Results – Dependent Variable: Public Social Expenditures (PSEX), 1980-2000

130 8 A Quantitative Study on the Determinants of Welfare Spending

Dh?olhel#ox-6hl6gxhl Ueeh_x!6?xth7oe6?x t06o o6e? ?dl?h?ol?o K/P ss /

K/ Ksss A/ dK/K+sss d /m dK/ dK/A dK/KP dK/
A dK/K s d/\
K/
K/+ d/ sss dA/mA A/K sss /\P m/ Asss /mm dP/+ss dP/mA /Kmsss /PK /mK /K
P/P+ss P/ /K
sss /
 dK/P dK/+ K/ K/\ d /+sss dm/mA K/ msss A/m
/P
/ P A/\sss A/P d/AAs d/\ d /K\ss dP/m+ K/\Asss A/PA mPA K/\PP
m\P/P+ K/KK K/\P

Dh?olhel#ox-6hl6gxhl t-/x!6?xth7oe6? t06o o6e? ?dl?h?ol?o K/P /P K/P+sss m/A dK/K
sss d /A K/ss P/mK dK/Km d/A dK/K ss d/
K/ K/ + d/
Asss d/m\ /Asss /
m/mAsss /A\ d/
d/A /A
/AK K/P K/AK P/APsss P/+ /K
sss m/KP K/AK K/  / P / + d /PKsss d /AA K/Amsss A/+ dK/++ dK/
P/
Kss P/K
d/
Pss dP/+ dP/PPs d/\P /msss A/
 mPA K/\ + m / K/KK K/


Dh?olhel#ox-6hl6gxhl Ueeh_xcoS#?xth7oe6? t06o o6e? ?dl?h?ol?o /
 K/P ss m/

K/ Ksss d /A dK/K
sss dK/ dK/K dK/
+ dK/KP d/+K dK/K s K/\P K/PK dA/
d/ msss /A m/
sss /mP m/ Asss dP/ + dP/\+ss /P P/

sss /Km /  P/P+ P/Pss /+m /K+sss dK/A dK/
K/A K/P dm/
d /+Psss A/mK K/ Psss /PA /P A/K\ A/Ksss d/
d/As d /K ss dP/m K/\ sss A/ mPA K/\PK m
+P/
P K/KK K/\P

Dh?olhel#ox-6hl6gxhl t-/xcoS#?xth7oe6?x t06o o6e? ?dl?h?ol?o K/K\ K/A K/ Ksss A/ dK/K
sss d /AP K/K+s /+ dK/K d/ dK/K s d/
P K/+ K/m d\/sss dA/m A/mAsss \/ A/+Psss /
A dP/Amss dP/PA m/P sss /\ P/A\s /\
/
mss /
\ /msss m/ K/PA K/PK P/Ps /\ dA/++sss dm/\ /K+sss A/
P P/\Kss P/\ \/K
sss m/+A d/\
ss dP/KA d /A
sss dP/+
/msss A/A
mPA K/\P\\ A
K/
K/KK K/\

Note: Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively.

D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e FUl?h_oh FUl?oh F%6_SoFthehgh FF6e-h FGoe_heg FGhe 6 F6-he, F46_heg F4?h_, F hhe FB6?#6_hegl FB0)h, F.)6g6e FJQ FJ.U 0el?he? B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

4eg66eg6e?xhoh7_6l

Table 24: Baseline Model Results – Dependent Variable: Social Security Transfer (SSTRAN), 1975-2000

8.4 Empirical Results

131

Dh?olhel#ox-6hl6gxhl Ueeh_x!6?xth7oe6?x t06o o6e? ?dl?h?ol?o K/\Ksss A/m
K/K K/
K dK/K dK/ dK/P dK/m
dK/s d/\\ dK/Psss dP/\m K/\ /PA K/K K/mP K/KKsss /+m dK/PA dK/m d/\
sss dm/A dK/ + dK/ K /  K/+
d/\ d/ +/ +sss \/+ m/msss /  / /P dK/PA dK/P dm/ Pss dP/\ d/\sss dm/\ d
/\Psss d/P +/Ksss /Km K/ sss \/P+ P/ Ksss P/PA d+/sss dA/
d
/\Psss dA/P P/\msss \/\K mPA K/+
PP P KA+/K K/KK K/


Dh?olhel#ox-6hl6gxhl t-/x!6?xth7oe6?x t06o o6e? ?dl?h?ol?o K/sss /K
K/KA K/+ dK/K dK/\ K/KA K/+m dK/ ss dP/K dK/Psss dP/+ K/ /A K/K K/AA K/KKsss /
 dK/P dK/A
d\/ sss dm/+m d/A dK/\\ / P K/
 d/P dK/\ \/sss A/m
m/KAsss /K /\K /  dK/ K dK/P d /
As d/
P d/
sss d /
 d
/APsss d/KK +/\Ksss /
/A+sss A/ /\sss / + d+/+Asss dA/\
d
/msss dm/
PP/\sss +/A mPA K/+
mP PA\PK/mK K/KK K/+

Dh?olhel#ox-6hl6gxhl Ueeh_xcoS#?xth7oe6? t06o o6e? ?dl?h?ol?o A/mm K/\Ksss K/
 K/K\ dK/+ dK/K dK/  dK/K\ d/+m dK/s dP/\A dK/Psss /P
K/
K/m+ K/KK /+m K/KKsss dK/m dK/PA dm/A d/++sss dK/  dK/m K/+\ /P+ d/ d/\ \/ +/ sss /P
m/ msss / /Am dK/Pm dK/P
dP/\ dm/ mss dm/PK d/\msss d/PP d
/\Ksss /KP +/A
sss \/PP K/Asss P/\ P/PPsss dA/\K d+/+sss dA/PP d
/\msss \/\K P/+Ksss mPA K/+

P 
\/A K/KK K/


Dh?olhel#ox-6hl6gxhl t-/xcoS#?xth7oe6?x t06o o6e? ?dl?h?ol?o K/\
sss A/P K/K\ K/

dK/K dK/
dK/KA dK/
 dK/s d/\K dK/Psss dP/
K/
/ K K/KK K/ \ K/KKsss /+ dK/Pm dK/P d/Psss d / dK/\ dK/AA K/m K/PP d/+
d/PA \/A
sss A/ K / ss P/  /\ /P
dK/  dK/P dm/Ass dP/Pm d\/Asss dm/K
d+/msss dm/KP +/msss A/PP
/\sss A/A\ /K+sss \/K d+/AAsss dA/m+ d
/
sss dm/
K P/ Psss \/  mPA K/+
 P m
/++ K/KK K/


Note: Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively.

D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e GF4x06ee6llxlLh6g hg6x06ee6llxlLh6g B6?x-oSh?o0exlLh6g FUl?h_oh FUl?oh F%6_SoFthehgh FF6e-h FGoe_heg FGhe 6 F6-he, F46_heg F4?h_, F hhe FB6?#6_hegl FB0)h, F.)6g6e FJQ FJ.U 0el?he? B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0x

4eg66eg6e?xhoh7_6l

Table 25: Baseline Model Results – Dependent Variable: Benefit Generosity Index, 1975-2000

132 8 A Quantitative Study on the Determinants of Welfare Spending

8.4 Empirical Results

133

Table 25 presents the baseline model results for the dependent variable Benefit Generosity Index. In contrast to expenditure data, this index does not look at the aggregate ratio of social spending relative to GDP, but rather takes net replacement rates, qualifying conditions, and coverage rates of the three major social insurance programs (pension, unemployment, and sickness payments) into account (Scruggs, 2004b, pp. 1-5). Again, the proxy variable Population over 65 reports a strong positive influence on the welfare state measure. All other explanatory variables of the domestic realm appear not to be significant; however, they report vastly the same signs as in the previous two output tables. With respect to globalization variables, Trade Openness seems to have a significant impact on the welfare state.140 Unlike in the PSEX and SSTRAN models, the causal relationship between the degree of Trade Openness and the Benefit Generosity Index is not linear and negative, but rather curvilinear. This curvilinear effect, however, stands in sharp contrast to the curvilinear view presented in Chapter 6.3.1.3, since an inverted U-shaped relationship between globalization and the welfare state cannot be shown to exist. On the contrary, Table 25 indicates sharp welfare retrenchment at the beginning of the globalization era and signs of stabilization and expansion of the Benefit Generosity Index at higher degrees of trade integration. As mentioned above, a similar effect has been reported by Brady et al. (2005) for FDI Openness on the dependent variable SSTRAN (p. 939). The results of Table 25 differ from Brady et al.’s (2005) earlier findings (model with de-commodification index as dependent variable) in two ways: Firstly, Brady et al. (2005) find that political partisanship has a significant influence on the level of de-commodification. Precisely, cumulative left and Christian-Democratic government parties are reported to enhance whereas cumulative right government parties are predicted to reduce de-commodification (p. 937). Secondly, Brady et al. (2005) detect a significant positive association between the globalization variable Trade Openness and de-commodification (p. 937). Given the fact that the author uses a more updated version of the de-commodification index, the Benefit Generosity Index, the differences might be fully attributable to the modification of the dependent variable. Finally, Table 26 (see p. 134) presents the baseline model results for the dependent variable Pension Generosity Index. As already mentioned in Chapter 8.1, the Pension Generosity Index is a sub-component of the Benefit Generosity Index. It takes net replacement rates, qualifying conditions and coverage rates of the national pension insurance system into account (Scruggs, 2004b, pp. 1-5). Similar to the results indicated in Table 25, the proxy variable Population over 65 reports a significant welfare-enhancing influence. The same relationship is also found for the variable Unem140

The proxy variable FDI Openness is not mentioned since it does not appear to impose a significant curvilinear effect on social spending.

Dh?olhel#ox-6hl6gxhl Ueeh_x!6?xth7oe6?x t06o o6e? ?dl?h?ol?o K/m sss m/K+K K/K
ss P/ 
dK/Kmmss dP/AKP dK/+ d/PA dK/K +ss dP/P
 K/KA /mP\ dK/PK\ d/K m /AA+sss m/+m m/\Pmsss
/
AP P/

sss / mA /A sss +/\m / msss /mm \/Asss
/  /m sss
/P+P K/ P K/ P/K\ss P/ A/K+sss A/\P
/KPPsss /\A
/P
Psss \/ /PmAsss /\ \/+ sss /K A K/K+\ K/
A/m\sss /+\ K/K

K/KA+ mPA K/\ + PK
K K/\P\mP

Dh?olhel#ox-6hl6gxhl t-/x!6?xth7oe6? t06o o6e? ?dl?h?ol?o K/m\
sss m/+K K/Asss P/A
dK/KmAss dP/AA dK/KAA d/A\P dK/K s d/
m\ K/K /A dK/ \ dK/+ /\Psss m/
\\ A/ \sss \/A\A P/+ sss /+A /KPsss /
+P \/KKmsss
/APm \/mmsss
/  /PPsss +/\ K K/Pm K/m
m /APP /+ A/K m m/Psss /+
P/\PPsss \/P + /\sss \/
K \/PA sss +/+K +/
Asss K/ P K/Pm A/\K
A/K sss dK/ dK/P\ mPA K/\  \PK K K/\

Dh?olhel#ox-6hl6gxhl Ueeh_xcoS#?xth7oe6? t06o o6e? ?dl?h?ol?o m/Km K/mP
sss P/  K/K
\ss dP/AAm dK/KmAss K/A  K/K
dP/ \ dK/Kmss /mA+ K/KA dK/

\ dK/PKP m/A\
/AKsss
/
mP m/mPsss / m P/
A+sss +/+mA /AA sss /m+
/ sss
/ K+ \/P\sss
/ K /mPsss K/K
K/P
+ P/
P/K\mss A/\\ A/msss /\\ /K \sss \/K
/P+Psss /
/
sss \/\
sss K/
K\ K/KA
K/K
/
 A/Asss K/KPm K/KmP mPA K/\ + PP K K/\mA

Dh?olhel#ox-6hl6gxhl t-/xcoS#?xth7oe6?x t06o o6e? ?dl?h?ol?o K/ \msss P/
 K/K
\ss P/ P dK/KmAss dP/AAP K/KP+ K/
KP dK/Kmss dP/mKP K/Km / P
dK/PK dK/
+A /Kmsss / KA m/+Ksss
/ Am /m+ sss / + / Psss +/+  /\ sss
/KA \/APAsss +/ K
/ Kmsss
/K K/ m K/ m P/PAPss P/PP+ A/
sss A/ P/ P
ss /
+ /Amsss \/  /
Asss +/ A+ +/mAsss \/+
dK/K dK/KPA A/ Psss /m  K/  K/
 mPA K/\  
K K K/\AKKP

Note: Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively.

D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e FUl?h_oh FUl?oh F%6_SoFthehgh FF6e-h FGoe_heg FGhe 6 F6-he, F46_heg F4?h_, F hhe FB6?#6_hegl FB0)h, F.)6g6e FJQ FJ.U t0el?he? B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0x

4eg66eg6e?xhoh7_6l

Table 26: Baseline Model Results – Dependent Variable: Pension Generosity Index, 1975-2000

134 8 A Quantitative Study on the Determinants of Welfare Spending

8.4 Empirical Results

135

ployment. The political partisanship indicators, by contrast, are all insignificant. In this sense, clear evidence is provided in favor of the new politics perspective: If the number of welfare recipients (particularly, the retired and the unemployed) increases, the political power of these interest groups also augments resulting eventually in a rise of per capita generosity of pension insurance programs (see Pampel & Williamson, 1992, pp. 165-166; Pierson, 1996, pp. 174-175; see Chapter 5.5 in this book for a more detailed explanation). Besides these findings, it is also worthwhile noting that the control variable GDP Growth reports an effect on welfare effort very similar to the one detected in Tables 23 and 24 (see Swank, 2002/ 2004, p. 81). Finally, with respect to the globalization proxy variables, FDI Openness appears to have a significant negative impact on the Pension Generosity Index. This result stands in sharp contrast to Table 25 (Benefit Generosity Index) which reports a curvilinear Ushaped relationship between Trade Openness and the welfare state. Overall, in line with Brady et al. (2005, p. 940, 944-945), the results of the baseline models support the view that advanced capitalist welfare states are primarily shaped by domestic political and economic factors. Among the post-industrial challenges (see Pierson, 2001a), the ongoing demographic shift appears to play the most prominent role. By contrast, globalization seems to have a minor influence on social spending activities. Furthermore, different dimensions of globalization (Trade Openness and FDI Openness) differ with respect to their association with the four welfare state measures. Scarce evidence is provided in favor of the competitiveness view.

8.4.2

Interaction Model Results

The context-contingent view assumes that global economic forces do not have a uniform impact on all advanced capitalist democracies (Kitschelt et al., 1999/ 2005, pp. 449-457) and their respective welfare states (Bonoli & Palier, 2001, p. 58; Ellison, 2006, p. 178; Leibfried & Obinger, 2001, p. 5; Scharpf & Schmidt, 2000, p. 335; Swank, 2003, pp. 60-61; Swank, 2002/ 2004, p. 5).141 Against this backdrop, the author – adhering to Rueda & Pontusson’s (2000, pp. 370-371) methodological approach – builds a series of interaction models that simultaneously control for the liberal (Anglo-Saxon countries), the conservative (Continental European countries and Japan), and the social-democratic (Nordic countries) welfare state cluster. Table 27 (see pp. 136-137) presents the results for the interaction model with PSEX as dependent variable. Confirming the general baseline model results stated above, most country-specific political and economic variables report significant coefficients for all three welfare state clusters. Population over 65, Unemployment, Cu141

See Chapter 6.3.2 for a detailed overview of the principal hypotheses prevalent within the context-contingent view.

136

8 A Quantitative Study on the Determinants of Welfare Spending

mulative Left Cabinet, and Cumulative Right Cabinet exert a significantly positive influence on welfare effort. Annual Right Cabinet in the conservative and socialdemocratic country group, by contrast, seems to roll back the welfare state as predicted by the power resources approach (see Chapter 5.5). Moreover, the macroeconomic control variable GDP Growth appears consistent with the findings retrieved in the baseline model (see Swank, 2002/ 2004, p. 81). With respect to the globalization Table 27: Interaction Model Results – Dependent Variable: PSEX, 1980-2000 A. Partisanship measured as Annual Left Cabinet 4eg66eg6e?xhoh7_6l D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

!o76h_ t06od ?d o6e? l?h?ol?o /\sss A/m+ K/PKs /\
dK/K d/P dK/K\ dK/A K/K K/ dK/KAs d/+P dK/A\ dK/\

t0el6[h?o[6 t06od ?d o6e? l?h?ol?o K/mKsss m/P K/msss m/+K dK/Pss dP/ \ K/mA /P dK/KP dK/
dK/KKP dK/K\ dK/ dK/mK mK K/
+ Am/PP K/KKKK K/AA

.0 oh_dF6-0 h?o t06od ?d o6e? l?h?ol?o K/K /K\ K/m sss /AK dK/+ss dP/ K/P\ K/A K/Ks /+ dK/PAsss dA/A / /A

Note: Model includes country dummies. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively. Country dummy estimates are not shown.

B. Partisanship measured as Cumulative Left Cabinet  4eg66eg6e?xhoh7_6l

D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

!o76h_ t06od ?d o6e? l?h?ol?o /Pmsss /Am K/+s /
dK/K\ d/mP K/PPsss /\ dK/K dK/A dK/KAs d/
K dK/Am dK/\

t0el6[h?o[6 t06od ?d o6e? l?h?ol?o K/ Psss m/PP K/ sss /P dK/ sss dP/\m K/PPsss /m dK/Ks d/+
dK/K dK/ + dK/ A dK/
A mK K/
P +
+
/
K/KKKK K/A

.0 oh_dF6-0 h?o t06od ?d o6e? l?h?ol?o K/Pm K/m+ K/
sss /  dK/\ss dP/ K/Pss P/A
K/K K/+m dK/PAsss dA/m K/+m K/+K

Note: Model includes country dummies. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively. Country dummy estimates are not shown.

137

8.4 Empirical Results

C. Partisanship measured as Annual Right Cabinet 4eg66eg6e?xhoh7_6l D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

!o76h_ t06od ?d o6e? l?h?ol?o /\sss A/K K/
s /+P dK/K d/P
K/Pm K/+m K/KP K/Pm dK/KAs d/
m dK/ dK/+P

t0el6[h?o[6 t06od ?d o6e? l?h?ol?o K/ +sss m/mm K/Asss A/
dK/ss dP/P dK/+\ss dP/P\ dK/KP dK/A dK/KK dK/m dK/K dK/PA mK K/
PP  
/mA K/KK K/Am

.0 oh_dF6-0 h?o t06od ?d o6e? l?h?ol?o K/\m / P K/mPsss / dK/+ss dP/ dK/

s d/++ K/Ks /\\ dK/PAsss dA/P /Am /A

Note: Model includes country dummies. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively. Country dummy estimates are not shown.

D. Partisanship measured as Cumulative Right Cabinet  4eg66eg6e?xhoh7_6l

D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

!o76h_ t06od ?d o6e? l?h?ol?o /Pmsss /Am K/+s /
dK/K\ d/mP K/PPsss /\ dK/K dK/A dK/KAs d/
K dK/Am dK/\

t0el6[h?o[6 t06od ?d o6e? l?h?ol?o K/ Psss m/PP K/ sss /P dK/ sss dP/\m K/PPsss /m dK/Ks d/+
dK/K dK/ + dK/ A dK/
A mK K/
P +
+
/
K/KKKK K/A

.0 oh_dF6-0 h?o t06od ?d o6e? l?h?ol?o K/Pm K/m+ K/
sss /  dK/\ss dP/ K/Pss P/A
K/K K/+m dK/PAsss dA/m K/+m K/+K

Note: Model includes country dummies. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively. Country dummy estimates are not shown.

proxy variables, three features have to be pointed out: Firstly, the social-democratic welfare regime cluster seems to be most affected by global economic forces whereas the conservative counterpart is hardly influenced. Secondly, Trade Openness exerts a significant welfare-retrenching influence thus providing evidence in favor of the competitiveness view. Thirdly, FDI Openness seems to be slightly positively associated with the social-democratic countries and partially negatively associated with the conservative countries.

138

8 A Quantitative Study on the Determinants of Welfare Spending

Table 28 presents similar results for the interaction model with SSTRAN as dependent variable. Most country-specific political and economic factors appear to determine the levels of social security transfers. Differences in significance levels and signs can be detected among the three welfare state clusters only with respect to the cumulative political partisanship variables and Population over 65. While the cumulative political partisanship coefficients show no significance in the case of Continental-European and Nordic countries, Cumulative Left Cabinet reports an unexpected sign for the Anglo-Saxon countries. As a consequence, one has to treat these results with caution. With respect to Population over 65, the coefficient indicates a significant positive effect on social transfer payments in the Anglo-Saxon and Nordic countries; however, the opposite is true for the Continental European countries.142 As mentioned in Chapter 5.5, Razin et al. (2002) have also found a negative relationship between population ageing and the size of the welfare state (p. 901). Yet given the fact that several scholars (for example, Bryant, 2003; Disney, 2007; Simonovits, 2007, 2003) have recently raised substantial concerns about this finding, the coefficient for the conservative country cluster has to be treated with caution. Further research is needed in order to shed more light on these conflicting results. This could involve the use of refined dependent variable measures, such as pension transfer payments as a percentage of GDP, which might help to elucidate the effect of population ageing on public welfare spending. Finally, with respect to globalization variables, the following features are worthwhile mentioning: Firstly, similar to the results presented in Table 27, liberal and social-democratic welfare regime clusters are more affected than their conservative counterpart. Secondly, while Trade Openness appears to exert a negative influence on both Anglo-Saxon and Nordic countries, Net Migration has a welfare-regressing impact on the former and a welfare-expanding impact on the latter. This finding can be explained by the fact that low-skilled migration is attracted to the most generous country cluster (Borjas, 1999, p. 609; Nannestad, 2007, pp. 528-529; Razin & Sadka, 2001, p. 79), thereby inducing the welfare state to further expand. Finally, FDI Openness appears to exert a negative influence on conservative welfare states. Reviewing these results in comparison with Swank’s (2002/ 2004, p. 104) earlier findings, no evidence is provided in favor of Swank’s (2002/ 2004) hypothesis stating that “international capital mobility should be unrelated or positively associated with cash income transfers in universal [social-democratic] and conservative institutional contexts and […] negatively associated with income maintenance in the liberal cluster” (p. 105). Looking now at the greater globalization picture, global economic forces do not ap142

Swank (2002/ 2004) has run a model for each welfare state cluster and finds similar results for the relationship between population ageing and income maintenance programs: While the size of the aged population reports a significantly positive coefficient in the case of the liberal and the social-democratic cluster, it appears to exert a negative, but insignificant effect on income maintenance in conservative welfare states (p. 104)

139

8.4 Empirical Results

pear to exert a singular retrenchment effect on liberal welfare states. Quite on the contrary, all welfare state clusters seem to be affected in a negative way by at least one facet of globalization. Overall, the findings of the models with aggregated welfare state measures suggest that globalization plays a role in welfare state development; however, in line with the baseline model results, domestic political and economic factors remain the most important driving forces of the contemporary welfare state. Table 28: Interaction Model Results - Dependent Variable: SSTRAN, 1975-2000 A. Partisanship measured as Annual Left Cabinet  4eg66eg6e?xhoh7_6l

D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

!o76h_ t06od ?d o6e? l?h?ol?o /
sss /m K/mss P/KK K/K K/AA dK/ m d/P+ dK/K dK/\+ dK/KAsss dP/
K dK/+Ps d/+

t0el6[h?o[6 t06od ?d o6e? l?h?ol?o dK/P s d/+ K/ sss P/\m dK/Kss dP/  K/mP K/
 dK/Pss dP/ dK/KP dK/ K/PP K/Am mPA K/\++ AP/\ K/KK K/A

.0 oh_dF6-0 h?o t06od ?d o6e? l?h?ol?o / Psss m/P K/mPsss m/\
dK/K dK/AA K/K\ K/P K/KP K/
 dK/K+ss dP/AP /+\ss P/ 

Note: Model includes country dummies. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively. Country dummy estimates are not shown.

B. Partisanship measured as Cumulative Left Cabinet  4eg66eg6e?xhoh7_6l

D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

!o76h_ t06od ?d o6e? l?h?ol?o /mmsss /
\ K/Ass P/+ K/K K/A dK/Asss d /+ dK/K dK/A+ dK/KAsss dP/
m dK/+s d/+\

t0el6[h?o[6 t06od ?d o6e? l?h?ol?o dK/P ss dP/K K/P+ss P/K\ dK/Kss dP/m K/KA K/ dK/msss dP/ dK/KP dK/A\ K/K
K/PP mPA K/+KP \/
 K/KK K/K

.0 oh_dF6-0 h?o t06od ?d o6e? l?h?ol?o / Psss m/KP K/m sss A/ dK/K dK/m+ dK/K dK/ K/K K/

dK/K+sss dP/m P/K+ss P/AK

Note: Model includes country dummies. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively. Country dummy estimates are not shown.

140

8 A Quantitative Study on the Determinants of Welfare Spending

C. Partisanship measured as Annual Right Cabinet  4eg66eg6e?xhoh7_6l

D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

6 !o76h_ t06od ?d o6e? l?h?ol?o /\sss /KA K/Ps /+ K/KP K/AK K/mmss P/ \ dK/K dK/\K dK/Kmsss dP/+ dK/
Pss dP/P

t0el6[h?o[6 t06od ?d o6e? l?h?ol?o dK/Pms d/
 K/ sss P/+\ dK/K
ss dP/ K dK/++ss dP/K dK/ss dP/Pm dK/KP dK/
K/P K/A mPA K/\++
m\+/ K/KK K/A

.0 oh_dF6-0 h?o t06od ?d o6e? l?h?ol?o / +sss m/P
K/mPsss m/\P dK/Km dK/+ dK/K d/P
K/KP K/+m dK/K+ss dP/mA /+Pss P/ P

Note: Model includes country dummies. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively. Country dummy estimates are not shown.

D. Partisanship measured as Cumulative Right Cabinet 4eg66eg6e?xhoh7_6l D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

!o76h_ t06od ?d o6e? l?h?ol?o /Ksss m/ + K/
s /+P dK/K d/  K/KP K/  K/KKm K/K dK/KAs d/+ dK/A
dK/\

t0el6[h?o[6 t06od ?d o6e? l?h?ol?o K/ /m\ K/ Asss /\ dK/ sss dP/\m K/
sss /A+ dK/K d/K K/KK K/K dK/P+ dK/+P mK K/
PP K\\/
P K/KK K/AA

.0 oh_dF6-0 h?o t06od ?d o6e? l?h?ol?o K/A K/PA K/
sss / + dK/
ss dP/ K/ ss P/PP K/Ks /+ dK/PAsss dA/ m /KP K/
\

Note: Model includes country dummies. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively. Country dummy estimates are not shown.

Turning to the interaction models with disaggregated welfare state measures, Table 29 presents the results for the dependent variable Benefit Generosity Index. The results across all three welfare state clusters appear largely consistent with the respective baseline model results. As proven above and mentioned in Chapter 5.2, population ageing seems to have a significantly positive effect on welfare spending. Annual political partisanship measures tend to have an influence on conservative welfare states consistent with the power resources approach (see Chapter 5.5). By con-

141

8.4 Empirical Results

trast, cumulative political partisanship measures report mixed results. All other country-specific political and economic variables are insignificant. Looking at the globalization proxy variables, the difference between the results found for this social spendTable 29: Interaction Model Results – Dependent Variable: Benefit Generosity Index, 1975-2000 A. Partisanship measured as Annual Left Cabinet 4eg66eg6e?xhoh7_6l D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e GF4x06ee6llxlLh6g hg6x06ee6llxlLh6g B6?x-oSh?o0exlLh6g B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

!o76h_ t06od ?d o6e? l?h?ol?o K/APss P/ K/K /K dK/Km dK/\ K/\ K/  K/ K/\ dK/s d/\ K/K K/\+ dK/KK d/P K/KKsss P/+ dK/ dK/

t0el6[h?o[6 t06od ?d o6e? l?h?ol?o K/m\sss /\m K/ /m
K/Km K/
K K/+\ss P/Pm dK/ \sss d /
dK/K\ d/ + K/ K/++ K/Ksss P/
K/KKss P/\ dK/+ dK/
mPA K/
+ A\K/P+ K/KK K/

.0 oh_dF6-0 h?o t06od ?d o6e? l?h?ol?o P/KAsss m/A\ dK/K\ dK/AP dK/ d/ \ dK/A d/A dK/P d/PA K/+ K/AK K/
K/K+ K/KK K/ P dK/KKP dK/+ dK/+ dK/KA

Note: Model includes country dummies. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively. Country dummy estimates are not shown.

B. Partisanship  measured as Cumulative Left Cabinet 4eg66eg6e?xhoh7_6l

D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e GF4x06ee6llxlLh6g hg6x06ee6llxlLh6g B6?x-oSh?o0exlLh6g B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

!o76h_ t06od ?d o6e? l?h?ol?o K/
ss P/mP K/ /PA dK/Km dK/ dK/K d/K K/ K/+P dK/Ps d/\
K/AA K/\ dK/K d/K\ K/KKsss P/\+ dK/m dK/PK

t0el6[h?o[6 t06od ?d o6e? l?h?ol?o K/APsss m/  K/K\ K/\P K/K K/\
K/Km K/A\ dK/ +sss d /K dK/K d/A K/m+ K/A K/Ksss P/
 K/KKs /
dK/ dK/ m mPA K/
P
mm
/ K/KK K/A+

.0 oh_dF6-0 h?o t06od ?d o6e? l?h?ol?o P/Psss m/m dK/K dK/m dK/P d/m dK/K dK/ + dK/m d/A K/P K/ P K/m K/PK K/KKP K/A+ dK/KK dK/AK dK/P+ dK/K\

Note: Model includes country dummies. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively. Country dummy estimates are not shown.

142

8 A Quantitative Study on the Determinants of Welfare Spending

C. Partisanship measured as Annual Right Cabinet 4eg66eg6e?xhoh7_6l D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e GF4x06ee6llxlLh6g hg6x06ee6llxlLh6g B6?x-oSh?o0exlLh6g B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

!o76h_ t06od ?d o6e? l?h?ol?o K/mms /\m K/K
/K\ dK/Km dK/\ K/mP /mA K/m /K dK/K d/m
K/m
K/m dK/K d/ K K/KKss P/AK dK/ dK/\

t0el6[h?o[6 t06od ?d o6e? l?h?ol?o K/m\sss /\ K/s /\A K/Km /KP d/sss dP/+ dK/ \sss d /+ dK/K\ d/m K/A K/
K/Ksss / K/KKss P/PA dK/+ dK/mK mPA K/
+ K\

/
P K/KK K/

.0 oh_dF6-0 h?o t06od ?d o6e? l?h?ol?o P/sss m/A dK/KA dK/  dK/ d/A dK/A dK/Pm dK/m d/A K/ K/P
K/P K/K
K/KK K/A dK/KK dK/m K/m K/Km

Note: Model includes country dummies. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively. Country dummy estimates are not shown.

D. Partisanship measured as Cumulative Right Cabinet 4eg66eg6e?xhoh7_6l D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e GF4x06ee6ll lLh6g hg6x06ee6llxlLh6g B6?x-oSh?o0exlLh6g B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

!o76h_ t06od ?d o6e? l?h?ol?o K/KA K/P K/K+ K/
 dK/Km dK/ K/P / K/K
K/\ dK/K d/A\ K/A K/\P dK/K d/ K/KKss P/AA dK/PK dK/ K

t0el6[h?o[6 t06od ?d o6e? l?h?ol?o K/\sss m/ K K/m /AP K/K K/\ dK/Pss dP/P dK/ sss d /PK dK/KA d/Km K/A\ K/+m K/Kss P/mP K/KKs /+ dK/+ dK/m mPA K/
++ PK /P K/KK K/P

.0 oh_dF6-0 h?o t06od ?d o6e? l?h?ol?o P/Asss A/P dK/K dK/PK dK/P d/AK dK/mss dP/Pm dK/K
d/KA K/\ K/m\ K/
A K/mP K/KK K/P dK/KKP dK/ dK/P dK/K\

Note: Model includes country dummies. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively. Country dummy estimates are not shown.

ing indicator and the previous two aggregated measures are quite interesting: Firstly, only liberal and conservative welfare state clusters seem to be affected by global economic forces. Secondly, Trade Openness appears to exert a U-shaped curvilinear in-

143

8.4 Empirical Results

fluence on the Benefit Generosity Index in Anglo-Saxon countries, whereas FDI Openness is associated with the same effect in the Continental European countries. Finally, Table 30 presents the results of the interaction model with the dependent variable Pension Generosity Index. Similar to prior findings, Population over 65 has a significantly positive effect on the welfare state across all three welfare state clusters. Table 30: Interaction Model Results – Dependent Variable: Pension Generosity Index, 1975-2000 A. Partisanship  measured as Annual Left Cabinet 4eg66eg6e?xhoh7_6l

D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e

!o76h_ t06od ?d o6e? l?h?ol?o K/ KAsss P/
K K/PKmsss m/KA+ dK/KP dK/Amm dK/K\
dK/P
K/KK K/K
 K/K ss P/K
+ dK/K+m dK/PK


t0el6[h?o[6 t06od ?d o6e? l?h?ol?o K/ msss /+
K/PP+sss /+KP dK/KP
d/KA K/ \m / \+ dK/K dK/\ K/KKP K/mP dK/P \ d/K+

B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

.0 oh_dF6-0 h?o t06od ?d o6e? l?h?ol?o K/\\ss P/mm dK/K+P dK/
+P dK/KP+ dK/A\ dK/\mss dP/Am dK/K\
sss dP/+
P K/KK
K/ K
dK/ m dK/m\m

mPA K/\

\\K K K/K

Note: Model includes country dummies. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively. Country dummy estimates are not shown.

B. Partisanship measured as Cumulative Left Cabinet 4eg66eg6e?xhoh7_6l D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

!o76h_ t06od ?d o6e? l?h?ol?o K/Psss m/
K K/P sss m/A

dK/K dK/mP dK/PKPsss dm/ A
K/KP K/A
P K/KP
ss P/A dK/K
\ dK/Pm

t0el6[h?o[6 t06od ?d o6e? l?h?ol?o K/ APsss /m K/Psss /K dK/K d/K K/KKP K/KAP dK/Km dK/A++ K/KK K/P dK/P+P d/P

.0 oh_dF6-0 h?o t06od ?d o6e? l?h?ol?o /KAsss /mmP dK/K\ dK/+A dK/Km dK/+\+ dK/P ss dP/Pm\ dK/K\Psss dP/A\\ K/K K/ +m K/
K/\A

mPA K/+K
Am K K/A+

Note: Model includes country dummies. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively. Country dummy estimates are not shown.

144

8 A Quantitative Study on the Determinants of Welfare Spending

C. Partisanship measured as Annual Right Cabinet 4eg66eg6e?xhoh7_6l D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e

!o76h_ t06od ?d o6e? l?h?ol?o K/P

sss P/
K/
\sss /+
 dK/KP dK/AP K/++ /KK K/KK K/\A K/K ss P/\+ dK/ m dK/ m

B-76 0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

t0el6[h?o[6 t06od ?d o6e? l?h?ol?o K/ mPsss /PK K/PP+sss /\ dK/KP+ d/KP\ dK/ K\ d/ dK/K dK/A\
K/KKP K/A dK/Pm d/Km

.0 oh_dF6-0 h?o t06od ?d o6e? l?h?ol?o K/m
ss P/P dK/K dK/\P
dK/K + dK/\\ K/ m K/
 dK/K+Psss dP/
AP K/KK K/
 dK/P dK/PAA

mPA K/\
A A\ K K/

Note: Model includes country dummies. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively. Country dummy estimates are not shown.

D. Partisanship measured as Cumulative Right Cabinet  6 4eg66eg6e?xhoh7_6l

D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

!o76h_ t06od ?d o6e? l?h?ol?o dK/Km+ dK/P++ K/
+sss m/PP dK/KPm dK/m+ K/K
\ss P/A dK/K dK/m
m K/K Pss P/m P dK/A dK/


t0el6[h?o[6 t06od ?d o6e? l?h?ol?o K/mA
sss /  K/PAPsss m/mP dK/K P d/
dK/K\s d/m dK/K dK/m + K/KKm K/P\P dK/Pmm d/+

.0 oh_dF6-0 h?o t06od ?d o6e? l?h?ol?o K/\+Pss P/ + dK/K
 d/K\P dK/K A dK/\\ dK/K + dK/ P dK/K
sss d / +\ K/K K/ A dK/
\ dK/P 

mPA K/+\ +mA K K/AA

Note: Model includes country dummies. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively. Country dummy estimates are not shown.

Unemployment, in turn, appears to be only positive and significant in the liberal and the conservative countries. Furthermore, in line with the baseline model results, political partisanship coefficients report mixed results that are not consistent with the power resources approach. With respect to the globalization variables, the following features are worthwhile emphasizing: Firstly, in contrast to prior assessments by Scharpf & Schmidt (2000, pp. 333-335), the generosity degree of conservative wel-

8.4 Empirical Results

145

fare economies’ pension systems appears to be least affected by global economic pressures. Secondly, Trade Openness happens to exert a positive influence on pension system generosity in Anglo-Saxon countries whereas the opposite effect has been reported for FDI Openness in Nordic countries. To summarize, in line with Brady et al. (2005) who have found “only limited support for the conclusion that globalization’s effects are systematically different across groups of countries or important only in European or liberal/uncoordinated countries” (p. 944)143, the author has only provided scarce evidence for the context-contingent view. Apart from Net Migration, none of the other globalization variables has reported a particular influence on welfare effort depending on the specific welfare state cluster. Thus, in contrast to the diverse hypotheses formulated by advocates of the context-contingent view (for example, Ellison, 2006, p. 178; Hall & Soskice, 2001/ 2004b, pp. 57-58; Scharpf & Schmidt, 2000, p. 335; Swank, 2003, pp. 60-61; Swank, 2002/ 2004, p. 5)144, none of the three types of welfare economies appears to be particularly prone to global economic pressures. Although the different facets of globalization appear to exert positive, negative, and curvilinear (U-shaped) effects on the welfare state, the negative relationship has predominated, thus stressing the notion of the competitiveness view. Overall, confirming Brady et al.’s (2005, pp. 944-945) ultimate conclusion, the findings of the interaction models suggest that globalization plays a role in welfare state development; however, in line with the baseline model results, domestic political and economic factors (in particular, population ageing) remain the most important driving forces behind the contemporary welfare state. In this context, evidence is partially provided for both the power resources and the new politics approach.

8.4.3

Structural Break Model Results

As depicted in Figure 8 to 11 in Chapter 8.1, the Nordic welfare states have experienced by far the most dramatic expansion and retrenchment process in recent decades. Against this backdrop, the author tests for a structural break in welfare state development. Evidence for the latter is found for two welfare state measures. According to the dataset, the structural break of PSEX took place in 1992 whereas the structural break of SSTRAN occurred in 1993. These dates fall into the period of economic and financial crisis afflicting all Nordic countries (with the exception of Denmark) mostly in the early 1990s (see Honkapohja, 2009, p. 7). 143

See Brady et al.’s (2005, pp. 940-943) sensitivity analyses for a more detailed review of their decomposing sample findings. 144 See Chapter 6.3.2 for a detailed overview of the principal hypotheses prevalent within the context-contingent view.

146

8 A Quantitative Study on the Determinants of Welfare Spending

Table 31 and 32 (see p. 148) present the results of the structural break models with the dependent variables PSEX and SSTRAN. Given the fact that the vast majority of Nordic government parties throughout the mid- and late 1990s belonged to the leftwing political spectrum, the author only includes Annual Left Cabinet and Cumulative Left Cabinet as proxy variables of political partisanship. As indicated, both variables report a positive association with welfare effort during the pre-structural break period and a negative association during the post-structural break period. This finding suggests a radical shift in social policy-making. Dahlström’s (2006) case study on Sweden offers some valuable insights to strengthen these results. According to his view, key state officials mainly in the Central Bank and the Ministry of Finance – the latter being loyal to the state in order to pursue their own career goals (p. 3, 7) – rather than political parties proved to be the principal architects of welfare state change (pp. 2-3).145 Precisely, Dahlström (2006) contends that given the severity and intricacy of the economic and financial crisis, political parties were willing to put ideology aside and abandon their usual strategy of shielding their constituents from welfare state cuts in order to enact the bureaucrats’ reform package as quickly as possible (pp. 2-3, 26-27). In this context, unemployment has played a crucial role. While in the pre-structural break period, the proxy variable Unemployment appears to have driven welfare state expansion, rising unemployment during the post-structural break period (period of economic crisis) seems to have had a welfare-regressing impact. Studying processes of welfare state retrenchment throughout the developed world, Huber & Stephens (2001a) have found similar evidence. Precisely, they have noted the following: “We found that the immediate cause of welfare state retrenchment was a large and apparently permanent increase in unemployment. With more people dependent on welfare state transfers and fewer people paying taxes to support the welfare state, budget deficits ballooned and governments moved to control and then reduce deficits by cutting entitlements” (p. 2). Before turning to the globalization proxy variables, it is worthwhile noting that the Nordic countries’ economic and financial crisis in the early 1990s was primarily the result of a series of domestic policy mistakes.146 While not triggering the crisis, certain forces of globalization certainly aggravated it. FDI Openness appears to have an inverted U-shaped relationship with SSTRAN and PSEX. The first steps of financial liberalization apparently did not have a negative influence on social spending. However, as Swank (2002/ 2004) notes, when financial deregulation was eventually fully 145

Dahlström’s (2006) case study analysis provides evidence in favor of the state-centric approach. See Chapter 2.2 for a more detailed explanation of this political-economic approach. 146 See for example Thakur, Keen, Horváth, and Cerra (2003, pp. 28-29) for a brief summary of the factors causing the Swedish economic crisis. Andersson & Viotti (1999) provide a thorough analysis of the Swedish banking crisis.

147

8.4 Empirical Results

implemented in the early 1990s, the “confluence of conditions […] – high international capital mobility and fiscal deficit approaching, and even exceeding, 10 percent of GDP – […] produced sufficient pressures for moderate policy reforms in the Swedish system” (pp. 140-141). Trade Openness, by contrast, seems to have exerted a U-shaped effect on PSEX and SSTRAN. Finally, the third proxy variable, Net Migration, indicates an inverted U-shaped relationship with the two welfare state measures. At first, net migration flows to generous welfare states such as the Nordic countries tend to have a welfare-enhancing effect as suggested by a number of scholars (Borjas, 1999, p. 609; Nannestad, 2007, pp. 528-529; Razin & Sadka, 2001, p. 79; Schou, 2006, p. 688). Interestingly, over the medium to long term Net Migration seems to reduce public social spending. An explanation for this development could be that an increasing number of immigrants integrate themselves into the domestic labor market Table 31: Structural Break Model Results – Dependent Variable: PSEX, 1980-2000

4eg66eg6e?xhoh7_6l D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e .% D0_h?o0ex0[6xA .% Je6-_0,-6e? .% FDxS0)?# .% Dh?olhel#o .% GF4x06ee6ll .% hg6x06ee6ll .% B6?x-oSh?o0e F6e-h Goe_heg B0)h, t0el?he? B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

.0 oh_dF6-0 h?o Ueeh_x!6?xth7oe6? t06o o6e? ?dl?h?ol?o K/ K/P
K/+sss A/A dK/PPss dP/A K/+ ss P/Km K/+ /m dK/ sss d\/ K /
/P dK/\ dK/+
dK/msss d /mm K/K+ K/AP d/ d/m dK/ dK/
 K/Kss P/K dK/
 dK/
dA/
sss dA/m dK/K\sss dm/\
d/
sss d+/A\ mm/AAsss A/AA +K K/
P A KA /AK K/KK K/P

.0 oh_dF6-0 h?o t-/x!6?xth7oe6? t06o o6e? ?dl?h?ol?o dK/+ d/Pm K/
sss m/+K dK/PKss dP/ K/P /A+ K/ K/
P dK/ Psss d\/
P/PK / K/m+ /mK dK/Asss dP/\A K/K\ K/m dK/P+ss dP/ A dK/m dK/+K K/K
/A dK/
 dK/ + d/ \sss dm/Km d/K
sss d /
P d\/Psss d+/
P AP/Psss A/\A +K K/
PP
KK/K K/KK K/PP

Note: SB refers to the period since the structural break has occurred. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively.

148

8 A Quantitative Study on the Determinants of Welfare Spending

Table 32: Structural Break Model Results – Dependent Variable: SSTRAN, 1975-2000

4eg66eg6e?xhoh7_6l D0_h?o0ex0[6xA Je6-_0,-6e?x FDxS0)?# Dh?olhel#o GF4x06ee6ll hg6x06ee6ll B6?x-oSh?o0e .% D0_h?o0ex0[6xA .% Je6-_0,-6e? .% FDxS0)?# .% Dh?olhel#o .% GF4x06ee6ll .% hg6x06ee6ll .% B6?x-oSh?o0e F6e-h Goe_heg B0)h, t0el?he? B-76x0 07l6[h?o0el cdlLh6g kh_gx #odlLh6 D07i #odlLh6 c#0

.0 oh_dF6-0 h?o Ueeh_x!6?xth7oe6? t06o o6e? ?dl?h?ol?o K/
msss /\ K/A
sss A/\m dK/+sss d /K K/A K/m K/A /P
dK/K
sss d /K
m/ msss /mm dK/PP d/A dK/K dK/ + K/K K/mm K/\ K/+ dK/m d/+ K/K /K d/m+ dK/\m dK/
dK/
m d/AP d/P d/+Asss dP/+m A/\ / P KK K/+P mA/KA K/KK K/


.0 oh_dF6-0 h?o t-/x!6?xth7oe6? t06o o6e? ?dl?h?ol?o K/m /P K/+sss A/+m dK/sss dP/+P K/ /KA K/PKs /+P dK/K
sss dP/
m/K
sss /PK K/A /P dK/ mss dP/\ K/K K/P dK/P+sss dP/
A dK/
s d/A K/Km K/\ dP/Km d/\ d/m
s d/ dP/AKs d/+ dP/ \sss d /mA K/m+ss P/  KK K/+ mK mP+/+ K/KK K/mK

Note: SB refers to the period since the structural break has occurred. Estimated using PCSE-OLS. “***”, “**”, and “*” mean that the coefficient is significantly different from zero at the 1%, 5% and 10% confidence levels, respectively.

over time, thereby reducing the total number of welfare recipients. Yet given the fact that the squared term does not reach significance and Schou’s (2006, p. 688) prior empirical results for the case of Denmark tell a different story, such an explanation needs to be treated with caution.

8.4.4

Shortcomings of the Quantitative Analysis and Further Research Needs

The author is aware that the above stated quantitative findings are not without limitations. Two critical issues in particular have to be pointed out: Firstly, although the multiple regression (MR) technique has become the standard research approach in

8.4 Empirical Results

149

comparative political economy, the author cannot discount the fact that MR models only allow for partial equilibrium analysis. Secondly, even if methods to control for conditional causal complexity are introduced, difficulties often remain with respect to the operationalization of “causal sequence, collective actors’ motivations, and rich historical and institutional contexts of strategic choices” (Swank, 2007, p. 366). Thus, the MR approach provides only partly a suitable framework to analyze the driving forces behind the contemporary welfare state. Consequently, in the next chapter the author will conduct several qualitative analyses to test and complement the findings derived from the quantitative study. Finally, before moving on to Chapter 9, it is worthwhile noting that further research is needed to assess if Scruggs’ (2004a, 2004b) two indices (Benefit Generosity Index and Pension Generosity Index) are reliable welfare state indicators.

9

A Qualitative Study on the Recent Development Process of Advanced Capitalist Welfare States

According to the findings of the quantitative study, domestic political and economic factors appear to be the principal driving forces behind the contemporary welfare state. Globalization, by contrast, seems to play only a subordinate role. Although global economic forces tend to exert significant positive, negative, and curvilinear effects on public social spending activities, the negative relationship – as suggested by the competitiveness view – slightly predominates. Building on these results, Chapter 9 conducts a set of qualitative analyses in order to examine the recent development process of advanced capitalist welfare states in more detail. The focus of the first part (Chapter 9.1) is laid on the revenue side of the welfare state. The author hereby undertakes a comparative exploration of how different types of welfare state revenue flows have evolved in the 17 advanced capitalist welfare states during the period 19752005. Special emphasis is hereby placed on Sweden for the following two reasons: Firstly, the country is regarded as the prototype of the modern welfare state (Thakur et al., 2003, p. 1). Secondly, it is the principal country of the social democratic welfare state cluster, the latter having by far experienced the most dramatic expansion and retrenchment of welfare effort in recent decades (see Figure 8 to 11 in Chapter 8.1). If tax competition is ongoing and harming advanced capitalist welfare states as suggested by the competitiveness view, these negative effects are likely to be found in the high-tax Nordic countries. Subsequently, following the comparative analysis of tax revenue flows, Chapter 9.2 examines the expenditure side of the 17 national systems of social protection. Studying principal welfare state indicators over a time span of maximum 24 years (1980-2003), the objective is hereby to analyze if expenditure patterns and underlying institutional configurations of each welfare state cluster have changed in recent times. Based on these insights, the author will be able to draw a well-founded conclusion regarding the future persistence of the three welfare capitalism clusters.

9.1

The Development of Welfare State Revenues in the Contemporary Era

9.1.1

The Competitiveness View and the Hypothesis of International Tax Competition

According to Musgrave (1997), the fiscal role of the government centers on the policy goal “to provide goods where externalities cause market failure, to address issues

152

9 A Qualitative Study on the Recent Development Process

of distribution, and to share in the conduct of macroeconomic policy. These functions, like those of the market, are not met with perfection, and their popularity varies over time. Nevertheless, they remain essential partners of the market system” (p. 156). While acknowledging the necessity of government intervention, several scholars adhering to the competitiveness view (see Chapter 6.3.1.1) have recently cast considerable doubt about the fiscal capacity of national policymakers in the contemporary era of globalization and locational competition.147 Tanzi (2002), for instance, contends the following: “With globalization, financial capital and highly skilled or highly talented individuals become more mobile because their options expand to other countries. High taxes or too constraining regulations create strong incentives for them to move elsewhere. The loss of highly talented individuals and the outflow of financial capital can have a negative effect on the growth rate and on the tax revenue of a country. In an open world where foreign competitors face lower taxes and fewer constraining regulations, it becomes more difficult and more costly for a country to maintain high taxes and more regulations” (p. 123).148 With contemporary welfare states facing increasing pressure to curtail taxation on mobile production factors (Razin & Sadka, 2005, p. 107), “it seems inevitable that the high-tax countries will have to succumb to the forces of tax competition and sharply cut their corporate tax rates” (Razin & Sadka, 2005, p. 4) to the lowest level of competing national tax systems. Consequently, if the tax base is not broadened capital income tax revenues are expected to fall entailing simultaneously a decline in public social provision (Razin & Sadka, 2005, pp. 35-36). Alternatively, governments – unwilling to roll back the welfare state – can either raise the tax burden on immobile production factors or finance social protection at the expense of a growing state deficit (Ganghof, 2000/ 2004, p. 601; see Schulze & Ursprung, 1999, p. 299; Sinn, 2001, pp. 15-16).149 However, both alternatives are hardly politically and economical147

For an elaboration of the mechanism of tax competition and a comprehensive overview of the literature on tax competition, see for example Genschel (2002), Pestieau (2006, pp. 52-60), Schulze & Ursprung (1999), Swank (2002/ 2004, Chapter 7), and Tanzi (2002). 148 Following Schulze & Ursprung (1999), it is worthwhile noting that countries do not only compete for mobile production factors by means of attractive taxes, but also through the public provision of a good infrastructure (p. 298; see also Siebert, 1999/ 2000, p. 245). Accordingly, “an isolated look at the relative tax burdens can therefore be misleading: If the additionally provided infrastructure increases a firm’s productivity by more than the tax levied to finance it, it pays off to locate in the country with the higher tax burden” (Schulze & Ursprung, 1999, p. 298). 149 With respect to the policy of shifting the tax burden on immobile production factors, Sinn (2001) distinguishes between two scenarios of tax competition: In the first scenario, “taxes that are paid on the factor capital would only suffice to pay for the infrastructure made avail(Continued at p. 153)

9.1 The Development of Welfare State Revenues in the Contemporary Era

153

ly feasible. As mentioned in Chapter 5.4, political tensions would likely ensue as an increasing part of the electorate could be expected to block efforts to impose a higher tax burden (Pierson, 2001a, p. 90). Furthermore, the market – acting as a sanctioning mechanism for public policy (Siebert, 1998, pp. 44-45; see Caspers & Kreis-Hoyer, 2004, p. 69; Tanzi, 1998, p. 13) – is likely to punish governments running large budget deficits by either pulling mobile production factors out of the country or charging higher interest rate premiums (Swank, 2002/ 2004, pp. 140-141). In light of these constraints, Tanzi (2002) holds the view “that a process of deep economic integration among countries will require a change in the role of the state in pursuing social protection. The end process would be a world where industrial countries will have to do less public spending, will reduce the use of tax expenditures for achieving particular social objectives, and will also have to reduce the role of specific socially-directed regulations. The state would have to reduce its role as a direct provider of social protection but, perhaps, it could (a) supervise compulsory systems of social insurance, as for example, in the area of pensions; and (b) it could play a larger role in developing markets where they do not exist. The state could also subsidize those who cannot afford buying privately-provided social protection” (p. 127). Finally, being aware of the harmful consequences of tax competition, the OECD (1998) recommended its member countries to intensify multilateral co-operation in order to combat welfare-destructing tax practices (p. 52). So far these policy initiatives are still at an early stage.150

9.1.2

The Hypothesis of Tax Competition in the Empirical Literature

In recent years, several scholars have empirically tested the hypothesis of international tax competition, reporting largely conflicting results (for detailed reviews, see Schulze & Ursprung, 1999, pp. 312-317; Swank, 2002/ 2004, Chapter 7; Swank & Steinmo, 2002, pp. 645-646; Ganghof, 2001, pp. 5-6): Studying the relationship between taxes and openness in 18 OECD countries for the time period 1965-1991, Rodrik (1997) finds “that taxes on labor respond positively to increases in lagged 149

(Continued from p. 152) able to this factor” (p. 15). Accordingly, all welfare state expenditures have to be shouldered by the owners of immobile production factors (p. 15). In the second scenario, taxes on the factor capital are equivalent to “the marginal congestion costs for a given infrastructure” (p. 16). Accordingly, “the owners of immobile factors, which include very many infirm, weak, or those who offer simple labour, receive no money from the rich capital owners but subsidise these capital owners in systems competition. This is not a race to the bottom but in a certain sense a race below the bottom” (p. 16). 150 See Dehejia & Genschel (1998) for a comprehensive analysis why it is hardly feasible to regulate tax competition through international co-operation.

154

9 A Qualitative Study on the Recent Development Process

openness, while taxes on capital respond negatively. […] in other words, there is strong evidence that as economic integration advances the tax burden of social insurance programs is shifted from capital to labor” (p. 63). Supporting this view, Razin & Sadka (2005) detect a positive relationship between trade openness and labor tax rates and a negative relationship between capital openness (foreign portfolio capital stock) and capital tax rates (p. 83). In contrast to the above findings, a number of studies have recently dismissed the hypothesis of international tax competition: Garrett & Mitchell (2001), for instance, do not find any significant empirical evidence for a negative effect of international economic integration on effective rates of taxation. Quite on the contrary, foreign direct investment appears to be positively associated with capital taxation while an increase in trade volume apparently induces the capital/ (labor + consumption) tax ratio to rise (pp. 171-173). Broadly similar results have also been found in other studies (Garrett, 2000, pp. 158-161; Garrett, 1998b, pp. 815-816; Swank, 2002/ 2004, pp. 254-256; Swank & Steinmo, 2002, p. 649151). Being puzzled by these findings, several scholars (for example, Ganghof, 2000/ 2004; Genschel, 2002; Swank & Steinmo, 2002) have recently started to analyze the causal forces behind the apparent absence of international tax competition. In this context, Genschel (2002) draws the following conclusion: “The race to the bottom has not taken place because tax competition was not the only problem that the welfare state faced during the 1980s and 1990s. There was also slow growth, high levels of precommitted budget expenditure, a mounting public debt, rising unemployment, and the threat of the shadow economy. These problems, which were not directly related to globalization, exerted countervailing pressures on tax policy that eclipsed the effects of tax competition. The welfare state is not trapped in a race to the bottom but boxed in between external pressures to reduce the tax burden on capital, on one hand, and internal pressures to maintain revenue levels and relieve the tax burden on labor, on the other” (p. 266; see also Ganghof, 2000/ 2004; Scharpf, 2000, p. 224; Swank & Steinmo, 2002, p. 651).

9.1.3

The Evolution of Welfare State Revenues: A Cross-Country Analysis

Building on the insights of the prior two sub-chapters and partly following Ganghof’s (2000/ 2004) methodological approach, the author conducts a comparative exploration of how different types of welfare state revenue flows have evolved in the 17 advanced capitalist welfare states during the period 1975-2005.152 Special emphasis is 151

Although Swank & Steinmo (2002) find a significant negative relationship between globalization (international liberalization of capital controls, trade) and statutory corporate tax rates, effective tax rates on capital appear to be unaffected by globalization (p. 649). 152 Besides Ganghof (2000/ 2004), a qualitative analysis of the evolution of welfare state revenues in OECD countries is also provided by Scharpf (2000).

9.1 The Development of Welfare State Revenues in the Contemporary Era

155

hereby placed on Sweden, for the following two reasons: Firstly, the country is regarded as the prototype of a generous welfare state (Thakur et al., 2003, p. 1). Secondly, it is the principal country of the social democratic welfare state cluster, the latter having by far experienced the most dramatic expansion and retrenchment of welfare effort in recent decades (see Figure 8 to 11 in Chapter 8.1). If tax competition is ongoing and harming advanced capitalist welfare states as suggested by the competitiveness view, the author expects to find these negative effects primarily in the high-tax Nordic countries.153 When analyzing welfare state revenues, the author looks primarily at the following five state revenue categories: corporate income taxes, personal income taxes, payroll taxes, social security contributions (both employers’ and employees’ contribution), and general consumption taxes. Before each of these categories is dealt with in more detail, Table 33 (see p. 156) depicts the development of total tax revenues (as a percentage of GDP) in all 17 OECD countries. As indicated, Sweden possesses the highest tax-to-GDP ratio among all advanced capitalist welfare states. At first sight, contrary to the predictions of the competitiveness view, the country has not experienced a race to the bottom with respect to the taxation system. This is proven by the fact that total tax revenues as a percentage of GDP are today approximately one fifth higher than the level reported for 1975. Besides Sweden, all other countries – with the exception of the Netherlands – have also witnessed a strong increase in tax revenues during the last three decades (compare Ganghof, 2000/ 2004, pp. 636-637; Scharpf, 2000, pp. 199-200). Interestingly, while total tax revenues have followed a steady growth path in most OECD countries, a different picture emerges in the case of Sweden. As depicted, Swedish tax revenues increased sharply until the early 1990s. Subsequently, the Nordic country experienced a dramatic fall (1990-1995), followed by a strong upswing for the remainder of the period under observation (1995-2005). With unemployment soaring and budget deficits ballooning in the early 1990s, “it was widely predicted that both the Swedish model and the Swedish economy were on their deathbeds” (Steinmo, 2002, p. 851; see Stein, 1991). Although the Swedish crisis started to unfold soon after major steps to liberalize the capital account had been implemented throughout the Nordic countries (see Figure 6 in Chapter 6.2.2), it is widely recognized nowadays that international economic integration was not the triggering force behind the economic downturn. Rather, as Swank (2002/ 2004) notes, the 153

See Steinmo (2003, 2002) and Thakur et al. (2003, pp. 81-93) for a more detailed assessment of the effects of globalization on taxation in the case of Sweden. Furthermore, a comprehensive case-study on the historic and contemporary development of the Swedish welfare state model is provided by Benner & Vad (2000/ 2004), Erixon (2000), Freeman, Topel, and Swedenborg (1997), Kjellberg (1998), Lindbeck (1997), Marterbauer (1998), Palme et al. (2002), Swank (2002/ 2004, pp. 133-142), and Thakur et al. (2003).

156

9 A Qualitative Study on the Recent Development Process

Table 33: Total Tax Revenues in Comparative Analysis, 1975-2005 t0e?o6l Ul?h_oh Ul?oh %6_Sothehgh  F6e-h Goe_heg  Ghe 6 6-he,P 46_heg 4?h_, hhe B6?#6_hegl B0)h, $C$A .)o?6_heg JQ J.U tFxd \ U[6hS6

0?h_xh]xc6[6e6lxhlxD6 6e?hS6x0xFD 
\A PA/+ /\
/A P/K +/m /A A/m m/ P+/\ PA/m PK/
m/P
/P G! Pm/A A/ PA/ /K


+K P/\
/K m/ /K m /K A/\ mK/ /m /K P
/\ PA/m m /m mP/m G !
PA/ A/P P/m A/P


+A P+/ mK/
mm/m P/A m/
/\ mP/+ / m/ / P\/m mP/ mP/ G;! P/ \/ PA/ \/K



K P+/A
/ mP/K A/
m/A m /A mP/K m/+ / \/+ P
/ mP/
m/K *4!; P/K / P\/ \/



A P+/+ m/ m / A/ m+/+ mA/\ mP/
\/P P/K mK/ P/+ m/A mK/
G! P\/+ m/\ P\/
\/


PKKK / mP/ mm/
A/ m
/m m\/P mm/m \/P /\ mP/ P\/K
/\ mP/ *4! K/A \/ P
/

/P

t#heS6xT $ PKKA K/
mP/ mA/m /m AK/ mm/K mm/ m/+ K/ m/K P\/m
/ m /\ *2!; P
/\ /A P\/ +/


\AdPKKA 
/\\ m/\ m/
m m/ + K/

PK/AA Pm/A+ /m /P /mP /K dA/K /m+ 4! P/PP /mK /m /Km

1. The total tax has been reduced by the amount of the capital transfer. 2. United Germany beginning 1991, Starting in 2001, Germany has revised its treatment of non-wastable tax credits in the reporting of revenues to bring it into line with the OECD guidelines. Source: Data compiled from OECD (2007d, pp. 74-76).

“confluence of conditions […] – high international capital mobility and fiscal deficit approaching, and even exceeding, 10 percent of GDP – […] produced sufficient pressures for moderate policy reforms in the Swedish system. […] international capital markets charge a premium in the form of higher interest rates on governments who run budget deficits” (pp. 140-141). Against this backdrop, the underlying causal forces behind the Swedish crisis are thus to be primarily found in a series of domestic macroeconomic policy mistakes which have been summarized by Thakur et al. (2003) in the following way: “Sweden made serious policy mistakes during 1974-92. It lagged other OECD countries in moving toward a low inflation environment and liberalizing capital markets by up to 12 years. The policy mistakes following Sweden’s postliberalization boom contributed to a banking crisis and sharp recession in the early 1990s. In the latter half of the 1980s, Sweden maintained a fixed exchange rate regime while having high inflation and rapid credit expansion. High inflation and tax effects interacted to produce negative real interest rates. Sweden then began a tax reform without suitable expenditure side adjustment. Real interest rates and the

9.1 The Development of Welfare State Revenues in the Contemporary Era

157

fiscal deficit skyrocketed, asset markets collapsed, most major banks’ capital base was virtually wiped out, and unemployment reached unprecedented levels. Public debt jumped by 30 percentage points of GDP during 1990-93, with the surge exacerbated by the substantial costs of recapitalizing the banking sector” (pp. 28-29).154 In light of the severe economic crisis, Sweden was under pressure to quickly stabilize the economy. Consequently, a restrictive monetary policy was pursued and a series of welfare state cuts introduced (Benner & Vad, 2000/ 2004, pp. 426-429).155 When the crisis was finally resolved in the second half of the 1990s, growth picked up again and so did tax revenues. Although the comparative analysis of total tax-to-GDP ratios did not reveal any evidence for a race to the bottom, further investigations are needed since total tax revenues data can obscure a potential shift of tax burden from mobile to immobile factors, the latter being central to the hypothesis of tax competition. Thus, Table 34 to 36 (see pp. 158-160) present the evolution of total government revenues derived from corporate income taxes, personal income taxes, and taxes on payroll and workforce. Interestingly, contrary to the hypothesis of tax competition, all countries – with the exception of Canada – have witnessed a dramatic increase in corporate tax revenues. Given the fact that statutory corporate income tax rates have on average declined from 43 percent in 1979 to 29 percent in 2002 across all OECD countries (Clausing, 2007, pp. 121-122), this finding is rather striking. However, as Clausing (2007) notes, “many factors other than tax rates affect the revenues that governments receive. Revenues depend not just on the statutory tax rates set by legislators, but also on the breadth of the tax base (e.g., the prevalence of loopholes and special arrangements), the opportunities for tax avoidance, the aggressiveness of corporate tax planners, the enforcement efforts exerted by government tax authorities, economic conditions that determine the profitability of firms, and the share of the corporate sector in the economy” (p. 117).

154

Andersson & Viotti (1999) provide a thorough analysis of the Swedish banking crisis. For a summary of the Swedish tax reform of the early 1990s, see Ganghof (2001, pp. 22-23), Thakur et al. (2003, p. 16). 155 Following Dahlström’s (2006) analysis, key state officials mainly in the Central Bank as well as in the Ministry of Finance – the latter being loyal to the state in order to pursue their own career goals (p. 3) – proved hereby to be the principal agents of welfare state change (evidence for the state-centric approach as explained in Chapter 2.2). Given the severity and intricacy of the economic crisis, the political parties were willing to put ideology aside and abandon their usual strategy of shielding their constituents from welfare state cuts in order to enact the bureaucrats’ reform package as quickly as possible (pp. 26-27).

158

9 A Qualitative Study on the Recent Development Process

Table 34: Corporate Income Tax Revenues in Comparative Analysis, 1975-2005 t0e?o6l Ul?h_oh Ul?oh %6_Sothehgh F6e-hx Goe_heg Ghe 6x 6-he, 46_heg 4?h_, hhe B6?#6_hegl B0)h, $C$A .)o?6_heg JQ J.U tFxd \ U[6hS6


\A /P / P/\ m/ /P /\ /+ /A /m / m/ /P / ! P/ P/P P/
P/

h]6lx0ext00h?6x4e 0-6xhlxD6 6e?hS6x0 FD 
+K 
+A 

K 

A PKKK PKKA /P P/ m/K m/P / A/
/m /m /m /m P/K P/ /
P/P P/K P/ /P /A / P/\ P/A P/
m/m /A /m P/P /\ P/ / /+ /P P/ P/K /m P/ A/A P/
A/\ ! / P/
P/+ P/A

/m /
P/P / / A/\ /K \/ !; /+ m/\ /
P/+

P/K P/P /\ / /+ /A /P /\ ! P/K / P/m P/\

P/ P/ /K P/\ /A m/ / /+ 4! /+ P/+ P/
P/\

A/
/ /+ /\ P/
/\ m/K +/
G!2 P/+ / P/ /


/ P/+ /\ /m P/+ m/ /+ /+ ! P/ /m / /


t#heS6xT $ 
\AdPKKA +m/ + m /\A P
/ d+/K P/\
m/P AA/A  / mP/+ \A/KK K/KK +/\A
\P/\ ! P /+ Am/AA /
K \K/m\

1. The total tax revenues have been reduced by the amount of capital transfer. The capital transfer has been allocated between tax headings in proportion to the report tax revenue. Source: Data compiled from OECD (2007d, p. 81).

Recently, several scholars have analyzed these factors providing empirical evidence that can help explain why corporate tax revenues have been increasing despite recent tax rate cuts: Firstly, Clausing (2007) finds that the operating surplus (relative to corporate value added) of the corporate sector in OECD member countries has been increasing while the corporate share of respective economies has largely remained constant (p. 123). Secondly, Devereux & Sorensen (2006) show empirically that many governments in developed countries have recently undertaken measures to broaden the corporate tax base (pp. 7, 52; see also Devereux, Griffith, and Klemm, 2002, p. 452). Finally, Ganghof (2000/ 2004) observes that countries are nowadays more inclined to strengthen anti-avoidance legislation and administrative capacity in order to combat international tax evasion (pp. 623-624).156 Turning now to the Swedish case, a picture consistent with the above findings emerges. Similar to other countries, Sweden has lowered statutory corporate income 156

For a detailed analysis of adjustment strategies governments apply to preserve their tax revenues, see Ganghof (2000/ 2004, pp. 603-624).

159

9.1 The Development of Welfare State Revenues in the Contemporary Era

Table 35: Personal Income Tax Revenues in Comparative Analysis, 1975-2005 t0e?o6l Ul?h_oh Ul?oh %6_Sothehgh  F6e-hx Goe_heg Ghe 6x 6-he, 46_heg 4?h_, hhe B6?#6_hegl B0)h, $C$A .)o?6_heg JQ J.U tFxd \ U[6hS6


\A / \/
P/
K/A P/m m/ /\ K/ \/P /+ A/K / P/m 
!4
/ m/ +/
K/+

h]6lx0exD6l0eh_x4e 0-6xhlxD6 6e?hS6x0 FD 
+K 
+A 

K 

A PKKK /\ P/+ P/P /\ /+
/K
/m +/ +/
/A A/K A/+  /m m/P m/K K/ /A m/\  /m  / PP/A P /m Pm/+ P/P PA/A P/\ m/+ A/ m/P m/A m/\ m/
m/A m/
+/K K/+
/
/
/P /m P/ 
!4
/
K/ K/ /m

K/ K/+
/K /+ +/
/ !* K/P
/+
/\ /A


/ K/
/
+/ K/ K/\ 42! K/ K/\ K/ P/K

K/P K/K K/m /K \/+ K/  ! K/ K/K K/K /m


/m
/A K/A A/\ /K K/ ;!* K/ K/
P/A /\

PKKA P/
/  /
/
Pm/A  /A \/ +/ +/ K/A A/K \/K
/\  !2 K/ K/
/ /

t#heS6xT $ 
\AdPKKA +/+A \/\P \/\A  / m/m
dm/P KA/m dP/  A/P+ \/ P K/KK d /
m dP/\\ @ ! ; K/mP dPm/+P \/+\ P/\

1. The total tax revenues have been reduced by the amount of capital transfer. The capital transfer has been allocated between tax headings in proportion to the report tax revenue. Source: Data compiled from OECD (2007d, p. 80).

tax rates while at the same time broadening the corporate tax base (Thakur et al., 2003, p. 84). As a result, corporate income tax revenues have increased during the period 1975-2005. Interestingly, for most of the period corporate tax revenue levels have been below the average level of all 17 OECD countries (see Table 34). Latter can be explained by the fact that Swedish tax policy has always been favorable to the production factor capital as Steinmo (2002) notes in the following: “Whereas large corporations and wealthy capitalists paid relatively low taxes (as long as they kept their capital in productive assets in Sweden), smaller, privately held firms and ordinary workers paid extraordinarily heavy tax burdens. […] Though rarely publicly stated, the explicit corporate tax policy goal of the Social Democratic governments in the postwar years was to squeeze capital into the large, internationally competitive manufacturing industries. At the same time, Sweden maintained an open international trade policy explicitly aimed at forcing Swedish firms to maintain international competitiveness” (p. 842). In line with these policy goals, it does not come at a surprise that the Swedish tax revenues derived from personal income as well as payroll and workforce (relative to

160

9 A Qualitative Study on the Recent Development Process

Table 36: Payroll Tax Revenues in Comparative Analysis, 1975-2005 t0e?o6l Ul?h_oh Ul?oh %6_Sothehgh  F6e-hx Goe_heg  Ghe 6x 6-he, 46_heg 4?h_, hhe B6?#6_hegl B0)h, $C$A .)o?6_heg JQ J.U

h]6lx0e Dh,0__xhegxk00 6xhl D6 6e?hS6x0 FD 
\A 
+K 
+A 

K 

A PKKK PKKA /A / / /\ P/K /m /m P/
P/\ P/ P/m P/\ P/ P/ d d d d d d d d d d K/+ K/+ K/\ K/\ d d K/ K/ K/P K/P K/P K/+ K/ K/P d d d d K/\ K/
K/
K/+ / /K /P K/ K/ d d d d d d K/ K/+ K/m K/m K/P K/P d K/P K/P K/ K/ d d d d d d d d d d d d d d d d d d d d d d d ! !4 ! ! !2 4! 4!G d d d d d d d K/K /A K/K d d d d d d d d d d d

1. The total tax revenues have been reduced by the amount of capital transfer. The capital transfer has been allocated between tax headings in proportion to the report tax revenue. Source: Data compiled from OECD (2007d, p. 85).

GDP) are by far higher than the average levels reported for the 17 OECD countries (see Table 35 and 36). Although these levels of tax burden normally appear exorbitant to people from outside Sweden, the citizens of the largest Nordic country are willing to shoulder them since they are aware “that a strong social welfare state, like the Swedish, helps finance a quality of life that low individual taxes cannot easily buy” (Steinmo, 2003, p. 43). Consequently, the so-called “exit option of skilled labor” has not been a threat to the Swedish welfare state in the past (Thakur et al., 2003, pp. 92-93). In light of these findings, it appears surprising that personal income tax revenues (relative to GDP) have been declining since the early 1990s. As indicated in Table 35, this phenomenon is not just limited to the Swedish case, but is also found in a range of other OECD countries. A recent study by de Mooij & Nicodème (2006) offers an explanation suggesting that the decline in personal income tax revenues is related to the recent fall in corporate income tax rates. Precisely, analyzing data on tax regimes, firm births, and legal forms of business across 20 European countries for the time period 1998-2003, de Mooij & Nicodème (2006) provide empirical evidence that when corporate income tax rates are cut, the resulting “tax gap between personal and corporate tax rates exerts a significant positive effect on the degree of incorpora-

9.1 The Development of Welfare State Revenues in the Contemporary Era

161

tion” of companies (p. 45). As entrepreneurs are changing the legal status of their firms, tax revenue is shifting from the personal towards the corporate income tax base (p. 2, 45). Having analyzed the development of tax revenues derived from corporate income, personal income, and payroll, the focus is now laid on the fourth sub-category of welfare state revenues, social security contributions. Table 37 (see p. 162) depicts the evolution of employees’ and employers’ social security contributions as a percentage of GDP in the principal country of each welfare state cluster (Sweden, Germany, and the USA). Reflecting the growth of the welfare state due to new spending pressures (for example, population ageing, rising unemployment, emergence of new household types) in recent decades, social security contribution levels in 2005 have by far exceeded the levels reported for 1975. This development stands in sharp contrast to the evolution of personal income and payroll tax revenues which have largely remained constant on average for the 17 OECD countries under observation. Following Ganghof (2000/ 2004), the shift in tax revenue structure towards social security contributions can be explained in the following way: “These contributions are not directly vulnerable to international tax evasion and avoidance since they are withheld at source. Governments may even have to act in order to keep them from rising because they are directly linked to social expenditures, the demand for which has increased strongly due to demographic change and rising unemployment. At the same time, the electoral costs of growing social security contributions may be low compared to other taxes because they are still, at least partly, perceived as insurance contributions” (pp. 630-631). Although social security contributions are less vulnerable to international pressures than tax bases of mobile production factors, it is worthwhile mentioning that governments have recently undertaken measures to limit or even reduce non-wage labor costs in order to foster private-sector employment and increase the international competitiveness of domestic companies in product markets (Ganghof, 2000/ 2004, p. 598).157 As indicated in Table 37, Germany and the USA have lately seen some modest reductions in social security contributions for both employers and employees. Sweden, by contrast, appears to be an awkward case: Employers’ social security contributions have considerably decreased while employees’ social security contributions have dramatically increased in recent years. Although a shift in tax burden from capital to labor has been occurring it is worth noting that unlike in Germany and the USA employers in Sweden still shoulder by far the largest share of social security contributions.

157

For a detailed analysis of the relationship between social security contributions and various types of employment, see Scharpf (2000, pp. 206-210).

162

9 A Qualitative Study on the Recent Development Process

Table 37: Social Security Contributions in Comparative Analysis, 1975-2005

$C$A 6-he, J.U

-_0,66l"x.0 oh_x.6 o?,xt0e?o7?o0elxhl D6 6e?hS6x0xFD 
\A 
+K 
+A 

K 

A PKKK PKKA @ 2!2 2! 2! ! 4! 4! A/ A/ A/\ A/ /m /m / P/P P/m P/\ /K /K / P/


$C$A 6-he, J.U

-_0,6l"x.0 oh_x.6 o?,xt0e?o7?o0elxhl D6 6e?hS6x0xFD 
\A 
+K 
+A 

K 

A PKKK PKKA ;! 4!
!G !; !G 2!; 2! / /\ /+ / \/P \/ /\ P/+ / /A /A /A /A /m

t0e?o6l

Source: Data compiled from OECD (2007d, pp. 83-84).

Finally, in order to complete this qualitative exploration of tax revenue flows, the author still sheds light on the general consumption tax category.158 As depicted in Table 38, all countries (with the exception of France and Norway) have witnessed a considerable increase in state revenues (as a percentage of GDP) derived from general consumption. This development can be explained in the following way: Although cross-border shopping has become easier, it is still negligible from a public finance perspective (see Ganghof, 2000/ 2004, p. 632; Thakur et al., 2003, pp. 88-90). Being aware of the low mobility of the general consumption tax base, several OECD countries (for example, Germany) have recently augmented their value added tax rate in order to offset the ongoing downturn of tax revenues observable in other tax categories. Precisely, as the OECD (2007d) notes, “the value-added tax has everywhere served to counteract the diminishing share of specific consumption taxes, such as excises and custom duties” (p. 24), the latter having fallen as a direct consequence of ongoing processes of trade liberalization (Ganghof, 2000/ 2004, pp. 632-633). In light of these findings, it does not come at a surprise that the general consumption tax revenues constitute the third most important source of welfare state financing in OECD countries. The aim of this sub-chapter has been to test the hypothesis of tax competition by analyzing the recent development of welfare state revenues in OECD countries. As demonstrated above and in line with prior studies (for example, Ganghof, 2000/ 2004; Swank, 2002/ 2004, Chapter 7; Steinmo, 2003, 2002; Swank & Steinmo, 2002), no empirical evidence can be found suggesting that advanced capitalist welfare states have suffered an erosion of fiscal capacity during the last decades. With the exception 158

Although the property tax constitutes an important source of welfare state financing, it is not included in this analysis since tax revenue levels have remained fairly constant on average in most OECD countries (see OECD, 2007d, p. 86).

163

9.1 The Development of Welfare State Revenues in the Contemporary Era

Table 38: General Consumption Tax Revenues in Comparative Analysis, 1975-2005 t0e?o6l Ul?h_oh Ul?oh %6_Sothehgh F6e-hx Goe_heg Ghe 6x 6-he, 46_heg 4?h_, hhe B6?#6_hegl B0)h, $C$A .)o?6_heg JQ J.U tFxd \ U[6hS6


\A /\ \/ /m m/K /

h]6lx0ex6e6h_xt0el-?o0exhlxD6 6e?hS6x0 FD 
+K 
+A 

K 

A PKKK PKKA /m P/P P/ P/A /\ m/ \/+ +/ +/P \/\ +/K \/
\/K \/K /
/\ \/ \/ / m/ A/ A/K A/ A/K
/
/
/A
/m
/A K/K

A/\ +/

/P +/A

\/ +/A

+/m \/


\/
\/A

+/P \/A

+/\ \/

A/K m/P /

/ m/ m/

A/\ \/ m/


A/
+/K *!2 P/ / /+ m/


/
\/\

! P/ A/P /
A/

/
\/+

!; P/+ /K P/K /

A/+ /+ A/ / \/ \/\ ;! /K / P/P /K

/A /
A/A /A /A +/\
! / / P/P /

/+ \/ /A P/m /
+/m !
m/K /\ P/ /m

/ \/\ /K P/ \/ \/

!G m/K /+ P/P /A

t#heS6xT $ 
\AdPKKA m/+ +/PP m/K PA/KK A/AP AP/ d+/m P/KK + / /\ P+/+ d/PA !22
K/m+ 
/ A PP/PP AK/m


1. The total tax revenues have been reduced by the amount of capital transfer. The capital transfer has been allocated between tax headings in proportion to the report tax revenue. Source: Data compiled from OECD (2007d, p. 89).

of the Netherlands, all countries under observation have experienced a strong increase in total tax revenues relative to GDP. Reviewing in more detail the analysis of the different sub-categories of welfare state revenues, a number of findings can help understand why international capital mobility has not exerted a significant downward pressure on mobile tax revenues: Firstly, although statutory corporate tax rates have been reduced, this has so far not affected tax revenue levels since governments have simultaneously broadened the corporate tax base (see Devereux et al., 2002, p. 452; Ganghof, 2000/ 2004, p. 640; Swank & Steinmo, 2002, p. 650). Secondly, the presence of domestic factors shaping tax policy has hampered tax competition from evolving. As mentioned in prior chapters, advanced capitalist welfare states have recently been facing increasing spending pressures from so-called post-industrial challenges (see Chapter 5 in this book; for a detailed overview, see also Pierson, 2001a). In the past, these expenditures were largely financed by augmenting the tax burden on labor. Nowadays, however, with unemployment rising and economic growth declining, governments have become aware of the necessity to limit or even reduce non-wage labor costs in order to overcome the unemployment problem and improve the international competitiveness of

164

9 A Qualitative Study on the Recent Development Process

domestic companies (see Ganghof, 2000/ 2004, p. 598; Scharpf, 2000, p. 224; Swank & Steinmo, 2002, p. 651). Against this backdrop, OECD countries cannot afford to engage in international tax competition that would further aggravate their alreadytense budgetary situation. Finally, a third explanation exists why tax competition has only had a limited influence on national systems of taxation: As the Swedish case study has demonstrated and Schulze & Ursprung (1999) have already noted long before, countries do not only compete for mobile production factors by means of attractive taxes, but also through the public provision of a good infrastructure (p. 298; see also Siebert, 1999/ 2000, p. 245). In line with Steinmo (2002), “Sweden currently appears to offer an example of some of the advantages of a high-tax system, one in which the costs extracted by the state from employers, workers, and consumers alike are indeed higher than in other competitive nations – but one in which the advantages may still outweigh the costs” p. 857). To summarize, the above stated findings largely confirm the results of the quantitative study. Global economic processes, in other words, tend to exert a small, constraining influence whereas domestic factors appear to be the driving forces behind the contemporary welfare state in OECD countries. Accordingly, contrary to the competitiveness view, but in line with prior findings (for example, Scharpf, 2000, p. 224; Steinmo, 2003, p. 44), the author believes in the persistence of diverse patterns of welfare capitalism. In order to substantiate this overall finding, a qualitative exploration of the expenditure side of the welfare state will be conducted in the following concluding sub-chapter.

9.2

The Development of Welfare Expenditure Patterns in the Contemporary Era

Similar to the comparative study of tax revenue flows, this chapter examines how different types of welfare state expenditures have evolved in the 17 OECD countries under observation. The objective is hereby to analyze if the expenditure pattern and the underlying institutional configuration of the three welfare capitalist systems have changed in recent decades. In keeping partly with Swank’s (2002/ 2004, Chapter 4-6) methodological approach, the development of each welfare state cluster is assessed according to a set of key indicators.159 The time span of the comparative analysis is up to 24 years, from 1980 to 2003; however, given the fact that capital account liberalization has only been enforced in a number of countries as late as the second half of the 159

Besides assessing the development of the three welfare state clusters based on key welfare state indicators, Swank (2002/ 2004, Chapter 4-6) also conducts a number of detailed country case studies. For a similar approach, see also Huber & Stephens (2001a, Chapter 7, Appendix).

9.2 The Development of Welfare Expenditure Patterns in the Contemporary Era

165

1980s (see Figure 6 in Chapter 6.2.2), the main focus is placed on processes of welfare state change occurring throughout the 1990s.

9.2.1

A Brief Summary Review of Welfare State Regime Analysis

Before conducting the comparative study, the author briefly reviews the historical core features of each welfare state cluster. Adhering to Esping-Andersen’s (1990/ 1998, 2000) welfare state regime approach, the three welfare state clusters differ along the following three key defining dimensions: the level of de-commodification, the modes of social stratification, and the degree of de-familialization.160 Based on this theoretical framework, the author has provided – largely in line with Esping-Andersen’s (1990/ 1998, pp. 26-29; 2000, pp. 74-86) prior findings – a classification of the 17 advanced capitalist welfare states for the year 1980 in Chapter 2.3.2. As indicated in Table 2, the three welfare state clusters incorporate the following core features: Firstly, the Nordic countries (Denmark, Finland, Norway, and Sweden) form the social democratic group of welfare states which is shaped by a generous, universally inclusive system of social protection. Precisely, its central characteristics involve high levels of de-commodification and de-familialization coupled with relatively low degrees of status differentiation. Secondly, the Continental European countries (Austria, Belgium, France, Germany, Italy, the Netherlands, and Switzerland161) and Japan cluster around the conservative welfare state regime. While they feature lower de-commodification and de-familialization scores than reported for Northern Europe, corporatism and etatism is much more pronounced than in the rest of the OECD world. Thirdly, the Anglo-Saxon countries (Australia, Canada, Ireland, UK, and the USA) constitute the liberal welfare state model. Given the modest universal provision of social protection, all indicators achieve much lower scores on average than in the previous two welfare state categories.162

9.2.2

The Recent Development Process of the Three Welfare State Clusters: A Comparative Analysis

Building on the prior theoretical groundings, the author will analyze the recent social policy developments in all three country groupings. The aim of this qualitative study 160

Chapter 2.3.1 provides a thorough review of Esping-Andersen’s welfare state regime approach. See also Esping-Andersen (1990/ 1998, pp. 21-26, Chapter 2 and 3) and EspingAndersen (2000, Chapter 4). 161 Although Switzerland does not appear to be the typical case of a conservative welfare state it is assigned to this group due to its relatively high level of de-commodification. 162 For a similar classification, see also Huber & Stephens (2001a, pp. 87-93), Huber & Stephens (2001b), Kitschelt et al. (1999/ 2005, p. 435), and Swank (2003, pp. 61-65).

166

9 A Qualitative Study on the Recent Development Process

is hereby to examine in how far the three welfare state clusters have preserved their regime-specific institutional configuration in the era of globalization. For this reason, a comparative exploration of principal welfare state indicators is conducted that takes a two-step analytical approach: Firstly, general measures of the size and structure of public social spending are studied; subsequently, the focus is laid on core features of the welfare state-labor market nexus.163 While all three welfare state clusters are simultaneously analyzed, special emphasis is again placed on the Nordic countries since the latter group has by far experienced the most dramatic development of public social spending in recent years (see Figure 8 to 11 in Chapter 8.1). 9.2.2.1 The Three Welfare State Clusters in Broad Perspective: Size and Principal Categories of Public Social Spending Giving a general overview of the recent development process of the three welfare state clusters, Table 39 depicts the evolution of total government disbursements (as a percentage of GDP), total public social expenditures (as a percentage of GDP), total private social expenditures (as a percentage of GDP), and the Gini coefficient for three time periods until 2000. As indicated and in line with Esping-Andersen’s (1990/ 1998, 2000) welfare state regime approach, the Nordic countries have developed by far the most generous system of social protection. Recently, however, the social democratic welfare state cluster (particularly Finland and Sweden) has witnessed a rollback of public social spending. As mentioned in prior chapters, Finland and Sweden (and to a lesser extent Norway) faced a severe economic crisis in the early 1990s. With unemployment rising and budget deficits ballooning, the Finnish and Swedish government were under pressure to stabilize the economy. Against this backdrop, policies were implemented that foresaw a tightening of eligibility criteria for major social protection schemes (Kuhnle, 2001, p. 5; Swank, 2002/ 2004, pp. 129-130). Furthermore, while holding onto the principle of universal welfare coverage, user charge fees were introduced in the area of public health services and a reduction of income replacement rates was enforced (Kuhnle, 2001, pp. 5-6). Although these measures succeeded in reducing existing disincentive effects of the welfare system on the overall functioning of the economy they entailed a rise in social inequality (see Lindbeck, 1997, pp. 1313-1314). Yet, as indicated in Table 39, the Nordic countries still appear to be the most equitable societies among the developed world. Furthermore, the high levels of public social commitments suggest that the social democratic model of the encompassing welfare state has not undergone any radical change (see Lindbom & Rothstein, 2004, p. 20).

163

Swank (2002/ 2004, pp. 126-127, 165-166, 187) provides a similar cross-country analysis of different welfare state indicators for the time period 1980-1995.

D7_o x.0 oh_x]6ego?6lP Do[h?6x.0 oh_x]6ego?6l

oeoxt06o o6e?m

7A*& X
*A X*7

); FG-TD[+QW-G*_FG-* Z.-TK_FG-TD-

7
*& XH*& XA*&

A+/m A
/A A /+ \/A 7X* X* &*?

A/ A /\ m/
 A7*A A*7 *

P+/+ P/+ PK/m P
/
?* AX*A &*H

/ / P/ /
A&*A AX*H H*

K/ P\/ Pm/\ K/A ?*
A*? *

K/m  h K/
/ * A*? X*

 /P /+ P/ *
A*? X*

/ /P P /P

AA* A* *

PP/+ PK/\ P /m P/

AA*H A&*A *H

P/ PP/+ PA/ P/

AX*& A&* *


PP/A P/ P/ Pm/

PKKK

Source: For all indicators see Appendix 4.

4. According to the OECD (2005d), "income distribution is measured by the Gini coefficient. […] The values range between 0 in the case of 'perfect equality' and 100 in the case of ‘perfect inequality’. An increase in the Gini coefficient thus represents an increase in inequality" (p. 94).

3. Private social expenditures (mandatory plus voluntary private) as a percentage of GDP (a: value for 1985).

2. Public social expenditures as a percentage of GDP.

1. Total government disbursements as a percentage of GDP.

A mm/A m\/P  /m


+Kxd
+


Kd

A 

dPKKK 
+Kd
+


Kd

A 

dPKKK 
+Kd
+


Kd

A 

dPKKK -ogd
+Kl -ogd

Kl

0?h_x0[6e-6e?xFol7l6-6e?l

F6e-h Goe_heg B0)h, .)6g6e

t0e?o6l

Table 39: The Recent Development of the Three Welfare State Clusters: Size of Public Social Spending and Income Equality, 1980-2000

9.2 The Development of Welfare Expenditure Patterns in the Contemporary Era

167

168

9 A Qualitative Study on the Recent Development Process

Turning now to the conservative and liberal welfare state cluster, social policy development during the 1990s seems to have been rather stable. As presented in Table 39, average public social spending levels have slightly risen in the former and slightly fallen in the latter group of countries. Income inequality has remained fairly constant.164 Although the overall public welfare effort has only marginally changed, two particular features of the recent development process are nevertheless worthwhile mentioning: Firstly, social security transfers relative to total public social expenditures have declined in the conservative country group (see Figure 8 and 9 in Chapter 8.1) providing scarce evidence that a shift towards more service orientation has been ongoing in recent years (for a detailed analysis, see Chapter 9.2.2.2). Secondly, given the modest universal provision of public social protection, the trend towards private social insurance in liberal welfare states has lately intensified. During the period 1996-2000, almost five percent of GDP was spent on the private welfare market (see Table 39). Following the comparative analysis of general welfare state measures, the author will now look in detail at the expenditure pattern of each welfare state cluster. For this purpose, Table 40 presents the development of expenditure levels (as a percentage of GDP) of the principal welfare state programs (namely, retirement, health, family, and unemployment) for three time periods from 1980 until 2000. In line with the prior findings, spending levels on three of the four categories of social protection have been highest in the Nordic countries. Only in the case of old-age expenditures, conservative welfare states have ranked first. Latter can be explained by the fact that the share of old people (aged 65 and above) in the total population has grown rapidly since 1990 while it has remained constant in the social democratic group (see Table 10 in Chapter 5.2). Apart from pensions, the provision of public health care has constituted the other major welfare state program. As presented in Table 40, the respective spending levels have recently converged across the three welfare state clusters. This development stands in sharp contrast to the evolution of public family and unemployment expenditures which indicate a growing divergence between the social democratic welfare states on the one hand and their conservative and liberal counterparts on the other hand. In particular, the gap in family-related spending that had already been quite high throughout the 1980s has considerably increased throughout the 1990s. In line with prior analyses (Table 13 in Chapter 5.3 of this book; see also Esping-Andersen, 2002c, pp. 34-46; Esping-Andersen, 2000, pp. 161-164; Taylor-Gooby, 2004, pp. 15-16), this empirical result suggests that the three welfare state clusters still vary considerably with regards to the role of the family as provider of social protection. In order to sub164

As mentioned in the introduction to this comparative analysis, special emphasis is placed on the Nordic countries. Accordingly, the author only presents average values for the conservative and liberal welfare state cluster. For a detailed analysis of the various countries comprising the conservative and liberal welfare state cluster, see Appendix 5 and 6.



P

m

H* H*7 X*

&*H &*& 7*A

&*X
*7 7*A

*& 7*7 7*A

Source: For all indicators see Appendix 4.

4. Public unemployment expenditures as a percentage of GDP (d: value for 1985).

3. Public family expenditures as a percentage of GDP (c: value for 1985).

2. Public health expenditures as a percentage of GDP.

*
* 7*H

* *A 7*H

A*
*& *A

*
*& *7

* *H *

*H *7 *H

A*
*7 *

A*A *X ?*


D7_o xf6h_?#x]6ego?6l D7_o xGh-o_,x]6ego?6l Je6-_0,-6e? ]6ego?6l D7_o x_gxUS6 ]6ego?6l 
+Kxd
+


Kd

A 

dPKKK 
+Kd
+


Kd

A 

dPKKK 
+Kd
+


Kd

A 

dPKKK 
+Kd
+


Kd

A 

dPKKK +/
+/+ \/ \ /
P/
/ /+ m/ m/\ /A /A +/A +/ A/A /A A/m P/A m /m / /m P/\ g \/ /
/ /\ /+ /m /m / K/ A/A7 /
K/A +/m
/

/\ + \/ \/P m/ m/A / K/\ P /+

1. Public old age expenditures as a percentage of GDP (b: value for 1985).

); FG-TD[+QW-G*_FG-* Z.-TK_FG-TD-

.)6g6e

F6e-h Goe_heg B0)h,

t0e?o6l

Table 40: The Recent Development of the Three Welfare State Clusters: Principal Welfare State Spending Categories, 1980-2000

9.2 The Development of Welfare Expenditure Patterns in the Contemporary Era

169

170

9 A Qualitative Study on the Recent Development Process

stantiate this finding, the analysis of the welfare state-labor market nexus conducted in the following chapter will partially revolve around the question to what extent work and family life are reconciled in each country cluster. 9.2.2.2 The Three Welfare State Clusters and the Welfare State-Labor Market Nexus Based on the comparative exploration of principal social spending indicators, the author will now examine the broader context of the three welfare capitalist systems. Precisely, the core features of the welfare state-labor market nexus in each cluster are thoroughly analyzed. In this context, Table 41 and 42 present detailed data on the recent development of the labor market structure, the employment legislation, and the prevalence of active labor market programs. With respect to the comparative analyses conducted in Table 41 and 42, the following findings are worth mentioning: Firstly, active labor market programs, being more abundant than in any other welfare state cluster, seem to constitute a vital element of the social democratic welfare economy (see Swank, 2002/ 2004, p. 124, 128). Sweden, for instance, tends to supplement “passive labor market measures with a wide array of active labor market programs […] that provide support conditional on some labor activity by the recipient. Programs of this kind in Sweden consist of selfemployment grants, subsidized on-the-job training, wage and employment subsidies, and training courses” (Thakur et al., 2003, p. 60). In comparison to the Nordic countries, labor market activation seems to be fairly underdeveloped in the Continental European and Anglo-Saxon world. Belgium seems to be the only country with a long tradition of active labor market programs. Only recently, starting in the early 1990s, several countries, including France, Germany, and the Netherlands, have begun to gradually shift their labor market policies from a passive stance towards a more active approach (see Appendix 5 in this book; Swank, 2002/ 2004, p. 164; Hemerijck, Unger, and Visser, 2000/ 2004, p. 186).165 Secondly, female labor force participation reaches far higher levels in the Nordic countries than anywhere else in the developed world. This phenomenon can be explained by an expanding public sector – with the exception of Finland, the average civilian government employment accounts for more than 20 percent of the workingage population – and a high degree of de-familialization. As mentioned in Chapter 2.3.1.3 and empirically confirmed in the prior comparative analysis of expenditure patterns, the state rather than the family or the market is considered to be the principal 165

See also Cox (1998) for a detailed case study on labor market activation in the Netherlands. Furthermore, Fitzenberger & Hujer (2002), Jacobi & Kluve (2006) as well as Seeleib-Kaiser & Fleckenstein (2007) provide a detailed overview of the historic development and recent changes in German active labor market policies.

*? ?*H ?*

*H ?*& ?*H

*7 ?*
?*


* X* 7?*H

&*H 7A* 7*7


*7 77*7 *

A7*H 
*X *X

A*7 
* A*

Source: For all indicators see Appendix 4.

4. Civilian government employment as a percentage of the working-age population (j: value for 1995).

3. Persons unemployed for 12 months or longer as a percentage of all unemployed (g: value for 1987; h: value for 1988; i: value for 1993).

2. Female employment as a percentage of the female working-age population (15-64) since 1980 (f: value for 1985).

H*7 X
*H ?*7


* &* ?*&


*

* ?*


*H
*A
*


G6-h_6x-_0,-6e?x to[o_ohex0[6e-6e?x !0eSd?6-xJe6-_0,-6e? Dh?o oh?o0eP -_0,-6e?m 
+Kxd
+


Kd

A 

dPKKK 
+Kd
+


Kd

A 

dPKKP 
+Ad

K 

d

 

\dPKKP 
+Kd
+


Kd

m 

Ad


 K/
/A /\ 
\K/\ P\/m P+/A PP/+ P/P P/ P/A \/m S C /m +/\  /A P/
P\/+ P/+  /
K/o 
 /m 6 # / K/
A/m \/ \ / P \/A \/
PK/
PP/\ K/ K/A # PK/\ P\/+ Pm/ P / P/P / P/A /+ \/\ \A \/P m/


U ?o[6x!h70xnh6?xD0Sh-l

1. Public spending on active labor market programs as a percentage of GDP (e: value for 1985).

); FG-TD[+QW-G*_FG-* Z.-TK_FG-TD-

.)6g6e

Goe_heg B0)h,

F6e-h

t0e?o6l

Table 41: The Recent Development of the Three Welfare State Clusters: The Welfare State-Labor Market Nexus (I)

9.2 The Development of Welfare Expenditure Patterns in the Contemporary Era

171

172

9 A Qualitative Study on the Recent Development Process

provider of social protection in social-democratic welfare states (see Esping-Andersen, 2000, p. 45, 61-64, 66, 80; Swank, 2002/ 2004, pp. 124-125, 129). Encouraging the reconciliation of work and family life through the provision of generous public family services, Nordic countries appear to be well adapted to the ongoing transformation of gender roles and household structures (see Chapter 5.3 of this book; Esping-Andersen, 2002b, pp. 20-21; Kuhnle, 2001, p. 5; OECD, 2007c, pp. 15-17). In contrast to the social democratic system, female labor force participation still appears to be fairly low in the conservative welfare economies. Although the traditional notion of the male bread-winner model is more and more on the wane, women have only gradually begun to enter the labor market in recent years. Two factors can explain why this is the case: Firstly, given the fact that family-oriented services provided by the state are still relatively underdeveloped (see Table 40), women are discouraged from working outside of the home (see Esping-Andersen, 2000, pp. 61-66; Swank, 2002/ 2004, p. 163, 167). Secondly, female labor force participation has also been hampered by sluggish employment growth, the latter being a direct consequence of the fact that conservative countries neither have developed a service-oriented welfare state nor established a private low-wage service sector for a long time (Scharpf, 2001, pp. 277-278; see also Chapter 5.1 in this book). Recently, however, several Continental European countries have launched a series of reform packages that have fostered the creation of a low-wage sector. Accordingly, the share of low-wage work has lately been on the rise. For instance in the case of Germany, it accounted for approximately 21 percent of total employment as of 2004, with approximately 64 percent hereof being low-wage female workers (Solow, 2008, p. 14; see also Bosch & Kalina, 2008). Turning now to the liberal group of welfare states, a different picture emerges. Female labor force participation tends to be significantly higher than in the conservative welfare state cluster. Latter can be reasoned by the fact that the market rather than the family is considered to be the main provider of social protection (see Esping-Andersen, 2000, pp. 61-65). Yet female employment does not reach the levels reported for the Nordic countries since a great number of households cannot afford private-market care. With family subsidies remaining low and public sector involvement as well as active labor market programs being insufficient, mothers in poorer households (particularly single parent households as demonstrated in Table 13 in Chapter 5.3) are forced to stay at home instead of entering the labor market (see Esping-Andersen, 2000, pp. 66-67). Thirdly, average long-term unemployment rates (measured as a share of all unemployed) in social democratic welfare economies have been significantly lower than the levels reported for the conservative country group. Precisely, they have fallen to an average level of 22 percent in Nordic countries while having remained at a fairly high average level of 40 percent in Continental European countries (including Japan) during the second half of the 1990s. According to Andersen & Svarer (2007), numerous commentators have attributed the outstanding labor market performance of the

A* A*A ?*H

A*& A* ?*H

); FG-TD[+QW-G*_FG-* Z.-TK_FG-*

A* A*? ?*&

/m P P/ P/P

PKK

H*7 *
X7*A

A
m  + &A* H*& 7*

\A + +
K

.oeS_6xD6l0ex .oeS_6xD6l0ex Te0x #o_g6e$ T?)0x #o_g6e$

4eo?oh_xD#hl6x0 Je6-_0,-6e?

X&*7 X*
*X

AK A mP A X*7 7
*7 7*X

\P  A AA

.oeS_6xD6l0ex .oeS_6xD6l0ex Te0x #o_g6e$ T?)0x #o_g6e$

!0eSd?6-xJe6-_0,-6e?

B6?x0?d0d)0xc6_h 6-6e?xch?6lxTPKKP$P

* A* *


A/ K/+  /\  /A

n6e

* ?*7 A*7

P/ A
/+ P/ P

k0-6e

6 ?o[6xc6?o6-6e?xUS6xT

\dPKKP$

Source: For all indicators see Appendix 4.

3. Average effective age of retirement (in years) during the period 1997-2002.

2. Net replacement rates for two phases of unemployment and two family types as a percentage of the average production wage level in the year 2002. The net out-of-work replacement rates for long-term unemployment are depicted after tax and including unemployment benefits, social assistance, family and housing benefits in the 60th month of benefit receipt. Excluding Italy from the conservative cluster the net replacement rates for long-term unemployment amount to 50.1 percent for the first group (single person without children) and 68 percent for the second group (single person with two children).

1. Aggregate indicator taking into account the employment protection regulation of regular employment, temporary employment, and collective dismissals. It ranges from 0 to 6, with higher values indicating a stricter employment protection legislation. For a detailed explanation of the scoring procedure of the aggregate indicator, see OECD (2004b, pp. 102-106).

/m P/ P/\ P/P

!h?6x

Kl

P/ P/ P/
/A

!h?6x
+Kl

.?o ?e6llx0 -_0,-6e?xD0?6 ?o0ex!6Sol_h?o0e

F6e-h Goe_heg B0)h, .)6g6e

t0e?o6l

Table 42: The Recent Development of the Three Welfare State Clusters: The Welfare State-Labor Market Nexus (II)

9.2 The Development of Welfare Expenditure Patterns in the Contemporary Era

173

174

9 A Qualitative Study on the Recent Development Process

Nordic countries “to the so-called flexicurity model combining flexible hiring and firing rules for employers with income security for employees” (p. 389). Yet, in their study of the Danish labor market, the two scholars find that the flexicurity model has not been at the root of the Nordic employment miracle. They argue as follows: “Whatever virtues this model may have, a low and stable unemployment rate is not automatically among them since the basic flexicurity properties were also in place during the 1970s and 1980s where high and persistent unemployment was prevalent. Labour market performance has changed due to a series of reforms during the 1990s, the main thrust of which was a shift from a passive focus of labour market policies to a more active focus on job search and employment. The policy tightened eligibility for unemployment benefits and their duration as well as introduced workfare elements into unemployment insurance and social policies in general. Thereby, policy makers attempted to strengthen the incentive structure without taking resort to general benefit reductions” (Andersen & Svarer, 2007, p. 389). Empirical evidence to support this line of argumentation is provided in Table 42. As indicated, both conservative and social democratic welfare economies have considerably loosened employment protection legislation (EPL) during the 1990s reporting an average level of EPL strictness of 2.0 and 2.1 respectively in 2003; yet as mentioned above and presented in Table 41, the gap in average long-term unemployment rates between the two clusters has significantly widened in recent years. Testing Andersen & Svarer’s (2007) above stated hypothesis, Table 42 presents a comparative analysis of data on net out-of-work replacement rates as a percentage of the average production wage for two hypothetical recipient groups (single person without children and single person with two children) for the year 2002. The dataset is hereby subdivided into two phases of unemployment: short-term unemployment and longterm unemployment (defined as unemployment starting in the 60th month of benefit receipt). At first sight, given the high generosity level of the social democratic welfare system, average net replacement rates for both recipient groups and periods appear to exceed the levels reported for the remaining two welfare state clusters. Yet it has to be noted that the average net out-of-work replacement rate for long-term unemployment in the conservative country group has been biased downwards since Italy apparently does not pay any long-term unemployment benefits.166 Consequently, 166

In this context, Italy appears to be a rather controversial case: While government expenditures on unemployment benefits are considerably lower than the Continental European average long-term unemployment rates are by far higher than the levels reported for the remaining Continental European countries (see Appendix 5). Following Lazear (1990, p. 699, 720721), the author holds the view that Italy’s poor labor market performance can be mainly attributed to the country’s job security rules, which have for long been one of the most restrictive throughout the OECD world. For a detailed case study on Italy, see Swank (2002/ 2004, pp. 201-211).

9.2 The Development of Welfare Expenditure Patterns in the Contemporary Era

175

if Italy is left out, a different picture emerges: Net replacement rates for long-term unemployment are on average lower in the Nordic than in the conservative countries (see Note 2 in Table 42; Appendix 5). Thus, unemployed people are more inclined to search for a job in the former than in the latter welfare state cluster. As a result, unemployment is less persistent in Northern Europe. Apart from less generous long-term unemployment compensation, retirement eligibility criteria also appear to be stricter in social democratic welfare states. As indicated in Table 42, both men and women participating in the labor market enter retirement on average one year later than in the conservative welfare state cluster. Finally, turning the analysis to Anglo-Saxon welfare economies, clear evidence is found in favor of the persistence of the residual welfare state model. As presented in Table 42, employment protection has remained fairly weak. Furthermore, in comparison to the social democratic and conservative welfare state cluster, net out-of-work replacement rates for both periods of unemployment are by far lower while effective retirement age is higher. Thus, social policy in liberal welfare states continues to incorporate a strong market orientation. To summarize, the aim of this sub-chapter has been to analyze the recent development of the expenditure pattern and institutional configuration of each welfare capitalist system. In line with Swank’s (2002/ 2004, Chapter 4-6) prior qualitative analyses, empirical evidence has been provided supporting the view that the three welfare state clusters have largely preserved their traditional features of social protection. Nevertheless a number of adaptations have been occurring that are worth mentioning: Firstly, while not converging towards Anglo-Saxon welfare state residualism, both social democratic and conservative welfare states have recently experienced a slight shift in the liberal direction. Precisely, as a direct consequence of the economic crisis of the early 1990s, social democratic welfare states have tightened eligibility criteria for major social protection schemes and enforced a reduction of income replacement rates (Kuhnle, 2001, pp. 5-6; Lindbom & Rothstein, 2004, pp. 5-7; Swank, 2002/ 2004, pp. 129-130). Conservative welfare states, in turn, have witnessed a change in labor market policies from a rather passive stance towards a more active approach (Swank, 2002/ 2004, p. 164; Hemerijck, Unger, and Visser, 2000/ 2004, p. 186). Secondly, with post-industrial challenges intensifying in all three welfare state clusters in recent decades (see Chapter 5 in this book; for a detailed overview, see Pierson, 2001a), spending on elderly people and public health has been on the rise. Furthermore, measures have been undertaken to raise female labor force participation; yet, as indicated above, spending on family-oriented services and transfer programs still differs widely between the three systems of social protection.

10

Summarized Findings and Conclusion

The aim of this dissertation has been to analyze the recent development process of advanced capitalist welfare states. Against this backdrop, the author has examined the central challenges of contemporary social policy. Precisely, two types of welfare state challenges have been distinguished: on the one hand, post-industrial developments involving processes of deindustrialization, population ageing, household structure transformation, and welfare state maturation (see also Pierson, 2001a); on the other hand, global economic forces originating from product, capital, and labor market globalization. As presented in Part II of the dissertation, various theories exist on how the developments in the domestic and global sphere shape social policy. Taking a holistic approach, the author has tested all these theories in order to identify the driving forces behind the contemporary advanced capitalist welfare state. Precisely, the empirical investigation in Part III has revolved around the following two research questions: Firstly, what has been the relative influence of post-industrial and global economic developments on recent welfare state change? And secondly, to what extent have different national systems of social protection preserved their core institutional features? In order to find answers to these questions, the author has applied a hybrid approach combining quantitative and qualitative analyses. Building on Brady et al.’s (2005) prior work, the quantitative study has involved a set of pooled cross-section time series model estimations (baseline, interaction, and structural break), with the dataset including up to 17 leading welfare economies for the time period 1975-2000. Besides controlling for all hypotheses on the determinants of social policy change, each model has been run with aggregated (PSEX, SSTRAN) and disaggregated (Benefit Generosity Index, Pension Generosity Index) welfare state measures as dependent variable. The findings of the quantitative study can be summarized in the following way: Firstly, largely supporting Brady et al.’s (2005, p. 940, 944-945) prior results, empirical evidence has been provided that advanced capitalist welfare states are primarily shaped by domestic political and economic factors. Among the post-industrial challenges, the ongoing demographic shift appears to play the most prominent role. By contrast, globalization seems to have a minor influence on the development of public social spending. Although global economic forces partly vary with respect to their association with the four welfare state measures, the negative relationship tends to dominate. Secondly, in contrast to the hypotheses formulated by advocates of the context-contingent view (for example, Ellison, 2006, p. 178; Hall & Soskice, 2001/ 2004b, pp. 57-58; Scharpf & Schmidt, 2000, p. 335; Swank, 2003, pp. 60-61; Swank, 2002/ 2004, p. 5), no empirical evidence has been found that a specific wel-

178

10 Summarized Findings and Conclusion

fare state cluster (liberal, conservative or social democratic) is particularly prone to global economic pressures. Only in the case of net migration flows, differences have somehow been observable between the Nordic and the Anglo-Saxon welfare economies. Finally, in line with the two prior findings, the dramatic expansion and retrenchment process of Nordic welfare states in recent decades has been mainly attributed to domestic political and economic developments. Complementing the insights of the quantitative study, two qualitative analyses have been conducted shedding light on both the revenue and expenditure side of the 17 advanced capitalist welfare states. Precisely, the first comparative exploration has aimed at examining how different types of welfare state revenue flows have evolved in the respective national systems of social protection during the period 1975-2005. In line with prior studies (for example, Ganghof, 2000/ 2004; Swank, 2002/ 2004, Chapter 7; Steinmo, 2003, 2002; Swank & Steinmo, 2002), no empirical evidence has been found that the welfare states have suffered an erosion of fiscal capacity during the last decades. In contrast to the hypothesis of tax competition, the rise in international capital mobility has neither induced a significant downward pressure on mobile tax revenues nor entailed a dramatic shift of tax burden from mobile to immobile production factors. Reviewing the academic literature, three reasons have been put forward to explain why this is the case: Firstly, although statutory corporate tax rates have been reduced, this has so far not affected tax revenue levels since governments have simultaneously broadened the corporate tax base (see Devereux et al., 2002, p. 452; Ganghof, 2000/ 2004, p. 640; Swank & Steinmo, 2002, p. 650). Secondly, the presence of domestic factors shaping tax policy has hampered tax competition from evolving. As mentioned above, advanced capitalist welfare states have recently been facing increasing spending pressures from post-industrial challenges. In the past, these expenditures were largely financed by augmenting the tax burden on labor. Nowadays, however, with unemployment rising and economic growth declining, governments have become aware of the necessity to limit or even reduce non-wage labor costs in order to overcome the unemployment problem and improve the international competitiveness of domestic companies (see Ganghof, 2000/ 2004, p. 598; Scharpf, 2000, p. 224; Swank & Steinmo, 2002, p. 651). Against this backdrop, OECD countries cannot afford to engage in international tax competition that would further aggravate their already-tense budgetary situation. Finally, a third explanation exists why tax competition has only had a limited influence on national systems of taxation. As the Swedish case study has demonstrated and Schulze & Ursprung (1999) have already noted long before, countries do not only compete for mobile production factors by means of attractive taxes, but also through the public provision of a good infrastructure (p. 298; see also Siebert, 1999/ 2000, p. 245). Similar to the comparative exploration of tax revenue flows, the second qualitative analysis has examined how different types of welfare state expenditures have

10 Summarized Findings and Conclusion

179

evolved in the 17 OECD countries under observation during the period 1980-2003. In this context, the following findings are worthwhile mentioning: Firstly, with post-industrial pressures intensifying in all three welfare state clusters, spending on elderly people and public health has been on the rise; yet, expenditure levels of family-oriented services and transfer programs still differ widely between the three systems of social protection. Secondly, Nordic and Continental European welfare states have recently experienced a slight shift towards more market orientation, particularly on the labor market. Nevertheless, despite these adaptations, the overall empirical findings suggest, in line with Swank (2002/ 2004, Chapter 4-6), that the three welfare state clusters have largely preserved their core institutional features in recent decades. In light of the summarized findings, the author draws the conclusion that advanced capitalist welfare states are still very much shaped by developments occurring within the domestic sphere. Globalization has so far only exerted a small constraining influence on social policy. Furthermore, the author believes in the future viability of the three welfare state clusters. Although revenue and spending pressures in national systems of social protection will intensify as a result of rising post-industrial and global economic challenges, the prior chapters have demonstrated that the advanced capitalist welfare state can adapt to a changing environment and simultaneously preserve its institutional configuration.

Appendix Appendix 1: Data Sources of the Production Regime Indicators in Table 6 . . . . 182 Appendix 2: Econometric Analysis – Descriptive Statistics, Data Sources, and Correlation Matrix of all Variables . . . . . . . . . . . . . . . . . . . . . 184 Appendix 3: Robustness Checks of Econometric Results . . . . . . . . . . . . . . . . . . 187 Appendix 4: Data Sources of the Qualitative Analysis conducted in Chapter 9.2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191 Appendix 5: The Development of the Conservative Welfare State Cluster in Comparative Analysis, 1980-2003 . . . . . . . . . . . . . . . . . . . . . . . 194 Appendix 6: The Development of the Liberal Welfare State Cluster in Comparative Analysis, 1980-2003 . . . . . . . . . . . . . . . . . . . . . . . 198

Fh?hx 0-o_6gx0-xf76x6?xh_/x TPKKmh$/

P/xt00goeh?6gx)hS6x7hShoeoeS

\/x!0eSd?6-x6-_0,-6e?xl6 o?,x Shhe?66gx7, o-l

/x!0eSd?6-x[0o 6d7hl6gx6_h?o0el#ox 76?)66ex #hl6xhegxl_o6xo-l

A/xU__ohe 6lxh-0eSx 0-6?oeSxo-l

m/x!0eSd?6-x[0o 6d7hl6gx6_h?o0el#ox 76?)66exo-lxhegx?#6oxoe[6l?0l

Fh?hx 0-o_6gx0-xf76x6?xh_/x TPKKmh$/

Fh?hx 0-o_6gx0-xf76x6?xh_/x TPKKmh$/

Fh?hx 0-o_6gx0-xf76x6?xh_/x TPKKmh$/

Fh?hx 0-o_6gx0-xf76x6?xh_/x TPKKmh$/

Fh?hx 0-o_6gx0-x%6 xF6-oSu1d Qe?xhegx!6[oe6xTPKKK$/x

Ugh?6gx0-xoll6xTPKKx/ mA$/

/xJeo0exg6elo?,

/x.?0 x-h6?x ho?h_oh?o0exhl 0 FD

to?h?o0e

D0g ?o0exc6So-6x4ego h?0

.66x4ego h?0xt00goeh?6gx)hS6x7hShoeoeS/

.66x4ego h?0xt00goeh?6gx)hS6x7hShoeoeS/

.66x4ego h?0xt00goeh?6gx)hS6x7hShoeoeS/

.66x4ego h?0xt00goeh?6gx)hS6x7hShoeoeS/

%6 x/ f/ !/xF6-oSu1dQe?xU/ R !6[oe6xc/xTPKKK$/xUxe6)xgh?h7hl6x0ex oehe oh_xg6[6_0-6e?xhegxl? ?6/xV+KS_!TQ_+Q+(Z_O-GZ-% x0_/xm B0/x xA
\dKA/

f76x6?xh_/xTPKKmh$x?h6x?#olxgh?hx0-x?#6x0__0)oeSxl0 6Ixfo lxU/xR Q6e)0?#,x!/ T

+$/xt006h?o0exhegxD0_o?o h_x 0e0-o xD60-he 6xoe U_6e?xF6-0 h?o xtho?h_ol-/xF(-ZTQ_+QTK_+ )+Z+K+D' x0_/xK  B0/x x d\P/

f76x/xchSoe t/ .?6#6elx / F/ %hg,xF/xR %6 o6_gx / TPKKmh$/x [+(TT5ZG-_V-K T-_)5T5-W_;T5T_)-5* B0?#)6l?6exJeo[6lo?,xJeo[6lo?,x0x B0?#xth0_oehxF6xJeo[6lo?,xhegx4egohehxJeo[6lo?,/xc6?o6[6gxnh,xKx PKKx0-x?#6xD0C6 ?xk67xlo?6Ix #??I)))/_ol0C6 ?/0S7_o h?o0el)6_h6gh?h)6_h6h 6ll/#?-x

oll6x /xTPKK$/xJeo0ex-6-76l#oxl?h?ol?o lxoe Pmx 0e?o6l/xY+Q5K'_T.+_ O-GZ-% x0_/xP
 B0/xx +dm
/

.0 6

Appendix 1: Data Sources of the Production Regime Indicators in Table 6

182 Appendix

to?h?o0e

Ugh?6gx0-xl?6[6dU76x4[6l6ex hegx.0lo 6xTPKKxPKKmx/x\K$/

Ugh?6gx0-xl?6[6dU76x4[6l6ex hegx.0lo 6xTPKKxPKKmx/x\K$/

D0g ?o0exc6So-6x4ego h?0

+/xn6gohex_6eS?#x0x?6e6


/x0 h?o0eh_x?hoeoeSxl#h6

Appendix 1: Continued

l?6[6dU76x6? h_/xTPKKxPKKm$x?h6x?#olxgh?hx0-x?#6x0__0)oeSxl0 6Ix JB.txT


$/

.66x4ego h?0xn6gohex_6eS?#x0x?6e6/

l?6[6dU76 6? h_/ TPKK PKKm$ ?h6 ?#ol gh?h 0- ?#6 0__0)oeS l0 6lI tF -_0,-6e?x?_00xT

\$/xG0xB0)h,IxtFx-_0,-6e?x?_00xT

$/

l?6[6dU76 n/ 4[6l6e / R .0lo 6 F/ TPKKm$/ .0 oh_ D0?6 ?o0e heg ?#6 G0-h?o0e 0 .o__lI U c6oe?66?h?o0e 0 ?#6 k6_h6 .?h?6/ 4e D/ U/ fh__ R F/ .0lo 6 Tgl/$ TZ-5Z-W + [TZ5TKZW(= - ,QW5Z55Z+QTK