The European Social Model under Pressure: Liber Amicorum in Honour of Klaus Armingeon [1st ed.] 978-3-658-27042-1;978-3-658-27043-8

Table of contents :
Front Matter ....Pages I-X
The European social model under pressure: Introduction (Romana Careja, Patrick Emmenegger, Nathalie Giger)....Pages 1-11
Front Matter ....Pages 13-13
Neo-corporatism and the responsiveness of democracy (Marius R. Busemeyer)....Pages 15-31
Social concertation in Europe during the Great Recession: A fsQCA-study of social partner involvement (Benedikt Bender, Bernhard Ebbinghaus)....Pages 33-50
Transformations of Swiss neo-corporatism: From pre-parliamentary negotiations towards privileged pluralism in the parliamentary venue (André Mach, Frédéric Varone, Steven Eichenberger)....Pages 51-68
Aufstieg und Auflösung der Deutschland AG (1896-2010): Unternehmensverflechtung in Deutschland, Frankreich und in den USA (Paul Windolf)....Pages 69-88
State-labour relations in illiberal times: The dissolution of social dialogue in Hungary (Rafael Labanino)....Pages 89-106
Front Matter ....Pages 107-107
Governing elites under pressure: Blame avoidance strategies and welfare state retrenchment (Markus Hinterleitner, Fritz Sager)....Pages 109-122
New evidence on the effect of partisanship on the welfare state (Carsten Jensen, Georg Wenzelburger)....Pages 123-137
The partisan politics of new social risks in advanced postindustrial democracies: Social protection for labor market outsiders (Duane Swank)....Pages 139-157
The politics of shifting burdens: German fiscal welfare corporatism (Christine Trampusch)....Pages 159-176
Social acceleration and social investment (Kees van Kersbergen, Barbara Vis)....Pages 177-191
Social investment, active labour market policies and migration (Giuliano Bonoli)....Pages 193-206
Die sozialpolitischen Folgen des Ersten Weltkriegs: Deutschland, Österreich und die Schweiz im Vergleich (Herbert Obinger)....Pages 207-227
Front Matter ....Pages 229-229
The varied sources of increasing wage dispersion (Evelyne Huber, John D. Stephens, Kaitlin Alper)....Pages 231-252
Lay explanations for poverty: A multilevel analysis of European public opinion (1976 – 2014) (Lionel Marquis)....Pages 253-275
Risk of technological unemployment and support for redistributive policies (Stefano Sacchi, Dario Guarascio, Silvia Vannutelli)....Pages 277-295
How does working time flexibility affect gender-specific work intentions? (Isabelle Stadelmann-Steffen)....Pages 297-312
Soziale Selektivität in der Bildung - kein politisches Problem? (Wolf Linder)....Pages 313-330
Welfare states, demographic transition and immigration policies (Philipp Lutz)....Pages 331-347
Staatstätigkeit in Demokratien und Autokratien: Befunde des internationalen Vergleichs (Manfred G. Schmidt)....Pages 349-362
Front Matter ....Pages 363-363
Polarisierung der Parteiensysteme in Zeiten des Populismus (Uwe Wagschal)....Pages 365-382
Income inequality, status decline and support for the radical right (Sarah Engler, David Weisstanner)....Pages 383-400
Wie „AfD-bewegt“ sind die Gewerkschafter? (Rüdiger Schmitt-Beck)....Pages 401-423
Indirect democracy in EUrope: The challenge of new political parties (Hans Keman)....Pages 425-445
The dilemma between bilateral agreements and immigration control: How would Swiss voters decide? (Pascal Sciarini)....Pages 447-460
Who cares? The limited effect of liberalization on satisfaction with democracy in Europe (Kai Guthmann, Anna Fill)....Pages 461-479
Exit oder Voice? Die Wahlbeteiligung politisch Unzufriedener (Anja Heidelberger, Marlène Gerber, Marc Bühlmann)....Pages 481-500
Policy or Office? The determinants of programmatic change in West European political parties (Reimut Zohlnhöfer, Frank Bandau)....Pages 501-518
Front Matter ....Pages 519-519
Die Politik der inneren Abwertung: Erzwungene Lohnsenkungen und ihre Alternativen (Martin Höpner)....Pages 521-536
The political economy of the German current account surplus (Andreas Busch)....Pages 537-554
Light at the end of the tunnel: Eurozone bailouts and satisfaction with democracy (Frank Schimmelfennig, Dominik Schraff)....Pages 555-566
Eurozone crisis management and the changing nature of the democratic deficit in the European Union (Yannis Papadopoulos)....Pages 567-581
More solidarity than meets the eye? Challenges and prospects for social Europe (Maurizio Ferrera)....Pages 583-598
Das Klima und die EU: Eine Diskursperspektive auf die deutsche und schweizerische Klimapolitik (Marlene Kammerer, Fadri Crameri, Karin Ingold)....Pages 599-623
Back Matter ....Pages 625-633

Citation preview

Romana Careja Patrick Emmenegger Nathalie Giger Editors

The European Social Model under Pressure Liber Amicorum in Honour of Klaus Armingeon

The European Social Model under Pressure

Romana Careja · Patrick Emmenegger · Nathalie Giger Editors

The European Social Model under Pressure Liber Amicorum in Honour of Klaus Armingeon

Editors Romana Careja Department of Political Science University of Southern Denmark Odense, Denmark

Patrick Emmenegger Department of Political Science University of St. Gallen St. Gallen, Switzerland

Nathalie Giger Department of Political Science University of Geneva Geneva, Switzerland

Additional material to this book can be downloaded from http://extras.springer.com. ISBN 978-3-658-27043-8  (eBook) ISBN 978-3-658-27042-1 https://doi.org/10.1007/978-3-658-27043-8 Springer VS © Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer VS imprint is published by the registered company Springer Fachmedien Wiesbaden GmbH part of Springer Nature The registered company address is: Abraham-Lincoln-Str. 46, 65189 Wiesbaden, Germany

Table of Contents

Tabula Gratulatoria Romana Careja, Patrick Emmenegger and Nathalie Giger The European social model under pressure: Introduction

IX

1

Part I Transformation of Neo-Corporatism Marius R. Busemeyer Neo-corporatism and the responsiveness of democracy

15

Benedikt Bender and Bernhard Ebbinghaus Social concertation in Europe during the Great Recession: A fsQCA-study of social partner involvement

33

André Mach, Frédéric Varone and Steven Eichenberger Transformations of Swiss neo-corporatism: From pre-parliamentary negotiations towards privileged pluralism in the parliamentary venue

51

Paul Windolf Aufstieg und Auflösung der Deutschland AG (1896-2010): Unternehmensverflechtung in Deutschland, Frankreich und in den USA

69

Rafael Labanino State-labour relations in illiberal times: The dissolution of social dialogue in Hungary

89

V

Part II Politics of Welfare State Reform Markus Hinterleitner and Fritz Sager Governing elites under pressure: Blame avoidance strategies and welfare state retrenchment

109

Carsten Jensen and Georg Wenzelburger New evidence on the effect of partisanship on the welfare state

123

Duane Swank The partisan politics of new social risks in advanced postindustrial democracies: Social protection for labor market outsiders

139

Christine Trampusch The politics of shifting burdens: German fiscal welfare corporatism

159

Kees van Kersbergen and Barbara Vis Social acceleration and social investment

177

Giuliano Bonoli Social investment, active labour market policies and migration

193

Herbert Obinger Die sozialpolitischen Folgen des Ersten Weltkriegs: Deutschland, Österreich und die Schweiz im Vergleich

207

Part III Political Determinants of Socio-Economic Outcomes Evelyne Huber, John D. Stephens, and Kaitlin Alper The varied sources of increasing wage dispersion

231

Lionel Marquis Lay explanations for poverty: A multilevel analysis of European public opinion (1976 –2014)

253

VI

Stefano Sacchi, Dario Guarascio and Silvia Vannutelli Risk of technological unemployment and support for redistributive policies

277

Isabelle Stadelmann-Steffen How does working time flexibility affect gender-specific work intentions?

297

Wolf Linder Soziale Selektivität in der Bildung - kein politisches Problem?

313

Philipp Lutz Welfare states, demographic transition and immigration policies

331

Manfred G. Schmidt Staatstätigkeit in Demokratien und Autokratien: Befunde des internationalen Vergleichs

349

Part IV Democracies in Hard Times Uwe Wagschal Polarisierung der Parteiensysteme in Zeiten des Populismus

365

Sarah Engler and David Weisstanner Income inequality, status decline and support for the radical right

383

Rüdiger Schmitt-Beck Wie „AfD-bewegt“ sind die Gewerkschafter?

401

Hans Keman Indirect democracy in EUrope: The challenge of new political parties

425

Pascal Sciarini The dilemma between bilateral agreements and immigration control: How would Swiss voters decide?

447

Kai Guthmann and Anna Fill Who cares? The limited effect of liberalization on satisfaction with democracy in Europe

461 VII

Anja Heidelberger, Marlène Gerber und Marc Bühlmann Exit oder Voice? Die Wahlbeteiligung politisch Unzufriedener

481

Reimut Zohlnhöfer and Frank Bandau Policy or office? The determinants of programmatic change in West European political parties

501

Part V Challenges for Europe Martin Höpner Die Politik der inneren Abwertung: Erzwungene Lohnsenkungen und ihre Alternativen

521

Andreas Busch The political economy of the German current account surplus

537

Frank Schimmelfennig and Dominik Schraff Light at the end of the tunnel: Eurozone bailouts and satisfaction with democracy

555

Yannis Papadopoulos Eurozone crisis management and the changing nature of the democratic deficit in the European Union

567

Maurizio Ferrera More solidarity than meets the eye? Challenges and prospects for social Europe

583

Marlene Kammerer, Fadri Crameri und Karin Ingold Das Klima und die EU: Eine Diskursperspektive auf die deutsche und schweizerische Klimapolitik

599

Contributors

625

VIII

Tabula Gratulatoria The following former and current colleagues of Klaus Armingeon wish to enter their names in this Tabula Gratulatoria as a tribute to his impact on the development of Political Science as a discipline and as a research community.

Professor Silke Adam, University of Bern Professor Hans-Jürgen Andreß, University of Cologne Professor Lucio Baccaro, University of Geneva and Max Planck Institute for the Study of Societies Professor Luciano Bardi, European University Institute Dr. Fabio Bertozzi, Department for Health and Social Intervention, Canton Vaud Professor Michelle Beyeler, Bern University of Applied Sciences Emeritus Professor Roger Blum, Ombudsman of SRG Deutschschweiz Emeritus Professor Dietmar Braun, University of Lausanne Emeritus Professor Francis Castles, Australian National University Professor Besir Ceka, Davidson College Professor Eileen Connolly, Dublin City University Emeritus Professor Thomas Cottier, University of Bern Emeritus Professor Michael Cox, London School of Economics and Political Science Professor Skyler Cranmer, Ohio State University Professor Roland Czada, University of Osnabrück Professor Manfred Elsig, University of Bern Professor Margarita Estevez-Abe, Syracuse University Professor Markus Freitag, University of Bern Professor Steffen Ganghof, University of Potsdam Dr. Simon Geissbühler, Ministry of Foreign Affairs of Switzerland Professor Anke Hassel, Institute of Economic and Social Research (WSI) Professor Silja Häusermann, University of Zurich Professor Kenji Hirashima, University of Tokyo Emeritus Professor Ursula Hoffmann-Lange, University of Bamberg Professor Simon Hug, University of Geneva Professor André Kaiser, University of Cologne Professor Ferdinand Karlhofer, University of Innsbruck Professor Niilo Kauppi, French National Center for Scientific Research (CNRS) Professor Herbert Kitschelt, Duke University IX

Professor Bernhard Kittel, University of Vienna Emeritus Professor Peter Knoepfel, University of Lausanne Professor Hanspeter Kriesi, European University Institute Professor Sandra Lavenex, University of Geneva Professor Lucas Leemann, University of Zurich Emeritus Professor Gerhard Lehmbruch, University of Konstanz Professor Philip Manow, University of Bremen Professor Cathie Jo Martin, Boston University Professor Wolfgang Merkel, Berlin Social Science Center (WZB) Professor Karolina Milewicz, University of Oxford Professor Wolfgang C. Müller, University of Vienna Professor Ferdinand Müller-Rommel, University of Lüneburg Emeritus Professor Dieter Nohlen, University of Heidelberg Professor Daniel Oesch, University of Lausanne Professor Anne Peters, Max Planck Institute for Comparative Public Law and International Law Professor Vello Andres Pettai, University of Tartu Professor Simona Piattoni, University of Trento Professor Jonas Pontusson, University of Geneva Professor Manuel Sanchez De Dios, Complutense University of Madrid Dr. Lisa Schädel, Kirchliche Kontaktstelle für Flüchtlingsfragen, Bern Professor Gijs Schumacher, University of Amsterdam Professor Jonas Tallberg, University of Stockholm Professor Kathleen Thelen, Massachusetts Institute of Technology Dr. Olivia Thoenen, Polsan Büro für Politikanalyse und -beratung, Bern Professor Jacob Torfing, Roskilde University Professor Jan van Deth, University of Mannheim Professor Adrian Vatter, University of Bern Dr. Emanuel von Erlach, Federal Statistical Office of Switzerland Dr. Reto Wäest, University of Geneva Dr. Jan Zutavern, University of Luxembourg

X

The European social model under pressure: Introduction Romana Careja1, Patrick Emmenegger2 and Nathalie Giger3

Introduction The central theme of this book is the challenges facing the European Social Model, its political institutions and the democratic elements embedded in its structures. Ever since the 1990s, discussions about the future of continental welfare states, the democratic deficit of Europeanlevel institutions or the emergence of right-wing populism as a mass political phenomenon have incorporated an element of ‘crisis’. Over the years, the tone has become even more severe following the Great Recession and its widespread consequences, not only on economic and social policy, but also on policy-making in general. In the aftermath of the 2008/09 financial crisis and the Euro crisis that followed, many commentators questioned the European Social Model and its viability. However, ten years later, and with the benefit of hindsight, it is worth asking: Is the European way of social concertation indeed facing a crisis, or are such views exaggerated? And, in a broader perspective, do recent events, when accumulated, threaten the very foundation of European democracies and the European Union in particular, or are they rather likely to prompt answers and solutions leading to more democratic systems, better equipped to cope with internal and external challenges? We invited renowned scholars to deploy their methodological and analytical tools in answering these questions. The general inquiry we proposed generated explorations in many directions. The contributions brought together in this volume raise numerous inspiring questions concerning the facets of neo-corporatist interest intermediation, the changing relationship between citizens and their polities, and the spill-over of EU politics into the political lives of its neighbours (such as Switzerland). In addition, they prompt reflections on the role of social investment policies, on quality and legitimacy of democracy, or on the social dimension of EU citizenship. This volume does not provide one, or ‘the’, answer. Yet, by painting a broader picture of the pressures facing the European Social Model, it produces a reliable and well-founded overview of the crisis-like events and their wide-ranging consequences, occurring in the last 30 years.

1 2 3

Syddansk Universitet [email protected] Universität St. Gallen [email protected] Université de Genève [email protected]

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 R. Careja et al. (eds.), The European Social Model under Pressure, https://doi.org/10.1007/978-3-658-27043-8_1

1

We believe that having such an overview, supplemented by detailed accounts on transformations of current practices, is a necessary step in becoming aware of the possible outcomes concerning social concertation, democracy, welfare states or social structures in the (not so distant) future of European states. All the contributors of this volume were also asked to connect their reflections to the work of a political scientist whose work, spanning over four decades, has influenced the thinking of generations of scholars pondering big and small questions concerning the political, economic and social trajectories of the European Social Model. Throughout his long and prestigious academic career, Klaus Armingeon has been particularly concerned with the European Social Model. Yet, while appreciating its positive sides, he has always had a keen eye for its challenges. As Martin Höpner observes in his contribution to this volume, for Klaus Armingeon, politics gets interesting when it hurts. Among his numerous books, chapters and journal articles, the (alleged) crisis of the European Social Model runs like a red threat through his work, be it in the form of painful reforms in the age of austerity, new social risks or political alienation. Yet, Klaus Armingeon has never been the one to see only the gloomy side of things. Typical for his work, as Kees van Kersbergen and Barbara Vis note in their contribution, is the conviction that controversies cannot be solved through endless theoretical disputes. Instead, rigorous and systematic (and comparative, one might add) empirical research must make the difference, identifying the real challenges to the European Social Model, while providing a badly needed dose of sober analysis in other cases. The contributions to this volume have all been written in this spirit. Before presenting the chapters in more detail, we will cast a limelight over the career of Klaus Armingeon.

Brief overview of Klaus Armingeon’s career and scholarly contribution On his webpage at the Institute of Political Science, Klaus Armingeon introduces himself as citizen of the European Union, of the Federal Republic of Germany and of Switzerland. These few words summarize the way he sees himself. However, it bears to emphasize that Klaus Armingeon is an active citizen, who, through his scholarly work as well as through his numerous interventions in mass media, has contributed significantly to the way we understand the workings of the political systems. After obtaining a master’s degree in Political Science and East European History at Eberhard Karls Universität Tübingen (Germany), Klaus Armingeon received his PhD in Political Science at the University of Konstanz (Germany) in 1982 with a thesis on Neocorporatist Income Policy: A Comparative Study of Income Policies in Western European Countries in the 1970s (Neo-korporatisiche Einkommenspolitik: Eine vergleichende Untersuchung von Einkommenspolitiken in westeuropäischen Ländern in den 70er Jahren, Armingeon 1983). Several research positions followed at reputed German universities (Konstanz, Mannheim, 2

Tübingen and Heidelberg). In 1993, he obtained his Habilitation in Political Science at the Ruprecht-Karls-Universität Heidelberg (Germany) with a thesis on State and Employment Relationships: An International Comparison (Staat und Arbeitsbeziehungen: Ein internationaler Vergleich, Armingeon 1994a), and soon after obtained an associate professorship (Extraordinarius) in political science at the University of Bern (Switzerland), where he was also appointed Director of the Institute of Political Science. In 1996, he became full professor at the same institute, a position he has held ever since. As professor of Comparative and European Politics, he has taught generations of students and has actively engaged with the scientific community. He has been a guest professor at Duke University (2002) and at the Leopold-Franzens University in Innsbruck (2003), Nannerl O. Keohane Distinguished Visiting Professor at Duke/UNC Chapel Hill (2010), and fellow of Collegio Carlo Alberto, Moncalieri/Turin (2014, 2017). He has been, and currently is, a member of numerous advisory boards of prestigious scientific institutions, such as ZUMA (Mannheim), Max Planck Institute for the Study of Societies (Cologne), FORS (Lausanne), Institute for Higher Studies (Vienna) and Hans Böckler Foundation (Berlin). He has been the first foreigner to hold the position of President of the Swiss Political Science Association, between 1999 and 2002. From 2004 to 2006, he has acted as the Chief Editor of the Swiss Political Science Review and he has been a member of editorial Boards of several high-ranking political science journals (currently Comparative Political Studies and West European Politics). Between 2006 and 2012, Klaus Armingeon has also been a member of the Board of the European Consortium for Political Research (ECPR), responsible for the Dossiers Research and Europe. Klaus Armingeon’s research themes illustrate his manifold interests and the breadth of fields to which he has brought a contribution. His scholarly focal points have been economic and social policy, social partners and political parties. These elements are present throughout his work, be that in the context of analyses focusing on countries, on supranational structures such as EU, OECD or trade regimes, or on contemporary processes such as economic crises. Some of the questions spurring his intellectual curiosity in recent years are: How successful are the reforms in direction of austerity or liberalization? How do national states react to OECD or EUled reform initiatives or advice? How do citizens inform themselves and how do they react to policies and reforms? How is the European democracy affected by economic crises, by the repositioning of various political actors, by interests of non-political actors or by supranational organizations and policy regimes? To seek answers to these and many other questions, Klaus Armingeon and his collaborators have launched many fruitful explorations. The many research projects bare witness not only of the breadth of his interests, but also to his continuous efforts to support the first steps of young scholars. Recent collaborations include the project Critical Junctures: An International Comparison of Institutional Reforms, which addressed the question of conditions for substantial reforms, while the project International Organizations and National Welfare States explored the link between the OECD as a supra-national organisation and its member states. The project 3

Social Integration and Political Participation focused on civic engagement of individuals as part and parcel of democratic societies, while the project Societal Conflicts and Forms of Government: A Comparison of 28 Eastern Countries collected data on the post-communist countries, tracking their institutional transformation on a multitude of political and institutional indicators. The project Political Consequences of Attitudes towards the Welfare State explored the conditions under which welfare state attitudes become consequential for political behaviour, while the project Constitutionalism and Multilayered Governance explored the process of “constitutionalisation” taking place within various regulatory trade regimes that emerge at the intersection of globalisation, international institutions and mutual commitments of individual states. Most recently, his projects Austerity: An International Comparison of the Implementation of Austerity Plans and their Social Consequences as well as Trajectories of Liberalization throw light on liberalization processes, the implementation of austerity plans and their social consequences. In the early years of his academic career, Klaus Armingeon focused on issues concerning working relationships, trade unions’ roles and impact, corporatism and neo-corporatism, initially at the state level, and – as the European level gained more emphasis – at the supranational level. For example, in several of his first articles, he analysed the political and institutional context explaining the ability of trade unions to represent their members, and secure the rights and work conditions for workers (Armingeon 1983, 1986, 1987, 1988, 1989a, 1989b, 1993, 1994a, 1997b; Armingeon and Schmitt 1986). As concerns about labour relations trickled up to the European level, with several discussions and proposals at the end of 1980s and beginning of 1990s, his analysis led him to conclude that short-term prospects for convergence in European industrial relations were low (Armingeon 1991, 1994b). The factors he emphasized then, such as diverse traditions in the European countries and the divergence in interests of political actors and social partners, remain relevant today when the limited progress in the area of a common EU industrial relation system is addressed. In the 1990s and years after, the scholarly interests of Klaus Armingeon broadened significantly. Pursuant his appointment at the University of Bern, he embarked upon what became an in-depth exploration of the functioning and specificities of the Swiss political system. His studies in this vein focus on the Swiss political economy (Armingeon and Emmenegger 2007; Armingeon 1998b), welfare protection (Armingeon, Bertozzi and Bonoli 2004; Armingeon 2001a), party politics (Armingeon 1995a, 1999), institutional makeup and its effects (Armingeon 2001b, 2002a), labour market (Armingeon 1997a, 2001b) or direct democracy (Armingeon 2000). Other studies have examined Switzerland or negotiation democracies more generally in a comparative perspective (Armingeon 1997b, 1998a, 2002b, 2004a, 2004b; Armingeon and Sciarini 1996; Armingeon and Freitag 1997; Armingeon and Hirashima 2004). Klaus Armingeon also addressed major societal challenges, notably in his highly influential work on the reform of the welfare state (Armingeon et al. 2004, 2016; Armingeon and Beyeler 4

2004; Armingeon 2005; Armingeon and Bonoli 2006; Armingeon and Giger 2008). His keen interest in institutional change (Armingeon 2004a; Armingeon and Careja 2008) and in the interaction between international organisations and nation states (Armingeon 1995b; Armingeon and Milewicz 2008; Armingeon 2007) has resulted in publications shining a critical light on these topics. Equally important, he focused on the political repercussions of socioeconomic changes, for instance in his work on political behaviour and preferences (Armingeon 1994c, 1995a, 2001b; Armingeon and Schädel 2015; Armingeon and Guthmann 2014) and the policy determinants of socio-economic outcomes (Weisstanner and Armingeon 2018). After 2008/09, a major focus has been the Euro crisis and its consequences. His work has addressed how the European states have managed the aftermath of the crisis (Armingeon 2012a; Armingeon and Cranmer 2017; Armingeon, Guthmann and Weisstanner 2016a) as well as the consequences of the austerity policies (Armingeon and Guthmann 2014; Armingeon and Ceka 2014; Armingeon, Guthmann and Weisstanner 2014, 2015, 2016b). Yet, Klaus Armingeon’s contribution does not rest with his publications alone. He has long recognized the value of data quality for social science research. Therefore, he has launched several projects aimed at creating large databases, which he fully opened to the scientific community. The large number of citations for the Comparative Political Data Set (see www.cpds-data.org), developed together with researchers at the Bernese Political Science Institute, initially including the OECD countries and subsequently incorporating the 28 postcommunist countries of Eastern Europe and former Soviet Union as well, clearly indicate the widespread appreciation of the dataset. More recently, he has joined forces with Lucio Baccaro to develop a database of liberalizing policy reforms – equally intended for the benefit of all scholars. Last but certainly not least, over twenty public appearances in the last two years alone consolidate Klaus Armingeon’s public image as an expert in European politics. His media appearances address a variety of topics – from current hot issues for Swiss politics, to explaining the positions of the EU, Germany and other major EU-level actors – both in written media and in radio and TV interventions. Far from being an academic hiding in an ivory tower, he has always engaged with the public and disseminated his research to the larger audience and has never shied away from sharing inconvenient truths.

This volume This collective volume honours Klaus Armingeon’s numerous contributions to the advancement of the field of political science. The chapters engage with his most important work and therefore we divided the volume into five sections, corresponding to some of the main themes of his work: the transformation of neo-corporatism, the politics of welfare state reform, the political determinants of socio-economic outcomes, democracies in hard times and the current challenges facing the European Union. 5

The first section of the volume includes chapters, which interact with Klaus Armingeon’s many scholarly contributions concerned with coordinated capitalism, labour relations and neocorporatism. Several authors take these reflections as their starting point. Building on Armingeon (1986), Marius R. Busemeyer explores the relationship between neo-corporatist interest intermediation and the responsiveness of democracy. Engaging with recent concerns about interest group influence and its effect on democratic policy-making, he critically examines the promises and challenges of neo-corporatist decision-making with regard to its ability to ensure the proper representation of interests that are difficult to organize and mobilize. Benedikt Bender and Bernhard Ebbinghaus focus on the responses of European democracies to the Great Recession. In particular, they examine the economic, political and labour relations conditions leading to tripartite social concertation efforts. They show that labour relations and political – but not economic – factors explain why social partners are included in crisis management. Several contributions are concerned with the transformation of neo-corporatism and labour relations (cf. Armingeon 2011). André Mach, Frédéric Varone and Steven Eichenberger document a reconfiguration of the Swiss neo-corporatist political regime toward a more pluralist system, in which interest groups more actively target political parties and parliament. Nevertheless, economic groups remain dominant in the domains of economic and social policies. Paul Windolf focuses on the rise and decline of corporate networks in Germany. Offering opportunities for the coordination of market strategies and the self-monitoring of management, the structure of these corporate networks has fundamentally changed in recent decades, leaving little discernible difference between Germany and the United States of America. Rafael Labanino’s examination of the dissolution of social dialogue in Hungary under Orbán governments lends support to the argument that parties’ ideologies and party competition play a role in the development of labour relations in Eastern Europe (cf. Armingeon 2012b). The second section focuses on the politics of welfare state reform. Markus Hinterleitner and Fritz Sager examine blame avoidance strategies and welfare state retrenchment. Categorizing the different blame avoidance strategies that governing elites can employ to avoid electoral punishment, they show that these strategies may have rather diverse effects and leave governing elites with several options to pursue reform. Carsten Jensen and Georg Wenzelburger explore the effect of partisanship on welfare state activities. Focusing on unemployment insurance and old age pensions, two paradigmatic examples of “old” social risks, they find that the partisan effect of governments has declined since the 1970s. Examining “new” social risk policies (cf. Armingeon and Bonoli 2006), however, Duane Swank shows that social democratic incumbency has a positive effect on social policies addressing the needs of labour market outsiders, in particular in the presence of encompassing labour organizations. Christine Trampusch explores the politics of burden shifting in Germany’s fiscal welfare corporatism. She demonstrates that the costs of industrial restructuring were offloaded onto the 6

social insurance budget, which brought short-term relief in several policy fields, yet triggered a negative feedback process and ultimately led to a transformative change of German welfare corporatism. Social investment policies have been a prominent topic in recent work on welfare state reform. Kees van Kersbergen and Barbara Vis examine the temporal dimension of these policies. They argue that the accelerating pace of technological change tends to complicate the assumptions of social investment policies and might ultimately result in higher inequality. Similarly, Giuliano Bonoli critically assesses the adequacy of social investment policies. Starting from the observation of increasingly multicultural societies, he suggests that social investment policy must be adapted to deliver on the promise of maintaining social cohesion. Finally, Herbert Obinger explores the role of the First World War in shaping the welfare state in Austria, Germany and Switzerland. He identifies important parallels in how the war affected social policies in these three countries, although the effects were somewhat weaker in Switzerland. Klaus Armingeon’s work has always been concerned with the political determinants of socio-economic outcomes. The third section of this book looks at several important policy fields also been prominently featured in his work. Evelyne Huber, John D. Stephens and Kaitlin Alper examine the varied sources of increasing wage dispersion. Among others, they identify several important countervailing factors such as union density, wage coordination, employment protection legislation and investment in human capital. Focusing on “lay” explanations for poverty, Lionel Marquis shows that self-interest considerations mostly drive poverty attribution. However, following the 2008 economic crisis, explanations relating poverty to social injustice have risen to unprecedented levels in most European countries. Stefano Sacchi, Dario Guarascio and Silvia Vannutelli examine the effects of technological change on public opinion. Using data on Italy, they show that occupational risk related to technological change is an important driver of redistribution preferences. Isabelle Stadelmann-Steffen analyses how flexible work arrangements affect the genderspecific allocation of time on paid work and care work. She finds that women spend more time in the labour market under flexible work conditions, although this effect does not apply to women with traditional childcare preferences. Wolf Linder examines differences in educational attainment in Switzerland, finding that family background explains a significant part of education success. He also investigates why there seems to be little willingness on behalf of the governing elites to deal with these inequalities and take advantage of the untapped potential. Philipp Lutz explores how welfare state regimes moderate the relationship between demographic ageing and labour immigration policies. He finds that while social democratic welfare states opt for the facilitated expansion of female employment, liberal and conservative welfare states feature elements of demography-induced liberalisation of labour immigration. Finally, Manfred G. Schmidt compares public policies of democracies and autocracies. He shows that performance levels vary within both democracies and autocracies. Most notably, 7

differences between “defective” democracies and advanced constitutional democracies explain a significant part of the “democracy advantage”. Throughout his long and distinguished academic career, Klaus Armingeon has regularly explored the political repercussions of socio-economic changes, for instance in his work on political behaviour and more recently political alienation and the crisis of democracy (e.g. Armingeon et al. 2016b; Armingeon and Schädel 2015). This book’s fourth section focuses on democracy in hard times. Uwe Wagschal examines the degree of polarization of party systems and observes an increase of polarization in recent years, which is particularly accentuated in countries with proportional electoral systems. Sarah Engler and David Weisstanner explore the effect of changes in the relative deterioration in material conditions on support for radical right parties. They show that increasing income inequality – both directly as well indirectly through status decline – substantially increases support for the radical right. Departing from his joint work with Klaus Armingeon on the political attitudes of trade union members (Armingeon and Schmitt 1986), Rüdiger Schmitt-Beck finds no effect of trade union membership on the support for the recently established right-wing populist Alternative für Deutschland. Yet, he finds several indirect effects, which show that trade union membership is no significant barrier to the adoption of right-wing populist attitudes. European integration processes raise particular challenges for democracy. Hans Keman explores indirect democracy in European Union member states. He discusses the emergence of “new” political parties and their effects on voter participation, electoral volatility and government survival. Focusing on the domestic politics of European integration, Pascal Sciarini analyses the dilemma Swiss citizens face between the reintroduction of some form of immigration control and the continuation of the bilateral agreements with the European Union. He finds that age, trust in government and left-right orientation are key determinants of the choice between bilateral treaties and immigration control. Recent decades have witnessed the far-reaching liberalization of several major policy fields. Did these processes have negative effects on citizens’ satisfaction with democracy? Kai Guthmann and Anna Fill examine this relationship, yet they find only limited support for the expectation that liberalization processes have undermined democratic satisfaction. Anja Heidelberger, Marlène Gerber and Marc Bühlmann analyse the political behaviour of citizens who are dissatisfied with democracy. They observe that a majority – despite their dissatisfaction – participate in elections, although participation varies with party identification. In addition, they do not find a strong relationship between dissatisfaction with democracy and protest voting. Finally, Reimut Zohlnhöfer and Frank Bandau explore the determinants of programmatic change in West European political parties. They observe that office-seeking concerns drive programmatic change rather than lack of policy success, because parties consider a lack of success at elections to be more directly related to their programmatic position. In recent years, Klaus Armingeon’s many research interests have culminated in his critical work on the Euro crisis (e.g. Armingeon et al. 2016b; Armingeon and Cranmer 2017). This 8

book’s last section therefore focuses on the European Union and the Eurozone. Martin Höpner examines the politics of internal devaluation within the Eurozone. Discussing the political limits of, and resistance to, internal devaluation, he argues in favour of a flexibilisation of the Eurozone in order to restore democratic autonomy. Andreas Busch discusses the causes and effects of the German account surplus, which reflects some of the imbalances of the Eurozone. He offers several remedies to ease the account surplus and explains why they should be considered even if one does not consider the account surplus to be an economic problem. While Germany was thriving, other countries were in need of a bailout. How did these bailouts and the accompanying restrictions of democratic autonomy affect satisfaction with democracy? Frank Schimmelfennig and Dominik Schraff observe a strong negative effect of bailouts on democratic satisfaction, but this effect diminishes substantially after a number of years. More pessimistically, Yannis Papadopoulos observes a dramatic decline in the quality and legitimacy of democracy, both in the countries of the Eurozone periphery as well as at the level of the European Union. He argues that much of the improvements in the alleviation of the democratic deficit in the European Union have been reversed during the management of the Eurozone crisis. Maurizio Ferrara calls for the enhancement of the social dimension of EU citizenship. Concerned about an increasing cleavage between mobile workers and “stay at home” people, he argues that there is greater potential for solidarity within public opinions than meets the eye. Finally, Marlene Kammerer, Fadri Crameri and Karin Ingold examine the effect of the European Union on German and Swiss climate politics. They show that developments at the EU level are important points of reference in both countries, although domestic political agents engage with the EU in different ways. Be they optimistic or pessimistic, empirical examinations or theoretical reflections, the chapters of this volume illustrate that the European Social Model is far from finished: it has challenges ahead, lessons to teach and hope to give. And all the more, it requires continuous critical assessment, not least from its own citizens and from the scholarly community. Online appendix: The volume has an online appendix, to which several of the contributions make direct reference. The appendix can be accessed on the webpage of the volume, hosted by Springer.

References Armingeon, Klaus. 1983. Neo-korporatistische Einkommenspolitik. Eine vergleichende Untersuchung von Einkommenspolitiken in westeuropäischen Ländern in den 70er Jahren. Frankfurt a.M.: Haag und Herchen. Armingeon, Klaus. 1986. “Formation and Stability of Neo-Corporatist Incomes Policies.” European Sociological Review 2 (2): 138-147. Armingeon, Klaus, and Rüdiger Schmitt. 1986. “Wie ‘friedensbewegt’ sind die Gewerkschafter? Einstellungen von Gewerkschaftsmitgliedern zu Sicherheitspolitik und Friedensbewegung.” Politische Vierteljahresschrift 27 (4): 423-433. Armingeon, Klaus. 1987. “Gewerkschaften in der Bundesrepublik 1950-85: Mitglieder, Organisation und Außenbeziehungen.” Politische Vierteljahresschrift 28 (1): 7-34. 9

Armingeon, Klaus. 1988. “Zwei Ansätze der Gewerkschaftsforschung. Antwort auf Hans-Hermann Hertle: Probleme der Gewerkschaftsforschung.” Politische Vierteljahresschrift 29 (2): 288-295. Armingeon, Klaus. 1989a. “Trade Unions under Changing Conditions: The West German Experience 1950-1985.” European Sociological Review 5 (1): 1-23. Armingeon, Klaus. 1989b. “Arbeitsbeziehungen und Gewerkschaftsentwicklung in den achtziger Jahren: Ein Vergleich der OECD-Länder.” Politische Vierteljahresschrift 30 (4): 603-628. Armingeon, Klaus. 1991. “Towards a European System of Labour Relations?” The Journal of Public Policy 11 (4): 399-413. Armingeon, Klaus. 1993. “Institutionelle Antworten auf wirtschaftliche Verwundbarkeit. Ein internationaler Vergleich der prozeduralen Regulierung der Arbeitsbeziehungen in kapitalistischen Industriegesellschaften.” Politische Vierteljahresschrift 34 (3): 436-454. Armingeon, Klaus. 1994a. Staat und Arbeitsbeziehungen. Ein internationaler Vergleich. Opladen: Westdeutscher Verlag. Armingeon, Klaus. 1994b. “Die Regulierung der kollektiven Arbeitsbeziehungen in der Europäischen Union.” In Staat und Verbände [PVS Sonderheft 25], edited by Wolfgang Streeck, 207-222. Opladen: Westdeutscher Verlag. Armingeon, Klaus. 1994c. “Gründe und Folgen geringer Wahlbeteiligung.” Kölner Zeitschrift für Soziologie und Sozialpsychologie 46 (1): 43-64. Armingeon, Klaus. 1995a. “Der Schweizer Rechtsextremismus im internationalen Vergleich.” Revue Suisse de Science Politique 1 (4): 41-64. Armingeon, Klaus. 1995b. “Comment: The Democratic Deficit of the European Union.” Aussenwirtschaft 50: 67-76. Armingeon, Klaus, and Pascal Sciarini, eds. 1996. Die politischen Systeme Deutschlands, Österreich und der Schweiz im Vergleich. Zürich: Seismo (Sonderheft der Schweizerischen Zeitschrift für Politikwissenschaft). Armingeon, Klaus. 1997a. “Karrierewege der Professoren und Professorinnen der Politikwissenschaft in der Schweiz, Österreich und Deutschland.” Revue Suisse de Science Politique 3 (2): 103-117. Armingeon, Klaus. 1997b. “Swiss Corporatism in Comparative Perspective.” West European Politics 20 (4): 164179. Armingeon, Klaus, and Markus Freitag. 1997. Deutschland, Österreich und die Schweiz. Die politischen Systeme im Vergleich. Opladen: Leske + Budrich. Armingeon, Klaus. 1998a. “Interregionale und internationale Unterschiede der Wahlentscheidung. Das Wahlverhalten in Schweizer Kantonen im westeuropäischen Vergleich.” Politische Vierteljahresschrift 39 (2): 282-300. Armingeon, Klaus. 1998b. “The Impact of Globalisation on Swiss Policy Making: A Comment.” Revue Suisse de Science Politique 4 (2): 104-111. Armingeon, Klaus. 1999. “Die Stabilität der eidgenössischen Regierungskoalition in vergleichender Perspektive.” Österreichische Zeitschrift für Politikwissenschaft 28 (4): 463-473. Armingeon, Klaus. 2000. “Ökonomische Erklärungen des Verhaltens bei aussenpolitischen Abstimmungen.” Schweizerische Zeitschrift für Volkswirtschaft und Statistik 136 (2): 207-214. Armingeon, Klaus. 2001a. “Institutionalising the Swiss Welfare State.” West European Politics 24 (2): 145-168. Armingeon, Klaus. 2001b. “Sozialer Wandel und politische Stabilität. Politische Einstellungen von Arbeitnehmern in der Schweiz.” Revue Suisse de Science Politique 7 (4): 65-86. Armingeon, Klaus. 2002a. “Consociationalism and Economic Performance in Switzerland 1968-1998: The Conditions of Muddling through Successfully.” In Consociationalism and Corporatism in Western Europe: Still the Politics of Accommodation?, edited by Jürg Steiner and Thomas Ertman, 121-138. Amsterdam: Boom. Armingeon, Klaus. 2002b. “The Effects of Negotiation Democracy: A Comparative Analysis.” European Journal of Political Research 41 (1): 81-105. Armingeon, Klaus. 2004a. “Institutional Change in OECD Democracies, 1970-2000.” Comparative European Politics 2 (2): 212-238. Armingeon, Klaus. 2004b. “Major Changes without Regime Shifts: Switzerland and Japan in Comparative Perspective.” Swiss Political Science Review 10 (3): 5-30. Armingeon, Klaus, and Kenji Hirashima, eds. 2004. Regime Shifts: Switzerland and Japan Compared. Chur: Rüegger (special issue of the Swiss Political Science Review). Armingeon, Klaus, Fabio Bertozzi, and Giuliano Bonoli. 2004. “Swiss Worlds of Welfare.” West European Politics 27 (1): 20-44. Armingeon, Klaus, and Michelle Beyeler, eds. 2004. The OECD and European Welfare States. Cheltenham: Edgar Elgar.

10

Armingeon, Klaus. 2005. “Die Ausbreitung der Aktiengesellschaft und der Wandel des Wohlfahrtsstaates und der Arbeitsbeziehungen. Ein internationaler Vergleich.” In Finanzmarkt-Kapitalismus, edited by Paul Windolf, 441-459. Wiesbaden: VS Verlag. Armingeon, Klaus, and Giuliano Bonoli, eds. 2006. The Politics of Post-Industrial Welfare States. Adapting Postwar Social Policies to New Social Risks. London: Routledge. Armingeon, Klaus, and Patrick Emmenegger. 2007. “Wirtschaftspolitik in der Schweiz: Erosion eines Modells.” In Wirtschaft Schweiz – Ein Sonderfall?, edited by Michael Nollert and Hanno Scholtz, 189-227. Zürich: Seismo. Armingeon, Klaus. 2007. “Active Labour Market Policy, International Organizations and Domestic Politics.” Journal of European Public Policy 14 (6): 905-932. Armingeon, Klaus, and Nathalie Giger. 2008. “Conditional Punishment. A Comparative Analysis of the Electoral Consequences of Welfare State Retrenchment in OECD Nations, 1980-2003.” West European Politics 31 (3): 558-580. Armingeon, Klaus, and Karolina Milewicz. 2008. “Compensatory Constitutionalization. A Comparative Perspective.” Global Society 22 (2): 179-196. Armingeon, Klaus, and Romana Careja. 2008. “Institutional Change and Stability in Post-communist Countries, 1990-2002.” European Journal of Political Research 47 (4): 411-435. Armingeon, Klaus. 2011. “A Prematurely Announced Death?” In Switzerland in Europe: Continuity and Change in the Swiss Political Economy, edited by Christine Trampusch and André Mach, 165-185. London: Routledge. Armingeon, Klaus. 2012a. “The Politics of Fiscal Responses to the Economic Crisis, 2008-2009.” Governance 25 (4): 543-565. Armingeon, Klaus. 2012b. “The Impact of Political Parties on Labour Relations: European Post-communist Democracies in Comparative Perspective.” In Party Government in the New Europe, edited by Hans Keman and Ferdinand Müller-Rommel, 113-136. London: Routledge. Armingeon, Klaus, Kai Gutmann, and David Weisstanner. 2014. “Politische Voraussetzungen von Austeritätspolitik: Ein internationaler Vergleich von 17 etablierten Demokratien zwischen 1978 und 2009.” Zeitschrift für Staats- und Europawissenschaften 12 (2-3): 242-271. Armingeon, Klaus, and Kai Guthmann. 2014. “Democracy in Crisis? The Declining Support for National Democracy in European Countries 2007-2011.” European Journal of Political Research 53 (3): 423-442. Armingeon, Klaus, and Besir Ceka. 2014. “The Loss of Trust in the European Union during the Great Recession since 2007: The Role of Heuristics from the National Political System.” European Union Politics 15 (1): 82107. Armingeon, Klaus, Kai Guthmann, and David Weisstanner. 2015. “Wie der Euro Europa spaltet. Die Krise der gemeinsamen Währung und die Entfremdung von der Demokratie in der Europäischen Union.” Politische Vierteljahresschrift 56 (3): 506-531. Armingeon, Klaus, and Lisa Schädel. 2015. “Social Inequality in Political Participation: The Dark Sides of Individualisation.” West European Politics 38 (1): 1-27. Armingeon, Klaus, Kai Guthmann, and David Weisstanner. 2016a. “Choosing the Path of Austerity: How Parties and Policy Coalitions Influence Welfare State Retrenchment in Periods of Fiscal Consolidation.” West European Politics 39 (4): 628-647 Armingeon, Klaus, Kai Guthmann, and David Weisstanner. 2016b. “How the Euro Divides the Union: The Effect of Economic Adjustment on Support for Democracy in Europe.” Socio-Economic Review 14 (1): 1-26. Armingeon, Klaus, and Skyler Cranmer. 2017. “Position- Taking in the Euro Crisis.” Journal of European Public Policy 25 (4): 546-566. Weisstanner, David, and Armingeon, Klaus. 2018. “How Redistributive Policies Reduce Market Inequality: Education Premiums in 22 OECD Countries.” Socio-Economic Review. https://doi.org/10.1093/ser/mwy018.

11

Part I Transformation of Neo-Corporatism

Neo-corporatism and the responsiveness of democracy Marius R. Busemeyer1 Abstract: In this chapter, I want to discuss the relationship between neo-corporatism as a form of interest group representation and prevailing inequalities in democratic representation and responsiveness. In much of the existing literature on the role of public opinion and democratic policy-making (most of it from the US), interest group influence is portrayed as something that distracts policy-makers from implementing the will of the people, catering to ‘special interests’ instead. In contrast, the core normative foundation of corporatist decision-making is based on the idea that corporatist institutions are superior to interest group pluralism in the sense that corporatism ensures a proper representation of those interests that are difficult to organize and mobilize. In my contribution, I want to provide a critical theoretical discussion of the promises and challenges of corporatist decision-making in relation to the normative goal of ensuring a responsive and representative democracy, drawing on the work of Klaus Armingeon and others.

1

Introduction

In recent years, public and scholarly debates about persistent inequalities in democratic representation have intensified (Bartels 2008; Elsässer, Hense, and Schäfer 2017; Gilens 2012; Gilens and Page 2014; Hacker and Pierson 2010; Schäfer 2010). The rise of new populist and anti-establishment parties on the left and right is often attributed to real or perceived deficiencies in the responsiveness of policy-makers to the demands of the mass public as policymakers respond to the particularistic concerns of their specific electoral constituencies rather than the wishes of the general public (Gilens and Page 2014). Interest groups tend to be depicted as ‘special interest groups’, i.e. organizations that specialize in the “mobilization of bias” (Schattschneider 1960). Through the process of mobilization and organization, so the argument goes, those societal strata and groups, which are already privileged due to their socio-economic position, get an even larger advantage as they face less obstacles in collective action, compared to groups and individuals with fewer socio-economic resources. This ‘pluralist’ perspective on the democratic process of representation is gaining ground in the scholarly debate about the relationship between public opinion, interest groups and party politics. Partly, this may be related to the fact that much of the research on political inequalities in representation and responsiveness is on the case of the United States, which is characterized by a highly pluralist interest group system as well as a particular set of political institutions. A

1

Universität Konstanz [email protected]

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 R. Careja et al. (eds.), The European Social Model under Pressure, https://doi.org/10.1007/978-3-658-27043-8_2

15

second reason, however, may be that previous contributions to the scholarship on interest groups in the tradition of the work on neo-corporatism tend to get side-lined at a time when the object of study in this literature – namely corporatist institutions of interest mediation – comes under pressure even in those countries where it was believed to be firmly entrenched. The purpose of this chapter is to juxtapose the contemporary debate about inequality in representation and responsiveness of democracy with earlier work on neo-corporatism. This endeavor is motivated by the idea that neo-corporatist forms of interest mediation could be both a more legitimate as well as more effective response to the various economic and political crises of advanced contemporary democracies, compared to pluralist interest mediation. Klaus Armingeon has made important and influential contributions to the scholarship on neocorporatism, some of which I will revisit later in the chapter. The bottom line of the exercise undertaken in this chapter is to renew the call for a more complex and sophisticated understanding of the relationship between interest groups, party politics and the responsiveness of democratic policy-making, which, in my view, tends to get lost in current-day debates about these issues.

2

Interest groups, party politics and the responsiveness of democratic policy-making

As stated above, there is sizable and still growing literature on the link between public opinion and policy-making (Burstein 2003; Erikson, MacKuen, and Stimson 2002; Hakhverdian 2010; Hobolt and Klemmensen 2008; Jennings and John 2009; Page and Shapiro 1983; Stimson, MacKuen, and Erikson 1995). Much of this literature focuses on the empirical case of the US, although there is increasingly work on other, mostly European countries as well. One influential strand of this literature argues that by and large and in spite of potential biases which I discuss further below, policy-makers in democracies are indeed responsive to public demands, if only for the simple reason that they want to get re-elected. Stimson, MacKuen, and Erikson (1995) have coined the notion of “dynamic representation” to describe a continuous process of feedback and responsiveness between citizens and policy-makers unfolding over time (see also Soroka and Wlezien 2010 for a related approach): According to this argument, “public opinion moves meaningfully over time” (i.e. citizens perceive changing socio-economic environmental conditions and policy-makers’ responses to these changes), “government officials sense this movement, and […] alter their behavior in response to the sensed movement” (Stimson, MacKuen, and Erikson 1995, 543). Admittedly, the signals that public opinion sends to policymakers are rather vague and unspecific, resembling particular “moods” on a scale from left (liberal in the US sense) to right (conservative) (Stimson 1991), but nevertheless, they are meaningful and consequential signals. This rather positive assessment about the responsiveness of democratic policy-making is challenged by a number of studies that identify inherent biases in representation and 16

responsiveness. Gilens’ work and other contributions in this tradition (Gilens 2005, 2012; Bartels et al. 2005; Page, Bartels, and Seawright 2013; Schlozman et al. 2005) show that the association between policy output and individual-level preferences is more pronounced for those in the upper income strata than in the lower strata, i.e. policy-makers tend to implement the preferences of the rich rather than cater to those of the poor. Increasingly, scholars are transferring the Gilens framework to European countries (Elsässer, Hense, and Schäfer 2017; Rosset 2013; Rosset, Giger, and Bernauer 2013; Giger, Rosset, and Bernauer 2012; Bernauer, Giger, and Rosset 2015), obtaining similar results. Critics of this work have pointed out that the preferences of the rich and poor often move in parallel to each other and that differences in attitudes related to income are less pronounced than ideology-driven differences (Ura and Ellis 2008; Soroka and Wlezien 2008; Stimson 2011). Both factors would, in the eyes of the critics, mitigate the problem of non-responsiveness. Independent of the eventual outcome of this debate, the claim about a direct connection between public opinion and policy-making has an empirical as well as a normative dimension. From an empirical perspective, the goal is simply to ascertain whether the preferences of the mass public are a significant determinant of policy change or not. From a normative point of view, the implicit claim that is made here is that the link between mass preferences and policy responses should be as direct and unmitigated as possible. This normative dimension is voiced more or less explicitly in the literature. For instance, after describing the basic contours of their “thermostatic model of democracy” (which is in many aspects similar to Stimson, MacKuen, and Erikson’s “dynamic representation”), Soroka and Wlezien (2010, 3) state: “This, we argue, is how a functioning democracy should work.” In the literature on inherent biases in democratic representation and responsiveness, the obvious implication is that the varying responsiveness of policy-makers to the concerns of the rich and poor violate the fundamental democratic norm of ‘one man/woman, one vote’. Why is the blending of empirical and normative dimensions regarding the linkage between public opinion and policy-making problematic? The core problem here is that this blending of perspectives and dimensions increases the danger of reverting to a rather simplistic and, in the end, pluralistic and majoritarian understanding of democracy. Going down this road in a way signals a return to the debates of the 1970s and 1980s, when scholars such as Lehmbruch (1979), Lijphart (1999), Schmitter (1979, 1985) as well as Armingeon (1986, 2002) had significantly broadened and expanded the empirical and normative perspectives of scholarship on democratic political institutions by highlighting the peculiarities of European democracies and thereby the variety of institutionalized forms of democratic representation different from the Anglo-Saxon version of majoritarian democracy. Given that much of the contemporary empirical research about the responsiveness of policymaking to public opinion studies the case of the US, there is again a certain tendency to conflate particular attributes of this case with generally desirable attributes of democratic decisionmaking in other countries. For instance, collective actors such as interest groups and political 17

parties that in other contexts might play a crucially important role in mitigating potential biases in representation and responsiveness are portrayed as negative influence as they potentially ‘distract’ policy-makers from implementing the ‘true will’ of the people. Burstein (2003), for example, describes a potential conflict between public opinion and party politics in the following way: “political parties may, when in office, enact policies favored by their most ardent supporters rather than the general public” (Burstein 2003, 30). In a context of coalition governments and multiparty-systems, however, the role of government parties is exactly to represent the interest of their respective constituencies rather than the general public. For instance, left-wing parties in European-style coalition governments could be more effective in representing the interests of low-income voters and labor market outsiders compared to the diffuse appeal of US parties to the median voters in particular electoral districts. In other words: The representation of the interests of the poor may improve rather than deteriorate, when left-wing parties in coalition governments try to implement the policies “favored by their most ardent supporters” rather than the median voter. And in fact, previous research (Huber and Powell 1994) has already found that the degree of congruence between citizen preferences and government positions is higher in the case of proportional representation (PR) systems compared to majoritarian systems. In a similar vein, interest groups are mostly regarded as organized manifestations of inherent biases in democratic representation to the benefit of the upper income classes (Gilens and Page 2014). Even though there are organized interest groups representing workers or general public interests, research finds that business groups representing the economic elites are more influential and powerful in setting the agenda (Baumgartner et al. 2009, 255-256; Gilens and Page 2014, 571-572). Hacker and Pierson (2010) come to a similar finding, using a different methodological approach. Of course, business groups might have significant advantages when it comes to the mobilization and organization of interests (as already explored by Offe and Wiesenthal (1980)). However, as is well-known from comparative research, the relative balance of power between unions and employers varies significantly across countries. Furthermore, as the recent work by Martin and Swank (2008, 2012) has shown, not all employers’ associations are alike. In a neo-corporatist framework, (formerly) pluralist interest groups are transformed into intermediary organizations or associations, which do not simply represent the interests of their members vis-à-vis the government, but also mediate between the public concerns expressed by governments and the particular concerns of their membership base (Streeck and Kenworthy 2005). Furthermore, when business interests are embedded in a network of crosssectoral and cross-class associations and governance frameworks, then employers’ associations are more likely to take collective concerns into account in their decision-making. The same holds true for unions: In decentralized and liberal settings, unions may pursue particularistic goals and lobby for higher wages and employment conditions for their members only, whereas in corporatist frameworks, unions are more likely to adopt a more holistic perspective (Armingeon 1986). 18

In sum, there is a large variety of interest group types, as well as, more importantly, institutional contexts, in which these associations are embedded and interact with each other. This is neglected in the literature cited at the beginning of this section. Therefore, in the following section, I would like to further explore the potential contribution of scholarship on corporatism and associated theories about political institutions (consensual and consociational democracy) to the debate about the link between public opinion, interest groups and party politics.

3

Neo-corporatism and the responsiveness of democratic policy-making

In this section, I would like to start off by re-visiting two influential contributions by Klaus Armingeon on the topic of neo-corporatism, before I broaden the focus of the debate to include further scholarship on this issue. In an early contribution, Klaus Armingeon explores the conditions for the formation and long-term stability of neo-corporatist incomes policies in a number of European countries (Armingeon 1986). Even though incomes policies – a rather direct instrument to intervene in labor markets – are no longer used in today’s liberalized and globalized economies, this article makes an important point: Rejecting simplistic and rash normative assessments about the role of organized interests in policy-making, this chapter develops a sophisticated and differentiated argument about the conditions under which organized interest groups contribute to joint problem-solving rather than maximizing their particularistic benefits. Thus, instead of depicting interest groups solely as ‘special interests’, which are only focused on maximizing benefits for their particular clienteles, Armingeon (1986, 139-140) analyzes the empirical variation in organizational attributes of interest groups, in his case mostly trade unions, as well as the institutional frameworks, in which they are embedded. A common point of reference on this issue is Olson’s (1965, 1982) work on the theory of collective action. As is well-known, Olson argues that the mobilization of individual actors for the purpose of collective action is heavily constrained by free-rider problems, which can be better solved by small, coherent groups, leading to the dominance of ‘special interest’ organizations over the organization of diffusive, collective concerns. The exception to this rule are ‘encompassing organizations’. These are organizations, which – for idiosyncratic reasons – already encompass large shares of the population in a particular country. If organizations are sufficiently encompassing, they develop an enlightened self-interest in maximizing collective concerns rather than pursuing particularistic short-term goals in a self-interested manner. Applied to the field of labor relations, this argument has been helpful in explaining why powerful trade unions – covering a large share of the working population – have not been particularly bent on maximizing short-term wage increases for their members, but rather accepted a strategy of wage restraint in crisis periods in order to maximize the long-term growth potential of the economy (see also Scharpf 1991). Armingeon (1986) takes issue with this argument. By studying the variation of income policies across a range of OECD countries, he finds that union organization is an important, but 19

by no means a necessary precondition for neo-corporatist policy formulation. This is an interesting point, because it proves that neo-corporatist forms of interest mediation are ultimately not dependent on a particular type of union organization, which can only be found in a very limited number of country contexts (i.e. the Scandinavian countries and, potentially, Austria). Rather, Armingeon (1986, 142) identifies a number of conditions related to the role of governments and employers’ associations in committing to and promoting neo-corporatist forms of interest mediation: Simplifying and generalizing somewhat, one of these conditions is that unions and employers’ association must share a basic and joint understanding about the goals and instruments to be used to address a particular policy problem (even though potentially disagreeing about the settings of these instruments). Furthermore, government actors must be willing to “realize those public policies agreed upon at the neo-corporatist level” (Armingeon 1986, 142). Thus, governmental and other state actors must be willing to delegate some of their decision-making authority to the “private interest governments” (Streeck and Schmitter 1985) of business and labor, accepting that private interest associations to some extent exploit this act of delegation to pursue their own particular goals, while at the same time – as an extension of the reach of government – mobilizing private actors to pursue public responsibilities. Also, employers and their associational representatives have to accept or at least consent to sharing a seat with unions at the bargaining table. In sum, by identifying these conditions, Armingeon (1986) shows that neo-corporatist forms of interest mediation are not self-sustaining institutional equilibria by themselves, but depend on a range of supporting behaviors and dispositions among the actors involved. However, and this is the more important take-away from this chapter, neo-corporatism is a matter of political choice. Even though centralized structures both on the union as well as the employer side certainly help, they are not a necessary precondition for neo-corporatist forms of interest mediation and policy formulation. And for sure, there are a number of examples in later years – in particular the social pacts of the 1990s – which confirm Armingeon’s intuition. Furthermore, if the involvement of societal actors is a matter of political choice, it is likely to be influenced by the particular ideologies of governmental actors. In line with this, Van Kersbergen (1995) has shown that Christian democratic parties subscribe to the “politics of mediation” (ibid, 28). This notion captures the idea that societal conflicts about values and/or the distribution of politico-economic resources should not be solved by majoritarian forms of decision-making, but by incorporating different political viewpoints and interests into an emerging political and societal consensus regarding the solution of a particular problem. Regarding the latter issue, there are obvious parallels to the debate about different types of democracies (Lijphart 1999). In a later contribution, Armingeon (2002) contributes to this debate by analyzing the “effects of negotiation democracy”. Here, Armingeon again rejects simplification by arguing that the term “negotiation democracy” is actually related to three distinct forms: (i) consociational democracy (i.e. institutionalized forms of cooperation between organizational representatives of diverse socio-cultural groups within the political 20

administration system); (ii) corporatism (i.e. institutionalized forms of coordination and negotiation between state actors, unions and employers); and (iii) strong systems of institutional veto players, who may thwart governments’ swift decision-making by exercising their constitutionally or otherwise enshrined veto rights. He then develops a differentiated set of empirical expectations regarding the associations between these different forms of negotiation democracy and various socio-economic and political outcomes such as political satisfaction, participation and trust, the integration of minorities, the size of the welfare state, inflation, economic growth and public debt. One important finding – for the purpose of this chapter – is that consociational democracies are superior compared to other types of democracy in social integration (ibid, 82), i.e. the integration and representation of “large minorities” as long as those minorities are sufficiently large in order to be included in the institutionalized forms of interest mediation in these democracies. In consociational democracies – for a long time, real-world examples had been the Netherlands and Switzerland – the elites representing different socio-cultural groups cooperate and regularly interact with each other in democratic decision-making in order to ensure that the interests of the different societal fractions are taken into account (Ibid, 83). A second important finding in Armingeon’s contribution is that stable corporatism is associated with better socio-economic outcomes, in particular lower levels of “industrial unrest” (Ibid, 84), i.e. strikes, as well as a more generous welfare state. Third, a high number of institutional vetoplayers is associated with a smaller public sector. What is the relevance of these findings for the topic of this chapter? Simply put, Armingeon’s work, as well as other classic contributions on neo-corporatism such as Lehmbruch (1979), Schmitter (1979, 1985) and Streeck and Schmitter (1985), has implications for understanding the role of interest groups and – eventually – government policy in addressing gaps in the representation of weakly organized sections of society. In the US-inspired literature on the link between public opinion and policy-making introduced above, interest groups are usually portrayed as actors deflecting policy-makers from pursuing the demands of the people (Gilens and Page 2014). At best, some types of interest groups – unions and civil rights groups – can boost the voice of low-income citizens, at least partly mitigating biases in representation and responsiveness (Baumgartner et al. 2009). However, given the dire position of unions in the United States and the central importance of monetary contributions in elections in this country, the success of these efforts is likely to be limited. Most importantly, the dominant perspective in the US literature remains wedded to a pluralist conception of the connection between interest groups and decision-making. From this perspective, interest groups vie for influence by marshalling economic and political resources, which can be used to finance electoral and other campaigns. In the ideal-typical system of pluralist interest mediation, the role of the state is limited: Explicit interventions from state actors directed at influencing the balance of power between interest groups are regarded as illegitimate and probably ineffective to boot. Direct forms of state intervention are deemed 21

illegitimate because they infringe on the basic right of freedom of speech, if the state tips the scales of political debates in favor of one particular type of interest versus another. These kinds of ‘partisan’ interventions – in the worst case – signal the ‘capture’ of state institutions by particularistic interests that abuse and exploit the powers of state institutions to promote their own agendas. From a pluralist perspective, the involvement of interest groups in agenda-setting and decision-making is associated with an increase in political inequalities and a decline in responsiveness to the concerns of the less well-off. This is because it is easier for interest groups representing the interests of the economic elites to accumulate power resources – money – to invest in political campaigns (Gilens and Page 2014). Interest groups representing non-elites or diffuse public interests have to compensate for a lack of economic resources by mobilizing a larger number of supporters, encountering the well-known challenges of collective action as identified by Olson (1965, 1982), even though they may attract the occasional big donor for ideological reasons. Furthermore, trade unions in particular have enjoyed non-pecuniary privileges supporting mobilization efforts, such as the automatic deduction of union membership fees from workers’ paychecks or ‘closed shop’ rules for collective wage bargaining, although these rules are eroding quickly in the case of the United States (Anzia and Moe 2016).

4

The potential contribution of corporatist institutions and practices to mitigating political inequalities

In the following section, I reflect on the normative underpinnings of corporatism and its implication for democratic responsiveness. Given the prominence of pluralist perspectives in the contemporary literature on public opinion and responsiveness, the potential contribution of neo-corporatist institutions in its various forms as discussed above to mitigating problems of political inequalities is at risk of getting lost. This risk is aggravated by the fact that even in those countries, where corporatist institutions and practices are well-established, they are coming increasingly under pressure to either reform (and liberalize) or wither away and become irrelevant. To some extent, these developments are likely driven by the broad process of individualization, which fuels more critical attitudes towards any forms of collectively organized interest representation in political parties and unions (Armingeon and Schädel 2015). Nevertheless and in spite of these developments, corporatist institutions and practices still hold the potential of mitigating some forms of political inequality under certain conditions as I will discuss in greater detail in the following paragraphs. By contributing both to social inclusion as well as economic development, neo-corporatism holds the potential of mitigating economic crises as well as democratic crises of legitimacy. First, and of central importance to the notion of neo-corporatism, is the idea that corporatist institutions and practices can balance out inherent imbalances in power resources between the 22

associational representatives of different sides of a given societal cleavage or conflict line. Most commonly, this conflict line is the class cleavage between business and labor, but other cleavages are relevant as well, e.g. latent conflicts between farmers and environmental protectionists or between health insurances, doctors and patients. The normative reference point for corporatism is the notion that all sides in a latent conflict have to be heard and given consideration. Rooted in and connected to very different “visions of liberal democracy” (Huber and Powell 1994), corporatism aims at achieving consensual policy choices supported by large parts of the population, whereas pluralism is associated with majoritarian decision-making. Furthermore, the corporatist understanding of democratic representation acknowledges more forcefully the fact that individual citizens participate in the political decision-making process as individual citizens (mostly via elections) and as members of collectives and social groups, which – as groups – deserve to be recognized and included into policy decisions. Applied to the issues of representation and responsiveness, corporatist institutions and practices can ensure that the interests of the ‘weaker’ side in a latent conflict are better represented, when policy decisions are contingent on the agreement between the organizational representatives of all concerned, i.e. traditionally the state, employers and unions. Since workers’ interests have commonly been regarded as the weaker side in the capitalist class conflict (Offe and Wiesenthal 1980), the establishment of corporatist institutions and practices has often been associated with powerful unions and social democratic parties (Rothstein 1987), although – as mentioned above – corporatism has also been promoted by Christian democratic parties (Van Kersbergen 1995) and – partly – employers and local representatives of workers themselves (Martin and Swank 2012). Independent of its historical origins, corporatism holds the potential of boosting the representation of the political and economic interests of societal groups and strata with weak organizational resources. This can – to some extent – compensate for the erosion of power resources and declining levels of political participation among the less well-off. However, it has to be emphasized that the balance of power prevailing at certain moments of historical development certainly shapes the responsiveness of corporatist institutions themselves, as Rothstein (1987) has argued for in the case of Scandinavian macrolevel corporatism. Second, corporatist institutions and practices can also help to develop organizational capacities of societal groups. Typically, the right to participate and to be recognized in corporatist decision-making arenas is made conditional on a number of attributes of the respective associations. More specifically, associations claiming to represent certain societal interests have to demonstrate that they are a legitimate and recognized representative of these interests. In exchange, the state supports the formation of functionalist representation monopolies by privileging certain interest groups over others. This can be done in a direct manner, in the extreme, non-democratic version of corporatism through the outright prohibition of interest groups. It can, however, also be done more indirectly by shaping the incentives for individual citizens or businesses to affiliate with one interest group rather than another. 23

The latter supports the development of organizational capacities of those interest groups that participate in corporatist institutions and practices since potential members have a strong incentive to become affiliated in order to influence decision-making processes more directly than they could as individual citizens. Furthermore, once “incorporated” into corporatist institutions and practices, private interest groups cease to be fully “private” lobbying groups, but turn into “private interest governments” (Streeck and Schmitter 1985), in particular when corporatist practices involve the delegation of quasi-public responsibilities to these associations. In this case, private interest groups are pressed – informally through politics and occasionally even via formal means – to look beyond particularistic needs and to take into account public concerns. Taken together, corporatism’s contribution to boosting organizational capacities is likely to provide further support to those societal and political interests, which are inherently more difficult to mobilize and organize, i.e. diffuse public interests. Third, and related to the previous point, corporatism can be a useful lever to promote organizational democracy, if the inclusion of interest groups in corporatist institutions and practices is made conditional on them establishing procedures of representative democracy for internal decision-making processes. Pluralist settings are more tolerant of a multitude of organizational forms: from social movements via formal organizations with clearly defined representational hierarchies to professionalized advocacy groups with little involvement of the membership beyond fiscal contributions. The legitimacy of corporatist decision-making, in contrast, more strongly depends on participating groups and associations subscribing to and complying with the normative standards of representative democracy. Vice versa, the outlook of participating directly in governmental decision-making can trigger, or at least consolidate, democratization processes within interest groups. Finally, fourth, a broader contribution of corporatism to mitigating political inequalities is that it tends to produce economic and social policies, which are more effective in mitigating socio-economic inequality, by itself associated with political inequalities (Anderson and Beramendi 2008; Solt 2008; Schäfer 2010). As shown by Armingeon (2002), corporatism is associated with a more generous welfare state. Other research (Hicks and Kenworthy 1998; Wallerstein 1999) has furthermore found that extensive collective wage bargaining is linked to lower levels of wage inequality. To the extent that corporatism is conceived as the “social democratic institutionalization of the class conflict” (Rothstein 1987), this may not be surprising: In countries, where the left is powerful, unions and social democratic parties establish corporatist institutions and practices to tame business interests. Thus, the ultimate driving force could still be differences in power resources and not corporatism as such. This argument holds some value, but neglects the fact that corporatist institutions and practices are much more wide-spread and not confined to the Scandinavian countries. Furthermore, once established, corporatism as a set of institutions develops a path dependency on its own, contributing to its long-term stability, but by no means ensuring its long-term survival. 24

5

The dangers of aggravating political inequalities and nonresponsiveness

In spite of its potential to mitigate political inequalities by improving the institutional representation of ‘weak’ segments of society discussed in the previous section, neo-corporatist institutions and practices also suffer from a number of weaknesses regarding representational inequalities and have been – rightly – criticized for those shortcomings. In the end, similar to Armingeon’s (1986) early contribution to the literature, the challenge is to identify conditions that would need to be met in order for neo-corporatism to contribute to mitigating political inequalities. In the following section, I will discuss corporatism’s weak spots, whereas the concluding section of this chapter summarizes and enumerates the aforementioned conditions. A first weak spot is its potential to contribute to the formation of new elites and ‘oligarchies’ within interest group associations. According to Michels’ “iron law of oligarchy”, there might be an inherent tendency in large-scale mass-based political organizations to establish oligarchies and hierarchies between political insiders and outsiders. But even independent of that, the danger of establishing oligarchical structures could be more pronounced in corporatist systems compared to pluralist settings. This is for several reasons discussed hereafter. First, by being directly ‘incorporated’ into joint decision-making processes with state actors and representatives of other associations, the political leadership of interest groups on the one hand receives political and even pecuniary privileges, which the membership base does not. On the other hand, the associational leadership has to defend and implement the joint decisions achieved in corporatist decision-making arenas, nurturing temptations to implement a hierarchical, centralized decision-making structure to minimize the potential of protest at the base (Armingeon 1986, 139). Obviously, there is a trade-off here: Improving the representational clout of low-skilled workers in political decision-making via the involvement of unions in corporatist institutions and practices might prove to be ineffective if the union leadership becomes detached from the interests and concerns of its membership base. Second, another weak spot concerns the scope of corporatist decision-making. As discussed above, the core idea of corporatist decision-making is to include representatives of all major factions in society, in order to achieve a politically and economically feasible consensus regarding policy responses to a particular political problem. A fundamental problem in this respect is that this consensus is rarely, if ever, truly comprehensive. One issue is representational inequality within associations discussed in the previous paragraph. Another issue relates to the scope of interests included in the ‘corporatist compromise’. In the heyday of industrial neocorporatism in the 1970s and 1980s, the major actors in corporatist decision-making arenas were employers’ associations and unions representing the core sectors of the economy. Thus, the scope of the ‘corporatist coalition’ was indeed encompassing. This might have changed in recent years, however: The structural transformation of economies to service-oriented knowledge economies has fueled the expansion of new types of employment in the service 25

sector. In some countries with well-established corporatist institutions, new associations representing these new sectors have come up, but are struggling to assert themselves against the dominant position of the established actors. In other (probably most) countries without a strong tradition of corporatism, the establishment of associations in emerging sectors of the economy has been even more difficult if not impossible. This holds true in particular for associations representing the interests of workers, i.e. unions. In country contexts, where unions nowadays only represent workers in the sheltered public sector or declining sectors of the economy, they are often regarded as defenders of the interests of the remaining labor market insiders to the detriment of outsiders (Rueda 2007; Emmenegger et al. 2012). Giving insider-oriented unions even more power by establishing privileged access channels via corporatist decision-making arenas could then severely aggravate political inequalities and problems of non-responsiveness. The question, however, is whether the response to this challenge is to completely dismantle any forms of collective representation of worker interests with the justification of “leveling the playing field” (on a low level) via liberalization or whether corporatist institutions and practices can still be used to improve the representation of outsider interests. A third weakness of the model of neo-corporatism is its often more implicit rather than explicit expectation that all actors involved in the process share at least some basic agreement about problem definition and potential policy responses (see also Armingeon 1986, 142). Of course, actors representing different constituencies with different socio-economic interests are likely to disagree on a range of issues, such as the adequate level of wages, social benefits and employment guarantees. Still, a basic precondition for corporatist institutions and practices to work is that one side accepts the other side as legitimate and largely reasonable negotiation partner. In fact, one might even argue that if this basic premise no longer holds, the resulting pattern of interest group interaction would no longer qualify as corporatism, but simply revert to pluralist bargaining. In times of increasing political polarization across OECD countries, this implicit assumption may no longer hold. Furthermore, a related implicit assumption for corporatist institutions and practices is that the corporatist partners involved in the process refrain from overtly tactical ‘warfare’, exploiting their hold-up power in corporatist arenas excessively in order to maximize short-term gains. The viability of corporatism depends on the mutual recognition among the involved actors that they are engaging in a long-term game and not in one-shot short-term bargaining. The fourth weakness of corporatist decision-making is related to the recognition of a longterm engagement, namely the fact that decision-making is likely to be much slower than in majoritarian, pluralist settings because of the requirement to achieve broad consensus before proceeding. Traditionally, research in public policy has summarized the differences between these institutional settings by stating that corporatist decision-making takes time, but produces policy solutions that are more stable and politically sustainable in the long run because all major stakeholders are involved. In majoritarian, pluralist contexts, in contrast, public policy-making 26

proceeds more in a stop-and-go manner with significant policy change in one direction, followed by significant reversals of direction in the next legislative period (Armingeon 2002, 83). Given the changing socio-economic context with an ever increasing pace of globalization and technological change, this particular weakness of corporatist decision-making might become more severe over time, however. For instance, the expansion of precarious forms of employment across European countries, which had already started before the crisis years (Eichhorst and Marx 2012), has further intensified in many countries since then. Corporatist institutions and practices might not be sufficiently reactive (or nurture particular ‘blind spots’, see previous point from above) to develop policy responses to this challenge, potentially foregoing some of its cherished contribution to promoting inclusive social and economic policies. Finally, and maybe most importantly, the fifth weak spot to be discussed here is the fact that the potential of corporatist institutions and practices to mitigate political inequalities and to enhance the responsiveness of policy-making to weak segments of society hinges on the political willingness of governmental actors to support these institutions. Again, there are several issues to be talked about. For one, governmental actors need to be willing to delegate significant powers in decision-making and potentially also the implementation of policies to non-state actors, i.e. private interest associations. Not surprisingly, state actors do not necessarily do this voluntarily as evidenced by the fact that corporatist institutions have often evolved historically in a context of weak statehood. This can, for instance, be seen in Katzenstein’s (1987) classical work on the “semi-sovereign” state of Germany. Furthermore, besides organizational capacities, a willingness to make use of corporatist institutions and practices depends on the particular ideologies of government parties. As shortly discussed above, social democratic and Christian democratic parties have historically been most supportive of neo-corporatism. Both party families, however, face daunting political challenges in contemporary European politics, where they had been most established. Social democratic parties are in a continuous state of crisis and decline, which has now even reached the strongholds of the Scandinavian countries. Whether the emergence of right-wing populist parties that increasingly cater to the traditional electoral constituencies of social democratic parties is a causal driving force of this development, or merely an accompanying symptom of it, cannot be discussed here in full. In any case, the increase of right-wing populism also puts pressure on Christian democratic parties. Hence, the political supporting coalition for corporatist institutions and practices is shrinking. Right-wing populist parties have started to come out as staunch critics of corporatism as can be seen in the recent conflict between the Austrian Christian Democratic Party (ÖVP) and the right-wing FPÖ.2 Even though Austria is one of the countries in which neo-corporatist institutions are well-entrenched and have 2

https://www.sn.at/politik/innenpolitik/koalitionsverhandlungen-oevp-und-fpoe-einigen-sich-bei-kammern21637711 (accessed August 8, 2018)

27

contributed significantly to economic prosperity and social inclusion, the ÖVP could barely fight off calls from the FPÖ during the recent coalition negotiations to dismantle the principle of compulsory membership in the Austria’s Chambers of Commerce – an essential pillar of the corporatist system.

6

Conclusions

To sum up, the goal of this chapter was to reflect on the potential contribution of classic scholarship on neo-corporatism – including first and foremost the work of Klaus Armingeon – to contemporary debates about the responsiveness of democratic policy-making. I argued that the perspective of this literature on the role of interest groups is predominantly influenced by pluralist conceptions of the relationship between public opinion, interest groups and the state. It neglects the potential contribution and value of corporatist institutions and practices to mitigating political inequalities and non-responsiveness of decision-making to the interests of the weak segments of society. In this concluding section, I would like to follow the example of Armingeon (1986) by providing an updated list of “conditions” that would have to be met in order for neo-corporatist institutions and practices to realize this potential contribution:3 1.

The actors involved in neo-corporatist institutions and practices need to share a basic agreement concerning problem definition and the range of potential (feasible) policy responses, in spite of the fact that they commonly disagree on the most preferred policy

2.

3.

4.

choice among the range of options. The actors involved need to accept associational representatives from the “other side” as legitimate negotiation partners and be willing to engage in goal-oriented negotiations with these partners rather than promoting their political destruction. The inclusive potential of corporatist institutions and practices needs to be regularly monitored in order to avoid the formation of insider-oriented oligarchies: Newly established stakeholders (associations) should be allowed to participate and contribute, and corporatist institutions and practices should be employed to compensate for organizational and capacity-related obstacles in the representation of the “weak” segments of society, i.e. those lacking political, social and economic resources. Government policies need to accommodate and support joint decision-making via corporatist institutions and practices, including the delegation of quasi-public responsibilities to corporatist bodies. At the same time, following the idea of negotiations in the “shadow of hierarchy” by Scharpf (2000, 323), the state needs to retain an essential capacity to act unilaterally in case corporatist institutions and practices result in grid-lock or are captured by particularistic interests.

3

It is important to emphasize that these “conditions” by no means imply a functionalist relationship between structures and policy output, but should rather be interpreted as factors that influence the potential of corporatist institutions to address non-responsiveness. 28

These ‘conditions’ are by no means easy to fulfill. The modest ambition of this chapter is certainly not to provide a final answer to the questions raised here, but to raise the awareness in contemporary debates about the responsiveness of democratic policy-making for earlier contributions by innovative and thought-provoking scholars such as Klaus Armingeon.

References Anderson, Christopher J., and Pablo Beramendi. 2008. “Income, Inequality, and Electoral Participation.” In Democracy, Inequality, and Representation, edited by Pablo Beramendi and Christopher J. Anderson, 278311. New York: Russell Sage Foundation. Anzia, Sarah F., and Terry M. Moe. 2016. “Do Politicians Use Policy to Make Politics? The Case of Public-Sector Labor Laws.” American Political Science Review 110 (4): 763-777. Armingeon, Klaus. 1986. “Formation and Stability of Neo-Corporatist Incomes Policies: A Comparative Analysis.” European Sociological Review 2 (2): 138-147. Armingeon, Klaus. 2002. “The Effects of Negotiation Democracy: A Comparative Analysis.” European Journal of Political Research 41: 81-105. Armingeon, Klaus, and Lisa Schädel. 2015. “Social Inequality in Political Participation: The Dark Side of Individualisation.” West European Politics 38 (1): 1-27. Bartels, Larry M. 2008. Unequal Democracy: The Political Economy of the New Gilded Age. Princeton: Princeton University Press. Bartels, Larry M., Hugh Heclo, Rodney E. Hero, and Lawrence R. Jacobs. 2005. “Inequality and American Governance.” In Inequality and American Democracy: What We Know and What We Need to Learn, edited by Lawrence R. Jacobs and Theda Skocpol, 88-155. New York: Russell Sage Foundation. Baumgartner, Frank R., Jeffrey M. Berry, Marie Hojnacki, David C. Kimball, and Beth L. Leech. 2009. Lobbying and Policy Change: Who Wins, Who Loses, and Why. Chicago: University of Chicago Press. Bernauer, Julian, Nathalie Giger, and Jan Rosset. 2015. “Mind the Gap: Do Proportional Electoral Systems Foster a More Equal Representation of Women and Men, Poor and Rich?” International Political Science Review 36 (1): 78-98. Burstein, Paul. 2003. “The Impact of Public Opinion on Public Policy: A Review and an Agenda.” Political Research Quarterly 56 (1): 29-40. Eichhorst, Werner, and Paul Marx. 2012. “Whatever Works: Dualization and the Service Economy in Bismarckian Welfare States.” In The Age of Dualization: The Changing Face of Inequality in Deindustrializing Societies, edited by Paul Emmenegger, Silja Häusermann, Bruno Palier and Martin Seeleib-Kaiser, 73-99. Oxford, New York: Oxford University Press. Elsässer, Lisa, Svenja Hense, and Armin Schäfer. 2017. “Dem Deutschen Volke? Die ungleiche Responsivität Des Bundestags.” Zeitschrift für vergleichende Politikwissenschaft 27 (2): 161-180. Emmenegger, Patrick, Silja Häusermann, Bruno Palier, and Martin Seeleib-Kaiser, eds. 2012. The Age of Dualization: The Changing Face of Inequality in Deindustrializing Societies. Oxford, New York: Oxford University Press. Erikson, Robert S., Michael B. MacKuen, and James A. Stimson. 2002. The Macro Polity. Cambridge, New York: Cambridge University Press. Giger, Nathalie, Jan Rosset, and Julian Bernauer. 2012. “The Poor Political Representation of the Poor in a Comparative Perspective.” Representation 48 (1): 47-61. Gilens, Martin. 2005. “Inequality and Democratic Responsiveness.” Public Opinion Quarterly 69 (5): 778-796. Gilens, Martin. 2012. Affluence and Influence: Economic Inequality and Political Power in America. Princeton: Princeton University Press. Gilens, Martin, and Benjamin I. Page. 2014. “Testing Theories of American Politics: Elites, Interest Groups, and Average Citizens.” Perspectives on Politics 12 (3): 564-581. Hacker, Jacob S., and Paul Pierson. 2010. Winner-Take-All Politics: How Washington Made the Rich Richer - and Turned Its Back on the Middle Class. New York: Simon & Schuster. Hakhverdian, Armen. 2010. “Political Representation and Its Mechanisms: A Dynamic Left-Right Approach for the United Kingdom, 1976-2006.” British Journal of Political Science 40 (4): 835-56. Hicks, Alexander, and Lane Kenworthy. 1998. “Cooperation and Political Economic Performance in Affluent Democratic Capitalism.” American Journal of Sociology 103 (6): 1631-1672. Hobolt, Sara Binzer, and Robert Klemmensen. 2008. “Government Responsiveness and Political Competition in Comparative Perspective.” Comparative Political Studies 41 (3): 309-337. 29

Huber, John D., and G. Bingham Powell. 1994. “Congruence between Citizens and Policymakers in Two Visions of Liberal Democracy.” World Politics 46 (3): 291-326. Jennings, Will, and Peter John. 2009. “The Dynamics of Political Attention: Public Opinion and the Queen's Speech in the United Kingdom.” American Journal of Political Science 53 (4): 838-854. Katzenstein, Peter J. 1987. Policy and Politics in West Germany: The Growth of a Semi-Sovereign State. Philadelphia: Temple University Press. Lehmbruch, Gerhard. 1979. “Consociational Democracy, Class Conflict, and the New Corporatism.” In Trends toward Corporatist Intermediation, edited by Philippe C. Schmitter and Gerhard Lehmbruch, 53-61. Beverly Hills, London: Sage. Lijphart, Arend. 1999. Patterns of Democracy. New Haven, CT: Yale University Press. Martin, Cathie Jo, and Duane Swank. 2008. “The Political Origins of Coordinated Capitalism: Business Organizations, Party Systems and State Structure in the Age of Innocence.” American Political Science Review 102 (2): 181-198. Martin, Cathie Jo, and Duane Swank. 2012. The Political Construction of Business Interests: Coordination, Growth and Equality. Cambridge, New York: Cambridge University Press. Offe, Claus, and Helmut Wiesenthal. 1980. “Two Logics of Collective Action: Theoretical Notes on Social Class and Organizational Form.” Political Power and Social Theory 1 (1): 67-115. Olson, Mancur. 1965. The Logic of Collective Action: Public Goods and the Theory of Groups. Cambridge, MA: Harvard University Press. Olson, Mancur. 1982. The Rise and Decline of Nations: Economic Growth, Stagflation, and Social Rigidities. New Haven, CT: Yale University Press. Page, Benjamin I., Larry M. Bartels, and Jason Seawright. 2013. “Democracy and the Policy Preferences of Wealthy Americans.” Perspectives on Politics 11 (1): 51-73. Page, Benjamin I., and Robert Y. Shapiro. 1983. “Effects of Public Opinion on Policy.” American Political Science Review 77 (1): 175-190. Palier, Bruno, and Kathleen Thelen. 2010. “Institutionalizing Dualism: Complementarities and Change in France and Germany.” Politics & Society 38 (1): 119-148. Rosset, Jan. 2013. “Are the Policy Preferences of Relatively Poor Citizens under-Represented in the Swiss Parliament?” Journal of Legislative Studies 19 (4): 490-504. Rosset, Jan, Nathalie Giger, and Julian Bernauer. 2013. “More Money, Fewer Problems? Cross-Level Effects of Economic Deprivation on Political Representation.” West European Politics 36 (4): 817-35. Rothstein, Bo. 1987. “Corporatism and Reformism: The Social Democratic Institutionalization of Class Conflict.” Acta Sociologica 30 (3-4): 295-311. Rueda, David. 2007. Social Democracy inside Out: Partisanship and Labor Market Policy in Industrialized Democracies. Oxford, New York: Oxford University Press. Schäfer, Armin. 2010. “Die Folgen sozialer Ungleichheit für die Demokratie in Westeuropa.” Zeitschrift für vergleichende Politikwissenschaft 4 (1): 131-156. Scharpf, Fritz W. 1991. Crisis and Choice in European Social Democracy. Ithaca, London: Cornell University Press. Scharpf, Fritz W. 2000. Interaktionsformen: Akteurzentrierter Institutonalismus in der Politikforschung. Opladen: Leske und Budrich. Schattschneider, Elmar Eric. 1960. The Semisovereign People: A Realist's View of Democracy in America. Chicago: Holt, Rinehart and Winston. Schlozman, Kay Lehman, Benjamin I. Page, Sidney Verba, and Morris P. Fiorina. 2005. “Inequalities of Political Voice.” In Inequality and American Democracy: What We Know and What We Need to Learn, edited by Lawrence R. Jacobs and Theda Skocpol, 19-87. New York: Russell Sage Foundation. Schmitter, Philippe C. 1979. “Still the Century of Corporatism?” In Trends toward Corporatist Intermediation, edited by Philippe C. Schmitter and Gerhard Lehmbruch, 7-52. Beverly Hills, London: Sage. Schmitter, Philippe C. 1985. “Neo-Corporatism and the State.” In The Political Economy of Corporatism, edited by Wyn Grant, 32-62. Houndmills, Basingstoke: Palgrave Macmillan. Solt, Frederick. 2008. “Economic Inequality and Democratic Political Engagement.” American Journal of Political Science 52 (1): 48-60. Soroka, Stuart N., and Christopher Wlezien. 2008. “On the Limits to Inequality in Representation.” PS: Political Science and Politics 41 (2): 319-327. Soroka, Stuart N., and Christopher Wlezien. 2010. Degrees of Democracy: Politics, Public Opinion, and Policy. Cambridge, New York: Cambridge University Press. Stimson, James A. 1991. Public Opinion in America: Moods, Cycles, and Swings: Westview Press. Stimson, James A. 2011. “The Issues in Representation.” In Who Gets Represented?, edited by Peter K. Enns and Christopher Wlezien, 347-360. New York: Russell Sage Foundation.

30

Stimson, James A., Michael B. MacKuen, and Robert S. Erikson. 1995. “Dynamic Representation.” American Political Science Review 89 (3): 543-565. Streeck, Wolfgang, and Lane Kenworthy. 2005. “Theories and Practices of Neocorporatism.” In The Handbook of Political Sociology: States, Civil Societies, and Globalization, edited by Thomas Janoski, Robert R. Alford, Alexander M. Hicks and Mildred A. Schwartz, 441-460. Cambridge, New York: Cambridge University Press. Streeck, Wolfgang, and Philippe C. Schmitter. 1985. “Community, Market, State – and Associations? The Prospective Contribution of Interest Governance to Social Order.” In Private Interest Government: Beyond Market and State, edited by Wolfgang Streeck and Philippe C. Schmitter, 1-29. London, Beverly Hills, New Delhi: Sage. Ura, Joseph Daniel, and Christopher R. Ellis. 2008. “Income, Preferences, and the Dynamics of Policy Responsiveness.” PS: Political Science and Politics 41 (4): 785-794. Van Kersbergen, Kees. 1995. Social Capitalism: A Study of Christian Democracy and the Welfare State. London, New York: Routledge. Wallerstein, Michael. 1999. “Wage-Setting Institutions and Pay Inequality in Advanced Industrial Societies.” American Journal of Political Science 43 (3): 649-680.

31

Social concertation in Europe during the Great Recession: A fsQCA-study of social partner involvement Benedikt Bender1 and Bernhard Ebbinghaus2

Abstract: The Great Recession, following the financial market crash of 2008, has been a major challenge not only to governments dealing with the crisis but also for state-society relations between government and social partners across Europe. This chapter addresses the question of why some governments engage in tripartite social concertation between governments, employer organizations and trade unions as part of crisis management, while others choose to unilaterally pursue their agendas. Drawing on an extensive dataset of 29 European countries (EU-27 plus Norway and Switzerland), we use fuzzy-set qualitative comparative analysis (QCA) to assess which factors can be identified as necessary or sufficient conditions for concertation efforts during the first five years since the crash. Specifically, the analysis studies whether economic pressure, governmental desire for political legitimacy, interest groups’ power resources, institutional context and corporatist legacy promote or prevent tripartite interest mediation. The main findings indicate that labour relations and political factors seem to dominate the overall pattern of including social partners in crisis management since 2008.

1

Introduction

The Great Recession that began with the global financial market crash of 2008 and subsequently led to the Euro-debt crisis with supranational intervention has called for crisis management.3 Are governments pushing through austerity measures unilaterally without consulting the “social partners”, i.e. trade unions and employer associations? Or do governments seek to negotiate with them encompassing “social pacts”, thereby coordinating changes in macro-economic steering, wage bargaining, labour market rules and social policies? To manage the economic and social repercussions of the crisis, some governments engaged in negotiations with trade unions and employer associations, whereas others chose to pursue austerity measures unilaterally. Our comparative study seeks to answer why certain governments across Europe aimed more or less at social concertation in the five years since the 2008 crash. 1

Universität Mannheim [email protected] University of Oxford [email protected] 3 The authors would like to thank Florian Andersen for research assistance in respect to data collection, coding and analysis. 2

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 R. Careja et al. (eds.), The European Social Model under Pressure, https://doi.org/10.1007/978-3-658-27043-8_3

33

Since the 1970s, the corporatist literature postulated that economic crises were a window of opportunity for macro-economic social concertation (Lehmbruch and Schmitter 1982; Schmitter and Lehmbruch 1981). Klaus Armingeon’s early oeuvre (1983, 1986) on neocorporatist income policy showed that variations in corporatism dealing with the economic crises of the 1970s was dependent on multiple factors, including centralized organizations and labour relations. After the end of corporatism had been pronounced by some observers, corporatist scholars hailed a “renaissance” of social pacts since the 1990s (Grote and Schmitter 1999). This occurred because EU member states needed to adapt to the Single Market of 1992 and particularly prior to joining the European Monetary Union (EMU) in 1999. A major comparative study analysing the 1990s found that political factors in addition to economic pressure led to social pacts across Europe (Avdagic et al. 2011), yet the data ended around 2008. The Great Recession since 2008 might be seen as a new window of opportunity for “crisis corporatism”, namely governments seeking concertation to respond to the crisis in a social inclusive strategy (e.g. Guardiancich and Molina 2017). In contrast, other scholars (e.g. Culpepper and Regan 2014) postulate a “crisis of corporatism” as governments side-line unions and employers. How have governments responded to the financial market crash of 2008 and subsequent Euro-debt crisis from 2010? We might expect that governments seek to use social concertation to coordinate macro-economic and social policies with the social partners. However, there might also be economic and political reasons why social partners are side-lined and governments eager to push through crisis measures unilaterally, not least because the “Troika” (EU, ECB, IMF) of supranational creditors left no space for domestic concertation. There have been thus far few studies describing social partners’ involvement during the Great Recession (Guardiancich and Molina 2017; Vaughan-Whitehead 2018). Our analysis is the first systematic comparison of the economic, political and labour relations conditions leading to social concertation efforts across nearly all European countries. In our comparative analysis we seek to map social concertation during the initial period of crisis management (2008-12). On the one hand, we seek to contribute to corporatist and political economy theories by analysing the conditions for social concertation in particularly difficult economic times. On the other, our empirical contribution is a systematic comparison across a large number of countries in Europe. We build upon, and extend the empirical research on social pacts carried out before the 2008 crash (Avdagic et al. 2011; Fajertag and Pochet 2000), reviewing their theoretical assumptions and empirically testing these under the recent crisis conditions. For this purpose, we seek to identify those conditions that promote (or hinder) social concertation, though we are less restrictive in our concept than in previous social pact studies. We use Qualitative Comparative Analysis (QCA), as developed by Charles Ragin (1987, 2008b), in order to systematically compare necessary and sufficient conditions that are concomitant with the (non)outcome across countries and over time, allowing to account for complexity, equifinality and asymmetry of the explanatory factors. By applying fuzzy set QCA (fsQCA) we calibrate theoretically defined and empirically informed set-theoretic fuzzy scores 34

with respect to the outcome (social concertation) and potential explanatory conditions, ranging from fully out to fully in. Based on annual indicators (2008-2012) collected for 29 European countries we identify necessary and sufficient conditions for the outcome of social concertation using fsQCA (Schneider and Wagemann 2010).

2

From corporatism to social concertation

Since the 1970s, (neo)corporatism has been a popular concept explaining interest intermediation in tripartite wage formation and policy-making between government and organized interests across advanced capitalist economies (Schmitter 1974). Schmitter (1974) defines corporatism as a system of interest intermediation with particular features of interest organizations, while Lehmbruch (1967) focuses more on the particular forms of policy-making and implementation, the particular style of social concertation that involves organized interests. We thus follow corporatist theory (Lehmbruch and Schmitter 1982; Schmitter and Lehmbruch 1981) by understanding tripartite interest intermediation as a form of policy-making in which organized interest groups are involved in the formulation and implementation of public policy with the consent of government. Social concertation describes both formal and informal consultation of the government with the social partners (employer associations and trade unions) with the aim of a joint agreement on public policy goals and the implementation of concrete measures (Berger and Compston 2002; Regini 2003). It can take place via formal tripartite institutions (such as a social and economic council) or via informal ad-hoc contacts (e.g. talks at the Chancellor’s office). Under certain conditions, concertation can also take place with divided responsibilities, i.e. the negotiations of the social partners are supported indirectly by the government. Concertation in our analysis is thus a continuum of coordination forms, ranging from exclusion due to unilateral government action to a social pact signed by governments and social partners. According to Avdagic et al. (2011) a social pact is defined as a formal agreement between government and social partners on economic objectives and measures that are implemented via legislation and/or collective bargaining. Social concertation in our view is a necessary precondition to successfully negotiate and sign a social pact. Earlier studies on social pacts focused solely on such negotiated social pacts signed by all three sides (Avdagic 2010; Avdagic 2011), while we use fuzzy sets to analyse the different degrees of social concertation, ranging from unilateral government action to a broad social pact. Considering the results of previous research on social pacts in Europe (Avdagic et al. 2011) there are indications that these are waning since the onset of the Great Recession (Guardiancich and Molina 2017; Vaughan-Whitehead 2015). In contrast to their 1990s renaissance, recent evidence shows that there have been fewer social pacts since 2008 and these occurred least in countries severely hit by the Great Recession. One reason for this discontinuity may be that governments grappled with urgent and harsh policy decisions without having much to offer to 35

the social partners in negotiations. Earlier short-term Keynesian demand-side policies allowed some compensation, but once public debt increased due to the financial and then Euro-debt crises, governments were pressured into fiscal consolidation, not least through international creditors in the case of the crisis countries. Governments have not much to offer to unions in return for their concessions, whereas business might have been unwilling to engage in any negotiations. During the Great Recession only few governments negotiated a picture-book social pact across broad policy issues with the social partners, but many more efforts failed. Ireland was known for social concertation since the Program for National Recovery in 1986 until the Towards 2016 agreement signed in 2000. These social pacts included fiscal, tax, unemployment, monetary and welfare policy, as well as measures dealing with disability, social exclusion, drug abuse, and other social problems (O’Donnell et al. 2011, 89). Shortly before and during the crisis, Ireland ran into severe problems with concertation (Culpepper and Regan 2014; Regan 2017). Romania signed a broad pact in 2008, but within a year it was no longer honoured by a newly elected government which declared it as too far reaching and inflexible (Guyet et al. 2012, 77). In contrast to earlier decades (Avdagic et al. 2011; Hassel 2006) most of the European countries shifted from broader social pacts to more limited, often “hand-shake” and singleissue-agreements during the Great Recession. A comparative study observes a discontinuity in social dialogue across several European countries since 2008 (Guardiancich and Molina 2017). Since the crash, governments sought rather short-term measures to be negotiated with the social partners as part of stimulus packages, anti-crisis labour market measures, and training initiatives fighting youth unemployment. Due to the changed problem load and limited opportunity structures, it would be rather strict to set the bar for social concertation too high by expecting only broad social pacts. Instead, we seek to study the phenomenon of social concertation by applying a wider definition and using fuzzy scores to capture degree of engagement. In terms of set theory, we see the stricter social pact concept as a smaller subset of social concertation, which is in turn a superset of the former.

3

Hypotheses on explanatory factors for social concertation since 2008

From an analytical perspective (Avdagic et al. 2011; Hassel 2006), there are three dominant explanatory approaches in the political economy literature explaining the occurrence (or absence) of social concertation, which will be discussed in the following. Economic factors From a functionalist perspective, social concertation is needed when it fulfils a specific economic function, in particular, overcoming (or preventing) an economic crisis. For the period 36

since 2008 we could expect that a (successful) tripartite intermediation should occur particularly in Europe following the financial market crash and/or subsequent Euro-debt crisis. After the crisis has been overcome, the functionalist expectation would predict that concertation is less likely to continue. We measure economic pressure by high unemployment, low growth, and severe government deficit. H1a: The severity of an economic crisis increases the need for social concertation; the higher the economic pressure, the more the government and social partners will seek social concertation. More specifically, the Varieties of Capitalism approach (Hall and Soskice 2001; Hall and Gingerich 2009; Regini 2003) expects that coordination of capital and labour would be advanced particularly in Coordinated Market Economies, given dominance of export-oriented sectors dependent on ensuring productive skills within firms. Social concertation is thus more likely in export-oriented economies that tend towards competitive corporatism (Rhodes 1997). We include therefore a measure of economic openness (exports plus imports as share of GDP). The competitive corporatism thesis would assume that openness also facilities more powerful unions and higher centralization, although the reverse relationship might also be possible (strong unions engaged in concertation facilitate economic openness). H1b: As forms of competitive corporatism, social concertation is more common to occur in open economies. Power resources approach and labour relations Following the Power Resources Theory (Korpi 1983), concertation is dependent on favourable (political) distribution of power. Trade unions are identified as pivotal actors in macroeconomic concertation, albeit they need sufficient power resources to convince the government and employer associations to enter negotiation. Their power is commonly measured by union (membership) density, internal centralization, and strike action (Ebbinghaus and Visser 2000; Visser 1987). Corporatist theory postulates that union movements should be organised comprehensively, following their membership logic, while they also need to be sufficiently centralized to commit their sub-organisations and their members to the negotiated outcome, following the logic of interest intermediation (Schmitter and Streeck 1999). In addition, the power resources approach implies that the union movement needs a partner in the government. In the past, especially left governing parties were considered a prerequisite for corporatism (Alvarez et al. 1991; Hicks 1988). The Varieties of Capitalism approach also stresses the importance of employer associations, in particular the ability to coordinate for achieving cooperative goods in an economy. We would thus expect that employer organizational density is facilitating concertation on the side of employers. 37

We are measuring union power by union density (share of unionized employees) and by the centralization of unions (Ebbinghaus and Visser 2000; Visser 1987). H2a: Social concertation is facilitated by strong and comprehensive unions (and employer associations) since they can reach broad and high levels of commitment. Comparative analyses stress the importance of corporatist institutions such as tripartite fora and corporatist traditions that can be activated during crisis management (Berger and Compston 2002; Ebbinghaus 2010). In order to account for path dependent developments, we also include corporatist legacy by using a measure of corporatism before 2008 based on the indicator developed by Jahn (2017) which goes beyond earlier indices (Kenworthy 2003; Siaroff 1999; Traxler et al. 2001). It measures the degree of hierarchical centralization, the degree of concertation, and degree to which agreements encompass broader segments of society. Jahn’s index is a multidimensional measure that goes beyond our outcome and captures a more general corporatist legacy. H2b: Countries with a corporatist legacy (before 2008) will tend to continue to engage social partners in policy making under conditions of crisis. Political institutions and actors Social concertation is only possible when a government seeks to include social partners in concertation and is willing to support a negotiated deal (Scharpf 1991). Governments may seek concertation when they are too weak to resort to unilateral action. We expect that they engage in concertation when their unilateral decision-making autonomy is compromised (Afonso 2014), such as under a minority government, a (grand) coalition with ideological conflicts or a fragmented party system. Moreover, the existence of a major veto point in federalist or bicameral political systems also limits the central government’s capability. We include several political indicators such as government weakness (from single-party majority to multi-party minority governments, party fragmentation and veto points. In addition, we include also the time to the next (scheduled) election because we presume that governments facing an imminent election are more willing to consult than those recently elected and not expecting to face an election in the near future. H3: Minority or non-partisan governments, divided coalitions and governments in political systems with veto points are more willing to engage in social concertation.

4

Fuzzy Set Qualitative Comparative Analysis

In our study we investigate the following research question: Under what conditions were social partners included in the government’s crisis management via social concertation during the Great Recession? In our comparative analysis we study the macro-configurations that fostered 38

or hampered social concertation across Europe. Since we study a large set of 29 European countries over five years (2008-2012) we use Qualitative Comparative Analysis (QCA), developed by Charles Ragin (1987). This logical approach allows accounting for an outcome across many cases through particular combinations of necessary or sufficient conditions. It is a method to examine whether a combination of factors is concomitant (or not) with the outcome (or the non-outcome) across observable cases (and non-observed logical implications) based on logical analysis (Ragin 2008b). The approach does not assume any general linear relationship like regression analysis, but instead QCA is a configurational analysis particularly well suited for complex, equifinal and asymmetric relationships. Under equifinality different paths could lead to the same outcome. For instance, governments in countries with a corporatist legacy might be more inclined to seek social concertation, while in countries with no corporatist tradition there might be other conditions that propel governments to do so. QCA allows for testing asymmetric relationships of the non-outcome in respect to the outcome, in contrast to the general linearity assumption of regression analysis (Abbott 1988). Given the small-N problem of too few observations for many potential variables, the logical analysis is more suited to deal with few observations. In order to go beyond binary coding, we use fuzzy set QCA (fsQCA) to analyse logical properties of outcomes and conditions coded in fs-scores ranging from 0 to 1, indicating different fine-grained degrees of set membership. We apply fsQCA using a set theoretic approach to overcome the binary coding of crisp QCA, which limits analysis to logical true and false only. Instead, fsQCA (Schneider and Wagemann 2012) assumes that outcome and conditions can be measured in fuzzy membership scores, ranging from “fully out” (0) to “fully in” (1) with a cross-over point (0.5), by carefully calibrating between the two theoretically derived and empirically informed extremes. The fsQCA procedure allows us to calibrate in a theoretically and empirically informed way the range of phenomena in their quantitative but also qualitative aspects. Our fsQCA-comparison includes 27 EU member states (EU-28 without Croatia) from 2008 to 2012, as well as Norway and Switzerland with corporatist traditions (Afonso 2013). These 29 countries cover a range of different labour relations (Crouch 1993), varieties of capitalism (Hall and Soskice 2001) and welfare state regimes (Esping-Andersen 1990). Our unit of analysis is not a period as in Avdagic (2010) or a cabinet duration (Vis 2011), but yearly observations for each country during the period (2008-2012). Annual analysis is crucial given the quickly evolving crisis: 2008-2009 was driven by the global financial crisis, whereas the Euro-debt crisis dominated many countries since 2010, albeit others started to recover (Guardiancich and Molina 2017). Hence, our unit of analysis are instances of social concertation in a given country for a particular year. Although we use corporatist legacy before 2008, we are not controlling in our fsQCA for concertation in the year before even if a pact might limit the need for further action for several years. 39

Calibrating the outcome: social concertation We conceive social concertation as a fuzzy set outcome ranging from unilateral government policies of adjustment (0, fully out) to comprehensive social pacts between government and both social partners to mitigate the crisis (1, fully in). We code degrees of concertation between these ideal typical extremes in six steps (see Table 1) with the cross-over point at 0.5, marking the divide between in and out of the set of social concertation). Based on the corporatist literature, we code social concertation as more or less fully in based on the criteria: (i) the breadth and depth of social concertation: how many policy issues are addressed and in which depth (in line with the concept of social pacts), (ii) the type of exchange: there has to be concessions resulting from corporatist bargaining, and (iii) an encompassing intermediation between government and social partners, employer associations and trade unions in policymaking and implementation (Afonso and Papadopoulos 2013; Avdagic et al. 2011; Baccaro and Simoni 2008). The last point refers to whether policy-making is unilateral, whether societal actors are only ad hoc consulted, or whether concertation is bi- or tripartite. We construct a fuzzy set called “tripartite crisis management” (social concertation) based on all criteria. Table 1: Fuzzy set calibration of outcome fs / Broadness of Concertation Score 1 fully in

Actors Involved

Condition

Broad concertation with

All three actors (main peak

Broad concertation AND

tripartite exchange of mutual strategic concessions leading to

associations) are involved in bargaining and signatory

tripartite negotiations

a formal pact

parties

0.9 nearly

Narrow concertation with the process of exchange of mutual

At least government and one social actor (unions or

fully in

strategic concession with or

employers) are concerting

0.6 partly in

0.4 partly

without a formal pact

Tripartite negotiations OR broad concertation (government with at least one actor)

Consultation (i.e. crisis

Government plus unions, or

For 0.6 fs, governments need

summits) with some form of outcome (even purely

government plus employer associations

to consult both social partners in some form, AND there needs

symbolical)

Consultation (i.e. crisis

to be some form of outcome Above 0.5 fs: political exchange Below 0.5 fs: information sharing only All three actors (government

Government consults both side

summits) without some form of

and peak associations) are

on crisis-related measures

out

outcome, not even a symbolical declaration

involved

AND no agreement

0.1

Crisis-related talks (i.e. crisis

At least two actors (peak

Only two of all three actors

nearly out

summits) without some form of outcome, not even symbolical

associations or government) are involved

discuss crisis-related measures

0.0

No consultation other than

Government unilateralism

fully

already institutionalized wage

out

bargaining

40

AND no agreement

Social concertation needs, firstly, to address more than one issue in relative depth, and there has to be a political exchange and, secondly, all three actors (government, trade unions, and employer organizations) should be equally and fully included. Accordingly, forms of social concertation that do not lead to a signed pact or take on a more informal form of agreement (like a handshake agreement) are considered to be forms of social concertation. Moreover, social concertation needs to be distinguished from normal wage bargaining, and must at least involve the government, thus purely bipartite bargaining is not considered social concertation. Table 2: Social concertation fuzzy scores, Europe, 2008-2012 Fully in (1)

Nearly fully in (0.9)

More in than out (0.6)

More out than in (0.4)

Nearly fully out (0.1)

2008

EE, NL, RO

CY, EL, UK

3 (10.7%) EE, NL

6 (21.4%) IE, MT, ES

3 (10.7%) CY, EL

0 UK

28

N 2010

2 (6.9%)

3 (10.3%) DK, FI, PL, SI

2 (6.9%) EL, ES,

1 (3.4%) CY, UK

29 (CH, MT)

N 2011

0 (0%) FI, NL

5 (18.5%) LV, PT

4 (14.8%) CZ, EE, DE, HU, MT, SI

2 (7.4%) CY, IE, IT, PL, RO

2 (7.4%) EL, UK

27 (CH)

N 2012

2 (7.1%)

2 (7.1%) FR, LT, NO

6 (21.4%) BG, CY, CZ, DE, PL, SI

5 (17.9%) HU, MT, ES, IE

2 (7.1%) EL, RO, UK

28 (CH)

N

0 (0%)

3 (10.7%)

BG, FR, IT, LV, ES, SE, NO, CH 8 (28.6%) BG, CZ, HU, PL, PT, RO, SI, SE, NO, CH 10 (34.5%) AT, BE, CZ, EE, DE, IE, IT, LV, LT, PT, RO, SK, SE, NO 14 (51.8%) AT, BE, BG, DK, FR, LT, LU, SK, ES, SE, NO 11 (39.3%) AT, BE, DK, EE, FI, IT, LV, LU, NL, PT, SK, SE 12 (42.9%)

FI, HU, LT, LU, MT, PL

N 2009

AT, BE, DK, IE, PT, SI, SK, DE 8 (28.6%) AT, BE, DK, FR, DE, IT, LV, SK, LT, LU, FI 11 (37.9%) BG, FR, HU, LU, NL

6 (21.4%)

4 (14.3%)

28

29 (20.7%)

55 (39.3%)

25 (17.8%)

16 (11.4%)

3 (10.7%) 8 (5.7%)

Year

7 (5.0%) (*) missing cases. All

Fully Out (0.0)

N (*) (CZ)

140

The mapping of the outcome coding for 29 countries and five years (see Table 2) shows some cross-national variation over time. There are only seven social concertation cases that are fully in (5% of all 140 country-year cases), particularly the Netherlands (three times) stand out next to Estonia (twice), Finland and Romania (once). The next category (.9, nearly fully in) has 29 observations (20.7%), while another 55 (nearly 40%) cases are in the more in than out category (.6). Thus, nearly two thirds of all annual observations are cases of social concertation (.6 or more), while only few countries tend to be nearly or fully out for more than one year, in particular the United Kingdom, Cyprus, Greece, Ireland, Romania and Spain. 41

From indicators to fuzzy set conditions For studying explanatory conditions we gathered indicators (as raw data) from various sources (OECD, ICTWSS, CPDS, Eurostat, ILO) and calibrated these to fuzzy scores (see online appendix). For most data, we used a combination of quantitative and qualitative calibration techniques. We used descriptive statistics of raw variables, substantial case knowledge, and theoretical considerations to derive fuzzy set calibrations. For many metric raw variables, we used a logistic transformation applying empirically and theoretically plausible thresholds to raw data, whereas for categorical variables, we used case knowledge and theory to directly assign fuzzy set values to individual categories (Thiem 2018).4 For the sufficiency analysis we combined sets of conditions for political, economic, and labour relations models, while also checking for some single condition models in combination with these models following our hypotheses. Although the choice between quantitative and qualitative calibration leaves room for debate, we make our coding explicit and transparent (for details, see online appendix) following the best standard in QCA (Ragin 2008a).

5

Necessary and sufficient conditions for social concertation

Following Schneider and Wagemann (2010), we search for necessary conditions (outcome Y is a subset of condition X) before focusing on sufficient conditions (X is a subset of Y). To take into account QCA’s asymmetric logic, we analyse the outcome of social concertation as well as its non-occurrence separately, acknowledging the possibility of different explanations for the outcome and non-outcome (Berg-Schlosser et al. 2008; Schneider and Wagemann 2010). Necessary conditions for social concertation Our first step is to test for necessary conditions, that is, conditions that are always present when we observe the outcome, applying a standard consistency threshold of 0.90 (Ragin 2008b, 118; Schneider and Wagemann 2012, 139). We find no single condition that is necessary for social concertation or for the non-outcome. In other words, there is no single condition that is nearly always present when social concertation is observed. Not even in countries with corporatist legacies do we find a necessary condition, which falsifies the legacy argument of corporatism (consistency 0.63, coverage 0.66). The only condition that comes close to the consistency threshold is labour relations, triggered by a high employer density (0.72) and the scarcity of veto-points (0.75). Since the coverage is relatively low for these quasi-necessary conditions and they also have such medium-high values for the opposite outcome, these conditions on their own seem rather trivial and undistinctive. Indeed, employer density is often present among all the cases of social concertation (coverage 0.72), but it also present with the non-outcome (only 4

We use the R package QCAPro, Version 1.1-2 (URL: http://www.alrik-thiem.net/software).

42

0.68 of X are concomitant with Y). Therefore, it seems ubiquitous independent of the outcome. For the non-outcome (absence of social concertation) a non-fragmented party-system (0.88) and a non-weak government (0.80) are minimally necessary conditions. However, while fsQCA recognizes asymmetric relationships, we actually need a single analysis for the non-outcome with different conditions, because the necessary conditions for the presence of an outcome are not the same conditions as those for the non-outcome. Moreover, the consistency values of both, outcome and non-outcome, are nearly the same; this makes a robust interpretation impossible. These patterns confirm that there is no single condition that is present for all cases of social concertation. Sufficient conditions for social concertation As a second step, we now analyse sufficiency, that is, whether a condition or combination of conditions is concurrent with the outcome, even if it does not necessarily apply to all outcomes. We follow Schneider and Wagemann (2012) in inspecting consistency to identify gaps in the distribution, resulting in a consistency threshold of 0.85. However, facing the problem of limited diversity (Ragin 1987) we have too many logical combinations of conditions (213 or 8192) to run the minimization process and compare those against only 140 observed cases. We therefore group three subsets comprising thematically and theoretically related conditions: (1) labour relations conditions including high union density, centralized union movement, corporatist legacy and high employer density; (2) political conditions: weak government, fragmented party system, near regular election, veto point system and left government; and (3) economic conditions, grouping low growth, high unemployment, government deficit and open economy. We apply parsimonious minimization (Baumgartner and Thiem 2017) in which logical remainders are used flexibly in order to minimize solution terms (consistency >0.85). We also add selected single (political or institutional) conditions to these groups in order to test interactions. We minimize these conditions with the Quine-McClusky algorithm (see Table 3). All solution paths are above the minimum (0.75) consistency threshold recommended for sufficient conditions (Schneider and Wagemann 2012, 127).

43

Table 3: Best sufficient models for social concertation5 Model (conditions) 1) Labour Relations high_ud*HIGH_UC

Hypotheses (H2a)

Consistency 0.858 0.858

Raw Coverage 0.315 0.315

Unique Coverage -0.315

0.729 0.737 0.818

0.666 0.604 0.374

2) Political

0.780

0.694

--

FRAG_SEATS

H3

0.820

0.442

0.201

(H3)

0.811

0.454

0.220

0.861

0.088

0.016

Labour Relations & HIGH_UC CORP_LEG*WEAK_GOV

veto*WEAK_GOV

H2a H2b/(H3)

VETO*ele_near*GOV_LEFT_S

0.763

0.593

--

H1a (H1b)

0.767

0.534

0.391

0.866

0.202

0.058

H1a / H2b

0.817

0.441

--

3) Economic high_deficit LOW_GROWTH*HIGH_UNEMP* OPEN_ECON

-0.291 0.062

Economic & high_unemp*CORP_LEG

(1) The best sufficiency model for social concertation (consistency 0.86) are labour relations conditions: a combination of weak unions AND centralized unions (partially confirming Hypothesis 2a). It indicates that a third of social concertation cases (country-years) are almost fully accounted by the joint conditions. It holds for Germany (2008, 2009), Latvia (2009, 2011), Slovakia (2008, 2009) and the Netherlands (2008-2011), though Germany 2010, Latvia 2008, 2012, Slovakia 2010-2012 and the Netherlands 2012 (see Figure 2) have conditions that suggest more than a 0.6 concertation score. Germany (2011, 2012) is the only inconsistent case with the condition (2% of all 140 cases). Governments find it easier to engage with centralized unions, while they are more worried about too powerful unions. The combination occurs particularly in Eastern Europe, where institutionalized participation is relatively low (Slovakia and Latvia) but gives these relatively centralized unions some voice, though their weakness remains an obstacle to more far-reaching policy-making inclusion. The government is willing to engage in social concertation with a centralized but not too strong union movement, thus maintaining its lead. In contrast to neo-corporatist theory, no equal involvement, political exchange or broad social pact seems possible. Such efforts seem to come close to ‘PR’ corporatism that serves largely public relations purposes, in fact, many reforms tended to be implemented unilaterally by the government following the negotiations with the unions as junior partners. Facing a ‘take-it-or-leave-it’ strategy by the government, the unions are too weak to oppose it. 5

Capital letters means the presence (HIGH_UC), small letters the absence (high_ud) of the condition, refuted hypotheses are crossed out, and in brackets if only partially confirmed. 44

Figure 1: Sufficiency of social concertation: Labour Relations (high_ud*HIGH_UC)

In the Netherlands, we see a combination of weak unions AND centralized unions leading to the most frequent social pacting since 2008. The Dutch consensus democracy is based on multiparty coalitions, mutual trust, and cooperative relations; this has facilitated negotiations between political and societal actors. Moreover, the Netherlands also fulfils two political condition paths for social concertation: (i) party fragmentation OR (ii) lacking veto points AND weak government. Both solution paths are in line with the legitimation argument (H3), because a party fragmentation and coalition government make swift and bold decision-making difficult, while the lack of a veto point makes decisions without social concertation less legitimate in the public’s eye. In contrast, a single party majority government allows unilateral action without social concertation. At the height of the crash, Germany initially used crisis management when trade unions and employer representatives were invited to the Chancellor’s office, agreeing on short-time work and economic stimulus packages. After overcoming the economic shock of 2008/09, the German social partners followed business as usual, conducting bipartite negotiations in the metal and electrical sectors. While social concertation as a device for political legitimation was a major reason in Latvia, Slovakia, and the Netherlands, Germany seems the best example of short-lived crisis corporatism.

45

Figure 2: Sufficiency of social concertation: political conditions model

(2) Party fragmentation is sufficient among the political conditions (0.82) and covers more than every fifth instance of social concertation (strongly confirming H3). Furthermore, it includes a path with lacking veto points AND weak government (only the latter is consistent with H3). In Belgium, anti-crisis legislation agreements as well as a bipartite pact with the government occurred in 2009. Also in Bulgaria, the government and social partners agreed on two national agreements in 2010, while in Lithuania the first collective agreement occurred in 2012, though it was rather symbolic. More than symbolic was an Estonian tripartite agreement on job security during the early crisis (2009) on measures to maintain job security, the “first of its kind” (Eurofound 2012, 23). Concertation started already in 2008 and nearly failed as the government adopted a “take-it-or-leave-it” strategy that led the unions to organize strikes and demonstrations. Rather weak, the union leader Harri Taliga denounced the government’s policy as “clearly an attack against the weakest” (Lindstrom 2015, 126). This reinforces the legitimation argument (H3) since weak government OR party fragmentation are based on the intuition that such a government has more difficulty to legitimate its decision and it therefore needs societal support, including social partners in tripartite concertation (Baccaro and Simoni 2008). A minority or multi-party government in a system without a veto point thus attempts to prevent extra-parliamentary opposition by broadening social consensus. If a veto point exists, the government would seek consensus with the veto player instead of the social partners. Weak government and low union density go together in the Czech Republic: one agreement on 38 short-term crisis-related measures was struck in 2009 and another a year later (Guyet et al. 2012, 75), even though the social partners were weak and the multi-party minority government used the two agreements in addition to the tripartite forum. However, shortly after the second agreement, the general election in May 2010 46

produced a new coalition which ignored these agreements, while the unions were too weak to mobilize strikes. The tripartite council now faces a stable majority government, limiting its influence and preventing any real concertation. (3) Surprisingly, economic factors (H1) prove not to be sufficient conditions of social concertation, refuting the claim of a functional necessity driven by problem load. The best combination is low unemployment (contradicting H1a) AND corporatist legacy (consistency 0.82, coverage 0.44). This joint condition holds for Austria (2008-2012), Belgium (2008, 2011), Denmark (2008-2012), Norway (2008-2012), Germany (2008, 2010), Ireland (2008), Italy (2008), Luxemburg (2008-2009), Netherlands (2008-2012), and Slovenia (2008-2009). These countries are conducting social concertation even without economic pressure, though because they engage in concerted action low unemployment maybe the beneficial outcome. The corporatist legacy (H2b) seems enough for tripartite cooperation since 2008 even if unions (and employers) are weak, while weak governments are attempting concertation. The few cases engaging in social concertation but inconsistent with this hypothesis are Sweden and Luxembourg (2008), Germany (2011-2012), Finland (2008-2009) and Slovenia (2010). The best single condition in our economic set is low government deficit (consistency 0.76, coverage 0.53) that is well below the Maastricht limit. Social concertation occurs where governments are less pressed to refinance themselves on international financial markets, such as Norway (2008-2012), Netherlands (2008, 2009), Switzerland (2008, 2009) or Luxemburg (2009, 2010), albeit Latvia (2011) is an exception. Generally, the absence of international creditors is a pre-condition for social concertation. The crisis countries (Portugal, Hungary, Slovenia, Romania, Spain, Ireland, Cyprus and Greece) were rarely using social concertation. Those assisted by the Troika of the European Commission (EU), European Central Bank (ECB), and the International Monetary Fund (IMF) had to follow substantial austerity and budgetary constraints. The budget consolidation and austerity targets were implemented in order to reduce the annual deficit (below 3 percent of GDP) in Portugal (2010), Estonia, Ireland, Italy (2011), and Greece (2010-2012); these were externally imposed without any social concertation. Portugal seems to be an exception as the social partners singed a tripartite agreement in January 2012, though it was rather symbolic to signal consensus to the financial markets. After deregulation and austerity reforms in October 2012, the Portuguese unions left the permanent social dialogue ending the only concertation among crisis-ridden countries. Even if we look at the individual economic factors in a single analysis, it becomes clear that high unemployment (coverage of 0.62) or low growth (0.51) are not individually sufficient conditions for social concertation. Even if economic pressure might be a plausible explanation for an involvement of social partners, it is neither a necessary nor a sufficient condition since 2008. Economic pressure alone could not bring the social partners and the government to the negotiation table, instead it was those countries with fairer economic conditions that continued to engage in social concertation. 47

6

Conclusion

The social partners certainly play a crucial role in economic hard times such as the Great Recession. Crisis management is reflected in various declarations including economic stimulus packages, crisis-related agreements, and emergency measures. Nevertheless, our analysis reveals that it is not the crisis-ridden countries with the highest economic pressure that engage in social concertation. Instead, our findings indicate that political conditions such as party fragmentation or weak government and no veto point as well as labour relations (weak but centralized unions) are important conditions. The only economic factor concomitant with social concertation does not follow our prior expectation: low unemployment in combination with corporatist legacy is sufficient to explain social concertation. Except for those countries less affected by the crisis which engaged in corporatist business as usual, the crisis-ridden countries have been marked by “corporatism in crisis” instead of crisis corporatism. The observed patterns indicate that social concertation has slowed down since 2008 compared to the heydays of social pacts in the 1990s. Labour relations and political factors seem to dominate the overall pattern of social concertation since 2008, almost independently of economic and other factors. Seen more closely, there are however additional factors that in combination with other conditions are sufficient and form possible explanatory paths. This suggests that political legitimacy is a major reason to include social partners in crisis management since 2008. Governments include the social partners in decision making under the shadow of state hierarchy with a “take-it-or-leave-it” strategy. In sum, against the assumption that only economic pressure would lead to cooperation between the government and the social partners, one of the primary reasons to engage in social concertation for the government is to mobilize support for policy reforms and foster acceptance of such reform measures especially among voters and other parties (Afonso 2013, 44; Ebbinghaus 2010, 256). Our analysis also indicates that it is not only economic or political factors, but also labour relations that matter in leading governments to resort to social concertation. Klaus Armingeon’s comparative approach to labour relations across Europe is thus still relevant to understand today’s corporatist arrangements.

References Abbott, Andrew. 1988. “Transcending General Linear Reality.” Sociological Theory 6 (2): 169-186. Afonso, Alexandre. 2014. “How to Please Voters Without Alienating Friends? Parties, Organised Interests and Advocacy Coalitions in Swiss Immigration Policy.” Comparative European Politics 12 (6): 568–583. Afonso, Alexandre. 2013. Social Concertation in Times of Austerity. European Integration and the Politics of Labour Market Reforms in Austria and Switzerland. Amsterdam: Amsterdam University Press. Afonso, Alexandre, and Yannis Papadopoulos. 2013. “Europeanization or Party Politics? Explaining Government Choice for Corporatist Concertation.” Governance 26 (1): 5-29. Alvarez, R. Michael, Geoffrey Garrett, and Peter Lange. 1991. “Government Partisanship, Labor Organization, and Macroeconomic Performance.” American Political Science Review 85 (2): 539-56. Armingeon, Klaus. 1983. Neo-korporatistische Einkommenspolitik. Eine vergleichende Untersuchung von Einkommenspolitiken in westeuropäischen Ländern in den 70er Jahren. Frankfurt: Haag & Herchen. 48

Armingeon, Klaus. 1986. “Formation and Stability of Neo-corporatist Incomes Policies: A Comparative Analysis.” European Sociological Review 2 (2): 138-47. Avdagic, Sabina. 2010. “When Are Concerted Reforms Feasible? Explaining the Emergence of Social Pacts in Western Europe.” Comparative Political Studies, 43 (5): 628-657. Avdagic, Sabina. 2011. “The Emergence of Social Pacts: Analysing Negotiation Process and Bargaining Outcomes.” In Social Pacts in Europe: Emergence, Evolution, and Institutionalization, edited by Sabina Avdagic, Martin Rhodes, and Jelle Visser, 45-60. Oxford: Oxford University Press. Avdagic, Sabina, Martin Rhodes, and Jelle Visser, eds. 2011. Social Pacts in Europe: Emergence, Evolution, and Institutionalization. Oxford: Oxford University Press. Baccaro, Lucio, and Marco Simoni. 2008. “Policy Concertation in Europe: Understanding Government Choice.” Comparative Political Studies 41 (10): 1323-1348. Baumgartner, Michael, and Alrik Thiem. 2017. “Often Trusted but Never (Properly) Tested: Evaluating Qualitative Comparative Analysis.” Sociological Methods & Research, online first, DOI https://doi.org/10.1177/0049124117701487. Berg-Schlosser, Dirk, Gisèle de Meur, Charles Ragin, and Benoît Rihoux. 2008. “Qualitative Comparative Analysis (QCA) as an Approach.” In Configurational Comparative Methods. Qualitative Comparative Analysis (QCA) and Related Techniques, edited by Benoît Rihoux and Charles Ragin, 1-18. London: Sage. Berger, Stefan, and Hugh Compston, eds. 2002. Policy Concertation and Social Partnership in Western Europe: Lessons for the 21st Century. New York: Berghahn Books. Culpepper, Pepper D., and Aidan Regan. 2014. “Why Don't Governments Need Trade Unions Anymore? The Death of Social Pacts in Ireland and Italy.” Socio-Economic Review 12 (4): 723-45. Crouch, Colin. 1993. Industrial Relations and European State Traditions. Oxford: Clarendon Press. CPDS. 2018. Comparative Political Data Set 1960-2016 (CPDS) by Klaus Armingeon, Virginia Wenger, Fiona Wiedemeier, Christian Isler, Laura Knöpfel, David Weisstanner, and Sarah Engler. Institute of Political Science, University of Berne, http://www.cpds-data.org/. Duşa, Adrian 2019. QCA with R: A Comprehensive Resource. Cham: Springer Nature. Ebbinghaus, Bernhard. 2010. “Reforming Bismarckian Corporatism: The Changing Role of Social Partnership in Continental Europe.” In A Goodbye to Bismarck? The Politics of Welfare Reforms in Continental Europe, edited by Bruno Palier, 255-278. Amsterdam: Amsterdam University Press. Ebbinghaus, Bernhard, and Jelle Visser. 2000. Trade Unions in Western Europe since 1945. London: Palgrave/Macmillan. Esping-Andersen, Gøsta. 1990. Three Worlds of Welfare Capitalism. Princeton, NJ: Princeton University Press. Eurofound 2012. Social Dialogue in Times of Global Economic Crisis. Irland: Dublin. Eurofound 2013. Industrial Relations and Working Conditions Developments in Europe 2012. Irland Dublin. Eurostat. Eurostat database, various statistics. Fajertag, Giuseppe, and Philippe Pochet, eds. 2000. Social Pacts in Europe: New Dynamics. Brussels: ETUI. Grote, Jürgen R., and Philippe C. Schmitter. 1999. “The Renaissance of National Corporatism: Unintended Sideeffect of European Economic and Monetary Union or Calculated Response to the Absence of European Social Policy?” Transfer 5 (1-2): 24-63. Guardiancich, Igor, and Oscar Molina, eds. 2017. Talking Through the Crisis: Social Dialogue and Industrial Relations Trends in Selected EU Countries. Geneva: ILO. Guyet, Rachel, David Tarren, and Claude-Emmanuel Triomphe. 2012. Social Dialogue in Times of Global Economic Crisis. Dublin: Eurofound. Hall, Peter A., and Daniel W. Gingerich. 2009. “Varieties of Capitalism and Institutional Complementarities in the Political Economy.” British Journal of Political Science 39 (3): 449-482. Hall, Peter A., and David Soskice, eds. 2001. Varieties of Capitalism: The Institutional Foundations of Comparative Advantage. New York, NY: Oxford University Press. Hassel, Anke. 2006. Wage Setting, Social Pacts and the Euro: A New Role for the State. Amsterdam: Amsterdam University Press. Hicks, Alexander. 1988. “Social Democratic Corporatism and Economic Growth.” Journal of Politics 50 (3): 677704. ICTWSS. 2015. ICTWSS Database, version 5.0. Amsterdam: Amsterdam Institute for Advanced Labour Studies (AIAS). http://www.uva-aias.net/en/ictwss ILO. Various Years. ILOSTAT Database. http://www.ilo.org/ilostat. Jahn, Detlef. 2016. “Changing of the Guard: Trends in Corporatist Arrangements in 42 Highly Industrialized Societies from 1960 to 2010.” Socio-Economic Review 14 (1), 47–71. Jahn, Detlef. 2017. “Distribution Regimes and Redistribution Effects During Retrenchment and Crisis: A Cui Bono Analysis of Unemployment Replacement Rates of Various Income Categories in 31 Welfare States.” Journal of European Social Policy. Online first. https://doi.org/10.1177/0958928717739249. Kenworthy, Lane. 2003. “Quantitative Indicators of Corporatism: A Survey and Assessment.” International Journal of Sociology 33 (3): 10-40. 49

Korpi, Walter. 1983. The Democratic Class Struggle. London: Routledge & Kegan Paul. Lehmbruch, Gerhard. 1967. Proporzdemokratie. Politisches System und politische Kultur in der Schweiz und in Österreich. Tübingen: J.C.B. Mohr. Lehmbruch, Gerhard, and Schmitter, Philippe C., eds. 1982. Patterns of Corporatist Policy-Making. London: Sage. Lindstrom, Nicole 2015. “Wither Diversity of Post-Socialist Welfare Capitalist Cultures? Crisis and Change in Estonia and Slovenia.” European Journal of Sociology 56 (1): 119-139. OECD. Various Years. Employment and Labour Market Statistics. Paris: OECD. http://Stats.oecd.org/Index.aspx. O’Donnell, Rory, Maura Adshead, and Damian Thomas. 2011. “Ireland: Two Trajectories of Institutionalization.” In Social Pacts in Europe, edited by Sabine Avdagic, Martin Rhodes and Jelle Visser, 89-119. Oxford: Oxford University Press. Ragin, Charles C. 1987. The Comparative Method: Moving Beyond Qualitative and Quantitative Strategies. Berkeley, CA: University of California Press. Ragin, Charles C. 2008a. Measurement Versus Calibration: A Set Theoretic Approach. In The Oxford Handbook of Political Methodology, edited by Janet M. Box-Steffensmeier, Henry E. Brady and David Collier, 174198. Oxford: Oxford University Press. Ragin, Charles C. 2008b. Redesigning Social Inquiry. Fuzzy Sets and Beyond. Chicago, IL: University of Chicago Press. Regan, Aidan. 2017. “Rethinking Social Pacts in Europe: Prime Ministerial Power in Ireland and Italy.” European Journal of Industrial Relations 23 (2): 117–33. Regini, Marino. 2003. “Tripartite Concertation and Varieties of Capitalism.” European Journal of Industrial Relations 9 (3): 251-263. Rhodes, Martin. 1997. Globalization, Labour Markets and Welfare States. A Future of 'Competitive Corporatism'? Florence: European University Institute. Scharpf, Fritz W. 1991. Crisis and Choice in European Social Democracy. Ithaca, NY: Cornell University Press. Schmitter, Philippe C. 1974. “Still the Century of Corporatism?” Revue of Politics 36 (1): 85-131. Schmitter, Philippe C., and Gerhard Lehmbruch, eds. 1981. Trends Toward Corporatist Intermediation. Beverly Hills: Sage. Schmitter, Philippe C., and Wolfgang Streeck. 1999. “The Organization of Business Interests. Studying the Associative Action of Business in Advanced Industrial Societies.” MPIfG Discussion Paper 99 (1). Schneider, Carsten Q., and Claudius Wagemann. 2010. “Standards of Good Practice in Qualitative Comparative Analysis (QCA) and Fuzz-Sets.” Comparative Sociology 9 (3): 397-418. Schneider, Carsten Q., and Claudius Wagemann. 2012. Set-Theoretic Methods for the Social Sciences. A Guide to Qualitative Comparative Analysis. Cambridge: Cambridge University Press. Siaroff, Alan. 1999. “Corporatism in 24 Industrial Democracies: Meaning and Measurement.” European Journal of Political Research 36 (2): 175-205. Thiem, Alrik. 2018. Advanced Functionality for Performing and Evaluating Qualitative Comparative Analysis. R Package Version 1.1-2. Traxler, Franz, Sabine Blaschke, and Bernhard Kittel. 2001. National Labour Relations in Internationalized Markets: A Comparative Study of Institutions, Change, and Performance. Oxford: Oxford University Press. Vaughan-Whitehead, Daniel, ed. 2015. The European Social Model in Crisis: Is Europe Losing Its Soul? Cheltenham, UK: Edward Elgar. Vaughan-Whitehead, Daniel, ed. 2018. Reducing Inequalities in Europe: How Industrial Relations and Labour Policies Can Close the Gap. Cheltenham, UK: Edward Elgar. Vis, Barbara. 2011. “Under Which Conditions Does Spending on Active Labor Market Policies Increase? An fsQCA Analysis of 53 Governments between 1985 and 2003.” European Political Science Review 3 (2): 229252. Visser, Jelle. 1987. In Search of Inclusive Unionism. Amsterdam: University of Amsterdam.

50

Transformations of Swiss neo-corporatism: From pre-parliamentary negotiations towards privileged pluralism in the parliamentary venue André Mach1, Frédéric Varone2 and Steven Eichenberger3 Abstract: Major economic peak-level associations, because of their various resources (in terms of membership, finance and institutional reconnaissance by public authorities) have become central political actors of the Swiss neo-corporatist regime. They were considered the dominant actors of the pre-parliamentary phase of the decision-making process (extra-parliamentary committees, consultation procedures), identified as the most important phase, whereas the Parliament only marginally modified the proposals of the Federal Council. However, since the beginning of the 1990s, the strategies of interest groups have profoundly changed, leading to a reconfiguration of the traditional neo-corporatist political regime toward a more pluralist system, in which interest groups more actively target the Parliament. Different factors explain these changes: the declining role of the pre-parliamentary phase, the revalorization of the Parliament, and the increasing role of the media. These changes have weakened the positions of traditional corporatist associations and favored the political rise of new citizen groups. They have also induced interest groups to develop new political strategies, privileging the parliamentary venue, especially the new permanent specialized committees. Despite the growing access of new citizen groups to the political system (pre-parliamentary and parliamentary venues), economic groups remain dominant in the domains of economic and social policies.

1

Introduction

Stressing the overall stability of the neo-corporatist character of the Swiss political system in a comparative perspective, Klaus Armingeon (2011) underlined the strong presence of economic groups – business interest associations (BIAs) and trade unions – in Parliament: “In 2010 the parliamentary commission on economy and taxes in the Swiss parliament – arguably one of the most important and powerful parliamentary commissions – is composed of members of leading representatives of interest groups representing the workforce, small and medium enterprises, farmers, employers and big business. If we use Lehmbruch’s criterion that the direct political representation of interest organizations in parliament is an indicator of corporatism, today Swiss

1 2 3

Université de Lausanne [email protected] Université de Genève [email protected] Université de Lausanne [email protected]

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 R. Careja et al. (eds.), The European Social Model under Pressure, https://doi.org/10.1007/978-3-658-27043-8_4

51

corporatism is in an extremely healthy state (Lehmbruch 1974, 1979)” (as cited in Armingeon 2011, 179). In another context, Binderkrantz and Christiansen (2015) also stressed the resilience of Danish neo-corporatism despite profound socio-economic and political changes. However, since the heydays of neo-corporatism during the 1970s, traditional European corporatist political systems, among which Switzerland, have been facing profound challenges. Among those are growing difficulties to reach compromises between business associations and trade unions in a context of social policy retrenchment, or the growing importance of postmaterialist values underlying the rise of new citizen groups (e.g. environmental and humanitarian groups) increasingly active at the political level. In addition, and concerning the Swiss case more specifically, the formalization of the pre-parliamentary phase of the legislative process and the revalorization of the Parliament (Sciarini 2014; Pilotti 2017) have also contributed to putting the traditional corporatist system under pressure. The aim of this contribution is to analyze in more detail how the traditional neo-corporatist structures have evolved over the last decades in Switzerland. Have they evolved towards a more pluralist system? Has the predominant position of business associations and trade unions during decision-making processes been challenged by citizen groups? Between the administrative and parliamentary venues, which is the most hospitable to economic groups? Are these actors more present in committees dealing with economic and social policy issues than citizen groups? After a first part presenting the main challenges facing traditional corporatist systems, we analyze the evolution of interest groups’ access to pre-parliamentary committees. Next, we focus on the parliamentary venue and study the evolution of MPs’ ties to interest groups, with a particular focus on members of the Committees for Economic Affairs and Taxation (CEAT) as well as the Committees for Social Security and Health (CSSH). We conclude that the Swiss system of interest intermediation has indeed evolved towards a stronger participation of citizen groups, but that major economic groups (BIAs and trade unions) still remain the most dominant players within the domain of economic policy.

2

Declining role of pre-parliamentary negotiations and revalorization of the Parliament: A challenge for Swiss neo-corporatism?

Interest groups, especially economic peak associations, have traditionally been considered as crucial and very influential political actors in Switzerland. In the context of an underdeveloped central state, weak national political parties and a weakly professionalized Parliament, major Swiss economic interest groups, because of their various resources (in terms of finance, membership, expertise and institutional reconnaissance by public authorities) and their early organization at the national level, became major and central political actors since the end of the 19th century. More precisely, interest groups appeared as the dominant actors of the pre-parliamentary phase of the decision-making process (extra-parliamentary committees, consultation pro52

cedures), identified as the most important phase, whereas the Parliament, with its militia character, only marginally modified the proposals of the Federal Council (Neidhart 1970; Kriesi 1980; Mach 2007). In this context, major interest groups, which had access to the pre-parliamentary phase, good contacts with the Federal Administration and were involved in the implementation of public policies, were clearly the dominant actors. These were the large peak level economic associations, business associations (USCI/Vorort – economiesuisse since 2000, USAM and UPS), linked to right-wing parties, and, to a lesser extent, trade unions (USS and CSCS) connected to the Social-Democratic Party (PSS) and social wing of the Christian-Democratic Party (PDC).4 However, there was a clear asymmetry in this corporatist structure, in which the trade unions had been only progressively integrated on a minority basis and in which business associations remained the dominant actors (Kriesi 1980; Katzenstein 1985). This power configuration remained largely stable since the Second World War for three major reasons: First, companies succeeded in setting up cohesive, well-coordinated and representative business associations; second, they maintained close connections to the dominant right-wing parties, especially the Radical-Democratic Party (PRD); and third, business associations developed technical expertise that helped them to decisively influence public policies (see P. Eichenberger and Mach 2011). Thanks to their representativeness, legitimacy and technical expertise, BIAs, among which the peak level associations, clearly remained the dominant actors in the decision-making process (P. Eichenberger and Mach 2011). Once they had been integrated, on a minority basis, into the corporatist structures in the 1930s, trade unions, as well as the Social-Democratic party, regularly tried to use the instrument of the popular initiative to promote their goals, and thus, to circumvent the power structure in which they remained in a minority position. However, they always lost these popular votes (e.g. co-determination initiative to give rights to employee representatives to sit on the boards of directors of companies in 1976, initiative to suppress banking secrecy in 1982, different initiatives to reduce the weekly working hours). Despite (or actually because of, as underlined by Neidhart (1970)) the existence of direct democratic instruments (especially the optional referendum), pre-parliamentary negotiations between the major economic associations have become the center of gravity of the decisionmaking process. In this context, the corporatist insiders with privileged access to the government and its administration were clearly the dominant actors in the “traditional model”. As in other small European neo-corporatist countries (Crepaz 1994; Rommetvedt et al. 2013), corporatist structures seem also to have lost influence in the Swiss context. Economic internationalization exacerbated the cleavage between export-oriented and sheltered industries.

4

We use the French abbreviations: ASB: Association suisse des banquiers (SwissBanking); USCI: Union suisse du commerce et de l’industrie (economiesuisse); USAM: Union suisse des arts et métiers; UPS: Union patronale suisse; USP: Union suisse des paysans; USS: Union syndicale suisse and CSCS: Confédération des syndicats chrétiens de Suisse (Travail.Suisse). 53

It also increased tensions between trade unions and BIAs, with the latter pushing for retrenchment policies (Mach 2006). This strain was accentuated through increasingly mediatized politics, which make political actors take more conflictive stances thus rendering classic closeddoor corporatist negotiations more difficult (Häusermann, Mach, and Papadopoulos 2004). On the basis of more than 300 interviews with political actors and a network analysis of the most important decision-making processes at the beginning of this century, Sciarini (2014) thus comes to the conclusion of the declining importance of both the pre-parliamentary phase and peak level BIAs and trade unions. At the same time, Parliament has become a less predictable, and thus more important, actor in Swiss politics. This is largely due to the rise of the Swiss People’s Party (UDC), which rendered parliamentary majorities less certain. Furthermore, citizen groups have become increasingly present within the Swiss society since the beginning of the 1970s. In fact, their number has more than doubled (Mach 2015). This largely parallels the evolution in other consolidated democracies, such as Denmark (Binderkrantz, Fisker, and Pedersen 2016) or the United States (Berry 1999). Even though the emergence of citizen groups represents a rather recent phenomenon, they have been able to consolidate organizationally (Zwicky 1993).

3

Seats in pre-parliamentary committees

Many authors have defined extra-parliamentary committees (EPC)5 as a kind of “militia administration” and as an unofficial system of representation existing in parallel to the Parliament (Germann 1981, 1985). Such “corporatist bodies” are very numerous and increased considerably after the Second World War. Their domain of competence is very specialized. Despite their central importance for the preparation of new legislation and the implementation of public policies, EPCs, as the heart of neo-corporatist negotiations, remained completely unregulated until the 1970s. It is only during the 1970s that a first ordinance regulating EPCs was adopted by the Federal Council. Under the impulses of the Parliament, which regularly criticized the composition of EPCs during the renewal of their composition, the ordinance was then reformed in 1996 and included in the new law on the administrative organization in 2008 (for more details, see Rebmann and Mach 2013). During this process of formalization concerning the composition and functioning of EPCs, which aimed to introduce more transparency in the functioning of EPCs, the number of EPCs was considerably reduced from around 300 in 1980 to less than 200 in 2000 (Table 1 below). 5

According to the official definition, extra-parliamentary committees are organs which assume tasks on behalf of executive authorities, but which are essentially composed of persons who are not civil servants. They have two main functions. First, they serve as a complement to the administration with regard to subjects about which the latter may not have enough knowledge and are thus considered as the best means to keep administrative costs down. Second, they can also be considered as “meeting places” for the Swiss elite, which help to find compromises between the major economic and political actors during the initial phase of the political decision-making process (Kriesi 1980; Germann 1985). 54

Table 1: Affiliations of members of extra-parliamentary committees to economic IGs and citizen groups (1980, 2000, 2010)1 1980 2000 2010 Trade unions Business interest associations

178 (24%)

108 (25%)

90 (24%)

417

208

183

(56%)

(49%)

(49%)

Citizen Groups

151 (20%)

109 (26%)

98 (26%)

Seats occupied by IGs

746 (100%)

425 (100%)

371 (100%)

Total number of committees

301

195

223

Source: S. Eichenberger (2017) on the basis of Swiss Elites in the 20th Century Database, University of Lausanne.

Table 1 displays the distribution of committee seats across trade unions, BIAs and citizen groups in 1980, 2000 and 2010. The total number of seats occupied by these three types of interest groups has declined (from 746 to 371), mirroring the decline in the number of committees (from 301 to 223). It must be noted though that the average number of seats occupied by interest groups per committee has also declined. Whereas each committee hosted, on average, 2.5 interest group representatives in 1980, this figure has declined to roughly 1.6 in 2010. Most importantly, Table 1 shows that the share of interest group seats occupied by citizen groups has indeed increased, from 20% in 1980 to 26% in 2000-10. This disguises, however, a decline in absolute terms: citizen groups occupied 151 seats in 1980 and 98 in 2010. This decline was much stronger in the case of BIAs and trade unions, which “lost” more than half of their seats. Nevertheless, it must be stressed that economic groups still occupy the vast majority of seats (i.e. 73%) attributed to interest groups in 2010. To some extent at least, the evolution of the distribution of EPC seats across group types reflects the increasing number of citizen groups within the interest group population (Mach 2015). The Swiss system of interest intermediation has thus adapted to changes within the interest group population. Of course, this does not mean that economic and citizen groups enjoy equal access to extra-parliamentary committees. But, from a neo-corporatist perspective the integration of citizen groups, which do not hold a monopoly of representation over certain societal interests, seems somewhat puzzling. Alternatively, and from a pluralist perspective, it might be argued that access to EPCs does not actually depend on representative monopolies, but more so on expertise, and that the rise and consolidation of citizen groups should accordingly translate into more (but not necessarily equal) access.

1

It should be noted that ‘ad hoc’ EPCs, which often play a crucial role in drafting a first version of legislative proposals, are not included in this table; only permanent EPCs have been considered. Neither does the table include committee seats occupied by institutional and occupational groups. 55

We then looked at the numbers of mandates in EPCs occupied by members of the executive committee of the seven major economic associations, including the elected members as well as the major paid officials of these associations (Table 2 below). The post-war period until the 1990s, covering 1957 and 1980, clearly represents the heyday of Swiss neo-corporatism, where the number of EPCs increased considerably and where the representatives of economic groups were the most numerous. During the second half of the century, the concentration of economic groups is strongest within some central EPCs dealing with social and economic policies developed by the Federal Department of Economic Affairs (DFE) and the Federal Department of Home Affairs (DFI). After 1980, the number of seats held by representatives of the peak level economic groups decreased substantially, to a larger extent than the general decrease in the number of EPCs, underlining again the declining role of economic groups within EPCs.7 There are two exceptions. The number of seats held by representatives of the Swiss Industry and Trade Association (USAM) has only slightly declined (from 19 to 17). And the number of seats occupied by representatives of the Confédération des syndicats chrétiens de Suisse (CSCS/Travail.Suisse) has actually increased during the 2000-10 period. In the former case, this might be related to the structurally weak position, in times of economic liberalization, of a BIA representing small and medium-sized enterprises rather focused on the domestic market. In order to avert a too strong and rapid liberalization process, this organization might have faced an additional incentive to remain represented within EPCs. In the latter case, this might simply be due to the fact that the CSCS merged with another peak level trade union (Confederation of Swiss Salaried Employees Organizations) in order to become Travail.Suisse in 2002. Table 2: Number of mandates in EPCs occupied by executive committee members of peak level economic groups (seats occupied by paid officials) Business interest associations

Trade unions

ASB

UPS

USCI

USAM

USP

USS

CSCS

Total

1957

12 (6)

24 (16)

39 (24)

28 (12)

56 (23)

54 (22)

18 (9)

231 (112)

1980

11 (5)

28 (14)

36 (29)

41 (22)

44 (21)

75 (38)

26 (11)

261 (140)

2000

4 (2)

22 (15)

20 (14)

19 (17)

14 (11)

29 (23)

17 (9)

125 (91)

2010

1 (1)

17 (15)

13 (9)

17 (17)

8 (6)

18 (10)

24 (14)

98 (72)

Total

28 (14)

91 (60)

108 (76)

105 (68)

122 (61)

176 (93)

85 (43)

715 (415)

Source: Swiss Elites in the 20th Century Database, University of Lausanne. Acronyms: See Footnote 1.

7

Besides the political will to formalize EPCs and to make them more transparent, the declining importance of the pre-parliamentary phase and of EPCs are also related to the declining capacity of corporatist actors to find compromises on policy reforms, notably concerning social policies, but also in other fields (Häusermann, Mach, and Papadopoulos 2004; Sciarini 2014). 56

4

Focus on pre-parliamentary committees in the economic and social policy domains

Finally, we propose a disaggregated view of the evolution of committee seats, in committees affiliated to the DFE and the DFI. These two departments are responsible for economic and social policy. We can thus test whether a pluralization has also taken place within these policy domains. As can be seen on Figure 1, the evolution of EPCs mandates in the economic and social policy domains is different than the one observed at an aggregate level (see previous section). At an aggregate level, the share of seats attributed to citizen groups has increased by 6 percentage points (see Table 1). At the level of economic policy – that is, within the committees affiliated to the DFE – the share of seats attributed to citizen groups has only increased by 2 percentage points, and 3 percentage points in the case of the committees affiliated to the DFI. Thus within committees affiliated to both the DFE and DFI the share of seats occupied by citizen groups has remained relatively stable. This stability must also be seen in light of a rather strong increase in the share of citizen group seats within committees affiliated to the Federal Department of Foreign Affairs (DFAE) and the Department of the Environment, Transport, Energy and Communications (DETEC) (not depicted in Figure 1). Citizen groups occupied roughly 20% of all DETEC seats in 1980, but 35% in 2010. Also, within committees affiliated to the DFI we mostly observe a shift in access from BIAs to trade unions. Whereas BIAs occupied 51% of all committee seats affiliated to the DFI in 1980, this share had decreased to 39% by 2010. Inversely, trade unions occupied 21% of all seats in 1980, as compared to 29% in 2010.

70% 60% 50% 40% 30% 20% 10% 0% Trade unions

BIAs

Citizen groups

Trade unions

DFE

BIAs

Citizen groups

DFI 1980

2000

2010

Figure 1: Share of DFE and DFI committee seats occupied by trade unions, BIAs, and citizen groups, 1980, 2000, and 2010 Source: S. Eichenberger (2017) on the basis of Swiss Elites in the 20th Century Database, University of Lausanne. 57

The pluralization of extra-parliamentary committee seats must hence be nuanced as the share of seats occupied by citizen groups has remained fairly stable within committees affiliated to both the DFE and the DFI. This might reflect a certain “resilience” of neo-corporatism within the economic and social policy domains. Neo-corporatist negotiations usually took place within committee affiliated to the DFE and DFI. The stability observed within the EPCs affiliated to these departments might reflect a certain neo-corporatist heritage. In these departments, civil servants might still turn to BIAs and trade unions, quasi by default, when developing public policies. And this might lead to the exclusion of citizen groups working within these policy domains. On the other hand, it might be argued that citizen groups remain marginally invested in economic and social policy. That is, the stability observed might simply reflect a certain stability within the interest group population concerned by economic and social policy.

5

Formal ties to elected Members of the Parliament

Traditionally considered as a less important phase in the policy process, in comparison to the pre-parliamentary negotiations, the Parliament underwent some institutional reforms since the beginning of the 1990s that reinforced its role. Despite the negative popular vote against the reform of the Parliament in 1992, except for the part about the creation of specialized parliamentary commissions, different authors have underlined the increasing role of the Parliament in the Swiss decision-making process (Lüthi 1997; Linder 1998; Jegher 1999; Sciarini 2014; Pilotti 2017). Different reasons explain this change. First, the rising role of Parliament is partly due to the partial institutional reform of 1992, based on the replacement of ad hoc parliamentary committees by permanent and specialized committees. This means that MPs have become increasingly specialized and competent in some policy fields, and might thus be less dependent on the expertise and pressure of interest groups. Second, the increasing difficulty of economic groups to find compromise during the pre-parliamentary phase has led to a more important role of the debates in Parliament. Deadlocks in the pre-parliamentary phase have often been overcome during the parliamentary debates (Kriesi and Jegen 2000 on energy policy; Häusermann, Mach, and Papadopoulos 2004 on social policy; Fischer 2005; Sciarini 2014). Third, despite the popular refusal of the major parts of the institutional reform of 1992 concerning the improvements of the remunerations of MPs, the Parliament, through different adjustments not subject to an optional referendum, increased the remunerations of MPs. Thus, the militia character of the Swiss Parliament has clearly diminished since the beginning of the 1990s. MPs have become more professional, and less dependent on other revenues stemming from other professional activities, such as member of cantonal or communal executive authorities, or membership in boards of directors, or paid positions in interest groups. It is not rare to find real professional MPs, without any main profession besides their political mandates (Pilotti 2017; Sciarini et al. 2017). The few studies on Swiss MPs (Gruner and Frei 1966; Kerr 1981; 58

Wiesli 2003) all emphasized historically the strong dependence of MPs with respect to economic interests, especially through leading positions in business associations and trade unions, or through memberships on the boards of directors of companies. Finally, the increasing polarization in the Parliament, related to the electoral success of the Swiss People Party (UDC), rendered the parliamentary debates more uncertain. Among the right-wing parties, the traditional cohesion of the ‘bourgeois bloc’, linking the main BIAs and right-wing political parties, was called into question because of the increasing divisions between BIAs, but also through the electoral success of the populist-conservative UDC to the detriment of the Radical-Democratic Party (PRD) and the Christian-Democratic Party (PDC). The PRD, which maintained very close links to business circles, lost much of its leadership in the political arena. This change in the power balance undermined the traditional channel of influence of BIAs via the PRD, rendering their lobbying activities more difficult. The revalorization of the Parliament should be nuanced for internationalized policy issues. The international literature generally underlines, on the contrary to what is argued in this section, the weakening role of national parliaments in the context of the increasing role of international organizations with a supranational dimension. This also apply to Switzerland: for “internationalized policy process”, meaning policy reforms related to international constraining agreements or norms (bilateral agreements with the EU or international pressure), the role of the Parliament is effectively weakened, whereas the government and its administration has been reinforced (Mach, Häusermann, and Papadopoulos 2003; Sciarini 2014; Gava and Varone 2014). In the context of the declining importance of the pre-parliamentary phase and the growing role of the Parliament combined with the increasing importance of the media, interest groups had to adjust their strategies of influence toward political actors. Similarly, to what happened in some Scandinavian neo-corporatist countries (see Öberg et al. 2011; Rommetvedt et al. 2013), Swiss interest groups have progressively modified their strategies of influence in direction of the Parliament. In relation to the growing diversity in interest group landscape, it seems that MPs are increasingly asked to occupy leading positions in various interest groups. This could be explained by the revalorization of the Parliament and the intensification of lobbying activities targeting the Parliament. On the basis of the official register of MPs private interests it is possible to document formal connections of MPs with interest groups.8

8

MPs’ private interests are accessible since the mid-1980s in the Registry of Interest Ties (see www.parlament.ch and particularly Gava et al. 2017, for a general overview). 59

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

66%

65%

70%

32%

31%

2%

4%

3%

1995 (N=270)

1999 (N=318)

2003 (N=582)

76%

76%

75%

21%

22%

22%

3%

2%

2%

2007 (N=866)

2011 (N=921)

2015 (N=1051)

27%

Unions

Business

Citizen

Figure 2: Evolution MPs’ ties to business associations, trade unions and citizen groups Source: S. Eichenberger (2017).

As can be seen in Figure 2, the number of ties to BIAs, trade unions and citizen groups has increased from 270 in 1995 to 1051 in 2015. It must be noted that the increase can be partially explained by stricter rules concerning the declaration of ties, which entered into force in December 2003. Particularly the more than twofold increase of ties between 1999 and 2007 must be, to a large extent, due to these stricter rules.9 Nevertheless, over the entire period we can observe an increase in the number of ties held by MPs. Between 1995 and 1999, an 18% increase in the total number of ties can be observed, and between 2007 and 2015 the total number of ties has increased by 21%. This clearly reflects the growing importance of the Parliament in Swiss decision-making processes. S. Eichenberger and Mach (2017) have further shown that MPs often develop ties in function of their membership in legislative committees. That is, members of the Environment Committees are more likely to develop ties to interest groups working on environmental policy than the members of other committees. This further illustrates how interest groups have increased their lobbying activities within Parliament. If we focus on the distribution of ties across group types, and particularly its evolution, we can observe a certain stability. At first sight, it appears as if the share of citizen groups has considerably increased. However, this increase intervenes between 1999 and 2007, when Parliament enacted the stricter declaration rules. In fact, MPs were required to declare all ties, not just those which they deemed “important”. It seems likely that ties to politically inactive citizen groups constituted the lion’s share of those ties which had not yet been disclosed.10 Thus the 9

In fact, these stricter rules of declaration had already been anticipated by the offices of both Councils in April 2001. Also, the registry of ties is updated at the beginning of each year. In order to isolate any effects related to these stricter declaration rules, both when studying the evolution of the total number of ties as well as the distribution across group types, we must limit ourselves to a comparison of the periods reaching from 1995 to 1999, and from 2007 to 2015. 10 Between 2000 and 2004, a fivefold increase in the number of ties to leisure groups has taken place. 60

increase in the share of ties to citizen groups should not be interpreted as an actual increase in the share of citizen groups actually seeking to influence public policy. Nevertheless, it must be underlined that the number of ties to citizen groups has starkly increased since the 1990s. And ties to citizen groups represented the majority of all ties already in 1995.11 Moreover, citizen groups clearly enjoy more access to Parliament than to extra-parliamentary committees. In 2010, citizen groups accounted for 26% of all EPC seats occupied by interest groups (see Table 1), but for 76% of all ties to MPs (see Figure 3). The increasing role of the Parliament, combined with the decline of the pre-parliamentary phase, has thus “pluralized” and opened up access to the decision-making process to a wider spectrum of interest groups, and thus weakened the traditional insiders. Decision-making takes more and more place within Parliament, which is relatively more hospitable to citizen groups. Even though economic groups have increased their ties to MPs, thus updating their lobbying activities to the increased importance of Parliament, this has not led to a crowding out of citizen groups. In other words, economic groups’ heavier investment in the parliamentary venue has not allowed them to obtain as dominant a position as the one previously enjoyed within the administrative venue. The sheer number of interest groups present within Parliament, as well as the increasing presence of citizen groups, most of which cannot be considered to retain a monopoly of representation, does not square well with a neo-corporatist understanding of policy-making, according to which only a handful of mostly economic groups should play an important role. The declining role of the pre-parliamentary phase does not mean that peak-level economic groups do not play any role. They remain major actors (in terms of finance or membership), but they had to adjust their strategies towards political actors. Table 3 below clearly indicates a growing presence of MPs sitting on the executive committees of peak level economic associations during the recent period. This is particularly true for the USAM and USP, but also for the trade-unions (USS and CSCS, which became Travail.Suisse in 2002); however, this does not apply to USCI-Economiesuisse and UPS. As underlined by Armingeon (2011), we can see a displacement in the involvement of corporatist actors from the pre-parliamentary phase to the Parliament. So, contrary to the decline in the number of mandates on the board of directors of companies, which can be interpreted as a professionalization of the Parliament, the involvement of MPs in the governing bodies of corporatist actors has on the contrary increased. It is also interesting to note that the associations, which were the most affected by liberal economic reforms of the 1990s (farmers – USP, Small and medium enterprises – USAM, and employees organized by the trade unions), are also those with the highest number of MPs in their executive committee. This can again be interpreted as an attempt of these associations to compensate for their perceived loss of influence through closer ties to MPs.

11

Even though this must be nuanced since many ties to citizens actually do not represent ties to politically active interest groups (S. Eichenberger 2017, 112). 61

Table 3: Number of MPs sitting in the executive committee of peak level economic associations Business interest associations

Trade unions

UPS

USCI

USAM

USP

USS

CSCS

Total

1957

1

1

2

8

6

2

20

1980

3

-

4

5

3

1

16

2000

3

1

6

8

3

2

23

2010

3

1

7

4

5

2

22

Total

10

3

19

25

17

7

81

Source: S. Eichenberger (2017) on the basis of Swiss Elites in the 20th Century Database, University of Lausanne. Acronyms: see Footnote 1.

This is further illustrated if we focus on those groups holding most accreditations to the Federal Palace.12 As can be seen in Table 4, the major peak level economic groups, the usual suspects so to speak, occupy the top four positions of groups holding most accreditations to the Federal Palace in 2015. Just to make this clear, the USCI (economiesuisse), USAM, USS and USP could rely on a total of 37 lobbyists enjoying free access to the Parliament’s anti-chambers. This represents roughly 10% of all accreditations handed out by MPs in 2015.13 At the same time, however, three citizen groups (Pro Natura, ASLOCA, and the WWF) also occupy one of the ten highest ranks. This suggests that access is not so much related to representative monopolies, but more to interest groups’ capacity to provide MPs with relevant political and technical information. Table 4: Interest groups holding most accreditations to the Federal Palace in 2015 (top ten)14 Rank

Interest Group

Accreditations

1

economiesuisse

10

2

union suisse des arts et métiers (USAM)

9

2

Union Syndicale Suisse (USS)

9

2

Union Suisse des Paysans (USP)

9

5

Pro Natura

5

6

Travail.Suisse

4

6

Association Suisse des Locataires (ASLOCA)

4

8

WWF

3

8

Union de propriétaires de maison (HEV)

3

8

swisscleantech

3

Note: Economic (citizen) groups are held in bold (normal) characters. Source: S. Eichenberger (2017). 12

Each MP can hand out two permanent access badges to “any two persons who wish to have access to the parts of the Parliament Building that are not accessible to the public” (Parliament Act, Art. 69.2). The accreditation lists had been made publicly available on the parliamentary website only since the beginning of 2012 (Schweiz am Sonntag 2012). 13 It must be noted that not all MPs make us of the opportunity to hand out such access badges. In fact, in 2015 there were 94 free slots. 14 Economic groups are held in bold characters. 62

6

Focus on parliamentary committees in the economic and social policy domains

Since 1992, both chambers of the Federal Assembly have nine permanent, specialized legislative committees and two permanent, supervisory committees.15 How have interest groups adapted their strategies towards the parliamentary venue in this new institutional context? Since the reform, decisive debates about legislative proposals by the government are considered to take place in these specialized committees, composed according to the strength of the different parliamentary groups. The plenary debates generally do not profoundly modify the solutions adopted by the specialized committees. It is thus crucial for interest groups to intervene in the deliberations of these committees. How are interest groups’ ties distributed across group types within different specialized parliamentary committees? For economic groups, two committees are of particular importance in both houses: The Committee for Economic Affairs and Taxation (CEAT) and the Committee for Social Security and Health (CSSH), often considered as the most important specialized committees (see Figures 3 and 4 below). First, a general increase in the number of ties held by CEAT and CSSH members can be observed, with the exception of the 2011-15 period for the CEAT. For instance, between 1995 and 1999, the number of ties held by CEAT members increased by 48% (from 40 to 59). This increase might be partially explained by CEAT members simultaneously holding mandates within other legislative committees (e.g. the Environment Committees). However, if the focus is put on CEAT members of the National Council, which usually do not occupy several committee seats simultaneously, we can still observe an increase in the number of ties. In fact, between 1995 and 1999, the number of ties held by CEAT members of the National Council increased by 81% (from 11 to 20). This shows that the increase observed at an aggregate level is not solely due to an increase in the number of ties held by the members of other legislative committees (most notably the Environment Committees). Secondly, economic groups account for a larger share of ties to CEAT members than to members of the entire Parliament: whereas they account for 34% of all ties to MPs, they account for roughly 50% of all ties to CEAT members (of both the National Council and the Council of States). This shows that the importance of citizen groups clearly varies across policy domains. When it comes to economic policy, economic groups remain, in terms of numbers at least, the most important actors, as it must be borne in mind that ties to citizen groups often involve groups which are not, or only weakly, politically active. Thirdly, except for the 1995-99 period, the share of CEAT members’ ties held by economic groups has remained stable, if not even slightly increased. Whereas economic groups accounted for 30% of all ties to CEAT members in 2007, they accounted for 38% in 2015. In absolute 15

In fact, permanent committees have existed since 1979, but their resources were limited, and ad hoc parliamentary committees remained very important. 63

terms, the number citizen groups holding ties to CEAT members has actually declined between 2007 and 2015 (from 112 to 85). This runs counter to the evolution observed at the aggregate level, suggesting that economic groups have adapted their lobbying activities to the growing importance of legislative committees, and the proportion of connections of economic groups to the CEAT actually increased more than for citizen groups. Nevertheless, it must be underlined that economic groups are not amongst themselves within the Economic Affairs Committees, but must face the company of numerous citizen groups. This is notably different from the classic pre-parliamentary corporatist structures with an exclusive participation of economic groups. In fact, our results suggest that both the CEAT and CSSH have never been the exclusive domain of economic groups. Citizen groups accounted for considerable share of ties already in 1995. Finally, the evolution within the Committees for Social Security and Health roughly parallels the evolution observed at the aggregate level. Citizen groups account for the vast majority of ties, and there has also been a general increase in the number of ties. Yet the distribution across group types has remained fairly stable over the entire period analyzed here. In spite of economic groups having supplemented their pre-parliamentary lobbying activities with a stronger investment into the Parliament, citizen groups appear to represent a force to be reckoned with. The number of groups present within both committees, as well as citizen groups’ capacity to increase their ties (in absolute terms) in spite of economic groups’ increased investment into the legislative venue, suggest that access does not solely depend on interest groups’ representational monopolies, but more so on expertise and legitimacy. Citizen groups, which to a certain extent are also illustrative of economic groups’ increasing difficulties to encompass various societal interests, have undergone a process of organizational consolidation since the 1980s.

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

21

42

82

112

33

44

108

85

61

47

18 15 1

2

6

3

4

6

1995 (N=40)

1999 (N=59)

2003 (N=121)

2007 (N=159)

2011 (N=173)

2015 (N=138)

Unions

Business

Citizen

Figure 3: Ties held by members of the Committees for Economic Affairs and Taxation (CEAT) Source: S. Eichenberger 2017. 64

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

37

47

6 3

9 2

1995 (N=46)

1999 (N=58)

77

135

136

154

24

36 7

51

2

32 8

2003 (N=103)

2007 (N=175)

2011 (N=179)

2015 (N=211)

Unions

Business

6

Citizen

Figure 4: Ties held by members of the Committees for Social Security and Health (CSSH) Source: S. Eichenberger 2017.

In addition, they have also reinforced their financial resources, particularly in the case of environmental groups. This consolidation allows them to acquire a broader access to decision-making processes. Moreover, citizen groups often represent concerns shared by large segments of the population. In more public venues, such as the parliament, citizen groups are thus more likely to be granted access. To be clear, this is not to say that citizen groups and economic groups nowadays enjoy equal access to the various decision-making venues. Financial and expertise resources are still unequally distributed across different group types, giving BIAs and trade unions a certain advantage over citizen groups. For instance, the USCI, USAM, USP and USS are represented among the interest groups holding most ties to CEAT members in 2015. The Swiss Farmers Union (USP) alone held five ties to CEAT members of both Councils.

7

Conclusion

As often in Switzerland, there is no brutal departure from traditional political structures. Changes follow a much more incremental dynamic, through progressive adjustments of existing structures and institutions. But, nevertheless, some changes have occurred concerning the system of interest intermediation and interest groups’ access to the decision-making process, amending the traditional neo-corporatist model. We come to nuanced conclusions. On the one hand, we can observe the growing presence of citizen groups in the various venues of the decision-making process, partly challenging traditional economic groups. On the other hand, business associations and trade unions still occupy dominant positions in some key specialized committees dealing with economic and social policy issues. While at the aggregate level, we can observe a growing access of citizen groups, 65

when we look at more specialized committees dealing with economic and social policies (in the pre- and the parliamentary venue), the overrepresentation of traditional economic groups remains stable. With the revalorization and growing professionalization of the parliament, the parliamentary venue has gained in importance. This venue is by definition more hospitable to citizen groups defending general causes and objectives. But, economic interest groups have also adjusted their political strategies, and have reinforced their efforts to be present and represented within the parliamentary venue, nuancing thus the strengthening of citizen groups. The corporatist heritage of Switzerland, marked by the dominant/entrenched positions of traditional economic groups, has certainly limited the capacity of citizen groups to access the decision-making process. However, the existence of direct democratic instruments, facilitating the political agenda-setting and the access of citizen groups to the decision-making process, has probably favored the recognition of citizen groups in Swiss politics (see Christiansen, Mach, and Varone 2017, who compare the Swiss and Danish cases). Such changes lead us to speak of a progressive transition from a neo-corporatist system toward a “privileged” or “biased pluralism” where citizen groups have gained more access to the decision-making process, more in the parliamentary than the administrative venue, but where tradi-tional economic groups, concentrating important resources (in terms of membership, finance and expertise) remain largely privileged actors in some key committees.

References Armingeon, Klaus. 2011. “A Prematurely Announced Death?” In Switzerland in Europe: Continuity and Change in the Swiss Political Economy, edited by Christine Trampusch and André Mach, 165–185. London: Routledge. Berry, Jeffrey M. 1999. The New Liberalism: The Rising Power of Citizen Groups. Washington DC: Brookings Institution Pr. Binderkrantz, Anne Skorkjær, and Peter Munk Christiansen. 2015. “From Classic to Modern Corporatism. Interest Group Representation in Danish Public Committees in 1975 and 2010.” Journal of European Public Policy 22 (7): 1022–1039. Binderkrantz, Anne Skorkjær, Helene Marie Fisker, and Helene Helboe Pedersen. 2016. “The Rise of Citizen Groups? The Mobilization and Representation of Danish Interest Groups, 1975–2010.” Scandinavian Political Studies 39 (4): 291–311. Christiansen, Peter Munk, André Mach, and Frédéric Varone. 2017. “How Corporatist Institutions Shape the Access of Citizen Groups to Policy-Makers: Evidence from Denmark and Switzerland.” Journal of European Public Policy 25 (4): 1–20. Crepaz, Markus ML. 1994. "From Semisovereignty to Sovereignty: The Decline of Corporatism and Rise of Parliament in Austria" Comparative Politics 27 (1): 45–65. Eichenberger, Pierre, and André Mach. 2011. “Organized Capital and Coordinated Market Economy.” In Switzerland in Europe: Continuity and Change in the Swiss Political Economy, edited by Christine Trampusch and André Mach, 63–81. London: Routledge. Eichenberger, Steven. 2017. “Interest Groups’ Access to Policy Venues: The Rise of Citizen Groups in Switzerland?” PhD diss., University of Lausanne. Eichenberger, Steven, and André Mach. 2017. “Formal Ties between Interest Groups and Members of Parliament: Gaining Allies in Legislative Committees.” Interest Groups & Advocacy 6 (1): 1–21. Fischer, Alex. 2005. Die Auswirkungen der Internationalisierung und Europäisierung auf Schweizer Entscheidungsprozesse: Institutionen, Kräfteverhältnisse und Akteursstrategien in Bewegung. Politikanalyse, Bd. 5. Zürich: Rüegger. 66

Gava, Roy, and Frédéric Varone. 2014. “The EU’s Footprint in Swiss Policy Change: A Quantitative Assessment of Primary and Secondary Legislation (1999-2012).” Swiss Political Science Review 20 (2): 216–22. Gava, Roy, Frédéric Varone, André Mach, Steven Eichenberger, Julien Christe, and Corinne Chao-Blanco. 2017. “Interests Groups in Parliament: Exploring MPs’ Interest Affiliations (2000-2011).” Swiss Political Science Review 23 (1): 77–94. Germann, Raimund E. 1981. Ausserparlamentarische Kommissionen: die Milizverwaltung des Bundes: avec un résumé français. Publikationen des Schweizerischen Nationalfonds aus den Nationalen Forschungsprogrammen, Bd. 8. Bern: Haupt. Germann, Raimund E. 1985. Experts et commissions de la Confédération. Publikationen des Schweizerischen Nationalfonds aus den nationalen Forschungsprogrammen. Lausanne: Presses polytechniques romandes. Gruner, Erich, and Karl Frei. 1966. Die schweizerische Bundesversammlung. Helvetia politica, Series A vol. 1-2, 4. Bern: Francke. Häusermann, Silja, André Mach, and Yannis Papadopoulos. 2004. “From Corporatism to Partisan Politics: Social Policy Making under Strain in Switzerland.” Swiss Political Science Review 10 (2): 33–59. Jegher, Annina. 1999. Bundesversammlung und Gesetzgebung: der Einfluss von institutionellen, politischen und inhaltlichen Faktoren auf die Gesetzgebungstätigkeit der Eidgenössischen Räte. Berner Studien zur Politikwissenschaft, Bd. 7. Bern: Haupt. Katzenstein, Peter J. 1985. Small States in World Markets: Industrial Policy in Europe. Cornell Studies in Political Economy. Ithaca; London: Cornell University Press. Kerr, Henry. 1981. Parlement et société en Suisse. Saint-Saphorin: Georgi. Kriesi, Hanspeter. 1980. Entscheidungsstrukturen und Entscheidungsprozesse in der Schweizer Politik. Frankfurt: Campus-Verlag. Kriesi, Hanspeter, and Maya Jegen. 2000. “Decision-Making in the Swiss Energy Policy Elite.” Journal of Public Policy 20 (1): 21–53. Linder, Wolf. 1998. Schweizerische Bundesversammlung: Ein Aktives Gesetzgebungsorgan. Eine Empirische Untersuchung Des Gesetzgebungsprozesses in den Jahren 1995-97. Bern: Institut für Politikwissenschaft. http://www.parlament.ch/d/dokumentation/berichte/weitere-berichte-und-studien/documents/ed-pa-gesetzgebungsprozess.pdf. Lüthi, Ruth. 1997. Die Legislativkommissionen der Schweizerischen Bundesversammlung: institutionelle Veränderungen und das Verhalten von Parlamentsmitgliedern. Berner Studien zur Politikwissenschaft 4. Bern [etc.]: P. Haupt. Mach, André. 2006. La Suisse entre internationalisation et changements politiques internes: la législation sur les cartels et les relations industrielles dans les années 1990. Analyse des politiques publiques, vol. 6. Zürich: Rüegger. Mach, André. 2007. “Interest Groups.” In Handbook of Swiss Politics, edited by Ulrich Klöti, Peter Knoepfel, and Hanspeter Kriesi, 359–380. Zürich: Neue Zürcher Zeitung Verlag. Mach, André. 2015. Groupes d’intérêt et pouvoir politique. Le savoir suisse Politique 111. Lausanne: Presses polytechniques et universitaires romandes. Mach, André, Silja Häusermann, and Yannis Papadopoulos. 2003. “Economic Regulatory Reforms in Switzerland: Adjustment without European Integration, or How Rigidities Become Flexible.” Journal of European Public Policy 10 (2): 301–18. Neidhart, Leonhard. 1970. Plebiszit und pluralitäre Demokratie: eine Analyse der Funktion des schweizerischen Gesetzesreferendums. Helvetia politica. Series B, vol. 5. Bern: Francke. Öberg, PerOla, Torsten Svensson, Peter Munk Christiansen, Asbjørn Sonne Nørgaard, Hilmar Rommetvedt, and Gunnar Thesen. 2011. “Disrupted Exchange and Declining Corporatism: Government Authority and Interest Group Capability in Scandinavia.” Government and Opposition 46 (3): 365–91. Pilotti, Andrea. 2017. Entre Démocratisation et Professionnalisation: Le Parlement Suisse et ses Membres de 1910 à 2016. Terrains des sciences sociales. Zurich: Seismo. Rebmann, Frédéric, and André Mach. 2013. “Commissions Extra-Parlementaires Fédérales.” In Manuel d’administration publique suisse, edited by Andreas Ladner, Jean-Loup Chappelet, Yves Emery, Peter Knöpfel, Luzius Mader, Nils Soguel and Frédéric Varone. Lausanne: Presses polytechniques et universitaires romandes. Rommetvedt, Hilmar, Gunnar Thesen, Peter Munk Christiansen, and Asbjørn Sonne Nørgaard. 2013. “Coping With Corporatism in Decline and the Revival of Parliament: Interest Group Lobbyism in Denmark and Norway, 1980–2005.” Comparative Political Studies 46 (4): 457–485. Schweiz am Sonntag. 2012. “Lobbyisten im neuen Parlament,” February 25. https://www.schweizamwochenende.ch/aktuell/lobbyisten-im-neuen-parlament-131048073. Sciarini, Pascal. 2014. “Eppure Si Muove: The Changing Nature of the Swiss Consensus Democracy.” Journal of European Public Policy 21 (1): 116–132.

67

Sciarini, Pascal, Frédéric Varone, Giovanni Ferro-Luzzi, Fabio Cappelletti, Vahan Garibian, and Ismail Muller. 2017. Étude sur le Revenu et les Charges des Parlementaires Fédéraux. Université de Genève: Services du Parlement. Wiesli, Reto. 2003. “Switzerland: The Militia Myth and Incomplete Professionalization.” In The Political Class in Advanced Democracies, edited by Jens Borchert, 374–92. Oxford: Oxford University Press. Zwicky, Heinrich. 1993. “Umweltaktivierungen in den 80er-Jahren.” Annuaire suisse de science politique 33: 185203.

68

Aufstieg und Auflösung der Deutschland AG (1896-2010): Unternehmensverflechtung in Deutschland, Frankreich und in den USA Paul Windolf1

Abstract: In the late 19th century, a relatively dense network arose among major public corporations. The network developed as an important economic institution of the emerging managerial capitalism. It was created by multiple directors who held many board positions (big linkers). The network offered an opportunity structure for the coordination of market strategies (regulated competition) and for the self-monitoring of management. The structure of the corporate network is analyzed for the period from 1896 to 2010 for Germany, France, and the United States. Network density was particularly high in Germany during the interwar period. During the 1990s network density continuously declined, and by 2010 the corporate networks were effectively dismantled, with little discernible difference between Germany, France, and the United States.

1

Produktionsregime

In einem Aufsatz über die Entstehung des deutschen Produktionsmodells vertritt Werner Abelshauser die These, dass die ökonomischen Institutionen dieses Regimes bereits vor dem Ersten Weltkrieg geschaffen wurden und bis in die 1990er Jahre weitgehend stabil geblieben sind.2 Zu diesen Institutionen zählen die Universal-banken, die Mitbestimmung, die Qualifizierung der Arbeitskräfte im dualen System und die Koordination von Produkt- und Arbeitsmärkten durch autonome Verbände. Im Kaiserreich habe ein „dramatischer Umbruch des deutschen Produktionsregimes“ stattgefunden. Diese Reorganisation habe sich aber nicht an traditionellen Organisationsformen der Wirtschaft orientiert. Vielmehr sei ein „neuer Realtypus von Kapitalismus“ entstanden, der die Unternehmen an die Verwissenschaftlichung der Produktion und an eine wissensbasierte Wirtschaft angepasst habe (Abelshauser 2001, 509510). Ein gemeinsames Merkmal, das die neuen Regulationsformen charakterisierte, war die unternehmensübergreifende Koordination wirtschaftlicher Transaktionen, die durch einen

1

Universität Trier [email protected] 2 Dieser Aufsatz fasst die Ergebnisse aus verschiedenen Projekten zusammen, die an der Universität Trier durchgeführt und von der Volkswagen-Stiftung, der Deutschen Forschungsgemeinschaft und der Hans-Böckler Stiftung gefördert wurden. Besonderen Dank schulde ich meinen (ehemaligen) Mitarbeitern/innen Jürgen Beyer, Karoline Krenn, Christian Marx und Martin Weber.

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 R. Careja et al. (eds.), The European Social Model under Pressure, https://doi.org/10.1007/978-3-658-27043-8_5

69

hohen Grad an Institutionalisierung erreicht wurde. Chandler (1990) hat diesen Typus als „kooperativen Kapitalismus“ bezeichnet und ihn vom „competitive capitalism“ in den USA abgegrenzt. Der hohe Kartellierungsgrad der deutschen Wirtschaft und die zentrale Rolle der Universalbanken bei der Finanzierung der Großunternehmen haben Hilferding (1915) veranlasst, das deutsche Produktionsregime als „organisierten Kapitalismus“ zu klassifizieren (Winkler 1974). Die Unternehmensverflechtung, die in den folgenden Abschnitten analysiert wird, war eine wichtige Institution des deutschen Produktionsregimes.3 Sie ermöglichte die Koordination wirtschaftlicher Transaktionen zwischen den Großunternehmen und war daher – neben dem Kartell – ein zentrales Element im System der „regulierten Konkurrenz“.4 Viele Manager hatten neben ihrer Führungsposition in einer großen Aktiengesellschaft zugleich ein Mandat in den Aufsichtsräten anderer Unternehmen. Dazu das folgende Beispiel: Carl Klönne hatte 1914 im Netzwerk der deutschen Großunternehmen insgesamt 22 Positionen: Er war Mitglied des Vorstands der Deutschen Bank und zugleich in den Aufsichtsräten von Siemens, Allianz, Gelsenkirchner Bergwerks-AG, Rütgerswerke (usw.) vertreten. Diese multiplen Direktoren erfüllten die Funktion eines go-between zwischen den Unternehmen, in deren Organen sie vertreten waren. Sie vermittelten Informationen über technische oder organisatorische Innovationen (Davis 1991). Banken entsandten ihre Direktoren in die Aufsichtsräte von Unternehmen, an die sie Kredite vergeben hatten (Ziegler 1998). Die Unternehmensverflechtung war eine wichtige Institution zur Konfliktschlichtung und zur Interessenaggregation. Multiple Direktoren, die im Zentrum des Netzwerks standen, hatten häufig eine leitende Position in Unternehmensverbänden.5 Useem (1986, 5) bezeichnet diese Manager als „big linkers“ und weist darauf hin, dass sie die Funktion hatten, das „generalisierte Interesse“ der Großunternehmen zu artikulieren. Bei Entscheidungen im Aufsichtsrat eines Unternehmens waren sie gezwungen, die Interessen der anderen Unternehmen zu berücksichtigen, in deren Organen sie ebenfalls vertreten waren.

2

Kartellgesetze

Produktionsregime bilden ein System von funktional interdependenten ökonomischen Institutionen, die zwischen den Ländern variieren (Hall und Soskice 2001). Zu diesen Institutionen gehören die Arbeitsbeziehungen, die Marktordnung (z.B. Kartellgesetze), das Bildungssystem, das Innovationssystem und die Unternehmensverflechtung. Es ist hier nicht möglich, die Unterschiede in den ökonomischen Institutionen zwischen Deutschland, 3

Zur Unternehmensverflechtung vor dem Ersten Weltkrieg vgl. Eulenburg (1906); Hilferding (1955, 276, Fn 1); Sombart (1928, 135). 4 Schmoller (1906, 249) bezeichnete die durch Kartelle geprägte Marktordnung in Deutschland als „regulierte Konkurrenz“. 5 Zur Rekrutierung des Führungspersonals der Verbände aus dem Netzwerk vgl. Windolf und Beyer (1995, 23, Tab. 5). 70

Frankreich und den USA systematisch zu analysieren. Die Darstellung beschränkt sich auf drei Kartellgesetze, die im 18./19. Jahrhundert verabschiedet wurden und die die Struktur der Marktordnung in den drei Ländern über einen langen Zeitraum geprägt haben. Zu diesen Gesetzen gehören das Urteil des Reichsgerichts von 18976, der Sherman Act von 1890 und das Gesetz Le Chapelier von 1791. Diese Gesetze verkörpern idealtypische Formen ökonomischer Regulierung. Ihre Verabschiedung kann nicht nur durch ökonomische Motive erklärt werden. Vielmehr werden in den Begründungen Leitbilder für eine politische Ökonomie sichtbar, die zeigen, dass institutionelle Variationen auch in kulturellen Unterschieden begründet sind. Es gibt einerseits das Leitbild eines sich selbst regulierenden Marktes und der unbegrenzten Freiheit des Individuums (Hayek 1963). Es gibt andererseits ein Marktmodell, das von der sozialen „Einbettung“ der Akteure ausgeht (Granovetter 1985): Die Individuen orientieren sich in ihrem Verhalten nicht nur an ökonomischer Rationalität, sondern auch an den normativen Standards von Fairness und Vertragstreue. (a) Das deutsche Reichsgericht verurteilte 1897 einen Unternehmer zur Zahlung einer vereinbarten Konventionalstrafe. Dieser hatte gegen den Vertrag des Holzstoffkartells verstoßen und seine Kunden – unter Umgehung des Verkaufsbüros – direkt beliefert.7 Mit diesem Urteil legalisierte das Gericht Kartellverträge und schuf einen stabilen rechtlichen Rahmen für die Kartellierung der deutschen Wirtschaft bis 1945.8 In ihrer Urteilsbegründung wiesen die Richter darauf hin, dass zwei Werteordnungen in Konflikt standen und das Gericht im Sinne einer Güterabwägung entschieden habe: Auf der einen Seite stand die Gewerbefreiheit, eine uneingeschränkte Marktkonkurrenz und die Pflicht des Gerichtes, die individuelle Freiheit der Marktteilnehmer zu schützen. Auf der anderen Seite stand das Recht der Gewerbetreibenden, im Wege der Selbsthilfe „einen Gewerbszweig durch Schutz gegen Preisunterbietungen Einzelner lebensfähig zu erhalten“ (158). Das Gericht argumentiert, dass es legitim sei, „im Wege genossenschaftlicher Selbsthilfe“ die Konkurrenz zu regulieren und „schädliche Ausschreitungen“ zu verhindern (162). Die Richter entschieden sich für das Prinzip einer durch (Kartell)vertrag „regulierten Konkurrenz“. Sie begründeten dies mit Argumenten, die Thomson (1971) in einem klassischen Artikel über die Moral Economy analysiert hat: Die Ökonomie wird nicht nur durch Angebot, Nachfrage und Konkurrenz geregelt. Es gibt daneben auch eine normative Ordnung, die sich an Konventionen orientiert und an dem, was – gemessen an den moralischen Standards der Marktteilnehmer – ein fairer Preis ist. 6

Das Urteil des Reichsgerichts war ein Urteil in letzter Instanz und hatte daher quasi-Gesetzescharakter (Richterrecht). 7 Der Vertrag des Holzstoffkartells sah vor, dass die Mitgliedsfirmen ihre Produkte nur über die gemeinsame Verkaufsstelle verkaufen durften. Der beklagte Unternehmer war Mitglied des Holzstoffkartells und hatte den Vertrag unterzeichnet. 8 Quelle: Entscheidungen des Reichsgerichts in Civilsachen, Bd. 38, Leipzig 1897, S. 155-162. Zur Kritik des Gerichtsurteils aus der Perspektive des Ordo-Liberalismus vgl. Böhm (1948). Zum Kartellierungsgrad der deutschen Wirtschaft 1907-1937 vgl. Fiedler (2002, 229). 71

Die hohe Verflechtungsdichte zwischen den deutschen Großunternehmen, die durch multiplen Direktoren hergestellt wurde, bildete einen institutionellen Rahmen für „Einbettung“ der Marktteilnehmer in diese normative Ordnung. In diesem Sinne ist Unternehmensverflechtung nicht nur ein Koordinationsinstrument, sondern auch

die die die ein

Kontrollorgan. (b) In den USA wurde 1890 der Sherman Act verabschiedet, der Kartelle als eine “conspiracy against the public” verurteilte. Kartellverträge verstoßen gegen das Prinzip der Marktfreiheit und sind daher nichtig. Preisabsprachen gelten als kriminelle Akte und werden entsprechend bestraft. Der Kartellvertrag, den die sächsischen Holzstoff-Fabrikanten 1893 unterzeichnet hatten, hätte vor keinem Gericht in den USA eingeklagt werden können. Im Gegenteil, das Gericht wäre verpflichtet gewesen, eine strafrechtliche Verfolgung einzuleiten. Der Sherman Act verteidigt das Prinzip der negativen Freiheit (Berlin 2002, 169): Hindernisse, die dem freien Wettbewerb im Wege stehen, sollen beseitigt werden. Die rechtliche Regulierung orientiert sich am Prinzip einer uneingeschränkten Marktfreiheit. Der Sherman Act wurde 1914 durch den Clayton Act ergänzt und erweitert. In § 8 werden interlocking directorates zwischen konkurrierenden Unternehmen verboten. Louis Brandeis, der 1916 als Richter an den Supreme Court berufen wurde, war ein besonders scharfer Kritiker der Unternehmensverflechtung: “Interlocking directorates must be prohibited, because it is impossible to break the Money Trust without putting an end to the practice [of interlocking directorates] in the larger corporation.” (Brandeis 1995, 76).9 In Deutschland wurden Kartelle legalisiert, und es gab kein Gesetz, das Verflechtungen zwischen konkurrierenden Unternehmen verboten hätte. In den USA standen die interlocking directorates unter Generalverdacht, als Instrument für Preisabsprachen missbraucht zu werden. In den USA wird die institutionelle Ordnung der Wirtschaft durch eine weitgehende Marktfreiheit der Individuen geprägt, die nur durch schwache (gesetzliche) Schranken begrenzt wird. Eine hohe Verflechtungsdichte zwischen den Unternehmen und eine hohe normative Erwartung an die Mitglieder des board of directors wäre mit dieser Wirtschaftsordnung nicht vereinbar. Der Vergleich der Unternehmensverflechtung zwischen den USA und Deutschland wird zeigen, dass die Verflechtungsdichte in den USA deutlich geringer war als in Deutschland. (c) 1791 wurde das Gesetz Le Chapelier von der französischen Nationalversammlung verabschiedet. In § 1 werden alle intermediären Assoziationen, die die Französische Revolution liquidiert hatte, verboten; ihre Neugründung wird unter Strafe gestellt. Intermediäre Assoziationen sind, unter anderen, die Zünfte, die die Berufsausbildung organisiert hatten; Gewerkschaften, Arbeitgeberverbände und Kartelle. Frankreich hatte damit bereits sehr früh ein gesetzliches Kartellverbot kodifiziert.10 9

Zum Money Trust gehörte eine Gruppe von Banken und Versicherungen in New York, die untereinander durch den Austausch von Direktoren verbunden waren (= interlocking directorates). Im Zentrum des Money Trust stand die Bank von J.P. Morgan. 10 Das Gesetz Le Chapelier richtete sich in erster Linie gegen politische Verbände aus dem Lager der Jakobiner, die die Französische Revolution fortsetzen wollten. Dieser Aspekt wird hier vernachlässigt, da er nicht zum Thema 72

Das Gesetz Le Chapelier orientierte sich an Prinzipien der politischen Ordnung, die Rousseau im Contrat social vorgezeichnet hatte. Es gibt nur zwei legitime politische Akteure: die Bürger auf der einen Seite und der Staat auf der anderen Seite. Alle intermediären Verbände, die das Vakuum zwischen Bürger und Staat ausfüllen könnten, werden als nicht legitime Assoziationen betrachtet, die nur gruppenegoistische Ziele verfolgen.11 Im Gesetz Le Chapelier wurden die liberalen und revolutionären Traditionen Frankreichs in einem rechtlichen Rahmen institutionalisiert. Wenn man versuchen würde, Frankreich aus dieser Perspektive in die von Hall und Soskice (2001) vorgeschlagene Klassifikation einzureihen, würde Frankreich den liberalen Marktökonomien zugeordnet werden. Neben den liberalen Freiheitsrechten wird in den Schriften Rousseaus jedoch ein zweites Prinzip formuliert, das dazu in Widerspruch steht und das Prinzip der negativen Freiheit verletzt. Der französische Staat repräsentiert eine aufgeklärte Autorität, die zu weitreichenden Eingriffen in die Wirtschaftsordnung legitimiert ist. Die Staatsbürokratie definiert das „Gemeinwohl“, das sie versucht, gegenüber den Marktteilnehmern, die nur partikularistische Interessen vertreten, durchzusetzen.12 Frankreich ist ein ambivalenter Fall, der einerseits radikal-liberale Traditionen verkörpert und daher dem angelsächsischen Marktliberalismus nahesteht. Andererseits gibt es starke Zentralisierungs- und Planungselemente, die eine Zuordnung zu den koordinierten Marktökonomien als sinnvoller erscheinen lassen (modèle dirigiste). Die Unternehmensverflechtung wird in Frankreich durch die Karrierewege der französischen (Wirtschafts)elite geprägt. Die französischen Direktoren, die viele Mandate im Netzwerk der Großunternehmen haben, sind zu einem hohen Anteil Absolventen der EliteUniversitäten, vor allem der Ecole Polytechnique. Sie haben einen Teil ihrer Karriere im Staatsdienst verbracht und sind später direkt in Führungspositionen der Großunternehmen berufen worden. Der französische Staat greift nicht direkt in die unternehmerischen Entscheidungen der Privatwirtschaft ein. Der Mechanismus der pantouflage garantiert, dass viele Spitzenpositionen in französischen Großunternehmen von Personen eingenommen werden, die einen staatsnahen Sozialisations- und Bildungsprozess durchlaufen haben und über enge Kontakte zur Staatsbürokratie verfügen.13

gehört. Vgl. dazu die Beiträge in Plessis (1993). Zur Verbreitung von Kartellen in Frankreich vor 1914 vgl. Freedeman (1988). 11 Der Unterschied zwischen Deutschland und Frankreich lässt sich am Gebrauch des Begriffs Korporatismus verdeutlichen: In Deutschland bezeichnet Korporatismus ein Verhandlungssystem, an dem Interessenverbände und der Staat beteiligt sind und das gruppenübergreifende Interessen definieren und durchsetzen soll (Lehmbruch 1985; Armingeon 1993). In Frankreich wird dieser Begriff im Plural gebraucht: „Les corporatismes“ sind intermediäre Verbände, die partikularistische Interessen vertreten (z.B. bei einem Streik der Lokomotivführer oder des Personals der Flugsicherung). Vgl. dazu Reynaud (1991). 12 Zur Kritik des Begriffs intérêt général vgl. Rangeon (1986). 13 Als pantouflage bezeichnet man den Wechsel eines Absolventen einer Elite-Universität aus dem französischen Staatsdienst (z.B. der Finanzverwaltung) in die Führungsposition eines Großunternehmens. Eine systematische Darstellung dieses Mechanismus findet sich in Bourdieu (1989). Für den Zeitraum 1880-1980 vgl. Charle (1987). Zur pantouflage in französischen Großunternehmen vgl. Bauer und Bertin-Mourot (1995). 73

Die institutionelle Ordnung der Wirtschaft wird in Frankreich durch zwei widersprüchliche Elemente geprägt: Die uneingeschränkte Marktfreiheit der Individuen wird durch einen zentralistisch organisierten Staat begrenzt, dessen Einfluss vor allem in den Unternehmen zu spüren ist, an denen der Staat einen hohen Aktienanteil hält und/oder die mit der Staatsbürokratie durch pantouflage verbunden sind. Daher finden wir auf der einen Seite eine hohe Verflechtungsdichte zwischen den französischen Unternehmen mit starkem Staatseinfluss, und auf der anderen Seite eine liberale laissez-faire Marktordnung in den staatsfernen Unternehmen. Es wird sich zeigen (Abb. 2, unten), dass die Verflechtungsdichte zwischen den französischen Großunternehmen höher ist als in den USA, aber deutlich geringer als in Deutschland.

3

Hypothesen

Wenn man die drei Gesetze, die bisher diskutiert wurden, aus einer vergleichenden Perspektive auf das Thema der Unternehmensverflechtung bezieht, kann man die folgenden Hypothesen formulieren, die in den folgenden Abschnitten überprüft werden: In Deutschland war die Unternehmensverflechtung während des 20. Jahrhunderts höher als in den USA und in Frankreich. Die Unternehmensverflechtung erfüllte in Deutschland eine zu den Kartellen komplementäre Funktion und entsprach einer Tradition, die durch das Urteil des Reichsgerichts einen rechtlichen Rahmen erhalten hatte (regulierte Konkurrenz). In den USA war die Unternehmensverflechtung deutlich niedriger als in Deutschland und in Frankreich. Eine hohe Verflechtungsdichte zwischen den Unternehmen widerspricht der Idee einer liberalen Marktordnung. Netzwerke stehen im Verdacht eine parasitäre Parallelökonomie zu etablieren (Luhmann 1996). Frankreich wurde als ein ambivalenter Fall zwischen Marktliberalismus und Dirigismus beschrieben. Die Verflechtungsdichte war niedriger als in Deutschland und näherte sich der Zeitreihe für die USA an. Frankreich nimmt eine intermediäre Position zwischen Deutschland und den USA ein (vgl. Abb. 2, unten). Die bisher formulierten Hypothesen beziehen sich auf einen statischen Vergleich der Unternehmensverflechtung zwischen drei Ländern. Die Dichte des Netzwerks verändert sich jedoch auch im Zeitablauf: Das Netzwerk entstand im späten 19. Jahrhundert und hatte während der take-off Phase eine relativ geringe Dichte. In der Zwischenkriegszeit steigt die Netzwerkdichte in Deutschland auf ein sehr hohes Niveau. Zwischen 1995 und 2010 kann man 74

die weitgehende Auflösung der Deutschland AG beobachten: 2010 liegt die Dichte unterhalb des Niveaus von 1896. Die einzelnen Zeitabschnitte werden in Abschnitt 5 genauer analysiert.

4

Funktionen und Strukturen

4.1

Ego-Netzwerke

Die Mannesmann AG war 1992 mit insgesamt dreißig anderen deutschen Großunternehmen durch interlocking directorates verbunden, darunter Finanzunternehmen (Deutsche Bank, Allianz), verschiedene Zuliefer-Betriebe (RWE, VIAG) und zahlreiche Kunden (VW, BMW, Hoechst). Abbildung 1 zeigt das Netzwerk dieser 31 verbundenen Unternehmen, mit Mannesmann im Zentrum des Netzes. Dieses Netz wird als das Ego-Netzwerk der Mannesmann AG bezeichnet.

Abbildung 1: Ego-Netzwerk der Mannesmann AG 199214

Funktionalistische Netzwerktheorien versuchen die Frage zu beantworten, warum Mannesmann gerade mit diesen – und keinen anderen - Großunternehmen verbunden war. Es werden darauf zwei Antworten gegeben: Erstens, Unternehmen kooptieren Akteure in ihre Netzwerke, auf deren Ressourcen sie besonders angewiesen sind (Reduzierung der Ressourcenabhängigkeit). Dazu gehören z.B. Zulieferbetriebe sowie auch wichtige Kunden und Banken. Nicht-Finanzunternehmen versuchen in ihre Ego-Netzwerke Banken zu integrieren, um ihren Kreditbedarf sichern zu können. Zweitens, Manager kooptieren befreundete Manager aus anderen Großunternehmen, um ihre eigene Position gegenüber kritischen Aktionären

14

Der größere Kreis in der Mitte repräsentiert die Mannesmann AG. Das Unternehmen ist mit 30 anderen Unternehmen verbunden, es hat also einen degree von 30 (vgl. dazu Wassermann und Faust 1994, 100). Dichte des Ego-Netzwerkes: 35,6%. Vgl. dazu auch die Ego-Netzwerke der Deutschen Bank in Abb. 3, unten. 75

abzusichern (Sicherung der Managerherrschaft). Sie schaffen sich auf diese Weise eine eigene „Hausmacht“ im Aufsichtsrat (Koenig et al. 1979). Funktionalistische Netzwerktheorien gehen davon aus, dass die Struktur und Zusammensetzung eines Ego-Netzwerks das Ergebnis der strategischen Entscheidungen der Mitglieder sind (Mizruchi 1996). Unternehmen können – innerhalb gewisser Grenzen – die Zusammensetzung ihres Ego-Netzwerks kontrollieren, d.h. sie können entscheiden mit welchen Unternehmen sie verbunden sein wollen (und welchen nicht). 4.2

Gesamtnetzwerk

Das Ego-Netzwerk der Mannesmann-AG ist eingebettet in ein Gesamtnetzwerk, das durch die Verflechtung aller deutschen Großunternehmen gebildet wird. Im Netzwerk dieser Deutschland-AG gab es 1992 insgesamt 1516 Beziehungen.15 Die Struktur dieses Gesamtnetzwerks – d.h. seine Dichte, der Grad der Zentralisierung und der Anteil der Mehrfachverflechtungen – kann von keinem einzelnen Akteur kontrolliert werden. Das Gesamtnetzwerk ist das emergente Produkt vieler Einzelentscheidungen. Durkheim (1984) hat dafür den Begriff der „sozialen Tatsache“ geprägt. Soziale Tatsachen haben eine eigene Form und Struktur, die sich nicht direkt aus den Handlungen einzelner Personen ableiten lassen. In den folgenden Abschnitten wird die Struktur von nationalen Gesamtnetzwerken analysiert (Makro-Ebene). Diese Analyse ist keine Funktionsanalyse, sondern eine Strukturanalyse. Es ist nicht möglich zu überprüfen, welche Funktion die 1516 Beziehungen im Netzwerk der 250 deutschen Unternehmen im Einzelnen hatten (und ob sie überhaupt eine Funktion hatten). Das Gesamtnetzwerk bietet eine Gelegenheitsstruktur. Für jeden Akteur ist das Gesamtnetzwerk eine gegebene soziale Tatsache. Die Mitglieder des Netzwerks können versuchen, ihre Interessen im Rahmen dieser Struktur zu verfolgen. Die Gesamtstruktur des Netzes steht nicht zu ihrer Disposition. Der soziale Tausch, der zwischen einzelnen Unternehmen und/oder Managern stattfindet, kann mit den hier verwendeten Erhebungsmethoden nicht beobachtet werden.16 Eine Funktionsanalyse von Ego-Netzwerken kann zuverlässig nur im Rahmen von Fallstudien durchgeführt werden. Die nachfolgenden Analysen beschränken sich auf die Strukturanalyse von Gesamtsystemen. Die Strukturanalyse nationaler Netzwerke versucht die Frage zu beantworten, ob es signifikante Unterschiede zwischen Deutschland, Frankreich und den USA in Bezug auf die Dichte, Redundanz und Zentralisierung der Netzwerke gibt. Es geht weiterhin um die Frage, ob 15

Vgl. dazu Tabelle A1 im Online-Anhang. In dieser Tabelle werden die Beziehungen doppelt gezählt (symmetrische Matrix): Wenn A eine Beziehung zu B hat, hat B auch eine Beziehung zu A. 16 Im Januar 2013 wurde das vertrauliche Protokoll einer Aufsichtsrats-Sitzung der Thyssen-Krupp AG der Presse zugespielt. In diesem Protokoll wird dokumentiert, dass Peer Steinbrück seine politische Unterstützung für niedrige Strompreise zusicherte (P. Steinbrück war der Kanzlerkandidat der SPD im Bundestagswahlkampf 2013; bis Herbst 2012 war er Mitglied des Aufsichtsrats der Thyssen-Krupp AG) (Murphy und Iversen 2013). Das Beispiel illustriert, dass der „soziale Tausch“, der in Unternehmens-Netzwerken stattfindet, nur selten beobachtet werden kann. 76

sich diese Unterschiede im Rahmen der Hypothesen erklären lassen, die in Abschnitt 3 formuliert wurden. Die diachrone Strukturanalyse bezieht sich auf die Frage, ob sich die Netzwerkparameter im Beobachtungszeitraum verändern (z.B. Zunahme der Verflechtungsdichte bis 1928; Auflösung der Deutschland AG). Veränderungen im Zeitablauf können durch Veränderungen des institutionellen Kontextes im jeweiligen Land erklärt werden. So kann die Abnahme der Verflechtungsdichte in Deutschland nach dem Zweiten Weltkrieg z.B. auf eine zunehmende Liberalisierung der Marktordnung zurückgeführt werden.

5

Die Dichte des Netzwerks

Die Dichte gehört zu den wichtigsten Strukturparametern eines Netzwerks. Sie gibt an, welcher Anteil der möglichen Verflechtungen in einem Netzwerk tatsächlich realisiert wird.17 Abbildung 2 zeigt die Dichte der nationalen Netzwerke für den Zeitraum 1900-2010. Für jedes Erhebungsjahr wurde die Stichprobe auf die 250 größten Unternehmen in jedem Land standardisiert. Die Dichte wird also nicht durch die Größe des Netzwerks beeinflusst und ist zwischen den Ländern und Jahren vergleichbar. Eine Übersicht über die Netzwerkparameter findet sich im Online-Anhang in Tabelle A1. Gegen Ende des 19. Jahrhunderts war die Netzwerkdichte in Deutschland noch niedriger als in den USA und Frankreich (1896: 1,8%). Bis 1914 steigt die Dichte auf 7,3% und ist in 2010 deutlich höher als in den beiden anderen Ländern. Die Zeit vor dem Ersten Weltkrieg ist die take-off Phase des deutschen Produktionsregimes. Zu den zentralsten Unternehmen im Netzwerk zählen 1914 die AEG (Emil Rathenau), die Deutsch-Luxemburgischen Bergwerkshütten (Hugo Stinnes) und die Deutsche Bank (Carl Klönne). Die 1920er Jahre sind gekennzeichnet durch den verlorenen Ersten Weltkrieg und die hohen Reparationsforderungen; eine Hyperinflation (1923) und die Besetzung des Ruhrgebiets durch die französische Armee (1923/24). Die Verflechtungsdichte steigt auf einen Spitzenwert von 16,2% - eine Dichte, die von keinem der drei Länder im 20. Jahrhundert wieder erreicht wird.

17

In einem Netzwerk mit N Mitgliedern sind die möglichen Verflechtungen definiert als: N(N-1). Die Dichte (in %) ist definiert als: [realisierte Verflechtungen/(N²-N)]*100 (symmetrisierte und dichotomisierte Matrix). Beispiel: 1896 betrug die Verflechtungsdichte zwischen den 211 deutschen Großunternehmen 1,84%. Dies bedeutet, dass von den möglichen Beziehungen (= 211*210 = 44310) nur 814 (=1,84%) realisiert wurden. 77

18 16 D

Dichte in %

14 12 10 8

F

6 4

U

2 0

1900

1914

1928

1934

1938

1976

1995

2010

Abbildung 2: Dichte der Unternehmensverflechtung 1900 – 2010 Quellen: Forschungsverbund Historical Corporate Networks, der von der Universitäten Utrecht und Lausanne koordiniert wurde. Datensatz ist verfügbar am Center for Global Economic History, Universität Utrecht (NL), Projekt: Corporate networks (data); http://cgeh.nl/cgeh/corporate-networks-data/

Eine wichtige Ursache für den Anstieg der Dichte ist darin zu sehen, dass das Netzwerk sich auf eine kleine Zahl von multiplen Direktoren (big linkers) konzentrierte, die in die Aufsichtsräte vieler Großunternehmen gewählt wurden. Die Anzahl der Mandate, die von den Top-20 multiplen Direktoren gehalten wurden, betrug 1896 im Durchschnitt 5,3 Mandate; 1914 steigt die Zahl auf 11,5 Mandate und erreicht 1928 einen Spitzenwert von 21,8 Mandaten. Je stärker die ökonomische Elite von einer kleinen Zahl von multiplen Direktoren mit sehr vielen Mandaten dominiert wird, umso höher ist die Netzwerkdichte (ceteris paribus). In einem Artikel, der 1909 in der Neuen Presse (Wien) veröffentlicht wurde, argumentierte Walter Rathenau: „Dreihundert Männer, von denen jeder jeden kennt, leiten die wirtschaftlichen Geschicke des Kontinents und suchen sich Nachfolger aus ihrer Umgebung. … im Laufe eines Menschenalters [hat] sich eine Oligarchie gebildet, so geschlossen wie die des alten Venedig.“ (1928, 350).18 Über die Zahl „dreihundert“ kann man sicherlich streiten, unstrittig ist aber die Beobachtung, dass die Großunternehmen von einer kleinen Gruppe von Managern kontrolliert wurden, die sich kannten, sich wechselseitig in die Aufsichtsräte ihrer Unternehmen kooptierten und sich in jeweils wechselnder Zusammensetzung in den Aufsichtsräten mehrmals im Jahr trafen.19 Die politischen und wirtschaftlichen Krisen, durch die die Weimarer Republik hindurchging, liefern eine weitere Erklärung für die hohe Verflechtungsdichte. Sie kann als eine Abwehrreaktion der Unternehmen gegen eine äußere und eine innere Bedrohung interpretiert 18

Diese Textstelle wurde von antisemitischen Freikorps-Mitgliedern in der Weimarer Republik als „Beweis“ für die Dominanz der Juden in der Wirtschaft immer wieder zitiert. W. Rathenau wurde am 24.6.1922 von Mitgliedern der Terrororganisation Consul ermordet. 19 Dazu ein Beispiel aus den 1990er Jahren: Dr. Wolfgang Röller war 1992 Vorstandsvorsitzender der Dresdner Bank, er hatte zehn Mandate in deutschen Großunternehmen. Dr. Werner Dieter war Vorstandsvorsitzender der Mannesmann AG; er hatte acht Mandate. Röller und Dieter trafen sich in den Aufsichtsräten der Allianz, Degussa, Hoechst und RWE, d.h. sie trafen sich 12-16mal im Jahr. Dieses Beziehungsnetz war eine Basis der Managerherrschaft. 78

werden (Ruhrbesetzung, kommunistische Partei).20 Der Anstieg der Verflechtungsdichte in Frankreich von 4,8% (1928) auf 7,3% (1938) lässt sich wahrscheinlich auf eine vergleichbare Abwehrreaktion zurückführen: 1936 kam eine Koalition von Sozialisten und Kommunisten in Frankreich an die Regierung (Volksfront). Die neue Regierung plante die Verstaatlichung zahlreicher Großunternehmen (Bodin und Touchard 1972). Zwischen 1928 und 1933 sinkt die Netzwerkdichte in Deutschland von 16,2% auf 11,1%. In diesen Zeitraum fällt der Beginn der Weltwirtschafts- und Bankenkrise. In einer Notverordnung vom September 1931 wurde die Anzahl der Mandate, die eine einzelne Person übernehmen konnte, auf 20 begrenzt (Lieder 2006, 223). Viele multiple Direktoren waren gezwungen, auf einen Teil ihrer Mandate zu verzichten. Weiterhin fällt in diesen Zeitraum die Liquidierung eines Teils der jüdischen Wirtschaftselite. Zahlreiche Direktoren, die im Zentrum des Netzwerks standen, waren jüdischer Abstammung und wurden aus den Aufsichtsräten vertrieben (Windolf 2011). Die Periode nach dem Zweiten Weltkrieg ist geprägt durch die Entflechtung der deutschen Großunternehmen und ein Kartellverbot, das 1945 von der US-Besatzungsmacht verhängt wurde. Mehr als tausend Kartellverträge – die meisten noch während der Zwischenkriegszeit geschlossen - mussten unter dem Druck der alliierten Besatzungsmächte aufgelöst werden (Wells 2002, 154). Das Gesetz gegen Wettbewerbsbeschränkungen, in dem ein generelles Kartellverbot kodifiziert wurde, trat erst 1958 in Kraft. Es gehörte zu den besonders kontroversen Wirtschaftsgesetzen der Nachkriegszeit.21 Insgesamt ist die Wirtschaftsordnung der Nachkriegszeit durch die Entflechtung der Großunternehmen und die Einführung einer liberalen Marktordnung gekennzeichnet. Dies spiegelt sich in Abb. 2 in einem Rückgang der Verflechtungsdichte nach dem 2. Weltkrieg wider. Bis 1976 sinkt die Verflechtungsdichte auf 6,5% und liegt damit deutlich unterhalb des Niveaus von 1938 (10,4%). Die Periode zwischen 1995 und 2010 markiert die Auflösung des Netzwerks, nicht nur in Deutschland, sondern auch in den USA und in Frankreich. In den USA sinkt die Dichte von 3,4% auf 1,6%, in Frankreich von 3,0% auf 1,5%; in Deutschland sinkt sie noch weiter ab, und zwar von 4,8% auf 1,2%. Sie liegt damit unterhalb des Niveaus, das zu Beginn der Untersuchungsperiode (1896) beobachtet wurde. Die Auflösung des Netzwerks kann auf verschiedene Ursachen zurückgeführt werden: (a) In den meisten OECD-Staaten haben die institutionellen Investoren ihren Einfluss auf die Unternehmensverfassung geltend gemacht. In Deutschland wurde ein Corporate Governance Code verabschiedet, der vorschreibt, dass ein Manager nicht mehr als drei 20

Firmenzusammenschlüsse, z.B. die Gründung der Vereinigten Stahlwerke (Reckendrees 2000, 86-88) und der IG-Farben (Plumpe 1990, 131-33), können ebenfalls als eine Reaktion auf den verlorenen Weltkrieg (Verlust aller Auslandstöchter) und auf die Krisen der Weimarer Republik interpretiert werden. Einige Unternehmer haben das Netz auch genutzt, um ihren Widerstand gegen die Wirtschafts- und Sozialpolitik der Weimarer Republik zu koordinieren (Marx 2013, 337). 21 Vgl. dazu Nörr (1994, 200). Zahlreiche Verbandsvertreter argumentierten, dass ein generelles Kartellverbot gegen die im Grundgesetz garantierte Vertragsfreiheit verstoße. Vgl. dazu Böhm (1956), der das Kartellverbot verteidigt. 79

(externe) Aufsichtsratsmandate übernehmen sollte.22 Im deutschen Datensatz für 2010 gibt es keine Person, die mehr als vier Mandate hat. Häufig wird ein Mandat im Vorstand des eigenen Unternehmens mit drei Mandaten in den Aufsichtsräten anderer Unternehmen verbunden. Eine einfache Berechnung illustriert die Wirkung dieser Bestimmung: 1928 hielt Jacob Goldschmidt 47 Mandate: Er schuf (47*46)/2 = 1081 Beziehungen im Netzwerk der Großunternehmen.23 Nach der Notverordnung von 1931 durften Manager nur noch 20 Mandate übernehmen. Sie konnten damit maximal (20*19)/2 =190 Beziehungen schaffen. Gegenwärtig sollten Manager insgesamt nur noch vier Positionen einnehmen: (4*3)/2 = 6 Beziehungen. Die Begrenzung der Mandate hat weitreichende Konsequenzen für die Dichte des Netzwerks und für die Kohäsion der ökonomischen Elite. (b) Viele Untersuchungen zur Unternehmensverflechtung haben gezeigt, dass Finanzunternehmen im Zentrum des Netzwerks standen. Banken spielen eine wichtige Rolle bei der Unternehmensfinanzierung. Hilferding hatte damit seine These von der „Bankenhegemonie“ begründet. Banken haben auch eine aktive Rolle bei der Verteidigung der Deutschland AG gespielt und die Mitglieder des Netzwerks gegen feindliche Übernahmen abgeschirmt. Mit der Transformation der Universalbanken zu Investmentbanken und der Verfügbarkeit des modernen Risiko-Management verändert sich die Interessenlage der Banken. Kredite, die Banken an Unternehmen vergeben haben, werden gebündelt und an den Finanzmärkten verkauft (securitization). Eine Bank kann sich auch durch den Kauf einer Kreditversicherung gegen ein Kreditrisiko absichern (credit default swap). Damit entfällt das Motiv, einen Bankvertreter in den Aufsichtsrat des Schuldnerunternehmens zu entsenden: Die Bank trägt kein Kreditrisiko mehr. Die feindliche Übernahme von Mannesmann durch Vodafone (2000) signalisierte eine weitere Interessenverlagerung: Banken schützen Unternehmen nicht mehr vor (feindlichen) Übernahmen, vielmehr initiieren sie derartige Übernahmen, weil sie damit viel Geld verdienen können. Ein Netzwerk wechselseitiger Verpflichtungen ist für diese Form des Investmentbanking nicht nur entbehrlich, sondern eher hinderlich (Beyer 2007). (c) Die Unternehmensnetzwerke, die in Abbildung 2 dargestellt werden, sind nationale Netzwerke. Das Netzwerk verbindet die Unternehmen einer nationalen Wirtschaft zu einer Interessengemeinschaft. Transnationale Verflechtungen waren selten und hatten nur einen geringen Anteil an der Gesamtzahl der Beziehungen. Mit der Globalisierung der Unternehmen verliert die nationale Vernetzung an Bedeutung. Die Deutsche Bank, Siemens oder

22

„Wer dem Vorstand einer börsennotierten Gesellschaft angehört, soll insgesamt nicht mehr als drei Aufsichtsratsmandate in konzernexternen börsennotierten Gesellschaften wahrnehmen …“ Deutscher Corporate Governance Kodex, Ziffer 5.4.5. Der französische Corporate Governance Code (Abschnitt 19) legt fest, dass ein Vorstandsmitglied (dirigeant mandataire) nicht mehr als zwei externe Mandate wahrnehmen soll und ein Verwaltungsratsmitglied (administrateur) nicht mehr als vier externe Mandate. 23 Jakob Goldschmidt (1882-1955) war von 1922 bis 1931 alleinhaftender Gesellschafter der Darmstädter und Nationalbank (Danat-Bank). Wegen seiner jüdischen Herkunft musste er Deutschland verlassen und emigrierte 1934 in die USA. 80

Volkswagen beschäftigen inzwischen mehr als 50% ihrer Arbeitnehmer im Ausland. Damit verschiebt sich der Fokus der Unternehmensstrategien. Der Aufstieg und die Auflösung der Deutschland AG werden am Beispiel der Deutschen Bank für den Zeitraum 1896-2010 in Abbildung 3 illustriert. Die Abbildung zeigt die EgoNetzwerke der Deutschen Bank für die Stichprobenjahre 1896, 1914, 1928, 1992 und 2010. 1896 war die Deutsche Bank mit sieben deutschen Großunternehmen verbunden (degree 7), das waren 3,3% der Unternehmen in unserer Stichprobe. 1928 war die Deutsche Bank mit 119 Großunternehmen verbunden (degree 119), das war fast die Hälfte der Großunternehmen in der Stichprobe (48%). 2010 ist die Deutsche Bank nur noch mit 15 Großunternehmen verbunden (degree 15), das sind 6% der Unternehmen in unserer Stichprobe. 1896: degree 7 (3,3%)

1914: degree 68 (27%)

1992: degree 62 (24,4 %)

2010: degree 15 (6 %)

1928: degree 119 (48 %)

Abbildung 3: Ego-Netzwerke der Deutschen Bank 1896-2010

6

Unternehmens-, Länder- und Perioden-Effekte

Abbildung 2 bestätigt zunächst die Hypothesen, die für die Netzwerkdichte in den drei Ländern formuliert wurden: Das Netzwerk der deutschen Unternehmen hat zwischen 1914 und 1995 eine deutlich höhere Verflechtungsdichte im Vergleich zu den französischen und den USUnternehmen. Die Netzwerkdichte der französischen Unternehmen ist nur geringfügig höher im Vergleich zu den US-Unternehmen (Ausnahme: 1938). In diesem Abschnitt werden die Differenzen zwischen den Ländern quantifiziert und von anderen Einflussfaktoren isoliert. Die Dichte des Netzwerks wird nicht nur von der Zugehörigkeit der Unternehmen zu einer nationalen Volkswirtschaft beeinflusst, sondern auch 81

durch Perioden-Effekte und durch die Merkmale der Unternehmen selbst. Die Ergebnisse der Analyse werden in Tabelle 1 zusammengefasst (OLS-Regression). Bevor die Variablen und Koeffizienten erläutert werden, muss zunächst der Begriff des degree genauer erläutert werden. Der degree eines Unternehmens ist definiert als die Anzahl der Firmen, mit denen das Unternehmen über gemeinsame Direktoren verbunden ist. In Abb. 1 wurde das Ego-Netzwerk der Mannesmann AG dargestellt. Mannesmann hatte einen degree von 30, d.h. es war 1992 mit dreißig anderen deutschen Großunternehmen verbunden. Der degree ist auch ein Indikator für die Zentralität eines Unternehmens im Netzwerk (degree centrality). Die Gelsenkirchner Bergwerks AG war 1928 das zentralste Unternehmen in der Stichprobe. Es hatte einen degree von 127, d.h. es war mit mehr als der Hälfte der Unternehmen verbunden (Stichprobengröße: N = 251). Wenn der degree über alle Unternehmen aufsummieren wird, erhält man die Gesamtzahl der Verbindungen in einem Netzwerk. Es gibt also eine direkte Beziehung zwischen degree und Dichte. In den Regressionsanalysen, die in Tabelle 1 dokumentiert werden, ist der degree die abhängige Variable. Die Analyse soll die Frage beantworten, welcher Anteil der Varianz in der abhängigen Variablen auf die Merkmale der Unternehmen zurückgeführt werden kann (Unternehmens-Effekt); welcher Anteil der Varianz auf den Länder-Effekt zurückgeführt werden kann; und welcher Anteil der Varianz auf historische Kontingenzen zurückgeführt werden kann (Perioden-Effekt). Da die nationalen Differenzen zwischen 1914 und 1938 relativ groß sind (und nach 1995 weitgehend verschwinden), wurde die Regressionsanalyse auf den Zeitraum 1914 – 1938 begrenzt. Unternehmens-Effekt: In Tabelle 1 werden in Modell 1 zunächst die UnternehmensVariablen eingeführt. Die logarithmierte Variable Kapital ist ein Indikator für das Eigenkapital; sie misst den Einfluss der Unternehmensgröße auf den degree und ist hoch signifikant. Die Variable Board misst die Größe des Aufsichts-/Verwaltungsrats: Je größer der Aufsichtsrat, umso größer sind die Verflechtungschancen eines Unternehmens.24 Wenn der Aufsichtsrat um eine Person erweitert wird, steigt der degree des Unternehmens um 1,2. Die logarithmierte Variable Region misst Zentrums-/Peripherie-Effekte. Viele Großunternehmen sind in Berlin, im Ruhrgebiet, in Paris oder in New York und Boston lokalisiert. Unternehmen in diesen Regionen haben mehr Verflechtungschancen verglichen mit Unternehmen, die in der geographischen Peripherie lokalisiert sind (z.B. Königsberg).25 Auch diese Variable ist signifikant. Die weiteren Variablen messen den Effekt der Wirtschaftssektoren: Ein Unternehmen aus der Montan-Industrie hat einen degree, der um 6,4 höher ist verglichen mit einem Nicht-Montan-Unternehmen. Der Effekt der Elektroindustrie ist etwas stärker (8,6).26 24

Beispiele: 1928 hatte der Aufsichtsrat der Deutschen Bank 69 Mitglieder; der Aufsichtsrat der Darmstädter- und Nationalbank hatte 52 Mitglieder. 25 Eine detaillierte Analyse dieser Variable (Ungleichverteilung des Sozialkapitals über Regionen) findet sich in Windolf (2007). 26 Die Koeffizienten für die Banken sind negativ. Das ist zunächst überraschend, kann aber darauf zurückgeführt werden, dass die Banken besonders große Aufsichtsräte hatten. Im Vergleich zu Unternehmen mit ähnlich großem 82

Die Dummy-Variable „Vorperiode“ hat den Wert 1, wenn das Unternehmen bereits im früheren Stichprobenjahr im Sample war, ansonsten hat sie den Wert 0.27 Das Bestimmtheitsmaß R² zeigt, dass die Variablen, die der Gruppe Unternehmens-Effekte zugeordnet werden, 36% der Varianz der abhängigen Variable erklären können. Tabelle 1: OLS-Regressionen (1914 - 1938) Modell Kapital (ln)

1

2

3

2,13

2,83

2,83

2,83

3

3

3

1,13

1,2

1,1

4

Board

1,2

Region (ln)

2,33

1,83

1,83

1,73

Bank

-2,51

-3,63

-3,0²

-2,7²

Montan

6,43

3,73

3,93

3,93

Elektro

8,6

3

4,8²

5,1²

5,5²

Vorperiode

1,0

1,21

1,41

1,51

Deutschland

19,13

19,23

15,23

Frankreich

7,43

7,63

7,23

1928

4,83

0,8

1938

2,3²

2,3² 11,23

D x 1928 R²

0,36

0,50

0,50

0,52

Abhängige Variable: Degree der Unternehmen (degree centrality). 3: α ≤ 0,000; 2: α ≤ 0,01; 1: α ≤ 0,05. Die Tabelle zeigt die nicht-standardisierten Regressions-Koeffizienten. Gepooltes Sample: N=2833 Unternehmen (für alle Modelle). In Tabelle A2 im Online-Anhang werden für Modell 4 die standardisierten β-Koeffizienten, t-Werte, robuste Standardfehler (Hubert-White) und die jeweiligen Referenzkategorien angegeben.

Länder-Effekt: In Modell 2 werden die Länder-Effekte in Form von Dummy-Variablen eingeführt. Die Referenz-Kategorie sind die USA. Im Vergleich zu US-Unternehmen haben deutsche Unternehmen im Durchschnitt einen degree, der deutlich höher ist (+19,1). Bei den französischen Unternehmen ist dieser Effekt schwächer (+7,4). Da viele Unternehmensvariablen kontrolliert werden (Größe, Region, Wirtschaftssektor), können die Länder-Effekte nicht auf eine unterschiedliche Zusammensetzung der Stichprobe zurückgeführt werden. Sie sind von diesen Effekten „gereinigt“. Anders ausgedrückt: Wenn man ein deutsches und ein US-Unternehmen mit annähernd gleichen Merkmalen vergleicht, dann hat das deutsche Unternehmen im Durchschnitt einen degree, der um 19,1 höher ist. Das Bestimmtheitsmaß steigt im Modell 2 auf 50%. Im Vergleich zum Modell 1 können also ca. 14% der Varianz auf den Länder-Effekt zurückgeführt werden. Perioden-Effekt: In Modell 3 werden Perioden-Effekte in Form von Dummy-Variablen eingeführt. Die Referenz-Kategorie ist das Jahr 1914. Im Vergleich zum Jahr 1914 haben Board haben sie einen unterdurchschnittlichen degree. Wenn man die Variable Board aus der Regression herausnimmt, wird das Vorzeichen für die Banken positiv. 27 Die Variable ist schwach signifikant (α ≤ 0,05). Sie wurde eingefügt, um die serielle Korrelation zwischen den Stichprobenjahren zu kontrollieren. 83

Unternehmen 1928 einen degree, der um 4,8 höher ist. Der Anteil der Varianz, der durch Perioden-Effekte erklärt wird, tendiert gegen Null. Ein Blick auf Abb. 2 zeigt, dass die Perioden-Effekte für die US-Unternehmen sehr schwach sind: Die Dichte verändert sich kaum von einem Stichprobenjahr zum anderen. Dies gilt eingeschränkt auch für Frankreich (Ausnahme: 1938). In Modell 4 wurde ein Interaktions-Effekt zwischen Land und Periode eingeführt: Deutschland (D) x 1928. Diese Interaktions-Variable misst den Effekt des Ausnahme-Jahres 1928 für Deutschland. Tabelle 1 zeigt, dass der Perioden-Effekt für 1928 nicht mehr signifikant ist und ersetzt wird durch den Interaktions-Effekt. Deutsche Unternehmen haben 1928 einen degree, der um 11,2 höher ist (ceteris paribus).

7

Transnationale Netzwerke

Seit mehr als zwei Jahrzehnten wird in den Sozialwissenschaften die Frage diskutiert, ob sich auf globalisierten Märkten auch eine globale Wirtschaftselite herausgebildet hat. Eine in der Literatur häufig vertretene These lautet, dass die nationale Unternehmensverflechtung durch transnationale Netzwerke ersetzt wurde.28 In diesen Netzwerken können die Manager der global agierenden Unternehmen ihre Strategien koordinieren und eine den nationalen Netzwerken vergleichbare Kontrollfunktion ausüben. Heemskerk (2013) hat diese These für die 300 größten europäischen Unternehmen für die Jahre 2005 und 2010 überprüft. Da die transnationalen Netzwerke nicht Gegenstand dieses Beitrags sind, werden hier nur die wichtigsten Ergebnisse seiner Studie zusammengefasst: 2005 betrug die Verflechtungsdichte zwischen den europäischen Unternehmen 0,69%; bis 2010 ist sie auf 0,98% gestiegen.29 Die Verflechtung hat in diesem kurzen Zeitraum zugenommen, die Dichte liegt aber auch 2010 noch deutlich unter dem niedrigen Niveau, das für die nationalen Netzwerke in Deutschland, Frankreich und in den USA für das Jahr 2010 berechnet wurde (vgl. Abb. 2 oben). Es ist unwahrscheinlich, dass die Manager auf europäischer Ebene eine dem nationalen Netzwerk vergleichbare Koordinationsaufgabe erfüllen können. Es gibt in diesem globalen Netzwerk keine „Oligarchie“ von wenigen Top-Managern mit vielen Mandaten (big linker). Die Beziehungen sind auf viele Manager mit jeweils wenigen Mandaten verteilt. Die Beschränkung der Mandate wurde in den meisten europäischen Ländern in den corporate governance code aufgenommen. Damit fehlt diesem Netzwerk ein charismatisches Zentrum, das über Einfluss und Autorität verfügen würde. Die europäischen Unternehmen mit dem höchsten transnationalen degree sind in vier Ländern lokalisiert: Frankreich, Niederlande, Großbritannien und Belgien. Das geographische Zentrum des Netzes wird also zwischen den Städten Paris, Amsterdam, London und Brüssel 28 29

Sklair (2001); Veen und Kratzer (2011); Heemskerk und Takes (2016). Quelle: Heemskerk (2013, 82, Table 2), eigene Berechnungen.

84

aufgespannt. Deutsche Unternehmen sind 2010 unter den 20 Unternehmen mit dem höchsten transnationalen degree nicht vertreten (Heemskerk 2013, 84, 90-91). Heemskerk kommt aufgrund der zunehmenden Verflechtungsdichte zwischen den europäischen Unternehmen zu folgender Schlussfolgerung: „... the central directors and the political geography suggests that there is a structural basis for overcoming the present Euro crisis that has been handled primarily at the political level.“ Angesichts der zahlreichen Skandale, die vor allem (aber nicht nur) die deutschen Unternehmen erschüttert haben (z.B. VW, Deutsche Bank), ist diese Schlussfolgerung sehr optimistisch. Mizruchi (2013, 264-265) kommt angesichts der Finanzmarktkrise von 2008 für die US-Wirtschaftselite zu einem anderen Urteil: „The American corporate elite, since the early 1990s, has become fragmented, without an organized group of pragmatic leaders capable of addressing the major issues with which the group has been confronted. … This is an elite that, rather than leading, has retreated into narrow self-interest…“ Diese Einschätzung beschreibt das Verhalten der gegenwärtigen Wirtschaftselite – sowohl in den USA als auch in Europa – wahrscheinlich realistischer als der optimistische Ausblick von Heemskerk.

8

Schlussbemerkungen

Netzwerke sind eine spezifische Form der Vergesellschaftung mit ambivalenten Funktionen und Folgen. Luhmann (1996, 35-6) weist darauf hin, dass viele Netzwerke „parasitär“ sind: Sie verschaffen den Mitgliedern Vorteile auf Kosten der Ausgeschlossenen. An die Stelle eines durch den Markt bewerteten Austauschs von Gütern und Dienstleistungen treten reziproke „Freundschaftsdienste“. Das old boys‘ network in den USA und die Deutschland AG sind eine nur dem Machterhalt der Manager dienende Institution. Brandeis (1995, 76) hatte den Money Trust angeprangert, weil er in seinen Augen korrupt war und den Wettbewerb beschränkte. Granovetter (1985) formuliert mit dem Konzept von „embeddedness“ eine Gegenposition: Märkte sind ohne eine Einbettung in soziale Beziehungen nicht funktionsfähig. Schmoller hatte schon 1905 in einer Rede vor dem Verein für Sozialpolitik auf dieses Problem hingewiesen und den Begriff der „regulierten Konkurrenz“ geprägt. Der Markt ist kein sich selbst regulierendes System. Ohne Einbettung in Institutionen paralysiert ein Konkurrenzmarkt sich selbst. Die Unternehmensverflechtung ist eine Institution, die die Koordination der Unternehmen und die Selbst-Kontrolle der Mitglieder erleichtert. Die Auflösung des Netzwerks verlief parallel zur Auflösung und Schwächung anderer Institutionen zur Produktion kollektiver Güter, z.B. der Gewerkschaften oder der Arbeitgeberverbände. Dafür gibt es verschiedene Ursachen: Die Deutschland AG war eine Gegenmacht gegen die Aktionäre. Sie schirmte Manager gegen feindliche Übernahmen ab und verhinderte einen direkten Einfluss der Finanzmärkte auf Unternehmensstrategien. Die institutionellen Eigentümer, d.h. die Investment-, Pensions- und Hedge-Fonds halten inzwischen die Mehrheit an den großen börsennotierten Unternehmen. Ihre 85

Interessen wurden bei der Formulierung des Corporate Governance Code berücksichtigt (Beschränkung der externen Mandate der Manager). Die Globalisierung der Märkte ist eine weitere Ursache für die Auflösung der Deutschland AG. Die Unternehmensverflechtung war auf die nationale Volkswirtschaft begrenzt; sie war ein Insider-Netzwerk. Die Großunternehmen sind inzwischen weltweit operierende Konzerne, die häufig mehr als die Hälfte ihrer Belegschaft nicht mehr im Inland beschäftigen. Unter diesen Bedingungen wird ein nationales Netzwerk zunehmend funktionslos. Banken waren zentrale Akteure im Netzwerk. Sie kontrollierten eine Ressource, auf die alle Unternehmen angewiesen sind (Finanzierung). Die Entwicklung neuer FinanzmarktInstrumente hat ihre Position im Netzwerk geschwächt. Unternehmen können sich über Anleihen selbst finanzieren. Kredite, die Banken an Unternehmen vergeben haben, werden gebündelt und an Finanzmärkten verkauft (securitization). Die Bank trägt nicht mehr das Kreditrisiko und kann auf eine Präsenz im Aufsichtsrat des Schuldnerunternehmens verzichten. Schließlich hat die Dominanz der neoliberalen Ideologie zur Auflösung der Deutschland AG beigetragen. Bourdieu (1998, 110) bezeichnet den Neoliberalismus als „ein Programm zur planmäßigen Zerstörung der Kollektive“. Die Deutschland AG gehörte zu jenen „Kollektiven“, die die Produktion von kollektiven Gütern erleichtert hat. Zwischen 1914 und 1938 waren die nationalen Unterschiede bezüglich der Netzwerkdichte relativ hoch. Sie verringerten sich nach dem Zweiten Weltkrieg, aber noch bis Anfang der 1990er Jahre gab es signifikante Unterschiede zwischen den drei Ländern. Die Deutschland AG war ein wichtiges Element im Institutionen-System des kooperativen Kapitalismus. Die Auflösung des Netzwerks kann jedoch nicht nur auf eine Anpassung des deutschen Kapitalismus an das angelsächsische Modell verstanden werden. Es gibt ein Bündel von Ursachen, die in allen Ländern präsent waren und in die gleiche Richtung gewirkt haben. Dazu gehören die Begrenzung der Managerherrschaft; der Aufstieg einer neuen ökonomischen Elite, die zu den Eigentümern der großen Aktiengesellschaften geworden sind (institutionelle Investoren); die Globalisierung der Märkte, die zu einer Verschärfung der Konkurrenz geführt hat; finanztechnische Innovationen, die die intermediäre Rolle der Banken obsolet gemacht haben; und last but not least, der Neoliberalismus, der das Programm und die Legitimation für die Strategien der neuen Eigentümer geliefert hat (z.B. feindliche Übernahmen, Aktienoptionen).

Literaturverzeichnis Abelshauser, Werner. 2001. “Umbruch und Persistenz: Das deutsche Produktionsregime in historischer Perspektive.” Geschichte und Gesellschaft 27 (4): 503-523. Armingeon, Klaus. 1993. “Korporatismus im Wandel: Ein internationaler Vergleich.” In Sozialpartnerschaft: Kontinuität und Wandel eines Models, herausgegeben von Emmerich Tálos, 285-309. Wien: Verlag für Gesellschaftskritik. Bauer, Michel und Bénédicte Berin-Mourot. 1995. L’accès au sommet. Les grandes entreprises françaises 19851994. Paris: CNRS. 86

Berlin, Isaiah. 2002. Liberty. Oxford: Oxford University Press. Beyer, Jürgen. 2007. “Primat der Finanzmarktorientierung. Zur Logik der Auflösung der Deutschland AG.” Berliner Debatte Initial 18 (4/5): 56-64. Bodin, Louis und Jean Touchard. 1972. Front populaire 1936. Paris: A. Colin. Böhm, Franz. 1948. “Das Reichsgericht und die Kartelle.” Ordo 1: 197-213. Böhm, Franz. 1956. “Verstößt ein gesetzliches Kartellverbot gegen das Grundgesetz?” Wirtschaft und Wettbewerb 6: 173-187. Bourdieu, Pierre. 1989. La noblesse d’état: Grandes écoles et esprit de corps. Paris: Minuit. Bourdieu, Pierre. 1998. Gegenfeuer. Konstanz: UVK Medien. Brandeis, Louis. 1995 [1914]. Other People’s Money and How the Bankers Use It. Boston: St. Martin’s Press. Burt, Ronald. 1992. Structural Holes. Cambridge: Harvard University Press. Chandler, Alfred D. 1990. Scale and Scope: The Dynamics of Industrial Capitalism. Cambridge: Harvard University Press. Charle, Christophe. 1987. “Le pantouflage en France (1880-1980).” Annales: Economies, Sociétés, Civilisations 42 (5): 1115-1137. Davis, Gerald. 1991. “Agents Without Principles? The Spread of the Poison Pill Through the Intercorporate Network.” Administrative Science Quarterly 36 (4): 583-613. Durkheim, Emile. 1984 [1895]. Die Regeln der soziologischen Methode. Frankfurt: Suhrkamp. Eulenberg, Frans. 1906. “Die Aufsichtsräte der deutschen Aktiengesellschaften.” Jahrbücher für Nationalökonomie und Statistik 32, III. Folge: 92-104. Fiedler, Martin. 2002. “Fusionen und Übernahmen in der deutschen Industrie, 1898-1938.” Jahrbuch für Wirtschaftsgeschichte 2002 (2): 209-239. Freedeman, Charles. 1988. “Cartels and the Law in France Before 1914.” French Historical Studies 15 (3): 462478. Granovetter, Mark. 1985. “Economic Action and Social Structure: The Problem of Embeddedness.” American Journal of Sociology 91 (3): 481-510. Hall, Peter und David Soskice. 2001. “An Introduction to Varieties of Capitalism.” In Varieties of Capitalism, herausgegeben von Peter Hall und David Soskice, 1-68. Oxford: Oxford University Press. Hayek, Friedrich. 1963. “Kinds of Order in Society.” New Individualist Review 3 (1): 3-12. Heemskerk, Eelke M. 2013. “The Rise of the European Corporate Elite: Evidence from the Network of Interlocking Directorates in 2005 and 2010.” Economy and Society 42 (1): 74-101. Heemskerk, Eelke M. und Frank W. Takes. 2016. “The Corporate Elite Community Structure of Global Capitalism.” New Political Economy 21 (1): 90-118. Hilferding, Rudolf. 1955 [1910]. Das Finanzkapital. Berlin: Dietz. Hilferding, Rudolf. 1915. “Arbeitsgemeinschaft der Klassen?” Der Kampf 8: 321-329. Koenig, Thomas, Robert Gogel und John Sonquist. 1979. “Models of the Significance of Interlocking Corporate Directorates.” American Journal of Economics and Sociology 38 (2): 173-186. Lehmbruch, Gerhard. 1985. “Concertation and the Structure of Corporatist Networks.” In Order and Conflict in Contemporary Capitalism, herausgegeben von John Goldthorpe, 60-80. Oxford: Clarendon. Lieder, Jan. 2006. Der Aufsichtsrat im Wandel der Zeit. Jena: Jenaer Wissenschaftliche Verlagsgesellschaft. Luhmann, Niklas. 1996. “Inklusion und Exklusion.” In Nationales Bewußtsein und kollektive Identität, herausgegeben von Helmut Berding, 15-53. Frankfurt: Suhrkamp. Marx, Christian. 2013. Paul Reusch und die Gutehoffnungshütte. Göttingen: Wallstein. Mizruchi, Mark. 1996. What Do Interlocks Do? An Analysis, Critique, and Assessment of Research on Interlocking Directorates.” Annual Review of Sociology 22 (1): 271-298. Mizruchi, Mark. 2013. The Fracturing of the American Corporate Elite. Cambridge: Harvard University Press. Murphy, Martin und Sönke Iversen. 2013. “Steinbrück bot Tyssen-Krupp Gefälligkeiten an.” Handelsblatt, 1 August, 2013. https://www.handelsblatt.com/politik/deutschland/protokoll-des-aufsichtsrats-steinbrueckbot-thyssen-krupp-gefaelligkeiten-an/7598904.html?ticket=ST-2687792-OgBYZwItIf9JZCoouFFF-ap5 Nörr, Knut. 1994. Die Leiden des Privatrechts: Kartelle in Deutschland. Tübingen: Mohr. Plessis, Alain. 1993. Naissance des libertés économiques: Le décret d’Allarde et la loi Le Chapelier. Paris: Institut d’Etudes Politiques. Plumpe, Gottfried. 1990. Die I.G. Farbenindustrie AG: Wirtschaft, Technik und Politik 1904-1945. Berlin: Duncker & Humblot. Rangeon, François. 1986. L’idéologie de l’intérêt général. Paris: Economica. Rathenau, Walter. 1928 [1909]. “Unser Nachwuchs.” In Nachgelassene Schriften, Band 2, 349-361. Berlin: S. Fischer. Reckendrees, Alfred. 2000. Das Stahltrust-Projekt: Die Gründung der Vereinigte Stahlwerke A.G. München: Beck.

87

Reynaud, Emmanuèle. 1991. Le pouvoir de dire non: Les corporatismes en entreprise. Paris: Éd. L'Harmattan. Schifeling, Todd und Mark Mizruchi. 2014. “The Decline of the American Corporate Network, 1960-2010.” In The Power of Corporate Networks: A Comparative and Historical Perspective, herausgegeben von Thomas David und Gerarda Westerhuis, 31-47. New York: Routledge. Schmoller, Gustav. 1906. “Das Verhältnis der Kartelle zum Staat.“ Verhandlungen des Vereins für Socialpolitik 116: 237-271. Sklair, Leslie. 2001. The Transnational Capitalist Class. Oxford: Blackwell. Sombart, Werner. 1928 [1911]. Die Juden und das Wirtschaftsleben. München: Duncker & Humblot. Thomson, Edward. 1971. “The Moral Economy of the English Crowd in the Eighteenth Century.” Past and Present 50 (1): 76-136. Useem, Michael. 1986. The Inner Circle. Oxford: Oxford University Press. Veen, van Kees und Jan Kratzer. 2011. “National and International Interlocking Directorates Within Europe: Corporate Networks Within and Among Fifteen European Countries.” Economy and Society 40 (1): 1-25. Wasserman, Stanley und Katherine Faust. 1994. Social Network Analysis. Cambridge: Cambridge University Press. Wells, Wyatt. 2002. Antitrust and the Formation of the Postwar World. New York: Columbia University Press. Winkler, Heinrich. 1974. Organisierter Kapitalismus. Göttingen: Vandenhoeck & Ruprecht. Windolf, Paul. 2007. “Sozialkapital und soziale Ungleichheit: Vergleichende Analysen zur Unternehmensverflechtung in Deutschland und in den USA (1896-1938).” Jahrbuch für Wirtschaftsgeschichte (2): 197-230. Windolf, Paul. 2011. “The German-Jewish Economic Elite 1900 -1930.“ Zeitschrift für Unternehmensgeschichte 56 (2): 135-162. Windolf, Paul und Jürgen Beyer. 1995. “Kooperativer Kapitalismus: Unternehmensverflechtungen im internationalen Vergleich.” Kölner Zeitschrift für Soziologie und Sozialpsychologie 47 (1): 1-36. Ziegler, Dieter. 1998. “Die Aufsichtsräte der deutschen Aktiengesellschaften in den zwanziger Jahren: Eine empirische Untersuchung zum Problem der ‘Bankenmacht‘.” Zeitschrift für Unternehmensgeschichte 43 (2): 194-215.

88

State-labour relations in illiberal times: The dissolution of social dialogue in Hungary Rafael Labanino1 Abstract: This chapter explains the dissolution of tripartite interest intermediation and the liberalisation of employment protection legislation in Hungary by comparing the 1999 partial reform and the 2012 full liberalisation. It argues that the causes for the success of this radical transformation of state-labour relations can be found both in the changes of the structural constraints and in the strategies of the different actors.

1

Introduction

How to make sense of the changes in state-labour relations in Hungary between 2010 and 2014? This is a decisive period, as tripartism was effectively dissolved and a completely new, radically pro-employer Labour Code was adopted, which aligned regulation of employment protection with contractual law, that is, it assumed equal power relations between employers and employees (Tóth 2012; Szabó 2013). Although there was an attempt in weakening tripartism in Hungary in 1999, it was rather incremental compared with the January 2012 reform. Similarly, the substantive changes introduced with the new Labour Code in January 2013 dwarf the liberalising amendments in 2001. These two reforms amounted to a path-breaking change in the system of industrial relations in Hungary. There are several papers which describe the changes, the strategies of traditional unions and new actors in coping with the crisis and legislative changes in Hungary (e.g. Komiljovics 2011, 2012; Prugberger and Szöllös 2012; Tóth 2012; Szikra 2014; Scharle and Szikra 2014; Szabó 2017; Neumann 2018), and even how unions gained some concessions from the government (Neumann 2012). However, there are only a handful of theory-driven explanations that make generalisable claims about the changes in state-labour relations in Hungary after the watershed elections of 2010, which gave a constitutional majority to the national-conservative Fidesz party. Szabó (2013), following the framework of Bohle and Grekosvits (2012), places these radical liberalisations in the broader context of the contradictions of post-Communist democratic capitalist regimes. Korkut et al. (2017), on the cases of Hungary and Latvia, explain the inability of unions to provide a strong voice, let alone alternative policies to the ongoing liberalisation with their declining and ageing membership base, outward migration and the weakened social dialogue mechanism. Bernaciak (2015, 37) emphasises the role of politics in

1

Goethe-Universität Frankfurt am Main [email protected]

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 R. Careja et al. (eds.), The European Social Model under Pressure, https://doi.org/10.1007/978-3-658-27043-8_6

89

the weakening of organised labour in Hungary and characterises it as a “conscious government strategy to eliminate all potential sources of opposition and independent social action”. I concur with Szabó (2013) that the inability of the “embedded neoliberal regimes” of the Visegrad countries to provide governments with non-zero-sum outcomes to the contradictory aims of fiscal stability, re-election and the provision of social protection goes a long way in explaining the overall weakness of organised labour and the impetus for liberalisation. Similarly, Korkut et al. (2017) rightly point out that the rapidly decreasing union density rates, outward migration of skilled workers, and weakened social dialogue mechanisms are very much behind the inability of organised labour to withstand the liberalisation of labour rights and industrial relations. Bernanciak (2015) rightly finds that the severity of the crisis alone cannot account for the radical liberalisation of industrial relations in Hungary. The paper points to the importance of the conscious strategy of Orbán’s self-declared “illiberal” regime in destroying one of the most important potential sources of opposition against his anti-welfare and anti-labour policies. None of these accounts, however, explain how this radical liberalisation took place in the first place. Why in Hungary and not in the other Visegrad countries? The very dilemma faced by Hungarian governments is shared by their Czech, Slovak and Polish counterparts. The social dialogue mechanism was weakened in the process of liberalisation; therefore, it cannot be treated as an independent variable. How was Orbán able to implement such radical policies smoothly, when during his first term as prime minister between 1998 and 2002 his much milder liberalisation led to political backlash? What changed? This chapter provides a comparative analysis of the changes in the structure of social dialogue and to the Labour Code adopted in 1999 and 2011 by the first and second Orbángovernments. By looking at two attempts at liberalisation with different outcomes by the same actors in the same country, the chapter highlights the importance of the strategies and the relative strength of the different actors. As the subject of the chapter is institutional change, the analysis is rooted in Avdagic’s (2005, 2006) dynamic framework, which goes beyond institutionalist accounts and explains long-term trajectories in industrial relations as contingent upon both institutional and behavioural variables. However, Armingeon (2012) found that not only did the ideological composition of party systems and the relative strength of parties affect labour relations in Central and Eastern Europe, but the direction was the same as in Western democracies. That is, the presence and relative strength of left, centre and right parties affect labour policy as one would expect based on the assumptions derived from the traditional partisan theory developed in the context of Western cases. My analysis lends further support to Armingeon’s (2012) findings, as it shows that to explain the variance between the labour policies of the first and the second Orbán-government, the dynamics of party competition must also be taken into account (Häusermann, Picot, and Geering 2012; Picot 2009; Green-Pedersen 2002; Kitschelt 2001). 90

The chapter continues as follows: The next section establishes the framework of analysis by bringing together Avdagic’s (2005, 2006) dynamic framework on the evolution on state-labour relationship with the partisan competition framework on the analysis of partisan effect in social policy outcomes. The third and fourth sections contain the actor-based analyses of the labour policy reforms of the first and second Orbán-governments, respectively. The last section concludes.

2

The framework of analysis – an actor-based model

2.1

The interactionist model of state-labour relations

Avdagic (2005) found that none of the traditional theoretic models could account for the observed cross-national differences in the strength of the tripartism and tripartite structures between the Czech Republic, Hungary, and Poland. Neither economic-structural explanations based on macroeconomic conditions at the time of transition, nor the organisation of industrial relations and the patterns of economic reform during socialism (Iankova 2002) were able to explain the strength and the level of institutionalisation of tripartite dialogue mechanisms, even less so actual outcomes of state-labour interactions. The role of the trade unions in the regime change could not explain the level of institutionalisation of tripartite social dialogue either. Similarly, the political orientation of the governments as explanatory variable did not yield consistent results across cases. Explanations built on institutional differences also come up short when applied to Central and Eastern Europe. Neither the level of fragmentation and polarisation of labour unions, nor formal ties between unions and political parties explain the level of coordination in the region. Avdagic (2005) points out that in East Central Europe unionparty ties can be best described as a form of inverse dependency relationship (Avdagic 2005, 40). Organised labour was the main loser of transition: it lost its welfare providing functions (Szikra and Tomka 2009), and the faster the economic liberalisation and privatisation were implemented, the quicker its societal base diminished. It also suffered from its Communist past where unions were a means of control. Therefore, formal ties to parties meant co-optation, not union influence. Avdagic (2006) instead introduces an interactionist approach, which rests on four pillars. First, the focus is on the interplay between the structural context and strategic action in the process of institutional development. That is, Avdagic (2006) does not assume a deterministic, linear causal relationship between structural constraints of the transition and the new institutional setup and actual outcomes. Second, Avdagic (2006) distinguishes between the actors’ core objectives and their preferences regarding corporatist-type deliberations and transformation policies. It is an important distinction, as the assumed core objective of any actor is to maintain their power under any circumstances, while the preferred policy and collective bargaining outcomes are relative to the context. Thirdly, Avdagic (2006) assumes that in the 91

context of transition, uncertainty is an even more serious issue than in the Western context. Therefore, preference formation of actors and their decisions on what strategies to follow are assumed to be informed by past interactions with other actors. Finally, because of the newness and instability of the institutional setup, the ability of actors to rely on “institutions, trust, norms, and habits to serve as mechanisms of social interaction” is limited (Avdagic 2006, 14). That is, in the context of transition political economies it is difficult to anticipate the responses of others, and actors’ choices reflect their perceptions of relative power. 2.2

The role of party competition in policy reform

However, policy reforms, even those affecting the structure of collective bargaining, union rights and employment protection legislation, are not solely depending on state-labour relations, but on political parties initiating and implementing legislative changes. Klaus Armingeon (2012) tested the traditional partisan hypothesis on labour relations on a sample of 25 established democracies (Anglo-Saxon + Japan, Continental, Mediterranean, and Nordic) and 10 post-Communist democracies. Adding the post-Communist cases did not falsify the theory, that is, a presence of a strong-left party was found to be a necessary condition for the following outcomes: high-union density rates; high collective agreement coverage; low employment protection flexibility (and conversely, strong right-wing parties for high flexibility); and high taxation. A strong centre party (Christian Democratic) was a necessary condition for low income inequality. Armingeon’s (2012) analysis is not based on linear causation but is set in the logic of fuzzy set analysis, which emphasises the importance of conjunctural causation in explaining individual outcomes – where the presence of a necessary condition, e.g. a strong left party, does not translate into an expected characteristic of state-labour relations. How to move beyond linear causation in measuring partisan effect on social and labour policy outcomes? It is important to note that Armingeon (2012) analysed party system properties (which type of parties are present and their relative strength) and not the colour of governments. This is a crucial distinction. As Kitschelt (2001) noted, traditional partisan explanations usually assume a uniform policy stance based on party family, neglecting party system properties and only considering the ideological composition of the governing coalition. He further argued that parties belonging to the same party family might hold very different policy positions and pursued different coalition building strategies based on the structure of competition in the party system. Based on his study on the politics of German unemployment insurance and social assistance reforms, Picot (2009) observed that parties were flexible with regard to their preferred policies – in order to appeal to important groups of voters – based on the spatial configuration of the party system at the time, and did not necessarily pursue static ideological policy programmes. Green-Pedersen (2002) stated that governments were also punished for the mismanagement of the economy (e.g. high unemployment, fiscal deficit, growth crisis). That is, there are electoral incentives for retrenchment and liberalisation 92

regardless of the colour of the government. The options for justification strategy for retrenchment are contingent on the spatial composition of the party system (Green-Pedersen 2002). Picot’s (2009) analysis of the German Harz IV reform of the unemployment insurance and social assistance schemes also drew attention to the fact that politics is dynamic. That is, a reform ex-post can alter the very structure of party competition that allowed for its implementation ex-ante. Based on the model of Avdagic, I expect that the government’s policy preferences and union strategies are developed and revealed during their interactions and informed by their former interactions in the context of structural constraints. The labour-government interactions during 1999 and 2011 are analysed through newspaper articles covering the negotiations (also statements and interviews) during the periods between the time when the intentions of the government for reform was revealed (announced) until the adoption of the new legislation. I used Népszabadság, which was the biggest circulation national political daily paper of the country (until its abrupt closure by a new owner close to Orbán in 2016). Népszabadság was, despite its left-liberal ideological stance, an independent paper adhering to basic editorial standards for quality journalism. The main conservative daily paper, Magyar Nemzet was owned by a long-time Orbán associate and the foremost oligarch of the country in both periods, functioning as a government mouthpiece. My analysis brings in party system properties as decisive contextual variables (structural constraints) in explaining the variance between the outcomes of the 1999 and 2011 reforms. The party system properties that are analysed in the two case studies are the relative strength of the left (MSZP) and the right (Fidesz) (following Armingeon 2012), and the credibility of the policy options represented by the opposition and its organisational flexibility (Kitschelt 2001). Kitschelt (2001) also considers the salience of economic-redistributive issues as an important enabling factor of retrenchment, as the salience of socio-cultural issues – based on his Western cases and Japan – signals a consensus around the welfare policy status quo. However, the salience of socio-cultural issues might just as well signal that liberalisation is a valence issue in the party system (as was the case in Hungary at least between 1990 and 1994 or in Poland between 1989 and 1993). The real question is thus not whether socio-economic or socio-cultural issues are (more) salient in a party system at a given time period, but to what extent the spatial composition of the party system allows for a departure from the status quo (regardless whether it is against or for liberalisation).

93

3

Dissolution of tripartism and liberalisation of the labour code in Hungary 1998-2002 and 2010-2014

3.1

Structural changes in social dialogue and liberalisation of the labour code 19982002

Between 1988 and 1998 tripartite social dialogue was institutionalised in Hungary. The first national level tripartite forum, the National Council of Reconciliation of Interests (OÉT – Országos Érdekegyeztetési Tanács) was set up in 1988 by the last Communist government. The first democratic government set up the Interest Reconciliation Council (ÉT – Érdekegyeztetési Tanács) on August 1. 1990 (Lux 2003). The ÉT proved to be essential in managing the transitional crisis (Bohle and Greskovits 2012). However, after 1994 the ÉT gradually lost its significance. The new socialist-liberal government and the social partners failed to conclude the Social and Economic Pact in December 1994, and a harsh stabilisation package was introduced without any prior negotiation with the social partners in March 1995 (Tóth 2001). However, in the subsequent years the Social and Economic Pact was concluded. Income policy agreements were concluded in 1991, 1992, 1993, 1994, and in 1997. The ÉT was responsible for minimum wage setting and it was the forum for national wage setting negotiations. Moreover, its statutes mandated the government to introduce the budget and tax bill proposals in the ÉT. Finance ministers, that is, introduced the budget and tax policy bills to the National Assembly only after the negotiations with social partners in the ÉT (Héthy 2000, 81-82). The first Orbán-government (July 1998 – April 2002) implemented profound changes to the tripartite interest intermediation structure, chipping away at the rights of organised labour and social dialogue. In April 1999 the ÉT was dissolved, and two new interest intermediation fora were set up, the National Labour Council (OMT – Országos Munkaügyi Tanács) and the Economic Council (GT – Gazdasági Tanács). The GT was not a tripartite body, as beside unions and employers’ organisations, chambers of commerce and industry, banks, investors and multinational corporations were also invited (Héthy and Ónody 2008; Nacsa and Tóth 2004). On paper the reform brought two main changes. First, budget and tax bills were not discussed in the OMT, the successor of the ÉT but in the GT. Second, an amendment to the labour code gave the opportunity to the government to set the minimum wage in lieu of a tripartite agreement (Héthy and Ónody 2008; Lux 2003). Otherwise, the structure of the OMT and its functions very much resembled those of the ÉT. However, interest reconciliation was emptied (Héthy 2001). Income policy was no longer negotiated, and the minimum wage was set by the government (Héthy 2001; Lux 2003). When the GT was convened, no interest reconciliation took place on macroeconomic policy issues. As one of the participants complained, cited by Héthy (2001), the GT lacked background structures and they saw no intentions from the government’s part to set one up. Indeed, the GT was not even a consultative body, its meetings 94

were nicknamed “audiences”. The participants’ role was to listen to ceremonial speeches held by government representatives (Héthy 2001, 62-63). The first Orbán government also amended the labour code and adopted a new wage setting mechanism. The April 1999 labour code amendment introduced three important changes. Works councils were given the right to conclude so-called works agreements when there was no collective agreement in place (these could not regulate wages and other forms of pay only the terms and conditions of employment). The obligation of employers to consult an employee before a labour lawsuit was abolished. Finally, the fine for the unlawful termination of an employment contract was reduced. A new wage setting mechanism was introduced, which linked real wage growth to GDP growth. Real wage growth was set not to be higher than 50 per cent of GDP growth on a yearly basis (Népszabadság, 8 April 1999). 3.2

Government-union interactions and partisan politics March-April 1999

The Hungarian union structure was very fragmented: there were six union confederations, four of which were successors to the Communist union confederation (SZOT – Szakszervezetek Országos Tanács, National Council of Trade Unions), and two anti-Communist ones. There was no single Communist vs. anti-Communist cleavage as in Poland, both the successor and the new union confederations had several conflict lines between them (Tóth 2001). As Avdagic (2005) described, one of the main conflicts between the successor unions was how to divide the valuable assets of the Communist union confederation. Moreover, old and new unions alike tried to grow by poaching each other’s members, which led to bitter conflicts (Neumann 2005). Among the new unions, the LIGA confederation was founded first, it took part in the democratic transition on the side of the opposition. LIGA was initially close to the liberal party, the SZDSZ (Alliance of Free Democrats). Meanwhile, the National Association of Workers’ Councils (MOSZ – Munkástanácsok Országos Szövetsége) was formed under the auspices of the leading party of the first democratic government – the centre-right MDF (Hungarian Democratic Forum) – in the summer of 1990 (Tóth 2001, 40). However, the two new confederations remained relatively small, particularly MOSZ. LIGA only began to catch up with the main Communist successor union confederation, the National Confederation of Hungarian Trade Unions (MSZOSZ – Magyar Szakszervezetek Országos Szövetsége) and its splinter confederations by the end of the 1990s. The MSZOSZ, consisting of industry and private-sector unions, remained the biggest confederation until the end of the decade, however it was then overtaken by the Co-operation Forum of Trade Unions (SZEF – Szakszervezetek Együttműködési Fóruma), which was established by public-service and civil servants’ unions. Another splinter confederation was the Association of Academic Employees’ Unions (ÉSZT – Értelmiségi Szakszervezetek Szövetsége), and the Federation of Autonomous Unions (ASZSZ – Autonóm Szakszervezetek Szövetsége), consisting of the Union of Chemical Workers and public utilities unions (Tóth 2001; Neumann 2005). 95

The structural positions of unions rapidly deteriorated. Union density rates were in a free fall (Figure 1). According to Visser (2016) the proportion of employees covered by collective bargaining and statutory regulations as proportion of all workers stood at 41,4 per cent in 2000. According to data of the MSZOSZ, the collective agreement coverage in the private sector was rapidly shrinking: while in 1995 local collective agreements still covered 23 to 25 per cent of employees, the share fell to a mere 19.2 per cent by 1997 (Tóth 2001, 50).

Figure 1: Union density rates, Hungary 1990-2001 Note: Net union membership as a proportion of wage earners in employment; Source: Visser 2016.

The negotiations between the unions and the government about the new structure of social dialogue and the amendment of the labour code took place through March and April 1999. On March 3 the unions stated that they were ready for a compromise with the government, but they could not accept a mere consultative role instead of social dialogue. The unions also emphasised that the peak level interest intermediation should include budget and tax laws, and social insurance. Unions were very critical of the plans to give works councils the right to conclude workplace agreements (Népszabadság, 4 March 1999). In the second round of negotiations on March 19 the government stepped up the pressure on unions. The government announced that if unions would accept that real wage growth was to be set at 50 per cent of GDP growth for the coming years, it would consider union demands regarding the Labour Code. It also announced that the Interest Mediation Council, the ÉT, should be dissolved on April 9 and the new economic council, the GT, would be founded. The unions rejected the proposal of the government and demanded that ÉT should be convened to conclude an income agreement (Népszabadság, 20 March 1999). Meanwhile, the public sector unions staged a huge demonstration in Budapest because the government refused to include them in wage setting – which the former government did between 1996 and 1998 (Neumann 2005) – and generally 96

depressed the wages in the public sector. All union confederations supported the demonstration (Népszabadság, 22 March 1999). On the next negotiation round, on April 2 it seemed as if there was a basic agreement on the new structure of social dialogue: GT would be the forum for macroeconomic issues, but the successor of ÉT, the National Labour Council (OMT) would be consulted on the budget and tax policy by the Finance Ministry. The unions still refused the proposed wage setting mechanism, but declared that they were ready to conclude an agreement on the new social dialogue structure (Népszabadság, 3 April 1999). However, the government arrived with an agreement to the next round on April 6, which contained the new wage setting mechanism. The social partners could not sign such an agreement. As there was no agreement until the April 6 government-set deadline, the government shrugged and introduced its original package about the reform of social dialogue and the amendments to the Labour Code in the National Assembly without any compromise. According to unnamed sources, the government pursued a noagreement scenario all along (Népszabadság, 7 April 1999). The first, founding session of the GT was convened on April 9 as originally planned. The day before the union confederations engaged in a fierce debate until late in the evening whether to participate. MSZOSZ and ASZSZ argued against participation because they did not want to legitimise the new forum. However, as the six confederations wanted to reach a common position, and the other four confederations wanted to participate, eventually they all decided to go. The trade unions declared once again that they were ready to sign the April 2 agreement on the new structure of the dialogue (Népszabadság, 9 April 1999). On April 9 the trade unions met with a triumphant Viktor Orbán during the founding session of the GT. Although he had been prime minister for almost a year, this was the first time Orbán met with the social partners. He compared the former structure of social dialogue to a “jungle war” and declared that social dialogue had a new structure. He pushed the responsibility of not reaching a compromise on the unions, and refused to reconsider the April 2 agreement (Népszabadság, 10 April 1999). A couple of days later MSZOSZ called for a joint action and a united front against the government, however, in vain (Népszabadság, 14 April 1999). On April 20, the day before the founding session of the successor forum to the ÉT, the OMT, the confederations again had a bitter debate about whether to participate. This time MSZOSZ stood alone with a firm no, the MOSZ leadership was divided, whereas the other four confederations argued for participation (Népszabadság, 21 April 1999). All the confederations participated on the founding of the OMT on April 21. However, the unions did not approve of the dissolution of the ÉT, they merely acknowledged the change. During the negotiations the government exploited the structural weakness of unions and the internal divisions of the confederations. The government was intent on taming social dialogue and give it only a consultative role, whereas trade unions were eager to have a seat at the table. Since they could not form a united front and boycott the government’s new institutions, they had no chance to come out as winners from the conflict. To give works councils the right to 97

conclude works agreements was a cornerstone of the government’s strategy, as it was a means of exacerbating the inherent conflict at the centre of the dual Hungarian bargaining structure. Works councils and unions had parallel consultation and information rights, but unions depended on the works council election results in order to have the right to conclude a collective agreement (Tóth 2000, 318-319). MSZOSZ took the hard stance, however, it was discredited because during the former socialist–liberal (MSZP-SZDSZ) coalition government, which introduced a harsh austerity package unilaterally in 1995 and a pension privatisation in 1997-1998, it was much more lenient. The successor union SZEF led the protest against the drop in real wages in 1995-1996 (Neumann 2005). MSZOSZ during the 1990s pursued the classic social democratic model. It had formal ties with the reformed post-Communist MSZP party, and union leaders from MSZOSZ were MSZP deputies in the National Assembly between 1994 and 2002. The socialists co-opted the MSZOSZ leadership during their term in government. The socialists, however, lost the elections in 1998, and one of the main reasons of Orbán’s enmity towards unions was the fact that the biggest confederation was allied with the MSZP. On the founding session of the GT, he openly told them that “they bet on the wrong horse” (Népszabadság, 10 April 1999). The party system properties, however, made Orbán’s anti-labour policies electorally riskier than they seemed at the time of implementation. The socialists heavily criticised the government and vowed to reverse all the reforms regarding social dialogue of the Orbán government. By April 1999 the socialists were closing the gap polling at 25 per cent compared to Fidesz’ 27 per cent (Népszabadság, 2 April 1999). Fidesz, that is, faced an equally strong, socially embedded left-wing contender. Moreover, the cleavage structure was not as sheltering either against an electoral backlash because of social and economic policy reforms. The political significance of conflicts over economic policy and redistributive issues was on the rise. Fidesz implemented a number of left-wing measures after taking power: it reversed the most symbolic measures of the MSZP-SZDSZ austerity package, such as means-testing for family allowance or modest university tuition fees, and partially reversed the 1997 pension reform. But otherwise it pursued a markedly conservative and right-wing economic policy, with permanent austerity for the private sector (particularly in real wage growth), restraint in pension indexation, and tax cuts for middle and high-income families. It also tightened the access to and reduced the generosity of unemployment benefit and social assistance (Armingeon et al. 2019). As a result, whereas in 1998, left-wing economic policy attitudes had a significant and negative effect on the vote choice for MSZP, in 2002, after four years of Fidesz-rule, having an economically leftist position had a positive and significant effect for the vote choice for the socialists. This relationship is all the more worth noting, because except for an overrepresentation of women in the MSZP electorate, only having strong anti-Communist or nationalist views and being a frequent churchgoer had a significant (and negative) effect on the 98

vote choice for MSZP. Age, class and rural living did not matter anymore – as the two big parties have become catch-all parties (Tóka and Popa 2013, 317-318, Table 10.1). To sum up, in 1999 the unions were fragmented, and their structural positions were rapidly deteriorating: collective bargaining coverage and union density have been steadily shrinking since the beginning of the decade. The Orbán government were able to exploit the differences between the unions. Lacking a united front and faced with a disciplined and coherent government, the unions had no choice but to acknowledge the changes to at least maintain consultative rights with the government. The formal alliance between the biggest confederation, the MSZOSZ and the socialists was, however, not only a burden, but also a power resource: union grievances were high on the socialists’ political agenda, and eventually became part of their electoral platform too. Although the main opposition party, the socialists were marred by their own liberalisations in the previous cycle, and Fidesz used the justification strategy of economic competitiveness and political efficiency (and even questioned the legitimacy of unions), the spatial composition of the party system did not allow for a full liberalisation. The socialist party proved to be a credible and viable pro-labour, left-wing competitor. 3.3

Structural changes in social dialogue and liberalisation of the labour code 20102014

The second Orbán-government went much further in its attack against social dialogue and employment protection. After the MSZP won the 2002 elections, one of the first things it did was to reverse the structural changes to social dialogue described above. A new tripartite body was set up, the National Interest Reconciliation Council (OÉT – Országos Érdekegyeztetési Tanács), with the functions of the dissolved ÉT (Armingeon et al. 2019). After eight years in opposition, Fidesz and its satellite party (KDNP) achieved a landslide victory in 2010 and enjoyed a two-thirds majority in parliament. In the summer of 2011, tripartite social dialogue was effectively dissolved. The law establishing the National Economic and Social Council (NGTT – Nemzeti Gazdasági és Társadalmi Tanács) was passed on 4 July 2011 in the National Assembly and entered into force January 2012 (Komiljovics 2011). NGTT is not a tripartite body: beside unions and employers’ organisations, it consists of representatives of trade and industry chambers, various social and scientific civic organisations, and even churches. It does not include government representatives; ministers only participate as observers. The NGTT can merely draft proposals for the government (Komiljovics 2011). In February 2012, a new consultative body was, however, launched by the government, the Permanent Consultation Forum of the Private Sector and the Government (VKF – Versenyszféra és a Kormány Állandó Konzultációs Fóruma). It is a tripartite body, where employers’ organisations and private sector trade unions are consulted on economic (including taxes) and labour market policy decisions that affect the private sector (Napi.hu, 22 February 2012). 99

The new labour code was adopted in December 2011 and came into full force in January 2013 (Krén and Rindt 2013). The most profound change it implemented was the reconceptualisation of employment relationships as a civil law contract, assuming equal power between employees and employers (Tóth 2012; Szabó 2013). That is, it allowed for individual work contracts and collective agreements to set lower standards than the Labour Code (Tóth 2012). The new Labour Code gave bigger discretion to employers over their employees, and it pushed responsibility for work-related risks from employers onto employees:  it increased the possible period of probation to three months, and within a collective agreement it allowed for a maximum of six months;  it allowed for a termination of employment during sick leave;  it was no longer obligatory for employers to reinstall unlawfully dismissed employees;  it stipulated lump-sum severance payment on the basis of the basic salary (not on average income);  the statutory maximum for overtime was raised from 200 to 250 hours a year, and within a collective agreement it could be increased to 300 hours;  in case of un-avoidable external influences, such as a power cut, employees are no longer entitled to their basic salary, and employers are only liable for damage occurring within their overseeing capacity;  employees handling money or valuables could be made to pay a guarantee deposit as high as a monthly salary (Krén and Rindt 2013). The labour code significantly cut the entitlements and rights of union activists. Just as in 1999, works councils were given the right to conclude works agreements (Krén and Rindt 2013). Together these changes amounted to a radical re-conceptualisation and liberalisation of employment relations and employment protection in Hungary. 3.4

Government and union strategies, and partisan politics April-December 2011

The structural position of unions deteriorated further during the 2000s. Density rate has fallen from 20 per cent in 2001 to just 10,7 per cent in 2011 (Figure 2). Total collective agreement coverage also shrunk from 41,4 per cent in 2000 to 23,9 per cent in 20112 (Visser 2016). Moreover, the union confederations could not find a solution for the fragmentation. The SZEF and the ÉSZT set up an umbrella organisation in 2000, but kept their own organisations and independence, and their co-operation was limited to the coordination of joint actions (Neumann 2005). The biggest unions were MSZOSZ, LIGA, and SZEF. However, SZEF as a public sector union confederation was shut out from collective bargaining and social dialogue by the Orbán government after 2010.

2

Employees covered by collective bargaining and statutory regulations as a proportion of all workers.

100

Rumours about government plans to dissolve tripartite social dialogue were circulating already in 2010. The government denied that it had such intentions. Nevertheless, Orbán himself visited the OÉT last in September 2010, and it had not been convened since the beginning of 2011 (Népszabadság, 11 October 2011). By the spring of 2011, however, government plans to amend the Labour Code were announced as part of an effort to cut red tape to make the economy more competitive. Moreover, by 2011 the government already implemented several laws criticised by the unions: with January 2011 the right to strike had been heavily curtailed (Rindt 2012).

Figure 2: Union density rates, Hungary 2001-2012 Note: Net union membership as a proportion of wage earners in employment; Source: Visser 2016.

Furthermore, the government reversed the 1997 pension reform completely, seizing the pension savings in private pension funds of millions of employees (Armingeon et al. 2019). The government strategy was already set in April: the government refused to discuss its planned labour law legislation with the social partners. Instead, it involved certain employers’ and business’ organisations, and companies in drafting the law. On the other hand, while effectively shutting down social dialogue, it executed direct control on income policy. On April 29 the government announced that the government will adopt legislation that will set an expected rate of wage increases, particularly for low earners (Népszabadság, 30 April 2011). The government also launched a “social consultation” directly with the people in a form of a questionnaire sent out to every citizen to provide legitimacy to the sweeping social policy and labour law changes it was about to introduce. The questions and the answers were all leading, formulated in a way that the respondents had no choice but to support the government workfare, anti-welfare agenda.

101

The government eventually successfully divided the union confederations by the summer of 2011. On 8 April 2011, as the ECOFIN had its meeting near Budapest the European Trade Union Confederation held a large, 40-50.000 strong demonstration downtown Budapest. All six Hungarian trade unions confederations marched together, and all of the union leaders but the president of MOSZ personally attacked Orbán and criticised government policies (Népszabadság, 9 April 2011). Yet, by May 1, the unity was gone. The confederations did not celebrate together. The president of LIGA, after meeting with Orbán later in April, said that the activism of MSZOSZ was not credible because it cozied up to the former MSZP-government (Népszabadság, 2 May 2011). In the second half of May the government promised to consult the trade unions after the “social consultation” ended (Népszabadság, 20 May 2011). At the end of May, however, the government announced that it will dissolve the OÉT and set up a new peak consultative forum, the National Economic and Social Council (NGTT – Nemzeti Gazdasági és Társadalmi Tanács) (Népszabadság, 28 May 2011). In June, the government introduced its final draft amendment to the Labour Code for the debate in the parliament, which was seen as a small win for labour at the time, as it was less radical than originally intended (6 June 2011). On July 4 Orbán announced that according to the “results” of the social consultation it is clear that the Hungarians did not want “free beer” (sic!) and approved of the government vision of a workfare society. The government was only negotiating at this point with LIGA, and the president of LIGA was confident that an agreement was close. Meanwhile, the National Assembly adopted the act founding the NGTT on 4 July 2011 (Népszabadság, 5 July 2011). The NGTT founding session took place on October 11, 2011. Orbán declared that it was a fallacy that only employees and employers would be allowed to have an opinion in the world of labour. The government announced that it will set the minimum wage and the guaranteed wage minimum unilaterally (Népszabadság, 11 October 2011; Népszabadság, 12 October 2011). The government held consultations about the new Labour Code with unions during November 2011 – there was also international pressure from the ILO to tone down the most extreme proposals (Komiljovics 2012). However, only LIGA, MSZOSZ and MOSZ were involved in the consultations. The agreement signed on December 1 gave some minor concessions, particularly on union rights (Népszabadság, 3 December 2011). Party system properties also were favourable for the government. In 2011 Fidesz faced a divided opposition: the MSZP, a new green party, LMP (Politics Can Be Different), and the far-right Jobbik. These parties were all much smaller in size than Fidesz, which received 52.7 per cent of the popular vote and won 173 of the 176 single-member districts in April 2010. The MSZP came in as a distant second with 19.3 per cent, Jobbik got 16.7 per cent, while LMP 7.5 per cent. The opposition parties were all mutually exclusive: both the LMP and the Jobbik considered MSZP to be corrupt and part of the “neoliberal” system, whereas both the LMP and the MSZP rejected cooperation with the Jobbik because of its overt racism. In the October municipal elections Fidesz struck a final blow to the socialists: Fidesz controlled 22 out of 23 of the biggest cities of the country (including Budapest) and every county assemblies (Várnagy 102

2011). Fidesz seized the opportunity to shape the country to its vision. Orbán talked about an electoral revolution and declared the founding of the “System of National Cooperation” (NER - Nemzeti Együttműködés Rendszere). The National Assembly became a law factory: it adopted a staggering 365 laws in 2011 (between 2006 and 2010 the complete legislative record of the National Assembly amounted to 591 laws) (Várnagy 2012). The many conflicts and the unilateral manner of Fidesz-rule had consequences. By May 2011 Fidesz lost 1/3 of its supporters (Népszabadság, 13 May 2011). However, the socialists – similarly to other opposition parties – were unable to profit from this, their support stagnated the whole year. The MSZP was in deep crisis. It was discredited because of corruption scandals and its handling of the 2008-2009 crisis. Furthermore, the MSZP was in an internal strife, which ended with their former president and prime minister leaving the party along with his supporters, founding his own party in October 2011. MSZP lost most of its social embeddedness, and not only because of its electoral losses. Fidesz spent the 8 years in opposition building up its societal base through grass root activism, and national campaigns and referenda, whereas the MSZP was in decline, its voter base rapidly ageing. While in government, the socialists’ reform agenda was not only opposed by Fidesz, but often by their own national and local politicians. Neither had the MSZP and MSZOSZ any formal or informal alliance anymore, in contrast with the period between 1998 and 2002. To sum up, by 2011 the positions of the unions deteriorated significantly compared to 19992000: both union density and collective bargaining coverage almost halved. Unions had not been able to mitigate the fragmentation either. During the period of negotiations, the government was able to play favourites with LIGA and MOSZ, and eventually split the unions by only negotiating with three private sector union confederations. Public sector unions were excluded from social dialogue altogether. This time there was real pressure for reform and austerity in wake of the Great Recession with the country still under EU and IMF conditionality. These external pressures together with the landslide victory of Fidesz in April 2010 with a mandate to implement profound changes, provided legitimacy to the reforms. Moreover, the ideologically coherent and disciplined anti-welfare government enjoyed a super-majority in parliament. Orbán faced a disintegrated and discredited left as a part of an ideologically fragmented opposition composed of mutually exclusive parties. In absence of a credible, viable left-wing contender nothing checked the government’s anti-labour political agenda, and the weakened and fragmented unions were not able to withstand the government’s push for radical liberalisation alone.

4

Conclusion

The findings of the paper lend support to Avdagic’s (2006) dynamic framework in explaining state-labour relations and actual labour policy outcomes. The comparative analysis showed that labour policy outcomes and the trajectory of change in the institutional setup of social dialogue 103

are not explained by static institutional or macroeconomic variables. Just as Avdagic (2006) claimed, no linear relationship could be found. Labour policy and state-labour relations are subject to the relative strength of governments and unions, which is a dynamic relationship and contingent on institutional and political constraints. The careful reconstruction of the government-union interactions with the help of contemporary newspaper articles, interviews, and statements in the case studies have shown that the government’s strategy and the unions’ responses were very similar in 1999 and 2011. However, in 2011 the union confederations were much weaker as both their structural and political positions deteriorated relative to 1999. Moreover, the 2010 strike law took care of their mobilisation capacities: between 2010 and 2014, facing the biggest threat against organised labour since 1990, there were 20 incidents of strikes, six of which were only warning strikes, whereas between 2006 and 2008 there were more than 40 (Neumann 2018). Unions were unable to overcome their fragmentation and the resulting collective action problem. The government strategy was informed by the previous interactions just as the model predicts: Orbán learned the lesson and divided the unions. Fidesz singled out and favoured LIGA and MOSZ over other confederations (Neumann 2018). While shutting down tripartite institutions, the government introduced direct state intervention in income policy, offering an alternative to social dialogue. However, by focusing on party system properties instead of the colour of government following Kitschelt (2001), Green-Pedersen (2002) and Pico (2009), the analysis showed that partisan politics proved to be a decisive structural factor in labour policy outcomes, just as Armingeon (2012) found. In 1999, Fidesz faced a similar-sized left-wing competitor with formal ties to the biggest union confederation, the MSZOSZ, which had deputies in the MSZP parliamentary caucus. The social and labour policies of the Fidesz-government had a measurable effect. The MSZP was able to exploit the grievances of these markedly right-wing policies and successfully politicise economic and redistributive questions. In fact, already in 1999, the socialists increased their support relative to the Fidesz. In contrast, in 2011, Orbán faced an already discredited MSZP in disarray, which was bogged down by internal strife with many key politicians eventually leaving the party and forming a new one. Furthermore, both the MSZOSZ and the MSZP gave up on the social democratic model after 2002. The MSZOSZ chose to emphasise its independence, and the MSZP aimed to be a catch-all party. Unions and workers’ rights were not highlighted in its program anymore (Neumann 2018). Purely mathematically, in 2011, the three opposition parties combined still commanded an equal force to Fidesz, however, neither the socialist, green and far-right parties, nor their voter base could be added together. None of the opposition parties was able to profit from the fact that the Fidesz lost 1/3 of its supporters by the summer of 2011. The findings of this chapter lend support to the importance of conjunctural causation in explaining individual cases, as Armingeon (2012) suggests. For the radical liberalisation of social dialogue and employment protection legislation (and even taxation, as Fidesz introduced 104

a flat personal income tax after taking power in 2010), a strong right-wing party was indeed a necessary condition, but the effect was contingent on the collapse of the left.

References Armingeon, Klaus. 2012. “The Impact of Political Parties on Labour Relations. European Post-Communist Democracies in Comparative Perspective.” In Party Government in the New Europe, edited by Hans Keman and Ferdinand Müller-Rommel, 113-136, London: Routledge. Armingeon, Klaus, Lucio Baccaro, Anna Fill, Jorge Galindo, Stefan Heeb, and Rafael Labanino. 2019. “Liberalization Database 1973-2013.” Bern and Geneva (forthcoming). Avdagic, Sabina. 2005. “State-Labour Relations in East Central Europe: Explaining Variations in Union Effectiveness.” Socio-Economic Review 3 (1): 24-25. Avdagic, Sabina. 2006. “One Path or Several? Understanding the Varied Development of Tripartism in New European Capitalisms.” MPIfG Discussion Paper, No. 06/5, Köln: Max-Planck-Institut für Gesellschaftsforschung. Bernaciak, Magdalena. 2015. “Beyond the CEE ‘Black Box’: Crisis and Industrial Relations in the New EU Member States.” Working paper 2015.05, Brussels: European Trade Union Institute. Bohle, Dorothee, and Béla Greskovits. 2012. Capitalist Diversity on Europe’s Periphery. Ithaca: Cornell University Press. Green-Pedersen, Christopher. 2002. The Politics of Justification. Party Competition and Welfare-State Retrenchment in Denmark and the Netherlands from 1982 to 1998. Amsterdam: Amsterdam University Press. Héthy, Lajos 2000. “Országos jövedelempolitikai megállapodások. A múlt és a lehetséges jövő.” Közgazdasági Szemle 47 (January): 81-93. Héthly, Lajos, and Irén Ónodi. 2008. “Országos társadalmi párbeszéd - Az Országos Érdekegyeztető Tanács és a Gazdasági és Szociális Tanács.” Budapest: Foglalkoztatási és Szociális Hivatal. Häusermann, Silja, George Picot, and Dominik Geering. 2012: Review Article: “Rethinking Party Politics and the Welfare State: Recent Advances in the Literature.” British Journal of Political Science 43 (1): 221-240. Kitschelt, Herbert. 2001. “Partisan Competition and Welfare State Retrenchment: When Do Politicians Choose Unpopular Policies?” In The New Politics of the Welfare State, edited by Paul Pierson, 265-302. Oxford: Oxford University Press. Komiljovics, Máté. 2011. “Government to Stop Consulting Unions on Minimum Wage.” Dublin: European Foundation for the Improvement of Living and Working Conditions, 1 September 2011 https://www.eurofound.europa.eu/publications/article/2011/government-to-stop-consulting-unions-onminimum-wage. Komiljovics, Máté. 2012. “Unions Slam New Labour Code.” Dublin: European Foundation for the Improvement of Living and Working Conditions, 30 January 2012 https://www.eurofound.europa.eu/publications/article/2012/unions-slam-new-labour-code. Korkut, Umut, Alex de Ruyter, Manos Maganaris, and David Bailey. 2017. “What Next for Unions in Central and Eastern Europe? Invisibility, Departure and the Transformation of Industrial Relations.” European Journal of Industrial Relations 23 (1): 65-80. Krén, Ildikó, and Zsuzsa Rindt. 2013. “New Labour Code Takes Full Effect.” Dublin: European Foundation for the Improvement of Living and Working Conditions, 7 April 2013 https://www.eurofound.europa.eu/publications/article/2013/new-labour-code-takes-full-effect. Lux, Judit. 2003. “Az Országos Érdekegyeztető Tanácstól az Országos Érdekegyeztető Tanácsig.” Eszmélet, 57 (Spring) http://www.eszmelet.hu/lux_judit-az-orszagos-erdekegyezteto-tanacstol-az-orszagos-e/. Nacsa, Beáta, and András Tóth. 2004. “Social Dialogue and Conflict Resolution in Hungary.” Dublin: European Foundation for the Improvement of Living and Working Conditions https://www.eurofound.europa.eu/sites/default/files/ef_files/pubdocs/2004/47/en/1/ef0447en.pdf. Neumann, László. 2005. “Trade Unions in Hungary: Between Social Partnership, Political Action and Organising Drive.” In Trade Unions Strategies in Central and Eastern Europe: Towards Decent Work, edited by Dimitrina Dimitrova and Jacques Vilrokx, 63-109. Budapest: International Labour Office. Neumann, László. 2012. “Hungarian Unions: Responses to Political Challenges.” Management Revue 23 (4): 369385. Neumann, László. 2018. “Hungarian Trade Unions from the Beginning of Transition to the New U-turn.” In Trade Unions in Transition - from Command to Market Economies, edited by Rudolf Traub-Merz and Tim Pringle, 317-346. Berlin: Friedrich Ebert Stiftung. 105

Napi.hu, 22 February 2012. “Megalakult a versenyszféra érdekegyeztető testülete.” Napi.hu https://www.napi.hu/magyar_gazdasag/megalakult_a_versenyszfera_erdekegyezteto_testulete.511398.html Népszabadság. 4 March 1999. “Alkut ajánlanak a munkavállalók.” Népszabadság. Népszabadság. 20 March 1999. “ÉT helyett Gazdasági Tanács?” Népszabadság. Népszabadság. 22 March 1999. “Tüntettek a közszféra dolgozói.” Népszabadság. Népszabadság. 2 April 1999. “Bisonytalanodó kormánypárti támogatás.” Népszabadság. Népszabadság. 3 April 1999. “Elvi egyezség a párbeszéd átalakításáról.” Népszabadság. Népszabadság. 7 April 1999. “Nincs alku a társadalmi párbeszédről.” Népszabadság. Népszabadság. 8 April 1999. “ÉT-megszüntetés: a szakszervezetek még egyeztetnek.” Népszabadság. Népszabadság. 9 April 1999. “A szakszervezetek ott lesznek a GT ülésén.” Népszabadság. Népszabadság. 10 April 1999. “Lezárult az érdekegyeztetés korábbi rendszere.” Népszabadság. Népszabadság. 14 April 1999. “Az MSZOSZ egységes fellépést sürget.” Népszabadság. Népszabadság. 21 April 1999. “Búcsú az ÉT-től, éljen az OMT.” Népszabadság. Népszabadság. 9 April 2011. “Európa ez egyáltalán?” Népszabadság. Népszabadság. 30 April 2011. “Elvárják a cégektől a béremelést.” Népszabadság. Népszabadság. 2 May 2011. “Fenn kell tartani az akcióegységet!” Népszabadság. Népszabadság. 13 May 2011. “Bisonytalan többség.” Népszabadság. Népszabadság. 20 May 2011. “Egyeztetés a nemzeti konzultáció után.” Népszabadság. Népszabadság. 28 May 2011. “Megszűnik az OÉT.” Népszabadság. Népszabadság. 6 June 2011. “Gyomros helyett füles.” Népszabadság. Népszabadság. 30 June 2011. “Felkészülés a forró őszre.” Népszabadság. Népszabadság. 5 July 2011. “A magyarok nem akarnak ‘ingyensört’.” Népszabadság. Népszabadság. 11 October 2011. “Biodíszlet a kormányszínházhoz.” Népszabadság. Népszabadság. 12 October 2011. “Orbán: Ha nincs kijárat, ki kell törni!” Népszabadság. Népszabadság. 3 December 2011. “Hatból csak hárman.” Népszabadság. Picot, George. 2009. “Party Competition and Reforms of Unemployment Benefits in Germany: How Small Changes in Electoral Demand Can Make a Big Difference.” German Politics 18 (2): 155-179. Prugberger, Tamás, and Andrea Szöllös. 2012. “The New Labour Code in Hungary in the Mirror of the Regulations of the European Union and EU Member States.” SEER Journal for Labour and Social Affairs in Eastern Europe 4/2012: 493-507. Rindt, Zsuzsa. 2012. “Amended Strike Law One Year On.” Dublin: European Foundation for the Improvement of Living and Working Conditions, 11 July 2012 https://www.eurofound.europa.eu/publications/article/2012/amended-strike-law-one-year-on. Szabó, Imre. 2013. “Between Polarisation and Statism - Effects of the Crisis on Collective Bargaining Processes in Hungary.” Transfer 19 (2): 205-213. Szabó, Imre. 2017. “Hungary: Inertia of the Old Actors, Constrained Innovation from the New.” In Innovative union practices in Central-Eastern Europe, edited by Magdalena Bernaciak and Marta Kahancová, 91-110. Brussels: European Trade Union Institute. Scharle, Ágota, and Dorottya Szikra. 2014. “Recent Changes Moving Hungary Away from the European Social Model.” In The European Social Model in Crisis. Is Europe Losing its Soul? edited by Daniel VaughnWhitehead, 289-338. Cheltenham-Northampton: Edward Elgar, Geneva: International Labour Office. Szikra, Dorottya. 2014. “Democracy and Welfare in Hard Times: The Social Policy of the Orbán Government in Hungary between 2010 and 2014.” Journal of European Social Policy, 24 (5): 486-500. Szikra, Dorottya, and Béla Tomka. 2009. “Social Policy in East Central Europe: Major Trends in the Twentieth Century.” In Post-Communist Welfare Pathways, Theorizing Social Policy Transformations in Central and Eastern Europe, edited by Alfio Cerami and Pieter Vanhuysse, 17-34. Palgrave Macmillan. Tóka, Gábor, and Sebastian Popa. 2013. “Hungary.” In The Handbook of Political Change in Eastern Europe. Third Edition, edited by Sten Berglund, Jokaim Ekman, and Kevin Deegan-Krause, 291-339. Cheltenham: Edward Elgar. Tóth, András. 2000. “Attempts to Reform a Workers’ Movement without Mass Participation.” In Trade Unions in Europe - Facing Challenges and Searching for Solutions, edited by Jeremy Waddington and Reiner Hoffmann, 305-338. Brussels: European Trade Union Institute. Tóth, András. 2001. “The Failure of Social-Democratic Unionism in Hungary.” In Workers after Workers’ States. Labor and Politics in Postcommunist Eastern Europe, edited by Stephen Crowley and David Ost, 37-58. Oxford: Rowman & Littlefield. Tóth, András. 2012. “The New Hungarian Labour Code - Background, Conflicts, Compromises.” Working Paper, Friedrich Ebert Foundation Budapest, June 2012 Várnagy, Réka. 2011. “Hungary.” European Journal of Political Research Political Data Yearbook 50: 991-998. Várnagy, Réka. 2012. “Hungary.” European Journal of Political Research Political Data Yearbook 51: 129-135. Visser, Jelle. 2016. “ICTWSS Data base. version 5.1.” Amsterdam: Amsterdam Institute for Advanced Labour Studies (AIAS), University of Amsterdam, Accessed at 17 September 2018. 106

Part II Politics of Welfare State Reform

Governing elites under pressure: Blame avoidance strategies and welfare state retrenchment Markus Hinterleitner1 and Fritz Sager2

Abstract: This chapter makes conceptual progress on an important puzzle in the research on welfare state retrenchment: Are welfare states retrenched because governing elites are astute blame avoiders or because important parts of the electorate have come to believe in the austerity medicine prescribed by mainstream economists? Existing research is, by and large, ill placed to address this question because it treats blame avoidance as a latent variable that is neither measured nor empirically observed. To cast some light on this important puzzle, this chapter develops a comprehensive categorization of blame avoidance strategies that governing elites can employ to avoid electoral punishment for retrenchment. The categorization identifies the mechanics of particular strategies and yields insights into their applicability, empirical traces, and effects on voters. In so doing, it can guide future research on the causal determinants of successful welfare state retrenchment. Progress on this puzzle promises important insights into how the governing elites of modern capitalist democracies juggle to reconcile domestic demands and supranational constraints.

1

Introduction

In 2011, the German chancellor, Angela Merkel, (in)famously remarked that democracies must be ‘market-conforming’ (Merkel 2011). While Merkel could have spared herself substantial trouble if she had made her point in a less blame-attracting way, she arguably described one of the major conundrums that the governing elites of modern capitalist democracies face on an almost daily basis. One, if not the, major reason for why governing elites – broadly understood here as heads of state, ministers, leaders of the parties in government, and top-level bureaucrats – have increasingly come under pressure in recent years is the far-reaching integration of modern states into the world economy. Their integration into global markets implies that governing elites must be responsive to two ‘constituencies’, instead of just one (Schäfer and Streeck 2013; Mair 2009). Traditionally, governing elites primarily had to be responsive to the demands of their voters. In democracies integrated into the world economy, governing elites must additionally address the demands of markets, which often conflict with those of voters

1 2

Kompetenzzentrum für Public Management, Universität Bern [email protected] Kompetenzzentrum für Public Management, Universität Bern [email protected]

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 R. Careja et al. (eds.), The European Social Model under Pressure, https://doi.org/10.1007/978-3-658-27043-8_7

109

(Hinterleitner et al. 2016; Sager and Hinterleitner 2016). Since the global financial crisis and the ensuing Euro crisis, governing elites must find ways to please and communicate with both constituencies in an especially difficult environment. As governing elites struggle to reconcile domestic demands with supranational constraints, a new kind of politics is emerging and awaits exploration (Schäfer and Streeck 2013). A major policy issue on which voter and market demands frequently collide is that of managing the welfare state. Welfare states, with their inherent tendencies to expand, are a bête noire in the eyes of global markets given that they contribute to national budget growth and thereby threaten complete and timely debt service – one of the key demands of global markets (Schäfer and Streeck 2013). To please the demands of global markets, governing elites must credibly signal their willingness for fiscal rectitude. This frequently involves welfare state retrenchment, i.e. the cutting of all kinds of social policy. However, unsurprisingly, voters are often skeptical about cutting back on public services, especially if they rely on them and have come to take them for granted. If governing elites do not meet their demands, voters tend to withdraw their loyalty (Schäfer and Streeck 2013). The large and diverse literature on welfare state retrenchment, to which Klaus Armingeon contributed substantially (Armingeon and Giger 2008; Armingeon et al. 2015), advances our understanding of how governing elites manage conflicting demands with regard to the welfare state. The literature suggests that governing elites employ a range of blame avoidance strategies to avoid electoral punishment for welfare state retrenchment. However, due to its large-N and comparative orientation, the literature does not usually zoom in on governing elites to study their actual strategic behavior under pressure. For the most part, blame avoidance remains a latent variable in research on welfare state retrenchment (Vis 2016). As a consequence, there are still major shortcomings in our understanding of how governing elites strike a balance between responsiveness to their voters and to supranational demands, and how this influences the trajectory of democratic welfare states. This chapter aims to address these shortcomings by better connecting the literature on welfare state retrenchment to the blame avoidance literature in public policy and administration research (Weaver 1986; Hood 2011; Hinterleitner and Sager 2015; Hinterleitner 2017). This literature adopts a different, more micro-level oriented approach that allows for the study of actual blame avoidance and its effects. Based on this literature, we develop a comprehensive categorization of blame avoidance strategies that deepens our understanding of the latent variable ‘blame avoidance’ and helps to advance the study of welfare state retrenchment on at least three fronts. The categorization reveals the conditions under which politicians can apply various blame avoidance strategies, the empirical traces that their application leaves behind, and their potential effects on voters. By zooming in on the applicability, empirical traces, and effects of particular blame avoidance strategies, our categorization offers the potential to create a more profound understanding of governing elites’ attempts to retrench the welfare state. Progress on this issue 110

not only helps illuminate one of the key challenges that governing elites face in modern capitalist democracies, but it also has broader implications for how to study and disentangle the determinants of policy outcomes. Political science, by and large, often struggles to determine the respective influences of voters and elites on policy outcomes (Hall et al. 2016). The next section provides a short overview of the literature on welfare state retrenchment. The overview concentrates on the blame avoidance variable and outlines the questions and puzzles that require deeper examination in order to study actual elite behavior. The second section introduces the blame avoidance literature in public policy and administration research and presents a categorization of blame avoidance strategies that governing elites apply to avoid electoral punishment for welfare state retrenchment. The third section elaborates on the payoffs of this categorization effort for future empirical work on welfare state retrenchment.

2

Blame avoidance as the ‘blind spot’ of the literature on welfare state retrenchment

Research on welfare state retrenchment originates in the seminal work by Paul Pierson (1994, 1996), which argues that the theories that explain welfare state expansion cannot automatically be applied to the study of welfare state retrenchment because the underlying political logic of retrenchment differs from a logic of expansion (Pierson 2001). In times of austerity, politicians struggle to reconcile policy necessities with their vote goals. When public budgets were not yet in dire straits, governing elites could claim credit for expansionary policy initiatives and thereby enhance their chances of getting re-elected (Weaver 1986). On the contrary, retrenchment initiatives, which are difficult to avoid in times of austerity (Armingeon et al. 2015), cause immediate and tangible losses for social groups, thereby threatening the vote goals of governing elites. Pierson sees blame avoidance as a way to reconcile the goals of governing elites in an altered fiscal context: Pursuing retrenchment while minimizing the likelihood and extent of electoral retribution (Hinterleitner 2017; Starke 2006). Pierson’s seminal work triggered a raft of research on welfare state retrenchment in different political and historical settings. Scholars particularly directed their attention to the conditions under which governing elites successfully pursued retrenchment initiatives (Hinterleitner 2017). Their work identified a number of factors that influence whether or not retrenchment is successful, such as the institutional fragmentation of a political system, the partisan composition of the government (Jensen and Mortensen 2014), or party political consensus (Green-Pedersen 2002). This research refines Pierson’s work by carefully determining the scope conditions of his theory. However, in doing so, it mostly adopts a comparative, macrolevel focus that studies welfare state retrenchment from a great distance, and the actual behavior of politicians slips from view. As a result, this research, for the most part, tacitly assumes that when politicians retrenched the welfare state, it was because they successfully applied blame 111

avoidance strategies. In cases where retrenchment did not take place, it was because governing elites applied blame avoidance in vain. The question resulting from this blind spot, and which was addressed by subsequent research, is whether politicians always, and under all conditions, (have to) engage in blame avoidance when they want or have to retrench the welfare state. In other words, are there situations in which unpopular reforms do not lead to electoral retribution because voters do not care, or are even appreciative of retrenchment? In a study of 18 OECD democracies, Klaus Armingeon and Nathalie Giger (2008) found that there is indeed no systematic punishment of governing elites that retrench the welfare state. In the wake of this finding, several studies explored the factors that condition electoral punishment for welfare state retrenchment, qualifying the assumption that retrenchment leads to electoral punishment under all circumstances and for all types of political actors. Accordingly, this finding raised doubts about whether or not politicians need blame avoidance strategies in order to successfully cut social policy (Giger and Nelson 2011, 2013; Schumacher et al. 2012; Nelson 2016). The two major findings from this research reveal that partisan composition of the government and various policy characteristics (as perceived by voters) condition electoral punishment (see the detailed discussion in Vis 2016). While religious and liberal parties face less electoral punishment (and can even obtain votes) for retrenching the welfare state, left parties are more likely to be punished if they cut back on social policy (Giger and Nelson 2011). Likewise, retrenching policies that are salient to or popular among constituents are especially likely to result in electoral punishment (Green-Pedersen 2002; Jensen 2012). However, plagued by the same ‘blind spot’ as previous work – failing to measure actual blame avoidance – this research does not answer the question of whether governing elites avoid electoral punishment because certain voter groups do not care about retrenchment (or even appreciate it) or because their blame avoidance efforts proved successful. The existing research does not provide a conclusive answer to this question because it mostly treats blame avoidance as a latent variable instead of actually studying blame avoidance and its effects at the level of individual politicians (Vis 2016). Armingeon and Giger (2008), for example, use the time frame during which entitlements are cut back in order to infer whether or not governing elites engage in blame avoidance. They assume that a long time span indicates that governing elites deliberately ‘phased out’ retrenchment. While this is a plausible conclusion, it is only possible to gather solid evidence for the posited relationship at the individual level. Likewise, Giger and Nelson (2013) study voters’ predispositions towards retrenchment in order to better understand under which conditions they punish governing elites at the ballot box. Using voters’ reactions to retrenchment, they infer the success prospects of the particular blame avoidance strategies adopted by the government. Again, while this is a plausible conclusion, it lacks strong evidence without the study of the actual blame avoidance behavior. Recent research on welfare state retrenchment zooms in on governing elites and studies their blame avoidance efforts by analyzing public statements and conducting interviews 112

(Wenzelburger 2011). This research reveals that governing elites are very creative when it comes to finding ways to avoid electoral punishment for retrenchment (see more on this below), and it suggests that it is not the actual but the perceived electoral risk that entices politicians to engage in blame avoidance (Wenzelburger 2014; Hinterleitner 2018). Focusing on governing elites’ perceptions helps to explain why they even engage in blame avoidance in situations where it would not be necessary because voters do not care or even acknowledge the need for retrenchment. Despite these important contributions, research on welfare state retrenchment still lacks answers to an important question. In situations where politicians manage to retrench the welfare state, is it because voters do not care (or even appreciate retrenchment) or because the politician successfully engaged in blame avoidance? In the words of Vis (2016, 123), “we do not know whether the lack of punishment is due to retrenchment not being electorally dangerous or because this electoral danger has been removed with a successful blame avoidance strategy”. While this question may seem like a minor obstacle for a research program that has made considerable progress over the last decades, it goes to the core of one of the main debates in current political science research: the relative influence of voters and governing elites on policy outcomes. Prominent scholars increasingly challenge the widespread assumption in political science that support from the majority of voters constitutes a sufficient explanation for policy outcomes (Beramendi et al. 2015). Instead, they emphasize the conditionality of the influence of voters on policy (Culpepper 2011; Hall et al. 2016). They urge the discipline to not only consider the ‘demand side’ of politics but also the ‘supply side’ of it, which consists of “strategic actors with communication and branding strategies” (Hall et al. 2016, 390; König and Wenzelburger 2014). To understand policy outcomes, like successful welfare state retrenchment, it is therefore necessary to consider how governing elites strategically react to and manage the demands of their voters.

3

What is blame avoidance?

For the literature on welfare state retrenchment, there is no avoiding the need to zoom in on governing elites to establish when, whether, and how they (successfully) engage in blame avoidance. Recent research identifies experimental research methods as a promising way to study the effect of blame avoidance strategies on voters’ perception of welfare state retrenchment (Wenzelburger 2014; Wenzelburger and Hörisch 2016). Meanwhile, there is also a need to develop an understanding of the applicability and application of blame avoidance strategies in real-life policy-making situations. Important preconditions for determining whether retrenchment is due to successful blame avoidance or to citizens’ indifference (or acknowledgement of the retrenchment) include knowledge about whether or not governing elites are able to apply a blame avoidance strategy in a certain context, and which empirical traces to look for in order to identify a given strategy. 113

To create knowledge about the applicability and application of blame avoidance strategies, it is important to first develop clear-cut categorizations of the various blame avoidance strategies that governing elites (can) employ to successfully retrench the welfare state. In recent years, scholars have developed several useful categorizations (e.g. Wenzelburger 2011; Giger and Nelson 2011; Vis 2016). However, most of them are not sufficiently comprehensive, do not yield clear indications on what to look for and where to look to identify and measure particular strategies, and they do not distinguish between anticipatory and reactive forms of blame avoidance (see below). To address these limitations, we draw on the blame avoidance literature in the fields of public policy and administration, which allows us to develop a categorization of blame avoidance strategies that maps and clearly distinguishes between the many strategies governing elites can apply in order to avoid punishment for cutting social policies. The concept of blame avoidance assumes that governing elites are reputation-conscious actors who apply various strategies to avoid and manage the blame that comes from their voters. According to Weaver (1986), who pioneered the study of blame avoidance, governing elites engage in blame avoidance because of their loss aversion and their concern toward the negativity bias of voters – i.e. the propensity of voters to be more sensitive to real or potential losses than to gains (Lau 1985; Hinterleitner 2017). We can thus expect governing elites to engage in blame avoidance whenever blame is generated by political opponents, the media, or the public threatens their reputation (Hinterleitner 2017; Busuioc and Lodge 2016; Weaver 1986). In this context, it is useful to think of reputation as a form of political capital that allows governing elites to pursue and reconcile their (re-)election and policy goals. Governing elites that have a strong reputation are less likely to become the target of blame attacks in the first place, and they are more resilient to blame attacks if they do occur. A strong reputation thereby facilitates staying in office even in situations where policy goals are incompatible with the preferences of voters (Hinterleitner and Sager 2018). In short, blame avoidance helps governing elites to protect their reputation, and a strong reputation helps them to reach their (re-)election and policy goals (and makes them compatible in case they are contradictory). A look at the reputation-protecting and reputation-enhancing effects of blame avoidance strategies reveals an important (and hitherto neglected) distinction between various forms of blame avoidance. Governing elites not only protect and enhance their reputation in crisis situations; they also do so before a crisis strikes. This is why it is important to distinguish between anticipatory and reactive forms of blame avoidance (Hinterleitner and Sager 2017; Sulitzeanu-Kenan and Hood 2005). If governing elites anticipate a potentially blame-generating political decision, they can prepare for it in a variety of ways. The strategies that governing elites employ to prepare for blameworthy events are very different to those that they apply in the midst of a ‘blame game’ (Hinterleitner 2018). So called ‘agency strategies’ and ‘policy strategies’ can predominantly be applied in an anticipatory fashion because they require distinct political resources and time for their successful application. ‘Presentational strategies’, on the contrary, mainly draw on political rhetoric, and politicians can apply them before and during a 114

blame game (Hood 2007, 2011). The following provides a brief description of these strategy types. Agency strategies seek to reallocate responsibilities and competencies in order to shift to others the risk of being blamed. In case of blame, the reallocated responsibilities allow governing elites to more credibly deflect responsibility and blame onto other actors involved in a policy-making situation (Hood 2007). An example of an agency strategy is the delegation of activities to actors lower down in the administrative hierarchy. By applying this strategy to contentious activities or responsibilities, governing elites hope to lower the risk of reputational damage once the activity attracts blame. Policy strategies address the policy as such. This strategy type seeks to make governmental activities less blameworthy by redesigning policies or changing the ways that they emanate (Hood 2011). If governing elites are able to choose between two policies, they may choose the one that is less likely to attract blame. Governing elites may also change the content of a policy by replacing particular blameworthy aspects with less controversial ones. Moreover, governing elites can try to manipulate the procedures through which policies emanate by, for example, limiting the opportunities for political opponents to voice their criticism. Presentational strategies follow a different logic. Instead of reallocating competencies or changing the substance of a policy, they seek to present policy decisions and governmental activities in a blame-avoiding or blame-limiting way (Hood 2011). The ways in which governing elites attempt to do so are manifold and include reframing an activity, finding excuses for it, deflecting blame onto other actors, distracting the public by pointing to other issues, or timing the political debate about a contentious activity.

4

A categorization of blame avoidance strategies

In the following, we use the distinctions of anticipatory and reactive forms of blame avoidance and agency, policy, and presentational strategies to categorize governing elites’ efforts to avoid electoral punishment for welfare state retrenchment (see Table 1 for an overview). Table 1: Blame avoidance strategies applied by governing elites to avoid punishment for retrenchment. Anticipatory blame avoidance

Agency strategies

Policy strategies

Presentational strategies

-

-

-

-

Engage in social concertation Assign retrenchment tasks to other actors

-

Reactive blame avoidance

Stretch retrenchment over a longer period of time Time retrenchment Spread burdens of retrenchment widely Make hidden cuts

-

-

Muffle resistance against retrenchment within (the) government party(ies) Prepare retrenchment in secrecy Reframe retrenchment Apply emergency rhetoric Deflect blame for retrenchment 115

4.1

Agency strategies

The literature on welfare state retrenchment discusses at least two agency strategies. First, governing elites can engage in social concertation during the retrenchment process. Social concertation spreads responsibility for policy outcomes among several actors (Afonso 2013). By involving trade unions and/or employers in the policy-making process, governing elites express their will to find socially acceptable forms of retrenchment (Wenzelburger 2011). Moreover, by involving additional actors in the policy-making process, governing elites transform potential opponents into accomplices who, being bound to adopt a constructive stance, are deprived of the opportunity to generate blame. Second, governing elites can assign various tasks associated with retrenchment to technocratic agencies or expert committees that are not directly accountable to voters (Wenzelburger 2011). Governing elites can use such agencies to rubber stamp their proposals for retrenchment, thereby endowing them with an ‘objective’ stamp of approval. Or, governing elites can instruct such agencies to propose concrete retrenchment proposals that they then accept. By integrating an ‘agency loop’ into the policy-making process, governing elites make retrenchment appear more ‘automatic’, conjuring the impression that their hands are tied (Weaver 1988). Another variant of this strategy is what Zohlnhöfer (2007) calls ‘institutional cooperation’, i.e. colluding with other parties in the political system to adopt retrenchment measures. 4.2

Policy strategies

The literature mentions at least four policy strategies. The first of these strategies, which appears in several papers, is to stretch retrenchment over a longer time span (Armingeon and Giger 2008). Gradually implementing retrenchment measures, instead of implementing them all at once, makes their perceived losses appear smaller. Smaller losses lead to less public outrage at specific points in time and thereby help to partition blame into more easily manageable parts (Pal and Weaver 2003). Over time, the affected social groups may get used to the cuts and public outrage may wane. Another policy strategy that receives considerable attention in the literature is the strategic timing of retrenchment (Zohlnhöfer 2007; Tepe and Vanhuysse 2010). Instead of slicing painful retrenchment measures into several more easily digestible parts, this strategy implements them all at once. However, in doing so, governing elites exploit the election cycle. For example, they try to implement social policy cuts directly after important elections, expecting that the public outrage associated with them will deaden by the next elections. A third policy strategy targets the scope and distribution of retrenchments. Instead of placing a heavy burden on a small number of social groups, governing elites can try to spread the burden of retrenchment among as many groups as possible (Wenzelburger 2011). The intended effect 116

of this strategy is similar to that of the ‘slicing’ strategy described above. By spreading the burden of retrenchment, governing elites calculate that smaller per capita losses will reduce the likelihood that affected social groups will voice criticism and actively oppose the planned retrenchment measures. Refined variants of this strategy seek to disproportionately distribute losses to weaker social groups that are less capable of mobilizing on retrenchment, and to compensate stronger social groups through side-payments (Pierson 1994; Bonoli and Palier 2007). The fourth policy strategy conceals retrenchment measures from voters and affected social groups as far as possible (Pierson 1994). Not all forms of retrenchment have an immediate and tangible financial effect. Some retrenchment measures only kick in over time, others, like the adjustment of indices (Wagschal and Wenzelburger 2008), are harder to identify as forms of politically imposed retrenchment. Another variant of this strategy is to ‘hide’ retrenchment from voters (and hinder political opponents from exposing it) by bundling retrenchment measures with other policy issues, for example in the form of omnibus laws (Wenzelburger 2011). By applying this strategy, governing elites can expect the perceived extent of retrenchment to be smaller than it actually is, lowering the probability that retrenchment will become a bone of contention. 4.3

Presentational strategies

Unlike agency and policy strategies, which clearly possess an anticipatory character, some of the presentational strategies identified in this section can theoretically be applied in both an anticipatory and reactive fashion. Nevertheless, it is possible to distinguish between strategies that primarily intend to preclude or limit political debate about retrenchment (anticipatory strategies) and strategies that aim to steer a political debate in a particular direction, or at least limit its negative (reputational) effects (reactive strategies). An anticipatory presentational strategy seeks to muffle resistance against retrenchment within the governing party or coalition (Parsons and Weber 2011). If the party or coalition in government is rather skeptical of retrenchment, governing elites can attempt to demobilize ‘internal’ opposition. During blame games about welfare state retrenchment, governing elites usually benefit from the support of their party/parties. Closed ranks allow governing elites to isolate the opinions put forward by political opponents and to dismiss their attacks as hypocritical vote-seeking behavior. Another anticipatory presentational strategy is to prepare retrenchment measures in secrecy in order to surprise opponents and limit their ability to scrutinize and deliberate on the proposed measures (White 2015b). Protracted retrenchment plans allow political opponents to position themselves and organize opposition to them. On the contrary, political opponents that are surprised by the retrenchment plans of governing elites have less time to generate blame and to organize opposition against retrenchment plans among affected social groups. 117

Other presentational strategies are predominantly reactive. First, governing elites can try to reframe retrenchment and cast it in a more positive light (Elmelund-Praestekaer and Emmenegger 2013). For instance, they can portray retrenchment as a precondition for economic sovereignty in the long run, or frame it as a form of ‘diet’ that guarantees a functioning welfare state in the future (Wenzelburger 2011). In this perspective, retrenchment does not appear to be an incision; it is simply a ‘maintenance’ measure that is needed from time to time to keep the system going. A related strategy that governing elites frequently apply when they face acute pressure from voters but cannot postpone retrenchment is to adopt various forms of emergency rhetoric. A prominent form of emergency rhetoric is known by the acronym ‘TINA’, standing for ‘there is no alternative’ (White 2015a; Séville 2017). The TINA rhetoric signals to the public that governing elites have no choice but to quickly implement retrenchment measures in order to avoid fiscal catastrophe (Vis and van Kersbergen 2007). This strategy aims to convince the public of the unavoidable and casts political opponents who rebel against the retrenchment measures as ‘irresponsible’. Finally, governing elites can deflect blame for retrenchment onto other actors (Wenzelburger 2011). If they had previously endowed an agency with the preparation or implementation of retrenchment measures, they would be able to deflect blame onto that agency. Governing elites can also accuse former governments of fiscal recklessness, which allegedly brought the welfare systems under strain in the first place, and which now makes retrenchment unavoidable. Other obvious scapegoats include global markets and their demands for timely ‘debt service’.

5

Insights into the applicability, empirical traces, and effects of blame avoidance strategies

The categorization presented in the previous section reveals a considerable number of blame avoidance strategies that governing elites (can) apply to avoid electoral punishment for welfare state retrenchment. The categorization makes distinctions and yields insights that help to advance the study of welfare state retrenchment on at least three fronts. First, the categorization makes important distinctions with regard to the applicability of certain strategies. Governing elites can only apply agency and policy strategies before blame for a policy issue has developed, because governing elites must mobilize considerable resources and invest considerable time to put them in place. For example, if governing elites decide to charge an agency with the design and implementation of social policy cuts when retrenchment is already prominently on the agenda, political opponents and the media are very likely to decipher this as a brazen blame avoidance move and will accordingly rebel against assigning responsibility to the agency. If governing elites set up the agency well in advance of an eventual blame game, the likelihood of political attention and resistance becomes smaller. In short, 118

agency and policy strategies lack credibility if politicians implement them swiftly (Hinterleitner and Sager 2017). Moreover, the application of agency and policy strategies is often at odds with both national institutional structures and the supranational constraints that prompt governing elites to apply them in the first place. In many situations, governing elites simply lack the institutional power to design welfare state retrenchment in a less blame-attracting way, because, for instance, they do not dispose of other actors with which they can collude (Zohlnhöfer 2007). Or, governing elites may intend to stretch retrenchment over a longer time span, but financial markets may insist on the immediate and timely implementation of retrenchment efforts, thereby depriving governing elites of the opportunity to apply a potentially effective blame avoidance strategy. Circumstances are different in the case of presentational strategies. Although ‘talk’ is not cheap for governing elites, and it requires a number of circumstances in order to be credible (Vis 2010; van Kersbergen and Vis 2015), politicians can apply presentational strategies to a greater number of political situations than they can with agency and policy strategies. A careful and systematic look at the conditions that determine the applicability of certain blame avoidance strategies is a first step in determining whether governing elites have the opportunity to engage in certain forms of blame avoidance in a given situation. Doing so allows research to systematically exclude strategies that are inapplicable in certain situations and concentrate on those that are applicable. If we conclude that it is not possible at all for governing elites to apply blame avoidance strategies in a given situation, we can hypothesize that a lack of punishment for retrenchment is mainly due to a lack of voter opposition. Second, the categorization yields clear-cut descriptions of the mechanics underlying certain strategies. An understanding of the mechanics helps to identify the empirical traces of blame avoidance strategies in real-life policy-making situations. Presentational strategies are relatively easy to identify because whenever governing elites apply them, they leave traces in the form of public statements. Identifying a reframing strategy or a blame deflection strategy is a straightforward endeavor. Agency and policy strategies also leave discernable empirical traces. Agency strategies cause institutional reshuffles that reduce the perceived responsibility of governing elites while positioning other actors more prominently. Policy strategies leave clearly identifiable imprints on aspects of policies, such as the processes through which they are adopted or the financial implications for different social groups. Nevertheless, identifying empirical traces of agency and policy strategies can often be difficult because their application occurs significantly before they come into effect, i.e. before they ‘work’ in the interest of governing elites. Research thus needs to show that at some point in time, governing elites were aware of a future threat and began to intentionally prepare for it (Wenzelburger 2014; Hinterleitner and Sager 2017). This is to say, not everything that serves a blame avoidance purpose was initially contrived for that purpose. Third, the categorization above also provides information about the concrete effect(s) of blame avoidance strategies, and thereby helps to measure the success of strategies (e.g. in 119

experimental research). Measuring the success of blame avoidance strategies requires a concrete understanding of the thoughts and opinions that politicians hope to alter among voters or specific social groups. For example, an agency strategy that reshuffles responsibilities is successful if, after its application, the level of blame directed at governing elites decreases, while the level of blame directed at actors at the ‘frontline’ increases. A presentational strategy is successful if the media adopts the framing and arguments put forward by governing elites. Knowledge about the concrete effects of certain strategies helps empirical research to refine its measurements of ‘successful’ blame avoidance. Existing literature, for the most part, uses relatively crude measures, such as election outcomes or approval ratings; measures that are frequently influenced by a large number of confounding variables. Thinking more systematically about the effects of blame avoidance strategies can help scholars to develop innovative and more precise measurements of success. Overall, knowledge about the applicability, mechanics, and concrete effects of various blame avoidance strategies facilitates our ability to hypothesize and conduct fruitful empirical research on when we can expect governing elites to successfully retrench the welfare state.

6

Governing elites as ‘master jugglers’

This chapter makes conceptual progress on an important puzzle in the research on welfare state retrenchment: Are welfare states retrenched because governing elites are astute blame avoiders or because important parts of the electorate have come to believe in the austerity medicine prescribed by mainstream economists? To cast some light on this important puzzle, the chapter develops a comprehensive overview and categorization of the blame avoidance strategies that governing elites can employ in order to avoid electoral punishment for retrenchment. The categorization identifies the mechanics of particular strategies, yields insights into their applicability, reveals the empirical traces that their application leaves behind, and predicts their likely effects on voters. In so doing, the categorization can guide future research on the causal determinants of successful welfare state retrenchment. This categorization helps researchers to assess the likelihood of whether or not a politician applied a blame avoidance strategy in a given situation, identify empirical traces in situations where its application is likely, and helps researchers to conceptualize and measure the concrete effects of blame avoidance strategies on voters’ perception of welfare state retrenchment. Progress on this puzzle is vitally important because in addition to helping us better understand the future form of democratic welfare states, it is representative of a broader range of difficult situations that the governing elites of modern, capitalist democracies find themselves in. Governing successfully in modern capitalist democracies requires making domestic demands and supranational constraints compatible. The careful pondering and strategic action of governing elites in this context reconfigure the political playing field (Schäfer and Streeck 2013; Mair 2009). The literature on welfare state retrenchment, provided it zooms 120

in on elite decision-making and addresses one of its most important research puzzles head-on, is ideally placed to be at the forefront of researching the nexus of where domestic demands ‘collide’ with supranational constraints.

References Afonso, Alexandre 2013. Social Concertation in Times of Austerity: The Politics of Unemployment and Labour Mobility in Austria and Switzerland. Amsterdam: Amsterdam University Press. Armingeon, Klaus, and Nathalie Giger. 2008. “Conditional Punishment: A Comparative Analysis of the Electoral Consequences of Welfare State Retrenchment in OECD Nations, 1980–2003.” West European Politics 31 (3): 558–580. Armingeon, Klaus, Kai Guthmann, and David Weisstanner. 2015. “Choosing the path of austerity: How parties and policy coalitions influence welfare state retrenchment in periods of fiscal consolidation.” West European Politics 39 (4): 628–647. Beramendi, Pablo, Silja Häusermann, Herbert Kitschelt, and Hanspeter Kriesi, eds. 2015. The Politics of Advanced Capitalism. Cambridge: Cambridge University Press. Bonoli, Guiliano, and Bruno Palier. 2007. “When Past Reforms Open New Opportunities: Comparing Old-age Insurance Reforms in Bismarckian Welfare Systems.” Social Policy & Administration 41 (6): 555–573. Busuioc, Madalina and Martin Lodge. 2016. “The Reputational Basis of Public Accountability.” Governance 29 (2): 247–263. Culpepper, Pepper D. 2011. Quiet Politics and Business Power: Corporate Control in Europe and Japan. Cambridge: Cambridge University Press. Elmelund-Praestekaer, Christian, and Patrick Emmenegger. 2013. “Strategic Re-framing as a Vote Winner: Why Vote-seeking Governments Pursue Unpopular Reforms.” Scandinavian Political Studies 36 (1): 23–42. Giger, Nathalie, and Moira Nelson. 2011. “The Electoral Consequences of Welfare State Retrenchment: Blame Avoidance or Credit Claiming in the Era of Permanent Austerity?” European Journal of Political Research 50 (1): 1–23. Giger, Nathalie, and Moira Nelson. 2013. “The Welfare State or the Economy? Preferences, Constituencies, and Strategies for Retrenchment.” European Sociological Review 29 (5): 1083–1094. Green-Pedersen, Christoffer. 2002. The Politics of Justification: Party Competition and Welfare-State Retrenchment in Denmark and the Netherlands from 1982 to 1998. Amsterdam: Amsterdam University Press. Hall, Peter A., Christoffer Green-Pedersen, Paul Pierson, and Jonas D. Kraft. 2016. “Review Symposium of Pablo Beramendi, Silja Häusermann, Herbert Kitschelt, and Hanspeter Kriesi (eds.). 2015. The Politics of Advanced Capitalism.” Socio-Economic Review 14 (2): 383–394. Hinterleitner, Markus. 2017. “Reconciling Perspectives on Blame Avoidance Behaviour.” Political Studies Review 15 (2): 243–254. Hinterleitner, Markus. 2018. “Policy Failures, Blame Games and Changes to Policy Practice.” Journal of Public Policy 38 (2): 221-242. Hinterleitner, Markus, and Fritz Sager. 2015. “Avoiding Blame - A Comprehensive Framework and the Australian Home Insulation Program Fiasco.” Policy Studies Journal 43 (1): 139–161. Hinterleitner, Markus, and Fritz Sager. 2017. “Anticipatory and Reactive Forms of Blame Avoidance: Of Foxes and Lions.” European Political Science Review 9 (4): 587–606. Hinterleitner, Markus, and Fritz Sager. 2018. “Blame, Reputation and Organizational Responses to a Politicized Climate.” In The blind spots of public bureaucracy and the politics of non‐coordination, edited by Tobias Bach and Kai Wegrich, 133–150. London: Palgrave Macmillan. Hinterleitner, Markus, Fritz Sager, and Eva Thomann. 2016. “The Politics of External Approval: Explaining the IMF's Evaluation of Austerity Programmes.” European Journal of Political Research 55 (3): 549–567. Hood, Christopher. 2007. “What Happens when Transparency Meets Blame-Avoidance?” Public Management Review 9 (2): 191–210. Hood, Christopher. 2011. The Blame Game: Spin, Bureaucracy, and Self-Preservation in Government. Princeton: Princeton University Press. Jensen, Carsten. 2012. “Labour Market- versus Life Course-related Social Policies: Understanding CrossProgramme Differences.” Journal of European Public Policy 19 (2): 275–291. Jensen, Carsten, and Peter B. Mortensen. 2014. “Government Responses to Fiscal Austerity: The Effect of Institutional Fragmentation and Partisanship.” Comparative Political Studies 47 (2): 143–170. König, Pascal D., and Georg Wenzelburger. 2014. “Toward a Theory of Political Strategy in Policy Analysis.” Politics & Policy 42 (3): 400–430. 121

Lau, Richard R. 1985. “Two Explanations for Negativity Effects in Political Behavior.” American Journal of Political Science 29 (1): 119. Mair, Peter. 2009. “Representative versus Responsible Government.” MPIfG working paper 9 (8): 1–19. Merkel, Angela. 2011. Press conference together with the Prime Minister of Portugal, Pedro Passos Coelho, Berlin. Nelson, Moira. 2016. “Credit-Claiming or Blame Avoidance? Comparing the Relationship between Welfare State Beliefs and the Framing of Social Policy Retrenchment in France and Germany.” Journal of Comparative Policy Analysis: Research and Practice 18 (2): 138–156. Pal, Leslie A., and R. Kent Weaver, eds. 2003. The Government Taketh Away: The Politics of Pain in the United States and Canada. Washington, D.C: Georgetown University Press. Parsons, Craig, and Till Weber. 2011. “Cross-Cutting Issues and Party Strategy in the European Union.” Comparative Political Studies 44 (4): 383–411. Pierson, Paul. 1994. Dismantling the Welfare State? Reagan, Thatcher, and the Politics of Retrenchment. Cambridge, New York: Cambridge University Press. Pierson, Paul, ed. 2001. The New Politics of the Welfare State. Oxford, New York: Oxford University Press. Sager, Fritz, and Markus Hinterleitner. 2016. “How Do Credit Rating Agencies Rate? An Implementation Perspective on the Assessment of Austerity Programs during the European Debt Crisis.” Politics & Policy 44 (4): 783–815. Schäfer, Armin, and Wolfgang Streeck, eds. 2013. Politics in the Age of Austerity. Cambridge: Polity Press. Schumacher, Gijs, Barbara Vis, and Kees van Kersbergen. 2012. “Political Parties’ Welfare Image, Electoral Punishment and Welfare State Retrenchment.” Comparative European Politics 11 (1): 1–21. Séville, Astrid. 2017. “From ‘One Right Way’ to ‘One Ruinous Way’? Discursive Shifts in ‘There is no Alternative’.” European Political Science Review 9 (3): 449–470. Starke, Peter. 2006. “The Politics of Welfare State Retrenchment: A Literature Review.” Social Policy & Administration 40 (1): 104–120. Sulitzeanu-Kenan, Raanan, and Christopher Hood. 2005. “Blame Avoidance with Adjectives? Motivation, Opportunity, Activity and Outcome.” Paper presented at the ECPR Joint Sessions, Granada, April 2005. Tepe, Markus, and Pieter Vanhuysse. 2010. “Who Cuts Back and When? The Politics of Delays in Social Expenditure Cutbacks, 1980–2005.” West European Politics 33(6):1214–1240. van Kersbergen, Kees, and Barbara Vis. 2015. Comparative Welfare State Politics: Development, Opportunities, and Reform. New York: Cambridge University Press. Vis, Barbara. 2010. Politics of Risk-Taking: Welfare State Reform in Advanced Democracies. Amsterdam: Amsterdam University Press. Vis, Barbara. 2016. “Taking Stock of the Comparative Literature on the Role of Blame Avoidance Strategies in Social Policy Reform.” Journal of Comparative Policy Analysis: Research and Practice 18 (2): 122–137. Vis, Barbara, and Kees van Kersbergen. 2007. “Why and How do Political Actors Pursue Risky Reforms?” Journal of Theoretical Politics 19 (2): 153–172. Wagschal, Uwe, and Georg Wenzelburger. 2008. “Roads to Success: Budget Consolidations in OECD Countries.” Journal of Public Policy 28 (3): 309-339. Weaver, R. Kent. 1986. “The Politics of Blame Avoidance.” Journal of Public Policy 6 (4): 371–398. Weaver, R. Kent. 1988. Automatic Government: The Politics of Indexation. Washington, D.C.: Brookings Institution, Wenzelburger, Georg. 2011. “Political Strategies and Fiscal Retrenchment: Evidence from Four Countries.” West European Politics 34 (6): 1151–1184. Wenzelburger, Georg. 2014. “Blame Avoidance, Electoral Punishment and the Perceptions of Risk.” Journal of European Social Policy 24 (1): 80–91. Wenzelburger, Georg, and Felix Hörisch. 2016. “Framing Effects and Comparative Social Policy Reform: Comparing Blame Avoidance Evidence from Two Experiments.” Journal of Comparative Policy Analysis: Research and Practice 28 (2): 157–175. White, Jonathan. 2015a. “Emergency Europe.” Political Studies 63 (2): 300–318. White, Jonathan. 2015b. “Politicizing Europe: The Challenge of Executive Discretion.” In Democratic politics in a European Union under Stress, edited by Olaf Cramme and Sara B. Hobolt, 87-102. Oxford: Oxford University Press. Zohlnhöfer, Reimut. 2007. “The Politics of Budget Consolidation in Britain and Germany: The Impact of Blame Avoidance Opportunities.” West European Politics 30 (5): 1120–1138.

122

New evidence on the effect of partisanship on the welfare state Carsten Jensen1 and Georg Wenzelburger2 Abstract: A core question of comparative political economy is whether government ideology still affects policy, or if government policies rather are directed by economic necessities – especially in times of crisis. This chapter revisits this fundamental question of welfare state dynamics by using new and more fine-grained data based on legislative changes to two major welfare programmes – unemployment insurance and old age pensions – in five Western industrialized countries during the last 40 years. The new dataset enables us to revisit empirically some of the major theoretical and most disputed claims put forward by the different strands of the welfare state literature. Our results support the ‘new politics’ thesis according to which the partisan effect of governments have declined from the 1970s until today.

1

Introduction

Since the heyday of the welfare state in the first post-War decades, western democracies have witnessed dramatic economic turmoil. Declining growth rates, mounting public debt and ageing populations are such some of the ingredients in what Pierson (1996) called the “era of permanent austerity”. It is in this climate of crisis that governments have to operate. No wonder, then, that one of the core questions of comparative political economy is whether government ideology still affects policy, or if all government policies are directed rather by economic necessities. This chapter revisits this fundamental question of welfare state dynamics by using new and more fine-grained data based on legislative changes to two major welfare programmes – unemployment insurance and old age pensions – in five Western industrialized countries during the last 40 years. The new dataset enables us to revisit empirically some of the major theoretical and most disputed claims put forward by the different strands of the welfare state literature. Specifically, the chapter contributes to the existing literature in two respects. First, we present new hand-coded data on welfare state reform events in Denmark, Finland, France, Germany and the United Kingdom based on legislative changes to the welfare state. This leads to a more complex picture of welfare state dynamics than previously known. Second, with this legislative data at hand, we revisit one of the most fundamental questions in the welfare state literature, namely whether the partisan ideology of the government affects welfare state change.

1 2

Aarhus Universitet [email protected] Universität Kaiserslautern [email protected] kl.de -

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 R. Careja et al. (eds.), The European Social Model under Pressure, https://doi.org/10.1007/978-3-658-27043-8_8

123

Our results support the ‘new politics’ thesis, according to which the partisan effect of governments have declined from the 1970s until today. The remainder of this chapter is structured as follows. The next section briefly recapitulates the achievements and weaknesses of the state of the art of the literature, before we develop a set of theoretical expectations about how parties affect welfare state change. The third section presents how the data was collected in the five countries under review, and it describes the development of the welfare state in these countries through the lens of our data on legislative changes. The empirical analysis of partisan effects appears in section four, and the final section discusses the results and concludes.

2

Parties and the welfare state

The literature on the partisan effects on welfare state change has developed in three major waves: In the beginning of the 1980s, various welfare state scholars pushed forward the idea that the balance of power between capital and labour (Korpi 1983; Stephens 1979) or the participation of Conservatives or Social Democrats in government was responsible for cross-country differences in welfare state development (e.g. Castles and McKinlay 1979; Schmidt 1982). Although this thesis has been nuanced over the years by emphasizing the role of religious parties (van Kersbergen 1995; Van Kersbergen and Manow 2009) or the constitutional structure (Huber et al. 1993; Kittel and Obinger 2003), the conventional wisdom until the mid-1990s held that parties were one of the most important explanations of the variance of welfare state generosity in the Western world. The second wave was spearheaded by Paul Pierson and his landmark study on welfare state politics in Britain and the United States (Pierson 1994). Pierson argued that the shrinking power of left parties and organized labour during the 1980s in several countries (and especially in Thatcher’s Britain and Reagan’s United States) “would suggest that welfare states are in deep trouble. […]. [H]owever, there is very little evidence that this decline has had a fundamental impact on welfare states” (Pierson 1996, 150). Following this argument, the ‘new politics’ perspective postulated that parties of the left and right could no longer explain welfare state dynamics. Instead, welfare state development was described as being locked-in due to institutional stickiness and electoral incentives. In fact, several authors argued in favour of a reversed partisan effect according to which left wing parties seemed better equipped to shrink the welfare state due to the configuration of political competition in some countries (Kitschelt 2001) or their traditional ownership of the welfare issue (Ross 2000a, 2000b). The availability of new data on the development of actual net replacement rates in major welfare state programmes that could be calculated for different model families set in motion the third wave of research on partisan effects and the welfare state. Based on different data collections, Korpi and Palme (2003) as well as Allan and Scruggs (2004) found not only much 124

more retrenchment than discovered by Pierson, but also substantial partisan effects which supported the ‘old politics’ view on welfare state dynamics. Right-wing governments were correlated with cutbacks and left-wing governments with expansions of the welfare state. At the most basic level, partisanship theory proposes that governments make legislative decisions to either expand or retrench the welfare state – the direction depending upon the ideology bent of the parties that form the government. In the quantitative literature, which has dominated the literature, these decisions are normally measured as either public spending on the welfare state (typically as a percentage of GDP) or as so-called replacement rates (typically the value of cash benefits compared to the average production worker’s wage). 3 Both of these measures have their strengths and weaknesses. Public spending captures the overall devotion of society’s fiscal resources to the welfare state, while replacement rates capture the quality of the welfare state from the perspective of the individual claimant. Yet, while both are clearly affected by political decisions, they are also a function of social factors outside the control of governments. Social demand drives public spending, so when, for instance, the number of jobless goes up, so will spending on unemployment protection (the “dependent variable problem” (Clasen and Siegel 2007; Green-Pedersen 2004)). Although less often considered, the same applies to replacement rates, which are calculated using two pieces of information: the cash benefits of claimants and the wage of the average production worker. The benefit level is directly affected by political decisions, but the wage of the average production worker is not. Wages are rather driven by the state of the economy and wage negotiation arrangements broadly defined (on the difficulties related to replacement rates, see Wenzelburger et al. 2013). These socio-economic drivers explain why both public spending and replacement rates move slowly and gradually rather than as swift breaks, which one argu-ably would expect if the measures only captured political decisions. Related to this, the effects of political decisions on spending and replacement rates can take a long time to manifest itself (Green-Pedersen 2004). This means that we run the risk of attributing changes to the wrong governments. A change today may – or may not – be due to the actions of the government in power. Indeed, given that governments will often try to avoid blame for enacting popular deci-sions, this kind of delayed implementation is a well-known blame avoidance technique (Pierson 1994). Given these challenges, we propose to focus on legislative decisions instead of aggregate outcome data. In order to do so, we have coded all legislative changes in the two main transfer schemes of the welfare state: unemployment and old-age pensions. The use of legislative data enables us to overcome some of the challenges that the current literature has to cope with and which have been described above. Legislation is the most proximate output of government activities and hence it is not influenced by other confounding variables such as demand for bene-

3

Other composite measures of generosity exists, but the main variable part of these are, in fact, replacement rates. 125

fits, the economic cycle or the calculation of net replacement rates and standard wages. Employing such a new dataset will allow us to discriminate between the two basic expectations of the literature: H1: (old politics): Government ideology continues to affect welfare state policies. H2: (new politics): Government ideology has a declining effect on welfare state policies.

3

Data and methods

The main innovation of the present chapter is that it uses a new dataset, which enables us not only to revisit the basic claims of the partisan theory of the welfare state, but also to look at whether parties matter in terms of instrument choice. The data-set on which these analyses rest encompasses all legislative changes concerning the social rights of citizens with regard to oldage pensions and unemployment protection in Britain, Denmark, Finland, France and Germany between the early 1970s and 2014.4 The research design therefore involves three decisions on case selection: on welfare programme, on countries, and on time. The two programmes were selected for two main reasons. First, they constitute a core part of what is conventionally understood as ‘the welfare state’ and, as a result, are prominently featured in a very long list of publications on the politics of the welfare state. Second, both programmes are transfer schemes, which means that we can code them using the same coding scheme. This is important if we want to compare the two programmes. The five countries were selected because they provide substantial variation in terms of their respective welfare state models (Esping-Andersen 1990; Korpi and Palme 1998; Scruggs and Allan 2006) and the popularity of the welfare state in general (Brooks and Manza 2006). Britain is a classic example of a residual welfare state where benefits are mostly means-tested, relatively low and public support of the welfare state is lower than in the two other countries. Denmark, in comparison, has substantially more generous benefits and less frequently employs a (strict) means test. France and Germany are archetypal insurance-based systems where benefits are tied to earnings history and membership of occupational schemes. And, finally, Finland is an interesting case as it unites features of the Scandinavian universal welfare state and the insurance-based Continental European model (see Kangas 2006). Moreover, in all five countries, support for the welfare state varies, with lower values for Britain and more support in Germany or Denmark (Brooks and Manza 2007; Svallfors 2012, 14). Our dependent variable for the test of the general partisan effect is the net change of welfare state legislation measured as the difference between the number of positive (expansive) and the number of negative (retrenching) changes to the legislative status quo in every year. To construct the dataset, we began by defining the universe of reforms we were interested in. We only wanted to focus on reforms that affected citizens’ social rights, and thus ignored changes to the 4

Due to data availability, the time series start in 1970 for UK, 1972 for Denmark and 1974 in Germany, France and Finland, and end in 2014 in all countries. 126

administration of benefits. Reforms had to either change the rules of access, rules of benefits, or rules of conduct of claimants (cf. the typology of change introduced by Clasen and Clegg 2007). Each legislative change was coded as a ‘reform event’. Reform events that curtailed the social rights of citizens were coded as ‘cutbacks’ and reform events that improved citizens’ social rights were coded as ‘expansions’. Within a given country-year, we then counted the number of negative and positive reform events, respectively. By subtracting the former from the latter, we arrive at the measure of ‘net change’.5 These country-year counts are the unit of analysis in the dataset we currently employ. To collect information on reforms, we started out by relying on secondary sources, which allowed us to establish a timeline of reform events. We then looked up all reforms in the official legislative databases when these were available. In this way, almost all reforms in the dataset have been double-checked using two sources. The coding was done by trained student assistants, but all coding decisions were verified by a senior researcher. There were very few instances of disagreements, and those that occurred were settled by the student assistant and senior researcher who examined the question in-depth together.

Figure 1: Net change of unemployment and pension legislation in UK and France

5

Calculating the ‘net change’ therefore comes down to simply taking the difference between the number of cutbacks and expansions as indication of welfare state development. While we agree that we do not measure the extent to which a certain reform event changes the welfare state, e.g. the size of a cutback, bigger reforms usually also concern a greater number of events. The face validity of our measure is also corroborated by the fact that the measure nicely takes up several reform events that are described in the qualitative literature (Lee et al. 2017, online appendix) 127

To illustrate the data structure, Figure 1 displays the net changes in unemployment and pension legislation in France and the United Kingdom. For both countries, we clearly see that our data takes up some of the major reforms that have been prominently discussed in the literature. For Britain, the data clearly identifies the major retrenchment implemented during the Thatcher years, and especially in 1980 and 1986, as well as the more expansive stance during New Labour, followed, again, by cutbacks in connection with the budget balancing efforts by the first Cameron administration. This pattern is in line with extant research, which often highlights the British experience as outstanding for the retrenchment it has witnessed (e.g., Allan and Scruggs 2004; Clasen 2005; Clasen and Clegg 2011; Jensen 2014) and shows the face validity of our data collection. The same is true for France, where the major cutbacks occurred during the early 1990s and the centre-right governments of Balladur and Juppé, as well as under the Presidency of Chirac and Sarkozy in the 2000s. The more expansive stance in the end of the 1990s and early 2000s can be attributed to the left-wing government headed by Jospin (1997-2001) which was made up of Socialists, Communists and the Green Party. Again, these results mesh well with what we know from the qualitative literature on the development of the French welfare state (Egle 2009; Hassenteufel 2008; Palier 2010). In sum, we are therefore confident that our dataset captures the overall tendency of the welfare state development quite well. Our major independent variable relates to the partisan complexion of the government. As we are primarily interested in the question whether it matters for social policy change that a certain party is part of the government (or not), we use the cabinet seat shares of the different party families to measure the partisan stance of the government. This measure is not without drawbacks and especially problematic in cross-country settings because the ideological position of parties of the same family may diverge from country to country (see Döring and Schwander 2015). At the same time, however, using some kind of ‘centre of gravity’ variable (Cusack 1997) based on measures of partisan ideology from manifesto or expert survey data also has its disadvantages (for a most recent overview of this critique, see Loftis and Mortensen (2017)). First, there is a wide debate about what measures of party position to use (Benoit and Laver 2007; Hooghe et al. 2010; Horn et al. 2017; Lowe et al. 2011). Second, using the ideological position of the government on a certain left-right scale reduces the test of a partisan effect to investigating whether parties translate what they say in their manifesto into actual policies. While this surely is an important research question (Budge et al. 2001; Klingemann 2006; Klingemann et al. 1994), our research question centres on the question of partisan differences, i.e. to investigate whether the parties actually still do what their labels promise to their voters. Clearly, voters do (maybe increasingly) switch between parties, but still, many voters identify themselves with a party (family) and therefore expect different welfare policies from social democrats and conservatives – regardless of what they actually have laid out in their last election manifesto. In order to address the challenge that parties of the same family may hold very different positions in different countries, we include country dummies in the regression equations, which take up the differences in ideological positions between parties. 128

Finally, the fact that we aim at comparing our results to previous landmark studies (like the one of Allan and Scruggs 2004) also calls for using cabinet shares. The coding of the individual parties follows the proposition of Armingeon et al. (2015), which distinguishes between right, centre and left parties. Table 1 summarizes the categorization of parties as either left- or rightwing. We use three operationalizations: The cabinet-share of right-wing parties, the cabinet share of left-wing parties, as well as the difference between both variables with high values indicating a right-wing government. The summary statistics of these three measures are reported in Table 2. Table 1. Coding of party families Country Left-wing Social Democrats (SD) Denmark Socialist People’s Party (SF)

Right-wing Liberals (Venstre) Conservative People’s Party (KF) Radical Liberal Party (Radikale Venstre) National Coalition (KOK) Swedish People’s Party (SFP)

Finland

Social Democrats (SSDP) Greens (VIHR) Left Alliance (VAS)

France

Socialists (PS) Green Party (EELV, Les Verts) Communists (PCF) Left Party (FI) Left Radical Party (PRG)

Conservatives (UMP, RPR, LR) Radical Party (RAD)

Germany

Social Democrats (SPD) Green Party (B90/Grüne) Labour Party

Liberal Party (FDP)

UK

Conservative Party Liberal Party

Source: Armingeon et al. 2015. Table 2. Summary statistics for the independent variables Country Cabinet share Cabinet share right-wing left-wing

Net ideological balance

Frequency

Denmark

50.76 (42.50)

44.95 (44.13)

5.82 (86.20)

263

Finland

37.06

35.95

1.11

358

(15.94)

(20.63)

(31.01)

France

39.22 (34.37)

35.74 (45.59)

3.48 (77.78)

281

Germany

15.34

36.20

-20.86

460

(10.51)

(42.82)

(50.05)

58.38

41.62

16.76

(49.44)

(49.44)

(98.87)

United Kingdom

173

Source: Armingeon et al. 2015.

Besides the partisan variable, we include some standard controls that have been used in the rich welfare state literature, and which may affect the relationship between government ideology 129

and welfare state policies. These include the unemployment rate, real GDP growth, trade openness, and fiscal balance (to capture the socio-economic situation) and union density and number of veto points (to capture the broader political context). We obtain all data for these variables from Armingeon et al. (2015). In order to assess the overall question whether the partisan composition of governments affects welfare state change as posited in H1 and H2, we run time series cross section (TSCS) regressions using the yearly net change of the welfare state legislation as dependent variable. In case of government changes within one year, we follow Armingeon et al. (2015) and weigh the variables for partisan complexion by month.6 Due to heteroscedasticity, we estimated panel corrected standard errors. Serial autocorrelation is not an issue in our data and unit heterogeneity is controlled by including country dummies. However, estimating the models without such dummies does not alter the results substantially.

4

Analysis

The main findings of the chapter are reported in Table 3 and Figure 2, which test the relationship between the ideological composition of governments and the net change-variable. Table 3 contains a total of six regression models. The first three explore the varying effect of the three operationalizations of government composition in a setup without country dummies. The last three regressions repeat these models, but now include country dummies. It transpires that there are consistent partisan effects with a higher right-wing cabinet share being associated with negative net change (models 1 and 4), and a higher left-wing cabinet share being associated with positive net change (models 2 and 5). The net ‘rightness’ of the government is similarly associated with negative net change. To get a better sense of the effects, we need to calculate predictive margins. We do so based on model 6 in Table 3. We opt for this model because the ideological balance measure allows us to employ a single summary measure of the composition of the government, while the country dummies allow us to zoom in on within-country effects. Figure 3 reports the results. Several things are noteworthy. First and most importantly, we see that the main action pertains to the expansionary efforts of left-wing governments, whereas the right does not stand out as particularly cutback-oriented compared to the left. Second, the country years where the net ideological balance is 40 or above sum up to around 35 percent of all country years. That is, the expansionary bent of left-wing governments relate to a large subset of the observations in our dataset. Turning briefly to the controls in Table 3, we see that an improving fiscal balance and higher economic growth both are associated with positive net change. On the flipside, a deteriorating economy will be associated with cuts, as witnessed in many European countries in the years 6

Hence, if a purely left-wing government is in office from January to June and a purely right-wing government from July to December, the overall share is 50% left and 50% right for the year, with the net balance being 0. 130

after the Great Recession. A higher number of veto players are negatively associated with net change. Table 3: Main findings

Cabinet share right wing parties

(1)

(2)

(3)

(4)

(5)

(6)

Pool,

Pool,

Pool,

FE,

FE,

FE,

PCSE

PCSE

PCSE

PCSE

PCSE

PCSE

-0.028***

-0.030***

(-3.33)

(-3.42)

Cabinet share

0.015*

left wing parties

(1.81)

(1.81) -0.012***

Net ideological balance of government Veto player index

0.016* -0.012***

(-2.83) -0.10**

-0.069*

(-2.72)

-0.086**

-0.090**

-0.082*

-0.089**

(-2.50)

(-1.75)

(-2.17)

(-1.98)

(-1.84)

(-1.98)

Union density

0.0028

-0.0040

-0.00061

0.011

-0.013

-0.00038

(0.16)

(-0.23)

(-0.04)

(0.23)

(-0.26)

(-0.01)

Trade openness

0.0030

0.00023

0.00063

0.0052

-0.0038

-0.00024

(export/import)

(0.09)

(0.01)

(0.02)

(0.13)

(-0.09)

(-0.01)

Fiscal balance

0.72***

0.73***

0.73***

0.74***

0.71***

0.73***

(5.94)

(5.93)

(5.98)

(4.92)

(4.77)

(4.85)

0.30*

0.32**

0.31**

0.30*

0.32**

0.31**

(1.95)

(2.08)

(2.01)

(1.93)

(2.04)

(1.98)

0.0012

-0.015

0.0052

0.016

-0.018

0.0047

(0.01)

(-0.10)

(0.04)

(0.09)

(-0.10)

(0.03)

-0.53

0.64

-0.011

(-0.15)

(0.18)

(-0.00)

-0.97

0.87

-0.038

(-0.28)

(0.26)

(-0.01)

Economic growth Unemployment rate Dummy Denmark Dummy Finland Dummy Germany Dummy UK Constant

-0.66

0.39

-0.070

(-0.40)

(0.24)

(-0.04)

-0.12

-0.032

-0.18

(-0.08)

(-0.02)

(-0.12)

2.55

1.12

1.59

2.42

1.48

1.71

(1.38)

(0.58)

(0.85)

(1.06)

(0.62)

(0.73)

R2

0.248

0.231

0.240

0.250

0.232

0.241

N

197

197

197

197

197

197

Notes: Dependent variable: Net change to welfare state legislation (sum of expansions – sum of cutbacks per year); t statistics in parentheses; CI: * p < .1, ** p < .05, *** p < .01; Reference for country dummies: France

131

3

Fitted Values

2

1

0

-1

-2 -100

-80

-60

-40

-20

0

20

40

60

80

100

Net ideological balance of government Figure 2: Predictive margins Note: Predicted margins (fitted values of “net change”) for different values of net ideological balance (-100: only left cabinet shares; 100: only right cabinet shares); 90% confidence intervals. Graph based on model 6 in Table 3.

So far, the results would appear to lend support to the old politics-argument of partisan effect. Yet, of course, the new politics-argument does not suggest that there were never partisan effects, only that these have declined since the heyday of the welfare state. To properly test the new politics-argument, we, in short, need to look at whether there is temporal variation in the effects. The three graphs in figure 3 investigate this based on model 6 in Table 3. The three figures report the effect of governments’ net ideological balance on net changes in three periods: 1970-1981, 1982-1993, and since 1993. Although, ultimately, these cut-off points are arbitrary, they do reflect the conventional periodization of the literature with the first period that constitute the final part of the welfare state’s Golden Age; the second part with the ideational transition of the 1980s and early 1990s where monetarism, supply-side economics, and new public management were introduced by mainly right-wing governments; and a third part where these new ideas transformed into a policy paradigm which is accepted by most mainstream parties. Several points are noteworthy. First, the direction of the relationship between the rightness of the government and the net change of the welfare state legislation is stable over time. The left is always associated to a more expansionist stance than the right. However, the impact of parties on welfare state legislation was clearly most pronounced in the years before 1982 and diminishes over time. In the post-1993 years, the predicted net change for a pure left-wing government (-100 on the rightness-scale) is only marginally and non-significantly above zero. 132

Finally, in the 1980s, there is a slightly positive and significant effect for governments from the centre, which is mainly due to the absence of left-wing governments during this period (basically, only several years in France).

8

4

4

Fitted Values

8

0 -4 -8

early 1970-1981 -100 -80 -60 -40 -20

0 -4 -8

0

20 40 60 80 100

Net ideological balance of government

1982-1993 -100 -80 -60 -40 -20

0

20 40 60 80 100

Net ideological balance of government

8 4 0 -4 -8

post 1993 -100 -80 -60 -40 -20

0

20 40 60 80 100

Net ideological balance of government

Figure 3. Predictive margins, three time periods Note: Predicted margins (fitted values of “net change”) for different values of net ideological balance (-100: only left cabinet shares; 100: only right cabinet shares); 90% confidence intervals. Graph based on model 6 in Table 3.

In sum, these results therefore indicate that the ‘old politics’ is still visible if we analyse the entire period from the beginning of the 1970s to today, but that this effect is mainly driven by a very strong expansionist stance of left-wing parties in the years before 1982. In the last two decades, it seems that partisan effects still show up to a certain extent, but they are weaker and no longer significant at conventional levels. With regard to our hypotheses, while our data generally supports the expectations of H1, the analysis of different sub-periods also shows that partisan effects have weakened considerably during the last two decades. Overall, this lends support to H2 and the ‘new politics’ argument.

5

Conclusion

The role of government ideology for welfare state development is a question that has attracted considerable attention from scholars for the past few decades. The backdrop of the discussion is the more or less constant climate of crisis that has existed in many western democracies since 133

the 1970s. In an era of permanent austerity, is there any room for governments to pursue their own ideological goals? This chapter adds to this literature by employing a new dataset on legislative decisions, which allows us to circumvent some of the more problematic aspects of existing data. With our data at hand, we get much closer to the dependent variable of interest, namely legislative decisions. Existing work has had to rely on either social spending or replacement rates, both of which can provide valuable information about the effects of legislative changes, but not on the legislative changes themselves. With our data it is, relatedly, also possible to more firmly link policy changes in a given year to the government in office that specific year. Taken together, this entails that our results arguably constitute one of the most valid estimates of partisanship on welfare state reforms that is available today. Our main finding is that left-wing governments used to be associated with expansions, but that this effect now has declined and possibly even vanished. This fits squarely with the ‘new politics’ argument of Pierson (1994, 1996). That said, a number of caveats need to be taken. In this chapter, we are interested in the general patterns of ‘the welfare state’, which we chose to measure on the basis of unemployment benefits and old-age pensions. We do so because the old versus new politics - which is debated - has been framed in this way, as far as we read it at least. However, in reality, the welfare state is made up of multiple policy areas that do not necessarily share the same political dynamics. Old-age pensions are, for instance, much more popular than unemployment benefits. Generally, the distinction between (popular) lifecycle programmes and (less popular) labour market programmes has been suggested to affect partisan conflict by authors who rely on the older data sets (e.g. Jensen 2012; Zohlnhöfer et al. 2013; Jensen and Petersen 2017). We suspect that we might find a similar diversity in our data, if we disaggregated it. Given that we have coded individual policy instruments in our dataset, it is also possible for us to explore the partisanship of instrument choice. That is, not simply to look at the direction of change, but also at the mode of change. In another article we have shown together with two colleagues that governments tend to use complicated, technical (“invisible”) policy instruments when introducing cutbacks, but easy-to-understand (“visible”) policy instruments when introducing expansions (Jensen et al. 2018). This aligns well with the blame avoidance argument of the ‘new politics’ literature: governments try to dissociate themselves with unpopular retrenchment, but associate themselves with the popular expansions. A key question arising from this chapter is to what extent the choice of visible and invisible policy instruments are driven by government ideology. With the new data set, we can study if government ideology matters for other policy choices as well. Clasen and Clegg (2007) differentiate between rules regulation access, benefits and behaviour of claimants. An obvious question for the future would seem to be if we can locate partisan differences in the usage of these diverse types of policy instruments.

134

Apart from the avenues of future research that emerge from re-conceptualizing our dependent variable, there is also the issue of the appropriateness of our independent variables. Naturally, we are adhering to conventions by using cabinet shares of right- and left-wing parties. Still, one reason why we find a declining partisan effect may be that this way of conceptualizing government partisanship is less and less representative of the real world. A burgeoning literature argues that the traditional class politics has given way to a new form of coalitional politics (Häusermann 2010; Beramendi et al. 2015). If so, government partisanship is less likely to manifest itself as clear-cut left-versus-right distinctions – even if the composition of governments remains important for policy-making. This is a serious concern, although our goal in this chapter never was to establish if governments matter at all, only if having a right- or left-wing government matters. At any rate, conceptualizing government partisanship in a meaningful – but not country idiosyncratic – way is a major challenge for comparativists. Finally, it is worthwhile returning to the visual outline of the data structure for the UK in Figure 1. Here we saw a relatively clear-cut partisan pattern in the UK for the entire period; i.e. also after 1993 when our statistical results would suggest these to have faded. As it is, the expansions of New Labour and the subsequent cutbacks of the Conservatives map onto the trends in the economic well-being of the country. Indeed, most places witnessed modest expansion before, and substantial retrenchment after the Great Recession of 2008. This is also reflected in the control variables reported in Table 2. Here the fiscal balance and real GDP growth were both positively associated with net changes. Welfare state generosity tends to track the state of the economy. In a narrow sense, it would appear that the largesse of New Labour and stinginess of the Conservatives are a reflection of the economic situation of these two governments. Yet statistical analysis only gets us so far. That a given economic state of affair entails either cutbacks or expansions is not preordained, but, in fact, a political choice. Probing more thoroughly how governments of different ideological orientation use the economic situation to pursue their economic goals, and how this meshes with the decline of partisanship effects, is an interesting question for the literature.

References Allan, James P., and Lyle Scruggs. 2004. “Political Partisanship and Welfare State Reform in Advanced Industrial Societies.” American Journal of Political Science 48 (3): 496-512. Armingeon, Klaus, Christian Isler, Laura Knöpfel, David Weistanner, and Sarah Engler. 2015. “Comparative Political Dataset 1960-2013.“ University of Bern Institute of Political Science. Bern. Benoit, Kenneth, and Michael Laver. 2007. “Estimating Party Policy Positions. Comparing Expert Surveys and Hand-Coded Content Analysis.” Electoral Studies 26 (1): 90-107. Beramendi, Pablo, Häusermann, Silja, Kitschelt, Herbert, and Hanspeter Kriesi. 2018. “Introduction: The Politics of Advanced Capitalism.” In The Politics of Advanced Capitalism, edited by Pablo Beramendi, Silja Häusermann, Herbert Kitschelt, and Hanspeter Kriesi, 1-66. Cambridge: Cambridge University Press. Brooks, Clem, and Jeff Manza. 2006. “Social Policy Responsiveness in Developed Democracies.” American Sociological Review 71 (3): 474-94. Brooks, Clem, and Jeff Manza. 2007. Why Welfare States Persist: The Importance of Public Opinion in Democracies. Chicago: University of Chicago Press. 135

Budge, Ian, Hans-Dieter Klingemann, Andrea Volkens, Judith Bara, and Eric Tanenbaum. 2001. Mapping Policy Preferences. Estimates for Parties, Electors and Governments 1945-1998. Oxford: Oxford University Press. Castles, Frank, and Robert D. McKinlay. 1979. “Does Politics Matter: An Analysis of the Public Welfare Commitment in Advanced Democratic States.” European Journal of Political Research 7 (2): 169-86. Clasen, Jochen. 2005. Reforming European Welfare States: Germany and the United Kingdom Compared. Oxford: Oxford University Press. Clasen, Jochen, and Nico A. Siegel. 2007. Investigating Welfare State Change. The ’Dependent Variable Problem’ in Comparative Analysis. Cheltenham/Northampton: Edward Elgar. Clasen, Jochen, and Daniel Clegg. 2007. “Levels and Levers of Conditionality: Measuring Change Within Welfare States.” In Investigating Welfare State Change. The ‘Dependent Variable Problem’ in Comparative Analysis, edited by Jochen Clasen and Nico A. Siegel, 166-197. Cheltenham/Northampton: Edward Elgar. Clasen, Jochen, and Daniel Clegg. 2011. “The Transformation of Unemployment Protection in Europe.” In Regulating the Risk of Unemployment: National Adaptations to Post-Industrial Labour Markets in Europe, edited by Jochen Clasen and Daniel Clegg, 333-345. Oxford: Oxford University Press. Cusack, Thomas R. 1997. “Partisan Politics and Public Finance: Changes in Public Spending in the Industrialized Democracies, 1955-89.” Public Choice 91 (3/4): 375-95. Döring, Holger, and Hanna Schwander. 2015. “Revisiting the Left Cabinet Share: How to Measure the Partisan Profile of Governments in Welfare State Research.” Journal of European Social Policy 25 (2): 175-93. Egle, Christoph. 2009. Reformpolitik in Deutschland und Frankreich. Wirtschafts- und Sozialpolitik bürgerlicher und sozialdemokratischer Regierungen. Wiesbaden: VS Verlag. Esping-Andersen, Gøsta. 1990. The Three Worlds of Welfare Capitalism. Cambridge: Polity Press. Green-Pedersen, Christoffer. 2004. “The Dependent Variable Problem within the Study of Welfare State Retrenchment: Defining the Problem and Looking for Solutions.” Journal of Comparative Policy Analysis 6 (1): 3-14. Häusermann, Silja. 2010. The Politics of Welfare State Reform in Continental Europe: Modernization in Hard Times. Cambridge: Cambridge University Press. Hassenteufel, Patrick. 2008. “Welfare Policies and Politics.” In Developments in French Politics 4, edited by Alistair Cole, Patrick Le Galès and Jonah Levy, 227-42. Basingstoke: Palgrave Macmillan. Hooghe, Liesbet, Ryan Bakker, Anna Brigevich, Catherine de Vries, Erica Edwards, Gary Marks, Jan Rovny, Marco Steenbergen, and Milada Vachudova. 2010. “Reliability and Validity of Measuring Party Positions: The Chapel Hill Expert Surveys of 2002 and 2006.” European Journal of Political Research 42 (4): 684703. Horn, Alexander, Anthony Kevins, Carsten Jensen, and Kees Van Kersbergen. 2017. “Research note: peeping at the corpus – What is really going on behind the equality and welfare items of the Manifesto project?” Journal of European Social Policy 27 (5): 403-16. Huber, Evelyne, Charles Ragin, and John D. Stephens. 1993. “Social Democracy, Christian Democracy, Constitutional Structure, and the Welfare State.” American Journal of Sociology 99 (3): 711-49. Jensen, Carsten. 2012. “Labour-Market- vs. Life-Course-Related Social Policies: Understanding CrossProgramme Differences.” Journal of European Public Policy 19 (2): 275-291. Jensen, Carsten. 2014. The Right and the Welfare State. Oxford: Oxford University Press. Jensen, Carsten, Arndt, Christoph, Lee, Seonghui, and Georg Wenzelburger. 2017. “Policy Instruments and Welfare State Reform.” Journal of European Social Policy 28 (2): 161-176. Jensen, Carsten, and Michael Bang Petersen. 2017. “The Deservingness Heuristic and the Politics of Health Care.” American Journal of Political Science 61 (1): 68-83. Kangas, Olli. 2009. “Finland: Labor Markets against Politics.” In The Handbook of West European Pension Politics, edited by Ellen Immergut, Karen M. Anderson, and Isabelle Schulze, 248-96. Oxford: Oxford University Press. Kitschelt, Herbert. 2001. “Partisan Competition and Welfare State Retrenchment: When Do Politicians Choose Unpopular Policies.” In The New Politics of the Welfare State, edited by Paul Pierson, 264-302. Oxford: Oxford University Press. Kittel, Bernhard, and Herbert Obinger. 2003. “Political Parties, Institutions, and the Dynamics of Social Expenditure in Times of Austerity.” Journal of European Public Policy 10 (1): 20-45. Klingemann, Hans-Dieter. 2006. Mapping Policy Preferences. Estimates for Parties, Electors, and Governments in Eastern Europe, European Union, and OECD 1990 - 2003. Oxford: Oxford Univ. Press. Klingemann, Hans-Dieter, Richard I. Hofferbert, and Ian Budge. 1994. Parties, Policies, and Democracy. Boulder: Westview. Korpi, Walter. 1983. The Democratic Class Struggle. London: Routledge. Korpi, Walter, and Joakim Palme. 2003. “New Politics and Class Politics in the Context of Austerity and Globalization: Welfare State Regress in 18 Countries, 1975-1995.” American Political Science Review 97 (3): 425-46. 136

Lee, Seonghui, Carsten Jensen, Christoph Arndt, and Georg Wenzelburger. 2017. “Risky Business? Welfare State Reforms and Government Support in Britain and Denmark.” British Journal of Political Science. Early view, DOI 10.1017/S0007123417000382. Loftis, Matt W., and Peter B. Mortensen. 2017. “A New Approach to the Study of Partisan Effects on Social Policy.” Journal of European Public Policy 24 (6): 890-911. Lowe, Will, Kenneth Benoit, Slava Mikhaylov, and Michael Laver. 2011. “Scaling Policy Preferences from Coded Political Texts.” Legislative Studies Quarterly 3 (1): 123-55. Palier, Bruno. 2010. A Long Goodbye to Bismarck. Amsterdam: Amsterdam University Press. Pierson, Paul. 1994. Dismantling the Welfare State? Reagan, Thatcher, and the Politics of Retrenchment. Cambrigde: Cambridge University Press. Pierson, Paul. 1996. “The New Politics of the Welfare State.” World Politics 48 (2): 143-79. Ross, Fiona. 2000a. “Framing Welfare Reform in Affluent Societies: Rendering Restructuring More Palatable?” Journal of Public Policy 20 (2): 169–93. Ross, Fiona 2000b. “Beyond Left and Right: The New Partisan Politics of Welfare.” Governance 1 (2): 155-83. Schmidt, Manfred G. 1982. Wohlfahrtsstaatliche Politik unter bürgerlichen und sozialdemokratischen Regierungen. Frankfurt a.M./New York: Campus. Stephens, John D. 1979. The Transition from Capitalism to Socialism. London: Macmillan Press. Svallfors, Stefan. 2012. “Welfare States and Welfare Attitudes.” In Contested Welfare States, edited by Stefan Svallfors, 1-24. Stanford: Stanford University Press. Van Kersbergen, Kees. 1995. Social Capitalism. London: Routledge. Van Kersbergen, Kees, and Philip Manow. 2009. Religion, Class Coalitions, and Welfare States. Cambridge: Cambridge University Press. Wenzelburger, Georg, Reimut Zohlnhöfer, and Frieder Wolf. 2013. “Implications of Dataset Choice in Comparative Welfare State Research.” Journal of European Public Policy 20 (9): 1229-50. Zohlnhöfer, R., Frieder Wolf, and Georg Wenzelburger. 2013. “Political Parties and Pension Generosity in Times of Permanent Austerity.” World Political Science 9 (1): 291-318.

137

The partisan politics of new social risks in advanced postindustrial democracies: Social protection for labor market outsiders Duane Swank1

Abstract: Advanced postindustrialization generates numerous challenges for the European social model. Central among these challenges is declining income, unstable employment, and inadequate training of semi- and unskilled workers. In this chapter, I assess the partisan basis of support for social policies that address the needs of these marginalized workers. I specifically consider the impacts of postindustrial cleavages among core constituencies of social democratic parties on the capacity of these parties to pursue inclusive social policies. I argue – and find support for in empirical analyses – that encompassing labor organization is the most important factor in strengthening the ability of left parties to build successful coalitions in support of outsider-friendly policies. I go beyond existing work on the topic by considering the full array of postindustrial cleavages facing left parties, by more fully elaborating why encompassing labor organization is crucial, and by considering a more complete set of measures of outsider policies than extant work. I compare my arguments and findings to important new work that stresses coalition building and partisan politics but minimizes the role of class organization.

1

Introduction

Advanced postindustrialization generates numerous challenges for the European social model. Central among these is the rise of ‘new social risks’ not well covered by core social insurance programs of the Twentieth Century welfare state. These risks include threats to income and job security for semi- and low-skilled workers generated by skilled-biased technological change and globalization, as well as new socioeconomic pressures on households that seek to balance work and family commitments and meet the skill/knowledge challenges of the postindustrial economy. Those new risks associated with employment and skills, often analyzed through the lens of labor market dualism, have been the subject of intense study.2 In addition, advanced postindustrialization creates pressures on the stability of political coalitions that undergird the European social model; this chapter will pay particular attention to this contemporary political challenge.

1

Marquette University [email protected] 2 See David Rueda (2005, 2007) for the first major analysis of the political ramifications of labor market dualism. See Emmenegger et al. (2012) for analysis of major features of labor market dualism and summaries of the large social science literature of the topic.

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 R. Careja et al. (eds.), The European Social Model under Pressure, https://doi.org/10.1007/978-3-658-27043-8_9

139

Generally, the literature on dualization stresses the bifurcation of postindustrial labor markets between insiders (e.g., skilled industrial sector workers) and outsiders, namely, younger, female, immigrant and less skilled workers increasingly concentrated in the service sector and on the unemployment rolls. Significant differences in wages and income, the incidence of involuntary part-time employment, and access to social insurance, services, job training and education characterize the divergent conditions faced by insiders and outsiders (e.g., Rueda 2007; Häusermann and Schwander 2012; Swank 2014). In this chapter, I assess the partisan basis of support for social policies that address the needs of marginalized workers. Following contributors in the seminal work of Amingeon and Bonoli (2006), I specifically consider the impacts of postindustrial cleavages among core constituencies of social democratic parties on the capacity of these parties to pursue inclusive social policies. (Indeed, one of the most important contributions of Klaus Armingeon has been to pioneer the study of the political consequences of the rise of new social risks.) I argue that encompassing labor organization is the most important factor in strengthening the ability of left parties to build durable and electorally successful coalitions in support of outsider-friendly policies. I go beyond existing work on the topic by considering the full array of postindustrial cleavages and attendant electoral challenges facing left parties, by more fully elaborating the logic of why encompassing labor organization is crucial, and by considering a more complete set of measures of social and labor market policies than existant work. In concluding, I compare my arguments and findings to important new work on social policy that stresses coalition building and partisan politics, but minimizes the role of class organization. 1.1

An overview of labor market outsider policies and outcomes

I focus on three dimensions of public policy specifically oriented to ameliorating the economic hardships of labor market outsiders. The first two policies are directed toward income protection for low income households: average income replacement rates of unemployment compensation across single and two earner households at 50 percent of the average production worker earnings and minimum income protection for outsiders. Minimum income protection is the sum of means-tested cash assistance plus family services and housing subsidies averaged across single, lone-parent, and two-parent households as a percent of the average production worker’s wage.3 In addition, I focus on national commitments for active labor market policy (hereafter ALMP), or ALMP spending as a percentage of GDP. AMLP programs include provision of an array of labor market services (e.g., job search assistance), training for major categories of labor market outsiders (e.g., unemployed youth, low-skilled workers displaced by postindustrialization), and publicly subsidized jobs. 4 I also include an analysis of the reduction 3

This measure taps the likely social wage available to younger and irregularly employed workers who do not meet or have exhausted eligibility requirements for normal unemployment compensation. 4 Models of ALMP and minimum income protection estimated here update and extend (with new controls, new 140

in the relative poverty rate through income transfers and taxes. I use 50 percent of median income as the poverty line and compute poverty reduction as the percentage change from pretax-and-transfer to post-tax-and-transfer poverty rates. The poverty rate itself is defined as the percentage of working-age households earning below 50 percent of median family income. 5 With this focus, I go beyond existing studies that have nearly exclusively focused on ALMP and general unemployment benefits as indicators of outsider policies (e.g., Rueda 2005, 2007; Iversen and Soskice 2015a). I display data for these four dimensions of outsider policies and outcomes in Table 1. The table provides values for the mid-1980s and early 2010s for the 18 focal nations of this study (grouped by welfare state regimes).6 Generally, as even casual observers would expect, Nordic social democratic welfare states score moderately higher in all areas. In both the 1980s and 2010s, the Nordic countries replace several percentage points more income for low wage workers through unemployment compensation and minimum protection programs than do the Continental (conservative) welfare states; the difference in these policies between Nordic and Anglo-liberal welfare states is in excess of 20 percentage points. A similar pattern holds with regard to all dimensions of outsider policies and outcomes. In addition, and surprisingly, change on each dimension in each welfare state type is, for the most part, remarkably modest. In fact, the aggregate averages for mid-1980s and early 2010s for all 18 nations differ by only one percentage point (i.e., in percent of income replaced) for low wage worker unemployment compensation and means-tested benefits. ALMP spending is identical in the mid-1980s and early 2010s while pre- to post-fisc poverty reduction only increases by four percentage points. That said, there are some notable changes within countries and welfare state types. For instance, while still relatively generous, minimum income support declines in the Nordic cases, unemployment compensation for low income workers increases notably in Italy and Switzerland and declines by more than 10 points in New Zealand, and ALMP spending declines by half or more in New Zealand and the UK. I now turn to the widely debated question of how partisan politics has shaped these programs and outcomes in the postindustrial era.

tests, corrected measures for some variables, and more years in the case of ALMP) initial analysis reported in Swank (2014). 5 See the online appendix for data sources for all variables used in this chapter. 6 I exclude both the Mediterranean countries because of sporadic data availability, as well as Central and Eastern European countries because of the time frame of this study. 141

142

Continental

Anglo-Liberal

81

76

71

45

72

76

74

60

77

67

78

74

Netherlands

Switzerland

Mean

Australia

Canada

Ireland

New Zealand

United Kingdom

United States

Mean

Japan (mixed)

Overall Mean

75

82

63

74

52

62

82

72

38

79

88

84

80

72

75

92

68

86

88

80

78

96

Early 2010s

47

51

43

28

39

48

54

51

38

47

59

43

48

45

44

38

51

63

69

59

69

56

Mid-1980s

48

72

44

20

48

47

66

41

41

46

55

44

53

49

40

34

49

53

55

55

54

48

Late 2000s

ALMP Spending

0.7

0.3

0.6

0.3

0.8

0.8

1.1

0.5

0.3

0.7

0.2

1.3

0.3

0.7

0.7

1.2

0.3

1

1.9

0.5

0.8

0.8

Mid-1980s

0.7

0.2

0.4

0.2

0.4

0.3

0.9

0.3

0.3

0.7

0.6

1.1

0.4

0.8

1

1.1

0.8

1.2

1.2

0.6

1

2

Early 2010s

(Percent of GDP)

42

NA

34

6

41

NA

56

38

30

40

11

57

NA

35

55

NA

NA

56

55

49

68

50

Mid-1980s

46

NA

38

28

51

NA

74

38

38

40

7

44

NA

46

61

NA

NA

62

68

63

58

57

Early 2010s

Relative Poverty Reduction

Table 1 reports unemployment compensation income replacement rates for 1985 and 2010 while minimum income protection is for 1990 and 2009. The table reports 1985-87 and 2010-2012 means for ALMP spending variable, and a mid-1980s and 2010 data points for reduction in relative poverty.

30

92

Belgium

Italy

73

Austria

74

88

Mean

Germany

4

Sweden

74

80

Norway

France

84

94

Mid-1980s

Cash and In-kind Benefits

Low Wage Worker

Finland

Denmark

Minimum Income Protection

Unemployment Comp RR

Table 1. Income protection for outsiders, active labor market spending, and relative poverty Rates in Postindustrial Democracies, 1980s – 2010s

Nordic

2

Theory: Parties, labor organization, and social protection of outsiders

Beyond the analysis of the nature of new social risks, contributors to Amingeon and Bonoli (2006) focused on the implications of these risks for parties’ electoral strategies, ideological orientations, and, most central to this chapter’s inquiry, partisan impacts on social policy itself. To begin, I consider the question of how partisan politics has influenced support for policies that aid labor market outsiders in light of the broader debate over whether or not “parties still matter” for social welfare policy. The first wave of studies on partisan government impacts on post-WW II social welfare policy consistently reported large positive impacts of social democratic (and to a more limited extent Christian democratic) government on social protection (e.g., Cameron 1978; Stephens 1979; contributors to Castles 1982; Hicks and Swank 1992; Huber, Ragin, and Stephens 1993).7 Yet, in the wake of 1970s and 1980s economic crises and postindustrialization, a group of scholars led by Paul Pierson (1994, 1996) questioned continued partisan policy effects. In this view, the new politics of welfare policy under postindustrial pressures for austerity and efficiency involved the politically difficult job of reducing concentrated benefits to welldefined, mobilized constituencies in return for future, diffuse benefits. In the face of continued political support for social programs, incumbent politicians had to engage in the politics of blame avoidance where the success of reforms depends on the character of program constituencies, the depth of lock-in effects of programs, and related programmatic traits. The partisan hue of governments has minimal impact on policy reforms. Beyond Piersonian ‘new politics’ dynamics, scholars have highlighted three important conflicts among postindustrial constituencies of left parties that may well undercut their ability to enact inclusive social policies. First, contributors to Armingeon and Bonoli (2006) as well as contemporary scholars (e.g., Gingrich and Häusermann 2015) highlight the divergent interests between new middle-class groups on the one hand, and traditional welfare state constituencies on the other. To elaborate, these scholars stress that younger, better educated, disproportionately female socio-cultural professionals, increasingly important numerically to social democratic parties, prefer social investment spending on family services and supports, education, and activation policies. Although lower-skilled blue-collar workers have an interest in ALMP programs, they and the broader working class have a strong preference for maintenance or expansion of income protection through the programs of the ‘traditional’ welfare state. Second, scholars have argued that a new conflict has emerged among key constituent groups of left parties as a function of changes in postindustrial occupational structures. As discussed above, David Rueda (2005, 2007) has argued that with respect to social democratic parties’ 7

As is commonly understood, scholars typically emphasized the balance of class power where labor mobilization through union organization, corporatist institutions, and left party rule signaled the relative advantage of low versus high income strata and the likelihood of redistribution (see Korpi 1983 for a classic statement). 143

constituencies, postindustrial labor markets increasingly consist of insiders (skilled industrial sector workers) and outsiders (younger, female, immigrant and less skilled workers increasingly concentrated in the service sector). Iversen and Soskice (2015a) make a similar point by noting that in the postindustrial production process, skilled and unskilled workers – heretofore intertwined in the structure of industrial production and politically integrated in a progressive social democratic-led coalition – are increasingly separated and dispersed across enterprises and sectors.8 For Rueda, the interests of insiders are in the maintenance of employment protection and the moderation of taxation of labor income, while outsiders favor both income transfers for shortterm security and active labor market programs for human capital development and, ultimately, secure employment. Rueda argues that in the absence of rises in unemployment and risks that threaten insiders, social democratic parties will maintain employment protection and moderate tax burdens but not increase transfers and active labor market programs (with their associated tax burdens) that enhance the incomes of outsiders. Generally, Rueda and collaborators (e.g., Lindvall and Rueda 2014) stress the complex electoral dilemma facing social democrats. Social democratic shifts to insider and middle-class appeals risk outsider defections (e.g, 1998 Swedish parliamentary elections); often left parties respond by returning to policy orientations that encompass outsider interests (e.g., 2006 and 2010 Swedish parliamentary elections). Third, Häusermann and Kriesi (2015) and Kitschelt and Rehm (2015), among others, adopt a different perspective and stress that while economic conflicts still divide left party constituencies, a strong univeralism-particularism dimension of cultural conflict creates additional, serious electoral dilemmas for social democrats. As noted, left parties increasingly rely on the support of socio-cultural professions who work in education, health, and family and cultural services; these citizens prefer policies that promote an inclusive, multicultural, and participatory society. Yet, social democrats must balance policies for this constituency with those for blue-collar wage earners who tend to favor exclusionary and traditionalist sociocultural policies.

8

For Iversen and Soskice (2015a), the challenge for social democrats is to forge a new electoral coalition between low-skilled workers increasingly concentrated in services, remaining high skilled blue-collar workers, and middle tier technicians and managers. As articulated in their past work (Iversen and Soskice 2006), the authors argue that this new lower- and middle-class coalition is most likely in PR systems, where left parties may credibly commit to a redistribution. Iversen and Soskice to do not fully account for programmatic divisions between new middle class and traditional working-class constituencies; they ignore the universalism-particularism cleavage discussed below. 144

Figure 1. Three sets of tradeoffs in Social Democratic party coalition building in advanced postindustrial societies

Figure 1 highlights these three conflicts – and sets of tradeoffs – as well as the positions of three central constituency groups of contemporary social democratic parties. While the degree of difficulty in coalition building will vary depending on the absolute and relative size of these groups in each nation, among other factors, left parties will, in all likelihood, have to strategically adjudicate multiple conflicts to create an electorally successful coalition that facilitates the development of an inclusive, solidaristic social policy. I now turn to an analysis of the conditions under which left parties are likely to be more successful in navigating these conflicts and, in turn, building coalitions around an inclusive social policy mix. 2.1

Labor organization

My central hypothesis is that encompassing labor organization, namely, high levels of union density, centralization of powers over constituent unions in one or a few peak associations, and state integration of unions (advisory boards, commissions, tripartite concertation, social program administration), provides a central mechanism through which social democratic parties can mitigate postindustrial cleavages and, in turn, build effective egalitarian coalitions. I make this argument for several reasons. In doing so, I build on Huo (2009) who has argued in his study of ALMP that encompassing labor movements moderate the tradeoffs faced by social democratic parties as they seek to balance the interests of insider and outsider groups. I also build on Thelen (2014), who has recently shown that labor organization is a key factor in Danish governments’ ability to embed market liberalization in policies that promote human capital development and preserve social protection. Encompassing labor organization effectively brought together white-collar, public sector, skilled manufacturing, and low-skill service workers (mitigating the insider-outsider and social investment-social insurance divides) in support of “embedded flexibilization.” I expand and elaborate the logic of these as well as my own recent arguments (Swank 2014; Martin and Swank 2012) and provide new evidence on 145

the role of labor organization in facilitating left party promotion of social protections for outsiders. First, union density is very important. As Huo (2009) and Iversen and Soskice (2015b) have noted, unions disseminate political information and provide forums for political discussion. This, in turn, fosters a better understanding of material interests by members and the capacity of members to make informed political choices. Mosimann and Pontusson (2017), moreover, make the crucial point that unions’ behavior and rhetoric create distributive norms (fairness, aversion to inequality) that members adopt as their own. In fact, Mosimann and Pontusson demonstrate in empirical analyses of individual-level data in 21 European democracies that union membership has a strong, positive effect on support for redistribution and that this effect increases with members’ income. Rennwald and Pontusson (2017) also show, and this is especially relevant to the universalism-particularism divide, that union membership significantly lowers the probability of working-class voters defecting from social democratic to right-wing populist parties. High levels of overall union density should also effectively incorporate many outsiders into the labor movement. Indeed, scholars have stressed that where unions remain relatively encompassing (Nordic countries and Belgium), public sector and low-wage service sector unions (and hence women, younger, and low-skilled workers) are represented alongside core sector industrial unions and white-collar workers (e.g., Ebbinghaus 2006; Huo 2009). In fact, the simple correlation between aggregate union density and the ratio of outsider to insider unionization rates is 0.56 in a sample of 15 European political economies in 2008 (Becher and Pontusson 2011). In addition, union centralization (that is, extensive control over political-economic strategies and behaviors of constituent unions by one or a few national peak associations) should reinforce the inclusiveness of relatively densely organized union movements. As Gordon (2015) has pointed out, highly centralized union movements have strong norms for the democratic representation of all constituent unions (including low-skilled unions) in peak association decision making. Moreover, centralized peak associations are more likely to promote coherence in norms and policy positions across the union movement crucial to the mitigation of cleavages, and the general disposition of such encompassing union peaks is to champion broad social justice goals (Hall 2017). The presence of unions within the state is also important. The integration of unions within policy-making forms (e.g., advisory boards and commissions) and administrative units in mature welfare states is not only likely to offer points of political leverage for the union movement (Swank 2002). It is also likely to foster support for redistributive programs among trade unions. As Gordon (2015) argues, the presence of Ghent system of union predominance in the administration of unemployment compensation, and to a lesser extent joint unionemployer administration of broad social insurance programs, is likely to cultivate union support of these programs as a vehicle of union maintenance and recruitment of members. 146

In sum, key features of encompassing union organization should integrate the core postindustrial constituencies of left parties, foster broad commitments to social justice and inclusive policies, and otherwise facilitate social democratic parties’ pursuit of policies that address the needs of outsiders. And, as foreshadowed in Walter Korpi’s (1983) seminal work on social democracy, encompassing union movements, at a rudimentary electoral level, still have large, significant effects on electoral mobilization (Radcliff and Davis 2000) and votes for left parties (Arndt and Rennwald 2016) in postindustrial democracies. 2.2

A note on employer organizations

I also account for employer organizations. Specifically, Martin and Swank (2012) have pointed out that high levels of employers’ organization cultivate a collective orientation to the long-run interests of employers (for instance, an interest in human capital). In addition, the repeated exchanges with labor and the state that occur at high levels of organization build trust, reciprocity and a commitment to the public interest among employers. These considerations, along with the tendency of highly organized employers to support social insurance that bolsters worker investment in specific skills, suggest that the greater the employer organization, the greater the social protection for outsider policies. Martin and Swank report clear evidence for this proposition in the case of ALMP.

3

An empirical analysis of the partisan basis of outsider policies and outcomes

In the following analysis, I focus on the four dimensions of outsider policies and outcomes discussed above: low-wage worker unemployment benefits, means-tested income protection, ALMP, and reduction in relative poverty. I draw on the large literature on the determinants of social policies and outcomes and develop and estimate empirical models of variations across space and time in outsider policies and poverty reduction for 13 to 18 developed capitalist democracies (see Table 1) for the years 1979-2012.9 The general model of outsider policies and poverty reduction is: [Eq. 1] Outsider Policy/Poverty Reductioni,t = α + β1 (Partisan Govt )i,mean t-10/5/3 to t-1 + β2(Labor Organization)i,t-1

+ Β3(Position of the Median Citizen/Income Skew)i,t-1 +

β4(Employer Organization)i,t-1 + β5(Trade Openness.)i,t-1 +

β6(Capital Mobility)i,t-1 +

β7(De-industrialization)i,t-1 + β8(Unemployment)i,t-1+ β8(Percent Change in Real per capita GDP)i,t-1 + εi,t

9

Data series begin between 1979 and the early 1980s and typically end around 2010-to-2012; Nelson’s data are available for 1990 to 2009 only. LIS household poverty data are typically available for time points every five years; the three low-income benefit measures are measured annually. 147

In the model of ALMP, I add a control for variations in skills demand of the economy in the form of a technological change variable. For the poverty reduction model, I add the lagged level of poverty reduction to tap autoregressive processes (see below for more on this point). For core political variables, I measure partisan control of government in the short and intermediate term. I do so for two reasons. First, effects of long-term cumulative partisan measures of left government are difficult to separate from those of labor union organization as they co-evolve across time. Second, much of the debate about the political sources of social protection and redistribution involves discerning the impact of alternation in left, center and right governments in the postindustrial era. That is, do partisan outcomes of democratic elections still fundamentally matter today? Long-term cumulative measures stretching back several decades give us limited power to address this question. Thus, I assess partisan impacts in the short term with a measure of the average ideological position of government over the last three years.10 Short-term (three-year) measures allow us to assess the relatively immediate policy effects of partisan policy choices. I also examine mean ideological scores of governments over the preceding five and 10 years. These measures of intermediate-term power in office are assessed because policy changes are often phased in over a few years and frequently take the form of discrete, limited reforms whose effects accumulate over an intermediate period of time. I use both short- and intermediate-term measures, alternatively, in each model, and, ultimately, report the most substantively and significantly important measure.11 The organization of labor is measured as a standard score index of union density, a 0.0 to 4.0 scale of centralization of powers in the largest national peak association (control of affiliate appointments, over strikes, of bargaining strategy, and of conflict funds), and a 0.0 to 1.0 scale of integration of unions into public policymaking forums. Employer organization is measured through a standard score index of the presence of a national employers’ peak association, 0.0 to 4.0 scale of powers of the peak (control of affiliate appointments, over lockouts, of bargaining strategy, and of conflict funds), and 0.0 to 1.0 scale of policymaking integration of employers. As to the economic position of the median citizen, I follow the conceptualization and measurement procedures of Lupu and Pontusson (2011) and use a ratio of ratios, or income skew: I compute the ratios of earners at the 90th and 50th percentiles and at the 50th and 10th percentiles. As suggested by theory, I use the ratio of these two numbers to capture relative closeness of median citizen/voter to low income earners. As to globalization, I control for imports and exports of goods and services as percentage of GDP and a general measure of capital mobility (0.0 to 100 scale of liberalization of capital 10

I measure the partisan control of government by weighting the ideological position of each governing party with that party’s share of total governing parties’ parliamentary (lower chamber) seats, and summing the score for all governing parties. The ideological position of each party is based on the party’s score on the 26-item scale of government intervention in markets and society developed in the Manifesto Data Collection project (Volkens et al. 2013). The resulting variable is scaled 0.0 to 100 (where 100 is the most left-leaning score possible). 11 In addition, I substituted the annual share of left party cabinet portfolios, a common alternative measure of partisan government, and I note these very similar results below. 148

flows); substitutions of trade with developing countries and actual capital flows produce results similar to those reported below for the main measures. Technological change in the ALMP equation is measured as patents per one million population. As to structural economic change, I follow Iversen and Cusack (2000) and measure deindustrialization as 100 minus industrial and agricultural employment as a percentage share of the working age population. Unemployment is measured as the percentage of the civilian work force unemployed (standardized across nations). As an additional control for economic conditions, I use growth rates in real per capita GDP (in international prices). I employ a one-year lag of all variables unless otherwise noted. To assess the effects of partisan government across levels of union organization, I estimate Eq. 1 with interaction terms; the equation for each measure of low-income policy and relative poverty reduction is as follows: [Eq. 2] Outsider Policy/Poverty Reductioni,t = α + β1 (Partisan Govt )i,mean t-10/5/3 to t-1 + β2 (Labor Organization)i,t-1 + … β9 (Social Democratic Govt × Labor Organization)i,t-1 + εi,t For estimation of outsider policy models, I use Prais-Winsten regression with first-order serial correlation and panel correct standard errors. In addition, as the time series for some nations occasionally begin after 1979 or end before 2012, I use a standard technique for unbalanced panels where elements of the variance-covariance matrix are computed with all available pairs of panels. As noted, for relative poverty reduction models that rely on data with somewhat variable numbers of years between observations, I use OLS regression with panel-corrected standard errors and a lagged dependent variable – a common alternative to Prais-Winsten regression (Beck and Katz 1996). To assess robustness of results (and check for potential bias from unmodeled country or time effects), I re-estimate all models with fixed and random effects estimators.12

4

Findings

The results of the analysis of the partisan basis of income protections for low-wage earners are presented in Table 2. As the table reveals, the general impact of left partisan government on unemployment benefit replacement rates and minimum income protection is indeed significant. In concrete terms, a shift, let’s say, in partisan government of 25 points on the 100-point ideological scale would increase unemployment compensation replacement rates for low-wage workers by 2.8 percent (25 × .1111) and means-tested benefits (as a percentage of an average production worker’s wage) by 4.5 percent (25 × .1809).13 This magnitude of partisan change 12

I do not use, for instance, a fixed effects estimator as my standard estimator because of the need to assess theoretically important variables with heavily cross-national variation (e.g., labor organization). In any case, fixed and random effects estimations reproduce the pattern of findings presented below. 13 Measures of party government over the previous five and 10 years produce modestly stronger findings than the three-year average score, I use the 10-year measure in all models as it proves slightly more robust than the five149

would be, for instance, equivalent to a shift from a moderate center-right government to a moderate center-left one. In addition, it is important to note that the direct impact of labor organization on unemployment benefits for low income workers is substantively large and statistically significant; a one (standard deviation) unit increase in the index of labor organization (for instance, a move from the degree of 1980s labor organization in New Zealand to the level in the Netherlands) would, itself, increase income replacement rates by three and a half percent. While positive, the substantive direct effect of union organization on minimum income protection is smaller and not significant at conventional levels. The second and fourth columns of Table 2 present the results for the basic assessment of the core hypothesis of the chapter, namely, that left government impacts on income protection for outsiders will differ across levels of labor organization. As the table suggests, the coefficient for the interaction between partisan government and labor organization is not significant for unemployment compensation; it is, however, clearly significant in the case of minimum income protection. Moreover, when one applies the advice of Kam and Franzese (2009) to the current context and examines marginal impacts of partisan government across specific levels of labor organization, it becomes clear that alternation in partisan governments has little impact on both types of outsider income protection when labor organization is low (e.g., Britain and the United States). Partisan impacts are, however, significant and substantively important for outsider policies at medium and high levels of labor organization.14 For low wage unemployment benefits, the impact of a 25 point upward shift in party government ideological position at medium levels of union organization (e.g., the Netherlands) is 2.8; at high levels of union organization (e.g., Austria and Denmark) the replacement rate impact is 3.4 percentage points. An identical pattern emerges with respect to minimum income protection benefits, and this is highlighted in Figure 2. As labor organization rises, the effect of social democratic government increases in substantive magnitude and significance. At high levels of union organization, benefits relative to median wages increase over seven percentage points under a left-leaning versus right-leaning government.

year average. 14 Recall that the effect of some variable X1 on Y at variable levels of X2 is given by β1 + β3[X2], where β3 is the coefficient for the interaction of X1 and X2. Standard errors for these marginal effects are readily computed (Kam and Franzese 2009). 150

Table 2. Partisan politics, labor organization, and income protection for labor market outsiders Unemployment Comp RR

Unemployment Comp RR

Minimum Benefits

Minimum Benefits

Left Government

.1111** (.0444)

.1083** (.0446)

.1809** (.392)

.1772** (.0376)

Labor Organization

3.5084**

2.5115

.9648

-3.4167

(1.1079)

(2.0960)

(1.1038)

(2.3140)

---

.0209 (.0356)

---

.0900** (.0410)

Left Government × Union Organization Economic Position of the Median Citizen Employers Organization

.3950

.5896

-7.7668

-6.4350

(4.770)

(4.7899)

(4.4055)

(4.3478)

.6798

.6348

3.9214**

3.9896**

(1.1820)

(1.1888)

(1.2332)

(1.1650)

Trade Openness

.0541** (.0184)

.0547** (.0185

.0017 (.0017)

.0054 (.0168)

Capital Market Openness

.0501**

.0508**

.0470

.0438

(.0302)

(.0302)

(.0378)

(.0373)

Deindustrialization

.0249

.0168

-1.0318**

-1.0876**

(.1369

(.1384)

(.1838)

(.1828)

-.1379** (.0588)

-.1394** (.0589)

-.0746 (.0702)

-.0810 (.0704)

Unemployment

-.1331 (.1496)

-.1288 (.1499)

.2818** (.1331)

.3134** (.1318)

Constant

59.8044

60.1783

123.3917

126.5812

.7976

.7979

.7873

.7903

434

434

340

340

Economic Growth Rate

R2 Observations

Models of the unemployment compensation replacement rate for low income workers are estimated with 19782010 data for 18 advanced democracies (see Table 1) by OLS regression; models of minimum income support are estimated with 1990 to 2009 data. The table reports unstandardized regression coefficients and panel-correct standard errors; *significant at the .10 level; **significant at the .05 level.

151

10 12 14 8 6 4 2 0 -4 -2

-2

-1

0 Union Organization Leftparty CI95Lower

1

2

CI95Upper

Figure 2: Left party effects on minimum benefits

The estimates of the effects of the models’ other variables are presented in subsequent rows of Table 2. While of secondary interest here, one might note the following findings: core features of globalization, trade and capital openness, bolster the level of unemployment benefits for low wage workers; globalization’s impacts are also positive and near significant in the minimum income protection models. In addition, there are substantively large need effects. For unemployment benefits, slow economic growth is associated with relatively more generous support for low-wage workers; for minimum income protection, high unemployment is associated with more generous assistance. Finally, two additional findings might be mentioned: employer organization is positively associated with both forms of outsider policy, but these effects are only significant for minimum income protection. Second, deindustrialization has a significant, negative impact on minimum income protection. One interpretation for this finding is that deindustrialization generally promotes the shift to an activation strategy and this, in turn, inherently entails reductions in the generosity of long-term income assistance for outsiders.

152

Table 3. Partisan politics, labor organization, ALMP, and the reduction of relative poverty for labor market outsiders ALMP Spending (% GDP) Left Government Labor Organization Left Government × Union Organization Economic Position of the Median Citizen

ALMP Spending (% GDP)

Poverty Reduction

Poverty Reduction

.0036**

.0031*

.1601**

.1577**

(.0023)

(.0020)

(.0766)

(.0795)

.0998**

-.1398

4.706**

3.7467

(.0506)

(.1173)

(2.4201)

(4.2945)

---

.0050** (.0021)

---

.0188 (.0730)

.0627

.0904

-2.1445

-1.7683

(.1712)

(.1701)

(9.2321)

(9.0349)

Employers Organization

.2599** (.0815)

.2550** (.0804)

2.5703 (2.0025)

2.5741* (2.0039)

Technological Change

-.0001 (.0002)

-.0001 (.0002)

---

---

Trade Openness

.0008

.0009

-.0823

-.0778

(.0011)

(.0011)

(.0872)

(.0659)

.0026**

.0026**

.0929

.0875

(.0014)

(.0013)

(.0886)

(.0967)

.0048 (.0053

.0042 (.0053)

.3832** (.2444)

.3762* (.2406)

-.0096**

-.0089**

-.0197

-.0256

(.0030)

(.0030)

(.2407)

(.2387)

.0211**

.0221**

1.3071**

1.3287**

(.0077)

(.0076)

(.4186)

(.4562)

---

---

.6405** (.0838)

.6416** (.0828)

Constant

-.2548

-.2403

-30.5050

-29.9139

R2

.2574

.2667

.8759

.8760

454

454

77

77

Capital Market Openness Deindustrialization Economic Growth Rate Unemployment Poverty Reduction Previous Period

Observations

Models of ALMP spending are estimated with 1979-2012 data for 18 advanced democracies by OLS regression; models of relative poverty are estimated with early 1980s to 2010 data for 13 nations. The table reports unstandardized regression coefficients and panel-correct standard errors; * significant at the .10 level; ** significant at the .05 level.

The results of the analysis of partisan and labor organization impacts on ALMP and poverty reduction are presented in Table 3. As in the case of income protection for low income workers, left party government is associated with significantly higher ALMP spending and poverty reduction than under center-right governments. The direct impacts of party government are substantively important. As shown in the first column, a shift of 25 points in the ideological position of the government would shift ALMP spending by .1 (25×.0036). Given that mean 153

ALMP spending as a percent of GDP is 0.8, this change in resource commitments for ALMP is not trivial. Similarly, a 25 point shift in government is associated with a 4 percent point increase (25× .1601) in pre-fisc to post-fisc relative poverty reduction. The direct impacts of labor organization (first and third columns) are also substantively important: ALMP spending and poverty reduction increase .1 and 3.7, respectively, for a one (standard deviation) unit change in labor organization. With regard to the mediation of partisan government effects by labor organization, Table 3 illustrates that higher labor organization, as expected, intensifies the impact of party government on ALMP. For instance, at average levels of labor organization (0.0 on the standard score index), the impact of a one unit change in party government is .0031.15 At moderately high labor organization (say a standard score index value of 1), the partisan government impact is .0081 (.0031+.0050×1.0). If we use a 25 point shift on the ideological position scale as the magnitude in change in party government, ALMP spending would increase by .1 at average levels and .2 at moderately high levels of labor organization with such a partisan change. On the other hand, the interaction between party government and labor organization is not significant in the poverty reduction model. Yet, recalling Kam and Franzese’s (2009) suggestion, when we examine individual marginal effects of partisan government, we do see important differences in party impacts across levels of labor organization: party alternation in power has no effect on the magnitude of poverty reduction at low levels of labor organization but substantively important ones at moderate and high amounts of labor organization (3.9 and 4.6 percent, respectively). With respect to other factors in the model, three findings deserve note. Employers’ organization has significant positive effects, as predicted by Martin and Swank (2012), on ALMP; it is also significantly associated with poverty reduction in one of two models (and nearly so in the other).16 Second, increases in international capital mobility are significantly related to ALMP spending, although international openness variables are not associated with poverty reduction. Finally, needs have substantial impacts on both ALMP and poverty reduction. In the case ALMP, a drop in economic growth and a rise unemployment rates both significantly increase ALMP spending. For poverty reduction, an increase in unemployment has a large substantive effect on the magnitude of poverty reduction: a one percentage point rise in the unemployment rate would, alone, increase poverty reduction by roughly 1.3 percent points.

15

Recall that the impact of x1 when the mediating variable, x2, is 0.0, is given by the coefficient for the direct effect of x1 in the equation. 16 It is relevant to mention that as predicted by Swank and Martin (2001), the impacts of left party government on ALMP also increase at higher levels of employer organization; this is not the case, however, with income protection variables nor poverty reduction. 154

5

Conclusions

This chapter speaks to a central question in the debate over the nature of the new politics of the welfare state, namely, does partisan government shape the state’s responses to the challenges of new social risks in postindustrial society? Specifically, I analyze the impacts of partisan government on policies to address labor market dualism from the perspective of coalition building by social democratic parties. I find systematic empirical evidence that left parties have significant, substantively important impacts on key features of income protection for low income workers, ALMP, and pre- to post-fisc relative poverty reduction in the 1980s to 2010s period. These results are in line with the contributors to Armingeon and Bonoli (2006), who stressed the general importance of the power resources model in explaining policies to address new social risks; the findings are also in line with other scholars such as Korpi and Palme (2003) and Huo (2009) who continue to stress the central importance of social democratic government to egalitarian social policy in the postindustrial era. This chapter also highlights the importance of the organization of producer interests to postindustrial welfare state politics. As Paul Pierson (2016) recently noted in a commentary on Beramendi et al.’s (2015) sweeping analytical assessment of trajectories of economic and social policy in advanced postindustrial capitalism, the role of electoral politics, coalition building, and partisanship is central to the explanation of policy change; the role of class organization is not. The arguments and findings presented here are both complementary to this recent work and implicitly critical of it. On the one hand, this chapter’s arguments and findings confirm the importance of electoral politics in shaping responses to the challenges of labor market dualism. On the other hand, my findings diverge from Beramendi et al. (2015), Iversen and Soskice (2015a), and others who minimize the importance of class organization in the contemporary periods. As this chapter shows, encompassing labor organization has a significant, substantively large direct effect on social policies for outsiders and, ultimately their relative economic status; labor organization also systematically mediates the capacity of partisan governments to adjudicate deep divisions between core constituencies and, in turn, pursue solidaristic policies that address the needs of outsiders. Moreover, as stressed by Martin and Swank (2012), one must also recognize the importance of employer organization for some forms of inclusive social and labor market policy. This is especially the case in the area of ALMP; earlier evidence offered by Martin and Swank on the salience of employer organization is reinforced here with new analyses of ALMP that extend over a decade into the 21st century. In sum, the analysis and conclusions of this chapter reinforce the initial assessments of Armingeon and collaborators as well as the array of like-minded scholars discussed above: the politics of new social risks, and especially the politics of inclusive social and labor market policy, is a story of the dynamics of coalition building in advanced postindustrial societies, the continued importance of partisan government, and the power of organized interests. This story 155

is not unlike, in broad structural terms, the well-established narrative of the politics of 20th century welfare state development.

References Armingeon, Klaus, and Giuliano Bonoli, eds. 2006. The Politics of Post-Industrial Welfare States: Adapting Postwar Social Policies to New Social Risks. London: Routledge. Arndt, Christopher, and Line Rennwald. 2016. “Union Members at the Polls in Diverse Trade Union Landscapes.” European Journal of Political Research 55 (4): 702-722. Beck, Nathaniel, and Jonathan N. Katz. 1996. “Nuisance versus Substance: Specifying and Estimating TimeSeries-Cross-Section Models.” Political Analysis 6: 1-36. Beramendi, Pablo, Silja Häusermann, Herbert Kitschelt, and Hanspeter Kriesi, eds. 2015. The Politics of Advanced Capitalism. New York: Cambridge University Press. Becher, Michael, and Jonas Pontusson. 2011. “Whose Interests do Unions Represent? Unionization by Income in Western Europe,” In Comparing European Workers Part B: Policies and Institutions. Research in the Sociology of Work [Volume 22], edited by David Brady, 181-211. Bingly, UK: Emerald Group Publishing. Cameron, David. 1978. “The Expansion of the Public Economy: A Comparative Analysis.” American Political Science Review 72 (4): 1243-1261. Castles, Francis, ed. 1982. The Impact of Parties. Beverley Hills, CA: Sage. Davidsson, Johan B., and Patrick Emmenegger. 2012. “Insider-Outsider Dynamics and the Reform of Job Security Legislation.” In The Politics of the New Welfare State, edited by Giuliano Bonoli, and David Natali, 206229. New York: Oxford University Press. Ebbinghaus, Bernhard. 2006. “Trade Union Movements in Post-Industrial Welfare States: Opening up to new social interests.” In The Politics of Post-Industrial Welfare States, edited by Klaus Armingeon and Giuliano Bonoli, 12-142. New York: Routledge. Emmenegger, Patrick, Silja Häusermann, Bruno Palier, and Martin Seeleib-Kaiser, eds. 2012. The Age of Dualization: The Changing Face of Inequality in Deindustrializing Societies. New York: Oxford University Press. Gingrich, Jane, and Silja Häusermann. 2015. “The Decline of the Working-Class Vote, the Reconfiguration of the Welfare Support Coalition and Consequences for the Welfare State.” Journal of European Social Policy 25 (1): 50-75. Gordon, Joshua C. 2015. “Protecting the Unemployed: Varieties of Unionism and the Evolution of Unemployment Benefits and Active Labor Market Policy in the Rich Democracies.” Socio-Economic Review 13 (1): 79-100. Hall, Peter A. 2017. “The Political Sources of Social Solidarity.” In The Strains of Commitment: The Political Sources of Solidarity in Diverse Societies, edited by Keith Banting and Will Kymlicka, 201-232. New York: Oxford University Press. Häusermann, Silja, and Hanspeter Kriesi. 2015. “What Do Voters Want? Dimensions and Configurations in Individual-Level Preferences and Party Choice.” In The Politics of Advanced Capitalism, edited by Pablo Beramendi, Silja Häusermann, Herbert Kitschelt, and Hanspeter Kriesi, 202-230. New York: Cambridge University Press. Häusermann, Silja, and Hanna Schwander. 2012. “Varieties of Dualization? Labor Market Segmentation and Insider-Outsider Divides Across Regions.” In The Age of Dualization: The Changing Face of Inequality in Deindustrializing Societies, edited by Patrick Emmenegger, Silja Häusermann, Bruno Palier, and Martin Seeleib-Kaiser, 27-51. New York: Oxford University Press. Heston, Allen, Robert. Summers, and Bettina Aten. 2012. Penn World Table Version 7.1. Center for International Comparisons of Production, Income and Prices at the University of Pennsylvania, July. Hicks, Alex, and Duane Swank. 1992. “Politics, Institutions, and Welfare Spending in Industrialized Democracies, 1960-1992.” American Political Science Review 86 (3): 659-674. Huber, Evelyne, Charles Ragin, and John D. Stephens. 1993. “Social Democracy, Christian Democracy, Constitutional Structure and the Welfare State.” American Journal of Sociology 99 (3): 711-749. Huo, Jingjing. 2009. Third Way Reforms: Social Democracy after the Golden Age. New York: Cambridge University Press. IMF. selected years a. Direction of Trade Statistics Yearbook. Washington, DC: IMF. International Monetary Fund. IMF. selected years b. Balance of Payments Statistics. Washington, DC: International Monetary Fund. Iversen, Torben, and Thomas R. Cusack. 2000. “The Causes of Welfare Expansion: Deindustrialization or Globalization?” World Politics 52 (3): 313-349. Iversen, Torben, and David Soskice. 2006. “Electoral Institutions and the Politics of Coalitions: Why Some 156

Democracies Redistribute More than Others.” American Political Science Review 100 (2): 165-181. Iversen, Torben, and David Soskice. 2015a. “Democratic Limits to Redistribution: Inclusionary versus Exclusionary Coalitions in the Knowledge Economy” World Politics 67 (2): 185-225. Iversen, Torben, and David Soskice. 2015b. “Information, Inequality and Mass Polarization: Ideology in Advanced Democracies.” Comparative Political Studies 48 (13): 1781-1813. Kam, Cindy, and Robert J. Franzese. 2009. Modeling and Interpreting Interactive Hypotheses in Regression Analysis, Ann Arbor: University of Michigan Press. Kitschelt, Herbert, and Philipp Rehm. 2015. “Party alignments: Change and Continuity.” In The Politics of Advanced Capitalism, edited by Pablo Beramendi, Silja Häusermann, Herbert Kitschelt, and Hanspeter Kriesi, 179-201. New York: Cambridge University Press. Korpi, Walter. 1983. The Democratic Class Struggle, London: Routledge. Korpi, Walter, and Joakim Palme. 2003. “New Politics and Class Politics in the Context of Austerity and Globalization.” American Political Science Review 97 (3): 425-446. Lindvall, Johannes, and David Rueda. 2014. “The Insider-Outsider Dilemma” British Journal of Political Science 44 (2): 460-475. Lupu, Noam, and Jonas Pontusson. 2011. “The Structure of Inequality and the Politics of Redistribution.” American Political Science Review 105 (2): 316-36. Martin, Cathie Jo, and Duane Swank. 2012. The Political Construction of Business Interests: Coordination, Growth, and Inequality, New York: Cambridge University Press. Mosimann, Nadja, and Jonas Pontusson. 2017. “Solidaristic Unionism and Support for Redistribution in Contemporary Europe,” World Politics 69 (3): 448-492. Nelson, Kenneth. 2007. Introducing SaMip: The Social Assistance and Minimum Income Protection Interim Dataset, S-WoPEc No. 11/2007. Swedish Institute for Social Research. Pierson, Paul. 1994. Dismantling the Welfare State: Reagan, Thatcher and the Politics of Retrenchment in Britain and the United States. New York: Cambridge University Press. Pierson, Paul. 1996. “The New Politics of the Welfare State.” World Politics 48 (2): 143-179. Peirson, Paul. 2016. “Review Symposium, The Politics of Advanced Capitalism, edited by Pablo Beramendi, Silja Häusermann, Herbert Kitschelt, and Hanspeter Kriesi, Cambridge University Press” (with Peter Hall, Christoffer Green-Pedersen, and Jonas Kraft). Socio-Economic Review 14 (2): 383-394. Radcliff, Benjamin, and Patricia Davis. 2000. “Labor Organization and Electoral Participation in Industrial Democracies.” American Journal of Political Science 44 (1): 132-141. Rennwald, Line, and Jonas Pontusson. 2017. “Paper Stones Revisited: Class Voting, Unionization, and the Electoral Decline of Mainstream Left.” Paper presented at the Annual Meetings of the American Political Science Association, San Francisco, August 31- September 3. Rueda, David. 2005. “Insider-Outsider Politics in Industrialized Democracies.” American Political Science Review 99 (1): 61-74. Rueda, David. 2007. Social Democracy Inside Out: Partisanship and Labor Market Policy in Industrialized Democracies, New York: Oxford University Press. Scruggs, Lyle, Detlef Jahn, and Kati Kuitto. 2017. “Comparative Welfare Entitlements Dataset 2. Version 201709.” University of Connecticut & University of Greifswald. Stephens, John D. 1979. The Transition from Capitalism to Socialism, Atlantic Highlands, NJ: Humanities Press. Swank, Duane. 2002. Global Capital, Political Institutions, and Policy Change in Developed Welfare States. New York: Cambridge University Press. Swank, Duane. 2014. “The Political Sources of Labor Market Dualism in Post-industrial Democracies.” Review of Keynesian Economics 2 (2): 234-57. Swank, Duane, and Cathie Jo Martin. 2001. “Employers and the Welfare State: The Political Economic Organization of Firms and Social Policy in Contemporary Capitalist Democracies.” Comparative Political Studies 34 (8): 889-923. Thelen, Kathleen. 2014. Varieties of Liberalization and the New Politics of Social Solidarity, New York: Cambridge University Press. Traxler, Franz, Sabine Blaschke, and Bernhard Kittel. 2001. National Labour Relations in International Markets. New York: Oxford University Press. Traxler, Franz, and Gerhard Huemer, eds. 2007. Handbook of Business Interest Associations, Firm Size and Governance. New York: Routledge. Van Vliet, Olaf, and Koen Caminada. 2012. “Unemployment Replacement Rates Dataset among 34 Welfare States 1971-2009.” NEUJOBS Special Report No. 2, www.neujobs.eu. Visser, Jelle. 2013. Data Base on Institutional Characteristics of Trade Unions, Wage Setting, State Intervention and Social Pacts, 1960-2011 (ICTWSS), Amsterdam Institute for Advanced Labour Studies AIAS, University of Amsterdam, Version 4.0, May. Volkens, Andrea, Pola Lehmann, Nicolas Merz, Sven Regel, and Annika Werner. 2013. The Manifesto Data Collection. Manifesto Project, Berlin: Wissenschaftszentrum Berlin für Sozialforschung (WZB) 157

The politics of shifting burdens: German fiscal welfare corporatism Christine Trampusch1 Abstract: Shifting fiscal burdens between the federal budget and social insurance funds served as an important stabilizer of Germany’s welfare corporatism and was supported by political parties of different colours as well as by social partners. On the one hand, it helped to offload the costs of industrial restructuring onto the social insurance budget, which removed from employers and unions the burden of coping with this challenge by wage policies; on the other hand, it relieved the Finance Minister from having to raise taxes in order to tackle labour market problems. However, over time the very success of these fiscal moves triggered the breakdown of German welfare corporatism. The chapter’s main lesson is that the fiscal perspective contributes to a better understanding of endogenously generated change in the welfare state because it helps to spotlight the exhaustion triggered by negative feedback as a mechanism of transformative change.

1

Introduction

For decades, the welfare state and its generous social benefits have been the most influential and stable institutions in Germany’s political economy, contributing enormously to the continuity of collective economic coordination in this country (Manow 2001, 2013; Manow and Seils 2000). Social policies were a consensus machine and matched very well the challenges of Germany’s production regime. Employers and unions were not only incorporated in the administration of social insurance schemes but also strongly affected social policy reforms. The welfare state helped the country to cope with several employment and economic crises as redundant workers were retired or redeployed in short term work or further training (Ebbinghaus 2006). The welfare state also served to restructure East Germany and finance the social costs of German unification. Various studies show that conditions have since changed (for an overview see Thelen 2012). After decades plagued by reform backlog (Reformstau) due to multiple institutional vetoes, vested interests, cross-party and cross-class coalitions, since mid-1990s German social policy has quite rapidly undergone a fundamental process of transformation and reforms. Benefits have been cut. The early retirement regime has been abolished and active labour market policy is now re-directed to workfare. Furthermore, reforms have also reduced the state’s role in providing social services. In pension, health care and labour 1

Universität zu Köln, Cologne Center for Comparative Politics, CCCP [email protected]

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 R. Careja et al. (eds.), The European Social Model under Pressure, https://doi.org/10.1007/978-3-658-27043-8_10

159

market policy, markets and private self-responsibility have been fundamentally strengthened and eligibility criteria for public benefits have been tightened. Although there are different explanations for this change – ranging from liberalisation (Streeck 2009; Vail 2008) and through dualisation (Palier and Thelen 2010; Seeleib-Kaiser 2002) to institutional exhaustion due to fiscal overload (Trampusch 2009; Busemeyer and Trampusch 2013) – there is consensus in the literature that the old German welfare state model has evaporated and some of its former institutional arrangements have been demolished (Thelen 2012, 138). How can we explain that for decades the welfare state was so exceptionally important to the German political economy and so strongly backed by all political parties, employers and unions? And why has Germany’s welfare corporatism broken down so rapidly since the mid1990s? An in-depth study of major welfare legislation between the mid-1960s and mid-1990s, this study addresses these questions from a fiscal-centred perspective (FCP). This FCP questions that social insurance schemes are “closed financial systems”. Social insurances are embedded in their broader political economic environment and state-society relations. In addition, the FCP follows a coalitional approach and argues that bargains among actors legitimize and stabilize economic, fiscal and financial functions of social insurance systems. Following this reasoning, this chapter argues that Germany’s social insurance system played an important economic and fiscal role, which not only explains why political parties and social partners were so strongly committed to it, but also helps to account for the recent erosion of the traditional welfare state consensus.2 Fiscal motives strongly affected the decision-making process in social policy, and contributed to both the continuity as well as the erosion of the German welfare corporatism. In detail, the chapter shows that the generous German welfare state was “financed” and stabilized by shifting expenditure between social insurance funds and the federal budget.3 This policy of shifting burdens (in German: Verschiebebahnhofpolitik) was backed by a cross-class and cross-party alliance between social politicians, employers, trade unions, works councils and public social policy administrators. The process was used not only to relieve the federal budget of cyclical social expenditures. Since the end of the 1970s, it was also implemented as a way to shunt expenditures between the budgets of different social insurance providers. The cross-class and cross-party alliance was in favour of offloading the costs of industrial restructuring, rationalization and German unification, first onto the social insurance funds, and later, with the rejection of a further rise in social insurance contribution rates, onto the federal budget as well. This arrangement of fiscal welfare corporatism helped the country to cope with various socioeconomic and political challenges. However, the offloading of costs of industrial restructuring onto the social insurance budget also resulted in severe fiscal resource constraints and prompted the rise of non-wage labour costs. Both (unintended, but over decades also partially accepted) 2

For cost shifting between the different levels of German federalism, please see Bonoli and Trein (2006). This chapter focuses on cost shifting within the social insurance system. 3 The data for the empirical analysis on the fiscal relationships between the German social insurance schemes have been taken from Trampusch (2003). 160

consequences resulted in fundamental tensions, which broke up German welfare corporatism in the mid-1990s. Thus, negative feedback effects (Weaver 2010) and exhaustion increasingly undermined fiscal welfare corporatism and became the seeds of its own destruction; exhaustion turned collaboration into confrontation (Streeck 2009; Busemeyer and Trampusch 2013). The aim of this chapter is to contribute to the development of a theory by suggesting that a fiscal-centred perspective helps to understand the state-society relations in advanced welfare states. Germany is a suitable case for this explorative objective because (before the transformative reforms) it constituted a typical (in the sense of representative) case of welfare state generosity. Defined in this inductive manner, according to Gerring (2007, 91), typical cases serve an “exploratory role” and “are helpful in exploring causal mechanisms”. Therefore, this within-case study does not “test” theories, but is aimed at generating theory, i.e. demonstrating the explanatory value of the fiscal-centred perspective which highlights exhaustion (prompted by negative, fiscal feedback effects) as a mechanism of transformative change. The data sources for this analysis are secondary literature on German social and fiscal policies (e.g. Manow and Seils 2000; Nullmeier and Rüb 1991; Mackscheidt 1990) as well as a database on the fiscal consequences of welfare legislation between 1949 and 2000. The study refers only to pension and unemployment insurance and leaves the health insurance budget aside. The period of the analysis is 1964 to 2000. In methodological terms, it is not only based on a comprehensive documentary analysis of the legislative acts gathered in the database, but also on a historical reconstruction of various legislative decision-making processes through process tracing. Process tracing is used as a means of discovering “causal-process observations” (Mahoney 2010, 124; Trampusch and Palier 2016). The rest of the chapter is divided into five parts. After this introduction, the second section presents the historical institutionalist approach of the study and argues that institutions may serve a multitude of interests, but require for their maintenance support coalitions (Mahoney and Thelen 2010; Busemeyer and Trampusch 2013). Section three presents data and describes the policies of shifting burdens. It shows how welfare legislation served to shift expenditure between social insurance funds and the federal budget. On the basis of an in-depth study of three typical pieces of welfare legislation, section four maps and explains how this policy has occurred. It analyses the underlying politics of shifting burdens. It shows that fiscal welfare corporatism was supported by politicians and social partners because these actors benefitted from the shift. While employers and unions appreciated the policy because it contributed to separating wage policies from the costs of industrial restructuring, the Finance Minister benefitted because a system financed by contribution tackled labour market problems without raising taxes. Section five describes the effects of these fiscal policies on German welfare corporatism and maps the system’s exhaustion due to negative feedback. The unintended consequences of shifting burdens resulted in fundamental tensions among social partners and political parties that triggered transformative change. The concluding section six discusses the 161

implications of the findings for further research on the fiscal-social nexus in advanced welfare states.

2

The welfare state from a fiscal perspective

For various reasons, a historical institutionalist perspective is particularly useful for analysing welfare states from a fiscal perspective. The first reason is that historical institutionalism is an inductive approach because it “scour[s] the historical record for evidence about why the historical actors behaved as they did” (Hall and Taylor 1996, 21). The inductive perspective implies that historical institutionalists are open-minded researchers who assume that actors may follow various and multiple preferences, interests and strategies. Institutions are viewed as “common carriers” (Pierson 2004, 109) for a multitude of interests. Consequently, in the case of welfare legislation, politicians and social partners may pursue social policy objectives, but it could also be that their policy-making is motivated by economic or fiscal concerns. To put it in other words: in terms of theory, historical institutionalism is an eclectic approach as it can be combined with various theoretical considerations. It compares these different hypotheses about actors’ motivations with the historical record. This procedure allows an in-depth analysis of welfare states beyond social policy concerns by highlighting their economic and fiscal functions. The second reason for drawing on historical institutionalism is its perspective on the distribution of power (for the following see also Busemeyer and Trampusch 2013). The analysis of power relationships makes this framework dynamic and appropriate for analysing endogenously generated change triggered by feedback effects. Historical institutionalists claim that the continuity and change of institutions depend on the distribution of power relationships. They contend that the stability of socio-economic institutions “requires the ongoing mobilization of political support as well as, often, active efforts to resolve institutional ambiguities in their favor” where “one important source of change will be shifts in the balance of power” (Mahoney and Thelen 2010, 11). However, in contrast to traditional power resource theory, historical institutionalism acknowledges that the interests of actors can change over time as they alter their strategies or react to changes in the environment. This assumption facilitates the analysis of dynamic processes triggering institutional change. Actors may initially support the same institutions for very different reasons because they benefit variously from institutions (Hall and Thelen 2009, 14). But in the long run, institutions may only approximately “hit the target” favoured by the originally politically dominant support alliance and thus generate unintended consequences. While an actor may benefit from some of these negative feedback effects, others may be harmed (e.g. Weaver 2010). As will become clear below, this chapter argues that this is exactly what happened in Germany during the last several decades. These unintended consequences lead us to a third reason, which makes historical institutionalism rather attractive for this study: the explanation of institutional change. 162

Historical institutionalists have identified various endogenous mechanisms of change – causal processes which signify how change is triggered by “actions undertaken by actors within the institution” and not by exogenous sources (external shocks), that is, events and processes originating outside the institution (Deeg 2001, 11). Besides drift, conversion, layering, and displacement, exhaustion is one variant of these mechanisms, as suggested by recent scholarship (Streeck and Thelen 2005; Trampusch 2005b; Streeck 2009; Busemeyer and Trampusch 2013). Busemeyer and Trampusch (2013, 294) refined the concept thus, “exhaustion describes a causal mechanism which drives processes of change as a consequence of endogenously generated negative feedback effects”. In the following sections, I provide a case study of how, over decades, the policies of shifting burdens were prompted and maintained by a cross-class and cross-party support coalition. In addition, I show that fiscal motives dominated in social policy decision making but, that in the end the policy of shifting expenditure burdens created negative feedback effects. Those feedback effects triggered the breakdown of German welfare corporatism by processes of exhaustion.

3

The policy of shifting expenditure burdens

Between 1964 and 2000, in nearly every year, pensions or labour market policy legislation explicitly altered the fiscal relationship between pensions insurance, unemployment insurance and the federal budget. This section sketches the policy of shifting fiscal burdens in order to demonstrate how fiscal motivations led to some of the most important social policy decisions in the heyday of German welfare corporatism. Before mapping the historical development of these policy measures, some technical notes on the modes of fiscal shifts and the study’s underlying database are necessary. After this presentation of data on the policy, the fourth section maps the cross-class and cross-party consensus behind fiscal welfare corporatism and thus the support coalition, which sustained this institution for decades. With regard to the concrete measures of shifting policy, two different modes of fiscal changes can be distinguished. On the one hand, there are changes which directly alter the relations between the different budgets. These direct shifts refer to alterations of allowances (Zuschüsse), the social insurance contributions of beneficiaries (Sozialversicherungsbeiträge der Leistungsempfänger) and reimbursements (Erstattungen). On the other hand, we find indirect shifts. These alter the streams by rule changes, which, although not explicitly directed to fiscal aspects, do generate shifts because they relocate benefits from one budget to another (Aufgabenverlagerung), change the regulation of the benefits (Änderung des Leistungsrechts) or alter the social contributions of employees (Sozialversicherungsbeiträge der Beschäftigten)4. While direct shifts are identified and quantified as stated in the bills (Begründungen zu den 4

The Sozialversicherungsbeiträge der Beschäftigten comprise equal social insurance contributions by workers and employers. 163

Gesetzentwürfen), indirect shifts are more difficult to discern because their discovery requires a profound knowledge of social policy law. That said, it becomes clear that a comprehensive collection of all extant shifts is impossible. This fact is also acknowledged by secondary literature on the Verschiebebahnhofpolitik (e.g. Klose and Schellschmidt 2001) as well as by the Federal Government’s Social Report (Sozialbericht) (BT-Drks. 13/10142, 253). Table 1 gives examples of direct and indirect fiscal shifts. Table 1: Direct and indirect shifts: Examples Shifts

Examples

Direct shifts Modification of allowances (Zuschüsse)

Modification of allowances of the federal government to the pension insurance scheme

Modification of beneficiaries’ social security contributions (Sozialversicherungsbeiträge der Leistungsempfänger)

Introduction/modification of pension insurance contributions of beneficiaries of the unemployment insurance scheme.

Modification of reimbursements (Erstattungen)

Introduction/modification of the reimbursement of the unemployment insurance scheme to the pension insurance for labour market-induced disability pensions

Indirect Shifts Relocation of benefits (Aufgabenverlagerung)

Relocation of the funding of follow-up unemployment assistance (1967) and language courses for resettlers (1989) from the federal budget to the unemployment insurance scheme; transfer of vocational rehabilitation to the unemployment insurance scheme (1977)

Modification of benefits’ regulation (Änderung des Leistungsrechts)

Extension of the period for receiving unemployment benefits leads to relief of the federal budget concerning the costs of unemployment assistance. Correspondingly, this burdens the budget of the unemployment insurance scheme by the costs of higher benefits

Modification of employees’ social insurance contributions (Sozialversicherungsbeiträge der Beschäftigten)

An increase in employees’ contributions to the unemployment insurance scheme relieves the federal government of its tax-based deficit liability towards the unemployment insurance scheme. An increase in employees’ contributions to the unemployment insurance scheme, combined with a simultaneous decrease in the contributions to the pension insurance scheme, uses resources of the unemployment insurance scheme for pension funding

Source: Own compilation.

This chapter is based on data gathered about direct and indirect shifts generated by welfare acts which were legislated between 1964 and 2000 (the year of announcement, not of enactment) and contains 100 cases (for details see: Trampusch 2003). Only those legal alterations for which the legislator specified the costs of the shift and which changed the law were included. Those 164

only prolonging changes made at an earlier point in time were not included. Table 2 provides an overview of the shifts. It contains only 83 cases because some laws brought more than one change in the same fiscal relationship and in the database such shifts were counted as one. Table 2 and Figure 1 show that in the historical development of the Verschiebebahnhofpolitik three different strategies of shifting burdens can be distinguished. Firstly, we find a strategy which can be called the consolidation policy (Konsolidierungspolitik) which was introduced in 1964 and has been implemented over the whole period. The policy of consolidation aims at removing the burden of fiscal costs from the federal budget by charging pension or unemployment insurance. Secondly, and in particular between 1977 and 1989, we also find a policy of shifting social insurance budgets (Belastungsverschiebungspolitik). This relocated fiscal costs between pension and unemployment insurance, and was invented in the wake of the fiscal crisis of pension insurance in 1977. The third strategy which can be discerned is called the re-consolidation policy (Re-Konsolidierungspolitik). This is signified by measures which remove a burden from pension and unemployment insurance by charging the federal budget. It was initiated in response to fiscal problems of the social insurance system, produced by German unification. Table 2: Historical development of the policy of shifting burdens Strategy Shifts Shifts Shifts Total 1964-1976 1977-1989 1990-2000 Remove the burden from 9 7 8 24 Consolidation policy federal budget – (Konsolidierungspolitik) charging pension insurance Remove the burden from (until 1976: stabilization 3 9 9 21 federal budget – charging policy) unemployment insurance scheme Remove the burden from 1 8 5 14 Policy of shifting social pension insurance – insurance budgets charging unemployment (Belastungsverschiebungsinsurance scheme politik) Remove the burden from 0 6 6 12 unemployment insurance scheme – charging pension insurance Remove the burden from 0 4 4 8 Re-consolidation policy pension insurance – (Rekonsolidierungscharging federal budget politik) Remove the burden from 0 0 4 4 unemployment insurance scheme – charging federal budget Total 13 34 36 81 Source: Own compilation; Note: The database contains 26 cases of Belastungsverschiebungspolitik, 45 cases of Konsolidierungspolitik, and 12 cases of Re-Konsolidierungspolitik. In the following section, these three policy modes are briefly explained.

165

Federal +

− +

ᬇ

+

ᬇ

ᬅ

+

ᬅ

− Unemployment insurance budget

− −

+ +

ᬆ

ᬆ

− −

Pension insurance budget

Figure 1: The three policies of shifting fiscal burdens Notes: ①: Consolidation policy; ②: Policy of shifting social insurance budgets; ③: Re-consolidation policy; →: Streams.

Table 2 indicates that the Verschiebebahnhofpolitik was implemented by different government coalitions (1964-1966: CDU/CSU-FDP; 1966: CDU/CSU-SPD; 1969-1982: SPD-FDP; 19821998: CDU/CSU-FDP; 1998-2000: SPD-Bündnis90/Die Grünen),5 which highlights the crossparty consensus. The following describes the three strategies of shifting fiscal burdens in more detail. 3.1

The policy of consolidation

In the first period of the policy of consolidation, namely until 1976, this strategy aimed at making the federal budget independent of the business cycle through the alteration of allowances, the relocation of benefits between budgets or the increase in social contributions of employees. After the experience of post-war economic growth, the government expected the increase to continue so that fiscal resources would feed the social insurance system; shunting fiscal costs from both pension and unemployment insurance to the federal budget should generate new and additional revenues which the government could use for an “active policy oriented towards the future” (quote from the Federal Government’s explanation of the Finanzplanungsgesetz of 1966, BT-Drks. V/1067, 12). The government intended to use the fresh money for economic stabilisation, and to this end the following acts were implemented: (i) Several important Budget Acts in 1964, 1965, and 1965 (Haushaltsgesetze 1964, 1965, 5

CDU is the Christlich Demokratische Union Deutschlands (Christian Democratic Party of Germany) and CSU is the Christlich-Soziale Union in Bayern (Christian Social Union of Bavaria), both are conservative parties. The SPD is the Sozialdemokratische Partei Deutschlands (Social Democratic Party of Germany) and the FDP is the Freie Demokratische Partei (Free Democratic Party) and is liberal. The party Bündnis90/Die Grünen (Alliance 90/The Greens) is a merging of the Greens (Die Grünen) and the party Bündnis90 (Alliance 90), which was founded in the new Bundesländer in the wake of the 1990 Volkskammer election in the German Democratic Republic. 166

1966), which diminished the Federal Allowance to pension insurance (Bundeszuschuss zur Rentenversicherung); (ii) the Financial Change Act (Finanzänderungsgesetz) of 1967, which raised the contribution to the pension insurance scheme; and (iii) the Financial Planning Act (Finanzplanungsgesetz) of 1966, which relocated the financing of unemployment assistance (Anschlussarbeitslosenhilfe) and of continuing vocational training (Berufliche Förderung) from the federal budget to the unemployment insurance fund. This – compared with later strategies – quite trivial policy of shifting costs away from the federal budget to the social insurance system lost its legitimacy when in 1976/1977 the pension insurance scheme slipped into a severe financial crisis and the Liberal Party (FDP) refused to agree on a further increase in the pension insurance contribution (Roth 1989, 162-163). The pension crisis and the Liberal Party’s refusal made the invention of a second strategy necessary, namely the policy of shifting social insurance budgets (Belastungsverschiebungspolitik), which moved fiscal costs between pension and unemployment insurance. 3.2

The policy of shifting social insurance budgets (Belastungsverschiebungspolitik)

This new strategy was aimed at re-assuring and strengthening the fiscal basis of the insurance system. On the one hand, these are direct shifts, namely the alteration of the social insurance contributions of beneficiaries (Sozialversicherungsbeiträge der Leistungsempfänger). On the other hand, the strategy also implies indirect ones, namely the simultaneous re-allocation of social contributions of employees to pension and unemployment insurance. This last measure means that there was a deliberate and explicit strategy to exchange the level of the social contributions in order to retain the social policy measures of the pension and unemployment insurance systems. As mentioned above, it was the pension crisis of 1976/1977, which catalysed this new kind of shifting burdens. Here, we have to note that this crisis was really severe because of the electoral and societal significance of German pension policy. In the campaign for the Federal Parliament’s Election of 1976 the pension crisis was a top issue. Political parties of all colours (as will be explained in the next section in more detail) agreed to cope with the fiscal problems of pension insurance by taking resources out of the unemployment insurance budget. In order to pump fresh money into the pension insurance budget, the 20th Pensions Adaptation Act (20. Rentenanpassungsgesetz) of 1977 introduced pension insurance contributions on behalf of beneficiaries which the unemployment (and health) insurance funds had to pay. The second sub-type, the simultaneous change in social insurance contributions to pension and unemployment insurance, had been regularly practised since the early 1980s (for details see Trampusch 2003, Table 6). This strategy also meant that the aggregated contribution rate of pension and unemployment insurance did not change very much between 1976 (21 per cent) and the beginning of 1991 (23 per cent) although the social insurance system coped with the 167

employment problems of the 1980s, the demographic challenges, and the tax reduction policy of the Finance Minister Gerhard Stoltenberg (1982-1989). With regard to the political motives behind this policy, the explanations of the bill given by the government clearly indicate that the shift of contribution rates aimed at sustaining the social insurance system as a whole (for more details see the next section). As an exchange of contribution rates between the different social insurance schemes helped to stabilize the aggregated contribution rate, it seems reasonable to argue that these fiscal movements also retained the legitimacy of the welfare state against liberal and employers’ criticism of high nonwage labour costs. 3.3

The policy of re-consolidation

From Table 1 it can be discerned that since the early 1990s the first two policy strategies have been supplemented by a third one, namely the policy of re-consolidation. Thus, besides shifting burdens from the federal budget to pension and unemployment insurance funds and unburdening pension insurance at the expense of the budget for unemployment insurance and vice versa, costs were also moved out of the social insurance funds to the federal budget. In detail, measures of the re-consolidation policy are more direct, such as the increase of federal allowances and federal reimbursements to pension and unemployment insurance, as well as indirect ones in the form of changes in the regulation of benefits such as tightening eligibility for benefits and shortening the duration of benefits. The increase of federal allowances to pension insurance was partially financed by tax increases (value added tax) or the introduction of a new ecological tax (Ökosteuer). In the following section, the in-depth analysis of representative welfare acts reveals the underlying politics of these various measures of Verschiebebahnhofpolitik. It is shown that the policy of shifting costs and burdens was maintained and supported by a cross-party and crossclass coalition and served a multitude of interests. This alliance also stabilized the state-society relationship and sustained the generosity of the German welfare state.

4

The politics of shifting expenditure burdens

The Verschiebebahnhofpolitik served a multitude of interests and was based on a cross-class and cross-party alliance between social politicians, employers, trade unions, works councils and public social policy administrators (see also Manow and Seils 2000, 265-266). The alliance was motivated by the social partners’ desire to transfer the costs of adjustment into public finance and away from wage bargaining, which coincided with the government’s interest in shifting the costs of industrial restructuring out of the federal budget and onto the social insurance funds (Manow and Seils 2000, 265-66). Strong biographical and structural links between politicians and social partners (Armingeon 1989, 17; Trampusch 2005a) gave the movements impetus and 168

reinforced not only cross-party consensus in social policy, but also the strong position of employers and trade unions in the making of social policy (Trampusch 2009, 87-103). This section investigates the politics of three welfare acts, which are representative of the policy of shifting burdens. Data are taken from an in-depth study of primary documents related to the bills (Gesetzentwürfe), the explanations by the government (Begründungen der Bundesregierung), the parliamentary debates, as well as the minutes of evidence of the Parliamentary Standing Committee for Labour and Social Policy (Ausschuss für Arbeit und Sozialordnung). The following acts have been analysed in more detail: (i) The 20th Pensions Adaptation Act (20. Rentenanpassungsgesetz) of 1977; (ii) the Act to Change the Rules in Employment Promotion and Public Pension Insurance (Gesetz zur Änderung der Vorschriften in der Arbeitsförderung und der gesetzlichen Rentenversicherung) of 1984; and (iii) the Act about the Change of the Contribution Rates to the Federal Employment Service and the Public Pension Insurance (Gesetz über die Änderung der Beitragssätze in der Bundesanstalt für Arbeit und in der gesetzlichen Rentenversicherung) of 1991. As the three acts are representative for the policy logics of the three different decades, the continuity of this collaboration between political parties and social partners can be demonstrated. 4.1

20th Pensions Adaptation Act of 1977

The 20. Rentenanpassungsgesetz of 1977 is the first welfare Act, which covered the fiscal needs of Germany’s post-war pension systems not by an increase in pension insurance contributions but by fresh money taken from the unemployment insurance scheme. It was enacted under the leadership of Labour Minister Herbert Ehrenberg (SPD, 1976-1982) and was supported not only by the SPD’s coalition partner, the FDP, but also by the unions, the employers, and the representatives of the pension insurance and unemployment insurance funds. As already mentioned, this Act is an outstanding example of the policy of shifting burdens between pension and unemployment insurance in order to retain the fiscal base of the insurance system and thus its capacity to internalize the costs of industrial restructuring. The charging of unemployment insurance was catalysed by a severe financial crisis in pension insurance to which the government could not react by increasing the pension insurance contribution rate because the FDP was opposed to a further rise (Roth 1989, 162-167). To resolve the stalemate, the reform introduced notional pension insurance contributions of beneficiaries actually paid by the unemployment (and health) insurance funds. It is significant that the government was fully aware of this trick. In the explanation of the bill, it argued that pension insurance “has to be made more independent from business cycles” and that the “social risks” should be distributed “more functionally among the social insurance institutions” (BT-Drks. 8/337, 83). This charging for unemployment insurance was also supported by the conservative parliamentary opposition (BT-Drks. 8/337, 83), which even suggested a heavier load for the unemployment insurance fund than the government (BT-Drks. 8/337, 83). 169

It is important to note that at that time the unemployment insurance system was in a comfortable fiscal condition. Its financial situation was so sound that a decrease in the unemployment insurance contribution would have been possible. By agreement with representatives of the unemployment insurance funds, however, the feasible decrease was restricted for the benefit of the pension insurance fund. In the parliamentary debate on the bill, the liberal Member of Parliament Schmidt (Kempten) argued that the unemployment insurance fund had to take care of the pension insurance contributions of unemployed persons in order to “attenuate the business cycle effects” on the pension insurance (Schmidt, FDP, 26. Sitzung des Deutschen Bundestages am 12.5.1977, 1867 [D]). The government estimated that in 1979 and 1980 unemployment insurance would be charged 4.84 billion Deutsche Mark (DM) by this measure, which would counterbalance pension insurance by the same amount (BT-Drks. 8/337, 6). We have to note that the 1977 Act not only introduced pension insurance contributions for the unemployment insurance fund, but also charged the unemployment insurance with a new benefit, whose financing was transferred from the pension to the unemployment insurance fund, namely, the measures of vocational rehabilitation (Berufliche Rehabilitation). The financial relief for the pension insurance (and likewise the charge for the unemployment insurance fund) was estimated at 1.7 billion DM for 1979 and 1980 (BT-Drks. 8/337, 6). Without doubt, this measure helped to sustain the legitimacy of the social insurance system at times when it became obvious that the pension insurance was undergoing a fundamental crisis and became politically and fiscally unsustainable. The main reason for the skewed position of the pension insurance at that time was increasing expenditures due to the early retirement system, which had been introduced by the pension reforms of 1957 and 1972, and was supported by a cross-party and cross-class coalition until the mid-1990s (Trampusch 2005a). It was in particular the employer’s association, the BDA, which highlighted these fiscal implications of Germany’s early exit regime (e.g. Dr. Doetsch (BDA), 5. Sitzung des Ausschusses für Arbeit und Sozialordnung am 23.3.1977, Wortprotokoll, 5/45-47). As the liberal governing coalition partner resisted any further increases in the pension contribution rate and the social partners used early retirement measures (financed by the pension insurances) to cope with industrial restructuring, the stream of fresh money from the unemployment to the pension insurance enjoyed widespread support. It is important to note that in 1977 the president of the chemical trade union, Hermann Rappe (SPD), chaired the Standing Committee of Labour and Social Policy. Furthermore, at the meetings of this Committee not only the members of parliament but also employers, unions, as well as the representatives of the social insurance schemes, expressed their strong commitment to the shifts between unemployment and pension insurance (for details see: 5. Sitzung des Ausschusses für Arbeit und Sozialordnung am 23.3.1977, Protokoll Nr. 5).

170

4.2

The Statutory Pension Insurance Act of 1984

The second typical Act, the Gesetz zur Änderung der Vorschriften in der Arbeitsförderung und der gesetzlichen Rentenversicherung of 1984, is doubly interesting for this study. First, the Act reduced the unemployment contribution rate from 4.6 to 4.4 percent and increased that of pension insurance from 18.5 to 18.7 (BT-Drks. 10/2176). It was estimated that this exchange of contributions rates increased the financial base of the pension system by 6.5 billion DM between 1985 and 1988 (BT-Drks. 10/2176, 7). As with all later exchanges in social contribution rates to pension and unemployment insurance, this exchange too was strongly supported by all political parties as well as social partners. In the parliamentary debate about the Act, Dieter-Julius Cronenberg, the leading social politician of the FDP at that time, argued that the simultaneous decrease and increase of the contribution premiums is an “enormous contribution” towards improving the financial situation of pension insurance without increasing the total contribution rate and he pointed out that it was strongly supported by the Liberal Party (Dieter-Julius Cronenberg, FDP, 108. Sitzung des Deutschen Bundestages am 6.12.1984, 8109 (D)). Besides this exchange policy, the 1984 Act secondly generated the fiscal preconditions for using the unemployment and pension insurance schemes to support early retirement policies. Since employment problems followed in the wake of the two oil crises across a range of industries, early retirement financed by pension and unemployment insurance became the preferred solution for managing large-scale layoffs. The prolongation of unemployment benefits for older workers (which was extended by legislation in 1985 and 1987) facilitated this early exit route and set strong incentives for older workers to use their pension benefits as early retirement benefits, an effect which employers and unions appreciated. This prolongation also relieved the federal budget of the costs of the means-tested unemployment assistance (Arbeitslosenhilfe), by shifting the costs from the federal budget to the unemployment insurance fund. This was an effect very much welcomed by the finance minister. The government estimated that the process cost the unemployment insurance fund 4.4 billion DM, but relieved the federal budget by 2.5 billion DM (BT-Drks. 10/2176, 7). The minutes of evidence on the 1977 act of the Parliamentary Standing Committee for Labour and Social Policy suggest that this cementation of the early exit strategy and the policy of shifting burdens were advocated by social partners and political actors. 4.3

The Act about the Change of the Contribution Rates to the Federal Employment Service and the Public Pension Insurance of 1991

The third typical Act by which the politics of shifting burdens can be illustrated is the Gesetz über die Änderung der Beitragssätze in der Bundesanstalt für Arbeit und in der gesetzlichen Rentenversicherung of 1991. This Act is an example of how pension insurance financed unemployment insurance. Like various Acts which were passed in 1981 and 1982 (see Table 3), the 1991 Act lowered the pension insurance contribution in order to “finance” an increase 171

in the unemployment insurance contribution. It raised the unemployment contribution rate from 4.3 to 6.8 (between 1.4.1991 and 31.12.1991; and then 6.3) percent and decreased that of pension insurance from 18.7 to 17.7 percent. The Act was passed in the wake of German unification and gave the unemployment insurance system enough room to expand further active labour market policies in East Germany in order to assist its transformation from a socialist to a capitalist economy. The minutes of evidence of the Parliamentary Standing Committee for Labour and Social Policy signify that this measure was strongly welcomed by social politicians, employers and trade unions. The government estimated that the 1991 act brought 41.3 billion DM into the unemployment insurance fund while the pension insurance funds were relieved of 19.2 billion (BR-Drks. 12/56, 3; these numbers were calculated for the years 1991 and 1992). In sum, our in-depth analysis of primary documents about the statements of political actors as well as those of employers’ associations and trade unions on legislation relating to shifting burdens suggests that this policy was supported by a cross-party and cross-class alliance. However, the consensual policies should not belie the tensions with which the German welfare state had already dealt by the end of the 1980s. Shifting burdens and welfare corporatism led to negative feedback effects because, despite shifting the burdens, the federal and social insurance budgets fell markedly. In addition, the early exit regime and German unification prompted the rise of non-wage labour costs. These negative feedback effects meant the German welfare corporatism was crumbling.

5

The breakdown of German welfare corporatism

This section shows that the policy of shifting burdens created negative and unintended feedback effects, which were ignored for some time but nevertheless accumulated and contributed to the erosion of the supporting coalition. The negative side effects set in motion tensions and conflicts and caused the breakdown of the formerly consensual coalition, and the tensions triggered the breakdown of German welfare corporatism. Until the beginning of 1991, the aggregate sum of social insurance contributions could be more or less restrained by the policies of shifting burdens, but the expansion of the labour market policy measures and the early exit in the wake of German unification (see also Manow and Seils 2000, 293) resulted in a top-up of the non-wage labour costs. By 1996, the aggregate sum exceeded the magic figure of 40 percent of gross wages (Trampusch 2009, 115) and between 1990 and 1998 alone the combined rate grew by as much as six and a half percentage points. It was estimated that about three percentage points were generated by German unification (Hinrichs 1998). Fiscal numbers indicate that the severe resource constraints of the federal budget were also quite real. As mentioned above, the social costs of German unification were largely financed by the social insurance budgets. As this caused a fiscal crisis in the social insurance system, its refunding since the mid-1990s through taxes and the federal budget became more and more important. For example, while in 1995 23 percent of pension benefits 172

paid were funded by federal subsidies from general taxes, by 2004 this share had risen to 37 per cent, or 77.4 billion Euros (Streeck 2007, 13). Without these public subsidies, the pension insurance contribution rate would have had to be 28.4 per cent instead of 19.5 per cent, producing a combined contribution rate of almost 50 per cent (Streeck 2007, 13). In order to avoid a collapse of the social insurance schemes and of non-wage labour costs, the government balanced the social budget by increasing the level of cross-subsidisation out of the federal budget. Streeck (2007), as well as Trampusch (2009) and Busemeyer and Trampusch (2013), argue that the rise of non-wage labour costs as well as the fiscal pressure on the federal budget resulted in three fundamental tensions which broke up the agreement between political parties and employers about the German social model: Firstly, unification resulted in “increasingly severe resource constraints” for public finance (Streeck 2007, 5) and dissolved the cross-party and cross-class welfare alliance. Secondly, strong tensions within the party system blew up as the liberal FDP and the economic wing of the CDU urged a fundamental change in social policy and demanded an end of the (mis)use of the pension system for early retirement (Trampusch 2009, 111-12). Thirdly, a divide developed within Germany’s employer associations as small firms and the Mittelstand (which allied with the liberal forces among political parties) were severely harmed by rising non-wage labour costs. Large firms’ early retirement practices massively contributed to this problem. A rift was opening between big firms and small firms. The smaller ones benefited less from the early retirement regime because they could not afford the voluntary redundancy payments (Mares 2003, 238). These tensions led to a revolt as small and medium-sized firms strongly opposed the traditional dominance of large companies. It was in the mid-1990s, when the welfare state had clearly become untenable due to the increasing non-wage labour costs and fiscal pressure and the tensions had become strong, that the cross-party and cross-class welfare consensus crumbled. Several major welfare state reforms and the reduction of social partners’ influence were the result. For example, in the mid-1990s, the rising costs of the early retirement regime caused employers and the liberal party to rebel. Facing increased pressure from small and mediumsized firms and his own liberal coalition partner, the FDP, Chancellor Kohl made early retirement reforms a government priority. The Federation of German Industry (Bundesverband der Deutschen Industrie) and the economic wing of the government coalition (a group of CDU/CSU members of the Bundestag based around Wolfgang Schäuble, together with the FDP) pressured the Chancellor to lower the aggregate social contribution rate through social policy reforms (Streeck 2003, 14-5). The result was the so-called “Action Program” of 1996, which was suggested by FDP Economics Minister Günter Rexrodt. This program resulted in comprehensive legislation that not only included the termination of the early retirement regime, but also encompassed the deregulation of the labour law through changes in the dismissal law and the cutting of sickness benefits, which trade unions perceived as part of their identity. In 173

1997, the trajectory of benefit cuts was continued by a pension reform which sought to cut down benefits significantly. Moreover, disability pensions were cut by actuarial deductions. Beginning with the 2001 pension reform, the Social Democratic Party, having won the 1998 Bundestag election, followed Kohl’s path of unilateral reform. The pension reform introduced private pension schemes and cut back public pensions. After this reform the new Chancellor Schröder restructured labour market policies and weakened social partners’ influences through the Hartz reforms and the reforms of the Agenda 2010. Schröder also enacted a fundamental restructuring of the self-administration, which contributed to a further decrease of trade union power (for an overview of these reforms see also Eichhorst and Kaiser 2006; Hassel and Schiller 2009; Trampusch 2009; Busemeyer and Trampusch 2013). At the same time the links between political parties and social partners fundamentally weakened as a profound elite change took place in the Bundestag’s Standing Committee for Labour and Social Policy (Trampusch 2005a).

6

Conclusions

This chapter investigated the state-society relation in the fiscal dimension of the German welfare state. It showed how the role of Germany’s social insurance system as a revenue raising institution was sustained and supported by a cross-class and cross-party alliance. This alliance advocated complex cost shifting between the federal budget and social insurance funds. Over time, however, the German welfare corporatism and thus the shifting of expenditure burdens lost legitimacy. Fundamental reforms and a breakdown of social partners’ influence and power, mainly those of trade unions, were the result. What are the broader lessons of this study for fiscal-centred welfare research at the state level? First, this study suggests that shifting burdens between different budgets may have shaped the policy as well as the politics of mature welfare states. Secondly, and connected to this insight, it seems reasonable to claim that programmatic and paradigmatic decisions which typify welfare regimes may likewise be driven by social policy as well as by fiscal objectives. A third lesson is that a fiscal-centred perspective contributes to a better understanding of endogenously generated welfare change because it helps to specify feedback-effects, which trigger transformative change (see also Trampusch 2009; Busemeyer and Trampusch 2013). Finally, and in line with previous studies on tripartite interactions between trade unions, employers and the state, as a mechanism to solve conflicts, (neo)corporatism is contingent and based on “cost-benefit considerations” (Armingeon 1986, 138). A fiscal-centred perspective clarifies how, due to negative feedback effects, these calculations may change over time and therewith also the stability of corporatism.

174

References Armingeon, Klaus. 1986. “Formation and Stability of Neo-corporatist Incomes Policies: A Comparative Analysis.” European Sociological Review 2 (2): 138-147. Armingeon, Klaus. 1989. “Trade Unions under Changing Conditions: The West German Experience 1950-1985.” European Sociological Review 5 (1): 1-23. Bonoli, Giuliano, and Philipp Trein. 2016. “Cost-Shifting in Multitiered Welfare States: Responding to Rising Welfare Caseloads in Germany and Switzerland.” Publius: The Journal of Federalism 46 (4): 596-622 Busemeyer, Marius R., and Christine Trampusch. 2013. “Liberalization by Exhaustion: Transformative Change in the German Welfare State and Vocational Training System.” Zeitschrift für Sozialreform 59 (3): 291–311. Deeg, Richard. 2001. “Institutional Change and the Uses and Limits of Path Dependency: The Case of German Finance.” MPIfG Discussion Paper 01 (6). Cologne: Max Planck Institute for the Study of Societies. Ebbinghaus, Bernhard. 2006. Reforming Early Retirement in Europe, Japan and the USA. Oxford: Oxford University Press. Eichhorst, Werner, and Lutz C. Kaiser. 2006. “The German Labor Market: Still Adjusting Badly?” IZA Discussion Paper No. 2215. Bonn: Institute for the Study of Labor. Gerring, John. 2007. Case Study Research. Principles and Practices. Cambridge: Cambridge University Press. Hall, Peter A., and Kathleen Thelen. 2009. “Institutional Change in Varieties of Capitalism.” Socio-Economic Review 7 (1): 7–34. Hall, Peter A., and Rosemary C. R. Taylor. 1996. “Political Science and the Three Institutionalisms.” MPIfG Discussion Paper 96 (6). Cologne: Max Planck Institute for the Study of Societies. Hassel, Anke, and Christof Schiller. 2009. “Bringing the State Back In: The Role of Fiscal Federalism for Welfare Restructuring.” 21st Annual Meeting of the Society for the Advancement of Socio-Economics, July 16–18, 2009. Paris: Sciences-Po. Hinrichs, Karl. 1998. “Reforming the Public Sector Pension Scheme in Germany: The End of the Traditional Consensus.” ZeS-Arbeitspapier No. 11/98. Bremen: Centre for Social Policy Research. Klose, Joachim, and Henner Schellschmidt. 2001. Finanzierung und Leistungen der Gesetzlichen Krankenversicherung: Einnahmen- und ausgabenbezogene Gestaltungsvorschläge im Überblick. WidOMaterialien Band 45. Bonn: Wissenschaftliches Institut der AOK. Mackscheidt, Klaus. 1990. “Finanzausgleichmaßnahmen zwischen dem Bund und den Trägern der Sozialen Sicherung und zwischen den Trägern der Sozialen Sicherung untereinander.“ In Finanzierungsprobleme der sozialen Sicherung, edited by Kurt Schmidt, 145–182. Berlin: Duncker & Humblot. Mahoney, James. 2010. “After KVV. The New Methodology of Qualitative Research.” World Politics 62 (1): 120–47. Mahoney, James, and Kathleen Thelen. 2010. “A Theory of Gradual Institutional Change.” In Explaining Institutional Change: Ambiguity, Agency, and Power, edited by James Mahoney and Kathleen Thelen, 1–37. New York: Cambridge University. Manow, Philip. 2001. “Welfare State Building and Coordinated Capitalism in Japan and Germany.” In The Origins of Non-Liberal Capitalism, edited by Wolfgang Streeck and Kozo Yamamura, 94–120. Ithaca, NY: Cornell University. Manow, Philip. 2013. “Coordinated Capitalism, Corporate Finance and the Pension System as a Source for Patient Capital: Germany and Japan Compared“. 20th International Conference of Europeanists, June 25-27, 2013. Amsterdam: Council for European Studies. Manow, Philip, and Eric Seils. 2000. “Adjusting Badly. The German Welfare State, Structural Change, and the Open Economy.” In Welfare and Work in the Open Economy. Diverse Responses to Common Challenges, edited by Fritz W. Scharpf and Vivien A. Schmidt, 264–307. Oxford: Oxford University Press. Mares, Isabela. 2003. The Politics of Social Risk: Business and Welfare State Development. Cambridge: Cambridge University Press. Nullmeier, Frank, and Friedbert W. Rüb. 1993. Die Transformation der Sozialpolitik: Vom Sozialstaat zum Sicherungsstaat. Frankfurt a.M.: Campus Verlag. Palier, Bruno, and Kathleen Thelen. 2010. “Institutionalizing Dualism: Complementarities and Change in France and Germany.” Politics & Society 38 (1): 119–48. Pierson, Paul. 2004. Politics in Time: History, Institutions, and Social Analysis. Princeton, Oxford: Princeton University Press. Roth, Richard. 1989. Rentenpolitik in der Bundesrepublik. Zum Verhältnis zwischen wirtschaftlicher Entwicklung und der Gestaltung eines sozialstaatlichen Teilbereichs 1957–1986. Marburg: Verlag Arbeiterbewegung und Gesellschaftswissenschaft. Seeleib-Kaiser, Martin. 2002. “A Dual Transformation of the Welfare State?” West European Politics 25 (4): 25– 48. 175

Streeck, Wolfgang. 2003. “From State Weakness as Strength to State Weakness as Weakness: Welfare Corporatism and the Private Use of the Public Interest.” MPIfG Discussion Paper 03 (2). Cologne: Max Planck Institute for the Study of Societies. Streeck, Wolfgang. 2007. “Endgame? The Fiscal Crisis of the German State.” MPIfG Discussion Paper 07 (7). Cologne: Max Planck Institute for the Study of Societies. Streeck, Wolfgang. 2009. Re-Forming Capitalism: Institutional Change in the German Political Economy. Oxford: Oxford University Press. Streeck, Wolfgang, and Kathleen Thelen. 2005. “Introduction: Institutional Change in Advanced Political Economies.” In Beyond continuity: Institutional Change in Advanced Political Economies, edited by Wolfgang Streeck and Kathleen Thelen, 1–39. Oxford: Oxford University Press. Streeck, Wolfgang, and Christine Trampusch. 2005. “Economic Reform and the Political Economy of the German Welfare State.” German Politics 14 (2): 174–195. Thelen, Kathleen. 2012. “Varieties of Capitalism: Trajectories of Liberalization and the New Politics of Social Solidarity.” Annual Review of Political Science 15: 137–159. Trampusch, Christine. 2003. “Ein Bündnis für die nachhaltige Finanzierung der Sozialversicherungssysteme: Interessenvermittlung in der bundesdeutschen Arbeitsmarkt-und Rentenpolitik.“ MPIfG Discussion Paper 03 (1). Cologne: Max Planck Institute for the Study of Societies. Trampusch, Christine. 2005a. “From Interest Groups to Parties. The Change in the Career Patterns of the Legislative Elite in German Social Policy.” German Politics 14 (1): 14–32. Trampusch, Christine. 2005b. “Institutional Resettlement: The Case of Early Retirement in Germany.” In Beyond Continuity: Institutional Change in Advanced Political Economies, edited by Wolfgang Streeck and Kathleen Thelen, 203–228. Oxford: Oxford University Press. Trampusch Christine. 2009. Der erschöpfte Sozialstaat. Transformation eines Politikfeldes. Frankfurt a.M.: Campus. Trampusch, Christine, and Bruno Palier. 2016. “Between X and Y: How Process Tracing Contributes to Opening the Black Box of Causality.” New Political Economy 21 (5): 437-454 Vail, Mark I. 2008. “From ‘Welfare Without Work’ to ‘Buttressed Liberalization’: The Shifting Dynamics of Labor Market Adjustment in France and Germany.” European Journal of Political Research 47 (3): 334–58. Weaver, Kent. 2010. “Paths and Forks or Chutes and Ladders? Negative Feedbacks and Policy Regime Change.” Journal of Public Policy 30 (2): 137–162.

176

Social acceleration and social investment Kees van Kersbergen1 and Barbara Vis2 Abstract: The starting point for our contribution on the (alleged) crisis of the European social model is David Weisstanner and Klaus Armingeon’s (2018) study on the cross-national variation in wage premiums to discuss the temporal dimension of social investment policy - futureoriented policy par excellence. Our guiding propositions are that: (1) “social acceleration”, that is, the accelerating pace of change, particularly but not exclusively in technology, tends to complicate and ultimately undermine the assumptions of social investment policy; (2) the result will be increasingly higher inequality. Social investment policies are future-oriented social policies, investing in human capital in the short term to yield a social return in the middle or long term (e.g., early childhood education). Social acceleration causes the temporal outlook of social investment politics to become more short-term. Simultaneously, it contracts the long-term governing capacity, making it increasingly difficult to determine precisely in what kind of skills to invest now, in order to maximize the chance of social return in the future.

1

Introduction

Throughout his impressive professional career, Klaus Armingeon has contributed significantly to a diversity of fields that all speak to the theme of this book: The (alleged) crisis of the European social model. The fields include welfare state politics and policies as well as the study of inequality (e.g., Armingeon 2007; Armingeon and Giger 2008; Armingeon and Schädel 2015), the Euro crisis and the politics of austerity (e.g., Armingeon 2014; Armingeon and Cranmer 2018; Armingeon, Guthmann, and Weisstanner 2016), institutional change and types of welfare states (e.g., Armingeon 2001; Armingeon, Bertozzi, and Bonoli 2004; Armingeon and Careja 2008), trust in and support for democracy (e.g., Armingeon and Guthmann 2014), and corporatism (negotiation democracy, consociationalism) (e.g., Armingeon 1986, 1997, 2002a, 2003b). Given our own field of expertise, we concentrate on some important new issues in the field of study on politics and policies of the welfare state and inequality. We indicate how these play a part in the crisis of the welfare state. In his recent work, co-authored with David Weisstanner (Weisstanner and Armingeon 2018), Klaus Armingeon has shown that economic forces such as globalization and technological change tend to increase the wage gap between the high-educated and low-educated labor force, 1 2

Aarhus Universitet [email protected] Universiteit Utrecht [email protected]

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 R. Careja et al. (eds.), The European Social Model under Pressure, https://doi.org/10.1007/978-3-658-27043-8_11

177

but that politics (determining levels of taxation and public spending) accounts for huge crossnational differences in inequality. This finding goes against the notion of a welfare state in crisis, broadly defined, because policies of the welfare state can still make a (positive) difference. In this paper, we follow up on this recent work, asking (1) whether the recent acceleration of, especially, technological change is likely to override the political counterforces Weisstanner and Armingeon identified, and (2) whether social acceleration is likely to further increase inequality. We argue that we expect the answer to both questions to be “yes.” If we are right, this would mean that the welfare state is in crisis because its policies can no longer make a difference. Of course, we fully realize - in Klaus Armingeon’s spirit - that only by collecting and analyzing the empirical evidence will we be able to tell truly whether this answer is correct or not. Such an empirical endeavor is, however, beyond the scope of this chapter and requires data that are not yet available.

2

Education premiums and redistributive policies: Explaining cross-national variation among developed democracies

Weisstanner and Armingeon (2018) examine why there is such large cross-national variation in the wage premium for higher education, that is, the degree to which the high-educated earn higher wages than the low-educated. From the economics literature, we know that the rise of economic inequality in the past decades can in no small part be contributed to skill-biased technological change (e.g., Acemoglu 2002; Goos, Manning, and Salomons 2014). Technological developments increase the relative demand for highly skilled labor in a globalized economy and, as a result, the price for that labor goes up. From a long tradition in comparative political science comes the well-established thesis that party politics, the type of policies parties promote in government, and the political-institutional setup for conflict mediation determine how equal or unequal the wage distribution and ultimately also the income distribution will be (see Jensen and Van Kersbergen 2017, especially chapters 7 and 10). So who is right: the economists or the political scientists? Typical for Klaus Armingeon’s work is the conviction that such controversies cannot be solved through endless theoretical disputes. Instead, empirical research needs to make the difference, and for empirical research, we need data. Therefore, the first step is to build a data set that includes information on the economic, political, policy, and institutional variables of theoretical interest. This is precisely what Weisstanner and Armingeon did: Compiling a new data set based on the Luxembourg Income Study surveys that allowed them to compare twenty-two OECD countries between 1989 and 2014. Their main findings can be summarized as “public policies matter”: (1) A large tax state, that is, high levels of taxation, and (2) high levels of public spending on education suppress the wage premium for higher education. “Hence, it is not only economic fate, but also political choice whether education premiums are modest and 178

stagnant, or whether they are large and increasing” (Weisstanner and Armingeon 2018, 16, emphasis added). As we mentioned in the introduction, the finding that the policies of the welfare state can still make a difference contradicts the notion of a welfare state in crisis, broadly defined. While Weisstanner and Armingeon do not systematically test explanations for the empirical pattern they identify, they list some plausible theoretical mechanisms. For one, they specify that “a large tax state reduces the material incentives and the attitudinal motivation of highly qualified workers to exploit their strong position in the labor market. [Since] a considerable part of additional income is taxed away” (Weisstanner and Armingeon 2018, 2). Moreover, they note (quoting Autor 2014, 850) that large public education expenditures “address inequality from two directions: (i) enabling a larger fraction of adults to attain high productivity, rewarding jobs, and a reasonable standard of living; and (ii) raising the total supply of skills available to the economy, which in turn moderates the skill premium and reduces inequality” (Weisstanner and Armingeon 2018, 2). In terms of the “who is right?”-question (the economists or the political scientists), the answer turns out to be “both.” The economists are right in stressing that there are general economic drivers, such as technological change and globalization, which put a wage premium on high education and penalize low education. They describe the general trend in the world economy very well. However, political scientists explain much better how higher levels of taxation and public education spending moderate the general trend. How economic developments, such as skill-biased technological change, play out in a country fundamentally depends on that country’s politics (that is, its political decision-making and public policies). That is a fundamental lesson of comparative political economy that Klaus Armingeon has helped draw from data collection efforts and subsequent careful empirical analyses. However, the lesson that politics matter might very well be a historical one that has perhaps been superseded – or will soon be – by more recent trends. First, the most recent technological changes greatly add to the already significantly increasing wage inequality and job polarization while certain advances in computerization are likely to radically restructure the labor market. These technology-driven changes challenge what traditional labor market institutions and governments are capable of doing, assuming they can adjust at all. Surely, in the past we have had doom scenarios spelled out on the damaging ramifications on the welfare state of rapid technological, demographic, and social change (e.g., Offe 1984), but labor market institutions and governments have demonstrated a remarkable and flexible adjustment capacity (e.g., Hemerijck 2013; Van Kersbergen and Vis 2014). Nevertheless, this time may be different to the extent that the major characteristic of contemporary technological advances (and other types of social and cultural changes) is their accelerated pace. It is not just that things change, but that changes occur at an increasing rate, captured by the term “social acceleration”. We argue that especially this latter phenomenon makes it increasingly difficult, if not impossible, for taxation and public 179

policies (or democratic problem solving more generally) to counteract speeding inequality-inducing and employment-destroying developments.

3

Artificial intelligence, robots and human skills

The empirically oriented economics literature, which Klaus Armingeon and David Weisstanner (2018) partly use and respond to, reports two major trends. The first trend concerns the already discussed increasing wage gap between high-educated and low-educated workers due to skillbiased technological change. The other, and more recent, trend adds to a “vanishing middle”trend in that jobs that had escaped the replacement of labor in earlier waves of mechanization and automation have been disappearing too. The reason is that intelligent computers and robots are increasingly capable of taking over also certain routine cognitive (rather than manual) tasks once thought to be the exclusive domain of humans (e.g., Avent 2016; Ford 2015; Kaplan 2015; Schwab 2016), such as repetitive customer service or information processing. The major effects of the trends are rising wage inequality and a U-shaped polarization in the labor market (Goos and Manning 2007). Many jobs in the low-skilled service sector are hard to replace and concern those occupations that involve “assisting or caring for others, for example, food service workers, security guards, janitors and gardeners, cleaners, home health aides, child care workers, hairdressers and beauticians, and recreation occupations” (Autor and Dorn 2013, 1555). Non-routine jobs at the high-end skill sector are hard to replace too, so highly educated professionals and managers, who perform “‘abstract’ creative, problem-solving, and coordination tasks” (Autor and Dorn 2013, 1559) can look forward to continued employment opportunities. The net result is job polarization – which for instance has characterized seventeen of the EU27 countries between 2000 and 2010 (Scarpetta et al. 2012) – and the hollowing out of the middle. The most recent literature is more speculative on the potential impact of further technological change, particularly automation and digitalization, but tends to predict further rises in wage inequality and the disappearance of employment and jobs. A well-known and widely discussed (and criticized) study (Frey and Osborne 2017), for instance, expects that “computerization” (e.g., machine learning, big data mining, mobile robotics, machine vision, and various other forms of artificial intelligence) will radically alter job prospects in various sectors. This includes occupations that – until recently – were widely believed not to be susceptible to the threat of machines, including legal writing and truck driving. The estimate is that in the next two decades almost 50 percent of total employment in the United States (US) will be at risk of being “computerized away.” Frey and Osborne (2017, 265) find that “most workers in transportation and logistics occupations, together with the bulk of office and administrative support workers, and labor in production occupations, are likely to be substituted by computer capital”. More surprisingly, they expect a large chunk of service jobs to disappear as well. Personal and household service robots are taking over jobs; such devices become smarter and more agile to deal with 180

non-routine tasks, interactive machines have already taken over many low-skilled jobs in sales (think of fully automated cashier systems in supermarkets), and advances in prefab techniques and 3D printing facilitate the computerization of construction work (Frey and Osborne 2017, 265–6). However, as said, technological change has always provoked anxiety over its negative consequences, especially with respect to the massive displacement of labor that it causes. As many point out (for an overview, see Mokyr et al. 2015; Ford 2015), such anxieties have proven well grounded when looking at the short-term disruptive effects but are exaggerated when taking into account the long-term benefits. For example, mechanization and other technological advances caused a massive loss of jobs in agriculture in most countries in the 1950s and 1960s, but this loss was more than offset by the rise of employment in other sectors (entirely new occupations, new sectors). Similarly, the remarkable growth of professional, personal, financial, and social service jobs more than compensated for the de-industrialization, that is, the loss of jobs in the industrial sector in the 1970s and 1980s. Computerization and digitalization since the 1990s destroyed many existing jobs, but also created many new ones. There is no reason to assume that the recent wave of technological innovation (including greatly improved artificial intelligence, machine learning, robotization, the platform economy and labor market, blockchain technology, etcetera) will be very different in its consequences: displacement of labor followed by growth of employment in previously unimagined occupations and sectors. So, why should this time be different? In his prize-winning book, The Rise of the Robots, Ford (2015) lists six well-researched and documented trends that indicate that this time might indeed be different. These include: 

stagnant wages since the 1970s in spite of huge increases in productivity;



a decreasing share of labor in national income and exploding corporate profits;



stagnating job creation and rising long-term unemployment;



increasing economic inequality;



skill-biased technological change accompanied by a decreasing education premium and underemployment of highly educated workers;



an increasing number of part-time jobs and further job polarization.

According to Ford (2015), more than globalization, financialization, and politics (e.g. deregulation, neo-liberalism) as well as technological advances, especially in information and communication technology, add up to an unprecedented disruptive force. Technological advances cause a change in the nature of work and the workplace. As Dobbs et al. (2016, 151) argue, “Technology is increasingly allowing employers to redesign and disaggregate work and to reassign routine tasks to lower-skill employees.” The workplace itself is disaggregated. Telecommunication technology makes flexible working from home easy and cheap while digital platforms entirely revolutionize employment relations and what it means to have a job. The OECD 181

(2016, chapter 4) has computed that 60 percent of employment growth since 1990 was in the form of “non-standard” work (temporary work, part-time work, independent contracting, or self-employment), whereas almost all losses in standard jobs occurred in middle-skill occupations. Non-standard work now represents over 46 percent of total employment in countries like the Netherlands and Switzerland. Even though the size of the so-called platform economy is still relatively small, the expectation is that its disruptive consequences will be formidable (Blix 2017, 79; Eurofound 2018). The number of registered users of the two biggest platforms that match professional service demand and supply, Upwork and Freelancer, grew from around 2.5 million in 2005 to over thirty-five million in 2015 (OECD 2016, 17).

4

Social acceleration: Is this time different?

One of the defining characteristics of recent technological change in the fields of computing, communication technologies, artificial intelligence, and robotics is not just that it is proceeding fast, but that it is occurring at an increasingly faster rate. As argued by Brynjolfsson and McAfee (2014, 7–8): “Computers and other digital advances are doing for mental power – the ability to use to understand and shape our environments – what the steam engine and its descendants did for muscle power.” Important as that may already be for the (potential) ramifications for taxation and public policies, the qualitatively different phenomenon of exponential change greatly adds to the problem, questioning to what extent political decision-makers have, in fact, the time at all to respond timely to its challenges. This explains why things may be very different this time around. In this sense, technological change is a key element of a broader phenomenon well captured in the social theory concept of “social acceleration”. Social acceleration refers to the progressively faster reduction of the amount of time necessary for a social activity or experience; broadly including production, consumption, communication, travel, jobs, relationships, etcetera (Scheuerman 2004, 9; Rosa 2015, 65). The speeding up of the rate of social change occurs in three dimensions: (1) technological acceleration, (2) the acceleration of social change, and (3) the acceleration of the pace of life. We discuss each dimension in turn. 4.1

Technological acceleration

The first dimension of the phenomenon of social acceleration is technological acceleration, that is, the speeded up and ever faster changes in transportation, information, communication, and production as well as the increasing rate of adoption of new technologies. The increasing speed in transportation, say from the first steamship and automobile to the modern jet aircraft, has caused a compression of space because experienced distances shrink when the speed of the means of travel accelerates. Similarly, information now literally travels at the speed of light 182

while also the amount of information that can be transmitted at an instance has multiplied immensely. Moreover, in interpersonal communication, “both asynchronous (i.e. through email or answering machine) and synchronous communicative interactions are possible at any time independent of the respective location of the conversation partners” (Rosa 2015, 73). Finally, also the production and distribution of goods and services have been accelerating. These three forms of technological acceleration merge and overlap because of digitization and virtualization. “The conventional transport of goods like recording media or books,” writes Rosa (2015, 74), “is replaced through digitization by a form of pure information transmission, and similarly material processes of production (for instance, the development of design or architectural models) can be transformed into information processing procedures through virtualization.” Again, key here is not that there is a technological change, but that there is an exponential innovation and an increasing rate of adoption of new technologies. Moore’s law famously exemplifies exponential innovation. It describes that – until recently – the processing power of computer chips has been doubling every eighteen months. To give an impression of what exponential growth in fact signifies: The number of transistors on integrated circuit chips was between 2,500 and 10,000 in the early 1970s; it was approaching a staggering twenty billion in 2017 (Rupp 2017). Such developments fuel rapid innovations in, for instance, big data analysis, cloud-computing services, wearable technology, the use of artificial intelligence and machine learning, and virtual reality applications. The increasing rate of adoption of technologies is exemplified by the ever-shorter time interval at which new technologies (and the applications that make use of these new technologies) are disseminated. For instance, it took thirty-eight years for the radio to reach fifty million users, thirteen years for the television, four years for the iPod, three years for the Internet, and just nine months for Twitter. Snapchat, introduced in 2011, has taken over from Twitter in terms of usage. Between 2014 and late 2018, the number of daily active Snapchat users grew from forty-six million to 191 million in 2018 (Statista 2019). Key here is the recognition that it is not just that things (such as computers) run quicker today, but equally important is that the rate of technological change as well as the pace of the adoption of new technologies increase. The mounting rate of adoption, in turn, accelerates further innovation (Dobbs et al. 2016, 41–3). 4.2

The acceleration of social change

The second dimension of the phenomenon of social acceleration is the acceleration of social change, that is, the gradually faster change in the forms and organization of production, sectors of production, division of labor, occupational patterns, and family structures. The acceleration of social change captures the experience of the increasing rate at which social structures, institutions, organizations, traditions, conventions, life styles, routines, and habits are destabilizing and are being replaced by new and different forms of social interaction. One major empirical development relevant for taxation and public policy concerns the shift in the tempo of social 183

change from an inter-generational to an intra-generational one. The social institutions of sexual relationships, for instance, have become less stable within one generation; both in terms of duration (rising divorce rates, remarriages, re-partnering, serial monogamy, etcetera) and form (cohabitation, religious and ethnic diversity, varying gender composition, single parenthood, etcetera). Similarly, the revolution of women’s roles (Esping-Andersen 2009) occurred within one generation. With regard to contemporary popular culture and media, an “intergenerational rift” has emerged. As Rosa (2015, 115) puts it: “Young and old people increasingly live isolated from each other in subworlds in the sense that they not only use different communications media but also play different games, frequent different locales, follow different TV and news programs, read distinct newspapers and magazines, prefer different styles of music in divergent milieus, wear distinct kinds of clothing that they buy in respective age-specific stores, and even eat distinct kinds of food and speak different languages.” A prominent example of the acceleration of social change concerns the increase in the speed and number of times workers and employees change jobs over the course of their lifetime and the retraining requirements that come along with this. At an accelerating rate of change, it seems very difficult to have public education policy follow. In theory, the lifelong learning approach (see e.g., Kaplan 2016 for a literature review) could be an answer to a more dynamic labor market. There is, however, likely substantial variation in the degree to which individuals are able to obtain the capacities that a lifelong learning approach requires, namely having “communication skills in both native and foreign languages, use technology effectively, be open to learn, be social, be active, be entrepreneurial and (…) have cultural awareness” (Kaplan 2016, 47). Another example of the acceleration of social change is the “constancy” of organizational change and the rapid replacement of familiar occupations by new ones. A major consequence of the acceleration of social change concerns rising insecurity about what today’s society will look like tomorrow. Social and political expectations and ambitions constantly need recalibration under conditions of accelerating social change, upsetting social integration and cultural reproduction, and leading to disorientation in social, cultural, and political life (e.g. Chesneaux 2000; Hassan 2009; Scheuerman 2004; Wajcman 2008). It can be expected that this has major ramifications, for instance for defining what the (future) tax bases must be and for the decision on what to focus social investment policies. 4.3

The acceleration of the pace of life

The third dimension of the phenomenon of social acceleration is the acceleration of the pace of life, that is, the “objectively measurable intensification of experiences and activities that we engage in during a given period of time” (Scheuerman 2004, 13). The speed of social activities increases objectively: We do things progressively faster, do more things continuously with fewer and fewer breaks, we do increasingly more things at the same time, and we connect with ever more others (via email, social media, the internet of things) all the time; professionally, 184

socially, and privately. Subjectively, this leads to a qualitative change in the experience of time and may express itself as the feeling of having no time for oneself, lagging behind, being behind the times. Being pressed for time all the time, in turn, can cause a fear of missing out, produce a compulsion to adapt (Nowotny 1989; Bertman 1998; Eriksen 2001; St. Clair 2011; Wajcman 2015; Rosa 2015), and cause a (or at least add to the already existing) work–life imbalance (e.g., Schöneck 2018). Rosa (2015, 134) identified the “slipping slope syndrome”; the existential feeling of standing on slipping slopes caused by the realization that “in a dynamic society almost all of one’s stock of knowledge and property is constantly threatened with obsolescence.” This syndrome reinforces the feeling of lagging behind and the compulsion to adapt, especially by even further increasing connectivity (see Korunka and Hoonakker 2014; Wajcman 2015).

5

Implications for democracy, problem-solving and policymaking

Against this backdrop, the question is whether public policies in advanced democracies can still make a difference as Armingeon and Weisstanner (2018) showed they could at least up until 2014. Generally speaking, governments’ room to introduce or expand policy measures is increasingly limited because a lot of spending is already pre-committed. In 2009, in the US, “every dollar of revenue was pre-committed before the new Congress walked through the doors of the Capitol” (Steuerle 2016, nn). The so-called fiscal democracy index, which measures how much of current revenues remain after the pre-committed spending, ranges for the US from >60 percent in the 1960s – leaving ample room of political decision-making – to minus 5 to 10 percent in 2009. Currently, in 2018, it is approximately 20 percent. The US is not the only country where the room to make political choices is rather limited (low fiscal democracy). Streeck and Mertens (2010, 10) show that the decline of fiscal democracy is also visible in Germany with the fiscal democracy index dropping from >70 percent in 1970 (including defense spending) to around 20 percent in 2009. In terms of problem solving, the effects of technological changes, such as robotization, are certainly on the radar of policymakers. To take the Netherlands as an example: between 2015 and 2018, formal and informal advisory bodies to the Dutch government and policymakers published four reports on robotization (SER 2016; Went, Kremer, and Knottnerus 2015; van Est and Kool 2015; Freese et al. 2018). These reports provide policymakers with useful information on robotization and its (possible) effects by, for instance, providing diverse scenarios of what working in a robot society would mean (e.g., Went, Kremer, and Knottnerus 2015). However, all reports also conclude that robotization is a process with uncertain outcomes and many variables that are not only unknown, but oftentimes also unknowable. Such a context severely inhibits the possibilities for effective policymaking. We turn to the case of social investment policies to illustrate our point. 185

5.1

Implications for social investment

Let us zoom in on one specific type of policy – social investment – and examine how social acceleration influences it. Our proposition is that social acceleration tends to complicate and ultimately even undermine the assumptions upon which social investment policy rests. Social investment refers to – and should be restricted to – that cluster of future-oriented social policies that are - literally - investments in human capital in the short term that are meant to yield a social return in the middle or long term. Social investment policies include early childhood education and care, life-long training (on the job), re-training, and specific active labor market policies. They are the opposite of the classical, present-oriented and compensatory policies, such as sickness benefits or passive labor market policies (disability pensions, unemployment benefits). Social investment policies are quintessentially the problem-solving tools for governments that govern for the future (Jacobs 2011, 2016). In this quality lies the justification for singling them out for further illustration in this chapter. Social investment is a sub-set of what Jacobs (2016, 435) calls policy investments, that is, policies that “translate a given amount of short-run welfare into greater long-run welfare.” By and large, social investment policies have been adopted by democratic governments across the world since especially the early 1990s (Garritzmann et al. 2016). But there is a lot of variation across countries and over time in the kind of policies that are adopted, how generous these policies are, when they were introduced, and whether and when they expanded or retrenched (Garritzmann et al. 2016). The thesis is that there is a heightening of the already existing temporal tension between, on the one hand, the demands of short-term and myopic politics of the electoral and budget cycle and, on the other hand, the need for long-term governing and broad-minded policy such as social investment. Given the increasing speed of economic and social change, the most likely outcome is that the temporal horizon of politics becomes even more short-term. At the same time, the long-term governing orientation of political system contracts is jeopardizing particularly social investment as it becomes increasingly difficult to determine precisely in what (e.g., what kind of skills) to invest today so as to maximize the chance on social return in the future. In addition, the political system tends to respond to social acceleration by fast-tracking social policy-decisions that, in turn, themselves may become obsolete and overhauled at an increasing speed. Paradoxically, social acceleration, on the one hand, requires more social investment but, on the other hand, it actually delimits the political opportunities to promote it because it encourages short-termism. The predicament is that social acceleration shortens both the temporal outlook of policymakers and citizens: Both become (more) myopic. Why is democratic politics oriented to the short term? Jacobs (2016, 438) discusses three factors that contribute to short-termism: (1) Voters know much more about today’s policy consequences than they do about those in the longer term, and they typically face informational obstacles to make prospective (rather than retrospective) judgments. What is more, political elites’ attention too focuses on current problems: 186

“Problems whose consequences have not yet emerged are less likely to emit attention-generating signals and are thus at a disadvantage in the competition for elite cognitive investment” (Jacobs 2016, 439). Moreover, the future is – per definition – more uncertain than the present. (2) Political commitments are fragile, especially in the long run. Finally, (3) a policy investment may mean the imposition of costs in an organized group. Leaving aside the extent to which policymakers actually have room to make decisions because of declining fiscal democracy (see section 5.1), citizens also realize that the high degree of uncertainty about future developments that policymakers face makes it very difficult, if not impossible, to decide which social investments to make, in order to obtain the highest social return in the future. Jacobs and Matthews (2017) show by means of several experiments that it is precisely this uncertainty that makes citizens reject social investments. Political uncertainty thus emerges as an important driver of citizen judgements of policies that impose current costs in exchange for the promise of future benefits. The findings suggest that a good deal of citizen opposition to costly public undertakings arises not from intrinsic hostility to taxation or indifference to public goods provision, but from the intertemporal and delegative character of these policy ventures; from a lack of confidence that public officials, having imposed policy costs, will later deliver the public goods that they have promised (Jacobs and Matthews 2017, 205). Governing for the long term in the area of social policy has always been problematic, but apparently it has also been possible as the case of pension policies proves (see Anderson et al. 2009; Jacobs 2011). From current research, we know very well how and under which conditions political actors have been able to overcome the temporal, political, and institutional obstacles for devising policies for the long term. Shortly summarized, long-term oriented social institutions and policies need to be organized such that they can “(a) enhance the relative quality of information about long-run consequences, (b) stabilize political commitments over time, and (c) minimize distributive opportunism by organized groups” (Jacobs 2016, 442). However, because of social acceleration, this approach needs to be turned on its head. Social acceleration (a) decreases the relative quality of information about long-run consequences, (b) destabilizes political commitments over time, and (c) maximizes distributive opportunism by organized groups. Elaborating on this, we hold that from a temporal perspective, we need to take into account the following time issues: a) “Frozen time.” Social policy is fraught with vast historical legacies and path dependencies, and the fiscal room necessary for social investment is extremely small and shrinking, given pre-existing “mandatory” social spending commitments from past policy agreements (see section 5.1). In other words, even though social acceleration requires faster responses to keep up with increasing inequality, the political and institutional capacity to do so is limited or even shrinking, hence adding to the crisis of the welfare state. 187

b) Timing. What Bonoli (2007) has convincingly shown to be the case regarding the timing of new social risk policies, a fortiori holds for social investment under conditions of social acceleration. Only a few countries, if any at all, are in the luxury position of having the perfect Goldilocks conditions that will allow them to redirect social policy efforts toward social investment in a period of budget troubles, adding another element of the predicament of the welfare state. c) Life course uncertainty. Social acceleration undermines the type of life-course perspective on which social investment policies are necessarily vested (Kvist 2015). Under such conditions, effective problem solving via social investment requires rapid amendments, but the problem of “frozen time” and the intricate issue of the timing of policy change make such adaptation of the welfare state extremely difficult, if not unlikely.

6

Conclusion

In this contribution, we asked (1) whether the recent acceleration of, especially, technological change is likely to override the political counterforces that have protected the welfare state until now, and (2) whether social acceleration is likely to further increase inequality. We have argued that while the thesis “public policies matter” was true for the period Weisstanner and Armingeon examined – 1989-2014 – we think this time is different. Consequently, we answer both questions with “yes.” This implies that – different from the period that Weisstanner and Armingeon studied – welfare state policies will lose effectiveness, and the European social model may soon truly be in crisis. For public policies to matter, there should be (enough) space to make political choices. This space has already been shrinking substantially since the 1970s, but it will continue to shrink exponentially because of the process of social acceleration. In particular, exponential technological development and adoption make it ever more difficult to make political choices that reduce inequality, for instance via clever social investment policies. We expect it to become more difficult to counteract reactively (or proactively) and effectively to inequality-inducing and employment-destroying developments of public policies and taxation when the future is not only unknown, but also increasingly unknowable. We illustrated our proposition regarding social acceleration by discussing how this process complicates and ultimately possibly even undermines the assumptions upon which social investment policy rests. Social acceleration heightens the short-term focus of public policymaking while seriously restraining long-term policymaking, thus reducing the extent to which social investment policies will be effective or even at all possible. Normatively speaking, we would like to conclude by stating that we hope that further empirical analyses will prove us wrong. We hope that our rather dismal expectations regarding the coming true crisis of the European social model, and welfare states outside Europe for that 188

matter, are the result of our lack of imagination or an unwarranted lack of trust in the ingenuity of policymakers and the adaptive capacity of the welfare state.

References Acemoglu, Daron. 2002. “Technical Change, Inequality, and the Labor Market.” Journal of Economic Literature 40 (1): 7–72. Anderson, Karen M., Ellen M. Immergut, and Isabelle Schulze. 2009. The Handbook of West European Pension Politics. Oxford: Oxford University Press. Armingeon, Klaus. 1986. “Formation and Stability of Neo-Corporatist Incomes Policies.” European Sociological Review 2 (2): 138–47. Armingeon, Klaus. 1997. “Swiss Corporatism in Comparative Perspective.” West European Politics 20 (4): 164– 79. Armingeon, Klaus. 2001. “Institutionalising the Swiss Welfare State.” West European Politics 24 (2): 145–68. Armingeon, Klaus. 2002a. “The Effects of Negotiation Democracy: A Comparative Analysis.” European Journal of Political Research 41 (1): 81–105. Armingeon, Klaus. 2002b. “Consociationalism and Economic Performance in Switzerland 1968–1998: The Conditions of Muddling through Successfully.” In “Consociationalism and Corporatism in Western Europe,” edited by Steiner, Jürg, and Tom Ertman Acta Politica 37 (1/2): 121–38. Armingeon, Klaus. 2007. “Active Labour Market Policy, International Organizations and Domestic Politics.” Journal of European Public Policy 14 (6): 905–32. Armingeon, Klaus. 2014. “The Politics of Fiscal Responses to the Economic Crisis, 2008–2009.” Governance 25 (4): 543–65. Armingeon, Klaus, Fabio Bertozzi, and Giuliano Bonoli. 2004. “Swiss Worlds of Welfare.” West European Politics 27 (1): 20–44. Armingeon, Klaus, and Romana Careja. 2008. “Institutional Change and Stability in Post-Communist Countries, 1990–2002.” European Journal of Political Research 47 (4): 411–35. Armingeon, Klaus, and Skyler Cranmer. 2018. “Position-Taking in the Euro Crisis.” Journal of European Public Policy 25 (4): 546–66. Armingeon, Klaus and Nathalie Giger. 2008. “Conditional Punishment: A Comparative Analysis of the Electoral Consequences of Welfare State Retrenchment in OECD Nations, 1980–2003.” West European Politics 31 (3): 558–580. Armingeon, Klaus, and Kai Guthmann. 2014. “Democracy in Crisis? The Declining Support for National Democracy in European Countries, 2007–2011.” European Journal of Political Research 53 (3): 423–42. Armingeon, Klaus, Kai Guthmann, and David Weisstanner. 2016. “Choosing the Path of Austerity: How Parties and Policy Coalitions Influence Welfare State Retrenchment in Periods of Fiscal Consolidation.” West European Politics 39 (4): 628–47. Armingeon, Klaus, and Lisa Schädel. 2015. “Social Inequality in Political Participation: The Dark Sides of Individualisation.” West European Politics 38 (1): 1–27. Autor, David H. 2014. “Skills, Education, and the Rise of Earnings Inequality among the ‘Other 99 Percent’.” Science 344 (6186): 843–51. Autor, David H., and David Dorn. 2013. “The Growth of Low-Skill Service Jobs and the Polarization of the US Labor Market.” American Economic Review 103 (5): 1553–97 Avent, Ryan. 2016. The Wealth of Humans. Work and its Absence in the Twenty-First Century. London: Allen Lane. Bertman, Stephen. 1998. Hyperculture: The Human Cost of Speed. London: Praeger. Blix, Mårten. 2017. Digitalization, Immigration and the Welfare State. Cheltenham: Edward Elgar. Bonoli, Giuliano. 2007. “Time Matters: Postindustrialization, New Social Risks, and Welfare State Adaptation in Advanced Industrial Democracies.” Comparative Political Studies 40 (5): 495–520. Brynjolfsson, Erik, and Andrew McAfee. 2014. The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies. New York: Norton. Chesneaux, Jean. 2000. “Speed and Democracy: An Uneasy Dialogue.” Social Science Information 39 (3): 407– 20. Dobbs, Richard, James Manyika, and Jonathan Woetzel. 2016. No Ordinary Disruption: The Four Global Forces Breaking All the Trends. New York: Public Affairs. Eriksen, Thomas H. 2001. Tyranny of the Moment. Fast and Slow Time in the Information Age. London: Pluto. Esping-Andersen, Gøsta. 2009. Incomplete Revolution: Adapting Welfare states to Women’s New Roles. Cambridge: Polity. 189

Eurofound. 2018. Automation, Digitalisation and Platforms: Implications for Work and Employment. Luxembourg: Publications Office of the European Union. Ford, Martin. 2015. Rise of the Robots: Technology and the Threat of a Jobless Future. London: Oneworld Publications. Freese, Charissa, Ronald Dekker, Linda Kool, Fabian Dekker, and Rinie Val Est. 2018. Robotisering En Automatisering Op de Werkvloer Bedrijfskeuzes Bij Technologische Innovaties. The Hague: Rathenau Institute. https://www.rathenau.nl/sites/default/files/2018-05/Robotisering en automatisering op de werkvloer.pdf. Frey, Carl Benedikt, and Michael A. Osborne. 2017. “The Future of Employment: How Susceptible Are Jobs to Computerisation?”. Technological Forecasting and Social Change 114 (January): 254-80. Garritzmann, Julian, Silja Häusermann, Bruno Palier, and Christine Zollinger. 2016. “WoPSI - the World Politics of Social Investment: A Collaborative Research Project to Explain Variance in Social Investment Agendas and Social Investment Reforms across Countries and World Regions.” Project Background Paper. Goos, Maarten, and Alan Manning. 2007. “Lousy and Lovely Jobs: The Rising Polarization of Work in Britain.” The Review of Economics and Statistics 89 (1): 118-33. Goos, Maarten, Alan Manning, and Anna Salomons. 2014. “Explaining Job Polarization: Routine-Biased Technological Change and Offshoring.” American Economic Review 104 (8): 2509–26. Hassan, Robert. 2009. Empires of Speed. Time and the Acceleration of Politics and Society. Leiden: Brill. Hemerijck, Anton. 2013. Changing Welfare States. Oxford: Oxford University Press. Jacobs, Alan M. 2011. Governing for the Long Term: Democracy and the Politics of Investment. New York: Cambridge University Press. Jacobs, Alan M. 2016. “Policy Making for the Long Term in Advanced Democracies.” Annual Review of Political Science 19: 433–54. Jacobs, Alan M., and J. Scott Matthews. 2017. “Policy Attitudes in Institutional Context: Rules, Uncertainty, and the Mass Politics of Public Investment.” American Journal of Political Science 61 (1): 194–207. Jensen, Carsten, and Kees van Kersbergen. 2017. The Politics of Inequality. Houndmills: Palgrave Macmillan. Kaplan, Jerry. 2015. Human Need Not Apply. A Guide to Wealth and Work in the Age of Artificial Intelligence. New Haven and London: Yale University Press. Kaplan, Alan. 2016. “Lifelong Learning: Conclusions from a Literature Review.” International Online Journal of Primary Education 5 (2): 43–50. Korunka, Christian, and Peter Hoonakker. 2014. The Impact of ICT on Quality of Working Life. Dordrecht: Springer. Kvist, Jon. 2015. “A Framework for Social Investment Strategies: Integrating Generational, Life Course and Gender Perspectives in the EU Social Investment Strategy.” Comparative European Politics 13 (1): 131–49. Mokyr, Joel, Chris Vickers, and Nicolas L. Ziebarth. 2015. “The History of Technological Anxiety and the Future of Economic Growth: Is This Time Different?”,The Journal of Economic Perspectives 29 (3): 31–50. Nowotny, Helga. 1989. Eigenzeit. Entstehung und Strukturierung eines Zeitgefühls. Frankfurt am Main: Suhrkamp. OECD. 2016. “New Forms of Work in the Digital Economy.” OECD Digital Economy Papers 260. Paris: OECD Publishing. Offe, Claus. 1984. Contradictions of the Welfare State. London: Hutchinson & Co. Rosa, Hartmut. 2015. Social Acceleration. A New Theory of Modernity. New York: Columbia University Press. Rupp, Karl. 2017. “42 Years of Microprocessor Trend Data” Accessed September 27, 2017. https://www.karlrupp.net/2018/02/42-years-of-microprocessor-trend-data/ Scarpetta, Stefano, Anne Sonnet, Ilias Livanos, Imanol Núñez, W. Craig Riddell, Xueda Song, and Ilaria Maselli. 2012. “Challenges Facing European Labour Markets: Is a Skill Upgrade the Appropriate Instrument?” Intereconomics 47 (1): 4–30. Scheuerman, William E. 2004. Liberal Democracy and the Social Acceleration of Time. Baltimore and London: The John Hopkins University Press. Schöneck, Nadine M. 2018. “Europeans’ Work and Life – Out of Balance? An Empirical Test of Assumptions from the ‘Acceleration Debate’”. Time & Society 27 (1): 3–39. Schwab, Klaus. 2016. The Fourth Industrial Revolution. Cologny/Geneva: World Economic Forum. SER. 2016. Mens En Technologie: Samen Aan Het Werk. The Hague: Sociaal-economische Raad. https://www.ser.nl/~/media/db_adviezen/2010_2019/2016/mens-technologie.ashx. Statista. 2019. “Number of daily active Snapchat users from 1st quarter 2014 to 4th quarter 2018 (in millions)” Accessed September 21, 2018. https://www.statista.com/statistics/545967/snapchat-app-dau/ St. Clair, Michael. 2011. So Much, So Fast, So Little Time. Coming to Terms with Rapid Change and its Consequences. Santa Barbara, CA: Praeger. Steuerle, C. Eugene. 2016. “The Federal Government on Autopilot: Mandatory Spending and the Entitlement Crisis.” 2016.https://www.taxpolicycenter.org/taxvox/federal-government-autopilot-mandatory-spendingand-entitlement-crisis. 190

Streeck, Wolfgang, and Daniel Mertens. 2010. “An Index of Fiscal Democracy.” MPIfG Working Paper 10/3. Van Est, Rinie and Linda Kool. 2015. Werken Aan de Robotsamenleving: Visies En Inzichten Uit de Wetenschap over de Relatie Technologie En Werkgelegenheid. The Hague: Rathenau Institute. Van Kersbergen, Kees, and Barbara Vis. 2014. Comparative Welfare State Politics: Development Opportunities, and Reform. Cambridge: Cambridge University Press. Wajcman, Judy. 2008. “Life in the Fast Lane? Towards a Sociology of Technology and Time.” British Journal of Sociology 59 (1): 59–77. Wajcman, Judy. 2015. Pressed for Time: The Acceleration of Life in Digital Capitalism. Chicago and London: Chicago University Press. Weisstanner, David, and Klaus Armingeon. 2018. “How Redistributive Policies Reduce Market Inequality: Education Premiums in 22 OECD Countries.” Socio-Economic Review. Early view. https://doi.org/10.1093/ser/mwy018. Went, Robert, Monique Kremer, and André Knottnerus. 2015. De Robot de Baas: De Toekomst van Werk in Het Tweede Machinetijdperk. Amsterdam: Amsterdam University Press.

191

Social investment, active labour market policies and migration Giuliano Bonoli1

Abstract: Over the last few years, social investment has been at the centre of the debate on how to deal with the crisis of the European social model. The notion of social investment is based on the idea that social problems can be dealt with and inequality can be counteracted by investing in the human capital of disadvantaged individuals and groups. Against this background, the objective of this contribution is to assess the adequacy of the social investment strategy in facing the challenges posed by a multicultural society, by looking at the example of active labour market policies. Relying on meta-analytical research, it shows that access biases to policy measures as well as labour market discrimination severely limit the potential of a social investment strategy to deal with the problem of mainlining social cohesion in increasingly diverse societies. The chapter concludes by arguing that social investment and particularly active labour market policies must be adapted to the emerging European multiethnic societies, otherwise they will fail to deliver on the promise of maintaining social cohesion.

1

Introduction

In the search for solutions to the crisis of the European social model, a lot of attention has focused on the notion of social investment as a strategy for reform.2 The basic idea behind the social investment strategy is that social problems can be dealt with, and inequality can be counteracted by investing in the human capital of (potentially) disadvantaged individuals and groups. Key policies in a social investment strategy are education, vocational training, childcare policy, and in particular, active labour market policies. The promise of social investment as a reform strategy is the ability to generate levels of social cohesion and equality that are similar to those achieved by the transfer heavy post-war welfare states, in a way that is compatible with the current demographic and economic context. Investment in human capital leads to higher levels of productivity. In addition, the strategy aims to increase employment rates, particularly of women and older workers. This results in a winwin situation in which social problems and inequality are fought with tools that favour, not hinder, wealth creation.

1

Université de Lausanne [email protected]

2

This chapter draws on research published in Bonoli (2013) and Bonoli and Liechti (2018).

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 R. Careja et al. (eds.), The European Social Model under Pressure, https://doi.org/10.1007/978-3-658-27043-8_12

193

Social investment is particularly promising in relation to migration and in the context of the emerging multicultural societies in Europe. In the absence of sizable investment in the human capital of (low skilled) migrants, these face the prospect of permanent exclusion form the labour market and from society. Preserving social cohesion in an increasingly multi-ethnic society requires massive investments in the human capital of the newcomers. The alternative is segmented (dualized) societies where life experiences and chances differ dramatically depending on place of birth. On paper, the social investment strategy is appealing like few other options. One crucial question, though, is whether it will be able to deliver on its promises. While the jury is arguably still out, some analyses suggest that the potential impact of a social investment strategy to fight poverty and contain the rise in inequality is limited. Social investment policies tend to benefit the middle classes and have difficulties reaching the most disadvantaged (Cantillon 2011; Bonoli et al. 2017). As a result, a social investment strategy is likely to underperform the traditional transfer-based welfare states in terms of promoting social cohesion and fighting inequality. Recent research on who has access to typical social investment polices indeed suggests that access biases could be a key weakness in this approach. With reference to St. Matthew’s gospel, these biases are typically referred to as “Matthew effects”, a term which describes social policy interventions that benefit the less disadvantaged often contrary to intentions.3 Studies on who uses childcare for example, have shown that access to this key service has a strong upward social bias. Children in formal group-based childcare come mostly from middle and upper class families (see e.g. van Lancker 2013; Abrassart and Bonoli 2014). There is less research on access biases in other typical social investment policies, but what we do have suggests that there too may be Matthew effects at play. Even in the case of active labour market policies, which are by definition targeting disadvantaged jobseekers, one can see some segmentation and the partial exclusion of the most disadvantaged (Heckmann and Smith 2004; Bonoli and Liechti 2018). These limitations are very relevant for the problem dealt with in this chapter: the suitability of the social investment strategy to a multicultural society. Often, the access biases mentioned above overlap with ethnic based stratification systems. Low skilled immigrants from culturally distant countries are those who are most likely to be excluded from the main social investment policies. Yet, these are the people for whom there is a crying need for human capital investment. There is a further complication to be considered. The social investment strategy assumes that once the investment in the human capital of the disadvantaged has been successfully achieved, these will find good quality jobs, adequate with regard to the skill level they have acquired. The problem is that, in a multicultural setting, at all skill levels, ethnic minorities suffer from substantial levels of labour market discrimination. In short, social investment policies may well 3

From St Matthews gospel, with reference to the following verse: “For unto every one that hath shall be given, and he shall have abundance: but from him that hath not shall be taken away even that which he hath.” 194

succeed in improving the human capital of disadvantaged individuals and groups, but this will be pointless if labour market discrimination prevents them from finding employment. Labour market discrimination against immigrants is a very serious problem for a social investment strategy in general, and for active labour market policies (ALMPs) in particular. To realise this, one needs simply to compare the (positive) effects produced by the best active labour market policies and the (negative) effect of discrimination. The relatively large literature that is available in relation to both these effects suggests that they are of comparable size. If anything, labour market discrimination generates stronger effects than active labour market polices (see below for details). In other words, the benefit produced by good ALMPs could be totally undone by labour market discrimination. Against this background, the objective of this contribution is to assess the adequacy of the social investment strategy in facing the challenges posed by a multicultural society. It does so by focusing on one key area of the social investment state: active labour market policy. This area constitutes a crucial test case for the social investment strategy in terms of its suitability to a multi-ethnic society. The chapter starts by discussing the social investment strategy in general terms. It then considers the role that social investment interventions can play in a multi-ethnic society, in particular in relation to the issues of access biases and labour market discrimination. The third section focuses on ALMPs and the issue of access for different groups of disadvantaged jobseekers. Finally, section four concludes and argues that the ethnic dimension must be considered in the current debate on social investment

2

The social investment strategy

Social investment emphasises investment in human capital and the promotion of labour market participation as a strategy to deal with social problems and to reduce inequalities. As a result, ALMPs, childcare and training are the key instruments of a social investment strategy. Investment in human capital is promoted throughout the life course. Much emphasis is placed on investing in children so as to maximise their chances of succeeding in education and in the labour market (Jenson 2002, 2009; Jenson and Saint Martin 2006; Esping-Andersen 2002, 2009). Other age groups are also targeted by investments in human capital, including older people who are often strongly disadvantaged in rapidly changing labour markets. Social investment also refers to measures that remove obstacles to employment or to career advancement. These include the provision of subsidised childcare to parents of young children so that they can enter or remain in the labour market, but also active labour market policies that are less human capital intensive and simply assist jobless people in their efforts to look for employment. The use of the term ‘investment’ suggests that these policies can generate ‘returns’ for the ‘investor’, i.e. society. Of course, the ability to impact positively on societal well-being is not a specificity of the social investment approach. Social policies, in general, do produce positive 195

returns for society in terms of improved social cohesion, which benefits all, or in terms of sustained demand for goods and services and thus promoting economic growth. These positive effects of social policies are well known, but are extremely difficult to measure. Benefits in terms of social cohesion may be particularly difficult to quantify, as the value that people attach to it may vary across individuals. The social investment perspective understands returns in stricter, quantifiable terms. Investing in education, childcare or in active labour market policy increases the productivity and earning capacity of individuals, who, as a result, are less likely to require benefits and more likely to be paying taxes. Esping-Andersen insists on this quality of some of the new policies, to argue that spending on them should not count as consumption but as investment. Spending on childcare, for instance, can produce net returns to the exchequer through increased labour market participation by the mother (Esping-Andersen 2009, 96). The notion of social investment is often used in a prescriptive / normative way. EspingAndersen (2002) calls for a new “social investment” welfare state. Its key features are the focus on investing in human capital in general. It emphasises the need for policy-makers to take a life-course perspective and hence a strong focus on child well-being and development. Vandenbroucke et al. (2011) call for a social investment pact for Europe, which, even in the current tough budgetary context, should prioritise investment in policies that support children, lifelong learning, work and family reconciliation, and so forth. Hemerijck (2012) calls for “affordable social investment”, or policies that can be developed in spite of the dire state of public finances. The notion of “social investment” tries to capture a way to help disadvantaged people, based on the improvement of their life chances, especially through investment in human capital. It is one among various notions that have been used to characterise the current transformation of European welfare states. Alternative formulations include references to “active social policy” (Bonoli 2013), flexicurity (Vierbock and Clasen 2009), “recalibration” (Pierson 2001; Ferrera and Rhodes 2000) or a “new social risks” welfare state (Armingeon and Bonoli 2006). These different notions refer to very similar transformations, but emphasise different aspects. The notion of social investment, as it is used by the authors mentioned above, puts more emphasis on human capital development and on a notion of equality of opportunity that entails also some equality in outcomes. In contrast, little attention is paid to work incentives. These aspects of the social investment perspective are strongly present in the academic discussions, but have been much less influential in policy developments. Like most influential ideas, that of social investment has attracted a reasonable amount of criticism. Hemerjick (2017) has reviewed the main critiques made against the social investment perspective and the related reform strategy. Various aspects have been highlighted in the literature. First, some authors have pointed out that the distinction between social consumption and investment is artificial and inadequate. It suggests that funds spent for income replacement benefits do not generate returns for society. For some policies, however, this is unlikely to be true. For example, programmes like unemployment benefits allow skilled unemployed people 196

to wait until a suitable job is available (Nolan 2013). Second, the “social investment” rhetoric suggests superiority of productive social spending. This may result in the downgrading of nondirectly provided activities, such as caring for the disabled or the elderly (Saraceno 2017). From the point of view of the topic explored in this chapter, however, the most crucial critique against the social investment strategy refers to its inability to reach the most disadvantaged in society. Cantillon (2011), for example, has shown that the employment gains of the pre-crisis 2000 have accrued essentially to households which already had a fair level of labour market participation. Jobless households, in contrast, did not gain to any significant extent. Van Vliet and Wang (2015) found that increases in spending on social investment policies are associated with increases in poverty rates (though not in the Nordic countries). These studies suggest that social investment policies may be failing to reach those who would need them most. Most likely, the failure of social investment-inspired reforms to improve the situation of the most disadvantaged is due to the social bias in access to this type of policies – the so-called Matthew effect. Matthew effects have been identified in public services such as health and education (Le Grand 1982) or family benefits (Deleeck et al. 1983), but also in typical social investment policies, such as subsidised childcare (Pavolini and van Lancker 2018) and ALMPs (Bonoli and Liechti 2018). Over the years, the social policy literature has discussed a number of situations that qualify as Matthew effects. In many cases, universal social policies are designed so as to benefit disproportionately the middle and upper classes relative to other more disadvantaged groups (Gal 1998). This effect, which may be intended or unintended, has been identified in a number of empirical studies. Among the first to point out this effect was Le Grand (1982, 129) in his analysis of the distributional impact of a range of public services, including health care, education, housing and transportation in the UK. His verdict was unequivocal: “Public expenditure, in about all the forms reviewed, is distributed in favour of the higher social groups”. Similar findings were obtained in Belgium by Deleeck et al. (1983) who found that child benefits provided a far greater advantage to middle and upper class families than to low income ones. Higher income families tended to have more children and their children tended to stay longer in full time education, a condition to receive the benefit up to the age of 25. Social investment policies are likely to be highly exposed to the risk of Matthew effects, for a number of reasons. First, since they imply investment in human capital, they often require some pre-existing level of cognitive and non-cognitive skills. In order to learn professional skills, one must generally master some basic skills like literacy and numeracy. In addition, noncognitive skills such as perseverance or self-confidence are needed in order to successfully undergo learning processes. These skills are not evenly distributed in society, and we can assume that the most disadvantaged are also those who tend to lack them. At the same time they are essential in order to profit from social investment policies. 197

Second, the social investment approach is based on the idea that disadvantaged individuals and groups will be able, after public interventions, to cater for themselves by participating in the labour market. The problem is that the labour market is an inherently selective institution. Firms compete against each other and in order to beat the competition they need to employ the best workers. This requirement conflicts with the aspiration for universality of social investment policies (and of social policies in general), where inclusion is dependent on successful labour market integration.

3

Social investment policies in a multi-ethnic society

Overall we can thus expect social investment policies to be plagued by access biases and Matthew effects. This is likely to be particularly bad news for migrants. The two reasons we have highlighted for suspecting the existence of substantial levels of Matthew effects in social investment policies are likely to play out particularly strongly in relation to migrants and ethnic minorities. Low skilled immigrants who received limited education in their country of origin and who lack language skills are at risk of being excluded from the most ambitious social investment intervention, those that unequivocally invest in human capital. In addition, the migration experience and the lack of knowledge of the cultural codes of the host country may impact negatively on some of the non-cognitive skills needed to succeed in education and training. If not everyone can participate and profit from social investment policies, then (low skilled) migrants are arguably among the most likely to be excluded (unless an intentional effort is made to compensate for the exclusion risk, more on this below). Second, social investment policies work by facilitating and promoting labour market entry. This is obviously the case for ALMPs. However, the same argument can be made also in relation to other social investment policies. The problem in relation to the application of ALMPs in multi-ethnic societies is that members of ethnic minorities face an uphill struggle in the labour market because of employer discrimination. Several decades of discrimination research have shown that in the vast majority of advanced industrial countries, immigrants have a more difficult time than natives in entering the labour market. Studies of labour market discrimination have shown that in OECD countries applicants with a migration/ethnic minority background, all else equal, are considerably less likely to go past the initial stages of recruitment. Several meta-analyses and literature reviews conclude unanimously that ethnic based discrimination is pervasive in European and North American labour markets (see e.g. Ruedin and Zschirnt 2017; Quillian et al. 2017; Bertrand and Duflo 2017). Employer discrimination is not only a major social and justice problem, it is also a problem for the social investment welfare state and for active labour market policy. Throughout western countries, migrants are over represented among the clients of social programmes (Barret and McCarthy 2008; Nannestad 2007). In some European countries, the risk of receiving unemployment insurance benefits for non-natives is twice as high relative to natives (Barret 198

and Maître 2013, 12). True, differences in the probability of being on welfare between immigrants and natives usually diminish if one controls for socio-economic status, but this observation is irrelevant to the point that is being made here, i.e. that among the targets of ALMPs, potential victims of discrimination are overrepresented. In this respect it is also instructive to compare effect sizes of ALMPs to the impact of discrimination. With regard to ALMPs, a meta-analysis of 15 impact studies found that participation increases the hazard rate of finding a job by 9% on average, the range of effects going from – 18% (n.s.) to + 38% (sig. 5% level; Filges et al. 2015, 39). Another meta-analysis, on ethnic-based discrimination in the US, found that the probability to be invited to a job interview for a white applicant is 36% higher than it is for an African-American and 24% higher than for a Latino applicant (Quillian et al. 2017). Effect sizes of ALMPs and discrimination are clearly within the same range, with, if anything, discrimination having a larger impact. Discrimination may entirely undo the hard work done by ALMPs. Labour market discrimination is a widely studied phenomenon, with contributions from several disciplines. Economists distinguish between two types of discrimination: taste-based and statistical discrimination (Guryan and Charles 2013). The former refers to situations in which applicants with given features (e.g. a certain ethnic origin) are excluded from recruitment procedures because of employers’ dislike, a sort of irrational preference against them (Becker 1957). The notion of statistical discrimination, instead, is based on the fact that many important qualities, like productivity, personality or trainability, can neither be observed directly nor in a short time span like a job interview. As a result, employers tend to make use of information that they believe to be associated in probabilistic terms to productivity or other important qualities. This type of statistical reasoning can explain discrimination and more in general employers’ reliance on “signals” when selecting candidates. Signals are understood here as easily observable features that are believed to be associated (always in probabilistic terms) with productivity (Akerlof 1970; Spence 1973). Empirically, the existence of statistical discrimination and the fact that employers use signals when selecting candidates have been demonstrated in a large number of studies. Methodologically, the “gold standard” is the correspondence test method. In this research design, pairs of fictitious applications are sent in response to real job openings. Within each pair, the two applications are identical (or equivalent) except with regard to the feature that is being investigated, for example ethnicity. Researchers then monitor the call back rate and interpret differences as a measurement of discrimination. This method has succeeded in documenting discrimination against ethnic minority and/or immigrant groups in most western countries (see e.g. Ruedin and Zschirnt 2017; Quillian et al. 2017; Bertrand and Duflo 2017).

199

4

Active labour market policy

Active labour market policy constitutes a key component of the social investment strategy (Armingeon 2007; Bonoli 2010, 2013). Rather than simply providing a cash benefit to those who are unable to work, active labour market policy aims to remove obstacles to employment, upskill workers or to provide access to work experience. Active labour market policy differs also form the neo-liberal approach to worklessness, based on strengthening incentives only through measures such as time limits on benefit recipiency, lower benefit rates or sanctions. The label active labour market policy (ALMP) is used to describe an extremely diverse range of policies, going from vocational retraining to welfare-to-work or workfare schemes for social assistance beneficiaries. The result is that discussions of ALMPs are often confusing. ALMPs have different origins. In Sweden, ALMPs were developed as early as the 1950s, with the objective of improving the match between demand and supply of labour in the context of a rapidly evolving economy, essentially by financing extensive vocational training programmes (Swenson 2002). At the opposite extreme, the term “active” has been used to describe the approach developed in various English speaking countries, which combines placement services with stronger work incentives, time-limits on recipiency, benefit reductions and the use of sanctions. This has also been referred to as the so-called “workfare” approach (King 1995; Peck 2001). In fact, as many have pointed out, active labour market policy is a particularly ambiguous category of social policy (Barbier 2001, 2004; Clasen 2000; Clegg 2005). In order to deal with this problem, various authors have tried to develop a more differentiated view of this policy area, based on the identification of different types of ALMPs. These classifications tend to draw a line between the ‘good’ activation policies, which are about improving human capital, and the ‘bad’ ones, which essentially use negative incentives to move people from social assistance into employment. Examples of such classifications are found in Torfing (1999) who distinguishes between “offensive” and “defensive” workfare. Offensive workfare, which is the term used to describe the Danish variant of activation, relies on improving skills and empowering jobless people rather than sanctions and benefit reduction, which is the “defensive” variant found in the US. In a similar vein, Barbier distinguishes between “liberal activation” (characterised by stronger work incentives, benefit conditionality and the use of sanctions) and “universalistic activation”, which is found in the Nordic countries and relies on extensive investment in human capital essentially through training (Barbier 2004; Barbier and Ludwig-Mayerhofer 2004). Barbier also hypothesises the existence of a third type, found within Continental Europe. This third type of activation emphasises “insertion sociale”, a notion that refers more to full participation in society and less to the fact of actually having a job. Its tools are job creation programmes in the public or non-profit sector (Enjolras et al. 2001; Barbier 2001). 200

Dichotomies between human investment and incentive-based approaches to activation are a useful starting point in making sense of an ambiguous concept. However, they constitute an oversimplification of the real world and run the risk of carrying value judgements. A different type of distinction is found in Clegg (2005), who identifies two policy mechanisms that can be subsumed under activation: circulation and integration. The idea behind circulation is to improve the chances of an unemployed person to enter in contact with a potential employer (i.e. through placement services). Instead, integration refers to instruments that more directly bring the jobless into employment, like benefit conditionality or sheltered employment (Clegg 2005, 56). Clegg’s approach is promising, insofar as it focuses on the mechanisms behind the active approach and avoids the risk of making value judgments by simply drawing a distinction between ‘good’ and ‘bad’ active policy. Elsewhere, I developed a two-dimensional classification of ALMPs (Bonoli 2010, 2013). The first dimension concerns the extent to which the objective of policy is to put people back into unsubsidised market employment, provided either by private or public employers.4 Many programmes have this objective, but some, especially in continental and northern Europe during the 1980s and early 1990s, looked more like alternatives to market employment. These took the shape of temporary jobs created in the public or in the non-profit sector. They were often used to re-create an entitlement to unemployment insurance rather than to increase the chances of landing an unsubsidised job. I call this dimension the “pro-market employment orientation”. The second dimension refers to the extent to which programmes are based on investing in jobless people’s human capital. Investment can take the shape of vocational training or help in developing the sort of soft skills employers look for when selecting candidates. Intersecting these two dimensions allows us to identify different types of labour market policy and to map the variety that exists under the label “active labour market policy”. Table 1: Four types of active labour market policy Investment in human capital

Pro-market employment orientation

Weak

Strong

None

Weak

Strong

(passive benefits)

Occupation -Job creation schemes in the public sector -Non employment-related training programmes Employment assistance -Placement services -Job subsidies -Counselling -Job search programmes

(basic education)

Incentive reinforcement -Tax credits, in work benefits -Time limits on recipiency -Benefit reductions -Benefit conditionality Source: Adapted from Bonoli (2010).

Upskilling -Job related vocational training

4

By “unsubsidised market employment” I mean jobs that are created as a result of a demand for labour by private or public employers, and not in order to absorb excess labour supply. 201

The classification of ALMPs presented in Table 1 allows us to generate hypotheses with regard to the type of ALMPs that are more and less likely to include/exclude ethnic minority members. More specifically, we can assume that “upskilling” will be the most difficult category to enter for migrants, because that is where more pre-existing abilities are likely to be required. These concern both cognitive and non-cognitive skills as well as language skills. Migrants might have an easier time accessing employment assistance measures which require less starting human capital. In fact, the level of human capital required to profit from these measures depends almost entirely on the requirement of the jobs that are being applied for. To the extent that a sufficient supply of low skill entry level jobs are available, low skilled migrants can theoretically profit from these interventions. Ethnic minorities, however are likely to be overrepresented in the third type of ALMPs discussed here, i.e. occupation. These are programmes like temporary jobs in the public sector, where very little pre-existing human capital is required. They are often the default option for jobseekers who cannot access other programmes. Finally, the forth category, incentive reinforcement, is likely to concern all jobseekers regardless of ethnic origin. To sum up, the reasoning presented above leads us to hypothesise that the distribution of immigrants among ALMPs participants will be skewed. They will be underrepresented in training, and overrepresented in the category “occupation” (Table 1). Their representation in employment assistance programmes will instead be somewhere in-between. But are these theory based expectations confirmed by empirical analysis? Fortunately it is relatively easy to answer this question. In fact, in the vast literature evaluating ALMPs, researchers often report information on who has accessed the programme under scrutiny. Bonoli and Liechti (2018) have surveyed this literature with meta-analytical techniques. They found that the distribution of migrants among ALMPs participants is indeed skewed. However, their results are only partly in line with my expectations. Basically immigrants are slightly underrepresented among participants in training programmes. They are strongly underrepresented in one type of employment assistance programme, i.e. job subsidies for private sector employers. Contrary to expectations, they are also underrepresented in occupation-type programmes (Table 2). Table 2: Access bias for migrants in ALMPs (representation relative to natives): Number of studies Immigrants are

Equal representation of

Immigrants are

underrepresented

immigrants and natives

overrepresented

Training

11

9

8

Employment subsidy

6

1

1

Employment programme

7

6

0

(Occupation) Source: Bonoli and Liechti (2018).

202

The results of the analysis by Bonoli and Liechti (2018) point in a few different directions. First, there are very few programmes in which immigrants are overrepresented. These are found almost exclusively in the training category. Unfortunately the literature reviewed by Bonoli and Liechti does not allow them to say anything with regard to the type of training delivered. However, it could very well be the case that migrant overrepresentation concerns only language training. If this were the case, then there would still be an access problem to the other training courses that are more directly targeted on professional skill. This hypothesis is plausible, given the fact that the majority of programmes surveyed do display an access bias, either positive or negative for migrants. Second, the biggest access bias problem is found in employment subsidies. This is at first sight surprising, because these could precisely be useful measures for migrants. In a metaanalysis of the effectiveness of different ALMPs for immigrants, Buteschek and Walter (2014) found that wage subsidies are the most effective category of measures for them. How do we explain the fact that immigrants are vastly underrepresented? The answer to this question could be labour market discrimination. In order to access a wage subsidy a job seeker needs first to find an employer who is willing to hire him/her (with a subsidy). It could be the case that immigrants targeted by this policy are considerably less likely to be offered a job, even with a subsidy. Third, contrary to expectation, there is no overrepresentation of immigrants in the “occupation” category. This suggests that these programmes, even though designed for labour market distant candidates, are not meant for migrants. It could be the case that in reality immigrants are overrepresented among non-participants. These very detailed results suggest that access biases and a Matthew effect may constitute a serious problem with regard to the ability of ALMPs to reach disadvantaged immigrants. The main reason for this may not be so much the lack of pre-existing human capital, as hypothesised, but more labour market discrimination and absence of suitable programmes. Given the above, it is at best unfortunate and at worst utterly inadequate, that active labour market policies have so far mostly ignored labour market discrimination, which could potentially be one of the biggest obstacles to labour market re-entry for a sizable proportion of their target population. True, some of the existing tools can be seen as indirectly targeting discrimination. This is the case of wage subsidies that aim to make disadvantaged candidates more attractive by covering part of the wage costs (Bernhardt et al. 2008), though given the high deadweight losses their effectiveness as an anti-discrimination tool is highly questionable (Marx 2001; Welters and Muysken 2006). Work experience schemes (e.g. internships, employment programmes) can also be seen as an indirect way to reduce the impact of discrimination, as they allow potential employers to observe applicants’ true qualities. In addition, it is quite possible that case workers provide jobseekers with advice that is aimed at limiting the putative impact of discrimination. However, in general the toolkit of ALMPs does 203

not include measures aimed at explicitly reducing the risk of applicants being excluded from jobs because of discrimination. Research on ALMPs has also largely ignored the connection between ALMPs and labour market discrimination, and there is little knowledge of interventions that are designed to reduce the impact of prejudice on labour market access. The bulk of research on how to counteract discrimination has been done by psychologists, in the lab and often with subjects selected among student populations. The extent to which these findings can be generalised to employers acting in real labour markets is probably rather limited (Bertrand and Duflo 2017, 338-339). More relevant to the real world of recruitment are some recent studies on blind hiring (e.g. anonymous CVs), who found interesting but rather mixed results (Krause et al. 2012; Behaghel et al. 2014).

5

Conclusion: ALMPs for a multi-ethnic society

The objective of this chapter was to assess the suitability of social investment as a strategy to deal with the crisis of the European social model in the context of increasingly multi-ethnic societies. The issue has been discussed in relation to a policy that can be considered as the backbone of the social investment welfare state, i.e. ALMPs. The evidence available in relation to both access biases to ALMPs and labour market discrimination, suggests that in order to make the policy suitable to the challenge posed by a multi-ethnic society the issue of discrimination needs to be tackled head on. ALMPs should be more explicit about targeting discrimination and migrant disadvantage in general. First, in relation to access, if anything, immigrants should be overrepresented among the users of ALMPs across the board, as they are likely to lack many features that are needed to succeed in the labour market: human capital, language skills, knowledge of the local culture, informal networks, etc. This would be perfectly rational from the point of view of budget sensitive governments, since immigrants are overrepresented among welfare clients. Second, policy should be more serious in fighting discrimination. This task is generally not considered part of what ALMPs have to do, and it is left to judicial authorities or general antidiscrimination bodies. However, the substantial impact that labour market discrimination has in terms of limiting the access to the labour market for immigrant jobless people suggests that ALMPs should incorporate fighting discrimination explicitly among their objectives. The reason presented here refers to ALMPs, but can to some extent be generalised to other aspects of the social investment strategy. As mentioned above, programmes such as childcare suffer from access biases. While these may not be due to discrimination, the fact that children with an immigrant background are tendentially excluded from group-based formal childcare constitutes a problem in the ability of the social investment strategy to reach the most disadvantaged. Social investment need to take into account the multi-ethnic quality of our societies and develop suitable interventions. 204

References Abrassart, Aurélien, and Giuliano Bonoli. 2015. “Availability, Cost or Culture? Obstacles to Childcare Services for Low Income Families.” Journal of social policy 44 (4): 787-806. Akerlof, George A. 1970. “Market for Lemons - Quality Uncertainty and Market Mechanism.” Quarterly Journal of Economics 84 (3): 488-500. Armingeon, Klaus. 2007. “Active Labour Market Policy, International Organizations and Domestic Politics.” Journal of European Public Policy 14 (6): 905-932. Armingeon, Klaus, and Giuliano Bonoli, eds. 2006. The Politics of Postindustrial Wefare States. Adapting PostWar Social Policies to New Social Risks. London: Routledge. Barbier, Jean-Claude. 2001. Welfare to Work Policies in Europe. The Current Challenges of Activation Policies. Paris: Centre d'études de l'emploi. Barbier, Jean-Claude. 2004. “Systems of Social Protection in Europe: Two Contrasted Paths to Activation, and maybe a Third.” In Labour and Employment Regulation in Europe, edited by Jens Lind, Herman Knudsen, and Henning Jørgensen, 233-254. Brussels: Peter Lang. Barbier, Jean-Claude, and Wolfgang Ludwig-Mayerhofer. 2004. “Introduction: The Many Worlds of Activation.” European Societies 6 (4): 424-436. Barrett, Alan, and Bertrand Maître. 2013. “Immigrant Welfare Receipt across Europe.” International Journal of Manpower 34 (1): 8-23. Barrett, Alan, and Yvonne McCarthy. 2008. “Immigrants and Welfare Programmes: Exploring the Interactions between Immigrant Characteristics, Immigrant Welfare Dependence, and Welfare Policy.” Oxford Review of Economic Policy 24 (3): 542-559. Becker, Gary S. (1971[1957]). The Economics of Discrimination. Chicago: Chicago University Press. Behaghel, Luc, Bruno Crépon, and Thomas Le Barbanchon. 2014. Unintended Effects of Anonymous Resumes. Bonn: IZA, DP No. 8517. Bernhard, Sarah, Hermann Gartner, and Gesine Stephan. 2008. Wage Subsidies for Needy Job-Seekers and Their Effect on Individual Labour Market Outcomes after the German Reforms. Nürnberg: IAB-Discussion Paper No. 21/2008. Bertrand, Marianne, and Esther Duflo. 2017. “Field Experiments on Discrimination.” In Handbook of Economic Field Experiments, edited by Abhijit V. Banerjee and Esther Duflo, 309-393. Amsterdam: North Holland. Bonoli, Giuliano. 2010. “The Political Economy of Active Labour Market Policies.” Politics & Society 38 (4): 435-457. Bonoli, Giuliano. 2013. The Origins of Active Social Policy. Active Labour Market Policy and Childcare in a Comparative Perspective. Oxford: Oxford University Press. Bonoli, Giuliano, Bea Cantillon, and Wim van Lancker. 2017. “Social Investment and the Matthew effect.” In The Uses of Social Investment, edited by Anton Hemerijck, 66-76. Oxford: Oxford University Press. Bonoli, Giuliano, and Fabienne Liechti. 2018. “Good Intentions and Matthew Effects: Access Biases in Participation in Active Labour Market Policies.” Journal of European Public Policy 25 (6): 894-911. Butschek, Sebastain, and Thomas Walter. 2014. “What Active Labour Market Programmes Work for Immigrants in Europe? A Meta-Analysis of the Evaluation Literature.” IZA Journal of Migration 3 (1). Cantillon, Bea. 2011. “The Paradox of the Social Investment State: Growth, Employment and Poverty in the Lisbon Era.” Journal of European Social Policy 21(5): 432-449. Clasen, Jochen. 2000. “Motives, Means and Opportunities: Reforming Unemployment Compensation in the 1990s.” In Recasting European Welfare States, edited by Mauriziio Ferrera and Martin Rhodes, 89-112. London: Frank Cass. Clegg, Daniel. 2005. Activating the Multi-Tiered Welfare State: Social Governance, Welfare Politics and Unemployment Policies in France and the United Kingdom. Florence: European University Institute, PhD Thesis. Deleeck, Herman, J. Huybrechs, and Bea Cantillon. 1983. Het Mattüseffect. Antwerpen: Kluwer. Enjolras, Bernard, Jean Louis Laville, Laurent Fraisse, and Heather Trickey. 2001. “Between Subsidiarity and Social Assistance. The French Republican Route to Activation.” In An Offer You can't Refuse: Workfare in International Perspective, edited by Ivar Lødemel and Heather Trickey, 71-104. Bristol: The Policy Press. Esping-Andersen, Gøsta. 2009. The Incomplete Revolution. Adapting to Women's New Roles. Cambridge: Polity Press. Esping-Andersen, Gøsta. ed. 2002. Why We Need a New Welfare State. Oxford: Oxford University Press. Ferrera, Maurizio, Anton Hemerijck, and Martin Rhodes. 2000. The Future of Social Europe. Recasting Work and Welfare in the New Economy. Oeiras: Portuguese Ministry of Labour and Solidarity/Celta Editora. Filges, Trine, Geir Smedslund, Due Knudsen, and Anne-Marie K. Jørgensen. 2015. “Active Labour Market Programme Participation for Unemployment Insurance Recipients: A Systematic Review.” Campbell Systematic Reviews 2. 205

Gal, John. 1998. “Formulating the Matthew Principle: On the Role of the Middle Classes in the Welfare State.” Scandinavian Journal of Social Welfare 7 (1): 42-55. Guryan, Jonathan, and Kerwin K. Charles. 2013. “Taste‐based or Statistical Discrimination: The Economics of Discrimination Returns to its Roots.” The Economic Journal 123 (572): F417-F432. Heckman, James J., and Jeffrey A. Smith. 2004. “The Determinants of Participation in a Social Program: Evidence from a Prototypical Job Training Program.” Journal of Labor Economics 22 (2): 243-298. Hemerijck, Anton. 2012. Changing welfare states. Oxford: Oxford University Press. Hemerijck, Anton. 2017. “Social Investment and its Critics.” In The Uses of Social Investment, edited by Anton Hemerijck, 3-42. Oxford: Oxford University Press. Jenson, Jane. 2002. From Ford to Lego: Redesigning Welfare Regimes. Boston: Annual Meeting of the American Political Science Association. Jenson, Jane. 2009. “Redesigning Citizenship Regimes after Neoliberalism. Moving Towards Social Investment.” In What future for social investment?, edited by Nathalie Morel, Bruno Palier, and Joakim Palme, 27-44. Stockholm: Institute for Future Studies. Jenson, Jane, and Denis Saint-Martin. 2006. “Building Blocks for a New Social Architecture: the LEGO (TM) Paradigm of an Active Society.” Policy and Politics 34 (3): 429-451. King, Desmond. 1995. Actively Seeking Work? The Politics of Unemployment and Welfare Policy in the United States and Great Britain. Chicago: University of Chicago Press. Krause, Annabelle, Ulf Rinne, and Klaus F. Zimmermann. 2012. “Anonymous Job Applications in Europe.” IZA Journal of European Labor Studies 1 (5). Le Grand, Julian. 1982. The Strategy of Equality: Redistribution and the Social Services. London: Allen & Unwin. Marx, Ive. 2001. “Job Subsidies and Cuts in Employers' Social Security Contributions: The Verdict of Empirical Evaluation Studies.” International Labour Review 140 (1): 69-83. Nannestad, Peter. 2007. “Immigration and Welfare States: A Survey of 15 Years of Research.” European Journal of Political Economy 23 (2): 512-532. Nolan, Brian. 2013. “What Use Is 'Social Investment'?” Journal of European Social Policy 23 (5): 459-468. Pavolini, Emmanuel, and Wim Van Lancker. 2018. “The Matthew Effect in Childcare Use: A Matter of Policies or Preferences?” Journal of European Public Policy 25 (6): 878-893. Peck, Jamie. 2001. Workfare States. New York: Guildford Press. Pierson, Paul. 2001. “Coping with Permanent Austerity: Welfare State Restructuring in Affluent Democracies.” In The New Politics of the Welfare State, edited by Paul Pierson, 410-456. Oxford: Oxford University Press. Quillian, Lincoln, Devah Pager, Ole Hexel, and Arnfinn H. Midtboen. 2017. “Meta-Analysis of Field Experiments Shows No Change in Racial Discrimination in Hiring Over Time.” Proceedings of the National Academy of Sciences of the United States of America 114 (41): 10870-10875. Saraceno, Chiara. 2017. “Family Relationships and Gender Equality in the Social Investment Discourse.” In The uses of social investment, edited by Anton Hemerijck, 59-65. Oxford: Oxford University press. Spence, Michael. 1973. “Job Market Signaling.” Quarterly Journal of Economics 87 (3): 355-374. Swenson, Peter A. 2002. Capitalists against Markets: The Making of Labor Markets and Welfare States in the United States and Sweden. Oxford: Oxford University Press. Torfing, Jacob. 1999. “Workfare with Welfare: Recent Reforms of the Danish Welfare State.” Journal of European social Policy 9 (1): 5-28. Van Lancker, Wim. 2013. “Putting the Child-Centered Strategy to the Test: Evidence for the EU 27.” European Journal of Social Security 15 (1): 4-27. Vandenbroucke, Frank, Anton Hemerijck, and Bruno Palier. 2011. The EU Needs a Social Investment Pact. Brussles: OSE, Opinion paper No. 5. Viebrock, Elke, and Jochen Clasen. 2009. “Flexicurity and Welfare Reform: A Review.” Socio-Economic Review 7 (2): 305-331. Welters, Riccardo, and Joan Muysken. 2006. “Employer Search and Employment Subsidies.” Applied Economics 38 (12): 1435-1448. Zschirnt, Eva, and Didier Ruedin. 2016. “Ethnic Discrimination in Hiring Decisions: A Meta-Analysis of Correspondence Tests 1990–2015.” Journal of Ethnic and Migration Studies 42 (7): 1-19.

206

Die sozialpolitischen Folgen des Ersten Weltkriegs: Deutschland, Österreich und die Schweiz im Vergleich Herbert Obinger1

Abstract: This contribution deals with the social policy consequences of the First World War in Germany, Austria and Switzerland. Although at the beginning of the war, there were initial cuts in the field of workers’ protection, the massive social upheaving induced by conflict eventually led to important groundwork with lasting impact on the development of social policy. A comparison reveals striking parallels between the three countries, even though the social political effects of war were much stronger and more sustained in the two warring countries.

1

Einleitung

Ist Krieg als eine existenzielle Krisensituation eine „Entwicklungsbedingung“ des Wohlfahrtsstaates (Titmuss 1958; Reidegeld 1989)? Diese Frage wurde von der vergleichenden Wohlfahrtsstaatsforschung lange stark ausgeblendet (Obinger et al. 2018). Hundert Jahre nach Kriegsende richtet dieser Beitrag seinen Blick auf die sozialpolitischen Folgen des Ersten Weltkriegs in Deutschland, Österreich und der Schweiz.2 Der Vergleich umfasst damit zwei kriegsführende Staaten und ein neutrales Land, das nicht unmittelbar in militärische Kampfhandlungen involviert war. Diese Varianz ist essentiell, um die sozialpolitischen Folgen des Ersten Weltkriegs abschätzen zu können. War der Erste Weltkrieg tatsächlich der „große Schrittmacher der Sozialpolitik“, wie der SPD-Politiker Ludwig Preller (1978 [1947], 85) feststellte, oder nicht eher die bestenfalls mit einem „Tropfen sozialen Öls“ versetzte Ursache von Verelendung und Massensterben, wie andere Zeitzeugen wie der deutsche Gewerkschafter Alexander Schlicke3 oder der österreichische Sozialdemokrat Ferdinand Hanusch4 behaupteten? Dieser Beitrag argumentiert, dass beide Positionen vereinbar sind, sofern zwischen kurzund langfristigen Effekten unterschieden wird. In beiden kriegsführenden Staaten führten die dramatischen sozialen Verwerfungen und Notlagen in der späten Kriegsphase zu bedeutenden 1

Universität Bremen [email protected] 2 Klaus Armingeon hat maßgeblich die Drei-Länder-Tagungen der politikwissenschaftlichen Fachvereinigungen der Schweiz, Deutschlands und Österreichs initiiert. Dadurch sollten nicht zuletzt vergleichende Untersuchungen dieser Länder angeregt werden. Dieser Beitrag ist ein entsprechender Versuch. 3 Zitiert nach Feldman (1985, 282). 4 Hanusch (1927, 16). Ferdinand Hanusch und Alexander Schlicke wurden nach Kriegsende Sozialminister bzw. Reichsarbeitsminister.

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 R. Careja et al. (eds.), The European Social Model under Pressure, https://doi.org/10.1007/978-3-658-27043-8_13

207

sozialpolitischen und weitgehend identischen Weichenstellungen, die zusammen mit den politischen und wirtschaftlichen Umwälzungen nach Kriegsende zu einem Quantensprung in der staatlichen Sozialpolitik führten. Der Weltkrieg bewirkte auch in der neutralen Schweiz sozialpolitische Kursänderungen, es kam aber nicht zu einer sprunghaften Entwicklung staatlicher Sozialpolitik, sondern der historisch gewachsene public-private mix wurde fortgeschrieben, wobei private Sozialschutzformen und Wohltätigkeit sogar gestärkt wurden.

2

(Sozial-)politische Ausgangslage

Zu Kriegsbeginn zählten das Deutsche Kaiserreich und die zisleithanische Reichshälfte der Habsburgermonarchie zu den Ländern mit den am höchsten entwickelten öffentlichen Sozialschutzprogrammen. Das Risiko des Einkommensverlustes durch Krankheit, Unfall und Alter (in Österreich jedoch nur für Angestellte) war durch obligatorische Sozialversicherungsprogramme abgesichert, was u.a. mit den Legitimationsnöten der Eliten in beiden autokratischen Monarchien erklärt wurde (Alber 1982). Die demokratische und industriell hoch entwickelte Eidgenossenschaft verfügte hingegen mit der Fabrikgesetzgebung (1877) sehr früh über eine im internationalen Vergleich sehr fortschrittliche Arbeiterschutzgesetzgebung, deren sehr hoher Standard noch am ehesten von der Donaumonarchie erreicht wurde, während Deutschland zwar Schutzgesetze für Kinder, Jugendliche und Frauen erlassen hatte, aber keinen gesetzlichen Maximalarbeitstag für erwachsene männliche Arbeiter kannte. Noch im Juni 1914 wurde in der Schweiz mit einer Revision des Fabrikgesetzes die Wochenarbeitszeit von 64 auf 59 Stunden reduziert und Kündigungsbeschränkungen erlassen (Emmenegger 2014, 83). Auf dem Gebiet der Sozialversicherung war die Schweiz auf nationaler Ebene hingegen ein Nachzügler. Neben dem Föderalismus, der mit Ausnahme der Arbeitsschutzgesetzgebung und der Kranken- und Unfallversicherung (seit 1890) dem Bund keine sozialpolitischen Kompetenzen überantwortete, verzögerte auch die direkte Demokratie die Sozialgesetzgebung (Obinger 1998; Armingeon 2001). 1900 scheiterte eine am deutschen Vorbild angelehnte Kranken- und Unfallversicherung. Lediglich die Militärversicherung (Kranken- und Unfallversicherung für Armeeangehörige) überstand das Abstimmungsdebakel und wurde 1901 als erstes Sozialversicherungsgesetz der Schweiz verabschiedet. Erst eine Revision der Kranken- und Unfallversicherung in Gestalt eines liberalen Rahmengesetzes überstand 1911 ein zweites Referendum (Krumbiegel 1913; Immergut 1992). Die Krankenversicherung blieb freiwillig5 und anstellte einer Beitragsfinanzierung beschränkte sich der Bund auf die Subventionierung der Krankenkassen. Die im Gesetz ebenfalls geregelte Unfallversicherung trat erst 1918 in Kraft. Für alle anderen Felder der Sozialpolitik blieben die Kantone und Gemeinden zuständig, die auf dem Gebiet der Alters- und Arbeitslosenversicherung vereinzelt 5

Allerdings konnten die Kantone eine Versicherungspflicht einführen, und machten davon auch zunehmend Gebrauch. 208

auch entsprechende Regelungen erlassen haben. Hinzu kam ein hoch entwickeltes Versicherungswesen, wodurch v.a. ökonomisch besser gestellte Bevölkerungsgruppen etwa im Bereich der Altersvorsorge die Lücken staatlicher Sozialgesetzgebung kompensieren konnten. In allen drei Ländern kam es mit dem Kriegsausbruch zu einer Mobilmachung der Armeen, die überall auf der allgemeinen Wehrpflicht basierte. Die Schweiz erklärte Anfang August ihre Neutralität und konzentrierte sich fortan auf die Sicherung ihrer Grenzen. In allen drei Ländern herrschte anfangs innenpolitischer Burgfrieden. Während aber in Deutschland und Österreich eine regelrechte Kriegseuphorie herrschte, war die Schweiz entlang der Sprachgrenzen gespalten. Die Deutschschweiz sympathisierte eher mit den Achsenmächten, während die Westschweiz mehrheitlich Frankreich und das überfallene neutrale Belgien unterstützte. Dieser Graben wurde durch die Wahl des Deutschschweizers Ulrich Wille zum General noch verstärkt. Mit den Notgesetzen vom 4. August (Deutschland), der Übertragung außerordentlicher Kriegsvollmachten auf den Bundesrat (3. August, Schweiz) und dem Inkrafttreten des Kriegsleistungsgesetzes und der Einschränkung staatsbürgerlicher Grundrechte (Österreich, siehe Redlich 1925, 82-95, 116-126) wurden mit Kriegsbeginn überall die Kompetenzen der Exekutive drastisch gestärkt. Trotz umfassender Kriegsvollmachten für den Bundesrat und Zensur blieb in der Schweiz die Demokratie aber intakt. Herrschte in der Schweiz ein „demokratischer Autoritarismus“ (Tanner 2015, 129), wurden in den Mittelmächten die Machtbefugnisse des Militärs stark ausgeweitet, und insbesondere Österreich, wo der Reichsrat bis Mai 1917 nicht mehr zusammentrat, mutierte quasi zu einer Militärdiktatur (Rauchensteiner 2013).

3

Kriegssozialpolitik

3.1

Mittelmächte

In Deutschland und Österreich kam es in sozialpolitischer Hinsicht zu Kriegsbeginn zu Einschnitten im Bereich des Arbeitsschutzes. Eines der deutschen Notgesetze vom 4. August (Gesetz betreffend Ausnahmen von Beschäftigungsbeschränkungen gewerblicher Arbeiter 6) eröffnete die Möglichkeit, den Arbeitsschutz zugunsten von Frauen, Kindern und Jugendlichen einzuschränken. Für die Zeit des Krieges konnten die Behörden auf Antrag von Unternehmen Ausnahmen von den bestehenden Arbeitsschutzbestimmungen erteilen. In der Praxis wurde davon stark Gebrauch gemacht, da nach kurzem Anstieg der Arbeitslosigkeit durch die Massenmobilisierung bald Arbeitskräftemangel herrschte. Auch in der Habsburgermonarchie wurden durch eine Reihe von Verordnungen die Arbeitszeit ausgeweitet, die Arbeiterfreizügigkeit eingeschränkt und bestehende Arbeitsschutzvorschriften gelockert (Stolper 1915, 101-102; Adler 1927). Bereits im Juli 1914 wurde für Dauer des Krieges die 6

RGBl 1914, 333-334.

209

Arbeit an Sonn- und Feiertagen zugelassen,7 während mit dem Kriegsleistungsgesetz arbeitsfähige und vom Militär zurückgestellte Männer zwischen 17 und 50 Jahren zu Zwangsarbeit in militärisch relevanten Betrieben verpflichtet werden konnten. 1915 wurde die zulässige Arbeitszeit nochmals ausgeweitet und ein Jahr später das Maximalalter des Kriegsleistungsgesetzes auf 55 Jahre angehoben. Obwohl der Großteil der Arbeiterschutzgesetzgebung formell in Kraft blieb, führte die Übernahme von Betrieben durch die Militärbehörden und die Drohung mit Fronteinsatz dazu, dass Arbeitsschutzstandards und das Koalitionsrecht in der Praxis ausgehöhlt wurden (Adler 1927, 69-77). Trotz dieser Einschnitte und eines erheblichen Rückgangs im Organisationsgrad hielten die Gewerkschaften zunächst aber an der Burgfriedenspolitik fest. Da weithin mit einem kurzen und erfolgreichen Krieg gerechnet wurde, hofften sie wie ihr deutsches Pendant auf politische Zugeständnisse als Anerkennung ihrer Kooperation (Mai 1997; Grandner 1992, 77, 107-111). Die Veränderungen im Bereich der Sozialversicherung in der frühen Kriegsphase bestanden zum einen aus kriegsbedingten Anpassungsmaßnahmen (Beibehaltung der Anwartschaften in der Krankenversicherung wegen Kriegsdienst, Anrechnung von Kriegsdienst auf die Rente). Zum anderen kam es in Deutschland zu einem Leistungsausbau. Die Ende 1914 eingeführte Reichswochenhilfe gewährte krankenversicherten Wöchnerinnen Wochengeld für 8 Wochen, einen Entbindungskostenbeitrag, Stillgeld und eine Beihilfe zur Deckung der Arzt- und Hebammenkosten. 1915 wurde der Bezieherkreis erweitert. In der Alterssicherung wurde 1916 das Rentenalter von 70 auf 65 Jahre abgesenkt, wodurch sich die Zahl der Altersrenten schlagartig verdoppelte. Zudem stellte die Sozialversicherung aus ihrem Vermögen erhebliche Mittel für die freiwillige Kriegswohlfahrtspflege bereit (Zahn 1916, 19), die Familien eingezogener Soldaten und Kriegsinvaliden unterstützte. Zwar kam es auch in der staatlichen Sozialhilfe (Kriegsfürsorge) durch Professionalisierung, Zentralisierung und die Schaffung von Rechtsansprüchen zu bedeutenden Weiterentwicklungen bei der Unterstützung dieser Gruppen (Starke 2018, 42), dennoch waren die Leistungen und Mittel für einen Krieg dieser Dimension unzureichend, sodass die meist von Frauenorganisationen getragene private Wohlfahrts- und Pflegetätigkeit zunehmende Bedeutung erlangte. Ähnliche Entwicklungen sind für Österreich konstatierbar (Hämmerle 2014). Die Hoffnung auf einen kurzen Krieg erwies sich aber als Chimäre. Mit zunehmender Dauer des Krieges und der alliierten Seeblockade spitzten sich die Versorgungsprobleme in fast allen Wirtschaftsbereichen zu. Lebensmittelengpässe, Lohnfragen und die Teuerung führten bald zu ersten Streiks und wachsender Unzufriedenheit im Hinterland. Die Regierungen der Achsenmächte standen daher unter wachsenden Druck, den Gewerkschaften konkrete Zugeständnisse für ihre Kooperation zu machen (Feldman 1985, 111-112). Ende Juni 1916 gab die deutsche Reichsregierung einer langen Forderung der Arbeiterbewegung nach, indem im Vereinsrecht der Passus gestrichen wurde, demzufolge Gewerkschaften als politische 7

RGBl 183 und 184/1914. Im August traten entsprechende Bestimmungen im Bergbau in Kraft (RGBl 219/1914).

210

Organisationen galten8 und das formell garantierte Koalitionsrecht praktisch einschränkte. Im April 1917 fiel der sogenannte Sprachenparagraf im Vereinsgesetz, der die Organisierung ausländischer Arbeitnehmer erschwert hatte.9 In Österreich versuchte die Regierung, unzufriedenen Arbeitern mit Änderungen im Arbeitsrecht entgegenzukommen (Adler 1927, 9596). So wurde 1916 für bestimmte eingezogene Angestellte der Kündigungsschutz10 verbessert und eine Änderung des Allgemeinen Bürgerlichen Gesetzbuches sicherte eine Lohnfortzahlung bei Erwerbsunfähigkeit wegen Krankheit oder Unfall bis zu einer Woche zu. Ironischerweise war es der Rohstoffmangel, der in den Achsenmächten zu partiellen Verbesserungen bzw. Wiederherstellung des Vorkriegsstandes im Bereich des klassischen Arbeiterschutzes führte. In Deutschland betraf dies das Nachtbackverbot (1915),11 das Verbot des Anstrichs von Häusern, Mauern und Zäunen mit Bleiweiß (1915),12 die Schließzeitregelungen für Geschäfte und Gaststätten (1916),13 Beschränkungen des Sonntagsverkaufs sowie Arbeitszeitbeschränkungen (1915) in Spinnereien und Webereien. In Österreich wurde 1915 die Sonn- und Feiertagsruhe im Handelsgewerbe wieder eingeführt14 und Anfang 1917 per Verordnung15 die Nachtarbeit in Bäckereien verboten, sofern diese nicht der Heeresversorgung dienten. Die Regierung gewährte zudem eine Arbeitslosenfürsorge für Personen, die wegen Rohstoffmangel erwerbslos wurden (Grandner 1992, 243, 255-256). Fortschritte gab es auch im deutschen Tarifwesen (Sinzheimer 1920). Entgegen mancher Befürchtungen wurden bestehende Tarifverträge im Krieg nicht ausgehöhlt. Auf Druck des Militärs kam es in militärisch relevanten Heimarbeitsbranchen sogar zu einer Ausweitung des Tarifvertragswesens (Reidegeld 2006, 313-314). So kam 1915 ein Reichstarifvertrag für die Militärausrüstungsindustrie zustande und durch einen Erlass des Oberbefehlhabers in den Marken wurde erstmals ein Tarifvertrag unabdingbar.16 Auf Ablehnung stießen Tarifverträge hingegen in der Schwer- und Textilindustrie, auch wenn hier – wiederum auf militärische Initiative hin – bis 1916 landesweit Schlichtungsstellen etabliert wurden (Preller 1978, 38, 7273). Die zweite Hälfte des Jahres 1916 brachte in politischer, militärischer und wirtschaftlicher Hinsicht grundlegende und auch für die Sozialpolitik bedeutsame Veränderungen. Maßgeblich dafür war die unvorhergesehen lange Kriegsdauer, für die keines der beiden Länder auch nur ansatzweise vorbereitet war. In Deutschland versuchte die Ende August 1916 installierte neue Oberste Heeresleitung (3. OHL) unter Paul von Hindenburg und Erich Ludendorff mit einem 8

Gesetz zur Änderung des Vereinsgesetzes vom 19. April 1908 (RGBL 1916, 635). Zuvor war es Arbeitern unter 18 Jahren verboten, politischen Vereinen (und damit den Gewerkschaften) beizutreten. Ein noch größeres Hindernis für das volle Koalitionsrecht war aber der § 153 der Gewerbeordnung (siehe unten). 9 RGBl 1917, 361. 10 RGBl 58/1916. 11 RGBl 1915, 8. 12 RGBl 1915, 671. 13 RGBl 1916, 1355. 14 RGBl 403/1915 15 RGBl 54/1917. 16 Konkret wurden Abweichungen von den vom Kriegsbekleidungsamt festgelegten Löhnen verboten (Zahn 1916, FN 16). 211

radikalen Programm den Krieg nun mit allen Mitteln zu entscheiden. In der Donaumonarchie führten die Ermordung des autoritären Ministerpräsidenten von Stürgkh und der Tod von Kaiser Franz-Josef zu einer politischen Kursänderung. In Deutschland forcierte die 3. OHL mit dem sogenannten Hindenburgprogramm die totale Mobilisierung aller Ressourcen und Kräfte für den Krieg. Für die Rüstungsindustrie sollten zusätzliche Arbeitskräfte mobilisiert und das bestehende Arbeitskräftepotenzial vollständig ausgenutzt werden, um die hochgesteckten Rüstungsproduktionsziele der OHL zu erreichen. Schon wenige Tage nach ihrer Berufung machte sich die OHL für die Einführung der Sonntagsarbeit stark17 und wies Überlegungen, die im August 1914 erlassene Ausnahmegesetzgebung zum Arbeiterschutz rückgängig zu machen, scharf zurück.18 Zur Durchsetzung ihrer ehrgeizigen Ziele schlug sie vielmehr einen zivilen Arbeitszwang (auch für Frauen), Maßnahmen zur Mobilisierung zusätzlicher Soldaten sowie die Schaffung einer Zentralstelle für die Arbeitskräfte-, Rohstoff- und Nahrungsmittelbewirtschaftung vor (Feldman 1985, 150-151). Die Regierung ging letztlich auf einige dieser Forderungen ein. Im November 1916 wurde das Kriegsamt unter der Leitung von Generalleutnant Wilhelm Groener mit dem Ziel geschaffen, die Beschaffung von Rohstoffen, Waffen und Arbeitskräften zentral zu koordinieren. Groener vertrat gegenüber dem Reichstag auch das „Gesetz über den Vaterländischen Hilfsdienst“,19 das Anfang Dezember 1916 mit großer Mehrheit beschlossen wurde. Es verpflichtete alle nicht im Fronteinsatz stehenden Männer zwischen 17 und 60 Jahren zur Arbeit in kriegsrelevanten Sektoren und schränkte zu diesem Zweck die Freizügigkeit der Arbeitnehmer ein. Ein Arbeitsplatzwechsel konnte nur mit Genehmigung des Arbeitgebers oder – im Fall zurückgestellter Wehrpflichtiger – mit Genehmigung der Militärbehörden erfolgen. Vor dem Hintergrund des Stimmungswandels und der Versorgungssituation im Hinterland waren aber zur Durchsetzung der Arbeitspflicht im Sinne einer quid pro quo-Logik Konzessionen an die Gewerkschaften notwendig, um das Stillhalten der Arbeiterschaft zu sichern (Feldman 1985, 175). Auch das Militär war bestrebt, Streiks in der Kriegswirtschaft mit allen Mitteln abzuwenden. Nach zähen Verhandlungen zum Hilfsdienstgesetz (HDG) kam die Regierung der Arbeiterbewegung mit der Einrichtung von gewählten Arbeiter- und Angestelltenausschüssen in kriegsrelevanten Betrieben mit mindestens 50 Beschäftigten sowie der Schaffung von paritätischen Schlichtungsstellen entgegen. Verweigerte der Arbeitgeber dem Hilfsdienstpflichtigen den Arbeitsplatzwechsel, konnte dieser einen Schlichtungsausschuss anrufen, der bei Vorliegen eines wichtigen Grundes eine Wechselbescheinigung ausstellen konnte. Als solcher Grund galt nach § 9 HDG „eine angemessene Verbesserung der Arbeitsbedingungen“. Zudem war bei einer Hilfsdienstverpflichtung nach Möglichkeit auf das 17

Schreiben des Chefs des Generalstabs des Feldheeres an den Kriegsminister vom 31. August 1916, abgedruckt in Ludendorff (1922, 64). 18 Schreiben des Chefs des Generalstabs des Feldheeres an den Reichskanzler vom 7.10.1916, abgedruckt in Ludendorff (1922, 77). 19 RGBl 1916, 1333. 212

Lebensalter, die Familienverhältnisse, den Wohnort, die Gesundheit und die bisherige berufliche Tätigkeit Rücksicht zu nehmen und es galt zu prüfen „ob der in Aussicht gestellte Arbeitslohn dem Beschäftigten und etwa zu versorgenden Angehörigen ausreichenden Unterhalt ermöglicht“ (§ 8 HDG). Diese Zugeständnisse bildeten nicht nur ein „Sicherheitsventil“ (Sinzheimer 1920, 37), sondern durch das HDG wurden erstmals die Gewerkschaften durch den Staat anerkannt. Das Hilfsdienstgesetz gilt daher zu Recht als „wichtigstes sozialpolitisches Gesetz der Kriegszeit“ (Preller 1978, 41), als „Wende in der Sozialgeschichte Deutschlands“ (Feldman 1985, 205) und als „Erfolg des Korporatismus, ja eine Revolution in der deutschen Sozialpolitik“ (Herbert 2017, 146). Schon während der Beratungen des HDG stellte der preußische Innenminister von Loebell fest, dass das HDG in der parlamentarischen Geschichte das erste Gesetz sei, dass „so schwerwiegende Änderungen in den vorliegenden sozialpolitischen Verhältnissen herbeiführe […]. Im Frieden wäre es einfach undenkbar, ein Gesetz mit solchen Eingriffen […] anzunehmen“. Nach Meinung von Staatsminister Karl Helfferich hätten gar „Sozialdemokraten, Polen, Elsässer und die Arbeitersekretäre das Gesetz gemacht“.20 In der Donaumonarchie kam es zu einer analogen Entwicklung. Zwar war hier mit dem Kriegsleistungsgesetz die Militarisierung des Arbeitsmarktes schon zu Kriegsbeginn erfolgt, aber die Angst vor Streiks in militärisch relevanten Branchen und die Unzufriedenheit der Arbeiterschaft erforderten auch hier Zugeständnisse. Im März 1917 wurden so genannte Beschwerdekommissionen eingerichtet.21 Ihre Aufgabe bestand darin, angemessene Löhne und Arbeitsbedingungen für Beschäftigte in der Kriegswirtschaft festzusetzen. Die Löhne sollten nicht unter jenen in der Privatwirtschaft liegen und auf die Ausbildung und die Familiensituation der Arbeiter Rücksicht nehmen. Gleichzeitig blieb die Arbeitnehmerfreizügigkeit erheblich beschränkt. Die Kommissionen setzten sich aus einem Vorsitzenden (meist einem hochrangigen Offizier), einem Richter, einem Beamten und je einen Vertreter der Arbeitgeber und Arbeitnehmer zusammen. Die Entscheidungen erfolgten mit Mehrheitsbeschluss und waren verbindlich. Mit diesen Beschwerdekommissionen wurden die Gewerkschaften erstmals als legitime Interessenvertreter der Arbeiterschaft anerkannt (Tálos 1981, 121; Grandner 1992, 379). Angesichts der langen Kriegsdauer und zunehmend desaströsen Versorgungslage rückte in den Achsenmächten die Bevölkerungspolitik in den Fokus. Der moderne Maschinenkrieg führte zu dramatischen Bevölkerungsverlusten, während Lebensmittelengpässe in den letzten beiden Kriegsjahren in Hunger umschlugen und Hunderttausenden das Leben kosteten. Der Tod von Millionen (zeugungsfähiger) Männer, kriegsbedingt sinkende Geburtenziffern, hohe Säuglingssterblichkeit und die Mangelernährung von Kindern und Jugendlichen bedrohten die (künftige) Wehrkraft und riefen das Militär auf den Plan. Tatsächlich war in Deutschland 20

Auszug aus dem Protokoll der Sitzung des preußischen Staatsministeriums, Beratung über das Ergebnis der zweiten Lesung des Hilfsdienstgesetzes im Reichstag, abgedruckt in Deist (1970, 527, 531). 21 RGBl 122/1917. 213

angesichts der langen Kriegsdauer das Potenzial an mobilisierbaren Soldaten weitgehend ausgeschöpft, zumal im Verlauf des Krieges fast 14 Millionen Soldaten, fast 85% aller Männer im wehrfähigen Alter (Herbert 2017, 136), eingezogen wurden. Bereits in einem Schreiben an den Kriegsminister vom 31.8.1916 bemerkte die 3. OHL, dass die Gegner über „ein fast unerschöpfliches Menschenmaterial“ verfügen, während „unsere Menschenvorräte beschränkt“ sind.22 Auf Anregung der OHL verfasste der Generalstabsarzt der Armee, Otto von Schjerning, eine Denkschrift über Bevölkerungspolitik.23 Sie ist insofern bemerkenswert, als Sozialpolitik für einen militärisch motivierten Pronatalismus und zur „Wiederherstellung und Hebung der deutschen Volks- und Wehrkraft“ instrumentalisiert wird. Die darin vorgeschlagenen Maßnahmen kommen einem sozialpolitischen Aktionsprogramm gleich und reichten vom Wohn- und Städtebau, Hygiene- und Ernährungsfragen, Arbeitsschutz, über die Familienpolitik bis hin zur Steuerpolitik. Zuvor hatten schon bürgerliche Sozialreformer und Sozialpolitikpraktiker im Kontext des Massenkriegs den militärischen Wert der Sozialpolitik klar erkannt. Der Präsident des Reichsversicherungsamts erkannte in ihr „eine Quelle deutscher Kriegsbereitschaft“ (Kaufmann 1914, 129-131). Der Schutz des „kindlichen Nachwuchs, dieses stärksten Quelles der Wehrkraft des Volkes“ (Zimmermann 1915, 18), habe ein „numerisch starkes Heer“ ermöglicht (Zahn 1917, 75) und die Sozialpolitik steigere nicht nur die physische Volks- und Wehrkraft, sondern auch die „psychische und moralische Beschaffenheit“ der Truppen (Zahn 1915, 100), die „politisch-rechtliche Kulturstellung“ und die Loyalität „des deutschen Arbeiters“ (Zimmermann 1915, 29). In der Realität fehlten aber in den letzten Kriegsjahren die finanziellen Mittel für eine kompensatorische Sozialpolitik (Herbert 2017), die für eine Stabilisierung der inneren Verhältnisse notwendig gewesen wäre. Stattdessen wurde die Sozialpolitik zu Zwecken der Kriegspropaganda instrumentalisiert. Einige Militärbehörden verwiesen im vaterländischen Unterricht auf die sozial- und bildungspolitischen Errungenschaften des Kaiserreichs zur Hebung und Stärkung der Kampfmoral der Truppen. So heißt es in einem Erlass des stellvertretenden Generalkommandos des XIX. Armeekorps aus dem Jahr 1917: „Die ganze soziale Gesetzgebung in ihren Grundlinien ist darzulegen. Es ist hinzuweisen auf die Millionensummen, die bisher von der Sozialversicherung an die Arbeiter gezahlt worden sind und täglich gezahlt werden. Vor allem ist hier zu betonen, dass die Zuschüsse, die das Reich zur Sozialversicherung beisteuert, ganz außerordentlich hohe sind. Hier ist nachdrücklich hervorzuheben, dass diese Leistungen Deutschlands für seine Arbeiter einzig in der Welt dastehen“.24 22

Ludendorff (1922, 63). Denkschrift der Obersten Heeresleitung über die deutsche Volks- und Wehrkraft, BArch-Militärarchiv PH 3/446, 1-52. 24 Auszug aus den Richtlinien für den „Vaterländischen Unterricht“ im Bereich des stellv. Generalkommandos des XIX. Armeekorps vom 15.12.1917, abgedruckt in Deist (1970, 892). 23

214

Die Situation spitzte sich mit der Russischen Revolution und dem Januarstreik von 1918 nochmals zu, denn die Oktoberrevolution veranschaulichte die Option einer gewaltsamen Erringung von Arbeiterrechten und verstärkte die Radikalisierung der Arbeiterschaft. Die politischen Konzessionen an die Arbeiterbewegung waren eher symbolischer Natur, aber durchaus von erheblicher Langfristwirkung. So erfolgte Ende Mai 1918 mit der Aufhebung des § 153 der Gewerbeordnung25 ein weiteres Zugeständnis an die Gewerkschaften, das ihnen nun faktisch das volle Koalitionsrecht einräumte. Die kriegsinduzierte Teuerung, die geringe Bautätigkeit im Krieg, höherer Bedarf nach Wohnraum durch Kriegsflüchtlinge und schließlich die militärische Demobilisierung führten auch zu ersten Staatseingriffen im Wohnungswesen.26 Hinzu kam der Schutz der Familien von eingezogenen oder gefallenen Soldaten, die mit der Einberufung oder Tod des Familienernährers die Mieten nicht mehr zahlen konnten. Bereits Ende 1914 wurden auf Verordnungsweg auf kommunaler Ebene Mieteinigungsämter geschaffen, 1915 der Mieterschutz für Hinterbliebene von Kriegsteilnehmern27 verbessert und 1916 Mieten begrenzt. Mit der „Verordnung zum Schutz der Mieter“ vom 26. Juli 1917 und einer Revision am 23. September 1918 wurde der Kündigungsschutz weiter ausgeweitet. Wiederum spielte der Druck der Militärbehörden (bis hin zur OHL) eine bedeutsame Rolle (Ludendorff 1922, 253-254). Durch Allerhöchsten Erlass von Kaiser Wilhelm II. vom 4. Oktober 191828 wurde ein Reichsarbeitsamt eingerichtet, das die sozialpolitischen Kompetenzen des Reichswirtschaftsamtes bündelte und unter der Leitung eines Staatssekretärs dem Reichskanzler unterstellt war. Dies war für die weitere Sozialpolitikentwicklung insofern ein wichtiger Schritt, als 1919 daraus mit dem Reichsarbeitsministerium das erste Sozialministerium Deutschlands hervorging, das fortan für praktisch alle Belange der Sozialpolitik verantwortlich war. Auch in der Donaumonarchie beschleunigten die hohe Zahl der Kriegsopfer, Hunger und Inflation, die auch die Mittelschicht hart traf (Redlich 1925, 168, 237) und zum Teil zu Kostgängern des Staates machte, die Unzufriedenheit. Auch hier galt es, die Loyalität und Kooperation der Arbeiterbewegung zu sichern und der Sozialpolitik kam dabei, oft mit Rückendeckung oder gar auf Initiative der Heeresverwaltung, eine wichtige Rolle zu (Grandner 1992, 310). Druck kam auch von Gewerkschaften, die 1917 zwei sozialpolitische Forderungsprogramme vorlegten (Grandner 1992, 341). Wie in Deutschland stiegen auch im Militär die Bedenken, dass die Situation im Hinterland die Kampfbereitschaft und Loyalität der Truppen mindern würde. Bereits im Oktober 1915 hatte das Armeeoberkommando in einer Note an die Regierung die Versorgungslage kritisiert und vor negativen Auswirkungen auf die Kampfbereitschaft der Truppen und den sozialen Frieden im Hinterland gewarnt. Zur 25

RGBl 1918, 423. Vor Kriegsausbruch enthielten weder das BGB (Deutschland) noch das ABGB (Österreich) nennenswerte Mieterschutzbestimmungen. 27 Verordnung über das Kündigungsrecht der Hinterbliebenen von Kriegsteilnehmern (RGBl 1915, 642). 28 RGBl 1918, 1231. 26

215

Abwendung von Streiks in der Schwer- und Rüstungsindustrie setzte sich die Militärführung auch für Lohnsteigerungen ein (Führ 1968, 123, 125). Die Heeresverwaltung griff zur Vermeidung von Streiks aber auch auf Repression zurück. So kam es im Juli 1916 zur Militarisierung kriegswichtiger Industriebetriebe, indem die Arbeiter militärisch vereidigt und militärischer Disziplinarregeln unterstellt wurden (Rauchensteiner 2013; Grandner 1992, 267268). Die Erosionstendenzen im Hinterland erhöhten den Reformdruck und der politische Wille dazu wuchs nach dem Thronwechsel. Wie hoch dieser Druck war, geht aus einem Brief von Außenminister Czernin an den ungarischen Ministerpräsidenten Istvan Tisza von Anfang 1917 hervor: „Sozialpolitik ist das Ventil, das wir öffnen müssen, um den überschüssigen Dampf hinauszulassen – sonst explodiert der Kasten“ (Grandner 1992, 309). Die Regierung unter Ministerpräsident Clam-Martinic stellte Ende 1916 soziale Reformen in Aussicht, die im Kabinett bezeichnenderweise am stärksten vom Landesverteidigungsminister unterstützt wurden (Rauchensteiner 2013, 744). Mit der Wiedereinberufung des Reichsrates im Mai 1917 sollten nicht zuletzt Ernährungsfragen und soziale Fragen vordringlich behandelt werden (Redlich 1925, 263). Tatsächlich nahm die Sozialpolitik in der Thronrede des jungen Kaisers anlässlich der Eröffnungssitzung des Reichsrats gewichtigen Raum ein. Die Bewahrung und Wiederherstellung der Volkskraft durch „zielbewusste Bevölkerungspolitik“ bildete das zentrale Motiv für sozialpolitische Reformen. Kaiser Karl kündigte Maßnahmen zur Eindämmung der Säuglingssterblichkeit, zur Bekämpfung von hygienischen Missständen und weitverbreiteter Krankheiten ebenso an wie Verbesserungen beim Arbeitsschutz für Frauen und Jugendliche, im Bereich der Jugendwohlfahrt, Sozialversicherung und im Wohnungswesen. Schließlich sollte der Sozialschutz für invalide Veteranen und ihre Angehörigen verbessert werden29. Zur Bearbeitung dieser Aufgaben und für die Übergangsplanung in die Friedenswirtschaft wurden neue Verwaltungsstrukturen aufgebaut. Mit Allerhöchstem Handschreiben des Kaisers wurde Ende 1917 ein Sozialministerium30 und 1918 ein Ministerium für Volksgesundheit31 etabliert. Die Sozialpolitik der letzten beiden Kriegsjahre folgte der in der Thronrede skizzierten Stoßrichtung. Zu den Maßnahmen gehörten u.a. die Erhöhung des Unterhalts für die Angehörigen und Hinterbliebenen eingezogener Soldaten zur Abfederung der Inflation32 und an Staatsbürgerschaft gekoppelte Fürsorgeleistungen für Kriegsopfer.33 In der Krankenversicherung34 wurde das Krankengeld für Wöchnerinnen von 4 auf 6 Wochen erweitert. Neu waren Stillprämien in Höhe der Hälfte des Krankengelds, die für 12 Wochen gewährt wurden. Das Krankengeld wurden erhöht und die maximale Bezugsdauer von 20 auf 29

1 der Beilagen zu den stenographischen Protokollen des Abgeordnetenhauses, XXII. Session, 1917, 3-4. „Ministerium für soziale Fürsorge“, RGBl 1917/499. RGBl 1918/297. 32 RGBl 1917/313 und RGBl 1918/126. 33 RGBl 119/1918. 34 Kaiserliche Verordnung (RGBl. 6/1917), später in Gesetz überführt (RGBl 457/1917). 30 31

216

26 Wochen verlängert. Durch Statut konnten die Kassen höhere Leistungen ausrichten sowie eine Familienversicherung für Versicherte mit geringem Einkommen einführen. Verbesserungen gab es im August 1917 auch in der Unfallversicherung.35 Sie erfasste nun auch Wegeunfälle, der Lohnersatz bei dauernder Invalidität wurde von 60 auf 66% erhöht, während Bedürftige bis zu 150% der Unfallrente beziehen konnte. Die Leistungen wurden an die Inflation angepasst und die Finanzierung wurde ausschließlich den Arbeitgebern überantwortet. Letzteres galt auch für die Unfallversicherung der Bergarbeiter,36 wo zudem der Versicherungsschutz ausgedehnt wurde. Mit einer Revision der Gewerbeordnung37 wurde das Beschäftigungsverbot für Wöchnerinnen von vier auf sechs Wochen nach der Geburt erhöht. Das Kriegsministerium forcierte mit einem Erlass aus dem Jahr 1917 die Einrichtung von Fabrikausschüssen und Vertrauensleuten. Sie bestanden aus gewählten Arbeitervertretern und sollten als Bindeglied zwischen Arbeiterschaft und Arbeitgebern fungieren. Erneut ging es um die Befriedung der Arbeitsbeziehungen und das Kriegsministerium räumte den Vertrauensleuten auch eine Sonderstellung ein (Grandner 1992, 298-299). Wie im Deutschen Kaiserreich wurde mit vier Verordnungen aus den Jahren 1917 und 1918 der Mieterschutzschutz verbessert, indem Mieterhöhungen begrenzt und der Kündigungsschutz von Mietern erweitert wurde. Neu geschaffene Mietämter sollten im Konfliktfall schlichten (Brügel 1919). Andere sozialpolitische Initiativen wie z.B. die Arbeitszeitbeschränkungen für Jugendliche und Frauen zogen sich hin und konnten durch den Zusammenbruch der Monarchie nicht mehr rechtzeitig behandelt werden. Wichtige Vorarbeiten konnten noch auf dem Gebiet der Heimarbeit und im Zuge der Demobilisierungsplanung auf dem Gebiert der Arbeitsvermittlung erledigt werden. Diese Maßnahmen blieben praktisch im Krieg zwar ohne größere Bedeutung, die entsprechenden Vorarbeiten beschleunigten aber nach Kriegsende die Sozialgesetzgebung der Ersten Republik. Insgesamt reichten aber wie in Deutschland diese Maßnahmen nicht aus, um die soziale Lage spürbar zu verbessern, sondern es handelte sich um den sprichwörtlichen Tropfen sozialen Öls, der noch dazu in erster Linie die Kriegsmaschinerie schmieren sollte. Diese Maßnahmen sollten primär das „Durchhalten“ sichern, brachten aber keine entscheidenden materiellen Verbesserungen. Der soziale und wirtschaftliche Niedergang und die Erosion des Vielvölkerstaates schritten daher unablässig voran, während die Oktoberrevolution Teile der Arbeiterschaft weiter radikalisierte. 1918 kam es zu hunderten Streiks, an denen fast eine halbe Million Beschäftigte teilnahmen. In einem letzten Akt der Verzweiflung erwog die Regierung die Einführung einer allgemeine Arbeitspflicht für Männer und Frauen und damit die totale Mobilisierung der Zivilbevölkerung,38 der Entwurf trat allerdings nicht mehr in Kraft. Im 35

RGBl 363/1917. RGBl 1917/523. 37 Kaiserliche Verordnung (RGBl 7/1917), später in Gesetz überführt (RGBl 475/1917). 38 1050 der Beilagen zu den stenographischen Protokollen des Abgeordnetenhauses, XXII Session 1918, 2. 36

217

Herbst 1918 scheiterte eine letzte Offensive in Italien, ehe im November die Militärmaschinerie implodierte und das Vielvölkerreich entlang seiner nationalen Bruchstellen auseinanderbrach. 3.2

Schweiz

Die Schweiz war zwar nicht in Kampfhandlungen involviert, dennoch war sie durch die Mobilisierung der Armee zur Grenzsicherung und als kleine offene Volkswirtschaft vom Krieg betroffen, da der Handel gestört wurde. Wie alle Länder war auch die Schweiz nicht auf einen langen Krieg vorbereitet und wie überall gab es Kriegsgewinner und vor allem -verlierer (Rossfeld et al. 2014). Der Bund verfügte zu Kriegsbeginn weder auf dem Gebiet der Arbeitslosenversicherung noch auf dem Gebiet der Alterssicherung über Gesetzgebungskompetenzen. Traditionell schwach war der Bund auch in der Fiskalpolitik. Der Löwenanteil seiner Einnahmen bestand aus Zöllen, die durch den kriegsbedingten Handelseinbruch aber stark rückläufig waren. Wegen dieser Einnahmenausfälle und kriegsbedingter Mehrkosten nahm die Schweiz Anleihen auf und führte eine außerordentliche Kriegssteuer39 sowie eine Kriegsgewinnsteuer ein. Diese Steuern waren aber zeitlich befristet, wobei eine 1921 eingeführte dritte Kriegssteuer erst 1932 abgeschafft wurde. Angesichts dieser Kompetenzlage waren es primär die Kantone und Gemeinden, die die sozialen Kriegsfolgen bewältigen mussten (Studer 2012, 935). Einige Kantone machten von ihrer bundesgesetzlichen Kompetenz Gebrauch, die Krankenversicherung für obligatorisch zu erklären (Walter 2018), während Glarus 1916 eine obligatorische Alterssicherung einführte. Dort wo der Bund, wie in der Fabrikgesetzgebung, Kompetenzen besaß, kam es zu Einschnitten. So beschloss der Bundesrat mit Kriegsausbruch, die erst im Juni 1914 beschlossene Arbeitszeitverkürzung zu suspendieren. Erst 1917 trat auf Druck der Gewerkschaften der 10-Stunden-Tag in Kraft (Degen 1983, 330). Wie in den Mittelmächten erlaubte der Bundesrat per Kreisschreiben vom 11. August 1914 Ausnahmen von den Arbeitsschutzbestimmungen des Fabrikgesetzes hinsichtlich Arbeitszeit, Nacht- und Sonntagsarbeit und der Beschäftigung von Frauen und Jugendlichen.40 Zu den kriegsbedingten sozialpolitischen Herausforderungen zählte v.a. die Absicherung der über 200.000 mobilisierten Wehrmänner41 und ihrer Angehörigen, zumal eine Verdienstausfallentschädigung nicht existierte und der ohnedies geringe Sold noch durch die kriegsbedingte Teuerung entwertet wurde. Ein erheblicher Teil der Soldatenfamilien war daher auf Notstandsunterstützung sowie private Wohlfahrtseinrichtungen angewiesen, die im Krieg erheblich an Bedeutung gewannen und maßgeblich von Frauen getragen wurden (Joris und Schumacher 2014). 39

Die Steuer wurde 1915 mit einer Rekordzustimmung von 94% vom männlichen Schweizervolk angenommen und gilt als erste direkte Bundessteuer. 40 Bundesblatt 1914 IV, 37-38. 41 Die Zahl der zur Grenzsicherung eingesetzten Soldaten war im Zeitverlauf stark rückläufig. 218

Der Bund musste angesichts fehlender sozialpolitischer Kompetenzen im Bereich der Einkommenssicherung notgedrungen andere Akzente setzen. Konkret nutzte der Bundesrat einen Teil der Kriegssteuereinnahmen zur Subventionierung privater Fürsorgeeinrichtungen wie den „Schweizer Verband Soldatenwohl“, der bedürftige Familien mobilisierter Soldaten unterstütze. Auch der private Verein „Für das Alter“ (heute „Pro Senectute“), 1917 als Hilfswerk für bedürftige Greise gegründet, erhielt kantonale und ab 1929 Subventionen des Bundes (Leimgruber 2018). Zudem erhob der Bund seit 1917 einen Zuschlag auf die Kriegsgewinnsteuer und richtete einen Arbeitslosenfürsorgefonds ein, aus dessen Mitteln die bestehenden privaten und dezentralen öffentlichen Arbeitslosenkassen rückwirkend Subventionen erhielten. Schließlich gewährte der Bund Steuervergünstigungen an Unternehmen für die Bereitstellung betrieblicher Sozialleistungen (Leimgruber 2008, 64-82). Als Folge entstanden hunderte betriebliche Pensionspläne, die insofern auf Unterstützung der Unternehmen stießen, als sie in Zeiten des Arbeitskräftemangels qualifizierte Arbeitskräfte an den Betrieb binden konnten. Auch betriebliche Wohlfahrtseinrichtungen gewannen an Bedeutung, während die schweizerischen Lebensversicherer von den Turbulenzen in den kriegsgebeutelten Ländern profitierten und ihren heimischen Marktanteil von 43% (1910) auf 95% (1925) erhöhen konnten (Leimgruber 2008, 78 Fn. 44). Weiterführende Maßnahmen auf dem Gebiet der staatlichen Sozialpolitik blieben aus, sodass im Krieg eine „unerträgliche sozialpolitische Lähmung“ (Schweingruber 1977, 52) herrschte. 1916 endete auch in der Schweiz der Burgfrieden und in den letzten beiden Kriegsjahren kam es zu zahlreichen Streiks (Degen 2012, 886) und schweren sozialen Spannungen, die sich durch Lebensmittelengpässe und Lebensmittelrationierungen in den Städten weiter zuspitzten. 1918 bezog ein Sechstel der Bevölkerung Notstandsunterstützung (Tanner 2015, 136), wobei die Quote in den Städten deutlich höher lag. Im gleichen Jahr mobilisierten Gewerkschaften und Sozialdemokraten gegen den Plan des Bundesrates, eine Hilfs- und Zivildienstpflicht für die männliche Bevölkerung einzuführen (Tanner 2015, 147). Das aus diesem Widerstand hervorgehende Oltener Aktionskomitee rief nach der militärischen Besetzung Zürichs im November 1918 zum Generalstreik auf und richtete ein Forderungsprogramm an den Bundesrat, das neben dem Frauenwahlrecht auch sozialpolitische Reformen (Einführung des Achtstundentags und einer staatlichen Alterssicherung) enthielt. Der Generalstreik brach letztlich unter massiven militärischen Druck zusammen.

4

Sozialpolitik in der Nachkriegszeit

4.1

Deutschland und Österreich

Der Zusammenbruch der Monarchie samt Delegitimierung der alten Eliten, die reelle Gefahr einer Revolution, kriegsinduzierte soziale Not, die Pandemie der Spanischen Grippe, hohe Arbeitslosigkeit durch Demobilisierung und Abwicklung der Rüstungsproduktion erzeugten in 219

Deutschland und Österreich ungeheuren Handlungsdruck zur Stabilisierung der prekären politischen, ökonomischen und sozialen Verhältnisse. Die ersten demokratischen Wahlen auf Basis des Verhältniswahlrechts brachten erstmals rot-schwarze Koalitionen an die Regierung, während der Machtgewinn der Gewerkschaften und die Revolutionsängste der Unternehmen zu einer neuen Machtbalance im Bereich der industriellen Beziehungen führten. Das Zusammenwirken dieser Faktoren eröffnete ein (kurzes) Gelegenheitsfenster für progressive Sozialreform, ehe ab den 1920er Jahren die politische Rechte wieder die Oberhand gewann und die Weltwirtschaftskrise und die damit verbundene politische Radikalisierung zunächst expansiver Sozialpolitik und schließlich der Demokratie ein Ende setzten. In Deutschland wurden am 12. November 1918 die Ausnahmebestimmungen zum Arbeitsschutz aufgehoben. Durch Verordnung des Reichsamts für wirtschaftliche Demobilmachung vom 13. November wurde eine aus öffentlichen Mitteln finanzierte Erwerbslosenfürsorge geschaffen. Die Gemeinden wurden verpflichtet, bedürftige, aber arbeitswillige Personen, die als Folge des Kriegs erwerbslos wurden, zu unterstützen. Zwei Tage später kam es zum denkwürdigen Legien-Stinnes Abkommen, das ohne die Revolutionsund Sozialisierungsängste der Unternehmerverbände (Paster 2013, 427-430) wohl keine Chancen gehabt hätte. Die Gewerkschaften wurden nun auch von den Arbeitgebern als Tarifpartner anerkannt und Kernelemente des Abkommens erfuhren in der Folgezeit eine gesetzliche Regelung. Dazu gehörten der Achtstundentag für gewerbliche Arbeiter (1918) und Angestellte (1919), die Ende 1918 erfolgte gesetzliche Regelung von Tarifverträgen, die u.a. dem Reichsarbeitsamt die Möglichkeit einräumte, Tarifverträge für allgemeinverbindlich zu erklären,42 die Koalitionsfreiheit und das Betriebsrätegesetz (1920), das auch wichtige Kündigungsbeschränkungen enthielt. Einige der im Krieg durch den Rohstoffmangel bedingten Arbeitsschutzbestimmungen wurden verstetigt. Dies betraf das Nachtbackverbot und die 19h-Schließzeiten von Geschäften. Ebenso wurden die provisorischen Bestimmungen zum Mieterschutz mit dem Reichsmietengesetz (1922) und dem Mieterschutzgesetz (1923) auf Dauer gestellt. Die Miete einer Wohnung am 1. Juli 1914 („Friedensmiete“) diente fortan als Berechnungsbasis für eine zulässige ortsübliche Miete. Auch die im Krieg 1915 eingeführte Reichswochenhilfe wurde nach Kriegsende verbessert und 1926 neu geregelt. Nachdem die Erwerbslosenfürsorge ab 1919 um eine produktive Arbeitslosenfürsorge – ein Aktivierungsprogramm mittels Notstandsarbeiten – ergänzt wurde, wurden ab 1923 zur Finanzierung der Erwerbslosenfürsorge von den Arbeitnehmern und -gebern Beiträge eingehoben. Von hier war es nur ein kleiner Schritt zur Einführung der Arbeitslosenversicherung im Jahr 1927, die auch eine Neuordnung der Arbeitsvermittlung mit sich brachte. Hohe Priorität nahm die Versorgung der rund vier Millionen Kriegsopfer ein. Während das Stinnes-Legien-Abkommen ein Rückkehrrecht von Kriegsteilnehmern auf ihren 42

RGBl 1918, 1456.

220

ursprünglichen Arbeitsplatz vorsah, war bei den 1,5 Millionen Kriegsinvaliden (Pironti 2015, 19) enormer Handlungsbedarf gegeben. Aufgrund des Gesetzes über die Beschäftigung Schwerbeschädigter (1920) mussten in Betrieben mit über 20 Beschäftigten zwei Prozent der Arbeitsplätze mit Kriegsinvaliden mit einer Erwerbsminderung von mindestens 50% besetzt werden. Die materielle Versorgung der Kriegsbeschädigten und Kriegshinterbliebenen erfolgte durch das Reichsversorgungsgesetz (1920). Die Renten bemaßen sich nach Erwerbseinschränkung, dem früheren Beruf und dem Wohnort. Hinzu kam ein Anspruch auf Heilbehandlung, Hilfsmittel (Prothesen, Blindenhund) und Pflegezulage bei Pflegebedürftigkeit. Die Hinterbliebenen erhielten davon abgeleitete Renten und auch Sterbegeld. Auf dem Gebiet der Sozialversicherung kam es hingegen nur zu wenigen Veränderungen. Im Mittelpunkt stand die Bewältigung der Hyperinflation, die neben Maßnahmen zum Teuerungsausgleich vorübergehend auch den Wechsel zum Umlageverfahren notwendig machte (Frerich und Frey 1993). Auch in Österreich kam es zwischen 1918 und Anfang der 1920er Jahre zu einem Quantensprung in der Sozialpolitik. Zu den wichtigsten Maßnahmen zählten die Einführung des Achtstundentags (Dezember 1918 und 1919),43 das Betriebsrätegesetz (Mai 1919),44 Gesetz über Einigungsämter und Kollektivverträge (Dezember 1919),45 Heimarbeitergesetz,46 das Kinderarbeitsgesetz47 (jeweils Dezember 1919), das Verbot Sonn- und Feiertagsarbeit, das Nachtarbeitsverbot für Bäckereien und Fleischereien,

das das der der

Erholungsurlaub für Arbeiter (1919), der Kündigungsschutz für Angestellte und die Einführung von Arbeiterkammern (1920). Im Bereich der Sozialversicherung gab es zwei Neuerungen. Die im Zuge der Demobilisierung (Pribram 1920, 631) erlassenen zahlreichen Verordnungen zur steuerfinanzierten Arbeitslosenfürsorge mündeten 1920 in der Einführung einer Arbeitslosenversicherung,48 die 1922 um eine Anschlussleistung (Notstandsaushilfe) und – wie in Deutschland – eine Art aktiver Arbeitsmarktpolitik (produktive Arbeitslosenfürsorge) ergänzt wurde. 1920 wurde die Krankenversicherung für Staatsbedienstete eingeführt. Die Alters- und Invalidenversicherung für Arbeiter kam hingegen über Gesetzesentwürfe nicht hinaus, während die Hyperinflation in anderen Sozialversicherungszweigen den Wechsel zum Umlageverfahren erforderte (Hofmeister 1981, 629). Weitgehend ähnliche Regelungen wie im deutschen Reichsmietengesetz fanden sich im österreichischen „Mietengesetz“ vom Dezember 1922, das neben dem „Friedenszins“ zur Richtwertbemessung (Stichtag 1. August 1914) neue Kündigungsbeschränkungen gesetzlich verankerte. 43

StGBl 138/1918 und StGBl 581/1919. StGBl 283/1919. StGBl 16/1920. 46 StGBl 140/1919. 47 StGBl 141/1919. 48 StGBl 153/1920. 44 45

221

Hohe Priorität nahm die Versorgung der ca. 450,000 Kriegsopfer (Kriegsinvalide und Hinterbliebene) auf dem Gebiet der Republik ein. Das Invalidenentschädigungsgesetz49 (1919) gewährte invaliden Veteranen und Kriegsleistungsbeschäftigten Invalidenrenten, medizinische Versorgung, Krankengeld, Prothesen und unterstützte Berufsausbildung. Im Todesfall hatten Hinterbliebene Anspruch an Hinterbliebenenrenten und Sterbegeld. Das Programm war steuerfinanziert und basierte auf Staatsbürgerschaft. Die Kopplung eines Leistungsanspruchs an eine staatlich oktroyierte Pflicht verankerte ein neues Anspruchsprinzip (Pawlowsky und Wendelin 2015, 61). Das Invalidenbeschäftigungsgesetz50 verpflichtete Betriebe, einen „Kriegsbeschädigten“ pro 20 Beschäftigte und einen weiteren für die nächsten 25 Beschäftigten einzustellen. Alternativ musste der Arbeitgeber eine Ausgleichstaxe in einen Fonds des Sozialministeriums zahlen, aus dem die Kriegsopferunterstützung finanziert wurde. Eine Vollzugsanweisung des Sozialministeriums verpflichtete Gewerbebetriebe mit über 15 Mitarbeitern, die Beschäftigtenzahl um 20% zu gleichen Lohn- und Arbeitsbedingungen zu erhöhen.51 Unternehmen mussten auch Angestellte nach abgeleistetem Militärdienst wieder einstellen, während Kündigungen beschränkt wurden. 4.2

Schweiz

In der Schweiz wurde von den Forderungen des Landesstreiks nur der Achtstundentag in der Industrie unmittelbar Realität. Anders als in den beiden Nachbarländern waren die Sozialdemokraten nach Kriegsende nicht an der Regierung beteiligt, wenngleich sie aus den Nationalratswahlen 1919 gestärkt hervorgingen. Nach Kriegsende wurden die Kriegsvollmachten des Bundesrates aufgehoben, sodass die sozialpolitische Bearbeitung der Kriegsfolgen wieder in dem durch Föderalismus und Referendumsdemokratie abgesteckten Handlungskorridor erfolgte. Zwar nahmen auch in der Schweiz nach Kriegsende die sozialpolitischen Reformaktivitäten zu (Lengwiler und Leimgruber 2018), allerdings trugen die institutionellen Vetopunkte wie vor dem Krieg zur Blockade oder Verzögerung staatlicher Sozialpolitik bei. Zwei Beispiele sind hier illustrativ.52 1920 scheiterte das „Bundesgesetz betreffend die Ordnung des Arbeitsverhältnisses“ vom Juni 1919 an der Stimmurne, das eine grundlegende Revision der Regelung der Arbeitsverhältnisse (z.B. Allgemeinverbindlichkeitserklärung von Gesamtarbeitsverträgen) mit sich gebracht hätte (Emmenegger 2014). Das gleiche Schicksal erlitt 1931 die Alters- und Hinterlassenenversicherung (AHV). Unter dem Eindruck des Ersten Weltkriegs und des Landesstreiks (Armingeon 2018, 48) nahm der Bundesrat die Einführung einer Invaliditäts-, 49

StGBl 245/1919. Das Gesetz wurde 1922 grundlegend revidiert. StGBl 459/1920. StGBl 268/1919. 52 In dieses Muster passt auf der Einnahmenseite die Ende 1922 gescheiterte Vermögensabgabeinitiative. Auch hier handelte es sich um eine Forderung des Oltener Komitees, das die Tilgung der Kriegsschulden durch die Reichen forderte. 50 51

222

Alters- und Hinterlassenenversicherung in Angriff (Bundesblatt 1919 IV, 1ff.), forderte aber dafür neue Steuern, die jedoch zum Teil an der Stimmurne scheiterten. 1925 erhielt zwar der entsprechende Verfassungsartikel die Zustimmung von Volk und Ständen, die prekären Bundesfinanzen führten aber dazu, dass die einfachgesetzliche Regelung („Lex Schulthess“) abgespeckt werden musste (Verzicht auf Invalidenversicherung, Absenkung des Rentenniveaus) und schließlich im Kontext der Großen Depression 1931 an der Stimmurne scheiterte. Neben dem Achtstundentag in der Industrie (Fabrikgesetzrevision 1919) und dem Verfassungsartikel zur AHV (1925) ist das Bundesgesetz zur Arbeitslosenversicherung aus dem Jahr 1924 das einzige sozialpolitische Vermächtnis des Ersten Weltkriegs auf Bundesebene. Letzteres Gesetz verstetigte aber lediglich das im Krieg geschaffene Provisorium der Subventionszahlungen des Bundes an private und öffentliche Kassen. Fehlende Bundeskompetenz und referendumspolitische Bedenken gaben aus Sicht des Bundesrats für diese Lösung den Ausschlag. Der Bundesrat hob zudem die schlanke Verwaltung der Subventionslösung hervor, denn wegen „der Abneigung unseres Volkes gegen die Vermehrung staatlicher Anstalten ist an eine eidgenössische Zwangsversicherung nicht zu denken“ (Bundesblatt 1923 II, 832-833). Die Regierung betonte schließlich, dass „das Vorgehen des Bundes keine abschließende Regelung bedeutet, sondern nur den Anstoß geben soll zu einer weiteren Entwicklung, die in die Hände der Kantone, der Gemeinden sowie der Verbände der Arbeitgeber und Arbeitnehmer gelegt ist“ (Bundesblatt 1923 II, 834-835). Tatsächlich verfügten 1933 bis auf Obwalden alle Kantone über eine freiwillige oder obligatorische Arbeitslosenversicherung (Tschudi 1989, 106) und das Bundesgesetz von 1924 blieb bis in die 1970er Jahre bestehen.

5

Schlussfolgerungen

Der Drei-Ländervergleich enthüllt einige erstaunliche Parallelen. Geradezu frappierend sind die weitgehend identischen sozialpolitischen Maßnahmen, die in Deutschland und Österreich getroffen wurden. Dies deutet auf enorme funktionale Handlungszwänge sowie Politiklernen und -nachahmung hin. Vieles was im Ersten Weltkrieg in den Mittelmächten im Großen stattfand, ist aber im Kleinen oder zumindest in Ansätzen auch in der neutralen Schweiz zu beobachten und zeigt die grenzüberschreitende Wucht eines totalen Kriegs. Die Gemeinsamkeiten betreffen etwa Zwangsbewirtschaftungsmaßnahmen und Rationierungen, die Militarisierung des Arbeitsmarktes durch Kriegsleistungs- und Hilfsdienstgesetze (in der Schweiz jedoch gescheitert), Einschnitte im Arbeiterschutz zu Kriegsbeginn, der Bedeutungszuwachs des Dritten Wohlfahrtssektors sowie die massive Zuspitzung der sozialen Lage, die politische Polarisierung und die Aufwertung der Arbeiterbewegung in den letzten beiden Kriegsjahren. Zudem kam es abgesehen von Einzelmaßnahmen in allen drei Ländern während des Kriegs zu keiner Expansion staatlicher Sozialpolitik, zu viele Ressourcen 223

beanspruchte selbst in der Schweiz der Militäretat. Durch die dramatische Verschlechterung der Lebensverhältnisse großer Teile der Bevölkerung hat sich dadurch ein enormer sozialer Problemdruck aufgestaut, der sich nach Kriegsende unter veränderten Rahmenbedingungen entladen und in Sozialreformen übersetzt hat. Im Hinblick auf die Intensität des Problemdrucks sowie die politischen Rahmenbedingungen zeigen sich jedoch drei gravierende Unterschiede zwischen den kriegsführenden Mittelmächten und der Eidgenossenschaft, die die unterschiedlichen sozialpolitischen Folgen des Kriegs in diesen Ländern maßgeblich erklären. Erstens war in den kriegsführenden Staaten der Handlungsdruck in dreifacher Hinsicht größer: (i) sozial durch Millionen an Kriegsopfern und Demobilmachung; (ii) politisch durch die Transformation und Instabilität des politischen Systems nach Kriegsende; und (iii) wirtschaftlich durch schwere kriegsbedingte makroökonomische Verwerfungen in Gestalt von Massenarbeitslosigkeit und Hyperinflation. Zweitens wurden die Mittelmächte autokratisch regiert und das Militär war ein zentraler innenpolitischer Akteur, der aus militärfunktionalen Erwägungen einige Sozialreformen unterstützte und partiell auch gegen den Widerstand der Arbeitgeber durchsetzte. Durch diese „militärische Sozialpolitik“ (Abelshauser 1987, 15) und die Demobilisierungsplanung erfolgten noch im Krieg bedeutende sozialpolitische Weichenstellungen und Vorarbeiten. Drittens bewirkte der Krieg eine Kräfteverschiebung zugunsten der Arbeiterbewegung, deren gemäßigter Flügel nach Kriegsende mit den Christdemokraten eine Koalition bildete. Das Zusammenwirken all dieser Faktoren führte binnen kurzer Zeit zu einer sprunghaften Entwicklung staatlicher Sozialpolitik, die sich in einem massiven Anstieg der Sozialausgaben widerspiegelte. Beliefen sich im Deutschen Kaiserreich 1913/14 die Sozialausgaben noch auf 4,9% des Volkseinkommens bzw. 26,8% der gesamten öffentlichen Ausgaben, so betrugen diese Anteile 1925/26 bereits 12,5 bzw. 44,3% (Frerich und Frey 1993, 175). Knapp ein Viertel davon war kriegsinduziert (Zöllner 1962, 43). In Österreich beteiligte sich der Staat nach Kriegsende erstmals an der Finanzierung der Sozialversicherung (Tálos 1981), was zusammen mit der Kriegsopferversorgung die Sozialausgaben ebenfalls nach oben trieb. Viele der sozialen Errungenschaften der Weimarer und Ersten Republik waren bereits von der Kriegssozialpolitik und Demobilmachungsplanung vorgezeichnet worden. Vom Hilfsdienstgesetz, den Fabrikausschüssen und Beschwerdekommissionen sowie der massiven Regulierung des Arbeitsmarktes war es nicht mehr weit zu den Betriebsrätegesetzen, Einigungsämtern und der staatlichen Arbeitsvermittlung. Die im Kontext der Demobilisierung eingeführte Erwerbslosenfürsorge mündete 1920 bzw. 1927 in Arbeitslosenversicherungen, während die provisorischen Staatseingriffe im Mietrecht in beiden Staaten ebenso verstetigt wurden wie einige der durch Rohstoffmangel bedingten Arbeitsschutzmaßnahmen. Erleichtert wurde die sozialpolitische Abfederung der Kriegsfolgen durch den Ausbau des staatlichen Verwaltungsapparats (z.B. die Schaffung von Sozialministerien) im Krieg, die umfangreichen Exekutivvollmachten, die im Rahmen der Kriegswirtschaft und Demobilmachung gesammelten Planungs224

und

Steuerungserfahrungen

und

den

korporatistischen

Unterbau

des

„Kriegssozialismus“. In diesen Veränderungen und den ungeheuren kriegsinduzierten sozialen Notlagen liegt der wesentliche Grund hinter der Prellerschen These einer sozialpolitischen Katalysatorwirkung des Ersten Weltkriegs: „Mit Ausnahme des Achtstundentags gibt es keine wesentliche sozialpolitische Neuerung in der Weimarer Republik, die nicht bereits während des Krieges auf der Grundlage des Anspruchprinzips eingeführt worden wäre: Arbeitslosenunterstützung, Kurzarbeitergeld, Kindergeld, Arbeitsvermittlung, ja de facto sogar eine Art Mindestlohn. Nicht die Verlautbarungen des Rates der Volksbeauftragten im November 1918, sondern das Hilfsdienstgesetz sowie das Notstandsrecht des Krieges und der Demobilmachungsplanung wurden zur Grundlage des Weimarer Sozialstaates“ (Mai 1997, 105). Nur wenig von dem kann in der Schweiz beobachtet werden. Zwar war auch in der Schweiz der Erste Weltkrieg ein „sozialpolitischer Umbruchmoment“ (Lengwiler und Leimgruber 2018, 27), der Ausbau staatlicher Sozialpolitik blieb auf nationaler Ebene jedoch im Wesentlichen auf den Achtstundentag beschränkt. Weder wurde ein neuer Sozialversicherungszweig eingerichtet noch das kollektive Arbeitsrecht grundlegend verändert. Der soziale und ökonomische Problemdruck war ungleich geringer und wie vor dem Krieg bildete das Referendum eine mächtige Hürde gegenüber den kriegsinduzierten Reformbestrebungen. Zwar kamen auch in der Schweiz einige tausend Soldaten durch Unfälle und Krankheit (spanische Grippe) im Aktivdienst ums Leben, dennoch war die Versorgung von Kriegsopfern nicht annähernd mit den Lasten zu vergleichen, die die Nachbarländer zu schultern hatten und dort zu einem massiven Anstieg der Sozialausgaben führten. Fortschritte gab es hingegen auf lokaler und kantonaler Ebene sowie im Bereich der (steuerlich geförderten) betrieblichen und privaten Sozialpolitik (Alterssicherung). Die sozialpolitische Landnahme durch dezentrale, private und betriebliche Lösungen und die darin eingelagerten föderalistischen und unternehmerischen Interessen blockierten aber in der Folge die Sozialpolitik des Bundes, wodurch wiederum die Expansion privater Sozialschutzformen weiter beschleunigt wurde (Leimgruber 2008, 2018). So verheerend im wahrsten Sinn des Wortes die kurzfristigen Auswirkungen des Ersten Weltkriegs auf die Bevölkerung waren, so zeigt der Drei-Ländervergleich auch, dass der Erste Weltkrieg langfristig nicht nur ein Schrittmacher für die Entwicklung des Wohlfahrtsstaates war, sondern auch seine Architektur nachhaltig geprägt und verändert hat. Während in den beiden kriegsführenden Staaten die massiven sozialen und wirtschaftlichen Kriegsfolgen die Weichen auf Ausbau staatlicher Sozialpolitik stellten, kam es in der von Kampfhandlungen und Hyperinflation verschonten Eidgenossenschaft zu einer Weichenstellung zugunsten privater und betrieblicher Sozialschutzprogramme, wenngleich mit finanzieller Unterstützung des Staates. Nur deshalb, weil die Schweiz nicht in beide Weltkriege involviert war, blieb dieser für das kriegsgebeutelte Kontinentaleuropa im so genannten Goldenen Zeitalter des Wohlfahrtsstaates so untypische welfare mix bis heute bestehen.

225

Literaturverzeichnis Abelshauser, Werner, Hrsg. 1987. Die Weimarer Republik als Wohlfahrtsstaat. Stuttgart: Franz Steiner Verlag. Adler, Emanuel. 1927. “Das Arbeitsrecht im Kriege.” In Die Regelung der Arbeitsverhältnisse im Kriege, herausgegeben von Ferdinand Hanusch und Emanuel Adler, 19-170. Wien: Hölder-Pichler-Tempsky. Alber, Jens. 1982. Vom Armenhaus zum Wohlfahrtsstaat. Frankfurt/New York: Campus. Armingeon, Klaus. 2001. “Institutionalizing the Swiss Welfare State.” West European Politics 24 (2): 145-168. Armingeon, Klaus. 2018. “Die Entwicklung der schweizerischen Altersvorsorge.” Swiss Political Science Review 24 (1): 43-52. Brügel, Ludwig. 1919. Soziale Gesetzgebung in Österreich von 1848 bis 1918. Eine geschichtliche Darstellung. Wien/Leipzig: Deuticke. Deist, Wilhelm. 1970. Militär und Innenpolitik im Weltkrieg 1914-1918. Zwei Bände. Düsseldorf: Droste. Degen, Bernhard. 1983. “Wie Bundesrat und Parlament einer elfstündigen Arbeitszeitverkürzung zustimmten: die Fabrikgesetz-Revision von 1919.“ Gewerkschaftliche Rundschau 75 (12): 329-337. Degen, Bernhard. 2012. “Arbeit und Kapital.” In Wirtschaftsgeschichte der Schweiz im 20. Jahrhundert, herausgegeben von Patrick Halbeisen, Margrit Müller und Béatrice Veyrassat, 873-922. Basel: Schwabe Verlag. Emmenegger, Patrick. 2014. The Power to Dismiss: Trade Unions and the Regulation of Job Security in Western Europe. Oxford: Oxford University Press. Feldman, Gerald D. 1985. Armee, Industrie und Arbeiterschaft in Deutschland 1914 bis 1918. Berlin/Bonn: J.H.W. Dietz. Frerich, Johannes und Martin Frey. 1993. Handbuch der Geschichte der Sozialpolitik in Deutschland. Bd.1. München: Oldenbourg. Führ, Christoph. 1968. Das k.u.k Armeeoberkommando und die Innenpolitik in Österreich 1914-1917. Graz u.a.: Böhlau. Grandner, Margarete. 1992. Kooperative Gewerkschaftspolitik in der Kriegswirtschaft. Die freien Gewerkschaften Österreichs im ersten Weltkrieg. Wien: Böhlau. Hämmerle, Christa. 2014. Heimat/Front. Geschlechtergeschichte/n des Ersten Weltkriegs in Österreich-Ungarn. Wien: Böhlau. Hanusch, Ferdinand. 1927. “Einleitung.” In Die Regelung der Arbeitsverhältnisse im Kriege, herausgegeben von Ferdinand Hanusch und Emanuel Adler, 3-18. Wien: Hölder-Pichler-Tempsky. Herbert, Ulrich. 2017. Geschichte Deutschlands im 20. Jahrhundert. München. C.H. Beck. Hofmeister, Herbert. 1981. Ein Jahrhundert Sozialversicherung in Österreich. Schriftenreihe für Internationales und Vergleichendes Sozialrecht, Band 6b. Berlin: Duncker & Humblot. Immergut, Ellen M. 1992. Health Politics and Institutions in Western Europe. New York: Cambridge University Press. Joris, Elisabeth und Beatrice Schumacher. 2014. “Helfen macht stark. Dynamik im Wechselspiel von privater Fürsorge und staatlichem Sozialwesen.” In 14/18. Die Schweiz und der Grosse Krieg, herausgegeben von Roman Rossfeld, Thomas Buomberger und Patrick Kury, 316-335. Baden: hier und jetzt. Kaufmann. 1914. “Die Arbeiterfürsorge – eine Quelle deutscher Kriegsbereitschaft.” Monatsblätter für Arbeiterversicherung VIII (10/11): 129-144. Krumbiegel, Kurt. 1913. Die schweizerische Sozialversicherung insbesondere das Kranken- und Unfallversicherungsgesetz vom 13. Juni 1911 verglichen mit der entsprechenden deutschen Gesetzgebung. Jena: Gustav Fischer. Lengwiler, Martin und Matthieu Leimgruber. 2018. “Im Schatten des Landesstreiks: Sozialpolitik nach dem Ersten Weltkrieg.” Soziale Sicherheit/CHSS 3/2018: 27-30. Leimgruber, Matthieu. 2008. Solidarity without the State? Business and the Shaping of the Swiss Welfare State, 1890-2000. Cambridge: Cambridge University Press. Leimgruber, Matthieu. 2018. “War and Social Policy Development in Switzerland.” In Warfare and Welfare, herausgegeben von Herbert Obinger, Klaus Petersen und Peter Starke, 364-392. Oxford: Oxford University Press. Ludendorff, Erich, Hrsg. 1922. Urkunden der Obersten Heeresleitung über ihre Tätigkeit 1916/18. 3. Auflage. Berlin: Mittler & Sohn. Mai, Gunther. 1997. Das Ende des Kaiserreichs. Politik und Kriegsführung im Ersten Weltkrieg. 3. Aufl. München: dtv. Obinger, Herbert. 1998. Politische Institutionen und Sozialpolitik in der Schweiz. Frankfurt u.a.: Lang. Obinger, Herbert, Klaus Petersen und Peter Starke (Hg.). 2018. Warfare and Welfare: Military Conflict and Welfare State Development in Western Countries. Oxford: Oxford University Press. Paster, Thomas. 2013. “Business and Welfare State Development. Why did Employers Accept Social Reforms?” World Politics 65 (3): 416-451. 226

Pawlowsky, Verena und Harald Wendelin. 2015. Die Wunden des Staates. Kriegsopfer und Sozialstaat in Österreich 1914-1938. Wien: Böhlau. Pironti, Pierluigi. 2015. Kriegsopfer und Staat: Sozialpolitik für Invaliden, Witwen und Waisen des Ersten Weltkriegs in Deutschland und Italien (1914–1924). Köln: Böhlau. Preller, Ludwig. 1978 [1947]. Sozialpolitik in der Weimarer Republik. Kronberg/Ts: Athenäum/Droste Pribram, Karl. 1920. “Die Sozialpolitik im neuen Oesterreich.” Archiv für Sozialwissenschaft und Sozialpolitik 48: 615-680. Rauchensteiner, Manfried. 2013. Der Erste Weltkrieg und das Ende der Habsburgermonarchie 1914-1918. Wien: Böhlau. Redlich, Joseph. 1925. Österreichische Regierung und Verwaltung im Weltkriege. Wien: Hölder-Pichler-Temsky. Reidegeld, Eckart. 1989. “Krieg und staatliche Sozialpolitik.” Leviathan 17 (4): 479-526. Reidegeld, Eckart. 2006. Staatliche Sozialpolitik in Deutschland. Band I: Von den Ursprüngen bis zum Untergang des Kaiserreiches 1918. 2. Aufl. Wiesbaden: VS Rossfeld, Roman, Thomas Buomberger und Patrick Kury, Hrsg. 2014. 14/18. Die Schweiz und der Grosse Krieg. Baden: hier und jetzt. Schweingruber, Edwin. 1977. Sozialgesetzgebung der Schweiz. Ein Grundriss. Zürich: Schulthess Polygraphischer Verlag. Sinzheimer, Hugo. 1920. “The Development of Labor Legislation in Germany.” The Annals of the American Academy of Political and Social Science 92: 35-40. Starke, Peter. 2018. “The Impact of War on Welfare State Development in Germany.” In Warfare and Welfare, herausgegeben von Herbert Obinger, Klaus Petersen und Peter Starke, 36-66. Oxford: Oxford University Press. Stolper, Gustav. 1915. “Die kriegswirtschaftlichen Vorgänge und Maßnahmen in Österreich.” Zeitschrift für Volkswirtschaft, Sozialpolitik und Verwaltung 24: 1-113. Studer, Brigitte. 2012. “Ökonomien der sozialen Sicherheit.” In Wirtschaftsgeschichte der Schweiz im 20. Jahrhundert, herausgegeben von Patrick Halbeisen, Margrit Müller und Béatric Veyrassat, 923-974. Basel: Schwabe Verlag. Tálos, Emmerich. 1981. Staatliche Sozialpolitik in Österreich. Rekonstruktion und Analyse. Wien: Verlag für Gesellschaftskritik. Tanner, Jakob. 2015. Geschichte der Schweiz im 20. Jahrhundert. München: C.H. Beck. Titmuss, Richard M. 1958. “War and Social Policy”. In Essays on the Welfare State, by Richard M. Titmuss. chap. 4, London: Unwin University Books. Tschudi, Hans Peter. 1989. Entstehung und Entwicklung der schweizerischen Sozialversicherungen. Basel: Helbling & Lichtenhahn. Walter, André. 2018. Federalism, Direct Democracy, and the Expansion of the Welfare State. Social Policy Development in Swiss Cantons 1930-2000. Dissertation no. 4738. Universität St. Gallen. Zahn, Friedrich. 1916. “Deutsche Sozialpolitik und der Krieg.” Annalen des Deutschen Reichs für Gesetzgebung, Verwaltung und Volkswirtschaft. München u.a.: Schweitzer, S. 1-58. Zahn, Friedrich. 1917. “Deutsche Kriegs-Sozialpolitik.” Allgemeines Statistisches Archiv. 10. Band 1916/17: 7482. Zimmermann, Waldemar. 1915. Krieg und Sozialpolitik (Soziale Kriegsrüstung). Berlin: Leonhard Simon. Zöllner, Detlev. 1963. Öffentliche Sozialleistungen und wirtschaftliche Entwicklung. Berlin: Duncker & Humblot.

227

Part III Political Determinants of Socio-Economic Outcomes

The varied sources of increasing wage dispersion Evelyne Huber1, John D. Stephens2, and Kaitlin Alper3 Abstract: Pre-tax and transfer, or market income distribution has grown more unequal over the past four decades and thus has put increasing demands for redistribution on the welfare state. We focus on one aspect of inequality, wage dispersion, investigating the factors that shape the ratio of wages at the 90th percentile to wages at the 10th percentile, and the ratio of wages at the median to wages at the 10th percentile. We analyze data for 21 advanced democracies for varying numbers of years between the 1960s and 2013, depending on data availability for both the dependent and the independent variables. We integrate a focus on structural changes in the economy with a focus on labor market institutions, partisan politics, and education policies. We argue that the same factors shape wage dispersion between both the top and the bottom and the middle and bottom, but some of those factors have a stronger effect on the distance between the bottom and the median income earner. Specifically, higher levels of capital market openness, Third World imports, and immigration are all associated with higher wage dispersion in the 90:10 and the 50:10 ratio, but these effects are much stronger for the 50:10 ratio. Among the countervailing factors, union density, wage coordination, employment legislation that regulates temporary work, and investment in human capital restrain the overall distance and the distance between the bottom and the middle of the income distribution. The effect of regulation of temporary employment, however, is stronger at the bottom.

1

Introduction

The rise in income inequality over the past quarter century has attracted wide academic and political attention. It has sharpened the debate about the state’s responsibility to address new and old social risks, at a time when fiscal pressures on the welfare state had intensified because of a combination of revenue constraints and demographic pressures (Armingeon 2006). As social scientists and social democratic policy makers began to see the limits of solely relying on redistribution through taxes and transfers as an antidote to rising inequalities, they began to focus on ‘predistribution’, that is, changing the distribution of market income through, for example, strengthening the hand of unions in wage bargaining, raising the minimum wage, and investing in skill development, especially at the bottom of the distribution (Chwalisz and 1 2 3

University of North Carolina, Chapel Hill [email protected] University of North Carolina, Chapel Hill [email protected] University of North Carolina, Chapel Hill [email protected]

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 R. Careja et al. (eds.), The European Social Model under Pressure, https://doi.org/10.1007/978-3-658-27043-8_14

231

Diamond 2015). Here we focus on factors that shape wage dispersion, investigating the factors that shape the ratio of wages at the 90th percentile to wages at the 10th percentile, and the ratio of wages at the median to wages at the 10th percentile. In doing so, we isolate the individual wage component from other determinants of household market income, such as composition of the household and the volume of work of the household. After declining since the beginning of the Great Depression, income inequality measured variously as household disposable income, household market income, wage dispersion, or top income shares, began to rise in the mid-1980s in almost all industrial/post-industrial democracies. Figures 1 and 2 show the trends in the 50:10 and 90:10 wage ratios respectively. While there is generally an unmistakable increase, there is a lot of cross national variation in the trends. The figures also show a picture of relatively greater stability in the 50:10 than in the 90:10 ratio (Figures 1 and 2). Nevertheless, there is a noticeable widening of the 50:10 ratio in some countries (U.S., Germany, the Netherlands, Denmark, Norway) and a decline in others (Finland, Austria, Belgium, France, and especially Canada), so it is worth exploring what drives these developments. Building on previous studies, we integrate a focus on structural changes in the economy with a focus on labor market institutions, partisan politics, and education policies. In the analysis that follows, we find that the same factors shape wage dispersion between both the top and the bottom and the middle and bottom, but some of them have a stronger effect on the distance between the bottom and the median income earner. Specifically, higher levels of capital market openness, Third World imports, and immigration are all associated with higher wage dispersion in the 90:10 and the 50:10 ratio, but these effects are much stronger for the 50:10 ratio. Among the countervailing factors, union density, wage coordination, employment legislation that regulates temporary work, and investment in human capital restrain the overall distance and the distance between the bottom and the middle of the income distribution. The effect of regulation of temporary employment, however, is stronger at the bottom.

232

Figure 1: Wage dispersion by country and welfare state regime, 50-10 ratio

233

Figure 2: Wage dispersion by country and welfare state regime, 90-10 ratio

234

2

Literature and hypotheses

The comparative politics literature on wage dispersion is not large and has primarily focused on the effects of variations in labor market institutions and politics. We confirm the finding of those studies in our analysis, and we add to that analysis tests of arguments about the effects of globalization and variations in human capital and skills on inequality. Globalization and the transition to the knowledge economy have created winners and losers, and it is important to understand which factors narrow or widen the gap between them. 2.1

Globalization

The economic explanation for rising inequality has been based on the impact of globalization and of technological change. Trade competition and international outsourcing caused by globalization have resulted in deindustrialization and a decline of demand for unskilled workers. Wood (1994) argued that manufactured imports from less developed countries (LDCs) are produced by low-skill labor, so an increase in such imports directly competes with the domestic production of goods that would employ low-skilled labor, thereby lowering the demand for, and the wages of, low-skilled labor. This argument spurred a huge debate in economics: as Wood (2018) notes, EconLit lists about 1,000 books and 2,400 articles with the JEL trade and wages classification. Essentially, the debate was about the relative contribution of trade versus technological change to the declining demand for low-skilled labor. However, when political scientists introduced labor market institutions as controls, the impact of trade appeared greatly reduced. Pontusson et al. (2002), for instance, found no effect of non-OPEC LDC trade on wage dispersion for the period 1973-1995. At the same time as trade with LDCs has reduced the demand for low-skilled labor, the transition to the knowledge economy has increased the demand for skilled workers, and to the extent that the supply has not kept pace with demand, the skill or education premium has increased. Thus, van Reenan (2011), Freeman (2009), and Chusseau et al. (2008) argued that globalization and technological change are complementary rather than competing explanations. Hope and Martinelli (2017) found a statistically significant positive effect of the growth of knowledge-intensive employment on the top 1% share and on the 90:10 ratio, but they also found that this effect was mitigated by union density (for the 90:10 ratio only), wage coordination, bargaining coverage, and employment protection legislation. In contrast, Huber, Huo and Stephens (2017) found a statistically significant negative effect of a different measure of employment in knowledge intensive sectors on the top 1% income share. In line with the above literature, our first set of hypotheses is concerned with testing globalization hypotheses. Economic globalization has several dimensions. The two most frequently highlighted in the literature are trade openness and capital mobility. Trade openness 235

in advanced countries is assumed to reward capital intensive production and reduce labor intensive production, thus raising wages of skilled and depressing wages of unskilled workers. However, to the extent that most trade is carried on with other advanced countries, this effect may not operate. Expectations are different for imports from LDCs. Such imports are typically produced by labor intensive methods using unskilled labor and thus substitute for similar kinds of production in the importing countries and thereby depress wages for unskilled workers. Accordingly, we expect imports from LDCs to increase wage dispersion at both levels. Capital mobility influences wage levels through two mechanisms. One is by companies threatening to move production abroad if wage demands are too high, and the other one is by companies actually investing in production abroad. Typically, the parts of production of goods and services (such as call centers) that can be moved abroad most easily are the routine, or lowskill parts. Threats by companies of moving production abroad are thus likely to depress wages of low-skilled workers. Such threats are more credible in the context of high capital openness, which is why we expect high capital market openness to increase wage dispersion. In a parallel manner, we expect high actual levels of investment abroad to increase wage dispersion. A final component of globalization is labor mobility. Immigration into advanced economies has increased significantly in the past quarter century. Only a small part of this migration is part of the brain drain, that is, attraction of very highly skilled people from less developed countries. The bulk of the migration consists of people fleeing violence or poor life chances because of lack of economic opportunities. Even though not all migrants are low-skill labor in their home countries, language problems typically force them into unskilled jobs, at least in the beginning. Thus, a large pool of migrants is likely to increase the supply of low-skill labor and thus depress wage levels for these groups. Therefore, we expect higher levels of immigration to be associated with higher wage dispersion overall and at the bottom. 2.2

Human capital and skills

Certainly, there have been significant differences in patterns of wage dispersion during the transition to the open post-industrial knowledge economy between countries that require explanation (Atkinson 2007, 2008; Gottschalk and Smeeding 1997). For instance, the loss of ground in the race between education and technology that was highlighted by Goldin and Katz (2008) for the United States has not been uniform. The Nordic countries have consistently invested more heavily in education at all levels and thus have increased the supply of skilled workers. Nickell (2004) found that skills dispersion, as measured by the ratio of test scores of adults at the 95th percentile to adults at the 5th percentile in the International Adult Literacy Survey (IALS), was a better predictor of wage inequality than variations in labor market institutions, the then dominant explanation of variations in wage dispersion. Weisstanner and Armingeon (2018) found that public expenditure on education depressed the education premium, conceptualized as the difference between the median wage of workers 236

with tertiary education and the median wage of workers without tertiary education, in 22 OECD countries between the early 1990s and 2014. Along similar lines, Busemeyer (2015) found that public spending on all levels of education was negatively related to the Gini of disposable household income in the period 1997-2008 in advanced industrial societies. In contrast, he found that the private share of education spending at all levels was positively related to the Gini of disposable household income in these same country/year observations. Similarly, Weisstanner and Armingeon (2018) showed a positive effect of private education spending on the education premium. In our exploration of scores of 15-year old students on the test of mathematical skills in all the waves of the PISA tests conducted between 2000 and 2015, we found that total education spending and public education spending raised the math scores of students at the 25th percentile and the mean, but that the effect of public spending was stronger. Moreover, the difference between the effects of total and public expenditure was larger for math scores at the 25th percentile than for math scores at the mean (Huber, Gunderson, and Stephens 2018). Finally, in our analysis of top 1% income shares, we found that the private share of total tertiary spending was strongly associated with greater top income shares (Huber, Huo, and Stephens 2017). So the sources of education spending matter because they shape the (in)egalitarian nature of educational systems and indirectly the distribution of wages. Another important characteristic of the educational system that has been associated with lower wage dispersion is the system of vocational education and training (VET). Opportunities for vocational education and training offer a path to medium to high skill jobs for secondary school students who are not going to enter tertiary education. Where such opportunities are wide spread, the education premium and wage dispersion are lower. Busemeyer (2015) found that the share of upper secondary students enrolled in VET was negatively related to wage dispersion in models with the appropriate labor market institutional and political controls. Our second set of hypotheses, then, focuses on the effects of skills and education policies. We expect lower skills dispersion to be associated with lower wage dispersion. The best correction to the imbalance between demand and supply of skilled labor is to increase the supply. This is clearly a medium to long-term proposition, as it starts with investment in education. Retraining initiatives of adults as part of active labor market policies can be helpful, but in general they reach workers with a stronger educational background to begin with (KilpiJakonen et al. 2015). Preparation for school starts in early childhood, so we would expect human capital spending, which includes public expenditures on early childhood education and care along with public education expenditure at all levels to be associated with lower wage dispersion. We have the same expectation for the impact of public education expenditures at all levels on wage dispersion. In contrast, we expect the private share of education expenditures to be associated with higher wage dispersion.

237

2.3

Labor market institutions and politics

Our third set of hypotheses is concerned with labor market institutions. It builds on power resources theory (Stephens 1979; Korpi 1983) that sees organization as a counterweight to wealth in democratic societies. Strong organization of labor in unions and political parties can increase the share of income going to labor directly through collective negotiations, and it can improve the overall position of labor in terms of rights and benefits indirectly through legislation. Where labor movements are able to organize a large share of employees and/or where they are able to coordinate wage bargaining at a high level, they not only improve the collective position of labor vis-à-vis employers but they also reduce income differentials among employees. So, we expect union density and coordination of wage bargaining at a high level to be associated with lower wage dispersion both overall (the 90:10 ratio) and in the bottom half (the 50:10 ratio). Left governments have seen the center of gravity in their social base shift from blue-collar workers to white-collar employees, but they have retained a commitment to equity as a defining characteristic (Gingrich and Häusermann 2015). Accordingly, we expect longterm incumbency of left governments to be associated with lower wage dispersion overall and in the bottom half. Political scientists have tested the effects of variables derived from power resources theory on wage dispersion. Wallerstein (1999) and Pontusson, Rueda and Way (2002) identified high levels of union density, of coordination of wage bargaining, of coverage by collective contracts, and left governments as key factors associated with comparatively low levels of dispersion. Rueda and Pontusson (2000) added varieties of capitalism to the analysis and showed stronger negative effects of coordination of wage bargaining on wage dispersion in coordinated than in liberal market economies, and positive effects of right government in liberal market economies, but an absence of such effects in coordinated market economies. Studies covering the more recent period confirmed the importance of union density as a factor restraining the growth of the top 10% income share (Jaumotte and Buiton 2015). Vlandas (2016) focused specifically on the 50:10 ratio and argued that only the combination of wage coordination with unionization of low-income workers depressed this ratio. As the share of temporary jobs has increased, the status of temporary work with regard to pay and benefits has become an important factor for wage dispersion. In some countries, contracts of temporary workers can be renewed time and again, and temporary workers have lower wages and benefits than permanent workers, whereas in other countries temporary employment is regulated as to duration and temporary workers are entitled to the same wages and benefits as permanent workers. Therefore, we expect regulation of the duration, remuneration, and benefits of temporary work to be associated with lower wage dispersion overall and in the bottom half.

238

Finally, some countries have laws or regulations that extend the coverage of collective agreements to sectors of the labor force that were not party to the negotiations. Such contract coverage extension works as a weaker substitute for coordination of wage bargaining at a high level. Accordingly, we expect expansion of contract coverage to reduce wage dispersion both overall and in the bottom half. The same structural and institutional variables that make up our three sets of hypotheses have figured prominently in the rapidly growing body of work that has explored the drivers of top incomes (Atkinson, Piketty, and Saez 2011; Alvaredo et al. 2013; Huber, Huo, and Stephens 2017; Tomaskovic-Devey and Lin 2011; Volscho and Kelly 2012). Indicators for the transition to the knowledge economy, deindustrialization, trade and foreign investment have been included in the models along with union density, union centralization, partisan government, top marginal tax rates, and education expenditures. Findings on the impact of labor market institutions, partisan politics, and policies have been quite consistent, whereas findings regarding the transition to the knowledge economy and trade and foreign investment have been more mixed. Studies of the top 1% income share over long periods of time (Atkinson et al. 2011; Alvaredo et al. 2013) have highlighted changes in top marginal income tax rates, along with financial crises, wealth accumulation, inheritance flows, and capital income Finally, economic transformations, labor market institutions, and policies, along with demographic changes have been the master variables used to explain inequality of market income of households, which is shaped both by wage levels and the volume of work, as well as self-employment and capital income (Bradley et al. 2003; Huber and Stephens 2014; Kenworthy and Pontusson 2005). Again, effects of deindustrialization and labor market institutions yielded consistent findings. To summarize, we shall follow previous literature here and include indicators for economic structural, institutional, and political changes. We shall pay particular attention to policy variables, as these are under control of policy makers and could be adjusted to counteract the rise in wage dispersion. Table 1 lists our variables and the direction of our hypotheses.

239

240

Seats of leftist parties as proportion of the seats of all governing parties, cumulative from 1945 to date of observation. Net union membership as a percentage of employed wage and salary earners. Coordination of wage setting. Coded 1 (most fragmented) to 5 (most centralized). Employment protection legislation limiting temporary status. Employees in workplaces or establishments covered by unions or works councils as a proportion of all wage and salary earners in employment.

Ratio of gross earnings received by a worker at the 50th earnings percentile to that received by a worker at the 10th percentile. Ratio of gross earnings received by a worker at the 90th earnings percentile to that received by a worker at the 10th percentile.

Definition

Visser (2013) Visser (2013) OECD Visser (2013)

Brady et al. (2014)

OECD

OECD

Original data source

All variables are available in Brady, Huber, and Stephens (2014)

4 or 5 = 1; 1, 2, or 3=0. Interaction term of union density*wage coordination dichotomy. Outward foreign direct investment as a % of GDP. IMF Sum of exports and imports as a percentage of GDP. Penn World Tables An index of capital controls in which high values mean few or no controls on cross Karcher and Steinberg (2013) border capital movement. Third World imports Manufacturing imports from developing countries as a % of GDP. OECD Immigration International migrant stock as a percentage of the population. World Bank Human capital spending Public spending on education, daycare, and active labor market policies as a % of GDP. OECD Public education spending Public spending on education at all levels except pre kindergarten as a % of GDP. OECD Total education spending Public and private spending on education at all levels except pre kindergarten as a % of OECD GDP. share of education spending as a % of total education spending. Private share of education spendingPrivate OECD Vocational education % of students enrolled in upper-secondary education who are in vocational or technical programs. OECD

Wage coordination dichotomy Union density*wage coordination Outward FDI Trade openness Capital market openness

Union density Wage coordination EPL temporary work Contract coverage

Left government

Independent variables

Wage dispersion 90-10 ratio

Wage dispersion 50-10 ratio

Dependent variables

Table 1. Variable definitions and sources

+ + + -

+ + +

-

-

Hypothesized relation

3

Data and measurement

3.1

Dependent variables

The variables, their measurement, and the data sources are shown in Table 1. We analyze data for 21 advanced democracies for varying numbers of years between the 1960s and 2013, depending on data availability for both the dependent and the independent variables. Wage dispersion at the bottom is measured as the ratio of wage and salary earnings received by the worker at the 50th percentile to those received by the worker at the 10th percentile. Wage dispersion across the distribution is measured as the ratio of wage and salary earnings received by the worker at the 90th percentile to those received by the worker at the 10th percentile. Both are measured among full time workers and thus are not affected by the volume of part time work. 3.2

Independent variables

The independent variables in our analysis are taken from the Comparative Welfare States Dataset (CWSD), though from a variety of original sources. Left government is defined as the proportion of seats held by social democratic, left socialist, communist, and green parties relative to the number of seats held by all governing parties. We use a cumulative measure from 1946 to the year of observation. Wage coordination, union density, and contract coverage variables come from Visser’s (2013) Database on Institutional Characteristics of Trade Unions, Wage Setting, State Intervention and Social Pacts (ICTWSS). The wage coordination variable is an updating of a variable originally developed by Kenworthy (2001). It is an indicator of the “degree, rather than the type, of coordination … based on a set of expectations about which institutional features of wage setting arrangements are likely to generate more or less coordination” (Kenworthy 2001, 78, 80). Vlandas recodes this variable as a dichotomy. Union contract coverage is the percentage of employees covered by collective (wage) bargaining agreements as a proportion of all wage and salary earners in employment. This is an extension of a measure originally developed by Traxler (1994). Temporary employment protection legislation (EPL) is the OECD measure of limiting the use of temporary work. High temporary EPL limits the extent of the use of temporary work and payment of lower wages to temporary workers. Thus, it does protect temporary workers but also protects permanent workers from replacement by temporary workers. Trade openness refers to the sum of exports and imports, as a percentage of GDP, and comes from the Penn World Tables. Outward FDI is measured as a percentage of GDP. Capital market openness is an index of capital controls coded by Karcher and Steinberg (2013) from IMF sources in which high values mean few or no controls on cross border capital movement. Third 241

world imports refer to manufacturing imports from developing countries as a percentage of GDP. Immigration is defined as the percentage of the population that are international migrants. Human capital spending is a composite measure we created by adding public expenditure on early childhood education and care, education and active labor market policies as a percentage of GDP. We use total education spending as a percentage of GDP and public education spending as a percent of GDP, and we calculate the share of education spending that comes from private sources. We look at funding sources rather than enrollment in public versus private schools because of the wide variation of the rules under which private schools operate. We measure vocational education and training (VET) effort as the percent of upper-secondary students enrolled in VET programs, which are defined by the OECD as educational programs which end in the necessary certifications to begin working in a given profession without further training. To test Nickell’s contention that skills inequality is the main determinant of wage dispersion, we use the numeric test scores on the OECD’s Programme for the International Assessment of Adult Competencies (PIAAC) Survey of Adult Skills which is an updated version (circa 2010) of the IALS data he used (OECD 2013). Following Nickell, we measure skill dispersion as the ratio of test scores of adults at the 95th percentile to adults at the 5th percentile.

4

Statistical estimation

We address serial correlation by correcting for first order auto-regressiveness rather than by adding a lagged dependent variable. Beck and Katz (2004, 2011) have shown that this strategy (AR1 corrections) actually does include a lagged dependent variable on the right hand side of the equation (PCSE and AR1 corrections are known as Prais Winsten estimations). Thus, it deals with the problem of serial correlation without, as our results show, suppressing the power of other independent variables. We hypothesize that our causes operate over long periods of time and changes in the dependent variable occur gradually, much like in the case of cumulative causes in Pierson's (2003, 198) typology of causes and effects. Moreover, in almost all pooled time series studies of the determinants of inequality, regardless of whether it is measured by wage dispersion, the Gini coefficient of household income, or top income shares, the dependent variable is measured as a level. Thus, it is appropriate to measure the dependent and independent variables as levels.4 We include period dummies in order to control for common economic shocks, such as oil price increases or global economic cycles. The periods selected are the latter part of the Golden Age of post war growth (1960-1972), oil shocks and stagflation years of the seventies (19731979), the period of deregulation up to the introduction of the single European market (19801992), and the global financial crisis and its aftermath (2008-2012). The reference period is 1993-2007, the transition to the knowledge economy. 4

For this reason, error correction estimation in which the dependent variable is measured as a first difference is not an appropriate technique to model the hypothesized causal processes. 242

We estimate our Prais Winsten models in Stata 14.1 using Vernby and Lindgren’s (2009) dvgreg package. Dvgreg is specifically designed to estimate dynamic panel data models with gaps in the dependent variable, but complete or nearly complete data on the independent variables. It generates an estimate of the value of the dependent variable at t-1 for each gap, based on the values of the dependent variable at the previous actual observation and the values of the independent variables. This then makes it possible to derive a corrected estimate of AR1. PIAAC data are available only for a cross-section of OECD countries circa 2010, including 17 of our 21 countries in Figures 1 and 2. Much of the data presented in the tables in OECD (2013) are the PIAAC scores not adjusted for literacy related nonresponses. Literacy-related nonresponses varied from a low of 0 (zero) per cent in Finland and Sweden to highs of 4.2 per cent in the United States and 5.2 per cent in Belgium (actually Flanders only). To adjust the 5th and 95th percentile scores for literacy-related nonresponses, we had to get the micro data for each country from the OECD. The OECD could not supply us with the micro data for Australia. The large literacy-related nonresponse rate in Belgium is being driven by an over 13 per cent literacy-related nonresponse rate in the province of Vlaams Brabant, as there are many French speakers and foreigners living there. Adjusting the Belgian data for literacy-related nonresponses resulted in a numeracy score of 85 at the 5th percentile, which is the lowest score possible and much lower than any other country. This also skewed the 95-5 skill ratio for Belgium. We decided to exclude Belgium from the analysis. We include the same variables that Nickell included in his analysis: the 95-5 skill ratio, union density and union contract coverage. The data for these variables and for the dependent variable are for 2010. The statistical estimator for the cross-sectional analysis is OLS regression.

5

Results

The results for the cross-sectional analysis with PIAAC data are shown in Table 2. With such a small number of cases, we report significance at the .1 level in addition to the more conventional levels. The skills dispersion variable and one of the two labor market institutions variables are consistently significant in the models for both dependent variables. Comparing adjusted R squares for the various model can give us an estimate of inner and outer bounds of the effects of skills dispersion and labor market institutions. Models 2 and 5 show the outer bound for skills dispersion; 28 per cent of the variation explained for the 50-10 ratio and 38 per cent of the 90-10 ratio. Comparing Models 1 and 3 gives the inner bound for the 50-10 ratio as 8 per cent; Models 4 and 6 show the inner bound for the 90-10 ratio as 18 per cent. For labor market institutions the outer bound is 37 per cent for the 50-10 ratio and 24 per cent for the 9010 ratio. The inner bound is 17 per cent for the 50-10 ratio and 4 per cent for the 90-10 ratio.

243

Table Dispersion Table2.2.Wage Wage Dispersion

Union density Contract coverage Numeracy 95-5 ratio Constant R2 Observations

Model 1 -.002 -.004 ** 1.951 .37 * 16

50-10 ratio Model 2 Model 3 -.001 -.003 * .310 ** .199 ┼ .990 ** 1.452 ** .45 ** .28 * 16

16

Model 4 -.016 ┼ -.006 3.874 ** .24 ┼ 16

90-10 ratio Model 5

1.326 ** .108 ** .38 * 16

Model 6 -.010 ┼ -.004 .992 * 1.384 .42 * 16

┼ significant at the .1 level* .05 level; ** .01 level

The results of our analysis for the 50:10 ratio and the 90:10 ratio are shown in Table 3 and 4 respectively. Model 1 (in both tables) reports the results of the re-analysis with turn of the century studies of labor market institution and political variables, the baseline for our analysis. Model 2 adds two institutional variables not tested in those studies. Model 3 tests Vlandas’ argument that wage coordination effects are contingent on union density. Model 4 adds the five globalization variables to the baseline. Models 5-7 add the various human capital effort variables and drop left cabinet, which is a determinant of human capital spending (Huber and Stephens 2014), from the baseline. We enter these variables one or two at a time rather than together as in the case of the globalization variables, because we lose so many observations due to the limited time coverage of the OECD education and vocational education variables. Both series begin in the mid-1990s.5 Model 8 combines the significant variables from Models 1-5. Again, variables from the last two education models are excluded because of limited time coverage.

5

There are earlier education data, from the OECD and other sources (see Brady, Huber, and Stephens 2014), but these data do not contain information on private education spending. 244

245

Model 1 -.002 ** -.058 *** -.006 *** .000

Model 2 -.002 *** -.031 *** -.004 *** -.078 *** -.064 * .001

-.006 ***

Model 3 -.003 ***

Model 4 Model 5 Model 6 -.001 -.002 *** -.001 -.057 *** -.037 *** -.067 *** -.006 ***

Period dummies included in all equations but not shown.

* significant at .05; **significant at .01; ***significant at .001; ^ significant opposite hypothesized direction.

Outward FDI -.003 ^ Trade openness -.001 ^ Capital market openness .057 *** Third World imports .030 *** Immigration .003 ** Human capital spending -.054 *** Public education spending -.055 *** Total education spending Private share of education spending Vocational education Constant 1.966 *** 2.002 *** 2.124 *** 1.715 *** 2.196 *** 2.160 *** Common ρ .90 .90 .90 .90 .90 .90 2 .47 *** R .60 *** .37 *** .52 *** .41 *** .45 *** Observations 506 366 506 430 367 234

Wage coordination dichotomy Union density*wage coordination

Left government EPL temporary work Contract coverage

Union density Wage coordination

Table 3. Determinants of Wage Dispersion (50-10 Ratio)

.028 * .007 .000 -.035 ***

-.064 ***

Model 8 -.003 *** -.021 **

-.025 * .069 ** ** -.002 * 1.919 * *** 4.503 *** .90 .90 .43 *** .51 *** 178 317

Model 7 -.002 * -.028 **

246

-.114 ***

* significant at .05; **significant at .01; ***significant at .001; ^ significant opposite hypothesized direction. Period dummies included in all equations but not shown.

-.003 -.003 ^ .077 ** .060 *** .008 * -.139 ***

Model 4 Model 5 Model 6 -.005 ** -.010 *** -.006 -.160 *** -.137 *** -.217 *** -.019 ***

3.925 *** 3.683 *** 4.543 *** 4.625 *** .90 .90 .90 .90 .50 *** .55 *** .41 *** .47 *** 518 430 365 238

-.019 ***

Model 3 -.015 ***

Outward FDI Trade openness Capital market openness Third World imports Immigration Human capital spending Public education spending Total education spending Private share of education spending Vocational education Constant 4.074 *** 4.098 *** Common ρ .90 .90 .55 *** R2 .59 *** Observations 518 362

Model 2 -.009 *** -.127 *** -.017 *** -.186 *** .003 ^ -.543 *** .009 ^

Model 1 -.007 *** -.166 *** -.018 ***

Wage coordination dichotomy Union density*wage coordination

Left government EPL temporary work Contract coverage

Union density Wage coordination

Table 4. Determinants of Wage Dispersion (90-10 Ratio)

-.044 .029 -.010 ** 4.317 * *** 4.503 *** .90 .90 .56 *** .51 *** 182 317

-.035 .001 .001 -.073 *

-.140 ***

Model 7 Model 8 -.008 *** -.012 *** -.114 *** -.096 ***

Model 1 in the two tables confirms that the findings of studies of 15-20 years ago still hold for both of our wage dispersion measures, which is not surprising. Model 2 shows that additional labor market institutions add to the variation in wage dispersion explained. Specifically, regulation of temporary work has a highly significant depressing effect on wage dispersion both at the bottom and overall.6 Our Model 3 does not support Vlandas’ finding that the effect of wage coordination is contingent on labor market institutions being inclusive, that is, on the workers at the bottom being organized. In fact, the interaction term is actually positive for both the 50:10 and the 90:10 ratio, and the main term for wage coordination is negative, indicating that wage coordination depresses wage dispersion even when union density is 0. Model 4 in both tables shows that reducing barriers to cross border capital movement, increasing cross border migration, and increasing manufacturing imports from developing countries increase wage dispersion. Models 5 and 6 show that increases in public investment in skills and education are associated with lower levels of wage dispersion, while Model 7 shows that a high share of private spending on education increases wage dispersion. Model 7 also shows that public vocational education decreases wage dispersion. Analyses by Weisstanner and Armingeon (2018) and Huber, Gunderson, and Stephens (2018) suggest that these effects are due to the fact that public investments in education and training are associated with lower education wage premiums. It is worth exploring further the effects of the globalization and social investment variables that are our primary theoretical interest. In most cases, we expect the globalization variables, especially Third World imports and immigration, and the public human capital spending variables to have a greater effect on the 50:10 ratio than on the 90:10 ratio. In the case of temporary employment protection legislation, the situation is somewhat more ambiguous as temporary EPL does protect temporary workers, who are certainly disproportionately at the bottom of the wage distribution, but it also protects permanent workers by limiting the volume of temporary work. To make these comparisons, one cannot compare the metric coefficients because the mean and standard deviation of the 50:10 ratio is much lower than the mean and standard deviation of the 90:10 ratio. Thus, we standardized both dependent variables and reran the models on the standardized variables (models available on request). In the case of the significant globalization variables, they do have a larger effect on the 50:10 ratio: 22 per cent larger in the case of immigration, 44 per cent for Third World imports, and 123 per cent for (lack of) controls on cross border capital movement. The size of the capital controls difference is somewhat surprising. The social investment variables also have stronger effects on the 50:10 ratio than on the 90:10 ratio, with public education spending having a 26 per cent larger effect and human capital spending a 44 per cent larger effect. The stronger effect of human capital spending makes sense, since it includes early childhood education spending 6

Contract coverage and temporary EPL are highly correlated (.61). When temporary EPL is dropped from Model 2, contract coverage is negative and highly significant in the models for both dependent variables. 247

and active labor market policy spending, which might be assumed to have greater effects at the bottom of the skill distribution. Finally, temporary EPL also has a strong effect at the bottom; the coefficient in the regression on the standardized 50:10 ratio is 29 per cent larger than the coefficient for the standardized 90:10 ratio.

6

Discussion and conclusion

Patterns of wage dispersion across advanced economies have been shaped by structural transformations of these economies and by institutional constraints. At the same time, governments with value commitments to equity have pursued policies that have counteracted the upward pressures on wage dispersion resulting from the structural transformations. In contrast, governments with value commitments to growth without equity have pursued policies that have enhanced wage differentials. We found that higher levels of capital market openness, Third World imports, and immigration all were associated with higher wage dispersion, both with the 50:10 and the 90:10 ratio. However, these effects were much stronger for the 50:10 than for the 90:10 ratio, suggesting that these variables affect people at the bottom of the skill and wage distribution particularly strongly. This is not surprising at all in the cases of immigration and Third World imports. With the exception of comparatively small numbers of highly educated individuals, immigrants compete for low-skill jobs and thus depress wages at the bottom. Manufacturing imports from LDCs compete with the domestic production of goods by unskilled labor and thus also depress wages at the bottom. The reasons for the stronger impact of low or no controls on capital movements across borders on wage dispersion at the bottom are not as obvious. One could argue that this is an indicator of shifts of production across borders, which likely would be production with low skill labor. However, such shifts in production would be captured by outward FDI, and outward FDI is not significant for the 90:10 ratio and significant but in the wrong direction for the 50:10 ratio, actually depressing wage dispersion at the bottom. A plausible interpretation is that capital openness makes threats of moving production abroad more credible, which gives employers particular leverage to keep wages at the bottom low. Interestingly, the effect of capital market openness remains significant, driving up wage dispersion at the bottom, even in the final model with all the labor market institutions and human capital spending. This suggests that even unions seem to be impressed by the threat of relocation of production and thus agree to keep wages at the bottom low. Labor market institutions have clearly worked as counterweights to the inegalitarian effects of structural transformations. Strong unions, high levels of wage coordination, and strong legislation protecting and limiting temporary work all have highly significant negative effects on wage dispersion overall and at the bottom. Moreover, there is no great difference between the magnitude of the effects of union density and wage coordination on the overall dispersion 248

and on the dispersion at the bottom. Thus, strong unions keep the bottom from falling too far below the median and the top from rising too far above the bottom. The same is true for high levels of coordination of wage bargaining. In contrast, the effect of legislation that limits temporary work is stronger at the bottom than for the overall distribution. High levels of legal regulation of temporary work protect temporary workers from becoming a permanent temporary underclass and thus depressing wage levels at the bottom. Strong systems of vocational education constitute a further institutional counterweight to the inequality-enhancing effects of structural transformations. Where large numbers of upper secondary students who for whatever reason do not have the option of pursuing tertiary education enter formal vocational training, the supply of skilled workers is higher and thus depresses wage dispersion. Again, this is true for both wage dispersion ratios. Finally, a strong public commitment to education at all levels, indicated by high levels of public spending on early childhood education and care, pre-school and school education, and active labor market policies, improves the supply of skilled workers and thus moderates wage dispersion. It is important here to emphasize the public commitment. Not all sources of education expenditures have a moderating effect on inequality. Indeed, a high share of private education spending has a highly significant enhancing effect on the 50:10 ratio. This finding squares with Busemeyer’s (2015) finding of the same effect on the Gini of household disposable income, Weisstanner and Armingeon’s (2018) finding of the same effect on the education premium, and our finding on the effect of the private share of education spending on skills at the bottom (Huber, Gunderson, and Stephens 2018). The importance of investments in skills can be more clearly seen in the cross-sectional analysis which includes the PIAAC scores, by far the best measure of the stock of skills of the adult population. Our analysis confirmed Nickell’s finding that the 95-5 skill ratio is actually a more important determinant of the 90-10 wage ratio than variations in labor market institutions and that it was also important in the case of the 50-10 wage ratio. The 95-5 skill ratio is driven by differences at the bottom: The correlation of numeracy scores at the 5th percentile and the 95-5 ratio is -.93. It is not surprising that the scores at the bottom are particularly high in countries that make large public investments in human capital, like the Nordic countries, and are particularly low in countries that do not, like the United States. In sum, structural transformations in advanced economies indeed are responsible for upward pressures on wage dispersion overall and at the bottom. In particular, manufacturing imports from the Third World, immigration, and high levels of capital openness widen both the 90:10 and the 50:10 ratios, but their effect is stronger at the bottom. Low skill/low wage workers are particularly disadvantaged by these transformations. However, governments have options to counteract these pressures. They can strengthen institutions and implement policies that protect people at the bottom of the skill and wage distribution from being pushed further below the median and further away from the top of the wage distribution. They can protect unionization efforts, promote bargaining coordination, and 249

regulate temporary work, or they can make unionization difficult, support decentralization of bargaining, and deregulate temporary work. They can promote vocational education and training and invest heavily in education at all levels and in active labor market policies, or they can shift the balance of responsibility in these areas from the public to the private sector. The first sets of these options are typically chosen by left governments, the second sets by secular center and right governments. Even with the European Social Model in crisis, consensus among potential supporters of left parties on these issues should not be too difficult to construct. Despite the much highlighted insider/outsider conflicts of interest, both groups should see their common interest in the protection of unions, coordination of collective bargaining, and regulation of temporary work, given that all three of them contribute to moderating wage dispersion at both levels. Support for investment in human capital at all levels is another area of common interest. And a less unequal distribution of wages will take some redistributive pressure off the overburdened welfare state, where conflicts of interest are arguably more pronounced and complicated by generational differences of interest.

References Alvaredo, Facundo, Anthony B Atkinson, Thomas Piketty, and Emmanuel Saez. 2013. “The Top 1 Percent in International and Historical Perspective.” Journal of Economic Perspectives 27 (3): 3–20. Armingeon, Klaus. 2006. “Reconciling Competing Claims of the Welfare State Clientele: The Politics of Old and New Social Risk Coverage in Comparative Perspective.” In The Politics of Post-Industrial Welfare States: Adapting Post-war Social Policies to New Social Risks, edited by Klaus Armingeon and Giuliano Bonoli, 100-123. New York: Routledge. Atkinson, Anthony B. 2008. The Changing Distribution of Earnings in OECD Countries. The Rodolfo Debenedetti Lectures. Oxford, New York: Oxford University Press. Atkinson, Anthony B. 2007. “The Distribution of Earnings in OECD Countries.” International Labour Review 146 (1–2): 41–60. Atkinson, Anthony B., Thomas Piketty, and Emmanuel Saez. 2011. “Top Incomes in the Long Run of History.” Journal of Economic Literature 49 (1): 3–71. Beck, Nathaniel, and Jonathan N. Katz. 2011. “Modeling Dynamics in Time-Series–Cross-Section Political Economy Data.” Annual Review of Political Science 14 (1): 331–352. Bradley, David, Evelyne Huber, Stephanie Moller, François Nielsen, and John D. Stephens. 2003. “Distribution and Redistribution in Postindustrial Democracies.” World Politics 55 (2): 193–228. Brady, David, Evelyne Huber, and John D. Stephens. 2014. Comparative Welfare States Data Set. Chapel Hill and Durham, NC, University of the North Carolina and Duke University. Busemeyer, Marius R. 2015. Skills and Inequality: Partisan Politics and the Political Economy of Education Reforms in Western Welfare States. Cambridge, United Kingdom, New York: Cambridge University Press. Chusseau, Nathalie, Michel Dumont, and Joël Hellier. 2008. “Explaining Rising Inequality: Skill-Biased Technical Change and North–South Trade.” Journal of Economic Surveys 22 (3): 409–457. Denk, Oliver, and Boris Cournede. 2015. “Finance and Income Inequality in OECD Countries.” SSRN Electronic Journal. Dünhaupt, Petra. 2014. “An empirical assessment of the contribution of financialization and corporate governance to the rise in income inequality.” Working Paper, Institute for International Political Economy Berlin. Feenstra, Robert C., Robert Inklaar, and Marcel P. Timmer. 2015. “The Next Generation of the Penn World Table”. American Economic Review 105 (10): 3150-3182. Freeman, Richard B. 2009. “Globalization and Inequality.” In The Oxford Handbook of Economic Inequality, edited by Brian Nolan, Wiemer Salverda and Timothy M. Smeeding, 575-598. Oxford, New York: Oxford University Press. 250

Gingrich, Jane, and Silja Häusermann. 2015. “The Decline of the Working-Class Vote, the Reconfiguration of the Welfare Support Coalition and Consequences for the Welfare State.” Journal of European Social Policy 25 (1): 50–75. Godechot, Olivier. 2016. “Financialization Is Marketization! A Study of the Respective Impacts of Various Dimensions of Financialization on the Increase in Global Inequality.” Sociological Science 3: 495–519. Goldin, Claudia, and Lawrence F. Katz. 2008. The Race between Education and Technology. Cambridge, MA: Belknap Press of Harvard University Press. Gottschalk, Peter, and Timothy M. Smeeding. 1997. “Cross-National Comparisons of Earnings and Income Inequality.” Journal of economic literature 35 (2): 633-687. Hope, David, and Angelo Martinelli. 2017. “The Transition to the Knowledge Economy, Labour Market Institutions, and Income Inequality in Advanced Democracies.” The London School of Economics and Political Science: International Inequalities Institute. Working Paper 18 (November). Huber, Evelyne, Jingjing Huo, and John D. Stephens. 2017. “Power, Policy, and Top Income Shares.” SocioEconomic Review, Early view. DOI: 10.1093/ser/mwx027. Huber, Evelyne, and John D. Stephens. 2014. “Income Inequality and Redistribution in Post-Industrial Democracies: Demographic, Economic and Political Determinants.” Socio-Economic Review 12 (2): 245– 267. Huber, Evelyne, and John D. Stephens. 2014. “Partisan Impacts on Social Policy and Distributive Outcomes in the Eras of Welfare State Expansion and Retrenchment.” Paper delivered at the meetings of the American Political Science Association, Washington, DC. August 28-31. Huber, Evelyne, Jacob Gunderson, and John D. Stephens. 2018. “Private Education and Inequality in the Knowledge Economy.” Paper delivered at the East Asian Social Policy Research Network Annual Conference, Bristol, UK, July 5-6. Jaumotte, Florence, and Carolina Osorio Buitron. 2015. “Inequality and Labor Market Institutions.” International Monetary Fund. IMF Staff Discussion Note (July). Jovicic, Sonja. 2016. “Wage Inequality, Skill Inequality, and Employment: Evidence and Policy Lessons from PIAAC.” IZA Journal of European Labor Studies 5 (1). Karcher, Sebastian, and David A. Steinberg. 2013. “Assessing the Causes of Capital Account Liberalization: How Measurement Matters.” International Studies Quarterly 57 (1): 128–137. Kenworthy, Lane. 2001. “Wage-Setting Measures: A Survey and Assessment.” World Politics 54 (1): 57–98. Kenworthy, Lane and Jonas Pontusson. 2005. “Rising Inequality and the Politics of Redistribution in Affluent Countries.” Perspectives on Politics 3 (3): 449-471. Kilpi-Jakonen, Elina, Daniela Vono de Vilhena, and Hans-Peter Blossfeld. 2015. “Adult Learning and Social Inequalities: Processes of Equalisation or Cumulative Disadvantage?” International Review of Education 61 (4): 529–546. Korpi, Walter. 1983. The Democratic Class Struggle. London, Boston: Routledge & K. Paul. Krippner, Greta R. 2011. Capitalizing on Crisis: The Political Origins of the Rise of Finance. Cambridge, Mass: Harvard University Press. Nickell, Stephen. 2004. “Poverty and Worklessness in Britain.” The Economic Journal 114 (494): C1–C25. OECD. 2013. OECD Skills Outlook 2013: First Results from the Survey of Adult Skills, OECD Publishing. Pierson, Paul. 2003. “Big, Slow-Moving, and …Invisible: Macrosocial Processes in the Study of Comparative Politics.” In Comparative Historical Analysis in the Social Sciences, edited by James Mahoney and Dietrich Rueschemeyer, 177-107. Cambridge: Cambridge University Press. Pontusson, Jonas, David Rueda, and Christopher R. Way. 2002. “Comparative Political Economy of Wage Distribution: The Role of Partisanship and Labour Market Institutions.” British Journal of Political Science 32 (2): 281–308. Rueda, David, and Jonas Pontusson. 2000. “Wage Inequality and Varieties of Capitalism.” World Politics 52 (3): 350–383. Stephens, John D. 1979. The Transition from Capitalism to Socialism. London: Macmillan Education UK. Traxler, Franz. 1994. “Collective Bargaining: Levels and Coverage.” OECD Employment Outlook 1994: 167-194. Tomaskovic-Devey, Donald and Ken-Hou Lin. 2011. “Income Dynamics, Economic Rents, and the Financialization of the U.S. Economy.” American Sociological Review 76 (4): 538–559. Van Reenen, John. 2011. “Wage Inequality, Technology and Trade: 21st Century Evidence.” Labour Economics 18 (6): 730–741. Vernby, Kåre, and Karl-Oskar Lindgren. 2009. “Estimating Dynamic Panel Models when there are Gaps in the Dependent Variable.” Studies in Political Economy and Welfare Working Paper, Uppsala Universitet, Department of Government. Visser, Jelle. 2013. ICTWSS Data base. Version 5.0. Amsterdam: Amsterdam Institute for Advanced Labour Studies AIAS. 251

Vlandas, Tim. 2018. “Coordination, Inclusiveness and Wage Inequality between Median- and Bottom-Income Workers.” Comparative European Politics 16 (3): 482–510. Volscho, Thomas W., and Nathan J. Kelly. 2012. “The Rise of the Super-Rich: Power Resources, Taxes, Financial Markets, and the Dynamics of the Top 1 Percent, 1949 to 2008.” American Sociological Review 77 (5): 679– 699. Wallerstein, Michael. 1999. “Wage-Setting Institutions and Pay Inequality in Advanced Industrial Societies.” American Journal of Political Science 43 (3): 649-680. Weisstanner, David, and Klaus Armingeon. 2018. “How Redistributive Policies Reduce Market Inequality: Education Premiums in 22 OECD countries.” Socio-Economic Review. Early view. DOI 10.1093/ser/mwy018. Wood, Adrian. 1994. North-South Trade, Employment, and Inequality: Changing Fortunes in a Skill-Driven World. Oxford: Oxford University Press. Wood, Adrian. 2018. “The 1990s Trade and Wages Debate in Retrospect.” The World Economy 41 (4): 975–999.

252

Lay explanations for poverty: A multilevel analysis of European public opinion (1976 – 2014) Lionel Marquis1

Abstract: Unlike ‘expert’ approaches to socioeconomic inequalities, ‘lay explanations for poverty’ are the explanations provided by ordinary people to account for the existence and persistence of poverty in contemporary societies. Depending on people, the alleged causes of poverty can stress either individual or social failures, while at the same time pointing out either a ‘blame’ or a ‘fate’ dimension. Following the 2008 economic crisis, ‘social blame’ explanations relating poverty to social injustice have risen to unprecedented levels in most European countries. However, there is no consensus on the individual and contextual underpinnings of lay explanations for poverty, and little is known about how these explanations change in response to socio-economic hardship. Using data from Eurobarometer surveys (1976‒2014), our analysis suggests that lay explanations for poverty are largely driven by selfinterest considerations, whereby less well-off individuals are more likely to explain poverty by social injustice. Nevertheless, the post-2008 trend toward social injustice explanations is remarkably similar in all social groups. If anything, there is evidence of a ‘normalization’ pattern, whereby the crisis seems to have brought social groups closer to one another in their poverty attributions.

1

Introduction

Why are some people poor? This apparently simple question has vexed generations of philosophers, economists, politicians, social scientists, and psychologists. Likewise, in their daily lives, countless ordinary people have been pressed or willing to speculate about the causes of poverty. These lay explanations of poverty are the main subject of this contribution. From research on the lay explanations of poverty, three broad questions frequently arise. First, what are the main types of poverty explanations, and how are these explanations structured among mass publics? Second, what are the main determinants of poverty explanations, and how much do these explanations differ across individuals, social groups, and contexts? Third, do these lay explanations of poverty change over time and why? In the empirical part of this contribution, we take up the second and third questions, asking more specifically how individual and contextual variables influence poverty explanations among

1

Université de Lausanne [email protected]

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 R. Careja et al. (eds.), The European Social Model under Pressure, https://doi.org/10.1007/978-3-658-27043-8_15

253

ordinary citizens, and whether and how the 2008 financial and economic crisis affected these explanations. To some extent, changes in popular perceptions of poverty go hand in hand with other attitudinal developments initiated by the crisis, such as decreasing trust in national and European institutions, weakened support for EU membership, or declining satisfaction with democracy (Armingeon and Guthmann 2014; Armingeon and Ceka 2014; Armingeon et al. 2016a). However, observing changes in poverty attributions brings additional light on the consequences of the economic crisis and suggests that these cannot be reduced to a trend toward increased euroscepticism.

2

Theoretical frame

2.1

The structure of lay explanations for poverty

One of the most discomforting experience for many people is to be confronted with the poverty of other people. A worse experience is being poor, of course, and a challenge of this paper is to take both situations into account – the understanding of others’ and one’s own poverty. In the early 1970s, Joe Feagin (1972) published an article under the provocative title ‘Poverty: We still believe that God helps those who help themselves’. Working on US data, Feagin thereby stressed the tendency of Americans to ascribe poverty to individualistic causes such as lack of effort. This pioneering work sparked relentless debate about lay explanations for poverty (hereafter LEP), asking how these explanations are structured among mass publics. Feagin (1972) identified three main types of explanations which lay people provide for poverty: (1) Individualistic: personal characteristics (such as a lack of effort and willpower, lack of thrift, loose morals and drunkenness) are assumed to be the main driver of poverty; (2) Structural: the larger socioeconomic environment (e.g., poor education system, low wages and unemployment, prejudice and discrimination against the poor) is considered primarily responsible for people’s welfare and their risk of falling into poverty; (3) Fatalistic: poverty is seen to be caused by circumstances which are beyond the control of both the individual and the society (bad luck, disability and illness, lack of ability and talent). Building on this early effort, later research (e.g., Feather 1974; Smith and Stone 1989) has confirmed the distinction between individual and social causes; however, it has also pointed out that items designed to measure the ‘fatalistic’ explanation of poverty do not cohere into a single dimension (e.g., Furnham 1982). Taking the issue further, van Oorschot and Halman (2000) made clear that a fourth type should be added to Feagin’s three-tier classification, namely a ‘social fate type’. The outcome of adding the social fate explanation to Feagin’s initial typology is shown in Table 1. Studies using a similar measurement strategy (e.g., Lepianka et al. 2009; Kainu and Niemelä 2014) confirmed the relevance of the four-tiered typology. Crucially for our empirical investigation, this fundamental structure of poverty attributions has been 254

implicitly adopted by the Eurobarometer survey since it started asking about LEP in the 1970s. Table 1: The two-dimensional structure of lay explanations of poverty

Blame (agency)

Fate (non-agency)

Individual level

Social level

Individual blame (the poor are lazy, lack thrift or good morals)

Social blame (the poor are victims of the actions of others/of social injustice)

Individual fate (the poor are unlucky)

Social fate (the poor are victims of uncontrollable societal and global developments)

Source: van Oorschot and Halman 2000, 7, Table 1.

2.2

The determinants of lay explanations for poverty

Over the years, considerable evidence has accumulated about the personal and structural determinants of LEP. In this overview, we will distinguish between three classes of determinants: sociodemographic, psychological, and contextual. Sociodemographic determinants. Already Feagin’s (1972) study suggested that LEP were associated with individuals’ background characteristics such as race, age, religion, income, or education (see also Kluegel and Smith 1986; Hunt 1996, 2002). We propose here that sociodemographic determinants of LEP fall into two broad classes. First, variables related to economic and social status (age, gender, ethnicity, social class, occupation, etc.) define individuals’ “objective position in the stratification system” (Kluegel and Smith 1986, 5). This position, together with feelings of relative deprivation among the disadvantaged groups, shapes an individual’s propensity to support or oppose different types of social policies and different accounts of poverty. In general, the higher one’s position in the social stratification, the more likely it is that poverty will be explained by individualistic factors; conversely, more deprived individuals usually emphasize structural or fatalistic factors (e.g., Furnham 2003; Saunders 2003; da Costa and Dias 2015). In other words, LEP are thought to be imbued with self-interest – just like other ‘stratification beliefs’ such as attitudes toward redistribution, inequality, or social justice. Accordingly, people who consider themselves poor, or who perceive themselves at risk of becoming poor, are more inclined to give a structural account of poverty (Kainu and Niemelä 2014). Importantly, however, empirical research has usually revealed only modest effects of sociodemographic variables. This, as we argue below, is mainly due to the intervening role of psychological (in particular ideological) factors. The second set of sociodemographic variables is related to socialization, a process that may either reinforce or counterbalance the effect of stratification variables. For the sake of simplicity, we will focus here on the roles of age and education. In general, it appears that younger people are more inclined to attribute poverty to structural causes, while older people 255

are more likely to stress individualistic causes (Feather 1974; Kluegel and Smith 1986; Kluegel 1987; Hunt 1996, 2002). To elucidate this issue, research has examined how children develop beliefs about socioeconomic inequality, about poverty and poor people, or about the causes and justification of poverty (e.g., Bullock 2006). In this respect, young children have been found to emphasize structural explanations more frequently than individualistic explanations (Chafel and Neitzel 2005), suggesting that later educational and professional experiences tend to foster a change in beliefs toward more individualistic accounts of poverty. However, investigating the effect of education leads to a set of contradictory expectations (see McClosky and Zaller 1984, 97-100; Kluegel and Smith 1986, 26, 297-299; Hunt 1996). On the one hand, education is a correlate of economic and social status and, as such, it should promote the endorsement of individualistic explanations for poverty. On the other hand, education may have an “enlightening effect”, by fostering attention to low-key information enabling individuals to discern their “real interest”, or by increasing “awareness of inequality and compassion toward the disadvantaged” (Hunt 1996, 296). Psychological determinants. Research has explored how attributions of poverty are influenced by values, norms, personality traits, and other well-entrenched dispositions. Overall, this research suggests that the effects of sociodemographic variables on LEP can be blurred because general worldviews, self-attributions for success and failure, and affect toward poor people do not overlap with membership in social groups. One psychological disposition that has received much attention is ‘ideology’. The influence of ideology on LEP is so strong and robust that it may offset the effects of many other variables – including the effects of the sociodemographic variables reviewed above. This seems to be especially the case in the United States where, in a nutshell, “the dominant ideology of individualism and the political events of the moment have virtually negated the effects of socioeconomic interests on beliefs about wealth and poverty” (Smith and Stone 1989, 103). So powerful is the effect of this “dominant ideology” 2 that it pervades every sector of society, including the more disadvantaged groups which might be expected to oppose the doctrine of individualism were it not for some kind of ideological “indoctrination” (e.g., McClosky and Zaller 1984, chap. 5). Even in times of economic strain, when whole economic sectors are plagued by unemployment and decreasing living standards, lower and middle strata do not engage in protest actions and show strikingly little disagreement with the economic system (Schlozman and Verba 1979; Kenworthy and Owens 2011). To be sure, the “dominant ideology” thesis is very much focused on the American context. There are reasons to expect strong cultural differences across countries; but whatever the type of ideology prevailing in a country, it may be expected to obscure the effects of sociodemographic variables. 2

The term “dominant ideology” was coined by Kluegel and Smith (1986), but many other designations found in the literature all point to the same phenomenon, whether it be “meritocracy”, the “American ethos” (McClosky and Zaller 1984), the “metatheory” of individualism (Smith and Stone 1989), or a “belief in the American Dream” (Schlozman and Verba 1979). 256

Other psychological variables of interest include the tendency to believe (or not to believe) in “a just world” (e.g., Lerner and Miller 1978), the “locus of control” (e.g., Heaven 1989), and, broadly speaking, the affective orientation people have toward the poor (e.g., Cozzarelli et al. 2001). Although they are important at a theoretical level, these variables are unfortunately not available in the dataset which will serve as the backbone of our empirical investigation. Therefore, they will not be used for hypothesis generation and in empirical analyses, and they will not be considered further here. Aggregate-level determinants. Most of the psychological determinants of LEP discussed above are related to general circumstances that can (and do) vary through space and time. For example, the ‘ideology of individualism’ prevailing in the United States may not be so dominant in other countries. Likewise, the severity of social inequality and poverty problems are themselves highly context-dependent. Although contextual variables have long been connected with the issue of poverty, they have only recently been incorporated into research on poverty attributions. In this brief review, we distinguish between four classes of contextual variables relevant to the study of LEP. The first category is made up of political and institutional variables, including the type of welfare regime and welfare tradition, and various indicators of the ‘quality of democracy’ (Larsen 2008; Lepianka et al. 2010; Kallio and Niemelä 2014; Kainu and Niemelä 2014). Overall, this category of variables yields poor results, with the possible exception of the level of social spending (Kallio and Niemelä 2014, Model 7). The second category is comprised of socio-economic factors which, by and large, ascertain the degree to which poverty is a serious issue in a given population (e.g., Lepianka et al. 2010; Merolla et al. 2011; Kainu and Niemelä 2014; Da Costa and Dias 2014). On the whole, these socio-economic indicators do a fairly good job at predicting the relative share of individual and structural attributions of poverty at the aggregate level; for example, individualistic explanations of poverty are more widespread when economic growth is greater and when income inequalities are more limited (Lepianka 2007, chap. 4; Kallio and Niemelä 2014). The third category is made up of cultural and ideological factors, such as shared sets of norms, values, and beliefs. Admittedly, empirical research on the influence of cultural variables is extremely scarce; one study (Lepianka et al. 2010) found an effect of Catholic presence (the more Catholics in a population, the more poverty is seen as caused by ‘social injustice’), but this result awaits empirical confirmation. Finally, the fourth category focuses on measures of exposure to (and awareness of) poverty, mostly at the local level. Available indicators include the actual size of the homeless or poor population (Lee et al. 2004; Hopkins 2009), and the perceived importance of poverty at the national level (Lepianka 2007). Studies using these measures have yielded promising results, allowing to see how the issue of poverty is constructed and framed by shared perceptions and narratives at the community level. By definition, a common feature of all these contextual variables is that they describe an aspect of the reality of poverty that lies outside of the personal realm. However, the entire 257

literature reviewed above has been concerned with variations across spatial units (usually between countries). In contrast, variations over time have rarely been considered; causes of changes in LEP are thus a blind spot in poverty-attribution research. 2.3

Changes over time in LEP: Lessons from the economic crisis

Changes over time in poverty attributions may result from several distinct mechanisms. To begin with, ‘compositional effects’ may occur when shifts in the composition of the population – for example, changes in the age, class, or educational structure – reinforce or diminish the size of social groups, which are biased toward a particular explanation of poverty. Alternatively, time variations in LEP may be caused by changes in macro-level variables of the sort examined in the previous section (e.g., economic growth, unemployment rate). This is the explanation favored by the few studies that compared LEP over time (e.g., Kluegel 1987; Saunders 2003). Importantly, while this ‘societal change’ explanation can be confounded with compositional effects when long-term change patterns are observed, it is the only relevant explanation of variations in LEP when abrupt changes (occurring in a few years or months) are recorded. The outburst of the 2008 global economic crisis is a crucial event that may help us to test the societal change explanation. Quite expectedly, the consequences of the crisis have been extensively studied in various fields. In the attitudinal domain, these investigations tended to find surprisingly little change in attitudes toward key aspects of democratic systems, welfare state institutions and social spending schemes (Brooks and Manza 2013; Bartels 2013; Soroka and Wlezien 2014; but see Armingeon and Guthmann 2014). Against this background, it is worth stressing that poverty attributions did change in the wake of the 2008 economic crisis. As we will show below, changes in LEP occurred quickly, they were significant and substantial, and they were robust, lasting for the whole 2009-2014 period. Thus, although they went almost entirely unnoticed, these shifts in LEP are more compatible with a ‘societal change’ explanation than with the effects of a compositional trend. An equally interesting question is how (rather than how much) the crisis affected LEP. In this regard, we might expect, first, that the aggregate variables of the sort discussed in the previous section will tend to shape individual perceptions of the poverty issue. Unfortunately, for reasons to be explained later, our model will not contain any psychological variable that could measure these changes in perceptions – instead, our model comprises only contextual and socio-demographic variables. However, it is likely that the crisis had a different impact on different socio-economic groups, and thus on the ways in which these groups conceive of the poverty issue. In turn, these group-specific shifts in perceptions of poverty should be reflected in variations – comparing pre-crisis to post-crisis periods – in the effects that group membership exerts on poverty attributions. According to the theories presented above, two contradictory hypotheses may be formulated to account for these variations. First, we may elaborate on the self-interest hypothesis presented 258

in section 2.2 and make the case for a polarization process. If poverty attributions are mainly driven by individuals’ private or group interests, then it follows that members of groups which were most severely hit by the crisis – in terms of unemployment, job insecurity, declining incomes, or shrinking social benefits – should be motivated to make external attributions of poverty such as the ‘social injustice’ argument. Conversely, individuals in social categories that remained well off or relatively unaffected by the crisis may continue to view poverty as “the problem of the poor” – had they put more energy in education and work, they would have found a way out of their personal problems (Kluegel 1987). In sum, the self-interest/polarization hypothesis predicts that the crisis has increased the gap between socially advantaged and disadvantaged groups as regards internal vs. external attributions of poverty. Second, we can specify a normalization hypothesis, stating that poverty attributions by various groups have tended to converge as a result of the crisis. This hypothesis is obliquely derived from the observation that many individuals, especially in certain social groups, are deeply ambivalent toward the causes of poverty and wealth (Hochschild 1981; Kluegel and Smith 1986). In short, ambivalent individuals relate poverty to both individual (e.g., laziness) and structural (e.g., social discrimination) factors. In the United States, this phenomenon of “split consciousness” (Kluegel et al. 1995) or “dual consciousness” (Hunt 1996, 2016; Merolla et al. 2011) is widespread, but it tends to be concentrated among disadvantaged groups such as racial minorities, whereas it is less prevalent among well-educated and high-income groups. In Europe, in contrast, we hypothesize that the most ambivalent individuals are not drawn from disadvantaged groups (which tend to see poverty as a structural problem), but rather from wide sectors within the middle strata or ‘classes’ (employees of the private sector, middle-income groups, well-educated younger cohorts in lower or precarious positions, etc.). Now, economic strain may be expected to reduce ambivalence by fostering support for structuralist views of poverty (Hunt and Bullock 2016, 95). Accordingly, it follows that the 2008 economic crisis should have bridged the gap between lower strata (which were already leaning toward structuralist attributions) and middle strata. Note, however, that a test of the dual consciousness hypothesis is not strictly possible in studies – like our own – which are based on the above fourtiered typology of poverty explanations and where the choices of such explanations are mutually exclusive. 2.4

Hypotheses

For reasons of space limitation, the remainder of this chapter will make two simplifying assumptions. On the theoretical side, we will formulate our hypotheses in rather general terms, rather than specifying the effects expected from every single variable used in the empirical analysis. For the hypothesis concerning changes in LEP, we stick to the selfinterest/polarization perspective – in full awareness that, as our above discussion has shown, other consequences of the economic crisis may be expected. On the empirical side, we will 259

content ourselves with a contrast between the ‘social blame’ cause of poverty (i.e., social injustice) and all other types of causes. As a matter of fact, the social blame category is the most important attribution of poverty causes in Europe since the Eurobarometer survey started investigating LEP in 1976; thus, it is a central component of poverty attributions in the European context. H1 (self-interest): Social groups with an objectively lower position in the stratification system (e.g., low-income groups, employees and workers as compared to managers and professionals) are more likely to express a social blame explanation of poverty. H2 (life cycle/socialization): Younger individuals are more likely to make social blame attributions than older individuals; likewise, the probability of making social blame attributions rises as a function of education. H3 (culture): Inhabitants of urban areas are more likely to make social blame attributions than are inhabitants of rural areas, where a more traditional culture (work ethic, religiosity, exposure to weather conditions in the primary sector, etc.) induces people to individualistic and fatalistic explanations. H4 (aggregate-level effects): Spatial and temporal variations in macro-level indicators indicating heightened exposure to socio-economic hardship (e.g., higher unemployment) favor the choice of social blame attributions, while variations indicating positive socio-economic outcomes (e.g., higher economic growth) will decrease such attributions. H5 (self-interest/polarization): An economic crisis widens the difference (assumed by H1) between social groups with a lower position and those with a higher position in the stratification system as regards their probabilities of attributing poverty to a “social blame” cause.

3

Methods

The Eurobarometer (EB) survey series, initiated in 1973 by the European Commission, offers a unique opportunity of monitoring changes in the lay explanations of the causes of poverty over a long period of time. We selected all surveys that contained a standardized question about the causes of poverty, leaving us with eight surveys spanning a period of nearly 40 years (19762014). 3.1

Dependent variable

Fortunately, the questions asking respondents for their explanations for poverty (LEP) and the answer options were stable (or changed only slightly) over the whole period. In addition, the four answer options offered to respondents closely correspond to the four-tier typology discussed in the theoretical section (in italics below):

260

Q: Why in your opinion are there people who live in poverty? Here are four opinions: which is closest to yours? 1. Because they have been unlucky (individual fate); 2. Because of laziness and lack of willpower (individual blame); 3. Because there is much injustice in our society (social blame); 4. Because it’s an inevitable part of progress (social fate). A fifth category (in addition to ‘don’t knows’) was created for respondents who spontaneously claimed that “none of these” options reflected their true opinion on the question. Figure 1 enables us to follow the evolution of the five answer categories through the years. As one can see, the ‘social injustice’ attribution attains 50 percent of respondents for the first time in the latest survey (2014), while the other three substantive categories receive less than 15 percent of responses. There is a sharp increase in injustice-based explanations (and decrease in other explanation types) between 2007 and 2009, which is obviously related to the onset of the global financial and economic crisis in late 2008 and early 2009. In the thirty months separating the 2007 and 2009 surveys, the share of social blame responses increased by a full 9 percentage point (and by more than 5 percent in all but three countries).

Figure 1: Lay explanations of poverty over time (EU member countries, 1976-2014) Note: For comparison purposes, Figure 1 is based only on data from the nine countries which participated in all eight surveys (F, B, NL, D, I, L, DK, IRL, UK). However, results are very similar if one considers data from all countries as they have joined the EC/EU, and thus also the Eurobarometer survey, through the years.

261

3.2

Independent variables

Individual-level determinants of LEP are measured in the EB surveys. The selection of indicators was constrained by their availability and comparability over time and across surveys. Unfortunately, this prevented us from using potentially relevant socio-demographic, psychological and political variables – such as religious affiliation, religiosity, social trust, and trust in political institutions – because they were not available in the whole period under investigation and were ruled out from the present analysis. This leaves us with a total of six indicators at the individual level: age, sex, education, income, occupation, and residence community size. As for macro-level determinants of LEP, we take into account ten of them, measured from various data sources and encompassing a broad range of socioeconomic issues related to poverty. Table 2 provides a description of all variables included in the model tested in the empirical section.

3.3

Empirical model and estimation methods

The structure of the data from the EB surveys corresponds to a three-level structure, where individuals (level 1) are nested in surveys (level 2), which are themselves nested in countries (level 3). The second level in this hierarchy corresponds to the 157 national surveys conducted at different moments in time, while the third level corresponds to the 27 EUmember countries having participated in several consecutive surveys.3 This country/survey/ individual hierarchy is admittedly less optimal than fully cross-classified models, but it is clearly preferable to two-level and other three-level models given our expectation that most variation in LEP occurs between countries rather than over time (see Schmidt-Catran and Fairbrother 2015). Our model does not ignore the time dimension, though. First, temporal developments are accounted for by the variations in level-2 aggregate indicators (e.g., unemployment rates). Second, as explained above, the ‘crisis’ variable is intended to capture a critical juncture in poverty attributions. We will assess both its main effect on LEP (i.e., expecting an overall rise in structural attributions) and its (cross-level) interactions with some sociodemographic variables (according to Hypothesis 5). In doing this, we also estimate random effects for all individual-level variables involved in the cross-level interactions. As we go along, we will start with an empty model specifying only random intercepts at the survey and country levels (Model 0), to which we will add individual-level variables (Model 1), then macro-level variables (Model 2), and finally cross-level interactions (Model 3). However, this latter model might be criticized on the grounds that it tends to compare the incomparable. 3

Croatia, although it accessed the EU in 2013, only participated in the 2007 and 2014 surveys (but not in 2009 and 2010); it will not be taken into consideration here. 262

As European countries joined the EU in successive accession waves, the population surveyed by the Eurobarometer has been changing, incorporating more and more countries (mainly from Central and Eastern Europe) confronted with more severe poverty issues. Hence, any change detected in poverty attributions and in their relationships with individual-level variables is arguably due, in part, to changes in the nature and composition of the Eurobarometer sample. In order to control for this possible bias, it is desirable to work on a “rectangular” [countries × years] matrix, to ensure that every wave of the Eurobarometer survey contains the same set of countries. Basically, there are two ways to obtain such a rectangular matrix. First, in Model 4 we select only countries that participated in all surveys from 1976 to 2014, which boils down to nine Western European countries (F, B, NL, D, I, L, DK, IRL, UK). Second, in Model 5 we select only the last four surveys (2007, 2009, 2010, and 2014), in which all 27 EU members participated throughout. Models 4 and 5 replicate the structure of Model 3 in every aspect excepting for the empirical basis of years and countries. Each of these two models has its own strengths and drawbacks, and they should be used in a complementary manner.

263

264

0(18.1%): up to 14 yrs.; 1(21.2%): 15-16 yrs.; 2(24.1%): 17-18 yrs.; 3(18.1%): 19-21 yrs.; 4(18.5%): 22 yrs. and more [-3.60 ‒ 17.30] M=0.05, SD=1.03

Education

[0.3 ‒ 26.5] M=8.6, SD=3.9

[14.1 ‒ 27.9] M=20.0, SD=2.4

[-14.8 ‒ 11.1] M=1.3, SD=3.2

[-4.5 ‒ 18.0] M=2.5, SD=2.8

[50.7 ‒ 92.6] M=79.5, SD=8.3

[70.9 ‒ 83.2] M=78.8, SD=2.9

Unemployment rate (%)

Government spending (%)

Economic growth (%)

Inflation (%)

Globalization (1-100 index)

Life expectancy (yrs.)

Crisis

Residence community size

Occupation

Income (standardized)

1(7.4%): self-employed; 2(6.5%): professional/manager; 3(12.7%): middle management; 4(27.0%): employee; 5(19.2%): skilled worker; 6(10.3%): unskilled worker; 7(16.8%): other 1(34.3%): rural; 2(39.6%): middle town; 3(25.8%): large town 0(48.8%): pre-crisis; 1(51.2%): post-crisis

—

Age

Sex

Distribution: % for categorical variables; [range] for continuous variables 0(50.7%): male; 1(49.3%): female

Variable

186’354 (100.0)

186’354 (100.0)

186’354 (100.0)

186’354 (100.0)

186’354 (100.0)

186’354 (100.0)

186’354 (100.0)

185’764 (99.7)

186’002 (99.8)

166’249 (89.2)

178’168 (95.6)

186’313 (100.0)

186’332 (100.0)

N (% valid)

KOF Swiss Economic Institute at the ETHZ: Composite indicator assessing the degree of globalization in three subdomains: economic, social, and political World Bank: “Number of years a newborn infant would live if prevailing patterns of mortality at the time of its birth were to stay the same throughout its life”

World Bank: “Percent change in the Consumer Price Index”

EB: Distinguishes the five surveys (1976, 1989, 1993, 2001, 2007) conducted before the onset of the economic crisis in 2008 and the three surveys (2009, 2010, 2014) conducted afterwards ILOSTAT: Seeking employment and “available to take up employment given a job opportunity” World Bank: Government’s “final consumption expenditure”, expressed in relative terms as % of GDP World Bank: Rate of change of real GDP

EB

EB: Measured at the level of households, standardized (with M=0 and SD=1) within each year and country EB

EB: Age at the time respondent finished her full-time education

EB

Eurobarometer (EB)

Main source and short description of measure (non-recoded)

Table 2: Descriptive analysis of all independent variables (design weights and population weights applied)

265

186’354 (100.0)

[47.4 ‒ 97.8] M=73.9, SD=8.3 [0.0 ‒ 97.8] M=51.2, SD=34.3 [0.0 ‒ 94.8] M=20.7, SD=23.1

Catholics (%)

Protestants (%)

World Health Organization: Number of suicides per 100,000 population, age-standardized United Nations Population Division: Percentage of “people living in urban areas as defined by national statistical offices” The Cline Center for Democracy: Number of people from the respective religious affiliation as a % of the total population Idem

Note: On occasion, auxiliary data sources had to be used to supplement missing data points in the main sources. All data sources and details about operationalization of variables can be obtained upon request to the author.

186’354 (100.0)

186’354 (100.0)

186’354 (100.0)

[0.0 ‒ 30.4] M=9.3, SD=4.1

Suicide mortality (per 100K) Urbanization (%)

4

Empirical results

The results of our empirical tests are shown in Table 3.4 Starting with Model 1 as a reference base, most fixed effects of individual-level determinants of LEP over the whole survey period appear quite significant, if not very substantial. First, women are more likely than men to support a socially-based explanation of poverty. Interestingly, the effect of sex is unaffected by the introduction of aggregate-level variables (Model 2) and by its (nonsignificant) interaction with the crisis period (Models 3‒5), suggesting that it is both robust and stable over time. Second, the polynomial function of age, when translated into probabilities, shows that social injustice attributions remain stable at a relatively high level until the age of about 60, before falling sharply among older populations. Because this relationship is preserved (with only minor changes) in all models, including in the recent period covered by Model 5, the effect of age is more consistent with the outcome of an ageing process, though an extended generation effect cannot be ruled out. Third, education has surprisingly no effect on LEP, at least in Models 1 and 2. However, Models 3 to 5 suggest that this is an aggregation bias and show a striking reversal in the effect of education between pre-crisis and post-crisis periods. Initially, the positive effect of education on the probability of providing a social blame explanation of poverty is broadly consistent with our socialization hypothesis. After the outburst of the economic crisis, however, this effect is inhibited, arguably as a consequence of the rapid changes in prevailing socio-economic conditions (see Figure 2A).5 In contrast, the effect of income is entirely congruent with the self-interest hypothesis. The higher the income, the less likely it becomes that poverty will be attributed to social injustice. This effect is remarkably constant over time and across models; if anything, it tends to be reinforced in the crisis period, according to the EU-27 sample for the years 2007-2014 (Model 5). As regards occupation, our results are also quite clear-cut. Self-employed individuals (including farmers and fishermen) are definitely less likely to explain poverty by structural aspects than are people in other occupations. To be sure, Models 1 and 2 suggest that professionals and managers (and, to a lesser extent, people from the residual ‘other’ category) are not so different from self-employed individuals. But closer inspection of Models 3 to 5 shows that all occupational groups were significantly different from the self-employed in the pre-crisis period, before the crisis apparently induced a convergence between the self-employed 4

All analyses performed in Table 3 are based on weighted data, using the design weight provided by the Eurobarometer survey. The population weight also provided by the EB is intended to correct for under- and oversampling, by adjusting national samples to each country’s population size in proportion to the overall EU population. Because such weights are, strictly speaking, level-2 weights applying to particular countries at particular moments of time, their use as a component of level-1 weights may lead to misspecifications (see Goldstein 2010); moreover, because our analysis does not aim at descriptive accuracy, we dispense with population weights in the present analysis. 5 Because education and other socio-demographic attributes are included in our models both as ‘main effects’ and in interactions with the crisis variable (respectively in the first and in the last rows of the ‘fixed effects’ section of Table 3), the main effects should be understood as the effects estimated for the pre-crisis years. 266

and both professionals/managers and the residual group (comprised of a large number of retired people), at least in Model 3. Although social blame attributions became more frequent in all occupational groups after 2008, the gap between self-employment and other occupations tended to shrink (see Figure 2B). This is obviously more consistent with a ‘normalization’ process (see section 2.3) than with the polarization process assumed in Hypothesis 5. Models 4 and 5, however, provide less conclusive evidence for a normalization pattern. Lastly, the type of residence area is also clearly related to social blame attributions. Such explanations of poverty are decidedly more frequent among inhabitants of urban areas (from both middle and large towns) than among people living in rural areas, as revealed by coefficients from all models (with the exception of middle-sized towns in Model 4). Large town dwellers are especially likely to provide social blame explanations of poverty, even more so than middle town dwellers.6 Besides, there exists no clear evidence of a crisis effect with respect to community size. Only in Model 5, which focuses on the 2007‒2014 period, does such an effect emerge — social injustice explanations increased less among large town dwellers than among rural inhabitants. Turning now to the effects of aggregate-level variables, most of them appear highly dependent on the data selected for analysis. Among the macro-level variables, only the unemployment rate has basically the same impact on poverty attributions in the whole sample (Model 3) and in the two reduced samples (Models 4 and 5). This effect corresponds to an average 0.9‒1.0 percentage points (depending on models) increase in the probability of providing a social blame attribution for every 1 percentage point increase in the unemployment rate – a strong effect indeed. The second most important contextual determinant of LEP is suicide mortality: the more suicides are committed in a particular context, the more likely it is that an individual from this context will explain poverty by social injustice reasons. Suicide mortality may not have a particularly strong effect, but it is rather robust as it fails to attain statistical significance only in Model 4. Next, economic growth is associated negatively with social blame attributions in Models 3 and 4. Again, the effect is modest but consistent with the idea that economic decline stimulates social blame attributions. The effects of other aggregate predictors are significant only in one model, typically in the long-term EU-9 sample used in Model 4. Such is the case of the globalization index, life expectancy, and the percent of Protestants in the population. Urbanization has some effect only in the short-term EU-27 sample used in Model 5. The three remaining variables (government spending, inflation, and the percent of Catholics) have no effect whatsoever. It is perhaps no coincidence that the aggregate-level variables most consistently predictive of the social injustice explanation of poverty – unemployment, economic growth and suicide rates – have all been related to the concept of ‘anomie’ in the Durkheimian tradition (e.g., Teymoori et al. 2016). Thus, while the economic and moral distress arguably reflected in these 6

Post hoc contrast in probabilities of giving a social injustice explanation (p