Download Book

Contemporary Switzerland This page intentionally left blank Contemporary Switzerland Revisiting the Special Case

Edited by

Hanspeter Kriesi Peter Farago Martin Kohli and Milad Zarin-Nejadan Selection and editorial matter © Hanspeter Kriesi, Peter Farago, Martin Kohli and Milad Zarin-Nejadan 2005 Individual chapters © contributors 2005 Softcover reprint of the hardcover 1st edition 2005 978-1-4039-4798-7 All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No paragraph of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, 90 Tottenham Court Road, London W1T 4LP. Any person who does any unauthorized act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The authors have asserted their rights to be identified as the authors of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2005 by PALGRAVE MACMILLAN Houndmills, Basingstoke, Hampshire RG21 6XS and 175 Fifth Avenue, New York, N.Y. 10010 Companies and representatives throughout the world PALGRAVE MACMILLAN is the global academic imprint of the Palgrave Macmillan division of St. Martin’s Press, LLC and of Palgrave Macmillan Ltd. Macmillan® is a registered trademark in the United States, United Kingdom and other countries. Palgrave is a registered trademark in the European Union and other countries. ISBN 978-1-349-52494-5 ISBN 978-0-230-52358-6 (eBook) DOI 10.1057/9780230523586 This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data Contemporary Switzerland : revisiting the special case / edited by Hanspeter Kriesi . . . [et al.]. p. cm. Includes bibliographical references and index. ISBN 1–4039–4798–8 (cloth) 1. Labor market—Switzerland. 2. Switzerland—Social conditions—1945– 3. Switzerland—Politics and government—1945– I. Kriesi, Hanspeter HN603.5.C64 2005 306′.09494—dc22 2004063296 10987654321 14 13 12 11 10 09 08 07 06 05 Contents

List of Figures vii List of Tables ix Notes on the Contributors xi Preface xviii

Introduction 1 Part I The Swiss Way of Life 1 Relative Deprivation and Well-being: Switzerland in a Comparative Perspective 9 Christian Suter and Katia Iglesias 2 What Pluralization of the Life Course? An Analysis of Personal Trajectories and Conjugal Interactions in Contemporary Switzerland 38 Eric Widmer, Jean Kellerhals and René Levy 3 Facets of Emotion Regulation in Families with Adolescents: A New Research Approach 61 Meinrad Perrez, Dörte Watzek, Gisela Michel, Dominik Schoebi, Peter Wilhelm and Yves Hänggi 4 The Impact of Social Inequalities on Personal Health 81 Monica Budowski and Annette Scherpenzeel

Part II The Swiss Labour Market 5 Long-Term Dynamics of Skill Demand in Switzerland, 1950–2000 105 Stefan Sacchi, Alexander Salvisberg and Marlis Buchmann 6 Information Technology, Workplace Organization and the Demand for Employees of Different Education Levels: Firm-Level Evidence for the Swiss Economy 135 Spyros Arvanitis 7 Young Adults Entering the Workforce in Switzerland: Working Conditions and Well-Being 163 Norbert K. Semmer, Franziska Tschan, Achim Elfering, Wolfgang Kälin and Simone Grebner

v vi Contents

8 The Erosion of Regular Work: An Analysis of the Structural Changes in the Swiss and German Labour Markets 190 Andreas Diekmann and Ben Jann

Part III Political Institutions 9 Internationalization and Domestic Politics: Evidence from the Swiss Case 221 Pascal Sciarini and Sarah Nicolet 10 Reforming the Swiss Municipalities: Efficiency or Democracy? 239 Andreas Ladner and Reto Steiner 11 The Lost Dimension of Swiss Federalism: Democracy Problems of New Regionalism in Metropolitan Areas 256 Daniel Kübler 12 Sustainable Fiscal Policy in a Federal System: Switzerland as an Example Lars P. Feld and Gebhard Kirchgässner 281

References 297 Index 329 List of Figures

1.1 Necessities: percentage of persons considering the listed items ‘absolutely necessary’ for decent living 17 1.2 Actual standard of living: percentage of persons who can afford the listed items 18 1.3 Level of deprivation (mean PDI) by quintiles of equivalent income 21 2.1 Men’s trajectory type ‘Dominant’ 43 2.2 Men’s trajectory type ‘Minority’ 44 2.3 Women’s trajectory type ‘Homemaker’ 45 2.4 Women’s trajectory type ‘Full-time worker’ 45 2.5 Women’s trajectory type ‘Part-time worker’ 46 2.6 Women’s trajectory type ‘Back-to-employment’ 47 3.1 Estimated time-related effects on emotional state, when situational and psychological factors are controlled 68 3.2 Influence of different settings on well-being, deviance from mean 69 3.3 Functionality of interpersonal emotion regulation behaviour in different settings 77 4.1 General three-wave causal model of health 89 5.1 Observed and expected trends in the job market for farmers and salespeople 117 5.2 Observed and expected trends in the job market for office and administrative staff 119 5.3 Observed and expected trends in the job market for unskilled workers 121 5.4 Observed and expected trends in the job market for highly skilled workers 124 A5.1 Advertised jobs for unskilled workers 131 7.1 Stressors and resources at work 167 7.2 Appreciation received at work 169 7.3 Well-being 170 7.4 Worries about the future 171 7.5 Task stressors for five professions 172 7.6 Social support at work for five professions 173 7.7 Job control for five professions 174 7.8 Resigned attitude towards one’s job for five professions 175 7.9 Positive attitude towards life for five professions 176 7.10 Sex differences in well-being 177 7.11 Sex differences in future plans 178

vii viii List of Figures

7.12 Resigned attitude towards one’s job: language differences 179 7.13 Irritability: language differences 180 7.14 Resigned attitude towards one’s job: language differences, controlling for work characteristics 181 7.15 Irritability: language differences, controlling for work characteristics 182 7.16 Self-esteem: language differences 183 7.17 Change of employer and job satisfaction 184 7.18 Change of profession and work characteristics 185 7.19 Change of profession and well-being 186 7.20 Well-being: influences from work, private life, and personality 187 8.1 Regular work in Switzerland from 1991 to 2003 196 8.2 Regular work in Switzerland from 1970 to 2000 198 8.3 Age-dependent rate of regular work in Switzerland from 1970 to 2000 200 8.4 Regular work in Germany from 1985 to 2002 203 9.1 Predicted probabilities of a successful launching of a facultative referendum, as a function of the importance and internationalization of a legislative act 229 10.1 Triggers of reform processes according to assessments by the municipal secretaries (case studies) 249 10.2 Goals of the reform projects (case studies) 253 List of Tables

1.1 Relative deprivation (PDI) in four European countries 20 1.2 Mean of individual and societal well-being and their components 24 1.3 Determinants of individual and societal well-being 26 A1.1 List of variables 31 A1.2 Factorial analysis for well-being 33 A1.3 Determinants of individual well-being 34 A1.4 Determinants of societal well-being 36 2.1 Results of cluster analysis based on responses from both partners 49 2.2 Conjugal conflict, conjugal quality and styles of conjugal interactions 52 2.3 Multinomial logistic regressions modelling the probability of exhibiting a given trajectory-type and a given style of conjugal interactions 54 3.1 FASEM-C information types and item types 65 4.1 Effects on health: standardized regression weights 92 4.2 Stability of the variables in the model 95 A5.1 Overall employment and advertised jobs by industry 130 6.1 Percentage of firms with decreasing, unchanged or increasing shares of employees with different levels of vocational education, 1997–2000 141 6.2 Formal education of employees in Swiss business sector, 1999 141 6.3 Full model: determinants of the employment shares of highly educated, middle-educated and low-educated employees, 1999 146 6.4 Determinants of the employment shares of highly educated, middle-educated and low-educated employees, 1999 150 6.5 Determinants of the employment share of highly educated, middle-educated and low-educated employees 152 6.6 Survey of recent empirical literature 155 A6.1 Composition of the data set 161 7.1 Project description 165 A8.1 Distribution of employment forms in Switzerland from 1991 to 2003 209 A8.2 Distribution of employment forms in Switzerland from 1970 to 2000 212 A8.3 Distribution of employment forms in Germany from 1985 to 2002 (West German states only) 214

ix x List of Tables

9.1 Importance of the phases of the three decision-making processes 226 9.2 Importance of state and non-state actors 232 10.1 Spread of reforms in Swiss municipalities 242 10.2 Explaining different reform activities through characteristics of the municipality 246 10.3 Initiators of reforms according to reform type 251 10.4 The different types of reforms and their goals 253 11.1 Four routes towards metropolitan governance 260 11.2 Institutional fragmentation of Swiss metropolitan areas (in 2000) 266 11.3 Communal tax rates according to types of communes 266 11.4 Types of metropolitan governance found in the four metropolitan areas 270 11.5 Predictors of service satisfaction 273 11.6 Predictors of loyalty towards local authorities 275 11.7 Predictors of satisfaction with local democracy 276 A11.1 List of interviews (qualitative studies) 280 A11.2 Interviews telephone survey 280 12.1 Quantitative impacts of the explanatory variables 289 A12.1 Descriptive statistics of the explanatory variables 295 A12.2 Expenditure, revenue, and deficit (per capita), 1980–1998 296 Notes on the Contributors

Spyros Arvanitis is Senior Research Economist at the Swiss Institute for Business Cycle Research (KOF) of the Swiss Federal Institute of Technology Zurich (ETHZ), where he is head of the Research Group ‘Market Dynamics and Competition’. Dr. Arvanitis holds a doctoral degree in economics from the University of Zurich and a doctoral degree in chemistry from the Swiss Federal Institute of Technology Zurich. He is a lecturer in economics at the ETHZ and has published extensively on the economics of innovation, technology diffusion, firm performance and market dynamics.

Marlis Buchmann is Professor of Sociology at the University of Zurich and the Swiss Federal Institute of Technology. Her major areas of research are sociology of the life course, social stratification and mobility, sociology of work, occupations and labour markets, and social change. She has published widely in these fields with, among others, the University of Chicago Press. Her work has also appeared in many journals, including the European Sociological Review, Kölner Zeitschrift für Soziologie und Sozialpsychologie, Work and Occupations, Poetics, International Journal of Sociology and Schweizerische Zeitschrift für Soziologie.

Monica Budowski is Professor of Social Work and Social Policy at the University of Fribourg (Switzerland). She was formerly senior researcher at the Swiss Household Panel. Her areas of research include social policy, family and gender studies, poverty and health. She has published in various books and journals on these topics, including the Journal of Comparative Family Studies and the Swiss Journal of Sociology.

Andreas Diekmann is Professor of Sociology at the Swiss Federal Institute of Technology in Zurich. He recently moved to ETH-Zurich from the University of Bern. His research interests include experimental game theory, environmental sociology, demography and the labour market, and methodology. His published writings in English include ‘Cooperation in an Asymmetric Volunteer’s Dilemma Game’, International Journal of Game Theory (1993), ‘The Social Inheritance of Divorce’, American Sociological Review (1999, with H. Engelhardt), and ‘The Wealth of Nations and Environmental Concern’, Environment and Behavior (1999, with A. Franzen). More recent publications have concerned reputation and Internet auctions (with D. Wyder) and the international comparison of divorce rates (with K. Schmidheiny). At present, he is working on a project on social norms, reciprocity, and intertemporal choice.

xi xii Notes on the Contributors

Achim Elfering received a PhD in Psychology from the University of Frankfurt, Germany in 1997. Since 1997 he has been a lecturer in Work and Organizational Psychology at the University of Bern, Switzerland. His research interests are concerned with stress and health at work, notably low back pain, and job satisfaction.

Peter Farago completed his studies in Sociology and Political Science at the University of Zurich with a dissertation on neo-corporatism in Switzerland (1986), since when he has been conducting empirical social research. His main focus is on programme and policy evaluation in the fields of public administration, social policy, housing, and transportation. From 1996 to 2003 he was the director of the Swiss National Science Foundation’s Priority Programme ‘Zukunft Schweiz – Demain la Suisse’.

Lars P. Feld is Full Professor of Economics, particularly Public Finance, at the Philipps-University of Marburg, a member of the Council of Scientific Advisors to the German Federal Ministery of Finance, and a member of the CESifo Research Networks. His research interests include public finance, in particular fiscal federalism and fiscal policy in open economies, new institutional economics and public choice; fiscal psychology, in particular tax evasion and tax morals. He has published numerous articles, including in the Journal of Public Economics, Public Choice, Kyklos, Regional Science, Urban Economics and the European Journal of Political Economy.

Simone Grebner received a PhD in Psychology from the University of Bern, Switzerland in 2001. Since 1995 she has been a research and teaching assistant for Work and Organizational Psychology at the Department of Psychology, University of Bern, Switzerland. Her research interests include applied stress research, with a special emphasis on physiological measures.

Yves Hänggi works as an assistant to the Chair of Clinical Psychology (Professor M. Perrez) at the University of Fribourg. His research interests include web-based prevention, in particularly a web-based training for parents in reducing stress (www.elterntraining.ch), studies concerning the method of internet surveys, and the evolutionary psychology of coping behaviour.

Katia Iglesias is a researcher at the Sociological Department of the University of Neuchâtel. Her areas of research include studies on social indicators, well-being and quantitative methods.

Ben Jann is a graduate research assistant at the Department of Sociology of the Swiss Federal Institute of Technology Zurich (ETH) and a PhD candidate at the University of Bern in Switzerland. His main areas of research are Notes on the Contributors xiii the labour market and inequality, environmental sociology, social research methods, and statistics. Recent publications have included ‘Old-Boy Network, Military Service and Professional Success in Civilian Life in Switzerland’ (Zeitschrift für Soziologie, 2003) and Einführung in die Statistik (2002).

Wolfgang Kälin received a PhD in Psychology from the University of Bern, Switzerland in 2004. Since 1996 he has been a research and teaching assistant for Work and Organizational Psychology at the University of Bern, Switzerland. His research interests are on coping, stress and well-being.

Jean Kellerhals is Full Professor of Sociology at the University of Geneva. He has contributed to various dimensions of sociological family research, including conjugal and family functioning, educational strategies within the family, kinship solidarity and distributive justice in small groups.

Gebhard Kirchgässner is Professor of Economics and Econometrics at the University of St Gallen and Director of the Swiss Institute of International Economics and Applied Economic Research at the University of St Gallen. He currently serves as president of the Economic Advisory Board of the Swiss government. He has received several fellowships and academic honours: he is a research fellow in the CESifo Network, a member of the German Academy of Natural Scientists, and of Leopoldina, Section Economics and Empirical Social Sciences. Since 2003, he has served as the president of the European Public Choice Society. His research areas include public choice, environmental and energy economics, applied econometrics, and the methodological foundations of the social sciences.

Martin Kohli is Professor of Sociology at the European University Institute, Florence (on leave from the Free University of Berlin). He is a member of both the Berlin–Brandenburg and the Austrian Academy of Sciences, and from 1997 to 1999 was President of the European Sociological Association (ESA). He has published extensively on the sociology of the life course, generations and aging, on family, labour markets and social policy, and, more recently, also on the emergence of a European society.

Hanspeter Kriesi is Professor of Political Science in the Department of Political Science of the University of Zurich. His current research interests include the study of social movements, political parties and interest groups, the formation of individual political opinions in political campaigns, and direct democratic procedures. He is the author of several books on these and related topics and has widely published in various journals. He has served as the President of the committee of experts of the Swiss National Science Foundation’s Priority Programme ‘Zukunft Schweiz – Demain la Suisse’ (1996–2003). xiv Notes on the Contributors

Daniel Kübler is assistant professor at the Institute for Political Science of the University of Zurich. His areas of research include Swiss politics, comparative urban and local government and governance, public policy analysis and evaluation. He is author of several books and has published in the International Journal of Urban and Regional Research, the Journal of European Public Policy, Addiction, Governance, the Policy Studies Journal and the Swiss Political Science Review.

Andreas Ladner is an Assistant Professor at the Centre of Competence for Public Management at the University of Bern. His areas of research include political parties, municipalities and institutional change. He has conducted several major research projects of the Swiss National Science Foundation and authored books and articles on these topics. He has published articles in leading journals, including West European Politics, Electoral Studies and Party Politics. He regularly comments on Swiss politics in the media.

René Levy is Full Professor of Sociology at the University of Lausanne. He is the director of the Center for Life course and lifestyle studies (Pavie). He has revisited various aspects of life in other research areas, especially stratification, mobility, and family dynamics, in the light of the life course. The relationship between institutional structuration and life course trajectories has also been one of his major research interests.

Gisela Michel is a Research Fellow at the Department of Social and Preventive Medicine at the University of Bern and former research assistant to the Chair of Clinical Psychology at the University of Fribourg. Her areas of research include the subjective experience somatic states, epidemiology of childhood cancer, experience sampling methods and multilevel analysis. She was in receipt of a PhD grant from the Swiss National Science Foundation to study at the New York University, and she achieved her PhD studies in the context of the programme ‘Zukunft Schweiz – Demain la Suisse’.

Sarah Nicolet is a Teaching Assistant at the Department of Political Science, University of Geneva, and also a doctoral candidate at the University of Michigan. Her dissertation focuses on the formation of opinion in European elections. She has published articles on internationalization and Swiss politics in the Journal of European Public Policy, the Swiss Political Science Review, and West European Politics.

Meinrad Perrez is Professor for Clinical Psychology and the head of the Department of Psychology at the University of Fribourg, Switzerland. His research topics concern stress, coping and social regulation of emotions in the context of health, prevention, and family psychology. One particular interest focuses on the development of new assessment tools using Notes on the Contributors xv computer-aided self-monitoring procedures. He is co-editor of different psychological journals and the author of several books and articles in international journals.

Stefan Sacchi is Senior Researcher at the Chair of Sociology at the Swiss Federal Institute of Technology. His research focuses upon labour markets; social stratification and mobility; political value change, and social movements. He is (co-)author of several books and journal articles on these topics in the European Sociological Review, Kölner Zeitschrift für Soziologie und Sozialpsychologie, Schweizerische Zeitschrift für Soziologie, Éducation permanente.

Alexander Salvisberg is researcher at the Department of Sociology at the University of Zurich. His main research interests are in labour markets, the sociology of culture, and social change. He is currently working on his dissertation about the shifting demand for ‘soft’ skills in the labour market.

Annette Scherpenzeel is a psychologist, specializing in methodology and statistics. She works as a freelance senior researcher for the Swiss Household Panel. Her areas of research include structural equation modelling, response effects, quality of life and life satisfaction. She has published in various books and journals on these topics, such as Social Indicators Research and Sociological Methods and Research.

Dominik Schoebi is Research Fellow for the Chair of Clinical Psychology at the University of Fribourg. His main research interests are in studies of interpersonal emotion regulation, conflict regulation and coping with stress in intimate relationships. He collaborated in several research projects on stress and emotion regulation in families, and he coordinates the EU project on ‘Family and Work’. He achieved his PhD studies in the context of the programme ‘Zukunft Schweiz – Demain la Suisse’.

Pascal Sciarini is Professor of Political Science at the Graduate Institute of Public Administration (IDHEAP), Lausanne. He has published on Swiss politics, European integration, comparative political economy and opinion formation in many journals, including the British Journal of Political Science, Comparative Political Studies, the European Journal of Political Research, and West European Politics. His most recent publications include a co-edited book on the 1999 Swiss federal elections.

Norbert K. Semmer received a PhD in Psychology from the Technical University of Berlin, Germany in 1983. He has been head of Work and Organizational Psychology at the University of Bern, Switzerland since 1987. His research interests concern stress at work and its relationship to productivity and health; and efficient work strategies of individuals and xvi Notes on the Contributors groups, how they are acquired spontaneously or taught systematically, and how they relate to issues of quality, errors and safety.

Reto Steiner is a Lecturer in Business administration at the Centre of Competence for Public Management of the University of Bern. He has participated in several research projects of the Swiss National Science Foundation and is an expert in intermunicipal cooperation and amalgamation of municipalities. He has published articles in the International Public Management Journal and in Public Management Review. He is currently consulting and lecturing for the Swiss federal government and various cantons and municipalities, particularly in the areas of public management, educational administration, and local governance.

Christian Suter is Professor of Economic Sociology at the University of Neuchâtel. His areas of research include studies on social inequality and poverty, social indicators and social reporting, health sociology, international political economy and the political sociology of Latin America. He has authored books on global debt crises, economic and political change in Latin America, social change and inequality in Switzerland and social support and health, and has published articles on these topics in Comparative Studies in Society and History, International Studies Quarterly, Social and Preventive Medicine, and in the Swiss Journal of Sociology.

Franziska Tschan received a PhD from the University of Bern in 1990. She has been a Professor of the Social Psychology of Work at the University of Neuchâtel, Switzerland since 1995. Her major research interests are group and team performance, group processes in (medical) emergency situations, social relationships at work, and time aspects in social psychology.

Dörte Watzek works as research fellow to the Chair of Clinical Psychology at the University of Fribourg. Her domain of research is work distribution in families, the assessment of housework and the consequences of workload on health and well-being. She is working towards her PhD studies in the context of the programme ‘Zukunft Schweiz – Demain la Suisse’.

Eric Widmer is Associate Professor of Sociology at the University of Lausanne, with an appointment at the Centre for Life Course and Lifestyle Studies (Pavie). His long-term interests include family relations, life-course research, social norms, and social networks.

Peter Wilhelm is Assistant to the Chair of Clinical Psychology at the University of Fribourg. His research interests concern empathic accuracy in couples and families and cross-cultural psychology. A third area focuses on research methodology. He is collaborating on a number of different research Notes on the Contributors xvii projects. The author of several books and articles in international journals, he achieved his PhD studies in the context of the programme ‘Zukunft Schweiz – Demain la Suisse’.

Milad Zarin-Nejadan is Professor of Economics at the University of Neuchâtel, Switzerland. He is a regular consultant to the Swiss State Secretariat for Economic Affairs, the Swiss Agency for Development and Cooperation and the OECD. His areas of teaching and research include economic policy, public economics, public finance and taxation. Preface

This volume presents a selection of the results of the largest social science research programme ever organized in Switzerland. The editors have chosen 12 from the more than one hundred projects that have been completed in the course of this programme with the intention of giving the international audience an idea of the present state not only of contemporary Switzerland, but also of the state of the art in Swiss social science. Our selection has been guided by considerations of quality, general interest and coherence. We have chosen contributions from excellent projects which concentrate on three salient aspects of contemporary Swiss society: the Swiss way of life, the labour market and political institutions. The programme included many other excellent projects, which are not documented here because they either dealt with other aspects of contemporary Switzerland, have already been documented in other publications of the programme, mainly address a specialized audience or simply because there was not enough space to accommodate them. We hope that this volume will whet the readers’ appetite to learn more about contemporary Switzerland. This publication has been made possible by the Swiss National Science Foundation, which has also financed the entire research programme. We would like to thank the contributors to this volume who have made a special effort to present their findings to an international general public. Karin Bartholomé has relentlessly cleaned and standardized the text material. Thanks also to Amanda Hamilton and her collegues at Palgrave Macmillan for their understanding and professional support.

HANSPETER KRIESI PETER FARAGO MARTIN KOHLI MILAD ZARIN-NEJADAN

xviii Introduction Hanspeter Kriesi, Peter Farago, Martin Kohli and Milad Zarin-Nejadan

Traditionally, outside observers have tended to idealize Switzerland, to admire its standard of living and to glorify its tradition of democratic government. More recently, they have focused rather on the particular qualities of Swiss political culture, to which they have attributed the peaceful cohabitation of different cultures in one and the same country (e.g. Deutsch, 1976; Segalman, 1986; or Steinberg, 1988). Against this background, a recent assessment by The Economist that ‘the Swiss island of calm is becoming a little more like other countries’ is rather new (Beck, 2004). The Economist notes that the Swiss economy has proved as vulnerable as any other to the fluctuations of the business cycle, and is not as uniformly efficient as the Swiss themselves like to believe. If there is one persistently special feature about Switzerland, it is, as far as The Economist is concerned, its political system, its particular mix of political institutions which sets it apart from any other country in the world. For Swiss social scientists, this outside assessment comes as a pleasant confirmation of what they have observed for some time. Over the past eight years, they have been engaged in a large-scale research programme to describe and analyse the transformation of Swiss society. The programme was, ambitiously enough, called Zukunft Schweiz – Demain la Suisse (Switzerland Towards the Future). In this volume, we present the key findings of 12 of the research projects involved in the programme. They do not cover the whole range of issues addressed by the programme, but focus on three key areas – the Swiss way of life, the Swiss labour market, and the country’s political institutions. The 12 chapters give an impression both of current trends in Swiss society and of the scope of this unique programme. How- ever, the contributions to this volume are not exclusively focused on Switzerland. All make some attempt to put their argument in a comparative perspective.

1 2 Introduction

Zukunft Schweiz – Demain la Suisse: a programme for the development of Swiss social sciences

Zukunft Schweiz – Demain la Suisse was intended to promote and reinforce the social sciences in Switzerland. In order to reach this goal, it developed two main types of activities: project-based problem-oriented basic research designed to contribute to the understanding of social, cultural, political, and economic changes in Swiss society and structural measures designed to improve the training of young researchers and data infrastructures. The programme covered the social sciences in a broad sense, but its main concerns were with the disciplines of anthropology, communication science, economics, education, political science, psychology, and sociology. The programme was launched in January 1996. During the eight years of its operation it funded approximately 100 research projects in six domains that were deemed to be particularly important for the future of the country: the research should address societal problems of great public concern. Accordingly, the focus was put on the following six domains or ‘modules’:

• Social inequality: Social inequalities and conflicts. • Labour: The dynamics of the work environment. • Individual and society: Individuality, human relations, and social structure. • Science and Technology: Knowledge production, development of scientific disciplines. • Information and media society: Public sphere and communication in the media society. • Institutional change: Institutional reform at the national, cantonal and local level.

The research activities were organized alongside projects which were, in turn, grouped into a limited number of more encompassing research networks. The networks were typically composed of participants from different disciplines and from various institutions. Each module included several such networks. In addition, the programme also launched a series of structural activities. These include training programmes for young researchers and measures to improve the data infrastructure for the social sciences. As far as training is concerned, three types of initiatives were launched:

• Postgraduate programmes: the programme supported eight postgraduate programmes covering topics in political science, psychology, the media sciences, education, and gender studies. More than 100 students participated in these programmes. • Summer school: The programme created a summer school that offers advanced courses on methods in the social sciences with the intention of Hanspeter Kriesi, Peter Farago, Martin Kohli and Milad Zarin-Nejadan 3

enhancing the methodological know-how of Swiss researchers. The courses cover both quantitative and qualitative approaches. The international fac- ulty consists of renowned specialists. The summer school brings together some 60–80 scholars yearly. Its courses take place at the Università della Svizzera italiana in Lugano. • Scholarships: Personal scholarships have been granted to more than 30 students allowing them to round off their scientific training abroad and to complete their dissertations successfully. Support was given mainly to researchers who contributed new aspects to the topic areas studied in the framework of the programme.

In the area of data infrastructure, the programme initiated and financed the construction of a Swiss Household Panel and it initiated Swiss participation in a series of international social science data collections:

• The Swiss Household Panel Survey: Since 1999 the Swiss Household Panel Survey has conducted annual interviews with the members of approximately 4,000 households in all parts of Switzerland with regard to their life situations (family, work, leisure time, health, etc.). Over successive years, this survey provides basic information on life opportunities, values, and behaviour of individuals and families of varying social backgrounds and, thus, on societal change at the level of individuals and households. From the start, the panel has been designed and conducted in a way that allows for internationally comparative analyses. • Participation in the European Social Survey: The European Social Survey was initiated by National Science Foundations in more than 20 European countries. This survey examines the development and change of attitudes and values in the population on important societal issues (such as the quality of life, national identity, and political orientation). In a representative sample, 2,000 persons in each country are surveyed every second year. • Participation in the International Social Survey Programme: This is a programme of comparative research that has been underway for about 15 years. More than 30 countries participate worldwide. Each year a set of questions on a particular topic is prepared. The topics (such as social inequality, work orientations, religion, political system) are repeated at regular intervals. This programme is currently the broadest comparative database of its kind. Switzerland is now a regular member and is entitled to take part in the discussion of the topics and the design of the modules. • Participation in the Luxembourg Income Study: This is an international database for the topics of income and fortune, in which about two dozen countries participate. Statistical data prepared by the individual countries are processed according to uniform criteria, making them available for comparative analyses. 4 Introduction

In addition to making basic data available, Zukunft Schweiz – Demain la Suisse made a contribution to the dissemination of the information about social transformations in Switzerland. For this purpose, a set of indicators has been designed, which allows for the systematic description of basic societal trends and which is regularly published in a Social Report. This report presents and interprets data covering five main domains: the distribution of resources, politics, social integration, cultural diversity, and ecological integration. The data and graphics are placed on a CD included with the book. They are also available on the Internet. Both the book and the CD are published in a German and a French version. The first edition of the Social Report was published in 2000, the second in 2004 as part of a series Gesellschaft Schweiz – Analyses sociales of the Swiss publishing house Seismo (Suter et al., 2000, 2004). This series has been initiated by the programme and publishes empirically oriented analyses of major trends in contemporary Swiss society (Bühler 2001, 2002; Perrig-Chiello and Höpflinger, 2001; Luginbühl et al., 2002; Eisner et al., 2003; Stamm et al., 2003; Widmer et al., 2003). The social sciences in Switzerland have long been characterized by an uneven development between disciplines and regions. While economics and psychology developed internationally competitive departments at the major Swiss universities across the country some years ago, the core disciplines of the social sciences – sociology, political science, communications science and anthropology – had difficulties in obtaining adequate resources and becoming fully recognized, autonomous disciplines, especially in the German-speaking part of the country. Zukunft Schweiz – Demain la Suisse has contributed to the improvement of the situation in the core disciplines. The current volume documents some aspects of the state of the art of Swiss social science.

The contributions to this volume

Part I of the volume deals with the Swiss way of life more generally. A good way to introduce this part is to note the relative decline of the Swiss economy, as indicated by the GNP per capita, corrected for purchasing power parity. Switzerland is still one of the richest countries in the world, and it is still richer than the average member of the EU or the OECD, but its relative advance over other OECD countries has declined considerably over the past twenty years. It was overtaken by the United States at the beginning of the 1990s and at the beginning of the present century, the Irish caught up with the Swiss and even passed them. This lacklustre performance is largely the result of lagging productivity growth and can be explained by insufficient albeit urgent structural reforms (OECD, 2004). If the Swiss are still quite privileged, their way of life is no longer significantly different from that of their European neighbours or, for that matter, of the population of any other country with a highly developed economy. We shall document the Hanspeter Kriesi, Peter Farago, Martin Kohli and Milad Zarin-Nejadan 5

‘normalization of the Swiss way of life’ in four chapters. First, Christian Suter and Katia Iglesias describe the relative deprivation and well-being of the Swiss in comparison with other Western and Central European experiences. The next three chapters show in more detail the personal trajectories in the Swiss labour market, the conjungal relationships (Eric Widmer, Jean Kellerhals and René Levy), the everyday experience of Swiss families (Meinrad Perrez et al.) and the determinants of personal health in Switzerland (Monica Budowski and Annette Scherpenzeel). The second part of the volume is concerned with the labour market. The Swiss labour market has traditionally been characterized by its flexibility – its social partnership has been exemplary, strikes very rare and unemployment virtually absent. Since the beginning of the 1990s, however, unemployment has also been rising in Switzerland and if it is still much lower than in most other European countries, it has now become a regular feature of the labour market. Two decades ago, Schmidt (1985) investigated the special circumstances that accounted for the particularly low level of unemployment in Switzerland and came to the conclusion that these circumstances were so specific that other countries could not learn from the Swiss experience in this regard. Ten years later, he came to the conclusion (Schmidt, 1995) that these very special circumstances had all disappeared, which he took to be the explanation why unemployment was now rising also in Switzerland. In this part of the volume, the contributions of Stefan Sacchi, Alexander Salvisberg and Marlis Buchmann and of Spyros Arvanitis consider the development of the demand for skills by Swiss employers and its determinants. The chapter contributed by Norbert K. Semmer and his collaborators reports on the experience of young adults entering the workforce. It shows that this experience is less stressful than generally expected and accounts for this result in terms of the peculiar features of the Swiss system of vocational training. The final chapter in this part, by Andreas Diekmann and Ben Jann, compares the normal working experience of the Swiss with the corresponding experience of the Germans and shows that in both countries, there is a tendency to replace standard employment with other forms of labour. Part III of our volume considers political institutions – that aspect of Swiss society which is still considered to be unique. At first sight, the Swiss uniqueness is confirmed: although Switzerland, as a typical case of a small, open economy, is highly integrated into the world market, it is politically a world apart. Traditionally neutral in world affairs, Switzerland joined the United Nations in 2002, but it is a non-aligned island in the midst of an ever-more encompassing European Union. Its population voted against joining the European Economic Area in 1992 and Swiss euroscepticism has been growing ever since. One of the main reasons for the Swiss refusal to join the EU is the attachment of the Swiss to their political institutions and their fear that these could lose their relevance once Switzerland became a member of the EU. 6 Introduction

However, as the four contributions to the last part of this volume show, Swiss institutions are currently changing at all levels of the federal structure, even if the majority of the Swiss are trying to stay apart. Pascal Sciarini and Sarah Nicolet show how political institutions at the national level change as a result of the Swiss embeddedness in its international environment. Andreas Ladner and Reto Steiner present the transformations of local government as a result of institutional reform. Given that local democracy has long been very highly developed, this reform is mainly focusing on improving the level of efficiency of local government. The chapter by Daniel Kübler is concerned with a new level of government – the regional government of metropolitan areas which is becoming increasingly important, although it has no place in the traditional framework of Swiss federalism. If democracy is not the problem at the local level, problems of democratic legitimacy come back with a vengeance at the level of metropolitan coordination between local authorities. Finally, given the far-reaching decentralization within the framework of Swiss federalism, institutions vary from one canton to the other to such an extent that it is possible to analyse the impact of the institutional diversity on fiscal policy in one and the same country. Lars P. Feld and Gebhard Kirchgässner do so in the final chapter of this volume, which can be construed as a defence of the efficiency and legitimacy of the specifically Swiss institutions. The upshot of the analyses in this volume is that while the Swiss are on their way to become Europeans just as their neighbours, they still live in a particular institutional context. The Swiss path to modernity is embedded in a set of institutions, which provide its citizens with a quality of life that is characterized by specific strengths, but also by certain weaknesses which the various chapters will highlight in detail. Part I The Swiss Way of Life 1 Relative Deprivation and Well-being: Switzerland in a Comparative Perspective Christian Suter and Katia Iglesias

Using data from a European welfare survey this contribution examines the actual Swiss standard of living, the degree and distribution of relative deprivation (the lack of socially perceived necessities) and their consequences for subjective well-being within a European context. Although Switzerland has maintained its high level of standard of living, its low level of relative deprivation and its high level of subjective well-being, the differences between Switzerland and the other European countries have become less pronounced. Despite disparities concerning the actual standard of living and a corresponding east–west gradient there is a large consensus among the examined four countries concerning the minimum standard of living regarded as absolutely necessary for a decent life. In all countries relative deprivation negatively impacts on individual well-being whereas societal well-being that concerns the broader social environments of the individuals remains largely unaffected by deprivation, income and other inequality measures.

1 Introduction

The decade of the 1990s ushered in profound changes for the Swiss society and economy. The most severe economic downturn of the postwar era resulted in rising unemployment, growing numbers of welfare recipients and the working poor and also provoked general feelings of economic uncertainty, fears of social declassification and of increasing social inequalities. Accordingly, current scientific and public debates are witnessing an increasing interest in topics of social exclusion, marginalization, deprivation and precarious living conditions. The objective of this chapter is twofold: first, it examines the standard of living and the level of deprivation in Switzerland as compared to neighbouring countries; secondly, it analyses the impact of these ‘objective’ living conditions on subjective well-being. In doing so, our study contributes to three, more general theoretical and methodological debates.

9 10 Relative Deprivation and Well-being

The first of these debates centres on the issue of diversity among the countries of Europe. Switzerland, in particular, has often been characterized as a unique case, not only with regard to its political institutions (e.g. direct democracy, federalism) and its cultural diversity (several languages are spoken), but also concerning its high living standards, comparatively low unemployment and its hybrid welfare state model. There are, however, several indications that Switzerland has experienced some kind of ‘normalization’ in the past years and that it has moved towards the more standard European example as a result of the adverse developments of the 1990s. In the context of our study we ask whether or not Switzerland has maintained its distinct model of standard of living and defended its comparative advantages. More generally, we examine country-specific patterns of actual standards of living and normative assessments of minimum living standards. The second debate entails the conceptualization and measurement of poverty and social exclusion, i.e. the debate on direct and indirect indicators of poverty. There is a large consensus among contemporary poverty and inequality researchers that poverty must be conceptualized in relative terms. According to the definition of poverty established by the European Union,1 the poor are individuals, families and groups whose material, cultural and social resources are so limited as to exclude them from experiencing the minimum acceptable quality of life in the member states in which they live. This definition implies that, because the prevailing standard of living within a community must always be taken into consideration, poverty cannot be defined in absolute, unequivocal terms. More specifically, the measurement of such a minimum standard of living depends upon the personal values of the individuals in question concerning their requirements for maintaining a decent standard of living. The concept of relative deprivation employed in this study allows for the inclusion of such judgements. Two traditions are prominent within the field of poverty research. The first makes use of indirect poverty indicators which are based on the economic resources (income) available to an individual. The second defines poverty directly and focuses on outcomes such as the level of consumption or the standard of living and records the results of the behaviour of individuals after they make use of the resources available to them. The relative deprivation indicators used in this study have several conceptual advantages: They take into account personal preferences, they are suitable for international comparisons and they conceptualize poverty in relative terms, referring to national standards of living. While there is abundant research on indirect poverty measures (e.g. Atkinson etal., 1995; Atkinson, 1998), direct deprivation measures have rarely been used in comparative research. This is due mainly to the scarcity of comparative data sources. Much criticism has arisen from arguments contending that an indirect notion of poverty should not be used to measure a model which employs a direct conceptualization of poverty in terms of exclusion from the customary level Christian Suter and Katia Iglesias 11 of consumption and the common standard of living. It is further argued that the actual living conditions of households with identical financial resources can differ markedly (Ringen, 1988: 358). Indeed, research has shown that the relationship between poverty measured indirectly as low income and measured directly as observed deprivation is rather weak. A sizeable proportion of households classified as poor because of low income do not suffer from deprivation, while some households classified above the official poverty line do experience deprivation (Townsend, 1979; Callan et al., 1993; Halleröd, 1995). This mismatch between income and deprivation measures of poverty can be explained to some extent by the temporal dimension of poverty. Research from Germany has shown that available household income can fluctuate over time so that income poverty must be regarded as a dynamic phenomenon (Leibfried et al., 1995). In many cases, shorter phases of poverty can be bridged by savings or with the aid of monetary or other support from the social environment. In contrast, deprivation is related to longer-term or structural poverty, and is likely to occur whenever individuals have to manage over a lengthier period of time on a low income. So deprivation indicators tend to measure a static situation of being ‘unable to afford things’, whereas income indicators yield a momentary picture of the financial situation. The third debate relevant to our study deals with the relationship between wealth – i.e. favourable ‘objective’ living conditions – and well-being. Are wealthier people happier and more satisfied with their lives than poor people? The relationship between income, poverty and well-being has been investigated by several cross-sectional studies (see Diener et al., 1999 for an overview). Although the effect of higher income on subjective well-being is positive, it is a surprisingly weak phenomenon. There are indications that people thrive well at the lower economic margin of society and that the effects of income increases are not always positive. Thus, several studies could not find a statistically significant relationship between income and well-being (e.g. Clark and Oswald, 1994; Cummins, 2000). This supports more general arguments on the decreasing relevance of traditional inequality and social stratification dimensions in contemporary post-industrial societies. Similarly, recent results of poverty studies suggest that subjective well-being does not increase significantly when the income of poor people is above the poverty line (Leu et al., 1997). While the relationship between income and subjective well-being is positive and strong at the national level, most studies have found a relatively weak correlation at the individual level (cf. Diener et al., 1999; Argyle, 1999). Thus, overall, wealthier societies seem much happier than poor ones, but wealthier people in the economically advanced western societies do not necessarily describe their individual lives as more satisfied than poor people. Based on a reassessment of ten recent studies, Cummins (2000) suggests a curvilinear relationship: the relationship between income and well-being should be stronger for the lowest income levels than for the middle- and 12 Relative Deprivation and Well-being upper-income groups. Furthermore, the impact of income on well-being obviously differs across different domains (such as standard of living, health, housing). Finally, a recent analysis by Christoph and Noll (2003) on variation in the level of satisfaction with the financial situation exhibited in different countries shows that both income-based, indirect measures and consumption-based, direct measures affect well-being. These results suggest that the impact of economic conditions on subjective well-being is more complex and that relative deprivation measures might be an important factor in explaining subjective well-being. Against this background, this chapter examines the level and distribution of deprivation in Switzerland as compared to other European countries and investigates the relationship between deprivation and well-being. More specifically, the following three questions are addressed:

1. What perceptions are there of a minimal acceptable standard of living in Switzerland versus in other European countries and how do these perceptions contrast with the living standards actually achieved by respondents? How does the Swiss concept of the minimum standard of living differ from that of other European countries? 2. What are the levels and distributions of deprivation in the surveyed countries and how do these patterns vary between nations? 3. Does relative deprivation have an impact on subjective well-being? Is relative deprivation a better predictor of well-being than income based, indirect measures?

We have selected three reference societies for our analysis of Switzerland:2 Austria, Germany and Slovenia. Austria and Germany have been chosen because they are the most similar countries to Switzerland with respect to several cultural characteristics, most notably their common language and history, but also similar patterns of consumption and lifestyles. By controlling these general cultural factors the hypothesis of Swiss exceptionalism (or normalization) can be tested more accurately. Compared to Switzerland, Slovenia represents a rather different case, although this country, for historical reasons, has a certain cultural proximity to the German-speaking region of Switzerland. Slovenia has experienced profound social transformations in the last decade (war, the break-up of the socialist society, national independence), its level of welfare is lower and the welfare state is only rudimen- tarily developed. The selection of Slovenia enables us to compare two transformation countries (eastern Germany and Slovenia). Furthermore, standard of living, level of deprivation and their consequences for well-being are examined against the background of a certain west–east gradient (Switzerland–Austria–western and eastern Germany–Slovenia). Thus, average income per capita at purchasing power parity in Switzerland is 10 per cent higher than in Austria, 15 per cent higher than in Germany and 40 per cent Christian Suter and Katia Iglesias 13 higher than in Slovenia.3 There is, however, less income equality in Switzerland than in the other countries.4 Generally, we expect a west–east pattern with only gradual differences between Switzerland, Austria and Germany but substantial differences to exist between these three German-speaking countries and Slovenia. The chapter proceeds as follows: section 2 presents the concept of relative deprivation upon which our analyses are based. Subsequently, we briefly describe the database and the construction of the measure of deprivation (section 3). Having established this, we move on to a descriptive analysis of the standard of living and relative deprivation in the four countries examined (section 4). Section 5 deals with the impact of deprivation on subjective well-being and section 6 summarizes the implications and conclusions of our findings.

2 The concept of relative deprivation

Whereas there is a long tradition of measuring poverty indirectly via income, in recent years there has been more interest in direct measurement. Townsend’s (1979: 413) concept of relative deprivation pioneers this ‘direct’ approach. According to his definition, people may be considered deprived if they lack the type of diet, clothing, housing, activities, facilities and environmental, educational, working and social conditions, which are customary – or at least widely encouraged and approved – in the societies to which they belong. In order to measure relative deprivation Townsend prepared a list of material and cultural indicators to determine whether individuals or households lacked certain commodities or did not participate in various common activities. A summed deprivation index was then generated by summing all the listed items deemed to be missing. Townsend’s study generated a multitude of comments, criticism and further research. One of the main criticisms directed at his approach was that, by giving them the power to select the indicators, it gave the researcher a privileged position (cf. Sen, 1981; Mack and Lansley, 1985). Another issue that has been raised is whether the lack of an indicator can be rated as a deprivation as it is conceivable that some people might choose to do without certain items. Furthermore, since different items may be assumed to hold varying degrees of importance to the welfare of a household, the omission of a weighting scheme has been criticized (Veit-Wilson, 1987). In their 1985 study, ‘Poor Britain’, Mack and Lansley extended Townsend’s approach, particularly in terms of survey technique. In an initial procedure, a random sample of individuals must first assess whether the items on a list of indicators are necessary for maintaining a decent standard of living. Then only those items which have been designated by at least 50 per cent of the respondents as ‘absolutely necessary’ are actually used. Thus, the normative decisions as to which aspects and features of daily life should be used as 14 Relative Deprivation and Well-being indicators for delineating an acceptable minimum living standard should not be made by experts but, rather, ‘by reference to the views of society as a whole’ (Mack and Lansley, 1985: 42). Hence, a consensual definition of poverty is at issue here. Deprivation is defined by Mack and Lansley in terms of an enforced lack of socially perceived necessities (1985: 39). In order to distinguish voluntary renunciation on the basis of personal preferences from enforced renunciation due to scarcity of income resources, in a separate step the respondents are asked not only to state whether they possess a particular item on the list (such as a washing machine), but also, if they do not, whether this item is missing for financial or for other reasons. Only items which are missing for financial reasons are included in the summed deprivation index. Halleröd (1995) objected to this procedure by claiming that it, at best, led to a majority definition of poverty, but it did not garner a consensual definition. He developed an alternative means of measuring deprivation, termed the ‘Proportional Deprivation Index’ (PDI), which is based on the same underlying assumptions as Mack and Lansley’s approach and which still defines poverty as a ‘lack of socially perceived necessities’ but which eliminates the rather arbi- trary classification of items as necessities or non-necessities. The PDI makes use of all items in the questionnaire, including those which only a minority regard as necessary. Each item is weighted depending upon the proportion of the population that regards it as absolutely necessary for the maintenance of a decent standard of living. Thus, the greater the consensus as to whether an item is absolutely necessary, the greater the weight of that item if it is missing. Since the advent of the methods described above, for the most part, deprivation indices have been used in national poverty studies. So far, only very few comparisons between countries have been performed (Halleröd, 1998; Böhnke and Delhey, 1999; Layte et al., 2000). This is due mostly to the scarcity of truly comparable data. Now, for the first time, the European Household Panel has acquired information on the living standards of respondents living within a larger framework. However, assessments by the respondents of which material goods or activities are required to satisfy the minimum standard of living in their respective countries is still lacking. The question now arises as to which deprivation index best satisfies the demands of comparative poverty research. A study by Lipsmeier (1999) showed that even complete renunciation of any form of weighting produces an index which shows such a high correlation with Mack and Lansley’s Consensual Deprivation Index (CDI) and Halleröd’s Proportional Deprivation Index (PDI) that identical results can be expected with their implementation. However, as long as living standards in the various countries still differ to a considerable extent, the use of an unweighted index makes no sense when attempting comparisons between countries. Deprivation, as indicated by the definition of poverty given by the Council of Europe, cited above, should Christian Suter and Katia Iglesias 15 be ascertained relatively in terms of the usual standards of living in the respective nations. Since its weighting can be adjusted according to national preferences, the PDI developed by Halleröd appears to be particularly well suited for use in cross-country comparisons. A missing item will be assigned a different weight in each country depending upon the importance attached to it by the people in that country.

3 Measuring relative deprivation

Our study made use of individual data collected in 1999 and 2000 in Switzerland, Germany, Austria and Slovenia by means of the standardized survey tool known as the Euromodule (WZB, 2002). The Euromodule contains questions both on the minimum acceptable standard of living and also the standard of living that has actually been attained. On the basis of this information, the Proportional Deprivation Index (PDI) defined by Halleröd (1995) can be calculated. This weighted index can then be used to determine the deprivation level of the respondents’ households relative to the common standard of living in the respective country. A list of 19 indicators was used in a first step to determine what the respondents felt is needed for a decent living standard in their respective country. The exact phrasing of the questions was as follows: ‘In your opinion, what items on this list should every household in your country be able to afford? What could be dispensed with, what is desirable but not necessarily needed, and what is absolutely necessary?’ The selection of indicators was chosen based on preceding studies from Great Britain, Sweden and Germany (Gordon and Pantazis, 1997; Halleröd, 1998). In a second step, respondents were asked whether the designated items or activities on the same list of indicators were owned/performed by the respondents personally or by their household: ‘Now if you consider your own living conditions, what do you have or can you do? What don’t you have or can’t you do because you cannot afford it? What don’t you have or do for other reasons?’ A missing item was considered to be an indicator of deprivation in the sense of the definition quoted earlier only if its absence is due to insufficient financial resources. These data on living standards were summarized by the weighted index PDI.5 To render the deprivation indices comparable, the scores were then calculated as percentages of the maximum possible score in each country using a simple linear transformation of the original score. This standardization has the advantage of being neither sample nor population-dependent (Cohen et al., 1999).6 Therefore, the indices constructed in this way vary between 0 and 100. Each value can be interpreted as the percentage of deprivation experienced by a household in relation to the theoretical maximum possible value which can only be obtained only if the household suffers simultaneous deprivation of all items included in the scale. 16 Relative Deprivation and Well-being

4 Standard of living and relative deprivation

4.1 The perceptions of the minimum and the actual standard of living The term standard of living is generally understood to refer to the ability of an individual, a group or a society to obtain goods and services. It thus comprises not only material components of supply but also factors such as educational opportunities, the environment, personal safety or the proportion of income which must be spent on health, recreation and social services. International comparisons often analyse variations in GDP, per capita income or other indicators such as life expectancy or access to clean water. To measure the standard of living, this study limits itself to 19 indicators which describe the level of consumption of individuals and households. To start with, we tried to establish what possessions and activities respondents assess as necessities for a decent standard of living in the respective countries. This step was necessary in order to gain some idea of the perception of the minimum acceptable standard of living in the various nations and to determine appropriate weights for those items which are missing for financial reasons. The respondents were asked to indicate which items on a list of standard of living indicators every household in their country should be able to afford, classifying items respectively as absolutely necessary, which items were desirable but not essential, and which could be dispensed with. Any mention in the following discussion of respondents’ views on minimum living standards refers exclusively to those items determined to be ‘absolutely necessary’. Figure 1.1 provides the main results, depicting the percentages of respondents who classified an item as an absolute necessity. The items are listed in order of descending degree of intrinsic necessity according to the Swiss sample. The first four items listed belong to the group of standard or basic goods: a large majority of the respondents consider a lavatory or bath in the apartment, one cooked meal per day, a washing machine and a telephone, to be necessities. It is remarkable to note that (with the exception of Switzerland), those items which more than 50 per cent of the respondents regarded as essential and which would thus, based on the majority criterion, be included in Mack and Lansley’s deprivation index, include only two further things – namely, a television and a car. Furthermore, a private pension plan was seen as a necessity by nearly 50 per cent of the respondents. There is also a large consensus regarding the last seven items. Neither a computer, nor a garden or balcony, nor a dishwasher, nor a family meal at a restaurant, new furniture or a video recorder are classed as being necessary for a decent standard of living. If the focus is now placed on the differences in assessments between countries, it is striking to observe that in Switzerland neither a television set nor a car are judged to be absolutely necessary by a majority of the Christian Suter and Katia Iglesias 17

100.00 90.00 80.00 70.00 60.00 50.00 40.00 30.00 20.00 10.00 0.00

Car Phone Television Computer Dishwasher Video-recorder Washing machine Garden or balcony Private pension plan

One cooked meal per day WC or bath in the apartment Subscription to a newspaper Save at least 50 Euro per month One room per household member Invite friends for dinner once a month To be able to buy newTake clothes outTo regularlyfamily be able for dinnerto replace once worn-out a month furniture A week’s annual holiday away from home

Switzerland Austria Germany Slovenia

Figure 1.1 Necessities: percentage of persons considering the listed items ‘absolutely necessary’ for decent living Notes: Country individual weighting (excepted for Slovenia). Source: Euromodule (1999/2000).

respondents – even though more than 80 per cent of them own these things themselves (cf. Figure 1.2). Thus, a picture is created of the Swiss as a rather modest people. On the other hand, the Swiss respondents regard a week of holiday away from home, monetary savings, a subscription to a newspaper and inviting friends for a dinner as more important than do respondents from the other countries. On the whole, Figure 1.1 shows some variations between the countries in perceptions of minimum living standards. The differences, however, are surprisingly small. Thus, contrary to our expectations Slovenia does not show a different pattern compared to Austria and Germany. Furthermore, there are no discernible differences between western and eastern Germany. The most important differences are not those between Slovenia and the three German-speaking countries, but those between Switzerland on the one hand and Germany, Austria and Slovenia on the other. In a second step, the respondents were asked to indicate in regards to the same list of items whether their household owns a particular commodity or can afford to perform a particular activity. If a negative answer was given, then respondents were also asked whether this lack was for financial or other 18 Relative Deprivation and Well-being reasons. A description of the quality and cost of items was not included in the survey (e.g. small old Fiat versus new Mercedes), so it is a fairly crude indicator of living standards. In the affluent countries, a large proportion of households possess all of the items surveyed. This gives the impression of a uniform living standard. However, the households may in fact differ considerably in terms of incomes and living standards. An overview of approximate living standards can nevertheless be given on the basis of the indicators used. Figure 1.2 shows the percentages of people who reported they were able to afford each of the items surveyed. It can be noted that the first four items, which were almost unanimously classified as indispensable, are present in nearly all households. Television sets are found in more than 95 per cent of the respondents’ homes as well. This discrepancy between the low level of importance ascribed to televisions and the fact that televisions are present in practically all households has also been observed in other studies (e.g. Van den Bosch, 1998). This result leads to the assumption that the respondents distinguish between what they want for themselves and what should be included in a community’s notion of the minimum standard of living. With regard to all items, the

100.00 90.00 80.00 70.00 60.00 50.00 40.00 30.00 20.00 10.00 0.00

Car Phone Television Computer Dishwasher Video-recorder Washing machine Garden or balcony Private pension plan

Switzerland Austria Germany Slovenia

Figure 1.2 Actual standard of living: percentage of persons who can afford the listed items Notes: Country individual weighting (excepted for Slovenia). Source: Euromodule (1999/2000). Christian Suter and Katia Iglesias 19 standard of living of the respondents is clearly highest in Switzerland and lowest in Slovenia. However, Germany and Slovenia differ only slightly. More specifically, eastern Germany and Slovenia demonstrate a practically identical pattern. This is confirmed by the percentage of respondents who reported that they could not afford the surveyed commodities (data not shown separately). Although the proportion of people who indicated that they could not afford certain items varied between countries, a clear pattern is discernible with regard to the items that are lacking. The households were mainly unable to afford commodities and activities which required a permanent surplus income, such as a private pension plan or the ability to save money. Items regarded as luxury goods – such as computers, dishwashers or new furniture – were also more frequently absent. In addition, in such cases, there is often insufficient money for taking holidays, entertaining friends or eating out with the family. When all of these items are considered, the west–east order of national rankings is again noticeable: Switzerland, Austria, western Germany, eastern Germany and Slovenia.

4.2 Relative deprivation The information obtained for individual indicators of living standards was used to calculate our measure of relative deprivation, i.e. the Proportional Deprivation Index (PDI). We begin our discussion by examining the proportion of respondents affected by deprivation and mean deprivation value. As described in section 3 above, the deprivation index varies between 0 (none of the 19 standard of living items are lacking for financial reasons) and 100 (all of the 19 standard of living items are lacking for financial reasons). Comparing the degree of deprivation in the four countries seems to confirm our expectation that deprivation would be lower in the more affluent countries such as Switzerland, Austria and Germany. As demonstrated by Table 1.1, the ratio of people being affected by deprivation is lowest in Switzerland (31 per cent), followed by Austria (39 per cent), western Germany (40 per cent) and Slovenia (52 per cent). In eastern Germany, however, 55 per cent of the population is suffering from deprivation – even more than in Slovenia. Thus, Germany still seems to be divided in some regards. These observations are mirrored in the mean deprivation values. Again, Switzerland demonstrates by far the lowest average deprivation value (3.6), followed by Austria (4.5). Surprisingly, and contrary to our expectations, the mean deprivation values of Germany and Slovenia are almost identical (6.1 and 6.5). The highest deprivation values are reported by the respondents in eastern Germany (7.9). The comparatively low mean values of deprivation do not tell us, however, how deprivation is distributed across the population. Is deprivation concentrated within a few social groups – or is it more equally distributed over a larger proportion of the population? In order to answer this question we have calculated the deprivation index values for each deprivation decile. In all 20 Relative Deprivation and Well-being

Table 1.1 Relative deprivation (PDI) in four European countries

Switzerland Austria Germany West East Slovenia Germany Germany

Proportion of the 30.8 39.0 41.7 39.9 55.2 51.5 population affected by deprivation (%) Overall mean of 3.6 4.5 6.1 5.7 7.9 6.5 deprivation (index values) Mean deprivation 9.3 10.7 16.2 16.1 16.4 15.2 of the 9th deprivation decile (index values) Mean deprivation 22.9 24.6 30.4 30.7 29.5 32.1 of the 10th deprivation decile (index values)

N 1570 502 2413 1941 472 1012

Notes: ‘PDI’ is the Proportional Deprivation Index, Minimum (no deprivation) = 0, Maximum = 100. Source: Euromodule 1999/2000. countries there is a gap between the ninth and the tenth deprivation deciles. Deprivation, therefore, is heavily concentrated. Regarding the level of deprivation in the most affected population groups, the exceptional status of Switzerland can again be demonstrated: the mean deprivation of the most highly deprived decile is 23 in Switzerland, but 25 in Austria, 30 in Germany and 36 in Slovenia. The next factor to be examined is the relationship between equivalent household income7 and level of deprivation. Figure 1.3 indicates the mean PDI value attained in each income quintile. It comes as no surprise that the average deprivation in all countries decreases with increasing income. In general, country differences are more pronounced when considering the lower income quintiles. Thus, in Germany and Slovenia the level of deprivation in the lowest income quintile is substantially higher than that observed in Switzerland (17 and 16 versus 8). Compared to the other countries, the Swiss deprivation pattern is less skewed. The relationship between income and deprivation is a highly contentious issue in poverty research. Townsend (1979) postulated that poverty starts at the point at which people become disproportionately excluded from social participation as their income decreases. According to this concept, the poverty line is drawn at the position where a threshold or break in the curve indicates that deprivation is increasing disproportionately. However, no such threshold can be observed in Switzerland. The deprivation rates in this country rise Christian Suter and Katia Iglesias 21

20.00

18.00

16.00

14.00

12.00

10.00

8.00

6.00

4.00

2.00

0.00 1st quintile 2nd quintile 3rd quintile 4th quintile 5th quintile

Switzerland Austria Germany Slovenia

Figure 1.3 Level of deprivation (mean PDI) by quintiles of equivalent income Notes: Country individual weighting (excepted for Slovenia). Source: Euromodule 1999/2000.

relatively continuously with declining incomes. In the other three countries the proportion of people suffering from deprivation jumps between the first and second income quintiles (Germany and Slovenia) and between the second and third income quintiles (Austria). In a comparative analysis of the relationship between income and deprivation in Sweden and Great Britain, Halleröd (1998: 299) noted that the overall level of deprivation decreased considerably when the unemployed were excluded from the sample. In these circumstances, the differences between the countries also disappeared. From this he concludes that deprivation is a dilemma particularly for the unemployed. However, this result could not be reproduced with our data. Corresponding comparisons of the overall sample and the sample excluding unemployed persons yielded practically identical deprivation values by income deciles for all countries. Therefore, in the examined countries, at the individual level additional factors must be considered in addition to unemployment. In a separate analysis we estimated the determinants of relative deprivation by computing multivariate logistic regression models.8 In addition to income and employment status we included other factors that are known to be relevant for poverty and deprivation (level of education, age, gender and as well as aspects of life-course transitions). In all four countries the classical inequality dimensions such as low income and low education 22 Relative Deprivation and Well-being turned out to be significant determinants of deprivation. In Switzerland, Germany and Slovenia a precarious employment situation (long-term unemployment) sickness, and disability are further important components. In Switzerland, the probability of being severely deprived is five times greater for those who are not employed due to illness and disability than for those who have been regularly employed. This finding supports the structuralist argument of the classical stratification perspective that social class and employment status have persisted as powerful predictors of poverty and living conditions. Nevertheless, aspects of lifestyles and life courses, as proposed by the individualization thesis, are relevant, too. Thus, separation and divorce is a predictor of (high) deprivation in Switzerland. Age, relevant in all countries but Austria, shows contradictory relationships: young people (under 35 years) show higher levels of deprivation than the reference group of the 45 to 54 year olds in Germany, but lower levels in Slovenia. In Switzerland, the midlife cohort (35–44 years) is affected by the highest deprivation values.

5 The consequences of relative deprivation for well-being

As demonstrated in the previous sections, there are people and households in Switzerland and its neighbouring countries who fall below the generally recognized minimum standard of living of their respective societies. But what are the consequences of this situation? Does relative deprivation really matter? Since the 1980s, the relevance of inequality and stratification in post-industrial societies has been highly debated. Thus, it is argued that the traditional dimensions of inequality – i.e. indirect, income-based indicators – have largely lost their explanatory power due to the ‘elevator effect’ of generally upward social mobility present during the postwar period and as result of the processes of individualization and pluralization (e.g. Beck, 1986). In the context of our analysis the questions arises as to whether this weakening of social stratification measures applies to direct consumption based measures as well. We will address the question of the significance of direct poverty indicators by examining the effects of relative deprivation on subjective well-being. More specifically, it is hypothesized that the relative impact of deprivation is one of the most important explanatory factors negatively affecting well-being.

5.1 Measuring well-being The concept of well-being is complex and contains many aspects. Thus, subjective well-being is generally measured by several specific questions concerning their level of satisfaction with different aspects of life. Our Euromodule data contain nine such questions dealing with the contentment of respondents with several areas of life. As in other surveys (e.g. the European Household Panel, the Swiss Household Panel) an 11-point scale was used Christian Suter and Katia Iglesias 23

(0 = completely dissatisfied to 10 = completely satisfied) to explore the following dimensions of well-being: (1) satisfaction with one’s apartment, (2) satisfaction with one’s current job,9 (3) satisfaction with one’s standard of living, (4) satisfaction with one’s household income, (5) satisfaction with one’s health, (6) satisfaction with one’s education, (7) satisfaction with one’s neighbourhood, (8) satisfaction with public safety and (9) satisfaction with the environmental situation. One of the main issues that arises in relation to these questions is whether each of these nine different aspects of well-being must be analysed separately or whether a general index should be constructed. We decided to create a general measure. But which dimensions of well-being should be taken into account? We answered this question by conducting a factorial analysis for each country to see if there is a common, latent structure underlying the different aspects of satisfaction. Our empirical analysis revealed two common factors for all of the four countries: the first is composed of satisfaction with one’s standard of living, one’s household income, one’s health and education, and the second refers to satisfaction with public safety and with the environmental situation.10 The first factor can be seen as a dimension of well-being which directly touches the subjects on an individual level and on which they can have an influence. We call this dimension individual well-being. The second factor can be seen as a dimension of well-being that concerns the broader social contexts and environments of the individuals. In contrast to the first factor it is more difficult for the subjects to directly influence this second dimension of well-being. We call this second dimension societal well-being.11

5.2 Individual and societal well-being and their determinants In a first step we examine the overall level of well-being among the four countries. As already mentioned, most empirical studies find a strong correlation between subjective well-being and income at the collective level of nations (cf. Christoph and Noll, 2003 for the European countries). We therefore expect the highest level of well-being to be observed in Switzerland and the lowest in Slovenia. Considering our empirical results on standard of living and deprivation, we expect to see an west–east gap – that is, higher degrees of subjective well-being for Switzerland, Austria and western Germany as compared to eastern Germany and Slovenia. The overall mean values of individual and societal well-being and the mean values of each component are listed in Table 1.2. As expected, Switzerland shows the highest values of well-being for both dimensions as well as for all components. Regarding individual well-being there is a pronounced west–east gradient. Mean scores vary between 6.5 (Slovenia) and 7.9 (Switzerland). Empirical evidence for societal well-being, however, is quite different. There is no west–east gap but, instead, two rather homogenous 24 Relative Deprivation and Well-being

Table 1.2 Mean of individual and societal well-being and their components

Switzerland Austria Germany West East Slovenia Germany Germany

Individual well-being 7.91 7.56 7.16 7.25 6.82 6.47 Standard of living 8.32 7.89 7.44 7.56 6.96 6.80 Household income 7.48 7.21 6.71 6.89 6.01 5.88 Health 8.06 7.74 7.31 7.36 7.12 6.71 Education 7.76 7.39 7.20 7.16 7.38 6.58 Societal well-being 7.23 7.04 6.13 6.29 5.50 6.23 Public safety 7.68 7.38 6.17 6.39 5.25 6.32 Environmental situation 6.79 6.73 6.10 6.19 5.73 6.18 N valid (listwise) 1000 486 2333 1862 470 916 n (unweighted) 1570 502 2413 1941 472 1012

Notes: ‘Individual Well-being’: weighted mean of satisfaction with standard of living, household income, health and education. ‘Societal Well-being’: weighted mean of satisfaction with public safety and environmental situation. Minimum (completely dissatisfied) = 0, Maximum (completely satisfied) = 10. Source: Euromodule 1999/2000. groups: index values for Switzerland and Austria are slightly over 7, and for Germany and Slovenia slightly over 6. A closer inspection of the German pattern reveals again that this country is deeply divided, with respondents in eastern Germany suffering from a lower level of societal well-being (index value of 5.5). For all countries, scores for individual well-being are higher than those for societal well-being. By far the greatest differences between the two dimensions of well-being can be found in Switzerland and eastern Germany. Obviously, the respondents are much more critical in regards to the situation of their society and their social environment than to their own personal situation. This can be explained by a stronger individual responsibility of respondents for their personal situation. In individualized societies people are, in some respects, forced to be happy because they feel fully responsible for their own life and would otherwise be fully responsible for their unhappiness. In our final analysis we will examine the relationship between relative deprivation and the two dimensions of well-being in each of the four countries (that is, at the individual level). We started by computing simple linear regression models for estimating the effects of relative deprivation on individual and societal well-being. In a second step, we added to our first straightforward models the factors that are known from earlier studies to determine well-being (cf. Diener et al., 1999 for an overview on the state of the art). Apart from socio-demographic variables (gender, age, marital status, household composition) we took into account income (i.e. the indirect Christian Suter and Katia Iglesias 25 measures of poverty) and the other traditional inequality measures (social class, employment status, education) as well as self-reported health, suffering from anxiety, intensity of perceived social conflicts in the country, incidents of crime, and membership of an association.12 The results are summarized in Table 1.3.13 Due to limited space available we will confine our discussion to the two most important findings. The first result shows that the Proportional Deprivation Index is a powerful determinant of individual well-being. It explains between 9 per cent (Austria) and 21 per cent (Slovenia) of the variance between the four countries. When compared to a model which includes relative deprivation as well as all of the other variables, the model using only the Proportional Deprivation Index enables us to explain almost half of the variance of the full model (total explained variance fluctuates between 29 per cent for Austria and 41 per cent for Slovenia). In Switzerland and Germany, high levels of deprivation reduce individual well-being by 1 point, in Austria they do so by 0.8 points and in Slovenia by 1.3 points (multivariate model) on the 11-point scale. The traditional dimensions of inequality – i.e. indirect income-based indicators, employment status, education, social class – also have an impact on individual well-being. These determinants, however, are less powerful (measured by the explained variance) than the consumption-based measure of relative deprivation. Our analysis also confirms the effects of other known factors – like the positive influence of good health or the negative impact of anxiety and crime. Interestingly, there are only minor differences between the four countries (this also holds true for western and eastern Germany), suggesting that the underlying mechanisms are similar in all countries despite the different levels of relative deprivation and well-being. Our second finding shows that the relationship between relative deprivation and societal well-being is rather different. As demonstrated by Table 1.3, relative deprivation is rather weakly correlated to societal well- being. Depending upon the country, relative deprivation only explains between 0.2 per cent and 1 per cent of the variance. The most important determinants of societal well-being are the intensity of perceived social conflicts in the country, the level of anxiety experienced and the incidence of crime (which is rather implicit since one of the two components of societal well-being is public safety). Compared to individual well-being the explanatory power of the whole model for societal well-being is rather weak (explained variance varies between 6 per cent for Slovenia and 18 per cent for Austria).

6 Concluding remarks

The goal of this chapter has been to examine the actual Swiss standard of living, the degree and distribution of relative deprivation and their consequences for subjective well-being within a European context. Our comparison points 26 2.19 −

0.43* − 2.06 4.63 n.s. 2.05 n.s. − − − 0.25* 0.36*** 0.15* − − − 2.65 − Societal Well-being 0.60* − 2.04 − 0.32* 0.32* − 8.40 n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. 8.40 15.26 9.51 n.s. n.s. 4.14 n.s. n.s. ia Switzerland Austria ia Switzerland Germany Slovenia − − − −

1.36*** 2.53*** 0.39*** 0.96***

− − − −

9.51 18.27 4.69 12.68 − − − − 0.99*** 1.96*** 0.30*** 0.86*** − − − − 3.75 6.67 3.21 4.35 − − − − Individual well-being 0.76*** 1.44*** 0.41** 0.61*** − − − − 7.75 7.75 12.55 2.74 2.74 5.62 5.62 − − − − Switzerland Austria Germany Sloven 1.00*** 1.00*** 1.64*** 0.24** 0.24** 0.51*** 0.51*** − − − − b t b t b t b t b t b t b t b t b b t t b t b b t b t b t b t

Determinants of individual and societal well-being well-being and societal of individual Determinants deprived deprived

n (unweighted) 1570 502 2413 1012 1570 502 2413 1012 1012 2413 502 (unweighted) 1570 1012 2413 502 1570 2.57 n 11.29 F 3.68 sig 2.27 weight 131.31 205.57 Deprivation 26.76 yes 85.13 yes yes no no yes yes yes 0,000 0,000 0,000 0,000 0.026 n.s. 0,104 0,000 n.s. 0.077 n (unweighted) 1570 502 2413 1012 1570 502 2413 1012 1012 2413 502 1570 1012 2413 502 1570 (unweighted) n F 16.56 sig 38.36 weight 18.94 20.5 51.5 68.24 16.7 yes 32.09 yes yes no no yes yes yes 0,000 0,000 0,000 0,000 0,000 0,000 0,000 0,000 Highly deprived Highly deprived Deprivation Not deprived Somewhat notinc. not inc. not inc. not inc. inc. not inc. not not inc. inc. not Not deprived Somewhat notinc. not inc. not inc. not inc. not inc. not inc. not inc. not inc.

bivariate model model bivariate model multivariate

Table 1.3 Table

27 ial situation comparedto licts in country,the Suffering ogie, Universitéde Neuchâtel. not included (saturation of regression model) model) regression of (saturation included not = 41.2 10.3 18.4 10.3 6.4 not significant; inc. not significant; not = 14.2 0.3 0.9 1.1 0.2 21.1 14.9 9.4 0.7 5.6 0.0 n.s. 9.5 0.0 n.s. 11.0 11.2 0.0 n.s. 30.2 29.2 36.2

): 2 adjusted R explained variance ( deprivation as explicative variable only incomeas explicative variable deprivation and other explicative variables(a) : Euromodule: 1999/2000. : Significant level: *0.05, **0.01, ***0.001; n.s. ***0.001; Significant*0.05, **0.01, : level: Percent ofPercent 1 year ago, Household income, Living with partner, Good Health, Crime incidents, ofMembership an association, Intensity of conf Sociol de Institut the of website at the A.4 and A.3 Tables see additional details for composition; Household and anxiety, from Source multiple linear regression:non-standardizedb regressioncoefficient, individual country weighting. financ Household’s status, Employment class, Social status, Marital degree, Educational Age, Gender, variables: explicative (a) Notes model with only with model with model 28 Relative Deprivation and Well-being to the following implications relevant for the three theoretical and empirical debates mentioned in the beginning of this chapter: (1) Similarities and differences: The similarities and differences between the four countries correspond only partly to our expectations. Three results are of special importance: First, as expected there is a west–east gradient in the level of standard of living, deprivation and subjective well-being. The pattern of Switzerland and Austria are quite similar, but we found substantial differences between these two countries and Germany and Slovenia. Notwithstanding the deep and prolonged economic stagnation during the 1990s, Switzerland has maintained its high level of standard of living, its low level of relative deprivation (particularly regarding the lowest income groups) and its high level of subjective well-being. The characteristics of the Swiss pattern of standard of living are a combination of modest expectations concerning the minimum acceptable standard of living and a high level of actual standard of living. However, Swiss modesty is not necessarily an expression of traditional, conservative attitudes and values (e.g. diligence, modesty, thrift). The comparatively low proportion of Swiss respondents who consider having a car as necessary can be also explained by the high quality of the public transport system or by post-materialistic values. Secondly, despite their structural differences (e.g. different levels of GNP per capita, cultural and political differences), standard of living, deprivation and subjective well-being are rather similar in Germany and Slovenia. This can be partly explained by the differences between eastern and western Germany: standard of living and subjective well-being are substantially lower in eastern Germany than in Slovenia. In a way, the pattern observed in eastern Germany represents the opposite of that observed in Switzerland: high expectation concerning the minimum acceptable standard of living combined with a modest level of actual standard of living and a pessimistic estimation of personal and, in particular, societal well-being. Thirdly, there is a large consensus among all countries concerning the minimum standard of living regarded as absolutely necessary for a decent life. Apart from a few cultural peculiarities – e.g. the Swiss assignment of high significance to having monetary savings and low significance to owning a television – the national hierarchies of necessities are identical. This empirical evidence holds true for other European countries as well (e.g. Spain and Hungary). This means that on the normative level we are witnessing the emergence of a consensual European minimum standard of living. Concerning the actual standard of living, however, the disparities between and within countries are still considerable. (2) Direct and indirect measures of poverty: The emergence of a consensual minimum standard of living in Europe – as well as the disparities concerning the actual standard of living, the level and distribution of relative deprivation – indicate the significance of the direct, consumption-based measures of Christian Suter and Katia Iglesias 29 poverty and social exclusion. Although there is a relationship between direct and indirect measures – e.g. between the degree of relative deprivation and income quintiles – the two measures produce different pictures. In the case of Germany and Slovenia, our data show a concentration of deprivation in the lowest income groups – but only small deprivation differences between the second and fifth income quintiles. Our empirical evidence suggests that direct and indirect measures are complementary rather than substitutive: They map different aspects of reality. Thus, indirect indicators and the traditional dimensions of inequality are important predictors of relative deprivation in all countries. Furthermore, both relative deprivation and indirect measures are relevant factors for explaining well-being. (3) Relative deprivation and well-being: Our empirical evidence concerning the relationship between wealth and well-being yields three results. First, subjective well-being is a two-dimensional concept. In all of the four countries examined, a dimension of individual well-being and a dimension of societal well-being can be distinguished. Individual subjective well-being – which is the most important dimension – concerns the micro-aspects of happiness – that is, all of those aspects which impact directly at the individual level. These include satisfaction with personal resources such as health, income, standard of living and education. Societal well-being refers to satisfaction with the broader social environment – e.g. the quality of life of municipalities and districts, as measured by public safety and the environmental situation. Secondly, the impact of wealth and objective living conditions on well-being is confined to the individual dimension of well-being. In all of the four countries analysed there is a strong relationship between individual well-being and both direct and indirect indicators of poverty and the traditional inequality and stratification measures. By contrast, societal well-being remains largely unaffected by deprivation, income or other inequality measures. Thirdly, relative deprivation – that is, a direct consumption-based measure – is an important determinant of individual well-being. Thus relative deprivation explains between 9 and 21 per cent of the variance of individual well-being in the four countries – the strongest effect can be found in Slovenia, the smallest in Austria. Hence, relative deprivation seems to be a more powerful explanatory factor than the income-based indirect measures. This leads us to the conclusion that deprivation rather than wealth or income in general is important for subjective well-being.

Notes

1. European Council of Ministers 1994 ruling. 2. Data have been available for seven countries: Austria, Germany, Hungary, Slovenia, Spain, Switzerland, Turkey (see WZB, 2002). 30 Relative Deprivation and Well-being

3. The figures for 2002 (GNP per capita, PPP US$) are: Switzerland: 31, 250, Austria: 28, 240, Germany: 26, 222, Slovenia 17, 690 (World Bank: http://www.worldbank. org/data/countrydata/countrydata.html). 4. The Gini coefficient is 0.31 for Switzerland (1992), 0.26 for Germany (1994), 0.28 for Austria (1995) and 0.29 for Slovenia (1993/94) (Jesuit and Smeeding, 2002 and World Bank, World Development Indicators 2001). 5. For a specific set of j items and a person or household i (i = 1, 2, . . ., n) in a specific country c (c = 1, 2, . . ., C), the non-standardized PDI would be:

J n ()c PDI()nstd = ∑ ∑ wj dij j =1 i =1

where dij is a binary variable that is assigned a value of 1 when household i can- = c not afford item j (j 1, 2, . . ., J) and 0 when it can, and wj is the proportion of respondents who consider item j to be absolutely necessary for a decent living in their respective country c.

6. observed PDI = ------× 100 ()std max

observed = the observed score for a single case max = the maximum possible score for the country of origin. 7. The income was determined by asking for the net monthly income of the household. In order to make the incomes of people from households of different sizes comparable, the equivalent per capita income was calculated. As the Euromodule data do not include information on the age of any children, the Atkinson scale was employed. The values on this scale are obtained from the square root of the number of household members (cf. Atkinson et al., 1995). The values on the Atkinson scale are very similar to those on the modified OECD scale. 8. Dependent variable: belonging to the highest deprivation quintile; for details see Suter and Paris (2002) and the web site of the Institut de Sociologie, Université de Neuchâtel. 9. The question concerning employment was not taken into account in our analyses, due to the fact that half of the respondents in our sample do not work. 10. See additional Table A.2 at the website of the Institut de Sociologie, Université de Neuchâtel. 11. The index values of the two dimensions have been calculated as weighted mean of the satisfaction components. We have used the correlation between each item and the factors to weight the components of our two indices. 12. For more detailed information, see additional Table A.1 at the website of the Institut de Sociologie, Université de Neuchâtel. 13. Detailed information on the regression models is given in the additional Tables A.3 and A.4 available at the website of the Institut de Sociologie, Université de Neuchâtel. 31

Appendix

Table A1.1 List of variables

Indicator Scaling and construction

A: Demographic characteristics Gender 1 = female; 0 = male Age 6 age groups: 18–24, 25–34, 35–44, 45–54, 55–64 and over 65 Educational degree ISCED 1997 scheme, 3 categories: ISCED 0–2: low (level of education below upper secondary level, no further qualification), ISCED 3–4: medium (upper secondary education, vocational training), ISCED 5–6: high (level of higher education, university). Marital status 4 types of marital status: ‘married’, ‘single’, ‘divorced or separated’, ‘widowed’ Social class 3 categories: ‘lower class/working class’, ‘middle class’ and ‘upper middle/upper class’ Employment status 6 categories: ‘employed’ (has been in work during the last five years), ‘short-term unemployed’ (less than six months without work in the last five years), ‘long-term unemployed’ (without work for more than six months in the last five years), ‘retired’ (both early and regular retirement), ‘sick, disabled’ (long-term incapacity for work), ‘momentarily not in the workforce’ (e.g. homemakers, people in retraining) and the category ‘has never been employed’. B: Well-being Satisfaction with standard 11-point scale (0 to 10); 0 = completely dissatisfied, of living 10 = completely satisfied Satisfaction with 11-point scale (0 to 10) household income Satisfaction with health 11-point scale (0 to 10) Satisfaction with education 11-point scale (0 to 10) Satisfaction with public safety 11-point scale (0 to 10) Satisfaction with 11-point scale (0 to 10) environmental situation Satisfaction with apartment 11-point scale (0 to 10) Satisfaction with actual job 11-point scale (0 to 10) Satisfaction with 11-point scale (0 to 10) neighbourhood Individual well-being weighted mean of satisfaction with standard of living, household income, health and education Societal well-being weighted mean of satisfaction with public safety and environmental situation 32

Table A1.1 (Continued)

Indicator Scaling and construction

C: Deprivation Actual standard of living List of 19 items; count ‘have’ Deprivation score List of 19 items; count ‘necessities’ and ‘cannot afford’,PDI cf. section 3 Deprivation categories 3 categories: ‘not deprived’ (PDI = 0), ‘somewhat deprived’ (0 < PDI ≤ 9th deprivation decile), ‘highly deprived’ (PDI = 10th deprivation decile) D: Economic situation Household’s financial 3 categories: ‘clearly/some what deteriorated’, situation, compared to ‘remained the same’ and ‘improved somewhat/clearly 1 year ago improved’ Household income 6 categories: 1st to 5th quintile of equivalent income and a category for the no response (respondents reported household income, open question, Atkinson scale) E: Others Living with partner 0 = no, 1 = yes Good health 2 categories: 0 = some extend/severely problems of health, 1 = no problem of health Crime incidents 2 categories: 0 = not subjected to crime, 1 = subjected to at least one of the following crimes during the last 12 months: get things stolen, be harassed or threatened, get sexual molested and be beaten and hurt. Membership of an association 1 = member of an association, 0 = no member Intensity of conflicts in the Mean of 6 types of conflicts: ‘between poor and rich’, country ‘between unemployed and people with jobs’, ‘between management and workers’, ‘between young and old people’, ‘between men and women’ and ‘between nationals and immigrants’. Scale 1 = no conflicts to 4 = very strong conflicts. Suffering from anxiety 2 categories: 0 = no anxiety; 1 = suffering from one or more of 5 types of anxieties: ‘exhaustion or fatigue’, ‘unhappy or depressed’, ‘shake or tremble’, ‘keyed up and jittery’ and ‘frightening thoughts’. Household composition 5 types of households: ‘single’, ‘couple without children living in the household’, ‘couple with 1–2 children younger than 18’, ‘couple with 3 and more children younger than 18’, ‘single parent family’ and ‘other household composition’. 33 extraction communalities factor 2 Pooled sample (4 countries) sample (4 Pooled factor 1

37.0 24.0 6119 0.000 extraction communalities Slovenia 0.83 0.70 — 0.85 0.72 0.83 0.70 — 0.81 0.66 — 0.80 0.68 0.80 0.68 0.81 0.66 — factor 2 factor factor 1 38.8 23.438.8 1012 1012 5497 no no yes 0.000 0.000 0.36 0.69 — 0.49 0.63 — 0.42 — 0.49 0.63 0.36 0.69 0.44 0.72 — 0.53 0.68 — 0.46 — 0.53 0.68 0.44 0.72 extraction extraction communalities Germany Germany factor factor 2

factor factor 1 36.2 36.2 24.1 3006 1199 yes 70 0.83 — 0.71 0.82 — 0.69 0.82 — 0.70 — 0.69 0.82 — 0.71 0.82 — 70 0.83 0.67 — 0.67 0.80 — 0.69 0.80 70 0.82 extraction communalities Austria 0.70 0.51 — 0.82 0.70 — 0.82 — 0.51 0.70 factor 2 0.61 — 0.61 0.41 0.66 — 32.7 23.9 factor 1 yes 502 502 2413 0.000 0.000 0.000 extraction communalities Switzerland 483.7 factor 2

— 0.82 0.68 — 0.78 0.61 — 0.84 0.72 — 0.84 — 0.61 0.78 — 0.68 0.82 — 0.81 — 0.81 0.67 — 0.76 — 0.83 0.60 0. — 0.83 0.

31.9 22.2 7 1 Factorial analysis for well-being for well-being analysis Factorial

: : Euromodule 1999/2000. 1999/2000. : Euromodule : Factor analysis: rotation: varimax, country individual weighting. with situation situation explained variance living living income satisfaction public safety safety public environmental — of percent 0.79 0.64 (unweighted) n 1570 — standard of household 0.28 — health 0.53 education — 0.62 0.38 — 0.42 0.47 0.59 — square chi 732.9 weight yes sig 0.000 yes weight Source Table A1.2 Table Notes factor 34

Table A1.3 Determinants of individual well-being

Individual well-being

Switzerland Austria Germany Slovenia

b t b t b t b t

Deprivation Not deprived not inc. not inc. not inc. not inc. Some what −0.24** −2.74 −0.41** −3.21 −0.30*** −4.69 −0.39*** −4.14 deprived Highly deprived −1.00*** −7.75 −0.76*** −3.75 −0.99*** −9.51 −1.36*** −8.40 Gender (woman) n.s. n.s. n.s. n.s. Age 18–24 not inc. not inc. not inc. not inc. 25–34 −0.24* −2.40 n.s. n.s. n.s. 35–44 n.s. n.s. n.s. −0.37*** −3.34 45–54 n.s. n.s. n.s. −0.40*** −3.41 55–64 n.s. n.s. n.s. n.s. 65+ n.s. n.s. n.s. n.s. Household’s financial situation Deteriorated −0.34*** −3.95 n.s. −0.50*** −6.38 −0.53*** −5.43 Same not inc. not inc. not inc. not inc. Improved n.s. n.s. 0.30*** 4.25 n.s. Income 1st quintile not inc. not inc. not inc. not inc. 2nd quintile n.s. n.s. n.s. n.s. 3rd quintile n.s. n.s. 0.36*** 3.52 n.s. 4th quintile n.s. n.s. 0.26** 2.81 n.s. 5th quintile 0.24* 2.36 0.44** 2.86 0.45*** 4.76 n.s. No answer n.s. n.s. 0.41*** 5.56 n.s. Household composition Single not inc. not inc. not inc. not inc. Couple n.s. 1.09*** 4.61 n.s. n.s. 1–2 children<18 n.s. 0.77*** 3.52 n.s. n.s. 3 and more n.s. 1.09** 3.47 n.s. 0.55* 2.06 children<18 Single parent n.s. 1.55*** 5.90 n.s. n.s. Other household n.s. 0.88*** 4.20 n.s. n.s. composition Marital status Married n.s. n.s. n.s. n.s. Single not inc. not inc. not inc. not inc. Separated or n.s. n.s. −0.29* −2.52 n.s. divorded Widowed n.s. n.s. n.s. n.s. Social class Low −0.54*** −5.14 −0.37** −2.67 −0.65*** −9.69 −0.40*** −4.04 Middle not inc. not inc. not inc. not inc. 35

Upper 0.30*** 3.13 n.s. 0.53*** 5.39 0.45** 2.74 Unknown n.s. n.s. −0.40*** −3.71 n.s. Living with n.s. n.s. n.s. n.s. partner (yes) Health (good 0.27*** 3.34 0.67*** 4.75 0.43*** 6.70 0.46*** 4.35 health) Employment status Employed not inc. not inc. not inc. not inc. Unemployed<6 n.s. n.s. n.s. n.s. months Unemployed>6 n.s. n.s. −0.56*** −6.25 n.s. months Retired 0.34*** 3.63 n.s. n.s. n.s. Sick, disabled n.s. n.s. −0.73** −2.91 n.s. Temp. not in 0.42*** 3.43 n.s. n.s. n.s. workforce Never been 0.47* 2.24 n.s. n.s. n.s. employed Educational degree ISCED 0–2 (low) n.s. −0.41** −2.95 −0.53** −2.93 −0.61*** −5.53 ISCED 3–4 not inc. not inc. not inc. not inc. (medium) ISCED 5–6 (high) n.s. n.s. n.s. 0.34** 2.66 Crime incidents −0.22* −2.46 −0.37* −2.39 −0.18* −2.19 n.s. (yes) Membership of an n.s. n.s. n.s. n.s. association (yes) Intensity of −0.21** −2.83 n.s. −0.23*** −3.98 −0.20** −2.62 conflicts in the country Suffering from −0.61** −8.12 −0.48*** −4.04 −0.40*** −6.52 −0.58*** −6.32 anxiety (yes) Constant 8.60*** 43.10 6.81*** 26.89 7.85*** 47.89 7.68*** 34.94 n (unweighted) 1570 502 2413 1012 F 32.09 16.7 68.24 51.5 sig 0,000 0,000 0,000 0,000 weight yes yes yes no percent of 30.2% 29.2% 36.2% 41.2% explained variance 2 (Radjusted)

Table A1.4 Determinants of societal well-being

Societal well-being

Switzerland Austria Germany Slovenia

b t b t b t b t

Deprivation Not deprived not inc. not inc. not inc. not inc. Somewhat n.s. n.s. −0.15* −2.05 n.s. deprived Highly deprived n.s. n.s. n.s. n.s. Gender (woman) n.s. n.s. n.s. n.s. Age 18–24 not inc. not inc. not inc. not inc. 25–34 n.s. n.s. n.s. n.s. 35–44 n.s. n.s. n.s. n.s. 45–54 n.s. n.s. n.s. n.s. 55–64 n.s. n.s. n.s. n.s. 65+ n.s. n.s. n.s. n.s. Household’s financial situation Deteriorated n.s. −0.35* −2.56 −0.19* −1.99 −0.38** −3.21 Same not inc. not inc. not inc. not inc. Improved n.s. n.s. n.s. n.s. Income 1st quintile not inc. not inc. not inc. not inc. 2nd quintile n.s. −0.41** −2.58 n.s. n.s. 3rd quintile n.s. n.s. n.s. n.s. 4th quintile n.s. n.s. n.s. n.s. 5th quintile n.s. n.s. n.s. n.s. No answer −0.32* −2.13 n.s. n.s. n.s. Household composition Single n.s. n.s. n.s. n.s. Couple not inc. not inc. not inc. not inc. 1–2 children<18 n.s. n.s. n.s. n.s. 3 and more n.s. n.s. n.s. n.s. children<18 Single parent n.s. n.s. −0.45** −2.75 n.s. Other household n.s. n.s. n.s. n.s. composition Marital status Married n.s. n.s. n.s. n.s. Single not inc. not inc. not inc. not inc. Separated or n.s. n.s. n.s. n.s. divorded Widowed n.s. n.s. n.s. n.s. 37

Social class Low n.s. n.s. n.s. n.s. Middle not inc. not inc. not inc. not inc. Upper 0.35** 3.13 n.s. n.s. n.s. Unknown n.s. n.s. n.s. n.s. Living with n.s. 0.29* 2.28 n.s. n.s. partner (yes) Health (good n.s. n.s. n.s. n.s. health) Employment status Employed n.s. n.s. n.s. n.s. Unemployed<6 not inc. not inc. not inc. not inc. months Unemployed>6 −0.54* −2.45 n.s. −0.24* −2.20 n.s. months Retired n.s. n.s. n.s. n.s. Sick, disabled 0.83* 2.22 1.06* 2.02 n.s. n.s. Temp. not in n.s. n.s. n.s. n.s. workforce Never been n.s. n.s. n.s. n.s. employed Educational degree ISCED 0–2 (low) n.s. n.s. n.s. 0.28* 2.26 ISCED 3–4 not inc. not inc. not inc. not inc. (medium) ISCED 5–6 (high) n.s. n.s. n.s. n.s. Crime incidents − −5.61 n.s. −0.79*** −7.71 −0.64*** −3.72 (yes) 0.63*** Membership of an n.s. n.s. n.s. n.s. association (yes) Intensity of − −7.69 −0.83*** −7.14 −0.67*** −9.45 −0.57*** −5.70 conflicts in the 0.71*** country Suffering from n.s. −0.52*** −4.18 −0.40*** −5.62 n.s. anxiety (yes) Constant 8.99*** 40.26 9.17*** 31.35 8.16*** 47.27 7.67*** 32.06 n (unweighted) 1570 502 2413 1012 F 20.5 18.94 38.36 16.56 sig 0,000 0,000 0,000 0,000 weight yes yes yes no percent of 10.3% 18.4% 10.3% 6.4% explained variance 2 (Radjusted)

Notes: Significant level: *0.05, **0.01, ***0.001; n.s. = not significant; not inc. = not included (saturation of regression model) multiple linear regression: b non-standardized regression coefficient, country individual weighting. Source: Euromodule 1999/2000. 2 What Pluralization of the Life Course? An Analysis of Personal Trajectories and Conjugal Interactions in Contemporary Switzerland Eric Widmer, Jean Kellerhals and René Levy

Sociological research has underscored trends towards a pluralization of social life in general and, more specifically, in the areas of family and work. These two central subsystems of modern societies are hypothesized by some to have become much more heterogeneous in recent years than had been the case hitherto; they are said to follow less recognizable patterns or models, either because of the weakening of social norms or because of the increasing structural complexity of contemporary societies. This chapter addresses the issue of the pluralization of the life course by focusing on personal trajectories (between professional and family activities) and family functioning of contemporary couples,1 based on a sample of the project entitled ‘Social Stratification, Cohesion and Conflict in Contemporary Families’. Does the highly post-industrialized context of Switzerland confirm the hypothesis that late modernity is associated with a strong diversification of work and the family and a weakening of their being socially structured? We shall see that, although variability does indeed characterize personal trajectories and family functioning in Switzerland, it is rather bounded and is quite strongly embedded in social structures.

Professional and relational dimensions of life courses in Switzerland

What types of personal trajectories characterize individuals having lived the main part of their life course in the second half of the twentieth century in Switzerland? Is there a single, predominant type, is there a small but varied group of types, or there is no clear typification at all? Further, if they do exist, what logic underlies these types? In this regard, several hypotheses can be

38 Eric Widmer, Jean Kellerhals and René Levy 39 found in the literature that can be summarized by two pairs of contradictory statements. The first pair concerns the standardization versus the individualization of trajectories; the second pair is about the presence of a unique model of trajectories versus two models, differentiated according to gender. The first opposition of theses is strongly connected with the debate on individualization as a corollary of the ‘second wave of modernization’, identified notably by authors such as Beck (1986), Giddens (1992 and 1994; Beck, Giddens and Lash, 1994), and more generally by postmodern theorists (for example, Bauman, 1992). In the same vein, some authors (for example, Beck and Beck-Gernsheim, 1994) stipulate that the process of individualization of the life course has been taking place over several decades, as a corollary of the decline of the elements of standardization. From an empirical point of view, this tendency would translate into a multiplication of professional and domestic trajectories, which would not be captured as easily as in the past by a limited number of ‘types of trajectories’. In contrast, Levy (1977; Levy et al., 1997) and Kohli (1985, 1986) formulate the hypothesis of a standardization of life courses, through a historical process which Kohli, referring to the period ante 1968, decomposes into three modalities: a ‘sequentialization’, or the emergence of distinct stages following each other in a defined order, such as schooling, employment and retirement; a ‘chronologization’ or the establishment of a link between life course transitions and age that can be related to the increasing bureaucratization of modern societies along lines already diagnosed by Weber; and finally, a ‘biographization’ or social attribution of the responsibility of the life course to each individual. If this thesis were true, strong regularities of common traits should appear in peoples’ life courses, or even clearly profiled types of life trajectories. The second opposition of theses concerns the extent to which life courses are gendered. This issue is of special relevance in relation to the standardization process. Some authors consider that the standardization of life courses essentially carries the imprint of the occupational subsystem, from which it draws its functional logic (Kohli, Rein, Guilemard and van Gunsteren, 1991).2 Based on this model, one may easily assume that women’s as well as men’s trajectories follow a single model, where possible sex-specific variations would be only sub-variants. The thesis of sex-typed life courses, on the other hand, postulates the existence of two very distinct models of trajectories, one for each sex, which are interconnected but follow quite different, although complementary logics (Levy, 1977; Becker-Schmidt, 1987; Myrdal and Klein, 1956). In its exclusive version, this thesis states moreover that differences existing within the female and male sub-populations are insignificant. Gendering of the domestic and professional fields is thus presented as the dominant factor in the orientation of trajectories which tends to level out the effects of other dimensions of the social structure. The second opposition of theses thus concerns the existence of one, gender-neutral, or rather two, gendered, trajectory types. 40 What Pluralization of the Life Course?

Between the homogeneity postulated by the hypothesis of standardization, the extreme variability postulated by the hypothesis of pluralization, the one-type hypothesis of the overwhelming structuration of life courses by the labour market, and the unambiguous differentiation between sexes suggested by the hypothesis of ‘sexuation’, we propose a fifth hypothesis, that of a differential impact of various profiles of social insertion. The following reasons underpin this additional hypothesis. The variability of female professional insertion, at least for women living in a couple, is a fact that has been known and analyzed for a long time. One finds, among its causal factors, the social status of the couple and the educational level of the women (Reskin and Padavic, 1994). Besides these factors associated with social stratification, the arrival of the first child has a significant impact on the employment of women (Held and Levy, 1974; Höpflinger, Charles and Debrunner, 1991; Klein and Lauterbach, 1994; Drobnic, 2000) i.e., a typical event belonging to the realm of the life course. On this basis, one can hypothesize that the differences between male and female trajectories, however important they may be, depend upon a variety of social insertions that modulate the life course. In this regard, the hypothesis of the profile of insertion states that well-differentiated models of trajectories exist within male and female subpopulations, depending upon the individuals’ location in the social structure. Similar issues were raised recently about family functioning. Researchers in the field of family interactions have been especially concerned about the problems or crises that today’s couples are increasingly presumed to meet. Nevertheless, there are no studies, in Europe at least, that systematically relate these problems to the structure of conjugal relationships. To be sure, some psychological approaches have opened the way (Reiss, 1971 and 1981; Kantor and Lehr, 1975; Olson and McCubbin, 1989). However, they are mostly based on the observation of clinical populations where the search for a solution to a crisis or an illness is of primary interest and, as a consequence, this relegates to the background the central issues of the social anchoring of domestic functioning and the representativeness of the observations. From this point of view, several current issues are particularly crucial. A first issue concerns the pluralization of family life since the 1960s. Most scholars have addressed this issue by focusing on household structure and composition. They have acknowledged the great changes in the distribution of household types (nuclear family households, single parent households, recomposed households, etc.) in recent decades. However, this is only one approach to the issue, as those changes are likely to be the expressions of more fundamental changes in intimate interactions within families. Some recent studies on the family and intimacy indeed suggest that contemporary conjugal functioning follows a single global logic, essentially capturing the characteristics of the ‘modern’ family as described as early as 1945 by Burgess et al. (1960) in the ideal-type of companionship family: a union of affinity, with individuals’ self-development Eric Widmer, Jean Kellerhals and René Levy 41 as a central (if not unique) function, with internal arrangements essentially based on contractual and egalitarian rules, freed from external constraints, be they institutional or relational.3 The profound demographic changes that have occurred since the 1960s pose a serious challenge to the now classic opposition between the institution and the companionship models, as proposed by Burgess, in characterizing the evolution of the family from traditional practices to modernity. This opposition, and the evolutionist perspective on which it is based, had been challenged by various attempts of definition of styles or models of family interactions, based on four conceptual dimensions: the degree of autonomy of spouses or partners within the couple; the degree of openness of couples towards their environment; their instrumental or expressive orientation; and the importance of negotiation in their daily functioning (Kellerhals, Troutot and Lazega, 1993). From there, various typologies were proposed that revealed the existence of a strong correlation between domestic functioning and social status, questioning the hypothesis of a general movement towards a unique model of family relationships. In particular, this was found for Switzerland (Kellerhals et al., 1982; Kellerhals and Montandon, 1991; Kellerhals, Troutot and Lazega, 1993), a context in which it was observed that high levels of economic and cultural resources corresponded to a strong emphasis on individual autonomy and on the opening of couples towards their environment whereas a low level of economic and educational resources was associated with an emphasis on traditional status, rigid norms, a priority of collective concerns over individual orientation, a relative distrust towards the environment by family members, and a rather gendered and unequal division of domestic roles and power. The present chapter will investigate whether the current situation of couples supports the hypothesis of a limited set of conjugal models, grounded in the social structure, if it confirms the hypothesis of a pluralization of conjugal relationships around a companionship model, largely disembedded from the social structure, or if there is a single type of trajectories according to the thesis of labour market predominance. So our purpose is not to retrace the possible historical change of male and female trajectories – our data do not sufficiently allow for this kind of analysis – but to ascertain their current setup; this shall allow to see which one of the five theses we formulated corresponds best to the trajectories we can actually observe.4

1 Data and method

The data presented in this chapter are drawn from the study ‘Social Stratification, Cohesion and Conflict in Contemporary Families’, a large survey of 1,534 couples and families living in Switzerland. Conducted in 1998, the study’s primary goal was to examine how the subjects’ social status and position in their life course influence cohesion and conflicts in couples. The sample for 42 What Pluralization of the Life Course? the project was drawn randomly with a non-proportional stratified design based on the three major linguistic areas of Switzerland. To be included in the sample, respondents had to have been living together for a least one year; the younger partner had to be aged at least 20, and the older partner no more than 70; they had to be living in Switzerland, but neither Swiss citizenship nor formal marriage was necessary for the couple to be sampled. We used a computer-assisted telephone survey questionnaire, translated into the three major idioms of Switzerland (German, French and Italian). Data collection took place between October 1998 and January 1999. For each couple, both partners (spouses or cohabitees) were interviewed separately, giving a total number of 3,068 interviews completed. For most questions, both partners had to provide an answer. Responses were weighted according to the population size of each of the three linguistic regions. This first phase included various measures of conjugal interaction, conjugal and family conflict, and conjugal quality. A detailed account of this analysis can be found in Widmer, Kellerhals and Levy (2003). In a second phase, carried out about three months later, a self-administered retrospective questionnaire was sent to all first wave respondents, with the aim of collecting information about the family and professional insertion of the partners or spouses throughout their life, from the end of compulsory education until the year of their participation in the study. Every period had to be dated on a yearly basis. The response rate for this second phase was 46 per cent, with 703 men and 717 women providing usable questionnaires. We found no significant differences between this subsample and the original sample. On the basis of the information thus gathered, we recoded all periods into eight mutually exclusive categories: periods of professional full-time activity, part-time professional activity, interruptions because of unemployment or because of problems of health (‘negative interruptions’), interruptions to travel or other intentional leave of absence, etc. (‘positive interruptions’), periods at home (homemaker), and periods of training (for details, see Widmer, Levy, Hammer, Pollien and Gauthier, 2003).

2 Results

Let us first consider the empirical results which address the issue of pluralization of personal trajectories and conjugal functioning.

2.1 Professional trajectories: plural but structured For reasons of validity, the analysis of the subsample of individuals who answered the retrospective questionnaire was restricted to those aged 30 and over, reducing the sample population to 677 men and 670 women. Indeed, short life courses can slant the classification because they do not allow the identification of any clear pattern. Regular multivariate statistical techniques do not apply here, because usual measures of distance, such as the Euclidean Eric Widmer, Jean Kellerhals and René Levy 43 distance, are ineffective for sequential data (Erzberger and Prein, 1997). We resorted, therefore, to using optimal matching analysis, a multivariate statistical method stemming from molecular biology (Waterman, 1995; Delcher et al., 1999), and which has been adapted, in several recent socio- demographic works, to the study of trajectories (Abott and Hrycak, 1990; Erzberger and Prein, 1997; Aisenbrey, 2000).5 Based upon measures of distance produced by optimal matching analysis, we ran a hierarchical cluster analysis applied separately to women’s and men’s responses. We found that two groups of sequences best characterize the male trajectories. Figures 2.1 and 2.2 present the relative frequencies of each category of activity between 16 and 64 years of age for each of the two groups. The first group of male trajectories includes very homogeneous life courses, characterized by professional full-time activity. This is the standard life course for men, followed by more than 8 out of 10 men. This is clearly the dominant model for men. The second group concerns only a minority of men (15 per cent) and presents a very dissimilar pattern from the dominant one. First, part-time employment is much more frequent than full-time activity: full-time work

100%

80%

60%

40%

20%

a16 a18 a20 a22 a24 a26 a28 a30 a32 a34 a36 a38 a40 a42 a44 a46 a48 a50 a52 a54 a56 a58 a60 a62 a64

Education/ Training Interruption (positive) Part time paid work Retirement Interruption (negative) Full time paid work At home

Figure 2.1 Men’s trajectory type ‘Dominant’ (n = 574, 85%) 44 What Pluralization of the Life Course?

100%

80%

60%

40%

20%

a16 a18 a20 a22 a24 a26 a28 a30 a32 a34 a36 a38 a40 a42 a44 a46 a48 a50 a52 a54 a56 a58 a60 a62 a64

Education/ Training Interruption (positive) Part time paid work Retirement Interruption (negative) Full time paid work At home

Figure 2.2 Men’s trajectory type ‘Minority’ (n = 103, 15%) represents only 9 per cent of the average trajectory. Part-time employment, however, concerns a proportion of 34 per cent of the average trajectory. Trajectories of this type also include more interruptions due to travel, breaks, etc. and problems of unemployment or health (about six months). The very heterogeneous character of these minority trajectories suggests that they do not correspond to a true model, but rather form a residual category. In contrast to men, women have quite heterogeneous models of trajectories. Their trajectories fall into four types. The first type of female trajectory (Figure 2.3) is centred on the home. After the period of initial education, the majority of women with this trajectory who enter the professional field via full-time employment remain there for only a short period of time. This is followed by an extensive period at home (66 per cent of the whole sequence considered). Women following this type of trajectory focus on family life and the education of their children after a short period of full-time professional activity, and do not re-enter the professional world. The second type (Figure 2.4) is composed of trajectories that are the opposite of the previous type on all accounts. Indeed, full-time employment dominates here. This type is very close to the predominant one observed among men. 45

100%

80%

60%

40%

20%

a16 a19 a22 a25 a28 a31 a34 a37 a40 a43 a46 a49 a52 a55 a58 a61 a64

Education/ Training Interruption (positive) Part time paid work Retirement Interruption (negative) Full time paid work At home

Figure 2.3 Women’s trajectory type ‘Homemaker’ (n = 220, 33%)

100%

80%

60%

40%

20%

a16 a19 a22 a25 a28 a31 a34 a37 a40 a43 a46 a49 a52 a55 a58 a61 a64

Education/ Training Interruption (positive) Part time paid work Retirement Interruption (negative) Full time paid work At home

Figure 2.4 Women’s trajectory type ‘Full-time worker’ (n = 133, 20%) 46 What Pluralization of the Life Course?

100%

80%

60%

40%

20%

a16 a18 a20 a22 a24 a26 a28 a30 a32 a34 a36 a38 a40 a42 a44 a46 a48 a50 a52 a54 a56 a58 a60 a62 a64

Education/ Training Interruption (positive) Part time paid work Retirement Interruption (negative) Full time paid work At home

Figure 2.5 Women’s trajectory type ‘Part-time worker’ (n = 159, 24%)

The third type of trajectory (Figure 2.5) corresponds to yet another model. As with the other types, there is a sharp take-up of employment after initial education, initially being generally full-time. As with the previous type, occupational activity is maintained throughout adult life, but full-time employment is rather quickly – towards the age of 25 – exchanged for part-time employment which becomes general at about 40. To summarize, part-time trajectories feature an early reduction in the rate of employment. The fourth type of trajectory (Figure 2.6) is similar to the type ‘homemaker’ but differs from it because its focus on the domestic sphere is only temporary. Indeed, from the age of about 32 years, there is a very significant increase in gainful employment, though this is generally part-time. This reaches its peak around the beginning of the forties. By this age, this type of trajectory is rather close to the previous one (‘part-time’), but it differs from it because periods in the home are much longer and full-time employment much shorter. The major characteristic of this type of trajectory is a late and partial reinte- gration into the labour market after a relatively long period spent exclusively at home. The last group is very different from all the others as it has an erratic, rather unstructured character. After a brief period of education and, Eric Widmer, Jean Kellerhals and René Levy 47

100%

80%

60%

40%

20%

a16 a19 a22 a25 a28 a31 a34 a37 a40 a43 a46 a49 a52 a55 a58 a61 a64

Education/ Training Interruption (positive) Part time paid work Retirement Interruption (negative) Full time paid work At home

Figure 2.6 Women’s trajectory type ‘Back-to-employment’ (n = 133, 20%) in a minority of cases, a short period of full-time employment, this group of trajectories is characteristic of women having known numerous interruptions due to problems of unemployment and health, or who are involved in a lot of different activities such as voluntary help. These atypical trajectories concern only a very small proportion of cases (less than 5%) and represent a little-typed residual category rather than a specific model of the life course. To summarize, four dominant models characterize the professional and family trajectories of women in Switzerland: ‘homemaker’ (33 per cent), ‘full-time worker’ (20 per cent), ‘part-time worker’ (24 per cent), and the ‘back to employment’ group (20 per cent). Therefore, our results show that there is a clear differentiation of female trajectories, with most of them falling into four distinct and recognizable models – one of which corresponds to the model predominant among men. A fifth group comprises less clear-cut trajectories, representing only 4 per cent of cases.

2.2 A limited number of styles of conjugal interactions Let us now consider conjugal functioning. We first want to know how many styles characterize the functioning of today’s couples. In order to construct such 48 What Pluralization of the Life Course? a typology, we have included six dimensions that are often used in empirical assessments of family functioning (Kellerhals, Troutot and Lazega, 1993):

(1) The degree of fusion, which designates the extent to which individual resources (time, money, ideas, feelings) are pooled by spouses or partners (Roussel, 1980; O’ Neil and O’Neil, 1972). (2) The degree of openness, which designates the extent of informational and relational exchanges occurring between the family group and its close environment (Kantor and Lehr, 1975; Reiss, 1971). (3) The main focus, either external or internal, of priority objectives assigned to the couple or the family (Parsons and Bales, 1955; Donati, 1985). Are internal and relational goals (such as affective security, support, tenderness) emphasized or rather external and instrumental goals, such as social integration and upward social mobility? (4) The degree of sex-typing of conjugal roles, which designates the extent of the gendered division of household labour and occupational activities, as well as that of relational roles (such as information, goal selection, support, emotions with regard to the spouse or partner (Aldous, 1977; Bott, 1971). (5) The normative strength of sex-specific master statuses in the couple. This dimension captures the differential investment of men and women in the domestic sphere relative to the occupational sphere, which is not only a question of time of presence but is also connected with the sacrifices that one can or is willing to make to it. It is based on the hypothesis proposed by Krüger and Levy (2001) that there is a priority sphere of investment attributed to sex categories, the domestic sphere for women and the public sphere for men, which subordinates the investments that the incumbents can put into the other sphere. (6) The degree of routinization, which designates the extent to which couples follow a fixed set of norms concerning family timetables, eating habits and territorial allocations (Olson and McCubbin, 1989).

The first three dimensions are concerned with the more general concept of cohesion, whereas the second three dimensions relate to the concept of regulation. All six dimensions have been measured by a set of indicators that are described elsewhere (Widmer, Kellerhals and Levy, 2003; Widmer, Kellerhals and Levy, 2004b). They were analysed together in a series of scales that were dichotomized and subjected to a cluster analysis. In order to determine the number of styles of conjugal interactions, we computed a sequence of hierarchical clusters (based upon Ward’s method of clustering). We carefully examined solutions between 3 and 7 clusters and found the solution with five to strike a good balance between within-cluster homogeneity, clarity and parsimony. The results of this analysis are presented in Table 2.1. 49 .66** .58** .35** .68** .50** .41** .34** .63** Cramers’ V Cramers’ (%)

42 40 57 .21** 63 .36** 22 .15** 44 53 34 35 50 27 (whole sample) 1534) = 5 9 9 95 19 58 65 92 67 76 56 27 7428 83 56 13 20 42 72 11 74 42 60

57 19 34 91 20 16 16 24 60 49 7974 38 91 4831 13 23 70 53 18 60 24 81 45 24 68 60 16 56 17 24 1624 15 17 29 17 6 Results of cluster analysis based on responses from both partners (column percentages, N

roles roles power Closure (women) Internal orientation (women) Internal orientation (men) Regulation Strong routinization Fusion (men) Closure (men) Strong differentiation of functional of relational differentiation Strong Strong differentiation of decisional Strong master status Table 2.1 Table Size of cluster (% of total) Cohesion (women) Fusion Parallel Companionship Bastion Cocoon Associative Average 50 What Pluralization of the Life Course?

Five strongly contrasted styles of conjugal interactions can be found in this sample of couples representative of residents in Switzerland:

Couples of style Bastion are characterized by a strong tendency to closure, fusion and gender differentiation. In these couples, contacts with the external world are not highly valued. On the contrary, some mistrust exists toward external actors, whereas internal relationships are highly valued and sought. The family as a group precedes individual interests or orientations. This rather close and warm world is sustained by a traditional division of labour between genders, and observes rigid norms about everyday routines. These are couples where each spouse or partner knows well what his or her contributions are supposed to be, depending on gender. This strong differentiation also has an effect on orientation, women being much more inter- nally oriented in this style than men are. 16 per cent of couples show this style of interactions. Couples of the Associative style are opposite to couples of style Bastion on all accounts. They are rather low on both fusion and closure: associative couples are both open and autonomous at the same time. They also demonstrate a fairly egalitarian power and role distribution, with no significant difference between spouses on instrumentality and expressiveness. Their openness is associated with high scores on social participation for men while women are close to the average. On this basis, one can state that the central values structuring this kind of functioning are the quest for personal authenticity and, at the same time, the negotiation of individual rights. Associative couples represent 29 per cent of the sample. Couples of Companionship style are characterized by a strong tendency toward fusion, contrary to associative couples. At the same time, they are very open. The extent of sexual differentiation is close to the mean. The dominant value in this style of interaction is the use of environmental resources in order to reinforce internal solidarity and communication. Companionship couples represent 24 per cent of the sample. High levels of fusion and closure also characterize couples of the Cocoon style. Contrary to Bastion couples, they do not, however, present a high-level gender division of domestic and relational roles. They show a strong tendency to emphasize the internal goals of the union for both spouses (in the Bastion couples, only women show such a tendency). Their functioning is concurrently warm, closed and relatively free of gender biases. They represent 15 per cent of the sample. Couples of the Parallel style are characterized by a strong differentiation of domestic and relational roles between spouses. They are strong on women’s expressiveness and on men’s instrumentality. Parallel couples have comparatively low scores of fusion and high scores of closure. These are couples who feel threatened by the external world (with respect to the family) while investing little in their own internal relationships. The idea of separate worlds for each Eric Widmer, Jean Kellerhals and René Levy 51 spouse seems to be at the centre of these couples’ functioning. This style includes 17 per cent of our couples. Various dimensions of conjugal conflict are significantly associated with these different styles of conjugal interaction. Conjugal problems were measured according to a set of twenty indicators that describe a wide variety of aspects of daily life.7 They were combined into two general scales, one dealing only with current problems, the other dealing with problems over the whole duration of the conjugal relationship.8 Multiple correspondence analysis (MCA) was then used in order to find out if those general scales could be decomposed into more specific kinds of problems. Three main types of problems were delineated by this analysis: problems related to violence and other deviant behaviours (such as drug use); problems related to the coordination of activities (i.e. how to fit together the spouses’ agendas, how to develop shared usage and rhythms within the couple, how to find a satisfactory division of labour, etc.); and, finally, relational or interactional problems (problems of communication, affective disillusion, significant difficulties with the spouse’s or partner’s personality, etc.). Table 2.2 shows that the different styles of conjugal interaction are indeed associated with different types of conjugal conflict. From Table 2.2, it is clear that the Companionship style of interaction is associated with lower conjugal conflict than any other interactional style. Companionship couples report tension and open conflict less often. When open conflicts do occur, they are much less serious and reconciliation is much easier than in the case of other couples. Companionship couples present conjugal problems of all kinds much less often than other couples. In contrast, Parallel and Associative couples score very highly on almost all indicators of conflict. Both men and women acknowledge a higher level of tension and more frequent open conflict than average. They show higher rates of problems of all kinds. Cocoon and Bastion couples, on the other hand, show quite similar profiles to those of Companionship couples, although they report slightly higher frequencies of problems and open conflicts. Conjugal quality was assessed using three sets of indicators: (1) Divorce proneness, which was measured by a single indicator; (2) Conjugal satisfaction, measured by a general indicator and also by four sets of sectorial indicators, rating the level of satisfaction with the division of labour, conjugal mood, mutual consideration and coordination between spouses or partners (see Widmer, Kellerhals, Levy (2003) for a detailed description of these indices); (3) Depressive symptoms of spouses or partners, measured by indicators such as sorrow, fears, loneliness, aggressiveness, etc. (Radloff, 1977). Table 2.2 shows that there is quite a strong correlation between styles of conjugal interaction and conjugal quality. Parallel couples are associated with the highest conjugal dissatisfaction and the most frequent thoughts of separation for both men and women. Companionship couples have the lowest scores on both measures for both 52 V .20** .24** .16** .23** .23** .20** .21** .18** .20** .18** .22** .22** .11** .16** .16** .22** .15** .13** .15** .26** .15** .17** .11** 24 37 28 47 42 36 33 31 29 21 59 29 23 26 36 39 25 36 22 49 20 58 17 8 7 9 11 25 24 25 17 12 20 20 26 14 15 18 15 20 26 16 21 13 13 20 26 45 34 1616 17 14 22 21 25 26 12 17 29 29 22 21 16 14 48 48 9 8 7 21 26 18 14 15 13 14 17 15 12 34 10 12 10 17 22 10 23 37

Conjugal quality quality Conjugal Conjugal conflict

41 33 33 30 17 31 57 34 19 35 Parallel Companionship Bastion Cocoon Associative Cramer’s 38 37 17 20 19 F 33 41 F 24 M 15 F 44 F 37 F 58 sig at .01. = Female, ** = Conjugal conflict, conjugal quality and styles of conjugal interactions (%) interactions conjugal of styles and quality conjugal conflict, Conjugal Male, F Male, = : M (current or past) M C Female shows signs of depression F 35 Current open conflicts are frequent open conflicts are serious Current 20 39 Divorce proneness Thoughts of separation % with relational problems % with coordination problems Coordination not satisfactory Male shows signs of depression MGeneral dissatisfaction is high 29 % with addiction or violence problems M Lack of mutual consideration Table 2.2 Table Notes C C M Division of labour not satisfactory M F 21 M C Poor conjugal mood Eric Widmer, Jean Kellerhals and René Levy 53 genders. Associative couples have a similar profile to parallel couples, although less extreme. Cocoon and Bastion couples lie in between. Correlations between signs of depression and conjugal styles of interactions confirm these results. Again, Parallel couples show the highest scores on the scale of depression, whereas Companionship couples show the lowest scores, the other three styles lying in between. Interestingly, the scores of women are more sensitive to conjugal styles of functioning than the scores of men, as if women’s subjective well-being depends more strongly on the family system (and less on other factors located outside) than men’s. The cluster analysis above shows that contemporary conjugal interactions in Switzerland do not follow a single model. To the contrary, the space defined by the various dimensions of cohesion and regulation considered is largely used, with five styles of conjugal relationships emerging, which are associated with unequal levels of conjugal conflict and conjugal quality.

2.3 Insertions in the social structure of Switzerland Are personal trajectories and styles of conjugal functioning sensitive to the different ways and degrees to which individuals are inserted in the social structure of Switzerland? This question is an important one, as it indirectly deals with the issue raised by the hypothesis of a pluralization of life courses. One main hypothesis about the life course affirms that individual lives have become more and more disembedded from local social structures and material contingencies, with an increasing impact from an overall planetary culture through the mass media. This ‘poststructural’ hypothesis is often set in opposition to the more classical sociological hypothesis according to which individual action spaces are socially structured and that actions reflect the options, constraints and resources an actor has to deal with in her/his situation. In this section, we shall show that both personal trajectories and conjugal functioning remain deeply embedded within the social structures. Let us first consider personal trajectories. Table 2.3 presents a multinomial logistic regression using number of children, level of education and birth cohort as predictors, and the types of trajectory as the dependent variable, with ‘Homemaker’ trajectories as the reference category. The predictors are considered as indicators of insertion into the social structures. Birth cohorts indicate membership in successive ‘generations’, and therefore their unequal susceptibility to participating in the hypothesized process of individualization. The level of education is a classic indicator of social status which corresponds, from the point of view of the life course, to an intermediate stage between ascribed and achieved statuses. The number of children can be considered as an indicator of the weight of the requirements of domestic life. The number of children in a family turns out to be a highly significant predictor of women’s trajectories. Having three children rather than two or one directs women to a home-centred trajectory, while this trajectory is practically non-existent among women who have no children. ‘Full-time’ 54 234.02 ** ———— 667 1530 — ——————— 1.03 1 1.12 .86 —————— —————— — — .51* 1.24 .65 .60 — —.38** .54** .73 .46** 4.30** 5.81**1.55 .80 1.7 1 252.7**

——— — — <.05 <.05 sig = Multinomial logistic regressions modelling the probability of exhibiting a given trajectory-type and a given style of conjugal and a given style trajectory-type given a of exhibiting the probability modelling regressions logistic Multinomial <.01, * <.01, sig women = N 677 None —None — — One .70 .18** .11** .43 .65 1.55 1.29 1.47 2.06* 1.23 1.47 1.71* 1.29 1.35 1.55 1.06 .28** interactions .46** .65 .12** .11** .93 .43 .37**2.76** 1958–1967 —.85 .38 .70 1948–1957 .96.48* .48** 1938–1947 .591.38 2.24**.11** .65** —.11** .49* .92 .37** .32** One .70 — .51** Two .84Three or more .18** of level Educational 2.87** .05** .71 .93 Low —Middle 1.43 .55High 2.08 1.62* 1.29 Women’s employment — 1.97** .01** 3.32**Part-time (<50%) Part-time (50–89%) 5.09** .08** — —Full-time Chi-square 11.2 — — — .21** — — — .10** — — .66 1.15 — 1.16 1.36 — .54** .58** .59* .29** Table 2.3 Table Minority Full-timeBirth cohort Part-time1968 and after Back-to-employment Erratic Parallel Companionship Bastion Cocoon ** Before 1938 children of Number —None .70 — .75 — .75 6.02** 4.34** .66 3.29** 1.78 2.55* Eric Widmer, Jean Kellerhals and René Levy 55 trajectories much more often concern childless women than others. Levels of education also play a leading role. The likelihood of home-centred trajectories significantly decreases as the level of education rises; ‘part-time’ trajectories are more likely for higher levels of education; however, ‘full-time’ trajectories have a relatively homogeneous distribution between the levels of education. These results show that women’s trajectories are shaped by multiple social insertions. Finally, female trajectories are also sensitive to birth cohorts. Compared with the ‘homemaker trajectory’ (reference category), ‘full-time’ and ‘part-time’ trajectories characterize more recent cohorts and ‘back to work’ trajectories older cohorts.9 In contrast, men’s trajectories are not sensitive to any of the above-mentioned social insertions. Neither birth cohorts, nor children’s presence, nor level of education have a significant influence on their trajectories. Do styles of conjugal interaction follow the same logic of social insertion? In this latter case, we consider not only birth cohort, number of children and level of education, but also women’s level of participation in the workforce, which obviously had to be dropped from the list of factors considered as predictors for personal trajectories.10 Table 2.3 also presents a multinomial logistic regression with the Associative style taken as the reference category. It shows that the level of women’s education has a high impact on styles of interaction. Low levels of education are associated with a Parallel, Bastion or Cocoon style, while couples with high levels of education develop significantly more often an Associative style of interaction. The number of children also has a significant effect on styles of conjugal interaction: one-child couples develop more often the Cocoon style while couples with two children are more often of a Bastion style. Most importantly, Table 2.3 shows that professional trajectories and styles of conjugal interactions are quite strongly associated, as employed women have a significantly weaker probability than non-employed women to be in a couple of Parallel, Bastion or Cocoon style. Full-time employment of women favours an Associative style of interaction whereas part-time employed women have a greater probability of belonging to a couple of Companionship style. Although it is hard to say what is the antecedent variable here, there is, to a large extent, a homology between the way couples organize their interactions and the way female professional carreers are shaped in Switzerland.

3Conclusion

Overall, the hypothesis of a pluralization of personal trajectories and conjugal interactions received mixed support from the data considered in this research. For men, statistical analysis revealed the presence of one dominant type of trajectory, characterized by almost continuous employment exercised full-time. The non-standard, more chopped and heteroclite trajectories, characterized by periods of training, part-time jobs and various interruptions, only concern 56 What Pluralization of the Life Course? a minority of men and do not represent a clearly identifiable model or trajectory type. So we can conclude from our findings that men’s trajectories are essentially centred on full-time employment, and that for them, the hypothesis of a general pluralization of work patterns is clearly rejected by the data. To the contrary, the persistence of a strong ‘standardization’ of men’s trajectories is confirmed. In clear contrast, women’s trajectories follow several distinct models, only one of which is similar to the dominant male trajectory. The four main models are: the ‘homemaker’, where there is a continuous focus on the home; full-time employment (with a decrease between 25 and 35); part-time employment throughout the trajectory; and, fourthly, women who re-enter the work environment after a period as homemaker. Only a very small minority of women do not follow one of these patterns. This ‘bounded diversity’ of female trajectories clearly follows a logic of social insertion, as the models we found are strongly associated with the number of children born to the couple, their level of education and birth cohort. The hypothesis of individualization corresponds only to a very small number of women’s trajectories. Therefore, one should emphasize the fact that both male and female personal trajectories follow a limited set of sequential models. A large proportion of men’s trajectories correspond to the tripartition model (Kohli, 1985). For women, several models exist; although more varied, they are limited in number and depend upon insertion in the social structure. Thus, in both cases, the hypothesis of far-reaching pluralization of personal trajectories has to be rejected for the couples residing in Switzerland. Our results also reject the hypotheses stipu- lating a single, gender-neutral trajectory type, along with a gender-typed dichotomy of trajectories, and clearly favour our fifth hypothesis of restricted variation dependent on elements of social insertion. How, then, are we to explain the distinct patterns characterizing men’s and women’s trajectories? The hypothesis of the master status (Krüger and Levy, 2001; Widmer, Kellerhals and Levy, 2003) may help us to understand these differences. This hypothesis advances that there is, for both sexes, a ‘privileged’ domain of activity to which other domains are subordinate. In many western countries, including Switzerland, the privileged domain of activity for men is still the occupational field, while the privileged domain of activity for women is still the domestic field (family work), whatever ‘modernization’ of gender relations may have taken place since the middle of the twentieth century. The dominance of one domain for each sex does not necessarily imply their exclusion from other domains – it simply means that the investment made by individuals in their secondary spheres of activity depends upon the demands of their major sphere of responsibility. So, the majority of the female trajectories are not characterized by an exclusion from the professional sphere – a situation of extreme sexual segregation, doubtless still predominant in some social circles but more and more openly criticized by public opinion and the mass media. Women’s trajectories rather reflect Eric Widmer, Jean Kellerhals and René Levy 57 a subordination of their gainful employment to the imperatives of family life. Therefore, it is necessary to emphasize the interdependence existing between male and female trajectories for individuals living in a couple or in a family: if most men can disregard domestic imperatives – often a structural necessity in terms of career – it is because their female partners essentially take care of them. There is a systemic twist here that has its roots deep in the economic organization of many capitalist societies. It represents a particularly clear case of the principle of ‘linked lives’ (Elder, 1995) which suggests that in order to understand male and female individual trajectories one should focus on the interdependences that associate them within the family. One can make the hypothesis that this interdependence between sex-specific master statuses has a highly resistant power in many western countries which works against the logic of the pluralization of personal trajectories. What about conjugal functioning? Our data show that in this respect also contemporary conjugal interactions do not fit a single model. To the contrary, the space defined by the various dimensions of cohesion and regulation is largely used. It is also worth underlining that these styles of conjugal interactions are associated with the social status of the spouses: the Bastion, Parallel and Cocoon styles are more frequent in couples of low social status. The frequency of the Associative style of conjugal interaction is significantly greater in couples with higher social status. Results also show that the different styles of conjugal functioning have other distinct properties: a strong emphasis on autonomy and closure of family life, and a rigid and gendered organization of conjugal roles are associated with increased conjugal problems and a deterioration of conjugal quality.11 Analysis of both personal trajectories and conjugal functioning reveal the probably real but certainly limited extent to which we may speak of a pluralization of life courses. A finite set of models, in both cases, captures a large part of the overall variance existing within our sample of couples living in Switzerland. Therefore, the hypothesis that we are in a time of extreme individualization of key aspects of life should be rejected. There are several models available which highly structure the ‘choices’ of individuals or couples. In fact, these models are not freely chosen, but rather depend on the resources and constraints associated with specific locations in the social structure. They also have specific consequences, which suggests that they play a structuring role of their own. In this perspective, one may consider models of trajectories and of conjugal functioning, along with social networks (Widmer, Kellerhals and Levy, 2004a), as key process-oriented variables for the production and reproduction of social inequalities in contemporary societies. This suggests that a life course perspective is likely to provide new understanding about social structures and their effects in late modernity. Are these results specific to Switzerland? To make such a diagnosis is, of course, a difficult endeavour as long as the comparison is not a common goal of the studies to be compared from the outset. There are so many parameters 58 What Pluralization of the Life Course? to be controlled in order to assure comparability in any strict sense that it is highly unlikely to find studies that fit together without having been planned to do so – and this is in fact the case for ours. However, one similar study of work career patterns of men and women exists in the United States (Moen, 2003; Moen and Han, 2001) and another in Germany (Born et al., 1996; Krüger, 2001). Note that their samples are not exactly the same as ours which excludes the possibility of answering some interesting questions and forces us to locate the comparison at the level of interpretations rather than of direct data comparison. Nevertheless four common features emerge: 1. life-course patterns are strongly gendered; 2. there is one single predominant male trajectory, while there is a higher number of female trajectories (including the male one, but without a clear numeric predominance); 3. only female occupational trajectories reflect variations due to the family life cycle; 4. part- time employment remains limited to women’s trajectories and seems to be part largely of their adaptation to the requirements of family life. The main difference we can see singles out the US where the ‘pure’ homemaker career, present in Switzerland (and in Germany), seems no longer to be present to a significant extent. How is this difference likely to be explained? Let us just mention the beginning of comparative research literature on the influence of the institutional framework in various societies on the rules relevant for the couples’ choices concerning the partners’ employment (and, thus, the shape of employment trajectories, especially of women). These comparisons are working with adapted typologies of regimes of social policy (Pfau-Effinger, 1999; Künzler and Walter, 1999; Korpi, 2001, Boje and Almqvist, 2000; Fux and Baumgartner, 2001). These studies show that the Swiss situation, that has been our focus, belongs to an especially ‘liberal’, i.e., non-interventionist and de facto inequality-fostering regime otherwise typical of the Anglo-Saxon countries. This makes us think that the modernization of family traditionalism which we have found instead of a simple disappearance of the traditional model, expresses a resilience of traditional elements that may be due to a large extent to the extrafamilial context of living and its ways of functioning. In this perspective, a considerable part of the family structure appears to be institutionalized indirectly by exogenous social rhythms and structures (see other contributions to this volume) as much as or even more than by the personal beliefs and interpersonal negotiations of individual partners (Levy, Widmer and Kellerhals, 2002). In terms of conjugal interactions, it was underlined that our study is unique in its inclusion of a variety of indicators and in its large and representative sample, as well as its inductive method, based on cluster analysis (Widmer, Kellerhals and Levy, 2003). Therefore, inter-country comparisons are even more difficult to undertake for styles of conjugal interactions than for professional trajectories. But, overall, the same diversity of family interactions and family structures seems to characterize other countries in Europe, with a variety of living arrangements, conjugal and intergenerational Eric Widmer, Jean Kellerhals and René Levy 59 relationships characterizing Italy, Spain, France, Portugal, Denmark, Holland, the United Kingdom and Germany (Rossi, 2003). Let us, however, emphasize once more that the large number of couples characterized by a high degree of fusion (i.e, the opposite of conjugal individualism), by strong sex typing of conjugal roles and by the presence of clear-cut master statuses, suggests that, despite a large commonality among western countries, including Switzerland, family change may have not gone as far in Switzerland as elsewhere. Again, the idea of the modernization of family traditionalism seems adequate to describe the Swiss situation. Finally, these results open up an interesting question. On the one hand, we begin to better understand how life-course regimes in Switzerland may relate to institutional features specific to the country, such as its peculiar daily school schedules, its lack of infrastructure for toddlers, its occupational training system (the apprenticeship), its gendered job market, the heretofore compulsory, men-only military service, and so on. On the other hand, we find that some rather basic aspects of men’s and women’s professional trajectories and family interactions are the same in Switzerland as in other western countries, such as Germany and the United States. The question – which we can only formulate at this time – is to consider whether these similarities are due to common cultural features between Switzerland and the other western countries that are sufficiently strong to homogenize life courses and conjugal interaction patterns whatever the country’s institutional outfit, or whether they are due to quite different – but functionally equivalent – forms of insti- tutionalization in Switzerland and in other countries that produce the same life-course and family patterns.

Notes

1. Many sociological studies focus either on professional work or on family interactions. This was criticized as reifying a gender bias within scientific research. This chapter is based on the observation that professional work and the family are interrelated spheres of activity and that it is scarcely possible to understand what happens in one without taking the other into account. 2. In a recent article however, Kohli (2003) acknowledges the importance of gendered life courses and thus signals some distance to the postulate of one single model. 3. The most notorious expression of such a perspective is represented by some of the latter work of Anthony Giddens. In his books Modernity and Self-identity (1992) and The Transformation of Intimacy: Sexuality, Love and Eroticism in Modern Societies (1994), the concept of a ‘pure relationship’, which is an archetype, according to Giddens, of late modernity, directly or indirectly refers to several key dimensions of the companionship model of family of Burgess et al. (1960), viz.: a focus of the relationship on the exploration of the self, the centrality of negotiation processes, a symmetry in power relationships, a weakening of the effects of external constraints on intimate relationships, etc. 60 What Pluralization of the Life Course?

4. It should be noted that, to various degrees, the hypotheses we mention postulate processes of change. These processes do presumably not cover the same periods. Kohli’s theses about standardisation and its different components concern the last two or three centuries whereas the thesis of postindustrial pluralisation and individualisation concern roughly speaking the last fifty years. This difference implies that there is no logical contradiction between these theses, since they concern subsequent rather than simultaneous historical periods. If our data do not allow for a historical analysis, they can show quite conclusively to what degree the one or the other of the processes implied by these hypotheses most strongly structure the present situation in life course dynamics. 5. A short presentation of this method, as well as some references, can be found in Widmer et al., 2003. 6. Measures of fusion, closure and orientation were included for men and women separately, as spouses may have quite different perspectives in this regard. The dimensions of regulation were measured using only the responses of a single informant per couple (women), as they are more factual. 7. If either of the two spouses or partners cited a problem, we assumed that the couple had this problem. 8. See Widmer, Kellerhals and Levy (2003 and 2004a) for a full description of the scales. 9. This effect, however, may be spurious, as women at younger ages have not yet had the opportunity to leave and go back to the job market. Analysis done on a much smaller subsample including women aged more than 50 (n = 277) revealed that no change concerning this type of trajectories occurred between cohorts born between 1938–1947, and before 1938. It is too early, however, to estimate changes of occurrence of this type of trajectories across younger cohorts. 10. A more elaborated list of factors is considered in Widmer, Kellerhals and Levy (2004b). 11. For an explanation of these associations, see Widmer, Kellerhals and Levy (2003 and 2004b). 3 Facets of Emotion Regulation in Families with Adolescents: A New Research Approach Meinrad Perrez, Dörte Watzek, Gisela Michel, Dominik Schoebi, Peter Wilhelm and Yves Hänggi

Family is the preferred place for the regulation of individual and relational affectivity, for the heating and tuning of emotions. Larson and Richards (1994) call it an ‘arena of emotions’. The project ‘Regulation of Emotions in Families with Adolescents’ explored the processing of emotional and somatic states, the perception of other’s emotional states and the social regulation of or coping with emotionally heated social situations in the family. We expected the sensitivity for family members’ emotions (i.e. accuracy of the perception) and the modalities of coping to be relevant prerequisites for the development and the well-being of the family members. An additional goal was the development of a new methodological tool to assess affective, cognitive and behavioural phenomena in the natural setting of family life. The questions we were focusing on concern (i) feelings and emotional states as a function of the time of the day, the day of the week, the social settings and activities, (ii) the role of accurate reading of the other’s emotional state in families and couples, (iii) functional and dysfunctional modalities of interpersonal emotion regulation, and finally (iv) the evaluation of the possibility for practical applications of the new assessment tool in counselling and clinical settings. The following text describes some core facets of the project.

1 Introduction

Although the modern family may have lost many of its former core functions, some of its other key functions have become even more significant. For Nave-Herz (1997) and many sociologists the emotionalization and intimization is the most important characteristic of modern families. This feature is more important than the pluralization of family-forms, which was already present in pre-industrial family structures. Family became, independently of the particular form, the place par excellence for the exchange of intimate

61 62 Emotion Regulation in Families with Adolescents feelings, emotions and other concerns. It is the privileged place for the regulation of individual and relational affectivity, and for the regulation of emotions. Larson and Richards (1994) call it an ‘arena of emotions’ – not only of positive, but also of aggressive emotions, and occasionally it is the site of both love and violence. Ariès (1975) has stated that the families’ function of emotion regulation, with the increasing levels of emotionalization and intimization, has become the dominant element during the last century. Nahrstedt (1992) showed how this trend to the intimization can be observed in a changing architecture of habitations in Europe since the seventeenth century. Luhmann (1982) considers the present family to be the only system equipped to address the emotional needs of modern society. How does this affective regulation work within the family? Summary accounts have taught us a lot for sociological and psychological exchange theory and social psychology research. We know something about satisfactory relationships as a function of the costs and the benefits of intimate interaction. We learned, among other important aspects, that social comparison and the perception of justice in relationships play a key role (Mikula and Freudenthaler, 2002). However, we know very little about how the interpersonal regulation of emotions, that is, about their mechanisms and processes. One major reason for this current situation is the methodological restrictions that are inherent in the traditional approaches to study within family processes.

2 Methodological aspects of traditional measurements in family psychology

Most research on emotion regulation, on conflict behaviour and stress in families and couples, has been undertaken using questionnaires on the basis of retrospective self-report data. Buehler’s (1990) review of assessment tools in the area of family stress and conflict research suggests that such methods account for more than 80 per cent of this type of study. Even today, any examination of journals on family or couple psychology will yield similar results. A crucial question in dealing with emotional states and behaviour performance of daily life, as an object of research, is how to assess these phenomena. Traditional questionnaires ‘ask participants to retrospect over weeks and months and provide summary accounts of their psychological states and experiences’ (Bolger, Davis and Rafaeli, 2003). However, over such a period of time people cannot recall reliably how they behaved. If retrospectively self-reported act-frequencies in a group discussion are compared with the results of coding by observers, the agreement is low (Gosling, John, Craik and Robings, 1998), even if the delays between behaviour or emotional states and retrospective assessment are relatively short (Käppler and Rieder, 2001). Smith, Leffingwell and Ptacek (1999) found that on average only 25 per cent Perrez, Watzek, Michel, Schoebi, Wilhelm and Hänggi 63 variance shared between the daily and retrospective accounts of coping behaviour. Studies from different fields of psychological research provide convincing evidence that retrospective data are not reliable (Margraf et al., 1987; Fahrenberg, 1994; Katschnig, 1980). Retrospective self-report data are impaired by memory bias, such as the mood dependency of memory, the need for consistency and so on. Similar problems arise if subjects have to infer summary accounts on frequencies of their behaviour as indicator for their present behaviour (Bolger et al., 2003). Typical summary accounts of self-report data mirrors the subjects’ beliefs about their behaviour, emotion and cognition rather than how they actually behaved, felt and thought in particular situations; the same problem arises in the estimation of frequencies of behaviours, emotions and cognition. This assessment approach may be useful and valid for the observation and measurement of the subjects’ beliefs about how they felt and behaved in the past. However, if our object of study is not beliefs on one’s own or other’s behaviour and emotional experiences, but, rather, their actual behaviour, emotional and cognitive performances, we need to employ better observational tools than traditional questionnaires (Perrez, Wilhelm, Schoebi and Horner, 2001). In order to avoid the impairing effects on validity of usual self-report data, we need other, more appropriate assessment strategies. One major goal of the Fribourg project on emotion regulation in families with adolescents was to develop and evaluate a new methodological tool for studying family processes. This objective took into consideration that in much traditional family research only one parent was the source of self-report data for the whole family (Buehler, 1990). The new procedure should engage with all family members. Furthermore, the data assessed have to be situated clearly within the theoretical framework.

3 Research questions

The project ‘The Regulation of Emotions in Families with Adolescents’ deals with the processing of emotional and somatic states, the perception of other’s emotional states and the social regulation or coping of emotionally heated social situations in the family. We expect those features of sensitivity for other’s emotions in the family (that is, the accuracy of their perception) and the modalities of coping are relevant conditions for the development and the well-being of the family members. A further and basic purpose was the development of a new methodological tool for assessing affective, cognitive and behavioural phenomena in the natural setting of family life. The questions we are focusing on concern: (i) feelings, emotional and somatic states as a function of the time of the day, the day of the week, the social setting (family) and activities; (ii) the role of accurately assessing the other partner’s emotional state in families and couples; (iii) the functional 64 Emotion Regulation in Families with Adolescents and dysfunctional modalities of interpersonal emotion regulation; and (iv) finally the applicability of the new assessment tool.

4 How to assess affective, cognitive and behavioural phenomena in the natural setting of family life

In order to avoid the impairing effects on validity that are usually present in self-report data, we need assessment strategies which meet the following three criteria: (i) The procedure should allow for an assessment of the information mentioned above, in daily life under the social and environmental conditions, in which these states and events occur; (ii) The time lag between the experienced states and behaviour and their recording should be as short as possible, in order to minimize memory distortions. Information should be recorded when the subjects are still in the state of emotional arousal for storing ‘hot emotions’ and ‘hot cognition’; (iii) The method should assess psychologically relevant data, not using a diary-free text self-description, but structured according to the theoretical framework. We have therefore developed, over the past few years, a systematic self- observation method based on the use of pocket computers. Our assessment procedure has its roots in the experience of the COMRES (COMputer REcording System), that Perrez and Reicherts (1996) developed and evaluated in earlier projects on the recording of individual stress experiences and coping. The pocket computer is used as an external memory for stress, which is applied directly in daily stressful situations. This allows a minimizing of the problems of memory and the subjective retrospective distortion. The purpose of our research method is to have access to the emotional life of families – this includes both positive and negative emotional and somatic states. The new method not only had to integrate a broader range of emotional states; it also had to involve the social contingencies of an individual’s emotional state in the form of the behaviour of other actors present in the family. The event-sampling method for the assessment of individual stress experience and coping was therefore adapted to a time-sampling instrument to assess positive and negative emotional states, related cognitive activities and the social regulative activities of emotions from all members of the family. What is self-recorded? The method is devised for couples and families with adolescents. Table 3.1 summarizes the different information types and item formats assessed by the family self-monitoring system (FASEM-C).

4.1 Experience and behaviour sampling The behaviour of other family members plays a major role in any analysis of the social regulation of emotions. This social relatedness of the assessed behaviour experience makes a time-sampling strategy indispensable. The first recording is made immediately after waking up in the morning, and Perrez, Watzek, Michel, Schoebi, Wilhelm and Hänggi 65

Table 3.1 FASEM-C information types and item types

Information type Item types

1. Duration and quality of sleep 5 Items (6- and 7-point scales) (1st observation of the day) 2. Emotional and somatic state 11 (6-point scales) 3. Causal attribution Internal or external attribution 9 Items (3-point scales) other persons 9 Items (3-point scales) 4. Control expectation by myself or by other persons (triggered by 2) which other persons 10 Items (3-point scales) 5. Somatic complaints/pain 10 Items (3-point scales) Attribution 13 Items (categories) 6. Place 6 categories 7. Setting 14 Items (categories) Presence of other persons 9 Items (categories) evaluation of pleasantness 1 Item (4-point scale) 8. Evaluation of partner’s affective 8 Items (categories and 6-point scales) state (only for parents) 9. Actual activity 16 Items (categories) valence of activity 2 Items (6-point scales) Consumption of drugs 7 Items (categories) 10. Conflicts with others Yes/No, triggers: Social coping 3 Items (3-point scales) description of own behaviour 12 Items (3-point scales) description of other’s behaviour 12 Items (3-point scales) 11. Individual stress 2 Items (category and 3-point scale) 12. Individual coping (triggered by 11.) 12 Items (3-point scales) 13. Social support Yes/No, triggers: by whom? 4 Items (categories) 14. Evaluation of own and other’s (triggered by 13) behaviour 2 Items (6-point scales)

the other five measurement points occur at intervals during the day. The handheld alerts the subject acoustically according to a random time point inside a time window of about three hours, five times per day during one week. At the given signal all family members had to record simultaneously their current emotional and somatic state and the other information mentioned above. This includes information on stressful individual or social events and coping with respect to the period between the previous recording and the present moment. The questionnaire programme is adaptive and the presentation of several questions is conditional upon previous answers. For example, if the question ‘Do you feel any pain or physical complaints?’ produces the answer ‘yes’, one has to answer supplementary menus that ask 66 Emotion Regulation in Families with Adolescents for the reason. The signal-contingent assessment assures the simultaneous recording of all members of the family (older than 13 years). In the same way social support and coping reactions can be assessed. Coping responses are only asked for if an individual or a social stressful situation has been previously observed. Questions on social coping depend upon the involvement of other people in the situation (e.g. conflicts). For situations, requiring adaptation without social involvement, individual coping items are presented. The programme then only presents questions concerning control expectations if a negative emotional state has been previously recorded. The total time spend on recording data depends upon the questions asked and whether more detailed information was required or not as a result of the answers given. For the shortest version the needed time for the protocol was 4.65 minutes (mean; SD =2.89) and for the complete version 7.28 minutes per protocol (SD = 3.82). After the self-observation week all subjects were questioned about their experience in using the assessment method. For most participants the duration of the recording task was acceptable (76 per cent), and nearly all participants (94 per cent) viewed the exercise positively. These results are in accordance with the experience of the previous study (Perrez, Berger and Wilhelm, 1998) and belies the expectation that ambulatory computer-aided self-monitoring disturbs participants, as suggested by Asendorpf and Wilpers (1999). For further details with respect to methodological criteria such as reliability, reactivity and validity and with regard to technical aspects the reader is referred to Perrez, Wilhelm and Schoebi (2000) and Perrez et al. (2001).

4.2 Description of the samples Swiss families with adolescent children were recruited by sending information brochures to families with schoolchildren at Fribourg and Bern in Switzerland. Only families with children in the seventh and eighth grades were recruited. Families interested in participating returned an application form and were later contacted by a member of the research team. We present analyses from the Second and Third Fribourg family project. In both projects data collection by pocket computers was supplemented by questionnaires. During the self-observation family members had to record information over the course of one week, six times a day simultaneously on a random time-sampling plan. Only families with father, mother and at least one seventh or eighth grade child took part in the studies. Siblings were also allowed to participate. In the Second project 96 families were recruited (N =314), in the Third project data from 77 families could be collected. As families took part with at least one adolescent, the whole data set was from 569 subjects within 173 couples. The average age of the parents was around 45 years of age; fathers were slightly older than the mothers. Most of them lived in a stable partnership with a length of 18 or 19 years. The mean age of the children was 15 (SD= 1). Perrez, Watzek, Michel, Schoebi, Wilhelm and Hänggi 67

Participants were middle or higher educated. More than 80 per cent of the fathers work full-time. Employment rate of women (60 per cent) is comparable to a representative Swiss sample (59 per cent) (Swiss Labour Force Survey (SLFS), 2002, 2003). Employment rate of women in Netherlands is about the same and in Germany or Austria 10 per cent lower. The rate of part-time working women is as high as the mean within these four central European countries (International statistics, 1999).

5 Some results

In the following section we refer to the results of the second and third Fribourg Family Studies. We first explain the rhythms of emotional and somatic states. What circadian pattern can be shown and what influences can be attributed to different days of the week or different settings? The accurate reading of the emotional state of family members is shown as an important factor in family life. In the third part of this chapter we look at functional and dysfunctional modalities of conflict behaviour. At the end of the chapter we show possible applications of the new assessment tool.

Statistical analyses We used multilevel analysis for most investigations to control the multiple dependencies present data (see Rasbash et al., 1999; Bryk and Raudenbush, 1992). Generalized linear multilevel models were used to analyse the binary data of symptom reporting, to estimate the influences of external factors on an emotional state and to get coefficients for different components of accuracy and assumed similarity (projection). For an extensive explanation of the appraisal of these statistic methods in our context see Wilhelm (2001) or Michel (submitted). In order to test the functionality hypotheses within conflict behaviour, we analysed data using hierarchical multiple regression models. In order to control for confounding the initial level of stress in the conflict situation and the experienced stress outcome (measured by emotional state), we additionally tested effects over three records subsequent to the conflict reported, controlling for the previous level of stress. Similarly, we controlled for concurrent relationship satisfaction to test long-term effects. We tested differences in the performance of functional and dysfunctional interpersonal emotion regulation using a multilevel ANOVA model.

5.1 Emotional states and symptom reporting over the day and over the week and in different settings Operationalization of emotional state and physical complaints In this section we refer to changes in emotional state during a normal week, and consider those situational factors that influence well-being. In the second 68 Emotion Regulation in Families with Adolescents and the third Fribourg Family projects we asked the family members to judge their own state for a list of bipolar adjective pairs: ‘Under tension–Relaxed’, ‘Sad/depressed–Happy’, ‘Concerned/anxious–Confident’, ‘Angry–Peaceful’. These items can be aggregated on the base of satisfying factorial parameters to ‘emotional state’ (see Perrez et al., 2000). The symptom report question in the self-monitoring procedure was phrased as follows: ‘Do you currently suffer from any somatic complaints or pains?’

Results Using multilevel analyses we could show that typical patterns exist over time. The pattern concerning emotional state is demonstrated in Figure 3.1. Within the course of a day the emotional state usually increases from morning to evening, with two turning points (significant linear and cubic trend). On Friday the increase from morning to evening is twice that of the other days. For most people the weekend really begins on Friday afternoon and this is associated with a substantially better emotional state. On Sunday the within- day pattern deviates from that of the other days, because in the afternoon the emotional state turns down and decreases until evening. This can be interpreted as the mental and emotional anticipation of work or school. The within-day patterns were different between females and males, but not between parents and adolescents. The pattern within days or within weeks provides an interesting insight into emotion regulation. Which factors beside the daily or weekly changes have an influence on the emotional state, is the question of this paragraph. For the effects of setting variables on emotional state see Figure 3.2: Being with other people is related to a better emotional state, than only being with family members. When an out-group person is present in the

80 males 78 estimated (controlled) Emotional state 76 very bad to 100 good

females 0 74 estimated (controlled) 72 Mon Sat

Figure 3.1 Estimated time-related effects on emotional state, when situational and psychological factors are controlled 69

Days (Ref: Mo-Th) Friday Saturday Sunday

Location (Ref: at home) at Work/school different place

Contact (Ref: only family) alone no familymembers families and others

Activities (Ref: spare time) paid work house work having a meal sleeping/resting other activities

Physical complaints

Conflicts with family members settled family members not settled others settled others not settled

other strain –20 –15 –10 –5 0 5

deviance from mean, unstandardised regression coefficients Range of T-lines indicate the 95% confidence interval

Figure 3.2 Influence of different settings on well-being, deviance from mean (week, all participants) 70 Emotion Regulation in Families with Adolescents family their well-being is better than in family intern situations. The same was observed in situations only with other people. In the first family project we found the same results using a conventional strategy to analyze the data (Perrez et al., 2001). Independent of local and social setting, activities have an influence on the emotional state. Following leisure activity or taking a meal is associated with a better emotional state, than other activities. The worst emotional state is associated with working for the job or the family. The control of setting, social contacts and activity reduces the weekend effect, as does the amount of increase in the emotional state from morning to evening. This could be expected because settings and activities as well as the social contacts are structured over time. The different effects are therefore confounded. But the structure remains after the statistical control for these setting-effects. Results concerning symptom reporting showed a similar pattern, with the probability of reporting symptoms being higher in the morning and in the evening and lower in the middle of the day. Men and women do not differ in this general pattern of symptom reporting, but for women the probability is higher throughout the entire day (average probability of symptom reporting: men = 0.149; women = 0.202; β = 0.365; SE = 0.160; p < 0.05). In contrast to this, parents and children do not differ in their general probability of symptom reporting, but in their diurnal pattern. Children show a higher linear increase of symptom reporting earlier in the day (β = 0.076; SE = 0.014; p < 0.001). But the average probability of symptom reporting does not differ between weekends and weekdays (Saturday: β = 0.026; SE = 0.064; p = 0.688; Sunday: β =−0.030; SE = 0.065; p = 0.648). If the setting and activity is taken into account, a difference between work/ school and other activities or setting is observable. During work or school less symptoms are reported than in other settings (β =−0.419; SE = 0.094; p < 0.0001). 20.8 per cent of all complaints are experienced at home. In contrast, only 13.3 per cent of all complaints are experienced at work or school (Michel, submitted).

Conclusions • Situational features such as the time of the observation, local setting, social contacts and current activity each have their own significant effect upon the emotional state and upon symptom reporting, even if the influence of the other variables was held constant. • Mean differences in well-being between family members were rare and rather small. Women report more symptoms than men, confirming the findings of previous studies. However, memory effects cannot be held responsible for this finding. Selective bodily attention might be a more fruitful hypothesis to explain this difference. Perrez, Watzek, Michel, Schoebi, Wilhelm and Hänggi 71

• Large differences in the variation of the emotional state could be demonstrated. • The curvilinear pattern of symptom reporting found in an experimental study by Goebel and Cordes (1990), was replicated in our real life field study. (i) Attention which is focused externally during working and school hours in the day can probably not explain this result. This finding was true even if the activity of work and school was taken into account. (ii) Another explanation for the increased symptom reporting at home might be that individuals suffering from a certain disease such as flu, do stay at home and do not go to work while ill. Thus, more symptom reporting can be expected at home than in other settings. • The difference in the symptom reporting pattern between parents and their children is a new finding. A possible explanation is a higher fatigue and tiredness in children at the end of the day, increasing various symptoms of exhaustion. However, this is so far only a hypothesis and has to be explored in future studies. • For emotional and somatic states several analogies were shown. But there was found a weekend-effect for well-being and none for symptom reporting.

Social emotion regulation in the family depends not only upon the manifestation of emotions in at least one actor – manifestation that can be influenced by circadian dynamics and many other factors and incidents. A further precondition is that emotions are recognized and read by other actors in the family. The following section examines the accurate reading of the other’s emotional state in the regulation of social emotion.

5.2 Accurately reading the other’s emotional state in families and couples Inferring correctly the current emotional state of another person is assumed as a central condition for successful social emotion regulation and interaction in the family. In the literature this ability is called empathy – or is at least seen as an important component of empathy. To emphasize the aspect of an accurate perception and inference Ickes and colleagues have introduced the concept of empathic accuracy, which is defined as the ‘ability to accurately infer the specific content of other people’s thoughts and feelings’ (Ickes, 1997, p. 3). They also have created experimental paradigms to study empathic accuracy (Ickes, 2001). Although these experiments are powerful tools, for the study of peoples’ accuracy in perceiving and inferring thoughts and feelings of their interaction partners, they have their limitations. Interactions were short and influenced by the laboratory context (waiting room situation, or the instruction to discuss 72 Emotion Regulation in Families with Adolescents a marital problem). Thus, the range of thoughts and feelings was restricted, and the intensity of the feelings rather low. Due to the brief interactions, empathic accuracy has only been studied at the micro level (changing states in a time frame of seconds and minutes). With our approach we could overcome limitations of such laboratory experiments and we were able to study empathic inference in couples’ and families’ daily lives.

Operationalization of accuracy and projection In addition to the emotional state of the self, parents should estimate the emotional state of their partner. Using multilevel methods we were able to get coefficients for different components of accuracy and assumed similarity (projection). We could therefore separate the spouses’ accuracy in judging their partners’ actual emotional state (at a single observation) from the accuracy in judging their partners’ daily emotional state (average over six observations per day) and the accuracy to judge their partners’ general emotional state (averaged across all 42 observations). Results reported below are based on the data of 95 husbands and wives of the Second family project (for an extensive report, see Wilhelm, 2004).

Accuracy is the degree of correspondence between the spouses’ partner judgements and their partners’ self-judgements. Assumed similarity or projection is the degree of equality between the spouses’ partner judgements and their own self-judgements. The husbands’ projection measures how similar husbands’ own feelings are to their judgements of their wives’ feelings. As accuracy can be inflated by projection, when similarity between the partners is high, the accuracy coefficients should be separated into the part that is due to correctly assumed similarity (indirect accuracy) and the part that does not rely on assumed similarity (direct accuracy) (Kenny and Acitelli, 2001). This can be achieved by estimating accuracy and projection in one and the same analysis.

Results Accuracy coefficients at all three levels were much higher than expected by chance, even when projection was controlled for. However, the level of projection was higher than the level of accuracy. The high degree of projection in judging the partners’ general emotional state seems to be due to the fact that spouses use the scale to judge their partners’ emotional state in the same way as they judge their own emotional state (response style). Accuracy in judging the general state after projection has been controlled decreased substantially (from .59 to .34). This indicates that spouses reach quite a substantial amount of accuracy through correctly assuming that the partner in general feels quite similar to the spouses. Indeed, the similarity between the general emotional state of partners was substantially correlated Perrez, Watzek, Michel, Schoebi, Wilhelm and Hänggi 73

(r = .44). For the judgement of the daily and actual emotional state a similar pattern could be observed, although decreases in accuracy were smaller. In line with the literature (Brody and Hall, 2000; Ickes, Gesn and Graham, 2000) women tended to be more accurate in judging their partners’ general emotional state than men (t = 1.92, p < .06). However, concerning the judgement of the daily emotional state and the judgement of the direct emotional state men were as good as women in judging their partners’ emotional state. We further examined accuracy and projection, when partners were together and when they were not together. The results confirmed our expectation that accuracy was significantly higher (t = 2.46 < .05), when the partners were together. Because they can directly perceive each other’s verbal and non-verbal behaviour, and they have an exact knowledge about the situational circumstances, they are better able to judge and infer their partner’s state. However, they also reached an accuracy much higher than expected by chance (t = 6.36, p < .001), when the partner was absent, indicating that content knowledge and inference are sufficient for a basic accuracy (see Wilhelm and Perrez, 2004, for a more extensive discussion). When spouses were together they tended to project much more than in situations in which they were not together (t=7.69, p <.001). This is surprising, because one could expect that partners base their judgement on the direct perception of their verbal and non-verbal behaviour and therefore do not need to project as much. Perhaps assuming that the partner has similar feelings, even if he actually does not, is a way to create the illusion of being in a common mood and therefore in harmony with the partner (Murray, Holmes and Griffin, 1996). Neither accuracy nor projection seems to be moderated by the couple’s relationship. This result does not confirm intuitive expectations, that accuracy should lead to better satisfaction (Noller, 1984; Gottman and Porterfield, 1981). However, there seems to be a more complex relationship between accuracy, projection and relationship satisfaction (see Murray et al., 2002; Ickes and Simpson, 1997; Murray et al., 1996).

Conclusions • Spouses are quite accurate in judging the emotional state of their partner during their normal daily life. • When judging the partners’ emotional state, spouses use their own emotional state as the main guideline. Content knowledge and inference are sufficient for a basic accuracy. • Accuracy increases substantially, when the behaviour of the partner can be directly perceived. • A part of improvement in accuracy, when the partner is present is due to correct assumed similarity (‘projection’). However, assumed similarity is exaggerated and is much higher, than real similarity. 74 Emotion Regulation in Families with Adolescents

• Individual differences in the amount of accuracy and ‘projection’ are large. • Sex differences are rather small. • There seems to be a complex relationship between accuracy, projection and relationship satisfaction.

To read accurately or inaccurately the emotional and somatic state of the partner or a child in the family can be expected as an important factor influencing the outcome of social emotion regulation (Swann, Hixon and De la Ronde, 1992). There may exist some exceptions concerning the benefits of accurate perception (Flury and Ickes, 2001) and some kind of errors (e.g. idealization) which are beneficial for the interaction, as has been discussed for example by Murray, Holmes and Griffin (1996). However, the way to perceive the other’s state is not the only factor influencing the outcome of social emotion regulation. In most cases, the accurate perception of the other’s emotional state may be a helpful condition for a functional social regulation of his of her emotional state, but it is not a sufficient condition. In particular, one needs specific competencies to regulate conflict situations, situations with divergent interests and negative emotional states. What are the modalities of social emotion regulation, which can be considered as functional? The next section is dedicated to this question.

5.3 Functional and dysfunctional modalities of interpersonal emotion regulation Another aspect of primary importance for successful emotion regulation is the quality of interpersonal interaction. Conflict issues, one example of a stressor inside a relationship, have been compared to stressors outside of a relationship or system (see Bodenmann, 1995). In particular, the way people deal with stressful events occurring within a family, or more generally within an intimate relationship, appears to be a critical issue concerning the developmental course of the family and its members. Research has broadly shown that marital conflict has detrimental effects on relationship quality, family functioning, and mental and physical health. A variety of conflict characteristics have been related to the panoply of outcomes on different levels of individual and interpersonal well-being. Research on interaction behaviour during conflict yielded a detailed view of the interaction process, leading to the conception of distinct interaction and communication styles that are predictive of outcomes in relationship satisfaction and divorce (Fincham, 2003). The standard approach to assess interaction behaviour in a conflict situation is the observation of conflict-related interaction in communication laboratories. Laboratory discussions are videotaped and coded for specific interactional cues (e.g. Gottman, 1994). However, research has also shown that the subjective evaluation of the conflict process and cognition during conflict equally account for detrimental effects of conflict (Holmes and Murray, 1996; Perrez, Watzek, Michel, Schoebi, Wilhelm and Hänggi 75

Karney and Bradbury, 1995). It is therefore reasonable to account for the multiple individual perspectives on conflict behaviour if setting the focus on interpersonal emotion regulation. The self-observation methodology allows us to meet this requirement by simultaneously tapping the subjective perception of the subjects own and their partner’s behaviour. In the Fribourg Family projects, we assessed the subjects’ perception of both intra-familial and extra-familial interpersonal conflicts. Among other aspects, we placed a major focus upon the question of the functionality of interpersonal behaviour in terms of emotion regulation strategies. We developed some theoretical criteria in order to determine the functional capacity of interpersonal behaviour for successful emotion regulation. Fur- thermore, we suggested differences in effects of conflict behaviour with regard to self-perception and partner-perception measures. We expected perceived conflict behaviour of the partner to be a more powerful predictor for emotion regulation outcomes than self-perception of conflict behaviour. We empirically tested our functionality hypotheses with regards to both proximal and distal effects. Furthermore, we considered the question of to what extent the family, as a particularly important social setting, predisposes to perform conflict behaviour of a particular type. We hypothesized that, in interactions with family members, more dysfunctional, i.e. antisocial interaction behaviour is displayed. This argument is based on the assumption of certain types of ‘display rules’ for interpersonal behaviour. We suggested that within the family as an arena of hot emotions, a wider variety of behaviours designed to cope with stress due to conflict is facilitated. While stress due to conflict primarily stimulates feelings of anger, the performance of negative behaviour is also likely. However, while negative behaviour is subdued because of the expectation of social sanctions in response to negative behaviour, the family creates a stable environment with a higher tolerance for negative interaction. It is therefore likely that in conflict interactions with partners outside the family, both perceived own and the partners’ interaction behaviour is of higher functionality. The results reported below are based on the samples of the second and third Fribourg Family projects.

Operationalization of functional and dysfunctional interpersonal emotion regulation The information on what behaviour the partner performed and what behaviour the participants themselves performed was recorded on the basis of an 11-item list. Additionally, the intensity of each behaviour was rated on a three-point scale. The items on conflict behaviour represented measures of functional conflict behaviour, i.e. behaviours that account for the concerns and self- esteem of both partners, and dysfunctional behaviour, i.e. behaviours that account for concern of only either of the partners and that is not conducive to the self-esteem of at least one interactant (for more details, see Perrez 76 Emotion Regulation in Families with Adolescents et al., 2000). Emotion regulation outcomes were operationalized by measures of emotional state over different reports at a proximal level. Relationship quality of the parents, measured using the Relationship Assessment Scale (Hendrick, 1988) was an operationalization on the distal level.

Results The results obtained largely supported the functionality hypotheses. We consistently found effects of perceived dysfunctional emotion regulation by the partner on proximal outcomes. The more dysfunctional the regulation of the partner was perceived to be, the more negative was the well-being. Effects of self-perception measures of dysfunctional emotion regulation turned out to be a weaker predictor of proximal outcomes than self-perception measures of functional emotion regulation. The more functional I perceive my regulation, the more positive is my state of well-being. Furthermore, the concept of functionality seems to apply less neatly for the emotion regulation of adolescents if interacting with their parents. With regard to distal effects, i.e. the relationship satisfaction of the parents, we consistently found that it is primarily the perceived functional partner behaviour that predicts changes in relationship satisfaction. In contrast, interpersonal emotion regulation behaviour turns out to be associated to concurrent relationship satisfaction with regard to dysfunctional behaviour, but not with regard to functional behaviour. In other words, one may assume that low levels of relationship satisfaction promotes dysfunctional interpersonal behaviour in conflict, while changes in relationship satisfaction are associated with the degree of functional behaviour performed during conflict episodes. Generally, effect sizes are small to medium (estimates range between β = .09 and β = .31 predicting emotional state and between β = .18 and β = .36 predicting relationship satisfaction), but are comparable to findings in other studies and considered substantial when taking into account the conservative analysis strategy as well as the fact that predictions base on few observations. The analysis of the performance of dysfunctional and functional interpersonal emotion regulation revealed consistent and significant differences. Within the family, both perceived own and partners’ interaction behaviour was less functional than while interacting with partners outside the family. This was true for the sample of the second Fribourg Family project (own behaviour: cohen’s d = .42, p < .02; partners’ behaviour: cohen’s d = .45, p < .02), and was replicated on the basis of the sample of the third Fribourg Family project (own behaviour: cohen’s d = .55, p < .01; partners’ behaviour: cohen’s d = .92, p < .0001). Neither a significant gender effect nor a significant generation effect could be found. Figure 3.3 shows the functionality of interpersonal emotion regulation behaviour inside and outside of the family with functionality of behaviours displayed on a scale from 1 to 100 (m = 50; sd = 25). Perrez, Watzek, Michel, Schoebi, Wilhelm and Hänggi 77

60.00 own behaviour 60.00 own behaviour partners’ behaviour partners’ behaviour

55.00 55.00

50.00 50.00

45.00 45.00 functionality of behaviour functionality of behaviour

40.00 40.00 with family with others with family with others Second Fribourg family project Third Fribourg family project

Figure 3.3 Functionality of interpersonal emotion regulation behaviour in different settings

Conclusions • Functionality criteria regarding individual and relationship concerns as well as conduciveness for self-esteem are a useful approach to adopt when analysing interpersonal emotion regulation behaviour concerning both proximal and distal outcomes. • Findings suggest that criteria of functionality do not apply equally in the case of adolescents. • Social settings, such as the family, are of significance in determining the performance of functional and dysfunctional interpersonal behaviour with regard to emotional regulation in situations of interpersonal conflict.

5.4 Applications: the automated FASEM evaluation as a help for professionals in clinical and health psychology One goal of the automatic, computer-aided evaluation of data collected during the period of self-observation was to create a software tool which could evaluate and display the output very quickly and easily. The evaluation, therefore, should be possible with minimal computer knowledge, making it suitable, for example, for experts in the field of family consulting or family therapy. The automation makes results available already in the first meeting after the self-observation by their clients. Such information can facilitate and effectively guide professional work. The automated evaluation of the palmtop data was carried out using the spreadsheet program Excel. In a first step the user has to transmit the data from the palmtops to the PC, which happens automatically once both devices are connected. In a second step the user must start Excel and enter 78 Emotion Regulation in Families with Adolescents the set-up information about participants in the self-observation. By clicking on the start button the evaluation appears on the screen, ready to be printed out. The evaluation contains information about the emotion of the participants during the self-observation period, the reasons for these emotions (causal attributions), physical disorders, the expected controllability in everyday situations, the individual and social stress experiencing and its coping, as well as information about social support. The representation of the results is done with tables, diagrams and arrow diagrams. To get the visual representation of this data, one only has to click on the button labelled ‘Graph’. The visualization of the emotional state during the week of the family, the so-called ‘emotion score’, is similar to Figure 3.1 (see above). While diagrams make data better comprehensible, arrow diagrams offer the possibility of representing complex relations within the family in an easy manner. As one example the expected controllability should be mentioned. The expected controllability can be computed between all family members and for each family member itself. Thus, a family of four persons results in the generation of 16 values, which are related to one another and which can also be interpreted independently of the other values. The representation of these values in an arrow diagram can help to understand the relationships which exist between the values. The automated evaluation of the FASEM data can be very important for clinical work, although the method is still in its developmental phase. A variant of the system has been applied in a pilot study with 20 psychiatric outpatients of a university hospital (Gschwandtner, Hersberger and Rauchfleisch, 2000) without the recording of the other members of the family. The relevance of the analysis software will increase, once a software tool is available for construction of questionnaires on palmtop computers. With the possibility of generating questionnaires, together with the automated evaluation of self- observation data, the practitioner will be have equipment available which he can optimally adjust to the individual conditions of his clients (see the new development IZY–Builder by Law, Rozun and Rozun, 2004).

6Discussion

This chapter reviews several findings of the Fribourg Family Project. The methodology of electronic time sampling with handheld computers (FASEM-C) overcomes some critical issues that have characterized paper and pencil questionnaires. Compliance with the protocol and adherence to the protocol is much higher when using electronic diaries than when paper booklets are used (Stone, Shiffman, Schwartz, Broderick and Hufford, 2002). Problems with retrospection were reduced considerably, because the delay from event to self-report was less than three hours. This allows us to capture everyday Perrez, Watzek, Michel, Schoebi, Wilhelm and Hänggi 79 behaviour much more reliably than with long- or medium-term retrospective reports. The limitations of our study chiefly concern the sample of subjects, who are not representative of the Swiss population. They represent upper-middle-class families, who are prepared to take part in a rather time-consuming study. But once they decided to participate, the level of acceptance is surprisingly high – we encountered only a few drop-outs in all three studies. The data have nevertheless a representative character concerning the measurement points, which allows us to consider the rhythms and reaction during a normal week. The relevance of FASEM-C as an assessment procedure for the characteristics of the interpersonal regulation and of emotions in families is supported by most of the results. Inter-individual processes can be analysed with respect to both interacting partners’ subjective views. A first result, which was replicated, concerns the circadian pattern of the emotional and somatic states. These observations are innovative in this research domain, and it is not possible to have access to these phenomena using traditional retrospective assessment procedures. The empathy- related results permit a better understanding of conditions influencing empathic accuracy as a condition for the social regulation of emotions in families. The dysfunctional interpersonal coping responses of one partner in a conflict situation is a moderately good predictor of the emotional state of the other partner. The self-monitoring parameters are furthermore significantly correlated with indicators for the quality of dyadic and family functioning (for more details, see Schoebi, 2004). The replicated result that family members show significantly more dysfunctional social coping behaviour in the family setting than in other settings suggest the existence of something like ‘display rules’ for social coping in analogy to the display rules for emotion expression. These results and the results in Widmer et al. (Chapter 2 in this volume) complement one another. Widmer et al. found less open conflict in companionship-couples. One characteristic of these couples is low internal orientation of men and women, they are open to others. We can assume that the presence of others, the openness to people outside the family has a positive impact. In our study the presence of other persons are associated with functional emotional regulation and better well-being. Widmer et al. found less conflict. The two different approaches – the micro-sociological and the psychological field-approach – show with respect to this question convergent results. These first experiences with the electronic time-sampling procedure evidence its practical usfulness. The automated FASEM evaluation will be applied in different studies in a counselling and clinical context. At present, the self-monitoring system is being used in the European project ‘Family Life and Professional Work: Conflict and Synergy’. In this 80 Emotion Regulation in Families with Adolescents project the stress and coping of young, professionally active couples with pre-school children is studied in seven different countries. To sum up, our research has uncovered a new and promising method to assess real life experiences as well as interactive behaviour. Several results underline the importance of such a method, and improve our knowledge about the position of emotions in family life. 4 The Impact of Social Inequalities on Personal Health* Monica Budowski and Annette Scherpenzeel

Health care costs in Switzerland are among the highest worldwide. Revealing the contribution of social inequalities on personal health might point to less costly alternatives for public health. This paper analyzes the effects of social environment on individual health. A growing body of literature posits that social position and relationships are important for well-being. Material circumstances (deprivation, poverty, financial situation) are among the most well-examined factors regarding their impact on personal health. Recent research indicates that health also varies according to the social and material circumstances experienced during youth. Different social environments are considered to hold different risks for the individuals involved and entail different social policy measures. The Swiss Household Panel survey provides a unique data set collecting information on a variety of issues for same individuals over time. It provides a short-term (panel) and a longer-term (social origin) temporal perspective, and allows for a dynamic analysis of changes in health of individuals in Switzerland. A structural equation model identifies causal pathways. Implications for public health policies for individual health are discussed.

1 Introduction

In 2001, Switzerland had one of the highest per capita expenditures on health care in the world (US$ 2,156), and these costs are growing annually. Switzerland is also one of those countries which place considerable emphasis on supporting healthy lifestyle values in the population. But is the level of

* This study has been accomplished using data collected in the ‘Living in Switzerland Survey’ (1999–2020) conducted by the Swiss Household Panel (SHP) and Université de Neuchâtel and financed by the Swiss National Science Foundation (Grants 5004-53205/5004-57894/5004-67304/10FI11-103293) (www.swisspanel.ch). Aubrey Gilbert helped with linguistic improvements to this text.

81 82 The Impact of Social Inequalities on Personal Health spending on health care the major determinant of health? As the case of the United States shows, health care expenditures do not necessarily correlate strongly with health indicators. Indeed, poor countries, such as Cuba, Albania or Costa Rica, which have much lower spending on health care, fare comparatively well. In addition, Marmot and Wilkinson (2001: 1233) provide evidence that ‘there is little relation between average income (gross domestic product per capita) and life expectancy in rich countries’. These deliberations have led to a discussion of whether material circumstances and poverty or rather social circumstances, relative disadvantage and social inequalities are more important factors in explaining levels of health. Such questions are crucial for health politics. The question posed in this chapter is: Which causal pathways addressed by the literature contribute to the explanation of health and health changes in Switzerland – and in what ways? In particular, we are interested in the relative contributions of material, financial, and social circumstances in the case of Switzerland. This interest arises from the fact that Switzerland is one of the richest countries in the world, with a high general standard of living, good basic education, and an encompassing, obligatory health care system characterized by equal access for all. Hence, Switzerland provides a good model to contribute to the recent discussion between a group of American researchers (Lynch and Kaplan, 2000; Berkman and Kawachi, 2000b) and British researchers (Wilkinson, 1996; Marmot, 2000) about the relative importance of the social and the material environment for health. The scientific literature has shown that the pathways explaining social inequalities in health are many and complex (see, for example, Goldman, 2001). In his overview on the role of social environment for health, Kaplan provides a graph in which he depicts different levels or components of health. These incorporate pathophysiology, genetic/constitutional factors, individual risk factors, social relationships, living conditions, neighbourhoods and communities, institutions, and social and economic policies. He further asserts the need to consider social environments ‘from a dynamic, multilevel, and upstream perspective’ (Kaplan, 1999: 117). Kaplan considers lifespan to be affected by all of the factors listed above. In the social sciences, research on social environment is generally constituted by the study of cultural, regional, social, or neighbourhood membership and actual material environment, also taking into account individual characteristics; consequently, research encompasses both the physical/material dimension and the social dimension of the environment (Yen and Syme, 1999). The term ‘social environment’ is not always defined clearly and the two components – the material and social – are not always clearly distinguished (see, for example, Wilkinson et al., 1998). If the implications of social environment for personal health are addressed, this implies, at least in part, social over and beyond biological causation as explanation for inequalities in health. Approaches for examining such social causation have often focused on health-impacting Monica Budowski and Annette Scherpenzeel 83 behaviours and material standards of living, such as poverty. However, as Elstad remarks, ‘health-related behaviour can hardly be regarded as the basic cause of heath inequalities’, because it is ‘seldom chosen “freely”, but is heavily influenced by social status and cultural milieus’ (Elstad, 1998: 39). In line with the literature reviewed, we have grouped the explanations of social causation into five pathways: (1) material and financial circumstances; (2) social and cultural circumstances; (3) self-assessment of personal situation and psychological mechanisms of coping; (4) time-lagged and/or cumulated effects of disadvantaged or advantaged circumstances over time; and (5) the degree of social inequality. (1) The material dimension of the social environment refers to material living conditions such as deprivation, poverty, wealth, or access to medical and educational services (the latter services are assumedly equal for all in Switzerland). The relationship between material and the social dimensions (e.g. socioeconomic status, race, gender) has been extensively researched (Moss, 2002; Bartley et al., 1998; Bartley et al., 1999; Gordon et al., 2000; Gravelle et al., 2002; Blaxter, 1991; Fox, 1989; Macran et al., 1996), producing the rather consistent finding that poorer living conditions result in poorer health. The relative material situation has been most often operationalized by level of income. Building on work of Townsend (1979) it has also been characterized by the level of deprivation (goods or activities inaccessible due to lack of finances (for Switzerland Suter and Paris, 2002)), or by a combination of both deprivation and income poverty (for Switzerland Budowski etal., 2002). (2) Social and cultural circumstances may be described by the structure of role-specific interactions among people as they are experienced, for example, in family relations, in living arrangements, neighbourhoods, at work, or in society at large (Journal of Marriage and the Family, 2000; Berkman and Kawachi, 2000a). For example, holding a married civil status has been shown to be beneficial for well-being when compared to those who are single, widowed, or divorced, states which all seem to affect health negatively, although these effects do diminish over time (Wallerstein and Kelly, 1980; Simpson, 1998). Over the past few decades, home living situations have become increasingly complex, providing a greater number of circumstances structuring everyday life. Instances of cohabitation or joint/step families are becoming much more common in society. Therefore, there is some empirical evidence that the characterizing power of civil status may have diminished (Willitts et al., 2004). For such reasons household composition may be an equivalent or better indicator than civil status for such comparative purposes. In Switzerland, a majority of young couples live together but do not marry until they are planning to start a family. Social relations also include social networks that involve interactions with neighbours, friends, acquaintances and co-workers, and participation within the community and other groups (Lynch et al., 2000; Evans et al., 1994). The structure of social relationships refers to aspects of formal relations and 84 The Impact of Social Inequalities on Personal Health social networks: it may cover, for example, such aspects as network size, composition of network, accessibility of network members, availability of specific kin, or contact frequency. They are considered to affect the flow of resources and to represent the quantity and diversity of potential support sources (Suter et al., 2003). Numerous studies do conclude that social support is beneficial for health (House et al., 1988a, 1988b). Being embedded in a social network is considered to lead to healthier behaviour, because the social network encourages positive health habits (Antonucci, 1990; Umberson, 1987), particularly in the presence of small children. However, studies have also demonstrated that because of stressful interactions, social environments, networks, and (too much) support may also lead to poorer levels of health (e.g. O’Brien Caughy et al., 2003; Rook, 1992). Such contrary findings are often attributable to variations in the conceptualization of social environment, its dimensions and components, the operational definition of health (whether it entails having a particular illness or disease, whether it is mental, physical, or self-rated health), how social support and the structure of social networks are delineated, and at which level – the micro, meso or macro – the relationships are investigated (Seeman and Crimmins, 2001; Melchior et al., 2003; Bolin et al., 2003; Helweg-Larsen et al., 2003). (3) As social beings, individuals are ‘structured’ by their environment. Their resulting interpretations and assessments of their situations are important elements in understanding how they perceive and construct their sociostruc- tural setting and, consequently, how it impacts upon their health (e.g. Thoits, 1982, 1995; House et al., 1988b; Berkman and Glass, 2000). Satisfaction with social ties, for example, reflects an assessment of their availability and perception of functioning. Social ties have been considered to be important in maintaining senses of social integration and belonging, support and identity (Wellmann et al., 1997). One’s material and financial situation also require an interpretation and assessment: it need not only be poverty that impacts on health. The opportunity to finance lifestyles by means of credit has become a reality in many richer countries (Bauman, 1998). Doing so may work in the short term but will eventually turn into a large strain when income sources can no longer make ends meet (Streuli, 2003). Self-assessment of one’s financial and material situation must therefore be considered as part of or as mediating the actual level of financial and material opportunities. (4) A growing body of research documents evidence for delayed (i.e., time- lagged) effects of social environments. This means that effects may only become visible in the long run (McDonough et al., 2001; Singer et al., 1999; Schoon et al., 2003; Case et al., 2001: 1; Hertzman, 1999; Pensola et al., 2003). As Singer and Ryff state (1999: 96), most studies do not evaluate processes over the long term, or cumulative experience for the prediction of health and well-being (in their case at old age). The same authors summarize a small, but increasing number of studies providing evidence for the impact of cumulative and long-term effects of poverty or elevated permanent income as well Monica Budowski and Annette Scherpenzeel 85 as of the cumulative effects of psychosocial experiences on health. An accumulation of stress over time increases health disadvantage. Singer and Ryff also point out that it is important to give ‘equal consideration to positive aspects of people’s lives and particularly to cumulative advantage’ (Singer and Ryff, 1999: 97). Consequently, they hypothesize that health outcomes depend upon both cumulative advantages and disadvantages across multiple life domains. (Berkman and Kawachi, 2000a; Janzen and Muhajarine, 2003) also stress the idea that a developmental and life-course perspective could yield important new insights. (5) The last causal pathway deals with the effect of either political factors (e.g. the Cuban health care system) or income inequality (Daniels et al., 2000; Marmot, 2000; Wilkinson, 1996; Ross et al., 2000). Taking research on these factors into account, the questions we will consider are: (1) To what extent does social environment (the material and social dimension) impact on health? and (2) What role does cumulative disadvantage, i.e. experiences beginning in childhood, play when considering health? Is it the actual social environment, the past experiences or both that impact on health? If the actual social environment impacts considerably, then are short-term changes in household circumstances, social networks, or material living conditions reflected in a change in health? (3) Are health differences more dependent upon the material or the social dimension of the environment? (4) Do personal interpretations and assessments of the actual situation impact upon health? To date, much research regarding the impacts of social position, social relationships, household type, and material disadvantage on health is based upon cross-sectional data, and conclusions regarding the direction of cause and effect are made on the basis of comparisons between groups with similar social characteristics (e.g. married versus single individuals, single parents versus two-parent families, men versus women, and so on). If the impact of social environment on health is really to be understood, changes in social environment and changes in health have to be measured at the same points in time and for the same people. Only in this way can the impact of social environment as a causative factor affecting individual health be empirically evidenced; changes in the components of the social environment should be reflected in changes in individual health in the short or the long run. For the first time in Switzerland longitudinal data are available that address various issues (health, social origin, household quality and characteristics, labour force, social position, etc.) in sufficient detail to allow for the examination of the short-term impact of social environment on health. As previous studies have shown, results from other countries cannot simply be applied to the Swiss context because the structural situation of population groups may vary considerably between countries (Budowski et al., 1999); comparative welfare state research indicates that each social context provides unique 86 The Impact of Social Inequalities on Personal Health advantages and disadvantages when compared with others (e.g. Esping- Andersen, 1990, 1996). We conceptualize three dimensions of health (their operational definitions are given in the next section): general health (everyday health), psychological health (psychological malaise or well-being) and health- related events (events such as accidents, long illnesses, number of visits to the doctor, etc.).1 Based on the literature reviewed, we expect to observe the following general associations: (1) Time: We expect a delayed impact of social disadvantage in youth (both material and social) on actual health. According to the literature, all three conceptualized dimensions of health should be affected. (2) Household situation: If social environments play a role (Berkman and Kawachi, 2000a: 7), interactions within the social environment of one-person and single-parent households are limited simply because of the fact that only one adult is present. Indeed, such situations need have no impact on health, as they may promote better health in certain environments or lead to poorer health in others.2 It goes without saying that interactions within the household may be agreeable, stressful or both. Therefore we do not necessarily expect the structure of the social dimension of the household to be associated with health, but rather the subjective perception of it – whether people are satisfied with their domestic situation. Depending upon this assessment, the relative structural limitations within one-adult households lead us to expect somewhat poorer health, particularly psychological health, in comparison with those household types that are constituted by at least two adults. Furthermore, most of the changes from households with more than one adult to one-adult households are often accompanied by rather stressful life events (divorce, death, children leaving the home) and fewer are based on choice. In addition, practical support is limited within one-adult households, leading more quickly to household external (or professional) help. Therefore, this household composition might have a higher level of changes in health-related events than other types, and changes into one-adult households would entail changes in that health dimension. The effect of living alone or in a single-parent household should be stronger still in combination with a low degree of satisfaction with one’s living arrangement. Household composition might also play a role for health when combined with work. (3) Social position: We expect social position to impact on health, in particular on general health; the higher the social position, the better general health is expected to be (e.g. Sacker et al., 2001); lower social positions within the labour market should be reflected in poorer ratings across all defined dimensions of health (Marmot and Wilkinson, 2000). If an individual changes his/her social position, we expect changes in health to follow. Such changes can take longer than a year to be realized, but to date, we do not know how long it will take. We expect the effects of social position of those individuals Monica Budowski and Annette Scherpenzeel 87 removed from the labour market to affect general health and health events for the retired, and psychological health for the unemployed. We expect a greater impact on psychological health for individuals who experience a change in position that allows them to enter into the labour market (housekeepers, students, and the unemployed). According to Arber and Khlat (2002), social position ‘may be a less discriminating indicator of health inequalities for women than men because of women’s more fragmented employment career’ (Arber and Khlat, 2002: 643). (4) The potential availability of social support and the feelings of identity and belonging they provide might buffer more disadvantaged household circumstances, reducing the level of vulnerability to illness; by the same token, low levels of social support might aggravate such situations. On the whole, we expect those people who feel well-integrated and embedded in a social network to assess their situation in a more positive manner. Regarding its effects on health, we expect the influence of social support to be associated with psychological health and possibly health-related events as social support cannot rule out physical illness but can help people to manage and/ or cope with it. (5) Material living conditions: We expect material living conditions to impact strongly upon health; experiencing a constrained financial situation and disadvantaged material living conditions, combined with a negative assessment of these conditions, is expected to lead to poorer health. (6) Socio-demographic characteristics: We expect to find health differences between men and women. Although women live longer, research shows that they experience higher levels of illness than men, particularly in the area of psychological health. However, there are longitudinal studies which fail to confirm this difference between men and women (Walters et al., 2002). In general, health, particularly general health, declines with increasing age, but health-related events also become more frequent. In summary, we expect that greater long-term disadvantage, in addition to currently disadvantaged situations (lower educational level, lower social position), a structurally limited household situation, a less favourable interpretation of one’s situation, less social support, and greater material disadvantage, will decrease well-being and increase illness. We expect changes in social and material circumstances to contribute to the explanation of the changes in health.

2 Methods

2.1 Data The data used come from the first three waves of the Swiss Household Panel, carried out in 1999, 2000 and 2001. The Swiss Household Panel was founded in 1998, supported by the Swiss National Foundation for Scientific Research, 88 The Impact of Social Inequalities on Personal Health the University of Neuchâtel, and the Swiss Federal Statistical Office. The goal of this large-scale panel study was to conduct annual interviews with about 8,000 people in 5,000 households in order to assess changes in their living conditions, attitudes and beliefs. The survey covers a broad range of social fields and topics and collects both objective data (resources, social position, participation, etc.) and subjective data (satisfaction, values, evaluation, etc.). The interviews were carried out by telephone (CATI – computer-assisted telephone interviewing). The complete longitudinal sample (individuals interviewed in all three waves) comprises 5,429 individuals. For our analysis, only individuals who had valid data for all variables in the model might be used. Individuals with missing data were excluded from the analysis, resulting in a final sample number of 4,414 individuals.

2.2 Effects of social inequalities on personal health: the model In this study we have formulated predictions with regard to changes in health and the effects of changes in living condition on health over time. We have then defined a dynamic structural equation model to test these predictions. Structural equation models may be seen as extended regression equations that allow measurement errors in the independent as well as the dependent variables and correlations between the error components. They can also be viewed as factor analyses that permit analysis of the direct and the indirect effects between factors. Furthermore, they can include multiple indicators and latent variables. In short, these types of models offer all possibilities to analyze the effects of change through time. In Figure 4.1, the model is depicted in general and simplified forms. Three health dimensions were constructed from the health measures in the panel survey: one dimension we called general health, a second health-related events and a third psychological health. Each of the dimensions was, as a latent factor in the model (see subsection 2.3 on Measuring Health and Social Environment), constructed from three different measures. In the model, change over time in these health dimensions is indicated. At each of the three points in time, the health dimensions are affected by the actual social environmental conditions: household situation, socioeconomic position (in combination with attachment or lack thereof to the labour force), material living conditions and perception of financial situation, social support, and the subjective interpretation of the living arrangement, operationalized as the degree of satisfaction with it. The conditions are allowed to change over time. In addition to these conditions of the social environment that are present in every wave, disadvantages in youth are specified to have a delayed impact on health in the first wave, also indirectly influencing health in subsequent waves through the stability effects of health over the waves. (Note the arrows from first wave health to second wave health and from there to third wave health in Figure 4.1.) Similarly, 89

Baseline 1999 youth: financial situation hh- type youth: household general type health hh- satis- sex faction

social health age support events

nationality social position psychological health financial situation

2000

hh- type general health hh- satis- faction

social health support events

social position psychological health financial situation

2001

hh- type general health hh- satis- faction

social health support events

social position psychological health financial situation

Figure 4.1 General three-wave causal model of health 90 The Impact of Social Inequalities on Personal Health the demographic factors – sex, age, and nationality – were included as control variables affecting first wave health directly, and health in the subsequent waves indirectly, through the effect of first wave health on second wave health and so on. Health in the first wave is explained by these demographic variables and the conditions of the social environment in the first wave only. The estimates of the effects of the conditions in this wave could, in principle, also be obtained with a normal regression or variance analysis (except for the use of the latent factors and the indirect effects), using cross-sectional data, and hence constitute a kind of ‘baseline’ estimation of static effects: how strongly do each of these conditions of the social environment and their perception relate to health status? Health in the second and third waves is explained by the conditions of the social environment in the second and third waves, but also by the previous health status. Furthermore, the actual conditions of a given wave are also influenced by the respective conditions in the previous wave. Consequently, the stable variance of both the health factors and the conditions is removed, and the effects of the conditions at the second and third waves represent the effects of changes in the conditions of the social environment on changes in health: how much does health change when, for example, there is a change in household composition? This basic model was varied somewhat during the estimation process: non-significant parameters were, in most cases, deleted from the model and, likewise, variables with no significant effects at all were left out. All model estimations were undertaken using the AMOS program for linear structural equations (Arbuckle, 1997), a detailed description of the estimation procedure can be found in the Technical Appendix to this chapter. The model in Figure 4.1 is a simplified, illustrative version of the model that, in reality, has to be specified for analysis. To make the model complete and estimable, unique components, correlations and restrictions have to be added (see Technical Appendix). Moreover, some measures were very highly correlated and could be viewed as different dimensions of a single concept. In such cases, we specified a latent factor with a measurement model. Latent factors make the structural part of the model more parsimo- nious and provide more possibilities for the correction of measurement error. By defining latent factors, we were able to reduce the original total number of approximately 90 observed variables analyzed to just 20 factors in the final model.

2.3 Measuring health and social environment Three latent factors were constructed from the health measures in the panel survey: the latent factor, general health, was measured by: medication needs, impediments to everyday activities and a subjective health evaluation. The second latent factor, health-related events, was measured by: number of days Monica Budowski and Annette Scherpenzeel 91 affected by health problems, number of doctor consultations in the first wave in addition to any illness or accident between two waves. The third latent factor, psychological health, was measured by: positive feelings, negative feelings, and the number of psychosomatic problems. The factor, disadvantage during youth, was, on a theoretical basis, assumed to consist of the level of education and the social status of one’s parents (according to the Goldthorpe classification), the nationality of one’s parents (Swiss or foreign), and the existence of any financial problems during youth. Furthermore, living arrangement at age 15 was specified, i.e. living in a two-parent family or in another situation. The social position of a person was measured by a modified Goldthorpe classification scheme representing occupational class, supplemented with four non-working classes, as follows: (1) white collar with educational level covering apprenticeship and higher; (2) white collar with educational level lower than apprenticeship; (3) petty bourgeoisie; (4) farmer; (5) skilled manual laborer; (6) unskilled manual laborer; (7) unemployed; (8) student; (9) housekeeper; (10) retired. Theoret- ical considerations about the development and sexual segregation of the labour market (the increase in the size of the service sector and women’s non-skilled white-collar jobs) as argued by Daniel Oesch led to the modification of the scheme (Oesch, 2003). The household type of the respondents was categorized by combining the data on living alone or as a couple with the data on having children or not and the age of any children, as follows: (1) couple, no children; (2) couple with children, youngest child older than 12; (3) couple with children, youngest child younger than 13; (4) one-person household; (5) lone parent household. Perception of the living arrangement was operationalized by satisfaction with household situation, a uni-dimensional observed variable. Social support is a latent factor constructed from two measures: the sum of all the practical support a person reports to get from a variety of networks (partner, children, relatives, friends, colleagues), measured on a scale from 0 to 10, and the sum of all emotional support one gets from the same networks. The latent factor, financial situation, was assumed to consist of both subjective assessments and objective situation with variables such as: problems paying bills (yes/no, self-reported); perceived manageability of financial situation (0–10, self-reported degree); and degree of deprivation (number of goods not possessed for financial reasons). Since these three measures were highly correlated, a latent factor was constructed from them in the model.

3 Results

3.1 Inequalities in health status at first interview Shown in Table 4.1 are the standardized effects as estimated in the model. These effect estimates have an interpretation comparable to that of the 92 0.07 0.21 − − w1 w1–2 w2–3w2–3 w1 w1–2 2 w2–3 2 w1–2

2 0.09 0.09 0.09 0.09 0.06 0.06 0.31 0.31 − − − dized regression weights : General health Health events Psychological health

12 13 ≤ ≥

Effects standar on health:

Couple no children Couple no children Couple, child Ref. Ref. Ref. Ref. Ref. Ref. Ref. Ref. Ref. Couple, child Oneperson Oneperson hh parent Lone Petty Bourgeois education high collar, White education low collar, White Petty workers agricultural and Farmers Ref. Ref. Ref. Ref. Ref. Ref. Ref. Ref. Ref. Table 4.1 Table w1 Base line Change Base line Change Base line Change Sex Nationality in youth disadvantage Financial Household type in youth Household type Age Goldthorpe and working status 93 The 3 2 0.08 − 0.29 − rs is the sum of these squares in squares these of sum the is rs w1: wave 1, 1999; w2: wave 2, 2000; w3: wave 3, 2001. 2 0.07 0.24 0.12 0.07 0.13 0.07 0.07 0.09 − − 4414. = 0.11 0.27 0.11 0.14 0.11 0.27 0.11 0.30 0.09

0.05 0.05 − − 22 0 8 1 2 21

0.08 3 .001 level are shown. ref.: Reference category; category; Reference ref.: shown. are level .001 =

Only effects significant at the p 1 : Swiss Household Panel, individual data, 1999, 2000, 2001; n 2001; 2000, 1999, data, individual Panel, Household Swiss : : Skilled manual workers manual Skilled manual workers Unskilled Unemployed Student Houskeeping Retired Satisfaction with household composition Total amount of variance explained by the factors (%) Social support situation Financial percentage of variance explained by one factor is the square of its standardized effect estimate. The total amount of all facto all of amount total The estimate. effect standardized its of square the is factor one by explained variance of percentage column. each Source Notes 94 The Impact of Social Inequalities on Personal Health

(standardized) B coefficients in a normal regression analysis: if we examine the effects of the two demographic variables, for example, it is clear that age has a much larger effect (−0.31) on general health than sex (−0.06). The standardized effect estimate for age indicates that if age increases by one standard deviation, general health decreases by 0.31 standard deviations, indicating that increasing age negatively affects general health. The small estimate for sex indicates that men have slightly better general health. The effect of sex is stronger on psychological health (0.21), suggesting that men have significantly better psychological health than women. Neither age nor sex has any significant effect on health-related events and age does not seem to affect psychological health either. Disadvantages during youth seem to affect all three health dimensions equally (effects around 0.09). As described above, the factor, disadvantages during youth, was originally assumed to consist of level of education and social status of one’s parents (according to the Goldthorpe classification), nationality of one’s parents (Swiss or foreign), the existence of any financial problems during youth and household type at age 15. However, tests showed that the single variable, financial problems during youth, was more strongly related to health than the cumulative index or any of the other indicators alone. There- fore, this variable was used as the sole indicator of disadvantages during youth. Household type at age 15 proved not to be a significant indicator. The most important social environmental condition seems to be the actual financial situation and its perception, with standardized effects between 0.24 and 0.30 on the three health dimensions in the first wave, meaning that people in poorer financial situations have poorer general and psychological health, and experience more health-related events. One of the eight categories of the social position factor has significant effects on first wave health: the category, retired negatively, affects baseline general health somewhat (effect −0.09) and increases the number of health-related events in wave 1 (effect 0.13). Another rather important factor is satisfaction with household situation, which affects general health (0.11) and psychological health (0.27). An effect of household situation on satisfaction with household situation was also specified and proved to be rather strong (standardized effect −0.30, not shown in Table 4.1). In fact, household situation turned out to have only indirect significant effects on health, mediated through satisfaction, and only for the category one-person household (not shown). The factor, social support, did not demonstrate any effects on health. However, the factor was highly interrelated with some of the other causal variables in the model: first, a hypothesized effect of the one-person household type on social support was specified and was indeed found to be rather strong (−0.31, not shown in Table 4.1), indicating that people living alone experience less social support than others. Secondly, the expectation that social support would influence satisfaction with household composition was also Monica Budowski and Annette Scherpenzeel 95 confirmed, although it was not a very strong effect: people who report higher levels of social support in the first wave are somewhat more satisfied with the way they live in their household (0.16). These were the static effects found at the first interview (wave 1 in 1999). We now turn to the effects of changes, which we can find in the second and third waves.

3.2 Do changes in circumstance explain changes in health? The dynamics of health at second and third interviews First, we consider changes in health over the waves (over the second and third interviews). The stability of general health is considerable, but some changes can be observed in Table 4.2 the stability estimate from the first wave to the second is 0.87, and from the second wave to the third, it is

Table 4.2 Stability of the variables in the model

Stability estimate w1 − w2 Stability estimate (% stable variance)* w2 − w3 (% stable variance)

General health 0.87 (76) 0.76 (58) Health-related events 0.59 (35) 0.53 (28) Psychological health 0.80 (64) 0.76 (58) Household type Couple no children ref. ref. Couple, child ≤ 12 0.92 (85) 0.90 (81) Couple, child ≥ 13 0.90 (81) 0.88 (77) Oneperson 0.93 (85) 0.91 (83) Lone parent hh 0.90 (81) 0.88 (77) Goldthorpe and working status: White collar, high education ref. ref. White collar, low education 0.67 (45) 0.71 (50) Petty bourgeois 0.51 (26) 0.38 (14) Farmers and agricultural workers 0.78 (61) 0.79 (62) Skilled manual workers 0.71 (50) 0.65 (42) Unskilled manual workers 0.63 (40) 0.54 (29) Unemployed 0.28 (8) 0.24 (6) Student 0.69 (48) 0.69 (48) Houskeeping 0.72 (53) 0.72 (52) Retired 0.87 (76) 0.90 (81) Social support 0.54 (29) 0.58 (34) Satisfaction with household composition 0.49 (24) 0.50 (25) Financial situation 0.98 (96) 0.98 (96)

Remarks: * The percentage of stable variance is the square of the stability estimate. Source: Swiss Household Panel, individual data, 1999, 2000, 2001; n = 4414. 96 The Impact of Social Inequalities on Personal Health

0.76. This means that there is 76 per cent stable variance (the square of the stability estimate, see note in Table 4.2) from the first wave to the second, and 58 per cent from the second wave to the third. The complement of the stable variance is the amount of change, thus general health changes 24 per cent between the first wave and the second wave, and 42 per cent from the second wave to the third wave. Psychological health also changes somewhat over the one-year periods: the stable variance from the first to the second wave is 64 per cent, and from the second to the third wave, 58 per cent, which yields 36 per cent and 42 per cent change, respectively. The least stable factor is health-related events, which changes 65 per cent and 72 per cent from the first to the second and the second to the third waves, respectively. Can these changes in health be explained by changes in the social and material environment and the perception of the situation? They do not seem to be explained very well. First of all, two of the factors change little over the waves: the one-person household category, for example, has stability estimates of 0.93 and 0.91 (Table 4.2), and even financial situation has stability estimates of 0.98 for both intervals. Moreover, the small change that we do find accounts for little of the change in the health dimensions. Changing from being a couple to being a one-person household causes some small change in the degree of satisfaction with household situation (effect −0.13 in both waves 2 and 3, not shown) and through that, a small change in psychological health. Changes in financial situation and its perception have a small effect upon changes in psychological health in the second wave (effect −0.08), whereas in the third wave they do not change psychological health, but do have a stronger effect on general health and health-related events (effects around −0.07 and 0.12). In general, we can say that changes in the financial situation of households can have some small effects on any of the dimensions of health, but which of the three dimensions is affected most might be a matter of coincidence. The relationship between social support and satisfaction with household composition shows that when social support increases from one wave to another, satisfaction with the household also increases slightly (effect 0.10 in both wave 2 and wave 3, not shown). Changes in social support have a small significant effect on changes on general health in wave 1. There is, however, one rather interesting effect of change: while the social position, unemployment, has no effect on the health status quo in wave 1, changes in unemployment status do have effects on changes in health. In the second wave, we see that a change in this category (that is: becoming unemployed) causes a small increase (effect −0.08) in health-related events. In addition, transitioning into the category retired slightly increases the incidence of health-related events (effect 0.07 in wave 2, 0.09 in wave 3). Satisfaction with the household situation changes noticeably (the percentage stable variance is 25 per cent) between waves and these changes can explain Monica Budowski and Annette Scherpenzeel 97 some of the change in psychological health (effects 0.11 in wave 2 and 0.14 in wave 3).

4 Summary of the results and discussion

In the model we constructed for this study, we aimed to examine the impact of social environment as measured by social and material circumstances and their perception on health. We also examined social support and satisfaction with the household situation as indicators for personal assessment of one’s situation. Our main focus was to identify whether changes in the social environment resulted in changes in health. Age, sex and nationality were taken into consideration and we exploited the possibility provided by the SHP data to take into account a longer-term, delayed effect of disadvantage during youth on health, as various information was gathered regarding social origin. Our contribution is unique for Switzerland in that the SHP data allow for the first time the formulation of predictions with regard to changes in health over time (with the exception of Zimmermann and Burton-Jeangros, 2004). We defined a dynamic model to test these predictions. As was said previously, substantive differences in health are often found between different groups of people in cross-sectional studies, the groups being defined on the basis of sex, age, living situation, socioeconomic status, and so on. However, cross-sectional studies, for example, consider differences in material living conditions and health between people and do not indicate whether change in a living condition will cause a change in the level of health. Finding differences between groups in a cross-sectional study cannot prove a causal relationship. Other, unknown variables might underlie the presumed relationships, or several variables might be confounded and might not be separable by comparing means between groups. To reveal the causality between health and other variables, change must be observed. If a change in a living condition is often accompanied by a change in health, the causal argument is much stronger. The uniqueness of a panel study like the one we use here is the possibility it provides to measure changes in living conditions and health, for a set group of people. In addition, by using the type of analysis we use, changes may be evaluated on the individual level by analyzing correlations over time instead of changes in aggregated distributions or group means. Coming back to the question set out at the start of this chapter, the answers we can give are the following: (1) To what extent does the social environment impact upon health? Our results indicate a causal relationship between age and general health. Sex has little or no effect on general health or health-related events, in line with recent longitudinal results indicating this finding (Walters et al., 2002). However, sex does have an effect upon psychological health, with men faring better than women. Surprisingly, the actual social position has no impact upon health except for the unemployed and retired. Interestingly, 98 The Impact of Social Inequalities on Personal Health being unemployed as such has no effect upon any of the three health dimensions, but becoming unemployed does. Indeed, a change from employment to unemployment increases health-related events, but has no impact on psychological health. A retired status has negative effects upon general health and health-related events, but not upon psychological health. Being or becoming retired diminishes general health and increases health-related events a little bit, but does not have any effect on psychological health. Household composition has no direct effect upon health and the only effect existing is an indirect one for one-person households for which satisfaction with the household composition is the factor that actually impacts general and psychological health. We observe the same pattern for social support, a factor which has nearly no direct effect on the three dimensions of health, but which, based on whether one lives alone or not, might have a small indirect effect on health, again through its effect on satisfaction with household composition. An increase in social support showed a small positive impact on general health in wave 1. The largest part of the variance in the three health dimensions is explained by material living conditions and their perception and by satisfaction (an indicator for the perception of the household situation). In summary, changes in the social dimension of social environment, as operationally defined in this chapter, explain little of the health changes observed, even though they take into account age, sex, nationality, household situation, place within the life cycle, location within the social structure, whether the individual is attached to the labour force or not, and social support. If the social dimension impacts on health, it does so mainly in the realm of psychological health. The material dimension of the social environment and its perception, by contrast, has a rather strong effect on health and impacts upon both general health and health-related events, much more forcefully than any effects noted from the social dimension. (2) What role does cumulative disadvantage play in health? Indeed, it appears as if social environment has a weak, delayed direct effect on health. Delayed effects of disadvantage play a role in all three dimensions of health. This confirms recent research that takes time into account (e.g. McDonough and Berglund, 2001; Singer and Ryff, 1999). However, interestingly, contrary to much research, we did not witness an impact of the social status of one’s parents (i.e. the social dimension measured by level of education, social status and nationality of one’s parents), nor of the living arrangement at age 15, but of the material dimension of social environment, measured by experience of financial problems during youth. (3) Are health differences more dependent upon the material or the social dimension of the environment? To summarize our findings, the analyses suggest that, as it is measured, including classical measures of stratification and everyday interactions within the household as well as in a broader social network, the social dimension of environment does not explain much of one’s health or Monica Budowski and Annette Scherpenzeel 99 health changes. The social environment indirectly affects health through people’s subjective assessment of their situation. Individuals in one-person households who are less satisfied with living alone also have poorer health, and a change in their household situation seems to entail a change in health. By contrast, the material dimension and its perception proves to be more important, impacting on health both with delayed and immediate effects. Reasons why we did not find greater explanatory power in the social dimension, might be: (i) Our measures are too crude to allow such an observation, and they require additional refinement. Although this may be true, prior (more global) measures did not yield substantially different results from the ones presented. (ii) Perhaps not only social position (hierarchy in the labour market and whether one is retired, in school or at home), but variables indicating different degrees of work involvement and domestic responsibilities should be incorporated. This was achieved, insofar as the household composition included the age of youngest child and the social position indicated whether someone was working or not. (iii) Another explanation might be that the factors affecting health are different or inter- act differently for different groups, such as for men and for women. By applying only one model for both groups, changes happening to one group and not to the other might not become sufficiently clear. However, we assured that this was not the case, by carrying out multi-group analyses, for example, according to sex. The results indicate that a single model for the case of men and women is sufficient. (iv) A further explanation for the weak impact of social environment on health is that the implications of short- term changes are not observed immediately, but are delayed for some time. Hence, a change from a white-collar to a blue-collar job might impact on health only at a later point in time. (v) An alternative account might be that we observed too short a time-span in which too few important changes or events occurred in the lives of the study group of people. It is possible that we observed changes in factors that are less important for health, so the changes we observe are small and the factors have little impact on health. Each additional year the panel is carried out provides more chances for such events to occur. Indeed, some of the conditions examined remain very stable. Others, however, are not (for example, certain social positions and social support are unstable), but these changes do not explain many of the changes in health either. Despite the fact that the health dimensions did change quite a lot over the short time-span observed, changes remain partly unexplained. What the model does indicate is, on the one hand, the importance of the subjective assessment and interpretation of one’s situation for psychological health. Results reveal that satisfaction with household composition is the most important variable for psychological health. On the other hand, the model suggests even greater importance of material circumstances and their assessment which influence health-related events and general health. The lack of 100 The Impact of Social Inequalities on Personal Health effect of social position at work may be due to the fact that the subjective assessment and the material situation were measured separately. Hence social position is ‘net’ of all other strongly correlating effects (such as income). Although initial correlations exist according to those found in the literature, our analyses indicate that it is important to control for baseline health and to interpret changes against this initial state, as baseline health appears to level out many effects (Adams et al., 2002 came to the same conclusion, albeit in the case of an older population). We conclude that it is necessary to continue with this type of analysis in order to achieve an understanding of the causality of the factors impinging upon health. Longitudinal data over longer periods of time are key for a precise understanding of the factors affecting health. Regarding the implications for social policy, the model suggests that measures the reduction of poverty and being able to manage the actual financial situation appear to be most important factors. This entails family politics and the transfer of practical knowledge. The model suggests that, despite the material environment being salient as a pathway to health, such inequalities require perception and assessment in order to have a strong impact. Measures supporting peoples’ assessment of their situations and their coping capacity (empowerment) might provide a less-costly direction to improve public health over medicalization of everyday life.

Notes

1. Details are given below in the section on Methods. 2. A literature review of the past two decades finds that people who are isolated experience an increased risk of mortality (Berkman, 1995). Lone mothers are in a situation which requires them to bear a greater burden of family maintenance, often on their own, and it has been evidenced that their health is disadvantaged (Lahelma etal., 2002).

Technical appendix: model estimation and model fit

Model estimation All model estimations were undertaken using the AMOS program for linear structural equations (Arbuckle, 1997), using the Maximum Likelihood estimation. All variables are standardized except the disturbance variables (which are not normally standardized). We assume, as is the convention, that the correlations between the disturbance variables and the explanatory variables in each equation and across equations is zero. The social and material living conditions at the time of the first wave are assumed to be exogenous variables, correlated with each other and having no unique component. At the time of the second wave they are considered endogenous, determined by the same conditions at the first wave (stability) and having unique components. A few social living conditions, household composition and socioeconomic status, were nominal in nature: the categories of these variable were recoded into dummy variables before they were used in the structural equation model, the value 1 indicating membership in the dummy category, the value 0 indicating lack of membership in the category. Checks were made for possible collinearities between the explanatory Monica Budowski and Annette Scherpenzeel 101 variables. As is logically assumed, the retired category of the social position variable was correlated with age and this relationship was specified in the model. Other checks were made for correlations between sex and the category, housekeeper, of the same variable and between financial situation and the category, unemployed, but these correlations were moderately low (all lower than 0.2). The expected correlations of social support with household type and satisfaction with household composition showed some low to moderate correlations. These relationships were therefore included in the model. Correlated errors are introduced in the model between identical measures of a latent factor in each of the three waves. Such correlations often improve the fit of measurement models, which is, as such, no justification for introducing them. How- ever, in this case there is theoretical justification for their use: they model the systematic error that may exist in repeated measures of the same variable (artificial correlation caused by use of the same method, measure, formulation, scale, etc.) Hence, we do not introduce correlations between all error components in the model, but only between the error components of identical measures from the three waves.

Model fit Our model does not seem to provide a good fit for explaining the effects of social inequalities on personal health and changes in social and material circumstances on changes in health (chi2 = 12373.58, df = 1349). However, this apparently poor fit can be partly attributed to the power of the chi-square test for an n = 4414. There- fore, we focus our primary concern on comparison of fit among potential models rather than absolute fit. If we compare the dynamic health model with several alternatives such as static models which do not allow for change over time, we clearly see that our model provides a better description of the available data and variables. It is, of course, still possible that some important variables were omitted (see the discussion in section 4). Part II The Swiss Labour Market 5 Long-Term Dynamics of Skill Demand in Switzerland, 1950–2000 Stefan Sacchi, Alexander Salvisberg and Marlis Buchmann

Our contribution investigates long-term shifts in skill demand in Switzerland with respect to both the level of formal skills and the type of occupational certificate. The former represents the most obvious ‘vertical’ dimension of changing job requirements, whereas the latter, no less important in a highly segmented labour market, reflects the ‘horizontal’ differentiation along occupational lines. We argue that two sources are responsible for any change in the aggregate composition of skill demand: The shifting sectoral composition of the economy and the intra-sectoral adaptations related to new technologies and the changing work organization. Against this background, we assess the relative impact of inter- and intra-sectoral change on the long-term trends in skill demand. We apply novel indicators of technological change within sectors to account for the latter. The empirical analyses are based on a representative random sample of job advertisements published between 1950 and the year 2000 in the German-speaking part of Switzerland.

Introduction

For various reasons, labour demand is a key indicator of societal development. At the macroeconomic level, prosperity and unemployment are decisively determined by the capability of firms to create new jobs and secure existing ones. At the individual level, employment opportunities and income chances are closely related to the volume of vacancies. However, a purely quantitative perspective of labour demand is shortsighted. Although quantitative swings in labour demand have serious macroeconomic and individual consequences, the impact of long-term, qualitative shifts in the employers’ demand for skills is scarcely less relevant. Typically, when filling vacancies, firms only take into consideration those applicants who are equipped with a well-defined set of skills and competencies. Despite a sufficient number of vacancies, the pool of workers with insufficient or, no less serious, the ‘wrong’ skills may be responsible for unemployment and low salaries. Employment opportunities

105 106 Long-Term Dynamics of Skill Demand in Switzerland and occupational career chances are therefore closely linked to the crucial question of whether firms indeed demand, in the medium and long run, those skills and competencies for which workers have been trained. With regard to both training jobs and entry jobs for young workers, trends in labour demand are key factors in the successful integration of younger generations into the labour market. Despite its great societal importance, the long-term overall development of skill demand has seldom been the topic of empirical studies, either in Switzerland or elsewhere. Due to the lack of appropriate data, the few studies conducted so far do not provide differentiated findings on the long-term dynamics of overall skill demand. These studies may be classified into two groups. The first group consists of econometric studies, investigating the long-term impact of technological change on the employment structure (OECD, 1988, 1996; Machin and van Reenen, 1998; Chennels and van Reenen, 1999; Hamermesh, 2001; for Switzerland, see Kugler and Spycher, 1992). The findings of this literature are astonishingly consistent in claiming a ‘skill bias’ of technological change, consisting of a dwindling demand for low-skilled labour and an increasing demand for more qualified workers. The explanatory power of these highly aggregated studies is severely hampered, however, by the undifferentiated operationalization of both technology and skill demand (i.e., ‘skilled’ and ‘unskilled’ labour). In addition, other potentially relevant factors of skill demand, such as production relocation abroad, work organization, and organizational ecology, are generally excluded from these studies. The second group consists of relatively detailed case studies of the demand for skills at firm level (e.g., Fernandez, 2001), industry level (e.g., Hauptmanns, 1995), and at sector level (e.g., Falk and Seim, 1999).1 As important as these findings are, they cannot be generalized across the entire economy. In some countries, including Switzerland, panel studies of firms have been established. In principle, they permit an accurate estimation of the overall skill demand and the respective determinants (Gerlach, Hübler and Meyer, 1998; for Switzerland, see Arvanitis et al., 2000, 2001, 2003). However, firm-based studies are often limited in that they are either cross-sectional or cover only a short time period (i.e., a maximum of ten years). Hence, short-term and cyclical shifts in skill demand cannot be separated from long-term structural trends. Against this state of affairs, it is no exaggeration to state that there is a lack of long-term, differentiated, empirical studies of skill demand. Our contribution attempts partially to fill this research gap by using data from a long-term empirical study on the overall skill demand in Switzerland for the time period 1950–2000 (Buchmann and Sacchi, forthcoming). In order to assess precisely the long-term dynamics of skill demand, we distinguish two relevant dimensions of change – namely, the occupational composition of skill demand and the requirements of firms regarding the level of formal skills. This contribution thus attempts to provide both a differentiated description and an explanation of the long-term shifts in skill demand. Stefan Sacchi, Alexander Salvisberg and Marlis Buchmann 107

The analysis of these two dimensions of change will be based on the much-neglected data source of job advertisements. Although indicators based on job ads (i.e., Help Wanted Indices) have been developed in several countries (for example, the United States for some years, they only provide information on trends in the number of vacancies and do not make use of the manifold characteristics listed in job ads. Job ads are a rich source of information about the firms recruiting personnel, the advertised vacancies, and the characteristics of prospective incumbents. A representative selection of job ads across all segments of the job market will thus shed light on employers’ efforts, at any given time, to recruit personnel equipped with particular skills and characteristics. Our contribution consists of four analytical sections. We begin by presenting the theoretical considerations about shifts in skill requirements. In this context, we shall also pay attention to the specifics of the Swiss system of vocational education and the Swiss labour market (1). We then describe our data set (2). Next, we examine empirically the extent to which shifts in skill demand in Switzerland may be attributed to technological change and to the dynamics of the underlying population of firms and employers (3). Separate analyses are conducted for the occupational composition of skill demand (3a) and the required skill levels (3b). We close with some conclusions on the long-term dynamics of skill demand and the research approach chosen here (4).

1 Theoretical considerations on the long-term dynamics of skill demand

Past research on the impact of economic and technological change, as well as changes in the organization of work, on the skill level, is the basis of the current debate about the dynamics of skill demand. This long-standing research is primarily associated with the work of Bell (1973), Braverman (1974), and, for the German-speaking context in particular, Kern and Schumann (1970, 1984). Based on the evidence referred to above, Braverman’s thesis regarding a general trend towards job deskilling has been widely rejected. Less unanimous has been the reaction to Bell’s thesis postulating a general trend towards job upskilling. Likewise, Kern and Schumann’s thesis on skill polarization (i.e., opposite trends for different job categories) still awaits clear-cut evidence (see, for example, Baethge, 2001b). Beyond the technology bias inherent in these discussions, the relevant literature often fails to distinguish precisely enough between the skills required for the performance of job-specific tasks and the formal skills relevant for personnel recruitment. Empirical studies, including ours, typically focus on formal skill requirements. Statements about changing skill requirements on the level of job tasks would presuppose the standardized and direct observation of workplaces. The Dictionary of Occupational Titles exemplifies an inventory 108 Long-Term Dynamics of Skill Demand in Switzerland of skill requirements based upon direct observation of tasks performed in the workplace (US Employment Service, 1977; Miller, Treiman, Cain and Roos, 1980). To our knowledge, however, no empirical study investigates long-term shifts in the overall demand for skills at the level of job tasks.2 Because of this ‘missing link’ in the research on skill requirements, it is impossible to assess the extent to which the often-cited upward trend in skill requirements is attributable to a growing share of jobs with more demanding tasks or, by contrast, to the educational upgrading of a pool of nearly unchanged jobs in the light of the growing pool of ever better educated workers (see also the discussion in Buchmann, 1989). In the literature, the latter thesis of ‘credential inflation’ (i.e., the progressive devaluation of educational credentials in the course of their proliferation) is discussed rather controver- sially. Empirical examination of the inflation thesis with respect to skill requirements for job tasks is lacking, however.3 Although empirical evidence on the overall shifts in the skill requirements for job tasks would, of course, be highly desirable, findings regarding formal skill requirements are nonetheless far from obsolete. Job seekers’ opportunities and personnel recruitment depend strongly upon the employers’ demand for particular formal skills – whatever the particular reason that lies behind any specific demand.

1.1 Determinants of the overall demand for skills Theoretically, the long-term dynamics of the overall demand for skills are to be understood in terms of both intra-firm processes and shifts in the economic structure. Both lines of argument are discussed in the literature and find empirical support. Within firms, technological innovation and concomitant organizational change are regarded as the major causes of shifts in skill requirements. With regard to the effects of technological change, the literature maintains that manual and craft skills lose some of their significance, while cognitive and abstract competencies come to play an increasing role. In view of the increasing automation and the broad application of computer technology across all industries and in the service sector, these expectations are well-founded (Baethge, 2001a: 36). With respect to the diffusion of computer work in particular, reflexive and analytical competencies, as well as communicative, organizational, and social skills, are also said to increase in significance (Mertens, 1974; Castells, 1996; Baethge, 2001b; Mayer, 2000; Buchmann, 2002). The literature has also well established that firm-specific and context- specific implementation strategies shape technology effects (Spenner, 1988; Edwards, 1995). This is to say that the characteristics of technology per se do not fully determine its impact. The literature suggests that the impact of technological change is to be primarily expected at the level of job-specific tasks. Finally, in the light of the postulated general upgrading of job skills, the expected devaluation of manual and crafts skills must be qualified as Stefan Sacchi, Alexander Salvisberg and Marlis Buchmann 109 rather ambiguous. Because of the aforementioned ‘missing link’ in research on skill requirements, the question of how shifts in job-specific task requirements affect personnel recruitment and the concomitant overall demand for formal skills in the labour market still awaits unequivocal answers (see, for example, OECD, 1996, vol. 2). A rare exception is the study by Autor, Levy and Murane (2003), which provides evidence of the claim that computer technology substitutes workers performing routine tasks and complements workers executing non-routine tasks. Given the limited evidence and the obvious affinity between formal schooling and intellectual competence, it is reasonable to assume that respective shifts in job-specific tasks, and the skills needed to perform these tasks, are at least partially responsible for the upgrading of job skills (for a similar argument, see Baethge, 2001b: 104). Another major trigger of shifts in skill demand is the changing pattern of work organization within firms. The large-scale introduction of teamwork, the flattening of organizational hierarchies and the concomitant change in the division of competencies have led to increasing levels of individual responsibility on the part of the employees. It seems that these changes in the organization of work have contributed to the upward trend in employers’ skill requirements (Caroli and van Reenen, 1999; Breshnahan, Brynjolfsson and Hitt, 2002; Eschuk, 2003). New forms of work organization and the broad application of information technologies are thus often viewed as complementary processes. According to this line of reasoning, the impact of new technologies should have been strengthened by the concomitant change in work organization, especially that occurring during the 1990s. In the case of Switzerland, Arvanitis (chapter 6 in this volume) provides some evidence supporting the notion of a skill bias pertaining to complementary organizational and technological change, including a positive joint effect on the demand for highly skilled workers. Although prominently discussed some decades ago by authors such as Dahrendorf (1964), Touraine (1969) and Bell (1973), the significance of the changing sectoral composition of the economy for shifts in skill demand has received only scant attention in more recent literature. Half a century after Fourastié (1952), the importance of the dynamics of tertiarization may be so obvious that any additional contributions would render themselves obsolete. In a nutshell, the argument is that the interplay between sector-specific differences in productivity development and increasing purchasing power engenders service-sector-directed shifts both in consumer demand (for goods and services) and labour demand. Although empirical evidence is not abundant, it basically supports this argument (for an overview, see OECD, 2000). In this perspective, the trend towards the service sector is quasi- automatically related to skill upgrading. Moreover, tertiarization and the concomitant reallocation of labour across industries are consensually regarded as the major force behind the changing occupational composition of labour demand. To date, it remains unclear, however, how much the shifting 110 Long-Term Dynamics of Skill Demand in Switzerland demand for occupation-specific credentials contributes to the upgrading of overall skill demand. Under close scrutiny, the relationship between tertiarization and skill demand reveals itself as being more complex than is commonly assumed, however (Baethge and Wilkens, 2001). First and foremost, there are ‘loser’ industries characterized by dwindling employment in the service sector and ‘winner’ industries with growing employment in the industrial sector (Singlemann, 1978; Littek, Heisig and Gondek, 1991). The impact of the reallocation of consumer demand and employment across industries can, therefore, be attributed only partially to the growth of the service sector. Rather, the entire range of shifts in the industrial structure of the economy needs to be taken into account. Likewise, the focus on consumer demand is not sufficient as business-to-business services have grown disproportionately over the last decades (Cornetz, 1991). This seems to be partially due to the outsourcing of services previously provided within a company, and partially attributable to the increasing ‘service intensity’ of production (e.g., see Klodt, Maurer and Schimmelpfennig, 1996). Along with the establishment of a market for specialized suppliers and contractors, both trends have substantially contributed to recent shifts in the economic structure. Hence, industry- and sector-related shifts in skill demand are hardly attributable to the dynamics of tertiarization driven by the interplay between sector-specific differences in productivity growth and shifts in consumer demand. For these reasons, the analysis of skill demand should not exclusively focus on shifts between economic sectors, but on the composition of the organization population in general. Moreover, the forces driving shifts in the organization population deserve more attention (Carroll, 1984). This question aside, any shift in the overall demand for skills may be broken down into two components. These are the inter-company component, involving changes in the population of organizations (i.e., the organizational ecology) on the one hand and the intra-company component on the other. Both technology and work organization may have a direct impact upon skill demand with regard to the second component. An approach focusing on organizational ecology, rather than on tertiarization alone, has the further advantage of also taking into account the relevance of organizational size. Net of industry, technology, and work organization, the ‘skill mix’ of firms is likely to depend upon firm size and organizational growth. In particular, the share of both administrative and control tasks are expected to increase with firm size (Carroll, Haveman and Swaminathan, 1990: 167). The growing internationalization of the economy should be at least be mentioned as an additional factor affecting skill demand. The international relocation of production and firm sites, accompanied by the proliferation of company activities abroad (e.g., business investments, formation of affiliates), and growing international competition are often said to exert negative effects upon the domestic labour market (for an overview, see Dicken, 1992; Stefan Sacchi, Alexander Salvisberg and Marlis Buchmann 111

Trinczek, 1999; for critical assessments, see Hirst and Thompson, 1996; Kurer and Pohl, 1999). Among other things, effects on domestic skill demand are expected as unskilled work is relocated to the periphery (i.e., countries with low wage levels) and ‘high-end jobs’ become concentrated in core countries (Wood, 1994). Consistent with this thesis, the findings of Arvanitis et al. (2001) for Switzerland in the 1990s suggest that the growing international engagement of companies has, by and large, resulted in an upgrading of domestic skill demand. This is especially true in the case of industrial enterprises. Unfortunately, because of the absence of appropriate data, we will be unable to integrate this aspect into our empirical analysis.

1.2 Specific aspects of the Swiss context In anticipation of the empirical analyses based on data for Switzerland, we will briefly describe those institutional and structural characteristics of the Swiss context that are most pertinent to skill demand. Although technology- induced shifts in task requirements are likely to be similar across countries, a number of strong contextual effects are to be expected with regard to adaptation strategies of firms and labor-market outcomes:

1. By international standards, Switzerland is characterized by a large share of small and medium-sized firms (OECD, 1994). In purely quantitative terms, internal labour markets therefore play a lesser role. Consequently, the inter-firm labour market as apparent in job advertisements represents a greater share of total skill demand. 2. A specifically vocational education system is characteristic of Switzerland (and the other German-speaking countries of Germany and Austria). Upon completion of compulsory schooling, the great majority of school-leavers serve an apprenticeship. This involves a two-to-four-year training programme, combining formal vocational education at a college with practical on-the-job training in a firm in the private or public sector. The state is mostly responsible for certification and quality assurance. When recruiting personnel, employers strongly rely on these state-sanctioned certificates. The great majority of vacancies are therefore accessible only to people equipped with the required occupation-specific credentials. At the macroeconomic level, this results in strong segmentation of the labour market, subdivided into numerous occupation-specific sub-segments (see Sheldon, 1996; Sacchi and Kriesi, forthcoming), similar to that described by Sengenberger (1987) for Germany. The skill demand of firms thus becomes predominantly operative within occupation-specific sub-segments of the labour market. For this reason, shifts in the occupational composition of labour demand assume a much greater relevance for Switzerland than for countries with other types of training systems. In this context, any analysis of skill upgrading requires appropriate control of the occupational composition of skill demand. 112 Long-Term Dynamics of Skill Demand in Switzerland

Our theoretical considerations lead us to advance two expectations about the long-term dynamics of skill demand in the Swiss labour market since 1950. With respect to its occupational composition, we expect profound shifts that primarily reflect the far-reaching change in the organization population. With regard to skill levels, we expect an upward trend in formal skill requirements. The long-term shifts between industries are expected to contribute significantly to the shifts in the occupational composition of skill demand. The latter, in turn, is largely responsible for the upward trend in formal skill requirements. To a considerable extent, this upward trend is also the result of technological and organizational change within firms. The relative weight of the determinants mentioned here needs to be assessed empirically.

2 Investigating skill demand – a representative survey of job ads

The empirical analyses presented here are based on a large, representative random sample of all job ads published in newspapers and advertisers in the German-speaking part of Switzerland between 1950 and the year 2000.4 The sample was drawn by a two-stage method whereby approximately 70 newspapers and advertisers (stratified by region and circulation) were randomly selected. For each year, 500 job ads published in the selected newspapers and advertisers were then selected. The full text of each ad was stored.5 The full sample thus includes 26,500 job ads with a total of 31,000 vacant positions. Using standardized classification schemes and computer- aided content analysis, we coded the information provided in the job ads regarding the recruiting firms, vacancies, and characteristics of prospective job incumbents (Werner, 1999).6 For the empirical analyses, the data set of job ads is supplemented by external data. These data are used to estimate the impact of technological change on skill demand. First, we use an industry-specific, annual indicator of the diffusion of computer work in the Swiss economy covering the time period 1977–2002 (Sacchi, 2003). An attempt to assess the effects of computerization on skill demand, this indicator measures the ratio between the number of computer-equipped workplaces and the number of employees for 11 industries. For the period 1960–98, the second indicator is a proxy for the technology-induced change in productivity per hour at the detailed level of 32 industries (Buchmann and Sacchi, forthcoming).7 This indicator captures approximately the component of productivity growth which is due to technical change in a very broad sense, encompassing, for instance, improvements in work organization. For our analyses, we will match both the technology indicator series with the data set of job ads based on survey year and industry classification. For the entire sample of newspapers, data on the development of the real costs of advertisement placement are also available. Stefan Sacchi, Alexander Salvisberg and Marlis Buchmann 113

2.1 Data validation Our validity studies indicate that the data set of job ads adequately represents the dynamics of the overall demand for skills. In this respect, four findings are especially worth mentioning:

1. The cyclical fluctuation of the total number of advertised vacancies corresponds closely to other labour-market indicators: the unemployment rate, for example, laterally mirrors the total number of advertised vacancies (for similar findings, see OECD, 1992). 2. Job ads cover a considerable proportion of the overall demand for labour. Our two representative firm surveys conducted in Switzerland in 1994 and in 2003 show that approximately 44 per cent of all vacancies were advertised in the print media (Bendel, Buchmann and Sacchi, 1995). This is in line with earlier findings for Switzerland and for other countries, confirming that job ads are the ‘silver bullets’ of personnel recruitment (for Switzerland, see Thom, Blum and Zaugg, 1996: 13; Wicki, 1998; for other countries, see Gossens and Degelmann, 1952; Reyher, Spitznagel and Kretschmer, 1990: 368; Deeke, 1991: 68–88; OECD, 1992: 22). Since the end of the 1990s, firms have also started to use the Internet as a recruitment channel (see Salvisberg, Klarer and Sacchi, 2001). The long- term trends of interest in this chapter are not yet affected by the advent of the Internet as a recruitment channel. To account for the online advertisement of vacancies over recent years, our monitoring of the Swiss labour market includes – in addition to job advertisements in the print media – job advertisements published on company websites starting in the year 2001. 3. Comparing the job advertisement data with official statistics, we found evidence that the job advertisement market includes all industries and all types of firms. Admittedly, some differences in the frequency of advertising vacancies do exist. However, the long-term dynamics of the firm population is proportionally represented in the job advertisement market. Albeit at different levels, Appendix A to this chapter illustrates the almost parallel development of the shares which industries hold in employment and in the job advertisement market. Hence, over the entire observation period and across all industries, the totality of firms present in the job advertisement market adequately reflects shifts in the composition of the Swiss firm population as awhole. 4. Our text-based indicator measuring the skill level closely corresponds to a second, independent indicator for the qualifications requested in job ads, which is based on Swiss Census data. Again, the long-term dynamics correspond almost perfectly in both series (see Appendix B to the chapter). 114 Long-Term Dynamics of Skill Demand in Switzerland

2.2 What types of vacancies are advertised? Although the job advertisement market adequately mirrors the long-term dynamics in organization population and skill demand, it would nonetheless be wrong to regard the pool of advertised vacancies as being representative of all jobs which are currently vacant. Rather, our findings, including those based on firm surveys and results from German studies conducted by Reyher et al. (1990; similarly, Cramer, 1990), suggest that the vacancies most likely to be advertised are those for which it is not easy to find prospective incumbents. Theoretically plausible, this comparatively expensive and time-consuming personnel recruitment strategy is primarily taken into consideration for vacancies that cannot be filled by unsolicited applications or less costly recruitment channels, including, for example, applications solicited via informal networks. Hard-to-fill vacancies are those characterized by especially demanding requirements; those requiring novel and therefore scarce skills; and, last but not least, those that are highly unattractive because of low salaries, undesirable working conditions, and other unfavourable job characteristics. In any case, the high frequency with which particular types of vacancies are advertised points to a specific mismatch between the skill demand of firms and the pool of sufficiently qualified, and – due to the lack of preferable alternatives – sufficiently interested applicants. Hence, job ads signal very precisely the unsatisfied labour demand of firms. When interested in examining shifts in skill demand, this is a highly desirable quality. As rapidly expanding job categories and those requiring novel skills are particularly hard to fill, they are likely to be well represented in the job advertisement market. Together with Dietzen and Kloas (1999), we argue that this type of data is also well suited for prospective analyses of skill demand, as firms are pressured to think ahead about future manpower requirements when creating texts for job ads. Here, in contrast to standardized surveys, firms decide themselves what kinds of requirements they want to include in job ads. ‘Killer criteria’ referring to formal, easily measurable requirements (for example, occupational experience over many years) are most likely to be explicitly mentioned in job ads. Skill requirements described in job ads thus assume the function of critical selection criteria. This is the case independent of the skills necessary for the performance of the job-specific tasks. In sum, longitudinal and representative surveys of job ads – which can also be used retrospectively – are low-priced instruments which are highly appropriate for monitoring skill demand over long time periods. They may thus be considered as a viable complement to firm panels.

3 Long-term dynamics of skill demand in Switzerland

Our empirical analyses of skill demand focus on the aforementioned conceptual distinction between occupational composition of skill demand (3.1) Stefan Sacchi, Alexander Salvisberg and Marlis Buchmann 115 and skill level (3.2). In line with the assumption presented in section 1.1 that the changing demand for occupation-specific credentials indirectly contributes to the increasingly higher skill level, we begin by examining the occupational shifts in labour demand. We then proceed to the analyses of skill level. In these models, we will introduce occupational shifts in labour demand as one of the determinants affecting skill level. This makes all the more sense, as the occupational label, prominently presented in job ads, is the major search criterion in the strongly segmented Swiss labour market.

3.1 Dynamics of the occupational composition of skill demand The analyses of long-term shifts in the demand for occupational credentials distinguish 14 occupational categories. Appendix C to the chapter summarizes the long-term shifts in the demand for these occupational categories. A strong increase in the share of job ads is observed for ‘Consultants, IT and media specialists’; ‘Salespeople’; ‘Hospital staff and health service workers’; and ‘Office and administrative staff’. In relation to the initial share observed for 1950–60, the following three occupational categories grow considerably as well: ‘Teachers and scientists’; ‘Engineers and technicians’; and ‘Not precisely defined occupations’. Sharp declines may be observed for ‘Home economics and cleaning personnel’; ‘Workers in other industries’; and ‘Farmers, growers or agriculturalists’. With regard to the remaining categories, the changes are relatively minor. Aside from the extremely strong and somewhat surprising demand for ‘Home economics and cleaning personnel’ in the 1950s, these findings were to be expected. More exciting is the question of the mechanisms that drive these shifts. To model the share of each of the 14 occupational categories in the job advertisement market, we use separate logistic regressions (Hosmer and Lemeshow, 1989).8 To assess the impact of shifts in the organizational ecology on the dynamics of skill demand, we use industry of the firm placing the ad and firm size as predictors, which measured at the categorical level, include 37 and three categories, respectively (see Appendix D). Advertisement- placement costs serve as the control variable. The considerable increase in advertisement-placement costs over the observation period may have affected the likelihood of publishing ads for various job categories differently. With regard to jobs for unskilled workers, in particular, firms may be most eager to substitute job ads with less costly recruiting channels. In addition, we use two technology indicators; one measures the diffusion of computer work and the other captures technology-induced change in labour productivity per hour. In order to account for industry-specific technology effects, the models also include interaction terms (Jaccard, 2001). In order to estimate the dynamics and the significance of the models’ unexplained time trends, we also compute a series of models, including a categorical variable indicating survey year. 116 Long-Term Dynamics of Skill Demand in Switzerland

Because of space limitations, we confine ourselves to summarizing the most important findings and presenting some figures. According to our findings, the dynamics of organizational ecology have had a decisive influence upon the occupational composition of labour demand. In this respect, the changing structure of industry turns out to be by far the most important factor. By comparison, the impact of firm size and advertisement- placement costs is relatively modest.9 When controlling for industry, firm size, and publishing costs of job ads, the contribution of the time variable to model fit is reduced by 68 to 88 per cent for eight out of 14 occupational categories.10 For four other occupational categories, the reduction ranges from to 25 to 59 per cent. For two further categories – office and administrative staff and sales persons – the changes over time remain more or less unexplained when controlling for these factors. When taking the technology indicators into account, they show significant and substantial effects in many instances. Overall, they contribute considerably less to the explanation of the long-term dynamics in the demand for different occupational qualifications, however.11 For illustrative purposes, we present some selected findings in more detail below. For the detailed presentation, we choose those occupational categories where the contributions of the dynamics of organizational ecology to the explanation of the observed long-term trends are maximally different. These are ‘Farmers, growers or agriculturalists’ on the one hand, and ‘Salespeople’ on the other. The category ‘Office and administrative staff’ is an outlier, which we will discuss later. Figure 5.1 shows the long-term trend of the share of all job ads for ‘Farmers, growers or agriculturalists’ and for ‘Salespeople’. The bold black lines represent the long-term trends in the observed annual shares. According to these trends, job offers for ‘Farmers, growers or agriculturalists’ have greatly decreased and those for ‘Salespeople’ have rapidly increased across the entire time period. The grey lines correspond to the long-term trends expected by a model controlling for organizational ecology (industry and firm size) and advertisement publishing costs. Appendix E describes in detail the computation of the trends. In essence, cubic trend functions are fitted to the observed and the predicted annual shares. For ‘Farmers, growers or agriculturalists’, the expected trend neatly fits the observed trend. The unexplained portion of the time trend is thus negligible and, in addition, no longer significant at the .01 level (as confirmed by a likelihood-ratio test). This amounts to saying that the long-term decline of vacancies for ‘Farmers, growers or agriculturalists’ is completely accounted for. For ‘Salespeople’, the occupational category at the other extreme, most of the observed trend remains unexplained. The model variables nonetheless contribute to some degree to the explanation of the long-term dynamics in skill demand. From 1985 onwards, the expected increase equals more or less the observed one, when controlling for the dynamics of organizational ecology. Although the observed increase in the time period before 1985 is obviously not attributable to shifts 117 Farmers, Growers, Agriculturalists Sales persons Year Observed share (Sales persons) Observed share (Farmers) Observed trends Model-based expectations (Organizational ecology) Observed and expected trends in the job for farmers and salespeople market

1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 5 0

20 15 10 Share of advertised jobs (%) jobs advertised of Share Figure 5.1 Figure 118 Long-Term Dynamics of Skill Demand in Switzerland in the internal composition of the organization population, such shifts nonetheless explain the more recent dynamics in the demand for salespeople. With regard to the explanatory power of shifts in the composition of the organization population, we emphasize again that ‘Farmers’ and ‘Salespeople’ represent the two extreme categories. The overwhelming majority of the 14 occupational categories are more closely situated to the category of ‘Farmers’, for which the long-term trend is fully explained. Figure 5.2 presents the development of the demand for ‘Office and administrative staff’. As mentioned before, this occupational category is a special case. Here, the trend is not uniform; rather, it shows a complex dynamic with several turning points. Organizational ecology also exerts a substantial effect. The repeated, pronounced shifts in the demand curve remain unexplained, however. The trends predicted by the model – including industry, firm size, and advertisement publishing costs – are obviously unable to explain the strong increase in the demand observed to the end of the 1960s. Likewise, the pronounced decline from then onwards to approximately the mid-1980s, must be attributed to other causes. To the more recent increase in demand for ‘Office and administrative staff’, organizational ecology has had its impact. When the diffusion of computer work is added to the model, however, its expectations fit the more recent development much better. This suggests that, in the wake of the diffusion of IT technology, the demand for these occupational skills has generally decreased since the mid-1980s.12 When adding the indicator for productivity increase to the model, the predictions show little improvement.13 According to these findings, technological change in the broad sense of the term has had almost no net effect on the demand for these occupations. However, the question of what hidden causes account for the very substantial residual development evident for this occupational group remains very much unanswered. In particular, none of the models explains the pronounced peak around 1970 and the subsequent decline. Possible interpretations might include, first, the transformation of the work organization (e.g., the rise and subsequent decline of Taylorism), which are not taken into consideration in our models; secondly, insufficient control of the organizational ecology and technology effects. However, it should not be overlooked that, for this occupational group, the explanatory power of our models is by far the lowest. Although we have only presented some aspects of our findings, we can nonetheless state that our models account for a substantial part of the long- term trend across almost all occupational groups. The major explanatory contribution is attributable to organizational ecology, while the effects of technology are often quite strong, but industry-specific so that they tend to cancel each other out. For some occupational categories, technological change seems to contribute considerably to the long-term trends in demand, however. With regard to the dominating impact of organizational ecology, we should stress again that the sectoral shift from manufacturing to services offers only 119 Observed values: Share Trend (polynom) Trend (mov. 5-y average) Model-based expectations: Organizational ecology + ICT + Productivity Year Observed and expected job trends in the staff market for office and administrative

1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 8 6 4 2 0

14 12 10 Share of advertised jobs (%) (%) jobs advertised of Share Figure 5.2 Figure 120 Long-Term Dynamics of Skill Demand in Switzerland a partial account for it (as confirmed by respective tests). This is to say that within-sector shifts in the organization population are nearly as important as tertiarization (i.e. between-sector shifts).14

3.2 Dynamics of skill level requirements The analyses presented in this section will provide answers to the question of whether the skill level of the advertised vacancies has increased over the observed time period and, if so, what factors account for the observed trends. We distinguish three skill levels: (1) no credential (i.e., unskilled); (2) occupation-specific vocational credential (i.e., apprenticeship or college- based training); (3) higher credential (i.e., tertiary-level vocational training or university degree). Our data impressively confirm the thesis of the long-term increase in skill level (see Appendix C). Over the time period of 50 years, we find a 26 per cent increase in vacancies requiring occupation-specific credentials and an 11 per cent increase in vacancies requiring tertiary-level credentials. Laterally reversed, the market share of vacancies for unskilled workers has dwindled considerably.15 Of interest here is the degree to which the observed long-term trends can be attributed to shifts in the organization population and to technological change. To answer this question, we ran a series of logistic regressions for each skill level. We limit ourselves to presenting the findings for vacancies with no skill requirements and those requiring tertiary-level credentials.16 In order to estimate the explanatory power of the organizational-ecology factors, the first series of models includes, in addition to the advertisement publishing costs, industry and firm size. The next series of models consider in addition the 14 occupational categories introduced in subsection 3.1. This will allow us to estimate the impact of the changing occupational composition of labour demand on the increasing skill level. The required formal qualifications will, last but not least, depend upon how the types of workers are matched with types of tasks to be performed on the jobs. Then, the third series of models additionally include the indicator for the diffusion of computer work and its interaction term with industry. These interaction terms will permit us to capture industry-specific effects of ICT. In the fourth and final series of models, we include the productivity indicator for the period 1960–98. Interaction terms between productivity and industry are considered as well. Figure 5.3 summarizes the main findings regarding the advertised job openings for unskilled workers. The black markers show their annual share of all job ads. The bold black line represents the observed long-term trend. The other lines symbolize the expected trends according to the aforementioned models. As shown, approximately half of the overall decline of almost 40 per cent in the share of vacancies for unskilled labour may be attributed to the dynamics in organizational ecology and, to a lesser degree, the advertisement-placement costs. 121 Year Observed values: Share Trend Model-based expectations: Organizational ecology + Occupation + ICT + Productivity Observed and expected trends in the job market for unskilled workers unskilled Observed and expected in for trends the job market

1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 0

80 70 60 50 40 30 20 10 Share of advertised jobs (%) (%) jobs advertised of Share Figure 5.3 Figure 122 Long-Term Dynamics of Skill Demand in Switzerland

Additional tests have shown that the impact of organizational ecology is attributable to some extent to shifts between economic sectors and to a greater extent to the changing structure of industry within sectors. The additional inclusion of the occupational composition of labour demand does not alter the model expectations in any significant way. Rather, the occupational structure gains significant explanatory power at the expense of the industry structure. The net contribution of the occupational composition of labour demand is thus rather modest. This amounts to saying that changing organizational ecology engenders shifts in the occupational composition of labour demand (see also section 3.1), which, in turn, contribute to the dwindling pool of vacancies for unskilled labour and, eventually, to the increase in the skill level. This unequivocally confirms our hypothesis that the impact of organizational ecology and tertiarization (i.e., a component of organizational ecology) on skill upgrading is mostly mediated through their effects on the occupational composition of labour demand. When the diffusion of computer work is added to the model, there is little change in the expected decline in vacancies for unskilled labour until the beginning of the 1980s. After the early 1980s, the corresponding trend line points downward, thus approaching the observed trend. Compared to the decline induced by organizational ecology, the sharp drop evident over the last twenty years can thus be accounted for if the diffusion of IT is taken into account. Hence, there is a pronounced negative relationship between the diffusion of computer technology and the demand for unskilled labour.17 By including it in our model, most of the declining demand for unskilled labour is accounted for. Thus, there is just a minor residual decline for the period up to the mid-1960s that needs to be explained. When we add the indicator for the technology-bound component of productivity growth (available for the time period 1960–98) to the model, we find a relationship analogous to the one observed in the diffusion of computer work. At the level of the individual model parameters, a significant, negative relationship between productivity increase and skill demand prevails.18 This is in line with the findings of numerous other studies regarding the ‘skill bias of technological change’. However, the negative impact of the productivity indicator does not alter the trend predicted by the model very much. Correspondence with the observed trend nonetheless improves a little. Compared to the model without the indicator for productivity growth, the slightly steeper curve for the decades of the 1960s and 1970s seems to suggest that the productivity indicator might also account for the unexplained decline in the 1950s – if it had been available for that time period. Overall, our analyses suggest that the long-term decline in the demand for unskilled labour is mostly attributable to the dynamics of organizational ecology and, for the last two-and-a-half decades, to the triumphal procession of IT. By contrast, technological change in the broader sense of the term seems to have had only modest effects. Stefan Sacchi, Alexander Salvisberg and Marlis Buchmann 123

Let us now consider the findings for job openings requiring tertiary-level credentials. The results are laterally reversed compared to those shown for vacancies requiring no credentials. We therefore limit ourselves to an even more compressed presentation. Analogous to the position in Figure 5.3, the observed trends and those predicted by the various models are presented in Figure 5.4. Overall, the share of vacancies requiring tertiary-level credentials increased from approximately 2 per cent to almost 20 per cent between 1950 and 2000. Shifts in the organization population explain about two-thirds of the increase in the demand for highly skilled workers. Again, the impact of organizational ecology turns out to be almost fully mediated through the changing occupational composition of labour demand. And once more, the net contribution of this latter factor is modest. The occupational mix within industry has thus not changed in a way that would have had major consequences for aggregate skill demand. A close positive relationship between the increase in the demand for highly skilled labour and the diffusion of computer work prevails. When this indicator enters the model, the long-term increase in the demand for tertiary-level credentials can be almost completely accounted for. This is shown by the almost perfect correspondence between the observed development and the trend predicted by the model. The increase in the demand for high-skilled labour is also positively related to the technological component of the productivity increase. The long-term trend is even slightly overestimated when this indicator enters the model. This is most likely due to the shorter availability (1960–98) of the productivity indicator. In sum, we may state that our full models for unskilled and highly skilled labour predict the observed long-term trends in skill requirements astonishingly well. When survey year is added to the model, it still has some significant effect. However, the model parameters clearly indicate that this effect is due to short-term fluctuations in the observed shares, which may be induced, for example, by economic cycles (and not by an unexplained long-term trend).

4Conclusions

Based on our analyses we shall first draw some methodological and fundamental conclusions before we move on to address policy implications. Methodologically, our results confirm that job ads are a rich, reliable, and – wrongly – much-neglected data source, well suited, among other things, to investigating employers’ changing demand for skills. The representative monitoring of advertised vacancies allows us, firstly, to observe employers’ demand for skills at the level of the workplace to be filled at a given point in time. By contrast, available studies based on firm panels do not usually 124 Year Observed values: Share Trend Model-based expectations: Organizational ecology + Occupation + ICT + Productivity Observed and expected trends in the job market for highly skilled workers skilled in for highly Observed and expected trends the job market

1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 5 0

20 15 10 Share of advertised jobs (%) (%) jobs advertised of Share Figure 5.4 Figure Stefan Sacchi, Alexander Salvisberg and Marlis Buchmann 125 dispose skill measures at the level of individual workplaces. Moreover, they focus on employees’ skills, rather than skill requirements (see Machin and van Reenen, 1998; and also the literature overview provided by Arvanitis in this volume). As a matter of course, the methodology applied in our study is not entirely problem-free. This applies, for example, to issues such as the question of how far the research design used here can be transferred to other countries. Furthermore, the question of whether the ways in which employers use job ads have changed over time deserves particular attention. The findings available for Switzerland, only partially reported in this contribution, do not suggest this. We therefore draw the conclusion that representative surveys of job advertisements are a useful complement to panel studies of firms. Fundamentally, our findings surpass the present state of research, especially with regard to the following four aspects. First, for the first time, they unequivocally confirm that in Switzerland employers’ skill requirements have substantially increased over long periods of time. Over the last twenty years, this trend has become even more pronounced. In contrast to previous evidence, exclusively referring to employees’ skills, our findings are based on employers’ skill requirements measured at the level of individual workplaces. Secondly, our study provides evidence that the upward trend in skill requirements is, to a large extent, attributable to shifts in the composition of the organization population. In accordance with the literature (see OECD, 1998), the changing organization population is, with regard to shifts between industries in particular, the major trigger for the change in the occupational composition of skill demand. This change, in turn, accounts for about half of the decline in the demand for unskilled workers and for the overwhelming majority of the pronounced increase in the demand for highly skilled workers. Thirdly, our findings lend further support to the hypothesis of the ‘skill bias’ of technological change in favour of better-qualified workers and to the disadvantage of unskilled workers. Using completely different methodology, our findings corroborate one of the major results reported in the respective research literature on Switzerland (see, especially, Kugler and Spycher, 1992; Arvanitis in this volume) and also in other advanced industrial countries. However, previous evidence is based either on cross-sectional and panel studies covering short observation periods only, or on highly aggregated analyses without any direct measures of technology. Our findings go beyond previous results because they provide evidence, for an unusually long time period and at the level of the workplace, of the impact of specific technologies. Accord- ingly, the upward trend in skill requirements is to a large part attributable to the diffusion of IT, increasing the pool of vacancies for highly skilled workers and decreasing the pool of vacancies for unskilled workers across all industries. This is exactly in line with the cross-sectional findings of Arvanitis (in this volume). In addition, we find clear longitudinal evidence that the rapid acceleration of this trend since the 1980s is mainly attributable 126 Long-Term Dynamics of Skill Demand in Switzerland to the diffusion of IT. Besides this, technological change in the wider sense of the term has had very little impact. Rather, the upward trend in skill requirements since the 1950s is almost completely accounted for by shifts in the composition of the organization population and the diffusion of IT. Fourthly, our findings allow us to estimate the relative impact of the dynamics of organization population and technology effects. Over the entire observation period, two-thirds of the increase in the demand for highly skilled workers and approximately half of the decrease in the demand for unskilled workers are attributable to shifts in the composition of the organization population. The diffusion of IT explains much less – about one-quarter and one-third, respectively – of the above-mentioned shifts. Considering the time period since the 1980s only, the relative impact of the diffusion of IT is obviously much larger, however. With respect to the demand for unskilled labour, the technology indicator explains more than half of the decrease, whereas approximately one-third is attributable to the changing organization population. The few studies known to us that make statements about relative impacts (see Machin and van Reenen, 1998; Autor et al., 2003) ascribe a smaller relative weight to the changing composition of the organization population. Autor et al. (2003: 1302) nonetheless share our conclusion that the significance of the ‘within-industry’ component has increased over the past few decades. One open question remains, however: why is the relative weight of organizational ecology greater in our study? Other methodological differences between the studies aside, the focus on employment as opposed to vacancies may play a major role. On the other hand, differences between countries and observation periods may also be relevant here. In particular, the indirect impact of the changing organization population, mediated through the occupational composition of skill demand, may be much stronger for countries that are characterized, like Switzerland, by a dual vocational education system and correspondingly tight links between occupation-specific credentials and industry. The major findings of our study also suggest some policy implications. First of all, they support the growing evidence of the increasing difficulties unskilled and low-skilled people encounter in the labour market. Moreover, they lend support to our assumption that job openings for low-skilled workers have largely diminished because of the increasing significance of IT. In addition, the changing structure of industry has favoured those industries that have traditionally hired skilled workers. Both the broad application of IT and the changing structure of industry are consensually viewed as irreversible and continuing developments that can hardly be influenced by political means. A future recovery of the demand for unskilled workers is therefore highly unlikely. Against this background, the difficult question arises of how we can minimize the foreseeable and growing difficulties of unskilled and low-skilled Stefan Sacchi, Alexander Salvisberg and Marlis Buchmann 127 workers in the labour market, and the concomitant burden on the social security system. The popular and apparently plausible answer to this question is continuous education. We do not think at all that this answer is wrong. However, three things are too often forgotten. First, it is well known that among those who do access continuous education, the skilled and highly skilled people are strongly overrepresented (see, for example, Buchmann et al., 1999). By optimizing the pertinent incentives (e.g., in the context of unemployment insurance), coupled with more tightly focused distribution of state subsidies, some success regarding the continuous education of unskilled and low-skilled people is likely to be achieved. Secondly, basic education that has been missed can hardly be compensated for later in life. In the long run, any educational policy needs to aim, therefore, at guaranteeing a core of basic education for the greatest number of young people possible. Finally, and very importantly, the decline of job openings for unskilled people is coupled with reduced employment opportunities for people who are unable to handle computer work and work tasks requiring cognitive skills and also for people who simply wish to perform undemanding jobs. In these instances, educational measures are unlikely to show any effect for people who are unable or unwilling to follow additional educational instruction. Similarly, the intention to save these types of workplaces by taking political and jurisdictional measures in defiance of ‘market laws’ is doomed to fail in the long run. Against this background, detailed information is needed about firms that are successful in holding or even creating these types of workplaces in Switzerland. This would permit the provision of good conditions and good incentives for such firms, as well as financial support for start-ups. From a societal perspective, a marketable business idea for a labour-intensive ‘low-tech’ firm is no less valuable than a new business promoting a ‘high-tech’ product that is likely to be successful on the world market.

Notes

1. Besides the panel studies conducted by the Swiss Institute of Business Cycle Research of the Swiss Federal Institute of Technology, Zurich (see Arvanitis et al., 2000), a few cross-sectional studies of regional labour markets (Marelli, 1985; Wicki, 1998) and some qualitative findings regarding personnel recruitment and staffing (Thom and Zaugg, 1995; Thom, Blum and Zaugg, 1996) are available for Switzerland. 2. A recent study conducted by Autor, Levy and Murane (2003) empirically examined the direct effects of computerization on the performance of job-specific tasks. 3. The major reason is the lack of appropriate, objectified data on job requirements. Thus, empirical analysis of the inflation thesis has primarily focussed on various labour-market outcomes, and on income effects in particular (e.g., Boylan and Morgan, 1993; Becker and Schömann, 1996; Müller, 2001). 128 Long-Term Dynamics of Skill Demand in Switzerland

4. The German-speaking part covers slightly more than 70 per cent of overall employment in Switzerland. 5. The size of the annual subsamples is kept constant in order to make estimation accuracy independent of the cycles within the job advertisement market. Apart from this, the sample is self-weighting as the different sampling fractions at the newspaper level are compensated for when selecting individual job ads within the selected newspapers. The annual samples are based on a fixed target week. Based on a pilot study, the second week of March was selected. 6. Testing the intercoder reliability, Cohen’s Kappa (for two coders) turned out to be .91 for occupational group; .74 for industry; .71 for educational level; and .61 for firm size. With the exception of firm size, these results suggest that the reliability of the coding is excellent in some instances and satisfactory in others (see Lombard, Snyder-Duch and Bracken, 2002). 7. Using productivity measures as indicators of technological advance is not without problems. From a long-term perspective, however, technological development is univocally regarded as the main motor of productivity increase (OECD, 1996). In addition, we counteract some of the well-known measuring problems by adjusting the raw per capita productivity series. First, we correct the underlying employment series for the industry-specific rate of part-time work. The adjusted indicator is thus a better approximation of productivity per hour. Secondly, we index the industry-specific productivity series (1960 = 100) to account for the fact that productivity series can hardly be compared across industries (see Warnken and Ronning, 1989: 51). Thirdly, this raw indicator is then used to extract, for each industry, a non-linear trend component. This was done to eliminate short-term, business-cycle-related fluctuations – induced, for example, through ‘labour hoarding’ or destocking – from long-term productivity trends (see also Müller, 2000). 8. Theoretically, the method of choice would really be multinomial logistic regressions (Hosmer and Lemeshow, 1989: 216pp). However, given a dependent variable with 14 categories, the number of parameters to be estimated and the number of empty cells increase rapidly, thus posing severe estimation problems. Note also that the technology indicators are not available for three occupational categories, namely, farmers, growers or agriculturalists; home economics and cleaning personnel; not precisely defined occupations. For this reason, we exclude these three categories and the respective industries from the analysis (i.e., when estimating the share of the other occupational categories in the Swiss job advertisement market). Separate estimates for the three excluded categories show that their declining share in the job advertisement market is mostly attributable to shifts in organizational ecology. This is to say that the exclusion of the three occupational categories hardly affects, mutatis mutandis, the residual trends to be explained for all other categories. 9. The findings support our hypothesis that the increasing publishing costs for job ads lower the share of ads for manual, low-skilled vacancies and the respective occupational categories. 10. These figures and the ones presented below refer to the reduction (in percent) of the χ2-contribution of survey year to the entire model fit (on the basis of likelihood ratio tests) when industry, firm size, and advertisement placement costs are included in the model. In this way, we can estimate the relative contribution of a given variable, or variable group, to the explanation of the long-term trend. The disadvantage is, however, that the effect of survey year may also include Stefan Sacchi, Alexander Salvisberg and Marlis Buchmann 129

short-term, cyclical fluctuations of skill demand. For this reason, we also compared, on the basis of a dissimilarity index, the observed trend in the share of the advertised vacancies across survey years with the respective trend implied by the expectations of a given model (see also Appendix E). The results of this comparison do not change in any way the conclusions we drew with regard to the impacts described above. 11. The partly positive and partly negative effects for different industries seem to cancel each other out. 12. The effect of the diffusion of computer work on the demand for office and administrative staff varies across industries. Besides the dominating negative effects, we also find positive effects. 13. Again, besides the dominating negative effects, we find weak positive effects for some industries. 14. As the long-term shifts in the organization population are more or less mirrored proportionally in our data set of job ads (see Appendix A), we are confident we have excluded the possibility that the respective findings represent artifacts produced by changes in firms’ personnel recruitment behaviour. 15. This does not necessarily mean that the total number of unskilled jobs has declined. It indicates, however, that the demand for unskilled workers has faded in relation to the respective supply, so that it has become easier to fill these vacancies without costly job advertisements. 16. The findings of the third series of models referring to the category ‘Occupation- specific vocational credential required’ are redundant as the shares of the three categories add up to 100. 17. The negative, highly significant effect of IT on job openings for unskilled workers varies by industry. It nevertheless shows up across all industries. 18. Again, we find significant variation in these effects across industries. Negative effects of different strength are found for slightly more than half of the industries. For the other industries, we find no effect.

Appendix A: development of overall employment and advertised jobs by industry

Similar long-term trends are observed for both overall employment and advertised jobs, documented by the respective ratios shown in the last two columns of Table A5.1. The only major discrepancy is observed for the category Transport/Communica- tion. We suspect that this is attributable to the unfortunate aggregation of modern telecommunication and traditional transport and news services. This aside, the figures show that all industries are present in the job advertisement market, although to variable degrees. Most important, the long-term shifts in employment broken down by industry are reflected astonishingly well in the job advertisement market. The partly substantial discrepancies between industries shown in the table are not necessarily caused by industry-specific rates of advertising vacancies. Rather, the number of vacancies may vary because of differing fluctuation rates with given levels of employment. This makes itself noticeable in the number of advertised vacancies. For the hotel and restaurant industry, well known for its high fluctuation rate, this explanation is most appropriate. 130 Long-Term Dynamics of Skill Demand in Switzerland

Table A5.1 Overall employment and advertised jobs by industry

Proportion (%) Ratio

1955 1998 1998/1955

Empl. Adv. Jobs Empl. Adv. Jobs Empl. Adv. Jobs

Machine/Metal/Electrical/ 19.8 18.5 14.7 11.4 0.74 0.62 Chemical/Energy Other industries 28.1 19.4 8.2 5.0 0.29 0.26 Building and Construction 13.9 14.7 8.5 8.8 0.61 0.60 Trade 15.3 19.8 15.4 18.7 1.01 0.94 Banks/Insurance 2.3 1.9 5.9 6.1 2.55 3.31 Transport/Communication 6.6 1.2 6.8 4.9 1.02 4.26 Hotels and Restaurants 7.1 17.5 6.5 15.8 0.92 0.91 Welfare Services/Culture/ 3.5 2.4 13.2 7.8 3.71 3.31 Sports Teaching/Science and 0.5 0.2 6.4 2.7 11.98 13.25 Research Public Administration/ 3.5* 2.8* 4.0 4.6 1.13 1.64 Social Security Insurance Advisory Services/ Business 1.2 1.6 7.7 10.2 6.64 6.30 Services Personal Services 1.7 2.3 2.9 4.0 1.73 1.75

Note: *Figures of 1975 instead of 1955. The proportion of advertised jobs is based on three survey years (1954–56 and 1997–99); advertised jobs in Agriculture and Private Households are excluded as comparative figures are missing. Advertised jobs for which no industry is indicated are excluded as well. Sources: Employment figures (a) Swiss Federal Office of Statistics (1962): Swiss Establishment Census August 1955, Trading Companies in Switzerland 1955, Table 2; (b) Swiss Federal Office of Statistics (1967): Swiss Establishment Census September 1965, Major findings for Switzerland; Establishments and Employees 1955 and 1965; Table 1.00; (c) Swiss Federal Office of Statistics (2000): Swiss Establishment Census 1998. Table 1.

Appendix B: validating the indicator of skill level

In job ads, the required skills are often not explicitly stated. Rather, Swiss firms take the respective formal credentials more or less for granted. Hence, when coding the requested skill level, job ads of earlier decades especially pose a problem. In order to validate the text-based indicator of skill level, we computed, based on Swiss Census data from 1960, 1970, 1980, and 1990, the proportion of unskilled workers for detailed occupational categories. This indicator was then matched with our data set of job ads. Figure A5.1 illustrates the linear and more or less identical long-term trends in both indicator series. The cyclical component superimposing the long-term trend in the text-based indicator seems to be slightly stronger, which is not relevant in the present context, however. In addition, the share of advertised jobs accessible to unskilled workers is invariably higher for the text-based indicator of skill requirements. This is most likely the result of our coding rules, according to which vacancies not specifying any skill levels were assigned to the category ‘unskilled’. These deviations are not problematic as we are interested in long-term shifts and not in absolute levels. 131 1990 1985 1980 Year 1975 1970 unskilled workers workers unskilled 1965 Text-based Indicator Census-based Indicator Trend (Text-based Indicator) Trend (Census-based Indicator)

1960 0

Advertised jobs for jobs Advertised 80 70 60 50 40 30 20 10 Share of advertised jobs for unskilled workers (%) workers unskilled for jobs advertised of Share Figure A5.1 132 (%) 2.5 5.9 0.2 0.0 6.4 0.9 2.4 4.5 3.8 9.5 1.2 0.1 5.1 − − − − 18.5 26.1 37.4 11.3 − − (1.–5. decade)

Advertised jobs (%) jobs ChangeAdvertised 9.7 10.0 11.9 9.5 8.9 7.4 6.2 4.0 8.4 10.3 10.1 13.4 1.2 1.2 1.6 2.5 3.4 3.7 6.0 4.9 9.7 6.8 6.1 5.3 1.7 1.1 1.0 1.3 0.6 1.1 1.6 3.0 2.6 5.0 6.8 11.3 1.7 2.8 4.4 6.7 2.8 3.8 6.3 13.2

1961–70 1971–80 1981–90 1991–00 .5 9.7 8.7 13.0 11.8 12.8 8.7 1.2 2.4 5.8 0.7 1.8 7.1 11.7 10.7 10.0 10.9 1.7 1.9 20.5 28.4 33.7 44.7 46.7 10.4 77.6 68.7 62.5 49.0 40.2 22.9 11.8 9.5 5.6 4.5 14.7 15.6 18.4 16.0 13.8 5943 6186 6370 6311 6131 5943 6186 6370 6311 6131 30’941) =

required chemical industrieschemical 2. Occupation-specific vocational credential 1. required credential Tertiary-level 6. Salespeople 7 3. No credential required 3.and electrical machine, metal, the in Workers 4. industries other in Workers 5. Engineers and technicians 7. Office and administrative staff 8. Consultants, IT and media specialists 2. craftsmen workers, Building 1. Farmers, growers or agriculturalists 9. Transport and(tele)communication employees 5.2 14. Not precisely defined occupations level Skill Sample size per decade (n 12. Hospital staff and health service workers 13. Home economics and cleaning personnel 11. scientists and Teachers 10. Hotel and restaurant service personnel 1950–60 Occupational categories Appendix C:Appendix 1950–2000 demand skill level of educational and composition occupational Stefan Sacchi, Alexander Salvisberg and Marlis Buchmann 133

Appendix D: coding of industry and firm size

Categorization of industry (the industrial codes are in brackets): 1. Agriculture (01, 02, 03, 04); 2. Food, alcohol, and tobacco (21, 22, 23); 3. Energy, water (11, 84); 4. Textiles (24); 5. Clothing (25); 6. Paper (27); 7. Printing (28); 8. Leather, synthetic materials (29, 32); 9. Chemical (31); 10. Stone Quarrying (12, 33); 11. Metal (34); 12. Machine (35); 13. Electrical (36); 14. Watch and Clock (37); 15. Timber (26, 38); 16. Building and Construction (41, 422); 17. Technical building installations (421); 18. Trade (51, 52, 53, 54, 55, 56); 19. Hotels and Restaurants (57); 20. Banks (71); 21. Insurance (72); 22. Public transport (61, 63, 64); 23. Haulage (62, 651); 24. Postal Services (66); 25. Medical Practices, Laboratories (831, 832, 834, 835); 26. Hospitals (833); 27. Real Estate/Property (73, 74); 28. Public administration, Social Security Insurance (91, 92); 29. Teaching (81); 30. Universities, Research (82); 31. Welfare Services (85, 86); 32. Culture, Sport (88); 33. Advisory Services (652, 75, 87); 34. Repair Services (58); 35. Personal Services (76); 36. Private Households; 37. Undefined industry.

Categorization of firm size: 1. Small firm (up to 10 employees), 2. large firm (more than 100 employees), 3. medium- sized firm (11–100 employees or number of employees unknown).

Appendix E: computation of observed and model-based trends

In principle, all trend lines correspond to a cubic function of the index of survey years t (1950 = 1, 1951 = 2, . . . , 2000 = 51) as expressed by the following formula:

= + + 2 + 3 Ti a b1t b2t b3t

The parameters of the function are estimated with OLS. For the observed trends, the π sample-based estimate ij for the share of advertised vacancies of category i (occupational category or skill level) in year j serves as dependent variable. The point estimates π for ij are also shown in the figures. The cubic function represents the underlying trends sufficiently well. Estimates of higher-order polynomials do not result in any substantial improvement. This also applies to office and administrative staff for which the cubic function does not fit well. For this reason, the respective trend in Figure 5.2 is based on a five-year moving average.

To compute the trend function TAi predicted by model variant A, the same cubic function is fitted to the model-based estimates of the annual shares (instead of the sample-based estimates) according to the following formula:

n j ()β + [β .x ] e 0 k jk ------∑ ()β + [β .x ] 1 + e 0 k jk ˆ 1 πAij = ------nj 134 Long-Term Dynamics of Skill Demand in Switzerland

Hence, we get a model-based estimate for the share of advertised vacancies of category i in year j (see Hosmer and Lemeshow: 1989), whereby xjk corresponds to β the model-specific vector of covariates, k to the vector of the respective regression β coefficients, 0 to the regression constant, and nj refers to an index of all jobs in the sample advertised in year j. 6 Information Technology, Workplace Organization and the Demand for Employees of Different Education Levels: Firm-Level Evidence for the Swiss Economy* Spyros Arvanitis

This chapter aims at investigating at the firm level the relationship between the demand for employees with different levels of (vocational) education and some of the factors claimed to be responsible for determining this demand in the 1990s: use of modern information and communication technologies (internet, intranet) and new forms of workplace organization (team work, job rotation, modes of decision decentralization). The study is based on data for about 1,400 Swiss firms in manufacturing, construction and in the service sector collected by means of a survey. The results of multivariate analysis show that the intensive use of technology and new forms of workplace organization correlate positively (negatively) with the employment shares of high-educated (low-educated) personnel, and further that these effects strengthen each other. These results are consistent with the hypotheses of (a) skill-biased technical change and (b) skilled-biased organizational change.

1 Introduction

There is long-term empirical evidence that both the number and the employment share of highly skilled (or highly educated) workers have grown over time in many OECD countries. In the past few years many prominent economists have been engaged in an intensive discussion on the reasons for

* This study was supported by the Swiss National Science Foundation (project number 5004-05446; SPP ‘Switzerland – Towards the Future’). Comments and remarks from the experts who evaluated this work on behalf of the Swiss National Science Foundation are gratefully acknowledged.

135 136 Information Technology and Demand for Employees the observed shift of labour demand towards highly skilled workers (see, e.g., Johnson, 1997 and the other contributions to the symposium in the Spring 1997 issue of the Journal of Economic Perspectives). One of the most popular explanations offered by the current economic literature is based on the so-called ‘skill-biased technological change’ hypothesis, according to which the reason for the upskilling of the labour force is the non-neutrality of technological change, which favours the use of skilled labour more than the use of other labour inputs. Due to the complementarity of skills (education) and technology, an acceleration of the rate of technological change would cause an increase of the demand for skilled labour.1,2 The reason for the most recent acceleration in technological change is assumed to be the diffusion of information and communication technologies (ICT) which seem to have given new impetus to the substitution process of low-skilled by highly skilled employees. Nevertheless, for some authors it is not clear whether the observed shift of labour demand is caused mainly by within-sector technological change or, rather, by sector-biased technological change – that is, technological change affecting only some specific sectors (see, e.g., Haskel and Slaughter, 2002). On the whole, the technology hypothesis cannot explain the entire magnitude of the observed shift in labour demand. This is the reason why some researchers have looked for other possible alternative or complementary explanations of the change of the composition of the labour force. For trade economists, such alternative explanations are related to the internationalization of economic activities (see e.g. Wood, 1995),3 while for industrial and managerial economists they are associated with the reorganization of production which took place parallel to the introduction of ICT. The latter is the approach we are also going to adopt in this chapter. The present study empirically explores the hypothesis that ICT and new organizational practices are important determinants of the demand for labour of different skills, and, further, that the joint use of these two factors leads to a mutual strengthening of their impact upon labour demand. The analytical framework is that of a demand function for employees with different education levels (heterogeneous labour) at firm level. The study’s new contribution to empirical literature consists in being the first empirical study of this type for Switzerland using a rich data set at firm level which was collected by means of a postal survey and giving particular attention to the complementarity issue and to the endogenization of organization.4 In addition, we focused on some statistical problems which are typically related to survey data; multiple imputations were used to substitute for missing values (the problem of item non-response). Despite these advantages, there are also some shortcomings in the study, the principal one being that it is only a cross-section analysis which does not allow the testing of causal relations, the use of lags between variables, etc. The chapter is structured as follows: section 2 sketches the analytical background of our contribution, related to new theories on the combined Spyros Arvanitis 137 influence of ICT and organizational factors on labour demand. Section 3 gives descriptive information on changes in the composition of labour demand as well as the existence and diffusion of ICT and new organizational practices across the Swiss business sector. Section 4 contains a description of our data. In section 5 the specification of the empirical model is presented and discussed. Section 6 contains the results of the econometric estimates. A survey of similar empirical studies for other countries is presented in section 7. Finally, we summarize the main findings, indicate some directions of future research and draw some policy conclusions.

2 Conceptual background

2.1 The ‘skill-biased technical change’ hypothesis: the role of ICT The shift towards more highly educated workers, which can be observed in many OECD countries since the late 1960s or possibly the early 1970s, appears to have accelerated in the last twenty years (see e.g. Berman et al, 1998; OECD, 1998). While many factors have contributed to this increase most authors feel that this effect is attributable primarily to skill-biased technological change. The size, breadth and timing of the recent shift in labour demand have led many observers to seek skill-biased technological change in the largest and most widespread new technology of the last years, ICT (see Bresnahan, 1999; Bresnahan et al., 2002). On the one hand, highly skilled labour is a precondition for the use of ICT; for example, training in problem-solving, statistical process controls and computer skills can increase the benefits of ICT. On the other hand, highly computerized systems not only systematically substitute computer decision making for human decision making in routine work, but they also produce a large quantity of data which needs highly skilled workers, managers and professionals in order for it to be fully exploited. The specific influence of ICT adoption and use on the composition of the workforce has been a particular subject of recent theoretical and empirical analysis. One important proposition is that ICT capital (i) substitutes for workers performing cognitive and manual tasks that can be accomplished by following given rules; and (ii) complements workers in performing non- routine cognitive tasks concerning generalized problem-solving and complex communications (see Bresnahan, 1999; Autor et al., 2003). This distinction of routine and non-routine tasks leads to a further differentiation related to the possibility of different effects on the skill mix depending upon the type of technology under study. In manufacturing firms with a large scope for factory automation technologies an increase of the share of highly skilled or highly educated workers may prove to be rather small in comparison to manufacturing or service firms in which computer investment is mainly related to the work of non-production workers with administrative or managerial tasks, etc. (see Doms et al., 1997). 138 Information Technology and Demand for Employees

2.2 The ‘skill-biased organizational change’ hypothesis: the role of new forms of workplace organization A further hypothesis put forward in the recent literature refers to the influence of the increasing diffusion and application of intra-firm reorganization processes on the observed change of firms’ skill requirements. The basic idea is that a gradual shift from rigid ‘“Tayloristic” organization (characterized by specialization by tasks) to “holistic” organization (featuring job rotation, integration of tasks and learning across tasks)’ (Lindbeck and Snower, 2000, p. 353) is taking place within firms. This phenomenon appeared first in the USA and Japan and has spread later throughout Europe, although at a different pace from country to country (see Aoki, 1986; Greenan and Guellec, 1994). The main elements of reorganization in the workplace according to economic, management and sociological studies on this field are (see Caroli, 2001 for a survey of the literature on this subject): (i) the decentralization of decision making by delegation of relevant competences from management to lower hierarchy levels, increased involvement and autonomy of employees at the shopfloor level; and (ii) new working practices such as teamwork (semi-autonomous workteams, quality circles, and so on), job rotation, other forms of multi-tasking, multi-skilling, etc. Many authors seem to share the idea that changes in work organization towards more ‘holistic’ structures, definitely require an upgrading of the skill content of most jobs related to these changes. Caroli (2001) presents a series of reasons for this. Current organizational changes increase employees’ responsibility for tasks and operations. This is not only the case for operatives but also for supervisors and technicians, whose roles, hence skills, are considerably modified by the new organizational practices. Interpersonal abilities also become more important owing to the increasing need for communication and co-ordination. Thus, an important precondition for the successful implementation of most of these new organizational practices is the availability of a higher skilled (or higher educated) workforce.

2.3 Complementarities: technical change related to organizational change The use of ICT, new organizational practices and human capital build a ‘complementary system’ of activities (Bresnahan et al., 2002, p. 341ff; Milgrom and Roberts, 1990 p. 191ff.). According to Milgrom and Roberts (1990, p. 514) ‘the term “complement” is used not only in the traditional sense of a specific relation between pairs of inputs but also in a broader sense as a relation among groups of activities’. In addition, in the case of organizational practices there exist interdependencies with other factors and inputs. Some of the changes of work design are associated with the introduction and diffusion of information technologies within the firm. For example, in a theoretical paper Greenan and Guellec (1994) show that the Spyros Arvanitis 139 relative efficiency of a centralized mode of firm organization in which knowledge is confined to specialized workers and a decentralized one in which every worker participates in learning depends on the technological level of the firm: ‘whereas the centralized style is more efficient when the technological level is low, the decentralized one becomes more efficient when the technological level is higher’ (p. 173).

2.4 Demand equations for specific labour skills The above discussion of the literature shows that there are some common testable hypotheses with respect to the contribution of ICT and new organizational practices to changes of the skill requirements of labour. These can be best put together formally in a framework of demand equations for categories of employees with different skills or education containing not only factor prices and physical capital but also ICT and organization capital (see Doms et al., 1997; Machin and Van Reenan, 1998):

— Hypothesis 1: there are considerable positive (negative) effects of ICT on the demand for highly educated (low-educated) employees; — Hypothesis 2: there are considerable positive (negative) effects of new forms of workplace organization on the demand for highly educated (low-educated) employees; — Hypothesis 3: there is a positive (negative) interrelationship between technology and organization leading to a mutual strengthening of the effects of these two factors on the demand for highly educated and low-educated employees respectively (complementarity hypothesis).

3 Descriptive analysis

3.1 Use of ICT and new organizational practices in the Swiss business sector in the 1990s Between 1995 and 2000, in common with many other OECD countries, the use of information technologies in the Swiss business sector increased at a tremendous rate. In 2000 94 per cent of all firms (with at least five employees) used a personal computer, 86 per cent e-mail and 78 per cent the Internet; about 55 per cent of Internet users utilized a homepage (see Arvanitis, 2004). Many firms also used more complicated networking technologies (electronic data exchange with other firms (EDI), firm computer networks (LAN/WAN), intranet and extranet). On the whole, Swiss firms are well-equipped with information technology; compared to other countries Switzerland is ranked (with respect to the overall diffusion of information technologies) behind the USA and the Scandinavian countries, but ahead of many other European countries (see Arvanitis and Hollenstein, 2002). 140 Information Technology and Demand for Employees

Parallel to the diffusion of ICT the dissemination of new forms of flexible work organization could be observed in some sectors of the Swiss economy. Two of the main forms of flexible organization are teamwork and job rotation. According to the results of a representative survey among enterprises with at least five employees conducted in 2000, 35.7 per cent of Swiss firms had introduced teamwork, 10.4 per cent job rotation. There has been a considerable acceleration of the adoption of such organizational practices in the Swiss economy since 1995. Up to 1995, 16.9 per cent of all firms had introduced teamworking, 7.4 per cent did it between 1995 and 1998 and 11.4 per cent between 1998 and 2000 (see Arvanitis, 2004). For job rotation the corresponding shares of firms are considerably lower, but also increasing; only 5.1 per cent of firms used job rotation before 1995, 1.8 per cent introduced this organizational practice between 1995 and 1997, and 3.5 per cent more between 1998 and 2000. These forms of flexible workplace organization could be found in all sectors of the economy, but in manufacturing at the strongest, particularly job rotation. At the same time a decentralization of decision making within enterprises has also taken place. 40 per cent of all firms declared in the survey already mentioned above that since 1995 management has delegated various competences to their employees or teams of employees, aiming at achieving a decentralization of firms’ decision-making process. Only 2.9 per cent of them found that a shift towards stronger competences (not of workers but) of managers has taken place since 1995, for 57 per cent of firms there was no change with respect to within-firm competence delegation. This decentralization effect was at strongest in manufacturing. The increasing use of flexible organizational modes was a particular characteristic of sectors of ‘niche’ medium-sized high tech manufacturing industries such as scientific instruments and specialized machinery, pharmaceuticals and financial and producer service industries in which the Swiss economy has traditionally specialized.

3.2 Shift of labour demand towards highly educated employees in the Swiss business sector in the 1990s A third remarkable tendency in the 1990s has been the increase of the share of employed who dispose of tertiary education from 19.2 per cent in 1991 to 23.4 per cent in 2000. This effect could be explained through the combined increase of both the demand for and supply of highly skilled workers in the 1990s, but there is some evidence that the demand effect was dominant. Table 6.1 offers some qualitative information on the extent of the shift of the skill-mix of Swiss firms in the period 1997–2000. The difference between the percentage of firms reporting an increase and the percentage of those reporting a decrease of the employment share of university graduates in the period 1997–2000 was 4.7 per cent, for employees with education at the tertiary level this percentage was 15.7 per cent and for employees with a vocational education ending with a formal degree 5.6 per cent. Of Spyros Arvanitis 141

Table 6.1 Percentage of firms with decreasing, unchanged or increasing shares of employees with different levels of vocational education, 1997–2000

H1/LT H2/LT M/LT L/LT

All Firms

(1) decrease 1.9 3.7 12.2 18.2 (2) no change 91.5 76.9 70.0 72.4 (3) increase 6.6 19.4 17.8 9.4 (3)−(1) 4.7 15.7 5.6 −8.8

Note: H1/LT: share of employees with university education; H2/LT: share of employees with technical or business college education (tertiary level); M/LT: vocational education ending with a formal degree; L/LT: vocational education without a formal degree/no vocational education; data for 2,648 firms; multiple imputations for missing values; the data were corrected for unit non-response bias and weighted in order to reflect the population of Swiss enterprises belonging to the two-digit industries listed in Table A6.1. course, during such a short period the shares remained unchanged for most firms. For all three employee categories we observe at the balance an increase of the corresponding employment shares; the extent of this increase is not the same for all categories. On the contrary, the overall change was negative for low-educated employees (vocational education without a formal degree/no vocational education), 18.2 per cent of all firms reported a decrease of the employment share of this employee group, only 9.4 per cent an increase; at the balance, 8.8 per cent of all firms reduced the share of low-educated employees. The differences among sectors (manufacturing, construction, services) are not so large, the direction and extent of the shifts are quite similar all over the economy. Table 6.2

Table 6.2 Formal education of employees in Swiss business sector, 1999

Manufacturing Construction Services Total

Formal education (average share of employees): – University 2.7 1.0 7.8 5.4 – Other tertiary-level education 12.6 10.7 16.5 14.7 – Vocational education; formal 45.0 52.0 44.6 44.8 degree – Vocational education without 34.9 25.5 24.2 29.0 formal degree; no vocational education – Persons in the process of 4.8 10.8 6.9 6.1 vocational education

Note: Data of 2,648 firms; multiple imputations for missing values; the data were corrected for unit non-response bias and weighted in order to reflect the population of Swiss enterprises belonging to the two-digit industries listed in Table A6.1. 142 Information Technology and Demand for Employees shows the educational composition of employed labour of the firms in our sample. Combining all elements, we obtain the following picture of the Swiss business sector at firm level at the end of the 1990s: high average educational level, intensive use of information technologies and widespread flexible organizational practices at firm level seem to be positively correlated with one another; this tendency has been accentuated in the second half of the 1990s.

4 Data

The data used in this study were collected in the course of a survey among Swiss enterprises using a questionnaire which included questions on the incidence and within-firm diffusion of several ICT technologies (e-mail, Internet, intranet, extranet, etc.) and new organizational practices (teamwork, job rotation, employees’ involvement, etc.), employees’ vocational education, the flexibility of working conditions and labour compensation schemes.5 The survey was based on a (with respect to firm size) disproportionately stratified random sample of firms with at least 20 employees covering all relevant industries of the business sector as well as firm-size classes (on the whole 28 industries and within each industry three industry-specific firm size classes with full coverage of the upper class of large firms). Answers were received from 1,667 firms – that is, 39.4 per cent of the firms in the underlying sample.6 The response rates do not vary much across industries and size classes, with a few exceptions (overrepresentation of paper and energy industry, underrepresentation of hotels, catering and retail trade; see table A6.1 in the appendix for the structure of the used data set by industry and firm size class). The non-response analysis (based on a follow-up survey of a sample of the non-respondents) did not indicate any serious selectivity bias with respect to the use of ICT and new organizational practices (teamwork, job rotation). A careful examination of the data of these 1,667 firms led to the exclusion of 123 cases with contradictory or non-plausible answers; this left 1,544 valid answers which were used for the econometric analysis. Further, we used the multiple imputations technique by Rubin (1987) to substitute for missing values in the variables due to item non-response (see Donzé, 2001 for a detailed report on these imputations). The estimations were based on the mean of five imputed values for every missing value of a certain variable. To test the robustness of this procedure we estimated the basic model for the original data without imputed values for every single set of imputed values as well as for the mean of them; finally, we calculated the mean and the variance of the parameters of the estimates based on the single five imputed values according to the method described in Donzé (2001) and compared the results. This comparison showed relatively high robustness of the estimated parameters. Spyros Arvanitis 143

5 Model specification and variable construction

In this study we considered three categories of employees: highly educated (employees with education at the tertiary level, including universities, technical and business colleges (‘Fachhochschulen’)), middle-educated (employees with a formal degree in vocational education) and low-educated (employees with some vocational education but without a formal degree or without any formal vocational education). The employment share for each of these categories was used as a dependent variable in our model estimations (equations for employment share).7 Since we did not dispose of data on capital use costs and wage data for each education category, we relied upon extensive industry controls to ascertain the influence of this important variable.8 As measures for technology input, particularly ICT input (‘ICT capital’), we used the intensity of use of two important network technologies: the Internet (linking to the outside world) and the intranet (linking within the firm). This intensity was measured by the share of employees using the Internet and intranet respectively in their daily work. The firms were asked to report this share not by a precise figure but within a range of twenty percentage points (1 per cent to 20 per cent, 21–40 per cent and so on). Based on these data we constructed five dummy variables for each technology covering the whole range from one per cent to 100 per cent (see note to Table 6.3). The idea behind this variable is that a measure of the diffusion of a certain technology within a firm would be a more precise proxy for ‘ICT capital’ than the mere incidence of this technology or some kind of simple hardware measure (e.g. number of installed personal computers, etc.). We expect in general a positive correlation of technology variables with the employment share of highly educated employees, in particular an increasing positive correlation with a higher percentage of employees using a certain technology. A further hypothesis is that technology variables correlate negatively with the share of low-educated employees. We have no a priori expectation for the relation between technology variables and the employment share of middle-educated employees. The measurement of organizational inputs, here restricted to inputs related to workplace organization, is an issue still open to discussion, since there is no agreement as yet among applied economists over the exact definition of ‘organizational capital’ (see Black and Lynch, 2002 and Lev, 2003 for a discussion of this matter). In order to choose the variables related to changes and/or introduction and the use of new organizational practices at the workplace level we draw on the definition offered by Black and Lynch (2002). They distinguished two components of organizational capital (in a narrow sense, that is, without training, which we view as part of the human capital of the firm): ‘work design’ and ‘employee voice’. Examples of practices that are included in the first component are re-engineering efforts that may involve 144 Information Technology and Demand for Employees changing the occupational structure of the workplace, the number of levels of management within the firm, the existence and diffusion of job rotation, and job share arrangements. The second component of organizational capital, ‘employee voice’, is associated with practices such as individual job enrichment schemes, employees being consulted in groups, employees having more decision competences, the existence and diffusion of work in (formally constituted) teams, etc. Our data enabled us to construct the following dummy variables covering most of the above-discussed aspects of organizational capital: intensive use of teamwork (project groups, quality circles, semi-autonomous teams, etc.); intensive use of job rotation; a reduction in the number of management levels; an overall shift of decision competences from managers to employees; employees having the competence to determine the sequence of performing tasks and/or the way in which tasks are performed; investment decisions being often discussed in work teams (see also note to Table 6.3).9 We expect an overall positive correlation of organizational variables with average labour productivity, but we do not have sign expectations for every single variable. We include three more variables which are related to workplace organization, but are not components of organizational capital per se. The first one refers to incentive-based compensation and is a dummy variable for the existence of employee compensation according to team performance. The other two variables measure labour flexibility (dummy variable for the intensive use of part-time work) and working time flexibility (dummy variable for working time considered flexible over the year). With respect to the compensation variable the sign of the correlation to the dependent variable is not a priori clear for middle- and low-educated employees; we expect that team compensation is considered as more adequate and/or is more often used for higher- than lower-qualified employees. In the empirical literature the relation between part-time work and education level of the employees is unclear and depends upon the overall conditions of the labour market as well as its institutional framework. We expect a positive effect for flexible annual working time for all three employee categories; this type of labour flexibility not only expands employee time sovereignty but also contributes to a more efficient combined use of labour and equipment. We use extensive controls for firm size (5 dummies) and industry affiliation (27 dummies) to account for firm- or industry-specific influences not taken explicitly into consideration. In a second model version we constructed composite indices for the ICT and the organizational capital. These indices were calculated as a sum of the stardardized values (average 0; standard deviation 1) of the constituent variables. For the technology variable (TECHNS) the original variables for the use of the Internet and intranet (measured on a five-point Likert scale) were used for the standardization procedure; the organizational variable (ORGANS) was constructed as a sum of the stardardized values of Spyros Arvanitis 145 the seven constituent variables.10 The use of these composite variables as overall measures of technology and organization enabled us not only to assess the relative importance of these factors with respect to employee education, but also to test the possible endogeneity of the organizational factors as well as the postulated complementarity between technology and organization.

6 Results of the model estimations

Table 6.3 contains the OLS estimates of the full model for the employment shares of the high-, middle- and low-educated employees.11 Since these results are only cross-section estimates, it is not possible to test directly the existence of causal relations between the independent variables and the dependent variable. Nevertheless, some robust regularities come out, which if interpreted in view of our hypotheses 1 to 3 (see section 2) could possibly indicate the direction of causal links.

6.1 Technological factors The coefficients of four of the five dummy variables for the intensity of use of the Internet are, as expected, positive and statistically significant in the equation for the employment share of highly educated employees, negative and statistically significant in the equation for the low-educated employees; for the category of middle-educated employees these four coefficients are positive but not statistically significant (at the test level of 10 per cent). Only the coefficient for the lowest intensity category of the Internet (1–20 per cent of employees using the Internet in their daily work) is not significant in any of the three equations. The general tendency is that the higher the intensity of use of these technologies among employees, the higher (lower) in case of the highly educated (low-educated) employees is also the positive (negative) correlation to employment share. The same effects are also found qualitatively for the intensity of use of the intranet. In this case the positive correlation to the share of the highly educated employees seems to be weaker than in case of the Internet; only the coefficients for the variables for high intensity of intranet use (over 60 per cent of employees using this technology in their daily work) are statistically significant. The estimates for the composite variable TECHNS (instead of the detailed information on Internet and intranet use) in Table 6.4 show once more clearly that the technological factors correlate positively with the share of highly educated employees and negatively with the share of low-educated employees; technology is not discernibly related to the employment share of the middle-educated employees. On the whole, the results for the variables for the ICT use are in reasonable accordance with hypothesis 1 of skill-biased technological change. 146

3 Low-educated 2.440* (1.376) 0.039 5.423** (2.462) 0.086 9.691*** (2.845) 0.106 6.251*** (2.197) 0.100 4.233** (2.196) 6.727*** (2.438) 0.072 0.088 3.962 (2.875)3.962 0.032 0.907 (2.319)0.907 0.013 − − − − − − 14.598*** (2.431) 0.186 13.500*** (3.062) 0.096 10.607*** (2.911) 0.093 −

− −

47.909*** (4.028) 2 0.009 0.149 (2.301) 0.003 0.009 0.149 0.074 0.001 0.012 0.003 0.004 0.023 −

(2.047) Middle-educated (2.252) (4.483) (2.801) (3.261) (2.611) 0.801 (1.338) 0015 0.630 0.241 0.099 0.412 (2.085) 0.026 (2.136) 0.000 2.838 (2.959) 0.027 0.372 1.539 4.040**

− − − −

1 0.101 0.150 0.0759 0.118 0.003 0.233 0.036

(2.699) 0.150 (2.108) (1.282) (1.003) (1.897) (2.713) (1.685) (1.126) (1a) (1b) (2a) (2b) (3b) (3a) Highly Highly educated 0.096 (1.062) 0.002

7.759** (1.858) 41.162*** (3.595) 3.386** (1.456) 0.063 3.203 (2.307) 0.048 1.593 (1.164) 0.038 3.291 (2.091) 0.064

− 4 4 0.222 5 3.344*** 5 Full model: determinants of the employment shares of highly educated, middle-educated and shares highly employment employees, of low-educated 1999 the of model: determinants Full : Job rotation 21–40 4.491*** : Workplace organization : Team-work 41–60 9.746*** (% of employees): Use of intranet 1–20 0.134 (1.043) 0.004 (1.043) 1–20 0.134 Intercept Table 6.3 Table (OLS estimates) variables Explanatory 1–20 21–40 1.603 61–80 9.599*** 81–100 15.015*** Technology (% of employees) Use of internet 41–60 61–80 81–100 13.047*** 147 0.247 (1.446) 0.004 0.651 (1.176) 0.012 2.441 (1.531) 0.036 1.992* (1.115) 0.039 4.878*** (1.458) 0.074 0.691 (1.111)0.691 0.012 1.781 (2.758) 0.013 2.934 (2.477) 0.024 4.940*** (1.545) 0.079 3.047**(1.490) 0.052 4.009** (1.731) 0.056 − − − − −

1.724 (1.441) 0.030 0.962 (1.097) 0.022 4.849* (2.484) 0.042 1.283 (1.220) 0.027 3.692**(1.263) 0.065 1.010 (1.167) 0.021 3.996* (2.419) 0.039 5.231*** (1.479) 0.096 0.439 (1.077) 0.010 1.719 (1.478) 0.034 3.245* (1.686) 0.052 − − − − − − − − − − 0.046 0.051 (0.841) (0.758) 0.328 (0.784) 0.009 2.136 (1.943) 0.027 0.671 (2.676) 0.007 2.743 (2.360) 0.025 2.111** (0.877) 0.052 0.106 (0.987) 0.003 2.622** (1.108) 0.054 0.496 (1.577) 0.008 1.932* (1.079) 0.041 2.137** (1.050) 0.046 2.810** (1.419) 0.049 (1.050) 0.046 2.810** 2.137** 0.805 (1.279) 0.016 1.332 (1.032) 0.030 2.102 (2.119) 0.025 1.721 (2.328) 0.018 −

− −

− 9 : :

6 1.860**

1.804** 11 8 10

7 : delegation of competences competences from of delegation of performing tasks workteams managers to employees: to managers performing tasks performing managers to employees to managers Employees competence for sequencethe Investment decisions are discussed in Compensation, working conditionsCompensation, Team compensation Flexible working time working Flexible Part-time work Delegation of competences from of competences Delegation Employees competence for way of the >999 employees 200–499 employees Firm size Firm 50–99 employees 100–199 employees Decrease of number of managerial levels managerial of Decrease number of Overall 500–999 employees 148 (the 3 is ‘important’ is ‘important’ managerial levels education (formal education Low-educated middle a Likert five-point scale)); b) contain the absolute values absolute the contain b) team performance team 0.307 sequence of performing tasks d from managers to employees); to managers d from 1544 52 21.023 enote statistical significance at the 2 ) is between 1% and 20%, 21% and 40%, 41% and 60%, and 41% 40%, and 21% 20%, 1% and between is ) part-time work (flexibleannual working time) is ‘widespread’ (values 4 and 5 on a five-point Likert scale)); Likert five-point a on 5 and 4 (values ‘widespread’ is Middle-educated job rotation internet (intranet use internet (intranet)); 5. dummy variable (value 1 for firms reporting that not 1 5.8*** 14.1*** formal education (vocational education without a formal degree; no vocational education); vocational no degree; formal a without education (vocational education formal low (1a) (1b) (2a) (2b) (3b) (3a) 0.372 0.140 0.372 52 52 52 14.215 20.338 14.215 1544 1544 1544 formal education (university, technical and business colleges); 2. share of employees with of employees share 2. colleges); business and technical (university, high education formal ) (values) 4 and5on five-point a Likert scale));9.dummy variable (value 1 for firms reporting that thenumber of (project groups, quality circles, semi-autonomous teams, etc.) or or etc.) teams, semi-autonomous circles, quality groups, (project (Continued) teamwork : 1. Share of employees with with of employees Share : 1. adj. 2 6. dummy variable (value 1 for firms reporting that in the period 1995–2000 (not further specified) competences were transferre were competences specified) further (not 1995–2000 period the in that reporting firms for 1 (value variable 6. dummy 7. dummy variable (value reporting that 1 for firms investment decisions discussed ‘often’ are work teams in (values 4 and 5 on (values4 and5five-point on a Likert scale)); 11. dummy variable (value 1 for firms reportingthat the use of use of the the determine to competence the have level employees at workplace the that reporting firms 1 for 8. (value dummy variables 4. dummy variables (value 1 for firms reporting that the share of employees using degree in vocational education); 3. share of employees with of employees 3. share education); in vocational degree way tasks are performed N to according compensation that employee reporting 1 for firms (value variable dummy 10. 1995–2000); period the in decreased Table 6.3 Table Notes Explanatory variables educated Highly (values 4 and 5 on a five-point Likert scale)); estimations include also two-digit industry controls (27 dummies); ***, **, * d * **, ***, dummies); (27 controls industry two-digit also include estimations scale)); Likert five-point a on 5 and 4 (values (3 and (2b) (1b), columns coefficients, original the contain (3a) and (2a) (1a), columns the respectively; level 10% and 5% 1%, procedure). (White errors standard heteroscedasticity-robust coefficients; standardized the of 61% and 80%, 81%and 100% respectively; reference group: firms which do DF SER F 18.6*** R Spyros Arvanitis 149

6.2 Organizational factors Job rotation and the decrease of the number of managerial levels, both of them associated to the ‘work design’ component of organizational capital, do not seem to be significantly correlated to the employment mix with respect to different education levels. In all three equations the coefficients of these two variables are not statistically significant. Closely related to employees’ education mix are various dimensions of ‘employee voice’. Teamwork is positively (negatively) correlated to the employment share of highly educated (low-educated) employees; we could not find a statistically significant relation of the variable for this type of organizational practice to the share of middle-educated employees. Further, we obtained statistically significant positive coefficients for the variables for high employee competence to determine the sequence and the way of performing tasks, and also for the variable for frequent discussions of investment decisions in workteams in the equation for the share of highly educated employees. On the contrary, we found statistically significant negative coefficients for teamwork and high employee competence for the way of performing tasks in the equation for the share of low-educated employees. For one organizational variable, high employee competence for the way of performing tasks, the estimates yielded a significant positive coefficient also in the equation for the share of middle-educated. There is no indication of significant effects for the overall delegation of competences from managers to employees in any of the three employment share equations. On the whole, teamwork in combination with the decentralization of important decisions (sequence and way of performing tasks) and increased participation of employees to high-level decisions related to investment are organizational measures which are realized in firms with relatively high shares of highly educated employees. The estimates for the composite variable ORGANS in Table 6.4 summarize the results with respect to the organizational factors. There is clearly a positive effect for the highly educated employees and a negative (but not statistically significant) effect for the low-educated employees. In the case of the middle- educated employees no significant effect could be found.

6.3 Endogeneity of organization Organizational changes may be endogenous, being introduced as a adustment needed to cope, for example, with technological changes.12 The construction of a composite variable for the organizational factors facilitates considerably the investigation of the important question of the endogeneity of organization (see Athey and Stern, 1998). It is, of course, not possible to settle definitely this matter based only on cross-section data, but some hints with respect to the robustness of the cross-section estimates can be gained through 2SLS estimates of the employment share equations. In the first stage the variables ORGANS was instrumented, the first stage estimates are shown in column 4 of Table 6.5. 150 2.547*** (0.396) 0.172 0.315 (0.252) 0.037 0.353*** (0.126) 0.077 1.017 (1.182) 0.018 2.335** (1.152) 0.045 − − − −

− 0.030 (0.389) 0.188 (0.137) 0.047 1.601 (1.151) 0.033 1.462 (1.215) 0.030 (1.309) 3.945*** 0.070 0.274 (1.075) 2.021 (1.478) 0.006 0.040 (1.509) 4.019*** 0.069 5.674*** (1.457) 0.104 6.761*** (1.533) 0.108 4.572* (2.342) 0.045 (ou2.624) 4.119 0.035 3.582** (1.663) 0.057 (1.777) 5.529*** 0.077 5.025** (2.509) 0.044 2.126 (2.780) 0.016 − − −

− − − − − − 0.207 0.387 0.165 (0.266) (0.106) 0.321* (0.1449) 0.053 0.102 (0.239) 0.014 1.223 (2.258) 0.015 1.295 (2.326) 0.014 2.668*** (0.866) 2.668*** 0.068 2.023*** (0.791) 2.023*** 0.055 2.221** (0.917) 0.055 0.745 (1.031) 0.018 0.046 (1.071) 0.001 0.507 (1.316) 0.010 − − − − : : 0.540*** 3 Determinants of the employment shares of highly educated, middle-educated and low-educated employees, 1999 (OLS estimates of estimates 1999 (OLS employees, low-educated and middle-educated educated, of highly shares employment of the Determinants

2.186*** 2 1 : ORGANS TECHNS*ORGANS a version of modelthe with composite indices for technology and organization based on stardardized values) Explanatory variables Intercept TECHNS educated Highly (1b) (2a) (2b) (1a) (3a) (3b) 11.193** (1.735) Middle-educated 43.262*** (2.836) Low-educated 42.300*** (3.287) Table 6.4 Table Compensation, working conditionsCompensation, Team compensation Part-time work Flexible working time size Firm 50–99 employees 100–199 employees 200–499 employees >999 employees 500–999 employees 151 es; estimations also ive-point Likert scale); Likert ive-point ; decrease of the number the of ; decrease eriod 1995–2000; employ- tively; the columns (1a), (2a) (1a), columns the tively; , investment decisions are dis- are decisions , investment ients; heteroscedasticity-robust 0.268 15.9*** 38 21.604

way tasks are performed andthe 0.140 1544 1544

sequence sequence performingof tasks : 1. Sum of the values of the stardardized variables valuesstardardized 1.: of the intensity of Internet Sum for(two variables and intranet user measuredon a f adj. 0.322 2 2. teamwork job rotation; for: variables dummy (seven organization workplace for variables standardized the of values the Sum of p in the employees to from managers competences (unspecified) of transfer overall 1995–2000; period the in levels of managerial the determine to level competence the at workplace the ees have cussed in work-teams); 3. Interaction term of ORGANS andcussed TECHNS; see notes to Table 6.3 for definition the of otherall variabl include respec level and 10% 5% 1%, significance the at statistical denote * two-digit **, ***, dummies); (27 controls industry and (3a) contain theoriginal coefficients,columns (1b),(2b) and (3b) containthe absolutevalues of the standardized coeffic procedure). (White errors standard DF 38 38 38 DF 38 N 1544 Notes F 20.3*** 7.6*** 7.6*** 20.342 SER 14.768 F 20.3*** R 152 (0.176) (0.176) (0.209) (0.282) (0.282) 0.092 0.092

0.713*** 0.713*** 0.826*** (0.159) (0.159) 0.826*** 5.06*** 5.06*** 0.115 (0.195) (0.195) 0.115 2.781 2.781 0.399*** (0.152) (0.152) 0.399*** 0.169 (0.171) (0.171) 0.169 0.401* (0.235) (0.235) 0.401* 0.857** (0.360) 0.857** 0.030 (0.208) (0.208) 0.030 0.502 (0.404) (0.404) 0.502 0.705* (0.420) (0.420) 0.705* − First stage estimate ORGANS (4)

1535 1535 38 ); 2. exports as a share of as a share exports ); 2. 5 for the definitions of all other 1%,5% and 10% level respectively. 3.886*** (1.367) 3.886*** 2.546*** (0.379) (0.379) 2.546*** 0.149 (1.622) (1.622) 0.149 (0.108) 0.413*** 1.128 (1.930) (1.930) 1.128 2.478* (1.363) (1.363) 2.478* Low-educated 2SLS (3) estimate − 0.441**

0.359 3.916** (1.545) (1.545) 3.916** − − − 5.402*** (2.026) (2.026) 5.402*** 42.391 (3.102) (3.102) 42.391 − 3.803 (3.264) (3.264) 3.803 6.867*** (1.627) (1.627) 6.867*** 2.050 (3.287) (3.287) 2.050 and low-educated employees (2SLS estimates of estimates (2SLS employees and low-educated (0.388) (0.388) 0.472 0.472 0.119 0.264 0.119 1.193 (1.975) (1.975) 1.193 6.5*** 15.5*** 15.5*** 6.5*** 1.089 (1.395) (1.395) 1.089 3.040* (1.659) 3.040* (0.111) 0.157 1.232 (1.580) (1.580) 1.232 5.292*** (1.665) (1.665) 5.292*** 1.772 (2.073) (2.073) 1.772 0.788 (1.398) (1.398) 0.788 1.161 (3.340) (3.340) 1.161 3.104 (3.362) (3.362) 3.104 42.700*** (3.174) (3.174) 42.700*** Middle-educated 2SLS Middle-educated (2) estimate

− − − − − − − 38 38 38 − 22.161 21.663 21.663 22.161 1535 1535 1535 (0.362) (0.103) (1.549) 1.847 (1.475) 1.847 0.705 (1.554) 0.705 2.022 (1.843) 2.022 3.391* (1.934) (1.934) 3.391* 3.194** (1.305) 3.194** 0.116 (1.302) 0.116 4.168 (3.117) 4.168 1.908 (3.138) 1.908 11.941*** (2.962) (2.962) 11.941*** estimate (1) estimate − Highly-educated 2SLS Highly-educated − − − − − − −

2 3 Determinants of the employment share of highly educated, middle-educated educated, middle-educated share of highly of the Determinants employment : 1. Dummy variable (value 1 for firms reporting the introduction of product and/or process innovations innovations in and/or process : period the 1998–2000 of product 1. Dummy (value 1 the introduction variable for firms reporting adj. 0.188 0.188 adj. 2 Intercept Intercept TECHNS 2.023*** 2.023*** TECHNS Firm size : employees 50–99 Export share Export Table 6.5Table variables Explanatory Notes sales; 3. dummy variable (value 1 for firms reporting that they belong to foreigna enterprise); notessee to Tables 6.4 and 6. the significance at statistical denote * **, ***, (27 dummies); controls industry two-digit also include estimations variables; ORGANS 5.455*** 5.455*** ORGANS Foreign firm Foreign TECHNS*ORGANS 0.596*** 0.596*** TECHNS*ORGANS logy and organization; variable ORGANS is instrumented) is instrumented) ORGANS variable and organization; logy for techno indices composite with version the model 100–199 employees employees 100–199 Team compensation 200–499 employees employees 200–499 Part-time work work Part-time Flexible working time working Flexible 500–999 employees employees 500–999 Innovative activities Innovative >9999 employees employees >9999 F 10.3*** 10.3*** F N 1535 N 1535 38 DF SER 20.684 20.684 SER R Spyros Arvanitis 153

As instruments we used besides the dummy variables for part-time work, annual flexible working time, team compensation and firm size three additional variables which were not included in the employment share model: the coincidence of innovative activities, the export share of sales and a dummy variable for foreign ownership. The overall statistical fit of the first stage estimates for ORGANS (R2 = 0.092) was rather poor. The 2SLS estimates in columns 1 and 3 of Table 6.5 for highly and low-educated employees respectively confirm the results of the model estimates without instruments, at least with respect to the signs of the coefficients. The coefficient of ORGANS in the equation for the middle-educated becomes negative and statistically significant at the test level of 10 per cent; in view of the rather poor performance of the instrument equation for ORGANS this result is not very robust. On the whole, hypothesis 2 (skill-biased organizational change) seems to be confirmed also by these results.

6.4 Complementarity effects The equations for the employment shares in Table 6.4 contain besides the composite variables TECHNS and ORGANS also the interaction term TECHNS*ORGANS. The coefficient of the interaction term in the equation for the highly educated employees is positive and statistically significant, in the equation for the low-educated employees negative and also statistical significant, in the equation for the middle-educated employees negative but statistically insignificant. The results can be interpreted as a hint for the existence of complementarity of ICT and workplace organization, which means that besides the direct effects of ICT and organization on the employment shares of highly and low-educated employees there are also indirect effects which can be traced back to the joint impact of these two factors on the employment shares. Hypothesis 3 receives some support from these results.

6.5 Other factors The variable for part-time work correlates positively with the share of highly educated employees and negatively with the share of low-educated employees; the variable for flexible working time also correlates with the shares of highly and low-educated but with exactly opposite signs (Table 6.3). These results reflect the relative importance of various dimensions of quantitative labour flexibility for different employee categories. Seemingly, part-time work is considered adequate primarily for low-educated employees, flexibility of working time primarily for highly educated. Also compensation according to team performance is relevant only for the highly educated employees. Compensation and working conditions are not related to the employment share of middle-educated. 154 Information Technology and Demand for Employees

Finally, firm size correlates negatively with the share for middle-educated and positively (up to 500 employees) with the share of low-educated employees; it does not correlate at all with the share of highly educated employees. The larger a firm is, the higher is also the share of low-educated (up to 500 employees)and lower the share of middle-educated employees.

7 Comparative evidence and survey of empirical literature

A comparison with other similar studies (see Table 6.6)13 shows that there are considerable differences among the firm samples of the countries under scrutiny. The German studies are particularly relevant in such a comparison because they also cover the case of middle-educated employees possessing a vocational degree from the ‘dual apprenticeship system’ which is similar to the system which operates in Switzerland. Positive effects for both the technological and the organizational factors for the highly educated are also observed in all surveyed German studies. Interaction terms for organization and technology were investigated in some cases but no statistical significant effects could be identified. The results for the middle-educated employees were mixed for both the technological and the organizational factor. On the whole, it seems that the employment of the German middle-educated employees is stronger affected from technological and organizational change than the employment of middle-educated employees in the Swiss economy (see also Jacobebbinghaus and Zwick, 2001). For low-educated employees, in most studies the empirical findings are similar to the results for Swiss firms but the effects, particularly those associated with the ICT use, seem to be weaker than those for Switzerland. Interaction terms were in most cases statistically insignificant. In the case of French and Italian firms the influence of technology on the employment shares of employees with different skills or education is less important than that of organization. Most studies found no discernible effects of technology. With respect to organizational factors there is a tendency for a positive impact for the highly educated and a negative one for the low- educated employees, in accordance with the Swiss results. Interaction terms were in some cases significantly positive for the highly educated and negative for the low-educated employees. Studies for the USA and the UK demonstrate clearly the expected effects of technology on the employment of highly and low-educated employees respectively. The impact of organization is less clear that that of technology, but in most cases as expected. On the whole, the results are indicative but not completely comparable because some of the observed differences can be traced back to differences with respect to the sectors and industries covered in the studies, the specification of the organizational variables and the nature of the investigations (cross-sectional versus longitudinal approach). 155 robust n.c. n.c. Complementarity

not n.c. n.c. ICT; TECH/ORG positive positive n.s. n.s. n.s. not robust not robust n.c. not robust not robust n.c. n.c. n.c. n.c. n.c. 1 1 1 1 pos./neg. n.s. n.s. pos./neg.

n.s./neg. pos./neg. n.s. negative negative robust not negative positive n.s. n.s. n.s. n.s. n.s. n.s. n.s. positive positive pos./neg.

n.s. n.s.

positive positive ICT; TECH ORG n.s. n.s. positive n.s. negative n.s. : : : positive positive positive : : workers positive workers manuals n.s. n.s. manuals workers n.s. workers managers/technicians managers/technicians manuals n.s. manuals managers n.s. n.s. managers unskilled manuals manuals unskilled average wages of human capital investmenthuman capital changes in the wage shares of bill changes in the wage shares of bill semi-skilled manualssemi-skilled clerical workers clerical managers/technical staff skilled manuals skilled Dependent variable : : : : Survey of recent empirical literature literature empirical recent of Survey supervisors supervisors managers/professionals managers/professionals senior senior unskilled skilled clerical middle office workers positive workers office technical workers positive workers technical USA: USA: (2000) Capelli/Carter longitudinal – Table 6.6 Table Study production Bresnahan et al. (2002) – cross-section UK: Caroli/Van Reenen (2001) longitudinal – positive supervisors/foremen France: Caroli/Van Reenen (2001) longitudinal –

156

ICT; TECH/ORG n.s. n.c. n.s. n.c. n.c. n.s.

negative n.c. positive n.c. 2 n.s. n.s. n.c. n.c. n.s. n.s. n.a. n.s. n.c. n.c. n.s. n.a. n.s n.s. n.c. n.c. n.s. n.s n.s. negative n.c. n.s. negative n.s. positive n.c. n.s. positive n.c. negative negative n.c. negative n.a.

pos./neg. n.s. : : : : management negative negative management unskilled blue workers growth rate of employment shares of employmentof share skilled blue workers probability of employment increase for : employment shares of : : : (Continued) university graduates positive positive n.c. positive positive graduates foremen/technicians university workers with vocational degree positive n.s. n.c.

middle management n.s. n.s. management – cross-section middle clerks Germany: Gerlach/Jirjahn (1998) longitudinal – skilled blue workers unskilled blue workers n.s. n.s. negative n.s. n.c. n.c. executives executives clerks clerks executives executives middle Table 6.6 Table Study Dependent variable ICT; TECH ORG Complementarity intermediate workers n.a. negative n.c. n.c. negative n.a. workers Caroli et al. (2001) : longitudinal – managers intermediate operatives Greenan (2003) – cross-section 157

n.s. n.s. n.s. n.s.

n.c.

n.s.

n.s. n.s. n.c.

negative n.s. n.s. negative negative n.s. n.s. negative n.s. n.s. n.s. n.c. n.s n.s n.s. n.c. positive n.s. n.s. n.s. negative n.s. n.s. negative

n.s n.s.

n.s n.s n.s n.s. negative n.s. n.s. negative positive

negative negative negative negative n.s. n.s. negative n.s. n.c.

n.s n.s

n.s. n.s. n.s n.s positive positive n.c. positive positive : : : : : : : : workers: workers: workers: workers: workers: workers: workers: workers: degree n.s. degree graduates positive graduates workers n.s. workers skilled positive skilled unskilled unskilled n.s. skilled skilled negative skilled n.s. unskilled n.s. skilled n.s. unskilled n.s skilled n.s unskilled unskilled n.s. probability of employment increase for employmentof share – highly skilled skilled highly – – highly skilled skilled highly – average wages of – highly skilled skilled highly – – highly skilled skilled highly – employmentof share : Falk (2002) : – cross-section university masters/technicians vocational highly low-skilled – – unskilled Hujer et (2002) al. : longitudinal – – – white-collar white-collar

– – blue-collar blue-collar

– – – white-collar white-collar – – blue-collar blue-collar – longitudinal longitudinal – Bauer/Bender (2001) 158 ICT;TECH/ORG positive negative negative egative coefficients; ICT: the test level of 10%) positive 10%) of level test the he models,he but the resultsarenot tistically not significant (at the test the (at significant not tistically n.s. negative negative

: workers n.s. workers workers n.s. n.s. workers log log of difference of the number of : (Continued) : 1. Positive: teamwork, reduction of management levels, regular meetings; negative: job rotation; 2. Partly positive, partly n blue-collar blue-collar white-collar – cross-section information and communication technologies; ORG: workplace organization; ‘positive’ (‘negative’): statistically significant (at sta n.s.: respectively; variables two these of term interaction the ORG and for ICT, variables(s) the of coefficient (negative) level of 10%); n.c.: not considered; n.a.: not available (for such cases in which the corresponding variables are included in t explicitly presented). Italy: Piva et(2003) al. Table 6.6 Table Notes Study Dependent variable ICT; TECH ORG Complementarity Spyros Arvanitis 159

8 Summary and conclusions

In this section the results are summarized and some implications for economic policy are briefly discussed; of course, the precaution with respect to the interpretation of cross-section results mentioned in section 6 above still holds. The technological changes of the last ten to fifteen years, particularly those driven by the increasing adoption and use of ICT, correlated significantly with the rising skill and education requirements of Swiss enterprises. The technological factors contributed to a clearly higher employment share for highly educated employees and an equally clear lower employment share for low-educated employees. The employment share of middle-educated employees seems to be unrelated to the extent of intra-firm ICT use. These results are in agreement with the hypothesis of skill-biased technical change, also with the results of similar studies for other high-tech OECD countries. Changes of work organization (particularly teamwork and the decentralization of decision making) also contributed to an increase in the demand for highly educated employees and a decrease in the demand for low-educated employees, but these effects were weaker than those of the technology factors. Even if the results are not so robust as in case of technology, there is a discernible tendency in favour of the hypothesis of skill-biased organizational change; the estimated effects of organization were considerably weaker than those found for example in French studies. A further finding was that the technology and organization effects strengthen each other (positive joint effect). This means that besides the direct impacts the changes of the employment shares of employees with different education levels were also driven by the indirect impacts of the combined use of technology and organization. Middle-educated employees were only slightly, if at all, affected from these changes. It is not clear if this will remain so in case of a further acceleration of skill-biased technical change in the near future. The main shortcoming of this study is that no data were available for a longitudinal study which would allow us to take into consideration possible lags between the variables and to test causal relationships between the explanatory factors and firm performance. We hope to be able to repeat the 2000 survey in the near future, so that data for an additional time point would become available. Are there any implications of the results of the study for economic policy? In general, policy makers have to pay special attention to the conditions favouring the formation and growth of human capital in the economy. This does not mean that only tertiary education should be promoted. The system of the ‘normal’ vocational education (‘Berufslehre’) – which is one of the two pillars of the Swiss ‘dual’ education system – has to be (further) upgrated, especially with respect to the content of education. Most of the reforms which 160 Information Technology and Demand for Employees took place during the 1990s were well-targeted to cover the additional skill requirements of the business sector: upgrading of technical and business colleges other than the classical universities (‘Fachhochschulen’), transparency with respect to the access conditions to the tertiary education system, upgrading of specific vocational degrees in technical and business professions, etc. Further, the existing – and, in international terms, rather high – flexibility of the labour market has to be maintained, as does the current legal framework which permits employers to form a flexible work organization. A more difficult problem is to offer employment perspectives to low-educated persons. Additional education and/or vocational training is one way of trying to tackle this problem, but it is not a way that is accessible to all involved persons, particularly not for older ones. For such cases social partners and policy makers have to coordinate efforts for specific solutions aiming at the social integration also of this category of employees.

Notes

1. Yet it is also true that an increase in the supply of skills could also induce an acceleration of technological change; it is mostly assumed that in the long run the driving force from the demand side is the dominant one (see e.g. Acemoglu, 1998); the empirical evidence seems to support this notion. 2. Some of the most influential empirical studies on this subject in the last few years are the following: for the USA: Krueger (1993); Doms et al. 1997; Autor et al. (1998); Chun (2003); for the UK: Chennels and Van Reenen (1999); Haskel and Heden (1999); for Germany: Kaiser (1999); Falk (2001); Falk and Seim (2001); for France: Goux and Maurin (2000); comparative studies for several countries: Berman et al. (1998); Machin and Van Reenen (1998). For recent surveys of the theoretical and empirical literature on skill-biased technical change see Sanders and ter Weel (2000) and Acemoglu (2002). 3. The main hypothesis is that the accelerated growth of world trade and foreign direct investment leads to a new international division of labour: the production of goods (and services) with a high content of low-skilled labour is dislocated to developing countries of the South, while activities with a high content of high-skilled labour are concentrated in the developed countries of the North. For an exploratory study on this subject for Switzerland at firm level see Arvanitis and Donzé (2000). 4. Recent Swiss studies on the influence of the skill or education level of workers on economic growth on an aggregated level are found in Kugler and Spycher (1992) and Sheldon (2003), the latter dealing particularly with the role of foreign workers; the essay of Sacchi et al., chapter 5 in this volume, is dealing with the long-term development of employers’ demand for skills in the period 1950–2000; an important finding of the study is that the changing organizational population, which is only partly attributable to the tertiarization of the economy, is the major trigger for the continuously increasing skill level in the overall economy; the use of ICT plays also a role in this process but a less important one. 5. The questionnaire was based to a considerable extent on similar questionnaires used in earlier surveys (see European Foundation, 1997; Francois et al., 1999; Vickery and Spyros Arvanitis 161

Wurzburg, 1998; Canada Statistics, 1999). Versions of the questionnaire in German, French and Italian are available in www.kof.ethz.ch. 6. The figures presented in Tables 6.1 and 6.2 in section 3 were based on a sample of 2,648 firms with at least 5 employees. The information on organization was raised only for firms with at least 20 employees (sample of 1,667 firms). As a consequence, we could only use the data for 1,667 firms for the econometric analysis. 7. An earlier version of this chapter also contained estimates for the change of the employment share of each of these categories (see Arvanitis, 2003). 8. Capital use costs vary little between industries. In case of wage data and their use as factor prices a further problem is their endogeneity which could cause serious econometric problems; some authors even proposed to omit altogether wage data in a one-equation framework (see e.g. Machin and Van Reenen, 1998). 9. Nonetheless, the conceptualization as well the measurement of workplace organization remains somewhat blurred; more elaborated concepts and more detailed information should be pursued in future work. 10. We also obtained similar results with a further version of a composite variable for organizational capital based on the factor scores of a one-factor solution of principal component factor analysis of the seven variables for workplace organization. 11. There are 84 observations with the value 0 in case of the highly educated, 18 observations in case of the middle-educated and 139 in case of the low-educated employees. Tobit estimates taking account of this peculiarity of the data yielded almost identical results with the OLS estimates for all three dependent variables. 12. In accordance to most economic literature we deal here only with this possibility of endogeneity; in this context technology is considered to be exogenous to the firm. 13. The choice of the studies reported in table 6 was based on following criteria: recent date of publication, consideration of both variable blocks technology and organization in the model specification, firm-level analysis, coverage of all sectors of the economy.

Appendix

Table A6.1 Composition of the data set

N Percentage

Industry: Food, beverage 69 4.5 Textiles 27 1.7 Clothing, leather 15 1.0 Wood processing 19 1.2 Paper 27 1.7 Printing 57 3.7 Chemicals 56 3.6 Plastics, rubber 31 2.0 Glass, stone, clay 31 2.0 Metal 17 1.1 Metal working 120 7.8 Machinery 137 8.9 Electrical machinery 37 2.4 Electronics, instruments 83 5.4 162 Information Technology and Demand for Employees

Table A6.1 (Continued)

N Percentage

Watches 27 1.7 Vehicles 17 1.1 Other manufacturing 33 2.1 Energy, water 27 1.7 Construction 169 10.9 Wholesale trade 162 10.5 Retail trade 94 6.1 Hotels, catering 37 2.4 Transport, telecommunication 70 4.5 Banks, insurance 60 3.9 Real estate, leasing 4 0.3 Computer services 22 1.4 Business services 88 5.7 Personal services 8 0.5 Firm size: 20–49 employees 495 32.1 50–99 employees 375 24.3 100–199 employees 311 20.1 200–499 employees 221 14.3 500–999 employees 77 5.0 >1,000 employees 65 4.2 Total 1544 100 7 Young Adults Entering the Workforce in Switzerland: Working Conditions and Well-Being Norbert K. Semmer, Franziska Tschan, Achim Elfering, Wolfgang Kälin and Simone Grebner

The ‘ÆQUAS’ study is about young adults entering the workforce after vocational training in five different occupations. Participants reported data on working conditions and well-being in the last year of vocational training and after one, two, and four years of professional experience. Both with respect to well-being and working conditions, the transition and the first years in the job were characterized by considerable stability; where there was change, it was mostly for the better. There were, however, differences between occupations. For instance, nurses reported comparatively high levels of stress, low levels of job control and high social support. Changing employer showed, on average, positive effects. Change of profession, however, often resulted in better working conditions but indicators of well-being remained unchanged. Besides levels of stress, social support, and job control, appreciation received at work emerged as a key factor for occupational health. The overall pattern of results indicates that the Swiss system of vocational training prepares well for the transition into ‘normal’ work.

1 Introduction

The transition into work is an important period for young adults, and it typically involves more than simply starting a job. For instance, there is the development of independence from one’s parents, and establishing new social relationships, including intimate ones. In other words, this period poses a number of ‘developmental tasks’ (Havighurst, 1972) for young people. It has therefore often been assumed that this period is a stressful one, characterized by ‘turmoil’ and instability (Ashforth and Saks, 1995; Brett, 1980; Louis, 1980; Nelson et al., 1995; Nicholson, 1984; Quick and Quick, 1984). However, empirical research has mostly not confirmed this expectation (Arnold, 1997). One of the factors that is likely to ensure a rather smooth

163 164 Young Adults Entering the Workforce in Switzerland transition is the system of vocational training that exists in a number of European countries, including Switzerland. This consists of a legally regulated status of apprenticeship, which takes place in what is called a ‘dual system’: people go to work for three or four days per week, and attend a vocational college for one or two days. The employer is obliged to give them an opportunity to learn things according to a plan that is legally binding. At the vocational college, theoretical backgrounds are taught, in addition to more basics skills like languages. After the training period, there is an exam, which consists of both a theoretical and a practical element. Having passed that, people receive a diploma that recognizes them as a ‘licensed salesperson’, ‘licensed bank accountant’, or the like. This system of vocational training thus provides a ‘bridge for non-college youths between the prevocational and occupational life course’ (Heinz et al., 1998, p. 77; cf. Kälin et al., 2000). The project ‘Work Experience and Quality of Life in Switzerland’ (called ‘ÆQUAS’, which is an acronym for the German title) has investigated the transition from vocational training into ‘real’ work (see Kälin et al., 2000). In addition, we followed the young people during the first three years of their professional experience. The focus of the study was on the type of work characteristics they report, the quality of their relationships, and their well-being, as well as personal characteristics, such as the methods they adopt in coping with stress. By contrast, much of the existing literature on transition into work has often focused on changes in work values and motives (e.g., Blickle, 1998; Lang, 1994; Lebo etal., 1995; O’Brien etal., 1994; Rosenstiel and Nerdinger, 2000), interests (e.g., Schmidt, 1984), worries or hopes concerning the future (e.g., Gallagher et al., 1992; Nurmi et al., 1995), the timing and patterning of transitions (Heinz et al., 1998; Marini, 1987; Reitzle et al., 1998) or career-enhancing strategies and role innovation (Feij et al., 1995; Nelson et al., 1988; Nicholson and West, 1988). We also have data on some of these variables, especially values and plans, but our focus in this chapter is more on stressful versus challenging experiences and the development of well-being. In this chapter, we will present an overview of important findings. After describing design of the study, we report data on working conditions and how they change over time, on well-being and how it changes over time, on the differences observed between the German- and the French-speaking parts of Switzerland, between males and females, and between people who stay with their employer and those who do not.

2 The study

The project ‘Work Experience and Quality of Life in Switzerland: Work, Stress, and Personality Development’ is a collaborative study conducted by five Swiss universities (the universities of Berne, Geneva, Lausanne, Neuchâtel, and the Swiss Federal Institute of Technology Zurich). The scope of the Semmer, Tschan, Elfering, Kälin and Grebner 165 project as a whole is quite broad, involving such issues as the socialization of values (Rimann et al., 2000) and selection effects (Elfering et al., 2004a), stability and change in work-related attitudes (Elfering et al., 2000), stress and health (Grebner et al., in press; Elfering et al., 2002), coping (Grebner et al., 2004; Kälin, 2004; Reicherts and Pihet, 2000), and the nature and quality of social relationships (Tschan et al., 2004).

2.1 Design and sample The study commenced in 1997, when participants were reaching the end of their vocational training. Waves two and three were conducted in 1998 and 1999. As our participants signalled that they were becoming tired of our questionnaires, wave 4 was conducted with a time lag of two years – that is, in 2001. Table 7.1 gives an overview of the project.

2.2 Sample bias and attrition One problem that plagues longitudinal studies is the issue of drop-out. As the study progresses, some people will decline to take part in further waves of measurement. This raises the question of whether those people that remain in the study (those we refer to later as ‘steady respondents’) are special. For instance, it could be that only those remain in the study that are doing well in life. If such differences between drop-outs and respondents are drastic, study results may be biased. It is therefore important to analyze differences between drop-out and steady respondents in order to assess any possible degree of bias.

Table 7.1 Project description

Project

‘Work Experience and Quality of Life in Switzerland’ (ÆQUAS)

Sample: Young people at the end of vocational training (mean age: 20.1 years; 59% female) Five occupations: Nurses, salespeople, electronic technicians, bank clerks, cooks German and French part of Switzerland Collaborative study of five Swiss Universities ‘Core project’: Whole sample, questionnaires Four waves of measurement: 1997, 1998, 1999, 2001 Number of people participating in all four waves of measurement: 423 Sub-projects: In-depth studies with subsamples, using diary methods (e.g., on stress, on social interactions), physiological assessments (e.g., stress hormones), job observation 166 Young Adults Entering the Workforce in Switzerland

These analyses show that the non-respondents were younger, and were disproportionately male, unmarried and living at home. Steady respondents reported higher task-related stressors and more freedom to decide about their own work (‘control’). At the same time, they reported lower social stressors. However, the differences are not extreme in most cases. For example, measuring social stressors on a scale that ranges from 1 (low) to 6 (high), drop-outs have a mean of 3.24, and respondents have a mean of 3.28. Although such a difference is larger than the ‘margin of error’, and therefore can be termed ‘statistically significant’, it is not very high numerically. The strongest differences are, however, found for occupations: Nurses show the highest (77.0 per cent), salespeople (35.1 per cent) and cooks (37.7 per cent) the lowest participation rates. With around 55 per cent, they are in between for electronic technicians and bank clerks (see Kälin et al., 2004). The most likely explanation for these occupational differences in the response rate is the level of education. This variable is often found to be associated with participation rates (Schnell, 1997). The length of vocational training (two years for salespeople, three years for cooks, and four years for the others) corresponds fairly well with the ordering of response rates, and the additional preparatory training of many nurses could explain the exceptionally high rate of response among nurses. Beyond differences in levels of different variables the most important issue from a research perspective is whether response bias affected the association between variables. As shown in Kälin et al. (2000), this was not the case: non-responders and responders do not show clear patterns of differences with regard to correlations between variables.

3 Results

The main focus of this overview is on the development of work characteristics and well-being in time. With respect to other important aspects of the project, we refer to work that is in progress, with special reference to an overview of findings by Kälin et al. (2004). Other publications focus on qualitative analyses of changes that people report (Elfering et al., 2004b), on the role of enduring (versus short-lived) unfavourable work characteristics (Semmer et al., 2004), and on the differences between the French and the German speaking part of Switzerland (Tschan et al., 2004).

3.1 Transition into work as a time of change As predicted, the transition into work is a period when many things change. Finishing vocational training implies a new status as a ‘licensed professional’. It often also means changing employer – in some cases because people want to leave, in many cases because the contract ends with the exam. Thus, more than half (56 per cent) change employer after the first year, and by the end of the study fully three-quarters of respondents have changed Semmer, Tschan, Elfering, Kälin and Grebner 167 employer at least once. Changes also occur in private life. In the first year, almost three-quarters (73 per cent) live at home with their parents. By wave four, this has dropped to somewhat more than one-quarter (28 per cent). Also, the percentage of people in a steady partnership rises from about one-fifth (19 per cent) to half of the sample (50 per cent).

3.2 Stability and change of working conditions The work situation is clearly characterized by more stability than change, and where changes occur, they are typically for the better (Figure 7.1). Thus, social stressors, that is, conflict with colleagues or supervisors, is rather low from the beginning and stays low throughout the five years. Values are around 1.95 on a scale that ranges from 1 to 5. Task-related stressors, a composite

Stable Changing

4.5 4.5 Appreciation received at work (***)

4.0 4.0 Social support at work (n.s.)

3.5 3.5 Job control (***) Task stressors (n.s.) 3.0 3.0 mean mean

2.5 2.5

Private stressors (***) Social stressors (n.s.) 2.0 2.0

1.0 1.0 t1 t2 t3 t4 t1 t2 t3 t4

Figure 7.1 Stressors and resources at work Note: For those who are interested in statistical details, we report here briefly what kind of analysis was run in obtaining the results presented. Stars in all figures denote differences that are statistically significant, that is, they are large enough to have confidence that they are not incidental. For all data presented Analyses of Covariance with time as a repeated measurement factor were performed. Stars that appear in parentheses behind the name of the dependent variable indicate a significant linear trend in time as tested with polynomial contrasts. Stars located beneath data points indicate significant differences between groups (deviation contrasts from mean), i.e. profession in Figures 7.5–9, gender in Figure 7.10 and Figure 7.11, cultural region in Figures 7.12 to 7.16, change of employer in Figure 7.17, and change of profession in Figure 7.18 and Figure 7.19. 168 Young Adults Entering the Workforce in Switzerland measure of time pressure, high concentration demands, a poor work organization, interruptions, uncertainty about tasks or feedback, are around 2.95, again on a scale from 1 to 5. This is at the higher end of the range typically found (see Kälin et al., 2000), but it is not drastically higher than in other studies, and – it hardly changes. Finally, social support at work shows values around 3.7, on a five-point scale. It is thus perceived as rather high, and it does not change significantly over time. Changes for the better are to be found with regard to decision latitude at work – what we term, control. Referring to the possibility of deciding for oneself how to carry out a task (method control) and how to organize one’s work day (time control), this aspect of work occupies a central role, as it has often been shown to be related to a higher degree of well-being (Frese and Semmer, 1991; Grebner et al., in press; Terry and Jimmieson, 1999). Lack of control has been found to predict the risk of cardiovascular diseases over time (Bosma et al., 1998), and we found, with data from this study, that lack of time control predicts the development of back pain in nurses (Elfering et al., 2002). Against this background, it is encouraging to see that control at work increases (from about 3.3 to about 3.6) over the four years of our study. The same positive picture can be said for recognition at work. Being valued and accepted, feeling appreciated and respected, is among the most important motives for most people (Epstein, 1998; Semmer and Jacobshagen, 2003; Semmer et al., in press). We therefore asked our respondents whether they where treated as apprentices vs ‘normal colleagues’ in the first year, and as ‘beginners’ vs ‘normal colleagues’ in the later years. The answers to this question are 3.9 in the first wave, and then rise constantly until wave 3, reaching a value of 4.3 on a five-point scale, where they stay. At this point, it may be sensible to clarify what is meant by the term ‘changes’ in this context. After all, moving from 3.9 to 4.3 with regard to ‘appreciation’ may seem a small change. If we look at what these developments mean in a more concrete way, however, these changes turn out to be quite significant. Thus, with regard to ‘recognition’, the average increase from 3.9 to 4.3 implies that the number of people that report their recognition to be ‘medium’ (3) or lower decreases from 86 to 62, or from 28 per cent to 16.9 per cent. At the other end of the scale, the number of people indicating that their recognition is ‘very high’ (5) almost doubles, increasing from 95 to 178, or from 30.9 per cent to 58 per cent. Figure 7.2 illustrates this finding. Put differently: If somebody said ‘medium’ in wave 1, that person was among the lowest 26 per cent in wave 1. A person saying ‘medium’ in wave four, however, was among the lowest 16%! Thus, these changes are more drastic than may at first appear. Overall, working conditions either remain stable (social support, task stressors, social stressors) or improve (control, appreciation). This picture does not support a picture of people in turmoil Semmer, Tschan, Elfering, Kälin and Grebner 169

150 150 144 Wave 1: Mean = 3.95 Wave 4: Mean = 4.33

125 125

102 100 100

80 75 75 63

50 45 50 Number of participants Number of participants

25 25 25 15 12 6 4 0 0 12345 12345 low high low high (n = 248)

Figure 7.2 Appreciation received at work but rather one of young people who have made a successful transition into the world of work. Of course, these are average values, which implies that there are individuals who do not conform to this picture. Nevertheless, the overall trend is quite clear. That this period of transition is associated with private changes is reflected in the fact that the only stress factor that increases refers to private stressors.

3.3 Individual well-being Well-being can be assessed in terms of potential stress symptoms or in terms of attitudes towards oneself and one’s life. Stress symptoms refer to irritability (reacting in an irritated manner towards others, having difficulties relaxing after work), to psychosomatic complaints, such as sleep problems, aches of all sorts (stomach, back, shoulder and neck), etc., and to depressive feelings (we talk here of ‘depressivity’ rather than ‘depression’ because the values we find are not in the range of ‘clinical depression’). Both psychosomatic complaints (2.1 on a five-point scale) and depressivity (2.2 on a seven-point scale) are low at the beginning of the study, and they even become lower over time (Figure 7.3). Irritability is the only symptom that is at the higher end of what is typically found (see Kälin et al., 2000), and it does not change very much. Attitudes refer to ‘Positive attitude towards life’, which is a general indicator, and to two work-related attitudes: job satisfaction, and a resigned attitude 170 Young Adults Entering the Workforce in Switzerland

Stable Changing

5.0 3.5 Positive attitude towards life (1–6) (n.s.) Resigned attitude towards one’s job (1–7) (***) 4.5 3.0

Job satisfaction (1–7) (n.s.) 4.0 2.5 mean mean Depressivity (1–5) (***)

3.5 2.0

Psychosomatic complaints Irritation (1–7) (n.s.) (1–5) (***)

1.0 1.0 t1 t2 t3 t4 t1 t2 t3 t4

Figure 7.3 Well-being towards one’s work. Positive attitude towards life shows rather high values, compared with other studies (Grob et al., 1991; see Kälin et al., 2000); and these values remain rather stable over four years, being around 4.7 on a six-point scale. Job satisfaction is towards the lower end of what is typically found (see Kälin etal., 2000; Grebner etal., 2003) with values around 4.3 on a seven-point scale, indicating no drastic change. Another aspect of job-related attitudes concerns job-related resignation (Elfering et al., 2000; Kälin et al., 2000; see Bruggemann, 1974). Job resignation is clearly decreasing with job experience from 3.0 to 2.6 on a seven-point scale. So, the picture with regard to well-being is similar to that regarding working conditions. There is a high degree of stability, and where changes occur, they are changes for the better.

3.4 Worries about the future We have already seen that the only stress factor that showed an increase over the period of the study was private stressors. When we contrast the work situation and the private situation with regard to worries, we also observe this discrepancy between work and private life (Figure 7.4) Worries about the future clearly decrease as far as they concern not having a job one likes. At the same time, however, worries about getting ill show an increase Semmer, Tschan, Elfering, Kälin and Grebner 171

3.0

to do a job I would not like (***)

2.5 mean

2.0

to get ill (**)

1.0 t1 t2 t3 t4

Figure 7.4 Worries about the future

between wave 2 and wave 4. That they were also fairly high in wave one probably reflects the fact that our participants were preparing for their final exams at that point in time.

3.5 Differences between occupations The results we have reported so far refer to averages in work characteristics and in well-being. Differences between various groups are, therefore, not taken into account. It is to be expected, however, that such differences exist. Regarding task stressors at work, there are two groups that stand out (Figure 7.5): nurses and electronic technicians. Nurses report higher stressors than the other groups in each wave, whereas electronic technicians report lower values. For the nurses, however, values do not change appreciably over time, whereas electronic technicians show a rising trend and move closer to the overall average. The data for nurses are compatible with trends that are being reported internationally, indicating increased workload and stress, often in connection with downsizing in hospitals (Aiken et al., 2001; Burke and Greenglass, 1999; Greenglass et al., 2003). 172 Young Adults Entering the Workforce in Switzerland

3.50

nursing: all waves ***

3.25

cooking 3.00 mean total

banking * *** ** 2.75 sales **

electronics: all waves ***

1.00 t1 t2 t3 t4

Figure 7.5 Task stressors for five professions

With regard to social support at work, the picture is more complex (Figure 7.6). Some groups show clear changes in year two and three but end up at a level similar to the one where they started. This is the case for both electronic technicians and people employed in banking. There are two occupations where there is a positive tendency, and one where there is a negative one: cooks start quite low but end up average. On the other hand, salespeople have approximately the same starting value but end up even lower. Nurses start quite well and improve constantly. They end up with the highest value of all, and their t4 value is clearly above the overall mean. Thus, by wave 4, nurses report the highest level of social support while salespeople report the lowest. This is especially important for nurses, who also report the highest levels of stress. Since social support helps people to cope with stress (Beehr, 1995; Beehr and Glazer, 2001), this constitutes an important resource. The second resource we wish to consider is control. As already mentioned, the possibility to take decisions about one’s own work is a very important resource, just like social support. Semmer, Tschan, Elfering, Kälin and Grebner 173

4.25

4.00 nursing banking ** * electronics * 3.75 mean

cooking total

* 3.50 ** sales

1.00 t1 t2 t3 t4

Figure 7.6 Social support at work for five professions

The overall trend in job control is quite clear (Figure 7.7) It increases, with bank clerks showing the greatest improvement. Nurses, however, report a problematic situation with regard to control. They start out with values close to the average, but they are the only group that subsequently shows a decline, and they move further and further away from the rest of our participants. We have shown in a recent analysis (Elfering et al., 2002) that job control has important implications for nurses: lack of control, especially time control, increases the risk for the development (or deterioration) of back pain. Comparing well-being across different occupations, we observe a similar picture with regard to many of the variables under consideration. As an example, we present the results for ‘Resigned attitude towards the job’. As shown above, this type of resigned attitude diminishes over time for the sample as a whole. With regard to differences between occupations, two aspects are especially noteworthy (Figure 7.8). First, there is a general tendency towards diminishing 174 Young Adults Entering the Workforce in Switzerland

4.00

banking *** ***

*** electronics 3.75

sales total

3.50 mean

cooking

3.25 *** *** *** nursing

1.00 t1 t2 t3 t4

Figure 7.7 Job control for five professions

differences between occupations. Evidently, over time many people become settled in their respective occupation. At the same time, nurses show considerable stability, being less resigned than the rest of the participants in each wave of measurement. This is especially noteworthy because they report the highest values in task stressors, and they report a deteriorating situation with regard to control. However, they also report a high level of social support. This support, as well as gratifying experiences from caring for people who are ill, may be a factor that protects them from developing a resigned attitude towards their work. Our data also indicate that nurses are quite effective in dealing with stress, as indicated by comparably high values in ‘problem-orientated’ coping as well as in considerable ability to calm down successfully. Whereas the first example referred to a work-related attitude, the second one concerns ‘Positive attitude towards life’ (Grob et al., 1991) in general (Figure 7.9). Semmer, Tschan, Elfering, Kälin and Grebner 175

3.75

cooking *** 3.50

sales 3.25 ** ** total

3.00

mean **

2.75 electronics banking

2.50 *** *

nursing * *** * 1.00 t1 t2 t3 t4

Figure 7.8 Resigned attitude towards one’s job for five professions

With regard to Positive attitude towards life, nurses are consistently high, salespeople are lowest throughout. Bank clerks and, especially, cooks, show an improvement. Thus, for these groups, results mirror those for resigned attitude towards work. There is, however, one notable exception, in the case of electronic technicians. They start at a middle level, rise after finishing vocational training but then show a sharp fall. We tried unsuccessfully to find variables that would explain this drop. For instance, variables like fears about not having a job, or social support at work could not account for this development. Thus, we have to report this result without really being able to interpret it in a stringent manner. Note that, unlike resigned attitude towards work, the development of positive attitude towards life is not simply one of diminishing differences. Rather, we see two groups emerging: nurses, 176 Young Adults Entering the Workforce in Switzerland

5.25

5.00

* electronics nursing * **

4.75 mean total

banking cooking 4.50 * * sales

1.00 t1 t2 t3 t4

Figure 7.9 Positive attitude towards life for five professions

bank clerks, and cooks are at the higher end, and electronic technicians and salespeople cluster at the lower end.

3.6 Differences between males and females With regard to the indicators of well-being and distress that we have assessed, many studies have found that women report higher levels of distress than men (e.g., Nelson and Burke, 2002). Our results yield surprisingly little support for these findings (Figure 7.10). We find no significant differences with regard to positive attitude towards life, or depressive tendencies. Women are more satisfied with their work, men are higher on ‘self-efficacy’, that is, the conviction that one is able to deal successfully with problems. Women are higher in irritability initially, but over time these differences disappear. One variable where we do find consistent differences is with regard to psychosomatic complaints, where women have higher values throughout. Semmer, Tschan, Elfering, Kälin and Grebner 177

5.0 Job Satisfaction (**) women 4.5

4.0 men women 3.5 Irritability (***)

3.0 mean

men 2.5

women

2.0

men 1.5 Psychosomatic complaints (***)

1.0 t1 t2 t3 t4

Figure 7.10 Sex differences in well-being

This lack of perceived differences may be partially due to the fact that we do not find consistent differences with regard to the work situation. Stress factors, control, or social support at work are not different for men and women. One interesting question with regard to males and females concerns future plans. Among other things, we asked our respondents about the intention to reduce work in order to gain time to care for children. As Figure 7.11 shows, women record somewhat higher values from the beginning, but by our last wave of measurement, the differences have become quite pronounced. As our participants reach the age where children become a serious issue, it appears that gender roles become more pronounced. The developments in society over the past few decades, while showing some increase in the time men spend with children and doing household chores, do not seem to have produced a fundamental alteration of gender roles in this respect. Women 178 Young Adults Entering the Workforce in Switzerland

2.50

wave x gender: Reduce work, increase child care (***) 2.25

women

2.00

1.75 mean

1.50

1.25 men

1.00 t1 t2 t3 t4

Figure 7.11 Sex differences in future plans are still those mainly responsible for child care (and also household work – see Nelson and Burke, 2002).

3.7 Differences between the French- and German-speaking regions of Switzerland Our study was conducted in both the German-speaking and the French- speaking regions of Switzerland. This raises the issue of whether differences exist these two regions, for instance, in terms of well-being, or in terms of working conditions. There are indications in the literature that recorded levels of well-being are somewhat lower among the French-speaking population (Grob et al., 1996). We observe a similar trend in our study. When compared to their French-speaking colleagues, German-speaking young job entrants reported a lower level of job resignation, less irritation and a higher positive attitude towards life. Values in self-esteem, however, were higher in the French-speaking region. Semmer, Tschan, Elfering, Kälin and Grebner 179

As examples, we present two indicators of well-being. One is work-related: resigned attitude towards work. The other is more general: irritability. As Figures 7.12 and 7.13 show, in both cases values are lower for the German- speaking sample, indicating a higher level of well-being. The question arises, of course, how these differences can be explained. One obvious possibility is that working conditions differ in the two regions. One way to test this is to control for working variables in statistical analyses. This yields values that are adjusted for work characteristics. If differences disappear, this indicates that they are not true regional differences but rather the result of regional differences in work characteristics. The analyses show, indeed, that most of the differences either disappear altogether, or diminish, when we control for work-related variables. Figures 7.14 and 7.15 show the values for resigned attitude towards work and irritability. They are adjusted for three important variables at work: feeling appreciated as an employee, social support, and job control. Young people in the French-speaking part of Switzerland experience less appreciation and support at

3.50

Language differences Resigned Attitude towards Job (***)

3.25

French 3.00 *** *** ***

2.75 mean

2.50

German 2.25

1.00 t1 t2 t3 t4

Figure 7.12 Resigned attitude towards one’s job: language differences 180 Young Adults Entering the Workforce in Switzerland

4.25 Language differences Irritability (***)

4.00

3.75 *** French *** *** 3.50 mean 3.25

German 3.00

2.75

1.00 t1 t2 t3 t4

Figure 7.13 Irritability: language differences work than do their German-speaking counterparts. This is associated with lower levels of well-being. However, where they do experience comparable appreciation and support, differences in well-being are less pronounced. As we mentioned earlier, there is one indicator of well-being where the participants from the French-speaking part of the country show more positive values than their German-speaking counterparts: they exhibit a higher level of self-esteem. The results are displayed in Figure 7.16. These differences are independent of work characteristics – that is, they do not disappear when they are adjusted for appreciation at work, social support, and job control (see Tschan et al., 2004).

3.8 Trajectories: changing jobs, changing professions Our sample consists of young people, an age cohort that is characterized by a considerable degree of change. We have already mentioned that 56 per cent change employer by the end of the first year, and that 75 per cent have changed employer at least once by the end of the study. So, the question Semmer, Tschan, Elfering, Kälin and Grebner 181

3.50 Language differences Resigned attitude towards job (n.s.)

3.25

3.00

French 2.75 * mean

2.50 German

2.25

1.00 t1 t2 t3 t4

Figure 7.14 Resigned attitude towards one’s job: language differences, controlling for work characteristics arises of whether such transitions are in any way indicators of a special development. It could be, for instance, that it is mainly those who were unsuccessful who change their employer. This might, in turn, imply that those who change employer have a poorer prognosis for their further development. On the other hand, it is also conceivable that a change of employer is due to a ‘misfit’ between the circumstances in a specific company that is ‘corrected’ by changing to a new employer. So far, the results available in the literature point to the latter explanation: those who change jobs are less satisfied to begin with (which is an important reason for changing) and, typically, their job satisfaction rises after the change (cf. Baillod and Semmer, 1994; Semmer et al., 1996). With regard to changing employer after vocational training, this question is complicated by the fact that many of our participants had to change their employer – not necessarily because they did not perform well, but simply because their contract ended. Unfortunately, we do not know why people changed between waves 1 and 2. For those changes that came later, we did ask participants if the change was based on their own choice or not. 182 Young Adults Entering the Workforce in Switzerland

4.25 Language differences Irritability (*) 4.00

3.75

** French 3.50 mean 3.25 German

3.00

2.75

1.00 t1 t2 t3 t4

Figure 7.15 Irritability: language differences, controlling for work characteristics

Similar questions arise for those who change profession. Of course, changing professions is a much less common event than changing employers. After one year (at time 2), 5.2 per cent have changed profession; this number rises to 7.3 per cent by wave two, and to 13.5 per cent by the end of the study. Does such a change indicate a successful correction of a choice that turns out not to be promising, or is it a sign of difficulties that may be more typical for the person rather than the job situation?

3.8.1 Changing employer If we first consider those who change their employer, we typically find the following pattern: compared with ‘Stayers’, ‘Leavers’ show a lower level of well-being before the change. After the change, however, their well-being improves and reaches a level that is comparable to, or even better than, that of stayers. This does not apply to all variables alike, but where changes occur, they tend to follow this pattern. It is most consistent for job-related variables (job satisfaction and resignation), and between waves two and three. Semmer, Tschan, Elfering, Kälin and Grebner 183

4.50

Language differences Self-esteem (***)

4.25

French: all waves ***

4.00 mean

3.75

German

3.50

1.00 t1 t2 t3 t4

Figure 7.16 Self-esteem: language differences

We wish to illustrate this pattern with a result concerning the job satisfaction of those who stay with their employer (‘stayers’) and those who change (‘leavers’). Participants were counted as ‘leavers’ only if they had initiated the change on their own, but not when they had to leave. The time frame relates to changes occurring between time 2 (about one year into ‘real’ work) and time 3 (one year later). Only participants are included in the analysis who did not change their profession altogether. As Figure 7.17 shows, job satisfaction is lower in leavers initially, but improves, and is somewhat higher afterwards. By contrast, job satisfaction of stayers decreases slightly. This is also the typical finding in other studies (cf. Semmer etal., 1996; Semmer and Schallberger, 1996). Two conclusions can be drawn from these results: (1) Those who change employer have a reason to do so: Their well-being is not as good as that of the stayers. (2) On average, it is beneficial to change: In many cases, well-being – especially work-related well-being – improves, and it often reaches values that are comparable to, or better than those of the stayers. 184 Young Adults Entering the Workforce in Switzerland

5.0

change

4.5

no change mean

4.0

1.0 t2 t3

Figure 7.17 Change of employer and job satisfaction Note: Participants are counted as ‘changers’ only if they left voluntarily. People who changed profession are not included in this analysis.

3.8.2 Changing profession We now consider those who have changed their profession. In this instance, the picture is similar in some respects, but quite different in others. Looking at working conditions (Figure 7.18), those who change profession initially report lower levels of control. By the end of the study, their possibilities to influence their own work have improved considerably, and has surpassed that of those participants that stay in their jobs. With regard to task stressors, their initial values are already better than those of ‘stayers’, and they get even lower over time – that is, the gap widens. So, ‘leavers’ with regard to profession improve their work characteristics over time, relative to those who do not change. (This difference in change over time is reflected in the fact that ‘wave x change’ is statistically significant.) If we consider their well-being, however, the picture is different (Figure 7.19): at the outset, those who change again have poorer values than those who stay. In contrast to the picture with regard to change of employer, however, the development of those who change profession is less clear. The difference to those who stay does not narrow significantly, their development over time being similar to that of the stayers. This shows, for instance, in ‘Positive Attitude Towards Life’, which is essentially stable in both groups, but at a Semmer, Tschan, Elfering, Kälin and Grebner 185

4.0

Job control Main effect: n.s.; Wave x change: ***

3.8

no change of profession 3.5 ** ***

change of profession 3.3 mean

*** *** 3.0 ** * no change of profession

2.8 change of profession

Task stressors Main effect: ***; Wave x change: **

1.0 t1 t2 t3 t4

Figure 7.18 Change of profession and work characteristics

lower level for those who change profession, and in ‘Resigned Attitude towards Work’, which improves for both groups in a similar way over time, with the gap diminishing only very slightly (see Figure 7.20). With regard to some variables (such as job satisfaction), the improvement is somewhat stronger than for those who do not change, but the difference still does not reach statistical significance. So, with regard to changing one’s profession, we find, again, that leavers are worse off in terms of well-being at the outset. Their change, however, does not pay off in the same way as in the case of changing employer; the picture here is much less consistent. It is interesting to observe that well-being often 186 Young Adults Entering the Workforce in Switzerland

5.0 no change of profession * ** * *

4.5 change of profession

Positive attitude towards life (1–6) Main effect: **; Wave x change: n.s.

4.0

mean Resigned attitude towards one’s job (1–7) Main effect: ***; Wave x change: n.s.

3.5 *** change of profession

** 3.0

no change of profession 1.0 t1 t2 t3 t4

Figure 7.19 Change of profession and well-being

does not approach the same level as for those who stay in their profession, even though there is an improvement in working conditions – less stress and more control are certainly what many people might seek. But those who change profession do not seem to be able to profit consistently from these developments in terms of well-being. At least for some people in this (rather small!) group, well-being indicators point to a more problematic situation in comparison to the ‘stayers’. Of course, these are mean values, so this statement does not refer to everybody within the group. Yet it is safe to say that in many cases change of profession is likely to be a sign of difficulties, and that these difficulties often do not get better following the change. Where they do, this development is most pronounced in job-related variables, such as job satisfaction. Semmer, Tschan, Elfering, Kälin and Grebner 187

3.9 Predicting well-being over time The final question that we want to address in our chapter deals with the prediction of well-being over time. Do working conditions affect well-being? Or is it those who have poorer well-being from the outset that manage to find challenging work, or improve their working conditions where they are? And what is the effect of other life circumstances? If stress influences well-being – is it really work stress, or is private stress the more important factor? The analysis we present here incorporates these ‘possible answers’, predicting well-being simultaneously through different ‘predictors’: initial well-being, work characteristics, and private stress. The variable we predict, for illustrative purposes, is resigned attitude towards work – a variable we have referred to throughout this chapter. As Figure 7.20 shows, one important predictor of resignation at the end of the study is resignation at the beginning of the study. This is a typical result: There is quite some stability in people’s well-being. So, knowing somebody’s resigned attitude towards his or her work yields some prediction of that person’s attitude in later years. A second predictor is a variable that we have mentioned several times already – namely appreciation. Feeling appreciated at work (that is, being

Resigned attitude towards one’s job (T1: Stability)

β = 0.10*

Appreciation received at work (mean T1–T4) Resigned attitude towards one’s job T4 β = –0.30*** 2 R = 0.22***

Private stressors 0.14** (mean T1–T4)β =

Figure 7.20 Well-being: influences from work, private life, and personality 188 Young Adults Entering the Workforce in Switzerland treated as a colleague rather than a beginner) is one of those variables that we consistently find to be important in our study (see also Semmer and Jacobshagen, 2003). And, as the figure shows, feeling appreciated predicts better well-being, in terms of lower resignation, at the end of the study. Note that appreciation represents the mean of all four waves, that is, the cumulative experience of appreciation over four years! Finally, we included private stressors. Again, we calculated the mean of all four waves, and this adds to the prediction of resignation at the end of the study. Its contribution to this prediction is smaller than that of appreciation at work; this makes sense, as the variable being predicted is job-related. It has to be emphasized that the statistical method we employed (multiple regression) displays the unique influence of each variable; the influence of the other predictors is controlled for. These results imply that the question raised above – is it the person, is it work, is it private circumstances – cannot be answered conclusively in favour of one or the other: it is, rather, a combination of all these influences. One cannot say that the work environment – which is at the focus of our investigation – is ‘the’ factor that influences well-being. People have personal characteristics that do not change easily, they have family constellations, social relations and activities outside work. All of these factors are important contributors to one’s well-being. It would therefore be naïve to attribute everything to work. But one cannot attribute everything to the person, either: work experience does play a role, and thus, trying to create a working environment that is as favourable as possible is an important task.

4 Summary and conclusions

Taken as a whole, our results present a positive picture. Many of the people who took part in our study seemed to have made a smooth transition from vocational training into ‘real work’, and they seem to have settled quite well in their work situation after a few years. Stability is more prevalent than change, and where change does occur, it is, generally, mostly change for the better. This pattern also emerged when we asked respondents for the perceived change in their general well-being and life satisfaction, with ‘worse’, ‘same’, and ‘better’ as possible responses. After each wave of measurement, the largest group always reported ‘no change’. Where participants did report change, it was much more often change for the better than change for the worse. Note, however, that the group we have studied is composed of people where one would expect quite a positive development. These are now skilled people who have gone through a system of vocational training that leads to respect and appreciation in society, and this may also help to buffer some of the uncertainties that are associated with the current economic changes (cf. Mills and Blossfeld, 2003). But it is important to note that, overall, the Swiss system of vocational training does seem to provide the smooth transition that one would desire. Semmer, Tschan, Elfering, Kälin and Grebner 189

But note that this is the overall picture. Of course, there are some results that may raise concern. Among them are the high task stressors and the decline of job control observed among nurses (Aiken et al., 2002), the low levels of social support for salespeople, and their rather low values on ‘Positive Attitude Towards Life’. Changing employer is, however, no reason for concern, as it is associated with improvements in many cases. The picture is less clear among those who change profession: Although their work charactistics improve after the change, their well-being does not show a clear picture of improvement. Thus, it is likely that there is a subgroup among those who change their profession for whom this may indicate a rather problematic situation, compared to the other participants. As the group of those who do change profession is small altogether, however, this concerns a rather small subgroup of our survey. Of course, it has to be emphasized that all of the results we have reported on here refer to averages: not everyone is stable in terms of job satisfaction; not all nurses report high stressors but also high social support; not everyone who changes employer will profit from this; and many of those who change profession do improve their level of well-being. The development of a given individual will always depend upon his or her very individual combination of personal characteristics, private circumstance, and working conditions. Nevertheless, our analyses shed light on developments that are found among these young people on average. And they underscore the importance of certain variables that influence the probability of experiencing a positive development. Among these variables is one which we wish to give special emphasis. We could show that appreciation is an important aspect of the work environment that influences how people feel about their work, and about themselves. This is something that employers should pay particular attention to. We know from a considerable amount of research that people who feel respected and treated in a fair manner will show loyalty and engagement in their work, while those who do not are at a higher risk of becoming demotivated, frustrated, and burned out (cf. Greenberg, 1990; Semmer and Jacobshagen, 2003; Semmer et al., in press; Siegrist, 2002; Taris et al., 2001). Note, however, that appreciating people involves more than simply praising them. It may also manifest itself in giving them interesting work, giving them responsibility, and granting them as much discretion as possible in deciding how to carry out their jobs, but also criticizing them fairly and constructively, supporting them in difficult situations, and the like. Appre- ciation therefore implies social behaviours (Semmer and Jacobshagen, 2003; Semmer et al., in press), but also job design and working conditions (Parker and Wall, 1998; Ulich, 2001). If applied in this sense, appreciating employees, and treating them fairly and respectfully, will profit both employers and employees. 8 The Erosion of Regular Work: An Analysis of the Structural Changes in the Swiss and German Labour Markets* Andreas Diekmann and Ben Jann

For the past few decade, the hypothesis on the ‘erosion of regular work’ has received a remarkable amount of attention, especially in the German literature. Usually, regular work (or standard employment) is characterized by full-time employment based on a non-temporary labour contract. Wages and salaries are standardized through collective labour agreements and include social security contributions. The hypothesis on the erosion of regular work states that standard employment is being displaced by more atypical forms of employment. Increasingly accepted forms of labour are modern, more flexible, yet more precarious and unprotected. However, whether the hypothesis of the displacement and substitution of traditional work forms is supported empirically, and if so, to what extent such labour-market transformations are taking place, has not yet been answered satisfactorily. We therefore analyse here the degree to which the different forms of employment currently exist in the Swiss and German labour markets and how labour-market structures have changed. Our analyses reveal that there is, in fact, a tendency to replace standard employment with other forms of labour. This is true for both countries despite the clearly more rigid labour-market regulations in Germany that could be expected to have a decelerating effect. For men, in particular, an erosion of regular work can be observed over all age groups and not only ‘at the edges’

* The support from the Swiss National Science Foundation through a research grant on ‘Working Conditions, Work Orientation and Labour Market Participation’ is gratefully acknowledged (Grant no. 5004-047887). We would like to thank Elisabeth Coutts and Rachel Matthey for their helpful comments and suggestions.

190 Andreas Diekmann and Ben Jann 191 of the age distribution. Furthermore, the process did not start until the 1990s. In the preceding years a decline in standard employment relationships can be observed in a relative sense only. This pretended erosion is primarily due to the expansion of (female) labour market participation and with it a disproportionate growth of atypical (that is, part-time) employment.

1 The end of regular work?

At the end of the 1990s, the German Kommission für Zukunftsfragen1 (Commission on Future Issues) produced several succinct hypotheses that reflected its concern about dwindling standard-employment opportunities and the consequences of the decline in so-called ‘regular’ or ‘normal’ work – primarily permanent full-time employment.2 According to the committee’s report, dramatic changes that had begun in the 1970s were then becoming increasingly apparent in Germany. The diagnosis and prognosis was: ‘At the beginning of the 1970s,...there were still five people employed under standard conditions for every one employed under non-standard conditions. At the beginning of the 1980s, the ratio was one to four, in the middle of the 1980s it was already one to three. In 1996, it was one to two ...If this trend continues, the ratio of standard to non-standard employment will be one-to-one within a few years’ (Kommission für Zukunftsfragen, 1998, pp. 43/48, translation by the authors). Some worry that this trend will, for example, give rise to an increase in income inequality. In addition, some jobs may no longer bring in enough revenue to ensure a basic standard of living, while transfer payments in general, and retirement pensions in particular, may no longer guarantee an income above the poverty line for a growing part of the population. Over the past two decades, the issues addressed in the report by the Kommission für Zukunftsfragen have been subject to intensive discussion and also much speculation. The basic idea is that the individualization and flexibilization, which is supposed to go along with the modernization and globalization of post-industrialized western societies, give rise to distinct changes in working life, restructuring of the labour market and a diversity of life courses (e.g., Beck, [1986] 1992; Sennett, 1998). On the one hand, many studies have focused on the emergence of new and modern forms of labour that are more flexible, yet also more precarious and unprotected. These studies are concerned with the rise of so-called non-standard or atypical employment forms such as part-time and marginal employment, temporary and subcontracted work, new forms of self-employment, and so on (e.g., Rodgers and Rodgers, 1989; Walwei and Werner, 1995; O’Reilly, 1996; Blossfeld and Hakim, 1997; De Grip et al., 1997; Smith, 1997; Lind and Møller, 1999; Kalleberg, 2000; Giesecke and Gross, 2003). On the other hand, ideas about the restructuring of labour markets have been discussed from the viewpoint 192 The Erosion of Regular Work of the corrosion of the traditional model of employment – that is, the ‘end’, ‘crisis’ or ‘erosion of regular work’, and even the ‘destruction of the standard employment relationship’ (e.g., Mückenberger, 1985; Zachert, 1988; Roy, 1989; Burgess and Watts, 1999; Dombois, 1999; Bernhardt and Marcotte, 2000; Wagner, 2000; Kaiser, 2001b; Appelbaum, 2002; Moulin, 2003).3 Whereas it is undoubtedly true that new forms of labour such as part-time employment have become increasingly important over the past few decades, it may not be taken for granted that this process leads automatically to a corrosion of the traditional forms of employment, as is often assumed. In the context of the hypothesis of the erosion of regular work it is commonly suggested that the decline in standard employment relationships is the result of full-time employment being replaced by atypical forms of employment to an increasing extent. The Kommission für Zukunftsfragen has suggested that ‘substitution of normal employment conditions with non-normal employment conditions’ is occurring (1998, p. 23, translation by the authors). However, it is also conceivable that, despite a decline in regular work in a relative sense, the absolute number of standard positions may have remained constant, or even increased. This would, for example, be the case if part-time work, marginal employment, and other ‘atypical’ forms of work were growing disproportionately, but not at the expense of full-time jobs. In the first case, the erosion of regular work occurs through substitution, in the second case through the higher growth rate of non-standard forms of labour. The relative decline in ‘normal’ positions can therefore be explained by two hypotheses: (1) the substitution and displacement hypothesis, which states that there are fewer regular jobs in an absolute sense; and (2) the hypothesis of disproportionate growth of atypical employment, which states that there are fewer regular jobs only in a relative sense. The question of which process is actually taking place in the labour market must be answered empirically. Considerable criticism of the hypothesis regarding the extensive erosion of regular work has been voiced from an empirical standpoint. For example, Hoffmann and Walwei (1998) analysed changing employment patterns using data from the German microcensuses. While their study did identify a decline in the proportion of workers employed full-time on a permanent basis, particularly in the 1990s, they found that this process was taking place only relatively slowly. Their finding is in contrast to the rapid rate predicted by analyses carried out by the Zukunftskommission and other proponents of the erosion hypothesis (for example, Beck, [1986] 1992; see also the review provided by Kress, 1998). The hypothesis on the substitution and displacement of standard employment has also been subject to criticism. Analyses by the German Institute for Economic Research, for example, also indicate that regular work is on the decline (Wagner, 1998; Kaiser, 2001a; for similar results see also Oschmiansky and Schmid, 2000). The primary reason provided by the authors was, however, that the labour force participation rate has grown overall and that formerly unemployed people have become employed part-time to Andreas Diekmann and Ben Jann 193 an increasing degree. In particular, female labour-force participation has increased and with it the number of part-time jobs. This trend alone would indicate a relative reduction in regular positions and give support to Hypothesis 1 above rather than to Hypothesis 2. However, another factor that may play a role (according to Wagner) is the differential development that takes place at the beginning and the end of an occupational career. If more young people delay their entry into the labour market in favour of education and older people retire earlier, shrinkage in full-time employment should, ceteris paribus, be a consequence (Wagner, 1998). In evaluating this claim, cohort size must also be taken into account. It may well be that full-time employment is increasingly being replaced by education among young adults. However, for demographic reasons the proportion of the population as a whole formed by younger cohorts is declining. To exclude the above effects, Wagner (1998) suggests that the developments in full-time employment be studied only in the 25–55-year-old demographic group. In this chapter, we will give a detailed analysis of the decline in regular full-time employment in Switzerland and Germany. On the one hand, we are interested in the differences in the speed of this development in the two countries. On the other hand, we are also interested in whether the labour-market developments in the two countries offer more support to the hypothesis of the mere expansion of ‘atypical’ forms of work or to that of the displacement of standard employment. Furthermore, the differential employment patterns for men and women and the age dependence of employment behaviour suggest a separate examination of the development in the labour market for men and women as well as for the various age groups. Comparing the two countries can also produce a better understanding of the development of labour-market conditions. More concretely, a comparative examination can provide some information on whether, and to what extent, the flexibility of labour markets accelerates structural changes in those markets. This is not necessarily the case: flexibility or rigidity only means greater or lesser freedom in specifying conditions of employment. To what extent these liberties are exploited depends upon both the interests and the power of the contractual partners. The introduction of new organizational structures (cost saving through outsourcing, concentration on the core business, and so forth) and the trend towards cutting labour costs may reflect the employers’ interest in reducing regular work. However, some kinds of employees may also actually prefer something less than full-time employment. For example, the demand for part-time jobs has grown particularly among women with children (see Kress, 1998, for the probable causes of the erosion of regular work). Clearly, the interests of some employers and employee groups will have better chances of being realized if the contractual liberty of the parties on the labour market is less strictly regulated by the government and the unions. According to this 194 The Erosion of Regular Work view, rigid labour markets slow down the trend toward the erosion of regular work. In flexible labour markets, in contrast, the change in labour relations will proceed more readily. The comparison of the developments in the Swiss and German labour markets is a touchstone for the hypothesis on the effect of labour-market flexibility. Although very similar in their level of societal development and modernization, the labour-market regulations of the two neighbouring countries differ in a significant way. These differences are a part of a general set of disparities in labour-market conditions among industrialized nations: According to a report by the OECD (1999), labour markets are most restrictive in southern Europe, France, and Germany. The least restrictive are in English-speaking countries: New Zealand, Canada, Great Britain, and the USA. According to the same report, Switzerland’s policies lie close to the Anglo-Saxon end of the flexibility spectrum. The OECD judged the flexibility of the various labour markets on the basis of three indicators: ‘strictness of employment protection for regular employment’ (for example, protection against dismissal), ‘regulation of temporary employment’ (fixed-term contracts, temporary work agencies) and ‘regulation of collective dismissal’ (OECD, 1999). In comparison with Germany, the Swiss labour market is clearly more flexible with regard to the first two indicators. Only with respect to collective dismissals are the constraints more rigid in Switzerland than they are in Germany. All things considered, the Swiss labour market holds one of the highest positions in the ranking of 30 OECD nations according to their labour-market flexibility (depending on the measurement method used, Switzerland ranks either sixth or seventh). Further, the Swiss held this position consistently throughout the 1980s and the 1990s. Germany, in contrast, was at place 14 in the late 1980s, but has since dropped down to place 18 or 20, depending upon the measurement procedure used. One may well ask, therefore, whether the very different labour-market regulation profiles of the two countries also affect the speed and extent of structural changes in the labour market.

2 The development of regular work in Switzerland

How one studies development and the prevalence of regular work depends to a large extent upon the definition of regular work, the inclusion of certain age groups in the sample being used and – as is usual for proportions – on how one defines the base. In keeping with the practices used in previous literature (Kress, 1998; Hoffmann and Walwei, 1998; Kommission für Zukunftsfragen, 1998; Kalleberg et al., 2000) we consider regular work to be (i) wage and salary employment that is (ii) based on a non-temporary labour contract, (iii) pursued on a full-time basis, and (iv) the earnings from which are subject to social insurance deductions.4 Our data source is the ‘Swiss Labour Force Survey’ (SLFS), which has been conducted annually by the Swiss Federal Statistical Office since 1991 (see Bundesamt für Statistik, 1996). The Andreas Diekmann and Ben Jann 195

SLFS data are collected by computer-aided telephone interview for a random sample of 16,000 to 18,000 subjects (with an increase to a sample size of 41,000 in 2002). The average response rate in the SLFS, which amounts to 70 per cent, is considered relatively high.5 We apply case weights (neglecting post-stratification) in our analysis to account for varying sampling probabilities (cf. Comment et al., 1996). Because the SLFS data sets provide information only about developments since the beginning of the 1990s, we also use data collected in the Swiss censuses as from 1970. To determine the proportion of standard positions as a fraction of the total number of positions, we employ a hierarchical employment classification that more or less reflects the definition of regular work given above. First, the population is divided into those people who are active in the labour market (the labour force, i.e. both employed and unemployed persons) and those who are not participating in the labour market (the non-employed). The employed are divided into the self-employed, trainees (apprentices) and wage and salary workers; those in the latter group are, in turn, assigned to a category according to the rate of their labour-market participation (i.e. full-time, part-time, or marginal). Finally, the standard-labour category comprises those wage and salary workers employed full-time with a non-temporary labour contract. There is no need for further differentiation according to whether or not earnings are subject to social insurance deductions because the deductions are mandatory for these workers in Switzerland.6 In the SLFS, respondents are considered active in the labour market if their workload is at least one hour a week (this amount also includes unpaid work for a family business). The distinction between the self-employed and wage and salary workers, as well as the distinction between full-time and part-time workers, is based largely on the respondents’ self-assessment. We start with an examination of the working-age population (ages 16 to 64). Information on the distribution of forms of employment in this population can be found in Table A8.1 (in the Appendix), which provides details of the various proportions for the period from 1991 to 2003. Let us first consider the share of wage and salary employment formed by regular work, which can be derived by a simple transformation of the values in the table (Figure 8.1). Among men, this proportion is 90.7 per cent in 1991, 88.0 per cent in 1992 and 86.4 per cent in 2003. Among women, the proportion is 44.4 per cent in 1991, 40.8 per cent in 1992, and 30.8 per cent in 2003. The substantial changes between 1991 and 1992 might possibly be a statistical artifact due to teething troubles in the first wave of the SLFS. We therefore choose the year 1992 as the starting point. In 1992, it is evident that regular work, with a share of almost 90 per cent, is clearly the dominant work form among male wage and salary workers. Furthermore, the proportion remains almost constant throughout the 1990s if the year 1992 is chosen as the starting point. These findings are in accordance with the results of Widmer, Kellerhals and Levy (in this volume) who reject the hypothesis of extreme 196

(a) Men (aged 16 to 64) 100 90.7 88.0 86.4 90

70 65.8 63.1 55.2 60

50 Percent 40

0 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

(b) Women (aged 16 to 64) 100

60 44.4 50 40.8 Percent 40 30.8 24.7 30 22.2 17.4 20

0 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

Denominator: Wage and salary workers Population

Figure 8.1 Regular work in Switzerland from 1991 to 2003 Source: Swiss Federal Statistical Office, Swiss Labour Force Survey (SLFS); calculation: the number of wage and salary workers employed non-temporarily on a full-time basis divided by the number of wage and salary workers in total or the total population aged 16 to 64; grey area: approximate 95 per cent confidence interval. Andreas Diekmann and Ben Jann 197 individualization in Switzerland and report that the life courses of 85 per cent of the men follow very homogeneous trajectories characterized by full-time professional activity. Among women, on the other hand, the share of those performing regular work – again with reference to all female wage and salary workers – has decreased markedly. Since 1992, the proportion of full-time female employees has been overtaken by the proportion of part-time employees. ‘Regular work’ for women seems increasingly to be part-time work (see also Baumgartner, 2003). However, the restriction of the analysis to wage and salary workers can be misleading if displacements have occurred between this group and the other categories. In order to paint a more differentiated picture, we also analyse the population shares listed in Appendix Table A8.1, which are based on the entire population of the 16-to-64 year-old age group. Because it is clear that the changes in employment forms are different for women and men, we focus on the sex-specific figures right away. We can now find a decrease in regular work for both women and men: these decreases are five and eight percentage points, respectively, between 1992 and 2003. If the 1991 values are instead used as the starting point, the decrease is even more pronounced: 7 per cent for women and more than 10 per cent for men. These values make it clear that the decline in regular work can no longer be explained solely by a trend towards part-time work among women. Instead, a decrease in regular work can be observed both for women and men. How can this decrease in regular work be explained? Obviously, not by the growth in fixed-term employment: Entirely at odds with the expectations of many, the relatively small share of temporary jobs has even fallen slightly. However, among men the proportion of the self-employed, part-time workers and the non-employed has grown. Among women, it is mostly part-time employment that has gained from an expansion of the female labour-force participation rate, but there has also been a shift away from full-time work. Now, let us extend the time line from 1991 further into the past. The Swiss census data record job characteristics with categories and methods different from those of the SLFS, so we cannot compare the absolute percentages of labour forms directly. For instance, those who work between one and five hours a week are considered to be non-employed by some of the Swiss censuses while the SLFS considers the same people to be a part of the working population. For the sake of consistency we treated these people as non- employed for all waves of the Swiss census. Furthermore, we cannot account for temporary labour contracts separately with this data. The basic trends are, however, visible in the census data. In the time from 1970 to 2000, a decrease in standard employment can be observed for men, down from 73.2 per cent in 1970 to 59.5 per cent in 2000. For women, the rate fell from 30.4 per cent in 1970 to 27.0 per cent in 2000 (see Figure 8.2 and Table A8.2 in the Appendix). The inclusion of 198

(a) Men (aged 16 to 64) 100 97.8 97.1 96.2 90 93.5 80

70 73.2 71.1 69.0 60 59.5 50

Percent 40

0 1970 1980 1990 2000

(b) Women (aged 16 to 64) 100

70 75.3 67.4 60 60.1 50 52.0 Percent 40

30 30.5 28.9 29.3 20 27.0

0 1970 1980 1990 2000

Denominator: Wage and salary workers Population

Figure 8.2 Regular work in Switzerland from 1970 to 2000 Sources: Swiss Federal Statistical Office, harmonized personal records from the 1970, 1980, 1990, and 2000 Swiss Censuses; calculation: the number of wage and salary workers employed on a full-time basis divided by the number of wage and salary workers in total or the total population aged 16 to 64. Andreas Diekmann and Ben Jann 199 this data also shows that the erosion in regular work occurred primarily in the 1990s. The census data can be broken down into single-year age groups, thereby permitting a detailed analysis of the distribution of forms of employment according to age. Figure 8.3 shows the development of the age-specific rates of standard employment over the four waves of the Swiss census (each curve represents one wave and displays the proportions of regular work within each single-year age group). Here it can be seen that the age-dependent curve of regular work decreased for men across all working-age groups particularly in the 1990s. For women, the pattern is different. For the 2000 census, we can observe a shift to the right and downward in the first part of the curve for women up to an age around the mid-thirties. This means that the percentage of regular work decreased in the 1990s for younger women, whereas it increased for women between their mid-twenties and mid-thirties. Both men and women are less likely to be employed in full-time jobs at a young age and towards the end of their occupational careers. The reasons are obvious: increased participation in education among the younger and earlier retirement among the older. These results are supported by an analysis of the SLFS data. These data also indicate that regular work has sharply declined for the 16-to-24-year-old age group. The proportion dropped from 34.6 per cent in 1992 to 20.0 per cent in 2003. This development, which mostly reflects an increase in young people undertaking an apprenticeship, can be observed for both men and women. However, because the relative size of this age group fell during the same period, the proportion of apprentices has remained constant in the overall population (see Table A8.1 in the appendix). In the 55-to-65-year-old age groups, on the other hand, the proportion of regular work has declined for men only. The reduction of about ten percentage points is due partly to higher self-employment in the group, but also to earlier retirement. Regardless of these changes amongst the younger and older age groups, the rate of full-time wage and salary work has also decreased for men in the 25-to-55-year-old category – from 77.0 per cent in the census of 1990 to 68.3 per cent in the census of 2000. According to the results of the Swiss census, the decline in regular work cannot, therefore, be attributed exclusively to changes ‘at the edges’ (Wagner, 1998) of the age distribution, although the extension of the length of education for young adults and earlier retirement for older adults have had some effect on the proportion of regular work. Again, such results are confirmed by the SLFS data. If the young and old age groups are excluded – that is, if only developments in the population of the 25-to-55 year-olds are analysed – the erosion of regular work observed may be less clear than across the whole sample, but the decline is still substantial. For men, the proportions are 72.5 per cent in 1992 (74.6 in the year 1991) versus 66.1 per cent in 2003. For women the proportion drops slightly from 21.3 per cent in 1992 (23.8 per cent in 1991) to 19.1 per cent in 2003. 200

(a) Men 100

Percent 40

0 15 20 30 40 50 60 70 80 Age

(b) Women 100

Percent 40

0 15 20 30 40 50 60 70 80 Age

1970 1980 1990 2000

Figure 8.3 Age-dependent rate of regular work in Switzerland from 1970 to 2000 Sources: Swiss Federal Statistical Office, harmonized personal records from the 1970, 1980, 1990, and 2000 Swiss Censuses; calculation: the number of wage and salary workers employed on a full-time basis divided by the total population for each single-year age group. Andreas Diekmann and Ben Jann 201

Compared to the results from the Swiss census, however, the erosion of regular work found in the SLFS data for the period from 1991 or 1992 to 2000 is surprisingly small. Although the general patterns are basically the same – a decline in standard employment for men, stability for women – the trend towards the erosion of male regular work is much more pronounced in the census data than in the SLFS data. The discrepancy between the results of the two data sources is probably due, to a large extent, to differences in the measurement and operationalization of the worker types. In particular, employees with fixed-term contracts cannot be accounted for separately with the census data. This seems to inflate the observed trend because the proportion of temporarily employed full-time workers declined for men in the period between 1990 and 2000. Thus, the results from the Swiss census probably overestimate the erosion of male regular work to some extent.

3 Development of regular work in Germany

In order to analyze the changes in regular work in Germany, we employ data from the German Socio-Economic Panel (SOEP; cf. SOEP Group, 2001). The SOEP data have been collected since 1984 using an initial random sample of 6,000 private households. All of the members of these households older than 16 years of age are interviewed on a yearly basis (this resulted in an initial yearly total of approximately 12,000 individual records).7 We use the weighted data and restrict our analyses to the old Bundeslaender (West Germany) for two reasons: first, no suitable data material is available on East Germany before 1990; and, secondly, East Germany underwent a radical transition from a planned to a market economy in the 1990s and thus cannot be directly compared to either Switzerland or West Germany. In our analyses for West Germany we distinguish again between those in the labour force (the employed and the unemployed) and those not included in the labour force, and separate the employed into the self- employed, trainees and wage and salary workers. A further differentiation separates those people employed full-time, part-time and marginally (based on self-assessment) and takes into account the length of the work contract. Here, too, we define regular work as permanent full-time wage and salary employment (as in the analyses for Switzerland, it does not seem to be necessary to perform a further differentiation with respect to social security deductions). In the SOEP, the classification into the employed and the non-employed is based primarily upon the respondents’ self-assessment. However, in order to preserve parity with the Swiss analysis, people who characterized themselves as non-employed are considered employed for the purposes of our analysis if they reported holding 202 The Erosion of Regular Work down a regular sideline job (with a workload of at least one hour a week) later on in the questionnaire. To begin with, we consider again the sample of the population aged 16 to 64. In the time span from 1990 to 2002, the proportion of regular work diminished by 5.5 percentage points, from 60.8 to 55.3 per cent for men in West Germany (see Figure 8.4 or Table A8.3 in the appendix). For women, the share of regular work remained roughly at the same level throughout the 1990s (1990: 27.0 per cent, 2002: 25.9 per cent). In contrast to Hoffmann and Walwei (1998), we thus find a fairly substantial reduction in the male rate of regular work. This discrepancy may be partially explained by the fact that Hoffmann and Walwei conducted their analyses with different data (the German microcensus) for a different time frame (1985 to 1995). They report a reduction in regular work for men of only one percentage point (from 66 per cent in 1985 to 65 per cent in 1995; see Table 1 on p. 416 and Figure 7 on p. 421). However, there is actually substantial agreement between our analyses of the SOEP data and their results, since the SOEP data similarly provide little evidence for an erosion in regular work for men between 1985 and 1995. The percentages reported by Hoffman and Walwei are larger than those in our study because they are based only on the total numbers of employees. If the SOEP results are translated to proportions based upon employees, tendencies similar to those observed by Hoffmann and Walwei are obtained for the 1985/1995 comparison: a slight decrease in regular work for men and a somewhat greater decline for women. The reason for the different results for the 1985–1995 time frame compared to the analyses of the years 1985 to 2002 is the inverted U-shaped development of the rate of regular work, which increased even up to the end of the 1980s. A clear decrease cannot be observed until the 1990s (see Figure 8.4; Table A8.3 in the appendix). How can the erosion of regular work in Germany be explained? As in the case of Switzerland, the decline may not be attributed to an increase in temporary jobs. In Germany as well as in Switzerland, and again confounding expectations, we found tendencies toward a dip in temporary employment. The shrinkage of permanent full-time work for men is instead due to a shift toward more part-time work, more marginal employment, more self- employment, more unemployment and more non-employment. Each of these domains accounts for roughly one to two percentage points of the reduction in regular work for men in the 1990s. In contrast, overall female labour force participation has grown (a reduction in non-employment can be observed from 37.9 per cent in 1990 to 32.2 per cent in 2002). For women, the proportion of part-time work, marginal employment, and unemployment has been on the increase (Table A8.3). In order to exclude the effects of changes at the beginning and end of the occupational career, we also perform the same analysis for the 25-to-55- year-old age group in West Germany. Here again, we find a clear decrease 203

(a) Men (aged 16 to 64) 100 90.4 89.9 86.5 90

70 59.5 60.8 55.3 60

Percent 40

0 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02

(b) Women (aged 16 to 64) 100

70 56.7 60 54.9 47.6 50

Percent 40 27.0 25.9 30 24.3 20

0 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 Denominator : Wage and salary workers Population

Figure 8.4 Regular work in Germany from 1985 to 2002 (West German states only) Source: German Institute for Economic Research (DIW), German Socio-Economic Panel (SOEP); calculation: the number of wage and salary workers employed non-temporarily on a full-time basis divided by the number of wage and salary workers in total or the total population aged 16 to 64; grey area: approximate 95 per cent confidence interval. 204 The Erosion of Regular Work in regular work for men of 5.4 percentage points – from 75.8 per cent in 1990 to 70.4 per cent in 2002. This shift occurred mainly during the first half of the 1990s. In contrast, regular work showed a slight increase for women in this age group, from 29.9 to 30.5 per cent, as the labour force participation rate for women aged 25 to 55 grew markedly, by 8.7 percentage points after 1990 and 16.1 percentage points after 1985. We are therefore unable to confirm the tendencies and hypotheses reported by Hoffmann and Walwei (1998), Wagner (1998) or Kaiser (2001a). During the 1990s, a substantial erosion of regular work was observed. It is true that labour force participation, part-time and marginal employment for women have grown during this period. It is also true that a decrease in labour-force participation may be observed for both young men and women (aged 16 to 24) and older men (aged 55 to 64). However, these three tendencies – more education for younger people, an earlier retirement age for older workers, and more part-time work for women – can by no means explain the decrease in regular work for men. This is so because we also find a reduction in regular work of approximately 5 percentage points for men aged 25 to 55 in the period 1990 to 2002.

4 Summary: parallelism instead of special paths

A comparison of labour-market activity in Switzerland and Germany in the 1990s shows clearly that developments in the two countries have more similarities than differences. In both countries, female labour-force participation is increasing while male labour-force participation has fallen slightly. The reasons for the declining employment rate of men are, among others, a longer period of formal education and an earlier retirement age. In both countries the proportion of part-time jobs has grown for both men and women. However, female part-time employment is higher than male part-time employment and has also grown more markedly. In contrast to what many have suggested, the share of temporary full-time contracts has not grown in either (West) Germany or Switzerland. Permanent full-time work is still the bread and butter of men’s labour: it remains the most common pattern of employment for men. Nevertheless, the proportion of male work that is regular is falling in both countries. This trend can be accounted for by shifts towards more part-time work, more self-employment, higher levels of unemployment and lower male labour- force participation. The process, if active before at all, gained in intensity in Switzerland and Germany primarily in the 1990s. According to the Swiss census, between 1990 and 2000 a decline in the share of regular work of 8.9 percentage points can be observed in Switzerland for men aged 25 to 55 (or, according to the SLFS data, 5.5 percentage points between 1991 and 2000, or 3.0 percentage points between 1992 and 2000). Translated into popula- Andreas Diekmann and Ben Jann 205 tion figures this means that around 100,000 fewer men in that age segment are employed in a permanent full-time job today, than would have been the case if the rate of regular work was still at the 1990 level. In Germany, the decline amounts to 4.9 percentage points. Here, we are talking about a ‘gap’ of around 700,000 regular jobs for men in the 25-to-55-year-old age group. By contrast, the aggregated quotas for women are not very informative because their labour-force participation pattern over the life and family cycle has changed significantly during the last two decades. Only an examination of age-dependent rates of regular work (see the curves plotted with the Swiss census data in Figure 8.3) would allow us to draw any mean- ingful conclusions about the rate of regular work for women. Here, changes in the aggregated measures result from the effects of forces that are, to some extent, in opposition. On the one hand, the proportion of regular work declined for younger women in the 1990s. On the other hand, there was an increase in the share of regular work for women in their mid-twenties to mid-thirties. The same is also observed in Germany: full-time employment has increased for women who have completed (a perhaps longer) period of education. The labour-force participation rate – that is, the proportion of people from the working-age population who are either actively working or actively seeking work – increased by one to two percentage points in Germany and Switzerland in the 1990s. This increase was the result of growing female participation in the workforce and mostly represented an increase in part- time employment for women; it does not, however, explain the decrease in the rate of regular work for men. A decrease in regular work can also be observed in separate analyses for men in the 16-to-64- and 25-to-55-year-old age groups. Men, in fact, substituted other kinds of work for full-time wage and salary employment during the 1990s.8 Thus, at least for the 1990s, it is the substitution and displacement hypothesis (Hypothesis 1 in the introduction) that is supported by our results. The attendant process may be described as follows: over the past few decades there has been a distinct growth in the female labour-market participation rate. However, the proportion of women employed in regular jobs has remained more or less stable in an absolute sense (i.e., as a fraction of the total female population of working age). This means that the expansion in female labour-force participation occurred primarily through the take-up of additional part-time jobs. Until the end of the 1980s, the proportion of men doing regular work also remained stable. Thus, the ‘erosion of regular work’ occurred only in a relative sense (Hypothesis 2) in that period. In the 1990s, however, not only did female part-time work continue to increase, but the male rate of regular work started to decline, thus lending further support to Hypothesis 1, which assumes ‘real’ erosion. Hence, one of the main findings of this study is a clear shift from 206 The Erosion of Regular Work regular to non-regular work occurring in both countries during the 1990s, one that does not, however, involve a shift to more temporary full-time work. And, despite this trend of erosion of standard work contracts, regular work remains the dominant pattern of work among the male labour force. Finally, what of the effect of the different labour-market regulations in the two countries? It is striking that, despite the different labour-market regimes, the development of different dimensions of employment and the decrease in regular work in particular have occurred largely in parallel in the two neighbouring countries. Of course, the comparison we have performed here does not provide any clear insight into the causal relationships at play (as would a controlled experiment). Furthermore, Germany and Switzerland differ not only with respect to the flexibility of their labour markets. For instance, the unemployment rate is considerably higher in Germany than in Switzerland. Unemployment could be a ‘push factor’, one which exerts its influence on the erosion of regular work according to the mechanism postulated by Hoffmann and Walwei (1998). Thus, one could argue that the flexible labour-market regulations in Switzerland have promoted the erosion of regular work there, whereas the erosion in Germany has occurred despite its more rigid labour market due to the high unemployment rates. This may well be the case, but it is a claim that cannot be studied using the data at hand. One can also speculate on a variety of other mechanisms. One mechanism may be that, despite its higher degree of rigidity in general, German labour market law provides for one important exception. There is a part-time, low-income sector in Germany that is at least partially exempted from paying taxes and social security. The increasing number of jobs in this segment has also contributed to the growth of atypical work in Germany. Apparently, the level of labour-market flexibility was not the only factor affecting the erosion of regular work in Germany and in Switzer- land in the 1990s. Going a step further, it may well be that overall labour-market flexibility has little or no effect on the extent of the shift from regular to non-regular work (Kress, 1998). There is also a thesis that, on the contrary, labour-market rigidity promotes atypical work relations. The reason is that employers in rigid labour markets try to evade strict regulation (and high labour costs) by shifting their employment structure to a workforce with a smaller proportion of regular work contracts (Appelbaum, 2002). In both countries, there are probably forces leading to the erosion of regular work on both the demand side and the supply side of the labour market. For instance, higher capital mobility has increased employers’ bargaining power. Reorganization, cost-cutting, the downsizing of the workforce, and the outsourcing of activities not defined as the company’s core business has Andreas Diekmann and Ben Jann 207 occurred in both countries. However, we should not confine our attention exclusively to employers’ strategies. It might well be the case that part of the change came about as the result of changing preferences on the supply side of the labour market. Of course, many atypical work relations belong to the category of poorly paid and insecure jobs, and many less-qualified workers have no alternative other than to apply for ‘bad’ jobs. As Kalleberg et al. (2000) demonstrate in the case of the United States, non-standard work is indeed correlated with unfavourable job characteristics. However, this is only a correlation – not all part-time jobs are insecure or yield earnings below the average. That women’s increasing labour-force participation was accompanied by an increase in the number of part-time occupations clearly illustrates that the changes towards more atypical employment may also be supply-driven. Labour-force participation patterns among married women have changed dramatically over the last few decades. Although they have maintained the main responsibility for household work in most cases, many married women have also started to seek work opportunities outside the family. In order to manage successfully both family responsibilities and paid labour, a large number of women do not apply for full-time employment. Possibly, a small fraction of male workers with family duties also prefer to work shorter hours. Employers have met these expectations by raising the proportion of part- time jobs, a process that has taken place irrespective of the degree of labour market flexibility.

Notes

1. The Kommission für Zukunftsfragen was formed by the German states of Saxony and Bavaria in order to analyse labour-force participation, unemployment and similar issues. 2. We will use terms like ‘standard employment’, ‘normal employment’, ‘regular work’, and ‘normal work’ interchangeably to refer to the ‘traditional’ model of full-time wage and salary employment based on a permanent labour contract. 3. Furthermore, there is a large body of literature on the changes in professional careers, for example Osterland (1990), Berger et al. (1993), Mutz et al. (1995), or Bernhardt et al. (2001) to name just a few, but we do not intend to address this issue here explicitly. See the contribution by Widmer et al. (chapter 2 in this volume) for a study on the pluralization of life courses. 4. Note that fixed working times and/or the spatial division of the private household and the workplace are sometimes considered as further characteristics of standard employment (e.g., Beck, [1986] 1992; Seifert, 1993). 5. It must be pointed out, though, that the 70 per cent response rate refers only to the yearly base sample of respondents who are being interviewed for the first time. 208 The Erosion of Regular Work

Because the SLFS employs a rotating panel (each year, one-fifth of the sample is replaced by new subjects, i.e., each respondent stays in the panel for five years), the impression given by this figure is actually too optimistic. As a matter of fact, the response rate drops down to approximately 55 per cent if the yearly drop-out rates for the panel cohorts are fully accounted for (year 2002 is an exception, since it had a response rate of approximately 65 per cent because of the large increase in the size of the sample). It should also be mentioned that proxy interviews are possible in the SLFS if a targeted person cannot be interviewed due to old age or illness. 6. Classifying all the respondents was somewhat problematic, given the incomplete records for some of the respondents. Simply excluding cases with missing information could result in serious distortions because missing values can occur at different branching points of the classification tree. Assume, for instance, that all full-time workers with missing information about the duration of their labour contracts are excluded. The number of full-time wage and salary workers as a fraction of all workers would then be underestimated systematically, since workers like the self- employed and non-employed would not be excluded under this regime, as their records cannot contain missing values for the fixed-term contracts variable. It is therefore not possible to represent the distribution of labour relations using a single classification variable. We solve the problem through the use of indicator variables that reflect the value of each level of the classification variable. The distribution of employment forms can then be estimated consistently by multiplying the means of different indicator variables, despite the fact that the missing values may be distributed among the levels of the classification tree in a systematic way. This solution rests on the assumption that missing information is unsystematically distributed among the different categories at a given level of the tree. The same procedure was applied in the analyses of the Swiss censuses and the German Socio- Economic Panel (see below). 7. Over the years, the SOEP’s initial household sample has been supplemented or increased by additional samples several times. In 1990, for instance, an East German sample of 2,000 households was added. In 2000, the SOEP sample was enhanced by another 6,000 households. 8. It must be noted that not every substitution necessarily has negative consequences. If, for instance, an employee becomes self-employed and creates an additional part- time job in his new small-scale firm, statistically this change will cost one regular position. The same is true for a full-time job that is divided into two part-time jobs to enable job sharing. In both cases, welfare may be increased despite the loss of a regular work job. 209 cases

Non-employed Total Population of No.

Self-employed Trainees Unemployed Full-time Part-time Part-time Full-time

Labour force force Labour

Part-time 1–14

a + 15

Wage and salary workers Distribution of employment forms in Switzerland from 1991 to 2003 Permanent Temporary Temporary Permanent Men aged 16 to 64 Appendix A8.1 Table Year Year Full-time Part-time Full-time 1991 65.80 2.85 2.76 1.17 12.78 0.79 4.50 1.06 4.50 100.000.79 2,263,5601991 8.29 12.78 1992 63.091.17 5,910 65.80 2.85 2.76 1993 61.30 4.47 1994 59.96 3.60 2.83 1.30 12.39 1995 59.85 3.60 3.08 0.87 1.41 13.36 5.13 1996 58.36 3.50 3.03 0.80 1.21 1.91 12.80 5.19 1997 57.32 2.73 3.44 1.13 1.17 2.65 13.84 5.28 1998 58.00 2.42 3.13 2,295,923 1.17 1.36 2.96 15.44 8.01 5.03 1999 58.32 2.70 3.81 100.002,314,825 1.20 6,260 1.28 2.35 15.422,329,325 8.61 5.43 2000 58.74 2.62 3.84 100.00 1.31 6,638 10.04 1.41 2.70 15.50 100.00 5.23 2001 58.94 6,619 1.77 3.96 2,340,825 1.31 1.33 3.26 15.53 9.64 5.31 2002 11,791 56.68 1.88 3.95 100.002,341,741 1.62 1.50 2.54 15.83 9.66 5.04 2003 55.19 2.30 4.48 100.002,359,195 1.69 5,937 1.74 2.15 15.01 9.95 5.14 2.34 4.27Women aged 16 100.002,364,105 to 64 1.74 6,049 1.44 1.67 14.92 9.39 5.17 4.821991 24.70 100.00 2,364,677 1.77 6,115 1.52 1.33 14.86 9.43 6.07 1992 22.25 100.00 2,371,852 1.65 2.09 6,539 2.08 9.72 5.92 11.44 17.38 1993 20.76 100.002,380,563 2.16 6,453 4.44 2.71 2,405,042 9.71 11.84 18.21 4.63 1994 20.31 100.00 1.40 6,688 10.47 14,531 3.87 2.40 100.002,433,196 11.64 18.38 6.04 1.58 10.99 14,722 1.75 4.36 3.00 100.00 11.96 19.03 6.37 2.37 3.722,265,638 3.88 6.16 31.16 3.05 100.002,286,130 2.99 6,902 30.26 2.88 100.002,299,081 7,238 30.17 100.002,329,387 7,665 31.35 100.00 7,616 210 cases

Non-employed Total Population No. of

Self-employed Trainees Unemployed Full-time Part-time Part-time Full-time

Labour force

Part-time 1–14

a + 15

Wage and salary workers (Continued) A8.1

Part-time Full-time Year Permanent Temporary Temporary Permanent 1995 20.40 1996 20.43 1.86 11.39 19.45 1997 19.37 1.42 3.95 10.77 20.32 5.59 1998 20.91 1.47 4.38 3.24 11.51 20.95 6.20 1999 20.14 1.33 2.54 4.68 3.35 11.65 21.03 6.50 2000 19.75 1.47 2.68 4.392,333,265 3.14 12.15 21.66 6.58 2001 31.57 18.90 1.65 13,103 2.68 100.00 4.582,340,395 3.19 11.39 22.38 6.77 2002 30.46 18.72 1.51 2.58 4.26 100.002,344,642 3.48 6,658 12.03 23.67 6.84 2003 29.70 17.36 1.51 2.28 100.00 4.282,346,800 3.76 6,616 12.11 24.72 6.64 28.32 1.51 Men aged 25 to 55 2.05 100.00 4.252,353,265 3.78 6,682 12.82 24.72 6.78 27.48 2.35 4.11 100.002,365,064 3.76 7,214 7.16 27.93 1.97 100.002,383,776 3.93 7,237 26.82 2.52 100.002,403,976 7,569 26.18 16,673 100.002,426,049 25.88 16,978 100.00 Table 1991 74.60 1992 72.16 2.97 1993 71.67 4.55 2.90 0.63 15.05 1994 71.25 3.52 3.06 0.73 0.56 14.95 0.07 1995 70.44 3.90 3.53 0.67 0.59 0.74 15.77 0.35 1996 69.50 3.64 3.32 0.57 0.54 1.57 15.22 0.19 1997 68.04 2.56 3.93 1,557,793 0.55 0.57 1.90 16.15 2.30 0.12 1998 68.33 2.16 3.27 100.001,575,246 0.81 4,287 0.61 2.68 17.75 2.12 0.17 1999 68.81 2.59 4.02 100.001,612,990 0.84 4,576 0.70 1.90 18.10 2.26 0.21 2000 69.16 2.62 4.03 100.001,631,930 1.07 4,910 0.61 2.64 18.10 2.42 0.17 1.52 4.16 100.001,642,098 1.11 4,892 0.40 3.03 17.99 2.39 0.12 4.30 100.001,652,428 1.33 8,651 0.74 2.22 18.34 2.61 0.18 100.001,666,817 1.34 4,359 1.77 2.71 0.24 100.001,661,045 4,460 1.26 2.90 100.001,655,277 4,505 2.74 100.001,659,331 4,820 3.10 100.00 4,741 211

(SLFS), the authors’ own calculations (weighted, post-stratification neglected). neglected). post-stratification (weighted, calculations own authors’ the (SLFS), Part-time, working 1 tohours 14 per week. b Swiss Labour Force Survey Force LabourSwiss Part-time, working 15 or more hours per week. week. per hours more or 15 working Part-time, a : Swiss: Federal Statistical Office, : 2001 70.02 2002 68.13 1.77 2003 66.14 2.15 4.66 0.80 17.18 2.12 4.85Women aged 25 to 55 1.30 0.67 17.49 0.21 5.41 1.49 0.63 0.83 17.60 0.09 1.29 1.55 0.20 1,654,683 2.27 3.23 100.001,661,455 4,846 3.58 10,357 100.001,669,200 4.33 10,300 100.00 Notes 1991 23.83 1992 21.30 1.37 13.66 21.13 1993 20.55 1.96 5.19 14.04 22.13 5.29 1994 20.32 1.26 4.36 0.08 13.38 21.99 7.29 1995 21.22 1.51 1.82 4.79 0.12 13.70 22.90 7.53 1996 21.45 1.60 2.53 4.231,519,564 0.24 13.14 23.50 7.09 1997 27.63 20.39 1.19 3.10 4.53 100.001,548,827 0.15 4,833 12.22 24.81 6.53 1998 26.27 22.77 1.18 2.82 100.00 4.611,579,081 0.11 5,178 12.92 25.50 7.10 1999 27.15 21.84 0.86 2.73 100.00 5.101,600,309 0.13 5,480 12.92 25.57 7.31 2000 27.27 21.87 1.09 2.94 4.92 100.001,620,044 0.09 5,515 13.19 25.97 7.25 2001 26.66 20.46 1.26 2.92 100.00 5.561,633,972 0.12 9,409 12.51 27.17 7.31 2002 25.56 20.57 1.12 2.69 100.00 4.871,639,456 0.09 4,816 12.88 28.78 7.57 2003 19.08 24.59 1.16 2.26 5.11 100.001,648,122 0.15 4,830 13.67 30.18 7.79 22.90 1.22 2.19 100.00 5.111,648,273 0.16 4,830 14.52 30.03 7.24 22.68 2.36 4.78 100.00 1,648,110 0.13 5,207 7.95 22.41 2.09 100.001,656,973 0.06 5,199 21.33 2.32 100.001,664,110 5,416 19.85 11,758 100.001,667,737 20.05 11,802 100.00 Source 212

6 1.56 4.54 2.82 11.62 100.00 2,433,588 4.54 2.82 6 1.56 100.00 11.62 3.17 2.58 3.37 0.48 47.54 100.00 2,076,787 2,285,526 3.37 0.483.17 2.58 100.00 2,407,303 3.47 47.54 1.653.60 3.23 100.00 3.45 39.21 3.585.18 4.29 100.00 31.61 13.19 0.73 4.80 1.60 7.98 100.00 2,364,289 4.80 1.6013.19 0.73 100.00 7.98 Full-time Part-time b Part-time 6–24 Part-time a +

Distribution of employment forms in Switzerland from 1970 to 2000 Table A8.2 Table Year Wage and salary workers Self-employed Labour force Trainees Unemployed Non-employed Total Population Full-time Part-time 25 Part-time Full-time 1970 30.45 1980 28.89 5.56 1990 29.34 4.66 2000 27.00 6.59 4.44 9.96 9.31 12.91 2.81 14.92 3.20 2.29 0.11 1,995,687 51.15 100.00 Women aged 16 to 64 Men aged 16 to 64 1970 73.24 1980 71.11 0.74 1990 68.98 0.96 2000 59.54 1.42 0.92 2.50 1.14 13.34 1.31 0.57 12.02 4.56 1.64 0.63 5.49 0.15 15.7 0.70 6.48 2,082,729 100.00 7.95 2,004,668 100.00 213 own calculations. calculations. own

Part-time, working 6 to 24 hoursper week (1970: 6to 19). b 5.84 5.12 0.21 3.87 25.91 100.00 1,673,730 0.21 3.875.84 5.12 100.00 25.91 3.12 3.87 0.04 0.08 54.87 100.00 1,257,238 1,344,5110.04 0.083.12 3.87 100.00 0.06 54.87 0.353.75 3.19 100.00 48.79

Part-time, working 25 or more hours per week (1970: 20 or more). more). or 20 (1970: week per hours more or 25 working Part-time, a : Swiss: Federal Statistical Office,harmonized personalrecords from the 1970, 1980, 1990and 2000 Swiss Censuses,the authors’ : Men aged 25 to 55 Notes 1970 81.24 1980 80.64 0.55 1990 77.04 0.88 2000 68.15 1.55 0.84 Women aged 25 to 55 2.83 0.97 14.78 1.20 0.42 13.87 0.03 1.44 0.54 15.11 0.10 0.13 0.71 17.71 0.14 0.61 1,288,132 1.53 0.25 1.51 2.02 1,363,497 100.00 2.63 2.39 1,615,084 100.00 2.75 1,697,669 100.00 5.44 100.00 1970 26.29 1980 26.31 6.44 1990 28.71 5.67 2000 28.90 8.15 12.10 5.30 11.88 16.22 18.04 4.29 4.07 0.11 1.71 1,535,819 36.73 100.00 Source 214 of cases

Distribution of employment forms in Germany from 1985 to 2002 (West German states only) Part-time Marginally Full-time Part-time Full-time Permanent Temporary Men aged 16 to 64 Wage and salary workers 198558.17 1986 5.67 59.48 59.62 1.51 4.78 1987 4.63 0.85 60.74 0.75 8.20 0.78 0.80 1988 4.02 0.784,841 3.72 1.29 61.08 4.68 20,422,095 8.27 0.39 0.83 7.86 4.00 1.14 1989 5.10 4,777 4.47 1.17 Self-employed 5.72 60.76 5.12 20,733,823 7.82 1.00 0.82 5.051990 16.09 100.004,534 4.23 0.73 Trainees60.43 4.82 20,726,705 7.26 1.45 0.86 15.84 100.001991 4.88 14.45 100.004,381 4.39 Unemployed 1.17 20,189,16761.06 3.85 20,902,168 1.14 7.71 0.96 1992 5.04 15.03 100.004,273 3.80 1.43 59.83 3.66 5,008 21,522,617 7.64 1.42 1.05 5.341993 15.87 100.004,256 4.25 1.87 57.50 3.32 22,007,277 7.66 1.37 0.58 1994 4.06 15.01 100.004,171 5.14 1.33 59.05 4.31 21,826,271 0.88 7.66 0.85 1995 3.77 15.27 100.004,142 3.20 1.94 57.60 5.56 21,952,073 7.67 1.20 1.44 3.311996 15.25 100.004,003 3.73 2.26 56.00 7.26 21,857,771 7.31 1.81 0.65 1997 3.50 15.37 100.004,233 4.23 2.43 56.59 7.17 21,979,018 2.43 7.36 1.32 1998 3.30 14.86 100.004,099 3.93 2.35 55.41 7.16 22,049,117 7.80 2.67 0.96 3.151999 15.66 100.004,009 5.04 2.36 56.16 7.69 22,087,813 7.71 2.22 0.97 2000 3.38 15.28 100.004,462 3.77 2.99 55.36 7.11 21,857,741 2.35 8.70 0.83 2001 3.52 15.39 100.004,260 3.27 2.68 55.32 5.98 21,906,136 8.25 3.12 0.94 2002 3.07 15.29 100.007,655 3.39 2.64 5.42 21,946,941 8.84 2.37 0.59 3.05 15.31 100.006,811 2.89 5.79 21,694,990 8.68 0.57 3.26 17.36 100.006,366 5.95 21,242,499 17.32 100.00 17.58 100.00 Table A8.3 Table Year force Labour Non-employed Total Population No. 215

1985 73.85 1985 3.20 73.51 0.84 1986 4.09 0.82 10.13 0.27 0.35 74.24 1.72 1987 3,305 3.78 0.28 6.08 75.20 12,860,007 0.59 9.71 0.55 1988 0.33 3.63 0.67 3,152 77.07 4.97 9.81 0.76 0.63 0.471989 4.45 100.00 4.13 0.6013,095,097 3,081 75.79 5.38 9.30 0.81 0.32 1990 0.46 3.97 13,176,070 0.15 4.84 100.002,909 75.03 4.80 0.99 9.00 0.57 1991 0.35 3.59 0.5313,289,473 4.44 100.002,825 74.22 3.51 9.56 1.01 0.66 0.821992 3.52 13,541,661 0.74 4.92 100.002,822 3.13 9.76 1.55 0.41 1.47 14,310,397 0.72 4.40 100.002,822 3.19 9.56 0.53 0.45 14,537,189 4.55 100.002,804 3.86 14,703,180 4.81 100.00 5.59 100.00 4,961 Women aged 16 to 64 20,862,784 4,755 24.26 2.83 13.142.63 2.57 2.59 4.31 2.69 1985 21,082,748 4,651 23.93 2.96 10.903.27 4.51 3.51 100.00 4.61 3.24 44.991986 20,947,940 4,433 24.36 3.00 13.144.26 3.01 3.74100.00 4.00 3.19 43.08 1987 21,091,4314,299 25.55 1.96 13.704.57 3.08 3.44 100.00 4.00 3.16 41.31 1988 21,174,633 4,183 27.51 2.40 13.884.05 3.04 3.25 100.00 3.62 2.72 40.531989 21,458,245 4,172 27.04 3.20 15.803.67 3.17 3.72100.00 3.07 2.42 39.53 1990 21,247,912 4,145 27.51 2.91 16.443.84 3.87 3.15100.00 2.55 2.31 37.91 1991 21,691,267 4,092 27.92 2.50 17.193.49 2.47 2.90 100.00 3.31 2.62 37.421992 21,931,273 4,039 27.55 2.93 17.243.27 3.04 3.04100.00 3.85 2.69 37.59 1993 22,066,750 4,285 25.87 3.27 16.292.57 4.32 2.70 100.00 5.34 2.49 36.39 1994 22,068,726 4,174 27.60 2.00 16.512.30 4.81 2.70 100.00 5.82 2.58 37.141995 22,071,282 4,077 27.11 1.58 17.492.12 5.20 3.24100.00 6.49 2.51 35.68 1996 22,182,854 4,564 25.96 2.12 16.902.61 4.77 2.25100.00 6.23 3.05 34.27 1997 22,285,640 4,355 26.54 2.09 16.792.28 4.97 2.45 100.00 5.68 2.61 36.111998 22,420,314 7,964 27.30 2.25 15.942.47 6.80 2.63100.00 4.77 2.85 36.59 1999 22,538,781 7,172 25.65 2.43 18.802.15 6.31 2.59 100.00 3.84 2.64 34.99 2000 22,663,833 6,761 26.43 2.20 19.273.19 6.06 2.55 100.00 4.73 2.84 35.592001 22,590,995 25.87 2.06 19.863.06 6.53 2.40100.00 5.67 2.39 32.73 2002 100.00 32.16 Men aged 25 to 55 216 of cases

Total Population No. Total Population

(Continued) 1985 26.29 1.83 18.81 2.89 3.10 3.27 0.09 4.00 39.74 100.00 12,948,072 3,232 12,948,0723,098 26.29 1.83 18.810.09 2.89 3.27 4.00 3.10 1985 13,024,2373,027 26.60 1.63 15.010.04 5.60 4.40 100.00 4.31 3.86 39.74 1986 12,997,2502,865 27.13 2.14 18.380.14 3.18 4.72 100.00 3.81 3.86 38.56 1987 13,216,4732,782 28.08 1.26 19.380.15 3.46 4.65 100.00 4.03 3.99 36.641988 13,448,1602,729 30.20 1.32 19.330.16 3.37 3.66 100.00 3.20 3.58 35.01 1989 13,854,7062,750 29.94 2.15 22.140.27 3.23 4.06 100.00 3.33 2.58 35.18 1990 14,029,205 31.12 1.96 22.140.57 4.44 3.77 100.00 2.38 2.77 32.291991 100.00 30.85 Full-time Part-time Marginally Full-time Part-time Full-time Permanent Temporary 72.82 1993 4.02 Wage and salary workers 69.66 1.44 1994 4.61 0.61 0.83 9.83 0.35 1995 0.55 70.40 1.38 2,799 5.15 9.70 1.23 3.23 1996 0.43 69.68 14,974,043 2,755 6.91 3.93 1997 67.65 15,083,064 1.45 5.25 100.00 1.73 5.04 1998 68.80 2.22 Self-employed 5.24 100.00 2.03 4.42 1999 69.10 9.23 2.51 0.56 Trainees 1.69 5.47 2000 9.15 70.93 0.72 Unemployed 3.06 1.23 1.53 3.85 2001 0.40 9.86 69.68 6.61 2.34 0.88 10.00 2.09 3.43 2002 0.77 70.40 0.43 6.17 2.11 11.47 0.77 2,941 1.63 15,105,805 3.87 0.37 7.57 3.19 6.08 Women aged 25 to 55 0.45 10.33 6.70 2,857 100.00 1.02 14,967,827 0.62 2.43 5.20 0.28 11.16 4.29 2,768 100.00 1.45 14,880,798 0.38 3,063 4.38 14,823,338 10.60 0.22 4.91 100.00 3.96 0.41 2,925 100.00 14,675,284 0.36 5.45 4.42 5,226 100.00 14,674,425 5.36 5.33 4,639 100.00 14,547,600 5.48 4,330 100.00 14,262,019 5.13 100.00 Table A8.3 Table Year Labour force Non-employed force Year Labour 217 : German Institute for Economic Research (DIW), German Socio-Economic Panel (SOEP), the authors’ own calculations (weighted). (weighted). calculations own authors’ the (SOEP), Panel Socio-Economic German (DIW), Research Economic for Institute German : 1995 31.49 1.83 21.20 5.36 3.28 2.72 0.32 5.44 28.36 100.00 14,968,278 2,946 14,968,278 31.49 1.83 21.200.32 5.36 2.72 5.44 3.28 1995 100.00 28.36 1992 31.81 2.03 22.03 2.70 2.93 3.40 0.52 3.33 31.26 100.00 14,412,387 2,765 14,412,3872,752 31.81 2.03 22.030.52 2.70 3.40 3.33 2.93 1992 14,817,4262,737 30.71 2.39 21.580.47 3.71 3.42 100.00 3.65 2.95 31.261993 14,961,653 29.72 2.73 20.180.44 5.29 2.94 100.00 5.28 3.03 31.13 1994 100.00 30.39 Source 1996 30.51 1.38 22.36 5.36 2.97 3.76 0.34 6.43 26.89 100.00 14,923,417 2,889 14,923,4172,844 30.51 1.38 22.360.34 5.36 3.76 6.43 2.97 1996 15,030,8973,166 31.15 2.14 22.100.17 4.75 2.21 100.00 6.05 3.88 26.89 1997 15,127,4583,003 31.93 2.08 21.840.40 4.80 2.53 100.00 5.10 3.45 27.541998 14,905,9755,462 32.46 2.42 20.450.55 7.13 3.19 100.00 4.61 3.88 27.86 1999 14,911,6474,909 30.68 2.29 24.530.15 6.31 2.92 100.00 3.63 3.49 25.31 2000 15,005,5684,616 32.29 1.99 24.670.27 6.19 2.52 100.00 4.72 3.88 26.012001 15,039,052 30.53 2.27 25.400.44 6.31 2.54 100.00 5.63 3.25 23.47 2002 100.00 23.61 Part III Political Institutions 9 Internationalization and Domestic Politics: Evidence from the Swiss Case Pascal Sciarini and Sarah Nicolet

This chapter contributes to the literature on the political consequences of the globalization process by analysing its effects on national decision-making systems. While most works deal with the policy dimension of internationalization, we provide evidence on its polity and politics implications. Furthermore, we pay close attention to the transmission mechanisms at work by differentiating between direct and indirect forms of internationalization and contrasting them with cases sheltered from international influences. Based on both quantitative and qualitative techniques, our empirical tests reveal that even a neutral and relatively isolated country like Switzerland cannot escape the consequences of internationalization. Internationalized decision-making processes display distinctive characteristics with respect to formal consultation, conflict level and power configuration. Our tests also highlight some clear differences between direct and indirect internationalization.

1 Introduction

This chapter provides a contribution to the burgeoning literature on the political consequences of the globalization process at the national level. Since the early 1990s a growing number of studies have highlighted the consequences of this process on national public policies, on the nation-state’s steering capacity, on the preferences of domestic actors, and on the political cleavages within national contexts (e.g., Andeweg, 1995; Bartolini, 1999; Garrett and Lange, 1995; Hix, 1999; Keohane and Milner, 1996; Lachat and Kriesi, 2003; Mair, 2000; Rhodes, 1998; Scharpf, 1998). To our knowledge, however, few studies have analysed the effects of globalization on national decision-making systems. In the current era of globalization internal and international politics are increasingly entangled. By lowering the boundaries among the nation-states the process of globalization – or, as some prefer to call it, the process of ‘denationalization’ (e.g. Zürn, 1998) – leads to the internationalization of

221 222 Internationalization and Domestic Politics domestic politics. Issues that used to be domestic in nature are now strongly influenced by economic, political or technological developments at the global level. This influence mainly takes two forms. The first and most straightforward influence results from an international negotiation or from a membership in a supranational organization. We call this form ‘direct internationalization’. But internationalization may arise even in the absence of formal international negotiations. This is typically the case when a state adapts unilaterally to existing international or supranational rules, or to less formal economic or political pressures – for example, when a state modifies its legislation in order to increase its locational attractiveness.1 We term this form of influence ‘indirect internationalization’. In the European context internationalization means first of all European- ization. With the deepening and widening of the European Union (EU), the domestic consequences of European integration have increased dramatically for EU member states (e.g. Anderson, 2002; Börzel and Risse, 2000; Green Cowles etal., 2001; Haverland, 2000; Héritier et al., 2001; Knill and Lehmkuhl, 1999) and non-EU member states alike (e.g. Ágh, 1999; Fischer et al., 2002; Lavenex and Uçarer, 2002; Mach et al., 2003; Sciarini et al., 2004). Our contribution to the internationalization literature is threefold. First, the bulk of this literature focuses on the extent to which national policies are adapted to internationalization, that is, it mostly treats the policy aspect of domestic change as a dependent variable in the analysis. By studying the impact of internationalization on the institutions of decision making, on conflict and on power configuration, we provide evidence on the polity and politics implications of internationalization. Secondly, by differentiating between direct and indirect forms of internationalization and contrasting them with cases where internationalization does not (or hardly) matter,2 we hope to deepen our understanding of the effects of internationalization and of the role played by the various transmission mechanisms. Finally, we focus on a country, Switzerland, which can be seen as a test case with respect to the impact of internationalization. As a neutral country that has only recently joined the United Nations (UN) and that is not a member of the European Union (EU) and will not become a member in the near future,3 Switzerland is – to a large extent – legally sheltered from the influence of international politics. If internationalization has an effect in the Swiss case, we can be confident that it also exerts an impact in other national contexts. This chapter is structured as follows. In section 2, we present the data and consider the methodology used. Methodologically, we innovate by combining quantitative analyses of more than two hundred legislative acts and a comparative case study of three key decision-making processes.The analyses of the influence of internationalization appear in section 3. We start by considering the impact of internationalization on the institutions of the decision-making process, and more specifically on the pre-parliamentary institutions. We proceed with its effects on the level of conflict and on the Pascal Sciarini and Sarah Nicolet 223 power configuration among domestic (public and private) actors. For each set of influences we first develop our hypotheses and then present our empirical findings. Section 4 concludes and suggests possible directions for further research.

2 Data and methodology

To test the effects of internationalization, we rely on a combination of a quantitative analysis and a comparative case study. On the quantitative side, we analyse a data bank of all legislative acts adopted by the Swiss Parliament during the 1995–99 legislature and subject to one of the three instruments of direct democracy (optional referendum, mandatory referendum, or popular initiative, i.e. 210 acts in total). This data set includes extensive information on these legislative acts (juridical type, importance, novelty, internationalization, etc.) and on the three phases of the decision-making process (pre-parliamentary, parliamentary and referendum).4 Regarding internationalization, our data bank enables us to distinguish between direct and indirect forms of internationalization. The former includes international treaties and related legislative adaptations. Among the latter, we can distinguish further between two categories: the cases of autonomous adaptation and a residual category of domestic legislative acts that are – to various extents – internationalized.5 We use statistical techniques such as multivariate linear and logistic regressions, and structural equation models (e.g. Arbuckle and Wothke, 1999; Byrne, 2001) to test the impact of internationalization on the institutions of the decision-making process and on the level of conflict. On the more qualitative side, we rely on an in-depth comparative analysis of three key decision-making processes that differ with regard to the nature and level of internationalization: the agreement on the free movement of persons with the EU (1994–98) as a case of direct internationalization, the reform of the telecommunication policy (1994–97) as a case of indirect internationalization (‘autonomous adaptation’), and the eleventh reform of the old age pension (1995–present), as an example of a ‘purely’ domestic political issue (control case).6 Resorting to the tools of network analysis (e.g. Laumann and Knoke, 1987; Wasserman and Faust, 1994), we analyse the importance of the different phases of the decision-making process, the networks of collaboration and conflict among the elite and the domestic power configuration in each of the three case studies. The data come from face-to-face interviews carried out with the main political actors in the three decision-making processes (over 150 interviews altogether, representing 65 organizations) between Summer 2001 and Autumn 2002. The quantitative and the qualitative analyses complement one another. The former helps us to identify the effects of internationalization while controlling for the impact of other variables, and to bring out some trends that may be 224 Internationalization and Domestic Politics generalized. The latter enables us to deepen and refine the findings of the quantitative analysis, and to test a number of influences that cannot be assessed on the basis of the quantitative data bank.

3 The domestic consequences of internationalization

3.1 The impact of internationalization on the institutions of the decision-making process The Swiss decision-making process is characterized by extensive pre- parliamentary consultation procedures. In order to overcome the ‘referendum threat’ – that is the threat that a federal law adopted by the Swiss Parliament is attacked by a referendum7 and/or the threat that a legislative act is rejected in a popular vote – elites are prone to compromise during the initial, pre-parliamentary, phase of the legislative process (Neidhart, 1970). A number of formal consultation mechanisms, such as commissions of ‘experts’8 and consultation procedures,9 have been developed to favour the integration of all political parties and interest groups able to threaten credibly the decision-making process. In addition, it is also possible to resort to informal consultation procedures that can take various forms (from an organized but selective ‘pre-consultation’ of interested actors to a simple phone call). The pre-parliamentary phase has often been described as the most important phase of the decision-making process (e.g. Kriesi, 1980: 589). According to our first hypothesis direct internationalization reduces the importance of formal pre-parliamentary consultation procedures, and increases the importance of informal pre-parliamentary consultation. The hypothesis of a reduced importance of formal consultation in the case of direct internationalization is based on a number of earlier works in the field of foreign policy. These works show that international negotiations tend to shift away the centre of gravity of the decision-making process from the domestic level towards the international arena (e.g. Moravcsik, 1994, 1998a, 1998b; Sciarini, 1994). Furthermore, the usual slowness of formal pre-parliamentary procedures does not fit the rapid and discontinuous rhythm of international negotiations (Germann, 1996). As a result, decision-making processes in foreign policy have overall a more “closed” character than in domestic policy (e.g. Sidjanski, 1966). The former tend to include fewer non-state actors and, therefore, to be more exclusive than the latter. This being said, the general rule that Swiss state actors have a strong incentive to include domestic actors in the decision- making process in order to prevent them from attacking a legislative act by referendum also holds for an international agreement: In the absence of such inclusion influential groups whose preferences are neglected are likely to campaign against the agreement in the ratification phase. As an alternative to formal consultation procedures state actors are hence likely to rely on extensive informal consultation in the case of direct internationalization.10 Pascal Sciarini and Sarah Nicolet 225

Our expectations are more ambivalent with regard to cases of indirect internationalization. On the one hand, the importance of formal consultations is also likely to be reduced in those instances, since similar constraints are at work in both direct and indirect internationalization – in particular, the constraints on the content of the legislation. On the other hand, owing to the absence of an international negotiation, the decision-making channel applying to indirect internationalization is similar to that observed with regard to domestic politics acts. Both the decision to adapt to the international pressure, and the extent of this adaptation, may be submitted to standard formal consultation procedures. Ambivalence also holds for informal consultations. If our hypothesis regarding the greater leeway for formal consultation in the case of indirect internationalization is correct, then there is presumably less need to resort to informal consultation in this situation than in the case of direct internationalization. But state actors may choose to use both formal and informal channels in order to increase the likelihood that a project arising ‘from outside’ is accepted by domestic non-state actors and successfully passes through the subsequent phases of the decision-making process. In sum, formal pre-parliamentary consultation procedures are expected to be less important in the case of direct internationalization than in the ‘standard’ domestic decision-making process. Conversely, informal consultation procedures are expected to be more important in the former case than in the latter. On both accounts indirect internationalization is presumably an ‘in-between’ case. To test these hypotheses we use two distinct empirical measures of the degree of formal consultation. First, the quantitative data bank of the objects of the 1995–99 legislature allows us to explore the effect of internationalization on the frequency of formal consultation procedures. Our analysis of the degree of formal pre-parliamentary consultation11 shows that highly internationalized acts are less likely overall to include an extensive pre-parliamentary phase than are more ‘domestic’ legislative acts (Sciarini etal., 2002: 17). This holds, in particular, for cases of direct internationalization (legislative treaties and the related legislative adaptations).12 However, formal consultation is higher in cases of autonomous adaptation for which one consults even more than for purely domestic acts (Sciarini et al., 2002: 17). Secondly, the comparative case study provides insights into the importance of formal pre-parliamentary consultation procedures. In each of the three cases we asked our interview partners to evaluate the importance of the different phases of the decision- making process. We then aggregated their answers for the three main phases of the process (the pre-parliamentary phase, the parliamentary phase, and a residual category comprising mainly the referendum phase, Table 9.1; see also Sciarini etal., 2004). Given our specific interest for the pre-parliamentary phase, we further divided it into two sub-phases: the consultation phase, which consists of procedures open to private actors such as, for instance, 226 Internationalization and Domestic Politics

Table 9.1 Importance of the phases of the three decision-making processes (percentage of respondents who see the corresponding phase as one of the three most important)

Free movement Telecommunication Eleventh pension of persons reform reform

Pre-parliamentary phase Administrative procedures 64 40 49 (including international negotiations) Formal consultation 419 16 procedures Parliamentary phase 13 38 35 Other phases (collection of 20 3 0 signatures, popular vote, etc.) Total 100 100 100 (number of respondents) (30) (37) (27) commissions of experts, and the administrative phase, which comprises the procedures in which only the federal administration can formally take part. The importance granted to formal pre-parliamentary consultation procedures confirms the findings of the quantitative analysis on the frequency of consultation, and strengthens our hypothesis. We see from Table 9.1 that the perceived importance of consultation procedures is lowest in the case of the free movement of persons, our case of direct internationalization (only 4 per cent of our interview partners see these procedures as one of the three most important phases), and highest in the telecommunication reform, our case of autonomous adaptation (19 per cent). In addition, Table 9.1 suggests that direct internationalization also reduces the importance of the parliamentary phase, whereas the administrative phase, including the international negotiation, is seen as highly important. By contrast, the importance of the parliamentary phase is as high in our case of indirect internationalization as in our domestic, control, case (overall importance of respectively 38 per cent and 35 per cent). Our case studies also allow us to assess the frequency of informal consultation – an aspect that we could hardly capture in our data bank on the decision-making processes. In the case studies, we use the percentage of non-state actors that had access to the administrative phase as an indicator for the existence of informal consultation. This phase, which according to Table 9.1 is of the utmost importance,13 is formally restricted to state actors but it can in fact be open – to various extents – to informal consultation of non-state actors. As expected, direct internationalization leads to more informal consultation than in a domestic decision-making process (Sciarini et al., 2004): In the case of the free movement of persons, 57 per cent of the non-state actors were informally involved in the administrative phase; this Pascal Sciarini and Sarah Nicolet 227 share is even higher in the case of indirect internationalization, where 69 per cent of non-state actors had access to this phase; the corresponding figure is only 33 per cent in our control case. These results point towards an increase in the resort to informal consultation as a consequence of both direct and indirect internationalization. However, it seems that the function attributed to these mechanisms differ between the two forms of internationalization. In the case of direct internationalization, informal consultations can be seen as a sort of ‘functional equivalent’ for the usual (formal) consultation procedures. In the case of indirect internationalization, formal and informal consultation mechanisms complement each other. In sum, our results confirm our first hypothesis of a reduced frequency and importance of formal consultation and the increased importance of informal consultation as a consequence of direct internationalization. In the case of indirect internationalization evidence is more clear-cut than expected and reveals that instances of autonomous adaptation are characterized by a high degree of both formal and informal consultations. In that sense, our results also confirm the relevance of the distinction between direct and indirect internationalization: Whether adaptation to international pressures occurs through an international agreement or ‘autonomously’ seems to matter much with regard to the role played by consultation procedures in the process of adaptation. In cases of autonomous adaptation our findings are in agreement with the argument that there is a need for federal state actors to use both the formal and informal consultation channels at their disposal in order to increase the acceptability of a legislative act; this need is all the greater since the act has not been initiated by an internal demand. Regarding direct internationalization, the reliance on informal and selective consultation presumably enhances the acceptability – and hence the efficiency – of state actors’ decisions, but raises interesting issues with respect to transparency and democratic legitimacy.

3.2 The impact of internationalization on the level of conflict in the decision-making process Scholars hold contradictory views with respect to the impact of internationalization on the conflict level within national contexts. On the one hand, a number of scholars, both in Switzerland and abroad, claim that internationalization eases the adoption of new domestic rules (e.g. Armingeon, 1998; Grote and Schmitter, 1999; Sciarini, 1994). This is because it enables domestic decision makers to implement reforms while shifting the blame for the related redistributive effects onto the international sphere (Armingeon, 1998: 108). In that sense, decisions that would be highly controversial in a domestic setting are more likely to be accepted under international pressure. This perspective corroborates Katzenstein’s (1984 and 1985) seminal work on small European states, which shows that economic openness and vulnerability favour the emergence of domestic consensus.14 Based on this rationale, the 228 Internationalization and Domestic Politics external pressures for change associated with internationalization are expected to result in a lower level of domestic conflict. Given that these pressures are especially constraining for directly internationalized acts, the level of conflict is likely to be lower in cases of direct internationalization than in cases of indirect internationalization. On the other hand, some authors argue that internationalization leads to a loss of control of the domestic elite over the political agenda, since it has to deal with acts that are in a way imposed from ‘the outside world’ (e.g. Germann, 1994; Hug and Sciarini, 1995). This should translate into a higher level of conflict in decision-making processes that are ‘internationalized’ than in those that are not. This holds especially for directly internationalized acts. Agreements negotiated at the international level have a strong ‘take-it-or-leave-it’ character that greatly reduces the room for manoeuvre of domestic actors. The necessity of reforming domestic policies without much amendment possibilities is likely to trigger conflict among the elite. In addition, international agreements have to be ratified within a fixed time period, which prevents the domestic elite from postponing its decision – a strategy often employed in domestic politics. The conflict-triggering effect is again expected to be more moderate in the case of indirect internationalization, where the domestic elite keeps a wider room of manoeuvre with respect to adaptation (extent and timing). We analyse the consequences of internationalization on the level of conflict in the decision-making process under three different angles. Our first two measures are based on our quantitative data bank. First, we examine how internationalization affects the level of conflict in the parliamentary arena. To that end, we use various indicators of conflict, based on the intermediary and final votes in the two Chambers of the Swiss Parliament as well as a measure of conflict between the two Chambers.15 Secondly, we test the impact of internationalization on the level of conflict in the referendum arena. As an indicator we use the measure of whether or not a referendum is successfully launched. Finally, based on our three case studies, we take a closer look at the nature (conflictual or consensual) of the relationship among the elite involved in the different decision-making processes. The quantitative analysis points towards a diminution of the level of parliamentary conflict as a consequence of internationalization (Sciarini et al., 2002: 19–21). Even though in general the coefficients are just below the significance level, the direction of the effect (negative) reveals a tendency towards conflict reduction within both Chambers (level of conflict of the intermediary and final votes) and between the Chambers. The results are particularly clear-cut for the lower Chamber of Parliament, the most representative of the country as a whole. Unlike expected, however, the consensus-triggering effect is not substantially stronger in the case of direct internationalization (international agreements) than in the case of indirect internationalization (autonomous adaptation). Pascal Sciarini and Sarah Nicolet 229

Evidence regarding the referendum stage tends to confirm the hypothesis of a reduction of conflict as a consequence of internationalization (Nicolet et al., 2003: 21, 32; Sciarini et al., 2002: 23–5; see Figure 9.1).16 Figure 9.1 shows that under the control of the importance of an act, the predicted probability of successfully launching a referendum varies according to the form and level of internationalization. While internationalization exerts no impact upon the launching of a referendum for acts of little or medium importance, its impact increases substantially past a medium level of importance. This result holds for direct internationalization (international treaties and related legislative adaptations) and even more so for indirect internationalization (autonomous adaptation and other internal legislative acts that are highly internationalized). Thus, for the most important acts, the difference in the probability of a referendum being launched as a consequence of internationalization amounts to 20 per cent between international treaties and legislative acts that do not arise from an international treaty (other things being held constant) and up to 80 per cent between cases of autonomous adaptation and cases of non-autonomous adaptation.17 The case studies confirm these findings. In order to analyze the conflict network among the organizations active in a decision-making process, our questionnaire included an item regarding both convergence and divergence of views with the other actors of the network.18 The average level of conflict

1.00

0.80

0.60

0.40

Predicted probability 0.20

0.00 1.00 2.00 3.00 4.00 5.00 Importance

dom. act wt minimal int. dom. act wt maximal int. no autonomous adapt. autonomous adapt. no treaty treaty

Figure 9.1 Predicted probabilities of a successful launching of a facultative referendum, as a function of the importance and internationalization of a legislative act (legislature 1995–99) 230 Internationalization and Domestic Politics is lower in the two internationalized cases than in our control case (Sciarini et al., 2004: 365):19 while conflict slightly dominates (−0.04) in our domestic case (the eleventh pension reform), the level of convergence is higher in our case of indirect internationalization (the telecommunication reform, 0.18) and is substantially higher in our case of direct internationalization (the free movement of persons, 0.43). Overall, then, our results suggest that pressures for change arising from the international level do not translate into domestic conflict. Rather, our findings are congruent with the view that internationalization tends to reduce conflict – or foster consensus – domestically. In that sense, the spectacular rejection of a number of internationalized objects such as, for instance, the participation in the European Economic Area (EEA) in 1992 or the creation of a Swiss UN ‘Blue Helmets’ corps in 1994 should not make us forget that next to these very conflictual issues, there are a great number of (directly and indirectly) internationalized acts that are readily adopted at the domestic level. Further research should investigate more carefully the likely causes of this interesting result: is the low level of conflict explained by a weak degree of inadequacy between the Swiss legislative order and international norms, by a willingness of the Swiss elite to stick together when facing international challenges, or by some sort of side-payments like those used in the case of the free movement of persons (Fischer, 2002, 2003; see also below)?

3.3 The impact of internationalization on the domestic power configuration 3.3.1 Empowerment of state actors The internationalization process modifies the opportunity structure in which actors evolve and leads to a subsequent redistribution of power among domestic actors. Some actors are expected to benefit more than others from this process, which should thus translate into a ‘differential empowerment’. The hypothesis of a differential empowerment has been made specifically with regard to the European Union (Börzel and Risse, 2000; Risse et al., 2001). In this respect, Moravcsik (1998a) argues that the phenomenon of internationalization/ Europeanization reinforces state actors: ‘International negotiations and institutions change the domestic context in which policy is made by redistri- buting four domestic political resources: initiative, institutions, information and ideas (the “four I’s”)’ (Moravcsik, 1998a: 2). In other words, international negotiations shift the control over domestic agendas (initiative), modify who can participate in domestic decision-making procedures (institutions), increases asymmetries in information and ‘alter[ing] legitimate domestic ideological justifications for policies (ideas)’ (Moravcsik, 1998a: 11). The changes in the decision-making channels brought about by the internationalization process are expected to favour those actors who are directly involved in international negotiations and institutions – that is, state executives. Pascal Sciarini and Sarah Nicolet 231

A similar argument has been made by Putnam (1988) in his work on the two-level game. Negotiation of a domestic issue at the international level reinforces state executives, which benefit from a reinforced autonomy with regard to domestic actors, since they are the only actors involved at both the international and the domestic levels. The observation of executive reinforcement as a result of direct internationalization has also been made specifically with regard to the Swiss case. Research undertaken on internationalized decision-making processes show that their more closed character leads to a reinforcement of the executive most closely involved in the international talks at the expense of both the Parliament and non-state actors (interest groups and political parties; see, e.g., Mach, 2003; Sciarini, 1992, 1994).20 The case of indirect internationalization again differs from the direct one only with regard to the intensity of the likely effects. Although they are not directly involved in international negotiations, state executives are still expected to be reinforced in an instance of autonomous adaptation. First, state actors are expected to benefit from asymmetrical information given that they have at their disposal official channels of access to the international sphere. Secondly, as already mentioned, the fact that the project emanates ‘from outside’ is likely to strengthen the executive by offering it the possibility of using international pressures as a leverage for implementing domestic changes, a strategy that is not available in a ‘purely’ domestic decision-making process. The three case studies outlined allow us to explore the impact of internationalization on the relative empowerment of domestic actors.21 To measure the importance of an actor in the decision-making process, we measure both its reputational power – that is the power that an actor is said to have according to the other actors in the network22 – and its centrality in the policy network.23 Our case of direct internationalization (the free movement of persons) clearly shows a reinforcement of state actors in comparison to private actors (Table 9.2). According to both reputational power and degree of centrality, state actors outweigh private actors in the case of the free movement of persons. As anticipated, the results are more ambiguous in the case of indirect internationalization. While state actors were perceived as having more power than private actors in the telecommunication reform, their degree of centrality in the network is lower than that of private actors. Their score on centrality is even lower than the one they obtain in the domestic control case (the eleventh pension reform). This can be explained by the fact that the Federal Office of Communications (OFCOM) acted as a transmitter of policy developments in the European context, which presumably accounts for its perceived importance (reputational power). However, in the absence of international negotiations, the centrality of state actors was reduced (Sciarini et al., 2004: 366). In sum, our results offer strong supporting evidence to the hypothesis of a reinforcement of state executives relative to private actors and the legislative 232 26) = Betweenness centrality (N 27) = Reputational power (N 31) = Betweenness centrality (N 27) = 18 10 Reputational power (N 28) = Betweenness centrality (N 26) = 112 6 16 0131 63 62 47 23 25 39 37 38 35 67 75 61 100 100 100 100 100 100 power (N

Importance of state and non-state actors (scores of reputational power and betweenness centrality, in %) centrality, power and betweenness of reputational (scores actors of non-state state and Importance

private firms firms private and federal railways) railways) and federal Table 9.2 Table Reputational Free movement of persons Telecommunication reform Eleventh pension reform Federal Council Council Federal Public administration Total state actors 51 12Others (experts, cantons) Total non-state actors 62 1 35 1 11 23 0 17 2 8 39 0 0 Political parties business Export-oriented Trade unions groups, interest Other 10 17 9 4 15 25 3 4 42 5 8 19 13 26 24 9 17 18 Federal companies (PTT (PTT companies Federal Total Pascal Sciarini and Sarah Nicolet 233 body as a consequence of direct internationalization. However, the findings regarding indirect internationalization are less conclusive. State executives do not emerge as clearly reinforced by the process of autonomous adaptation.

3.3.2 Empowerment of export-oriented business over trade unions Our hypothesis regarding the strengthening of state executives does not imply that private actors are all equally weakened by internationalization. We assume a ‘differential empowerment’, and this for two main reasons. First, the existence of formal institutions at the international level creates a new political opportunity structure that is susceptible to offer additional resources to domestic actors.24 Yet not all domestic actors have the same access possibilities to these institutions. This asymmetry is likely to favour the export-oriented business sectors to the detriment of trade unions. In fact, the former have traditionally had more competencies in the international arena and have enjoyed more channels of access to international or supranational institutions. This is explained by the fact that employers’ organizations – including the Swiss ones (Mach, 1999: 429) – have been able to organize themselves at the international level much earlier than unions.25 Following on from this, we expect this effect to apply primarily in the context of direct internationalization, where the possibility of access to formal institutions is of key importance. Secondly, when international changes are congruent with their own goals, it is likely that domestic actors will use international pressures as a way to support their claims and strengthen their position domestically.26 Thus, one can argue that the process of internationalization (both direct and indirect) is expected to benefit the actors whose preferences are the most in phase with the changes taking place at the international level. In the context of Europeanization, scholars commonly agree that, so far, in the EU so-called ‘negative integration’, that is liberalization and deregulation, has overall taken the lead over ‘positive integration’, that is the adoption of common European policies shaping the conditions under which markets operate (Scharpf, 1997: 210–11; Streeck and Schmitter, 1991; Traxler and Schmitter, 1995). This situation should therefore be more favourable to export-oriented business sectors than to trade unions. This hypothesis applies to both our cases of direct and indirect internationalization, which are clear cases of ‘negative’ integration (liberalization). Therefore, we anticipate a reinforcement of these groups compared to trade unions. A comparison of the reputational power and centrality scores for our three case studies provides mixed evidence in this respect (Table 9.2). In the case of indirect internationalization (the telecommunication reform), export-oriented organizations clearly outbalance trade unions. Contrary to our hypothesis, the situation is not as clear in our case of direct internationalization, where both measures of power point towards a balance between business organizations and trade unions. Thus, our hypothesis that direct internationalization 234 Internationalization and Domestic Politics reinforces the actors who have greater access to international institutions and/or whose preferences are the most in accord with changes taking place in the international arena is not confirmed by evidence from our case studies. The explanation for this unexpected result lies in the successful strategic use of the referendum threat by the unions: knowing that the export-oriented sector of the Swiss economy desperately wanted the bilateral agreements, the unions threatened to undermine the whole package by a referendum unless a safety net against social dumping was adopted in the free movement of persons. This turned the domestic power configuration on its head and resulted in legislative innovations – such as the possibility of introducing minimal wages in case of salary dumping – which the unions would have been unable to obtain in a purely domestic setting (Fischer etal., 2002: 160–1). The evidence presented here is more illustrative than generalizable given that it is based on a limited number of case studies. However, results from the ‘Free movement of persons’ study run counter to the hypothesis usually made in the literature – including the literature on the impact of economic globalization on competition among states (e.g. Scharpf, 1998) – of a reinforcement of the more mobile factors (capital) compared to the less mobile ones (labour). Further research is needed on this topic in order to establish whether the features found here are specific to the issue at stake or to the country under study, or whether they have a broader validity and, if this is the case, under which conditions.

4Conclusion

Based on a mixture of quantitative statistic methods and a comparative case study, we have analysed the effects of internationalization on the institutions of the decision-making process (‘polity’), on the level of conflict in the process and on the domestic power configuration between the actors (‘politics’). Switzerland served as a test case for this study. Our chapter first underscores the importance of paying closer attention to the degree and nature of internationalization. As we have revealed, while direct and indirect internationalization have some similar effects, there are also some clear differences in their impact. On the one hand, direct and indirect internationalization likewise affect the resort to informal consultations, contributing to their greater use. They also both result in a lower level of conflict throughout the decision-making process. On the other hand, the resort to formal consultation procedures is different between the two forms: directly internationalized acts are characterized by a reduced role of formal consultation mechanisms, in terms of both their frequency and their importance; the opposite is true for indirectly internationalized acts, for which pre-parliamentary consultation is even more important than for ‘purely’ domestic legislative acts. The distinction between direct and indirect internationalization also appears useful when studying the effects on the Pascal Sciarini and Sarah Nicolet 235 domestic power configuration. Direct internationalization results in a clear reinforcement of state executives compared to private actors and the legislative body, whereas our findings are less clear-cut in the case of indirect internationalization. A divergent pattern is also found with regard to the differential empowerment of private actors. While the export-oriented business sector is visibly reinforced compared to trade unions in the case of indirect internationalization, a balance of power holds in the case of direct internationalization. Thus, whether a state adapts as a consequence of an agreement negotiated at the international level or as a result of a more autonomous decision is reflected in differences in the adaptation process. Secondly, as our results show, the internationalization phenomenon exerts a substantial effect on the ‘polity’ and ‘politics’ dimensions of the decision- making process – an aspect that has been neglected in the literature to date to the benefit of the ‘policy’ impact of internationalization. The influences identified here are all the more spectacular since they concern a country, Switzerland, which is not strongly integrated at the international level. As our analysis has clearly shown, even a neutral state that only recently became a member of the United Nations and remains outside the European Union cannot escape the effects of the internationalization phenomenon. All of the dimensions of its decision-making process are affected in one way or another by developments at the international level – and especially by those at the European level.27 Thirdly, from a methodological point of view, our contribution highlights the benefits that can be gained from combining a quantitative approach with in-depth case studies. On the one hand, the statistical analysis of a high number of legislative acts allows us to establish some general trends. It also strengthens our confidence in the validity of the results from our comparative case study: the strong convergence existing between our quantitative and qualitative results suggest that our two internationalized cases are representative of the effects of internationalization in general. On the other hand, the detailed comparison of three decision-making processes permits us to deepen our findings. For instance, while the frequency of the resort to formal consultation mechanisms can be studied using the data set, their importance can be assessed on the basis of the case studies. In addition, the comparative case study also allows us to explore some new aspects of the influence of internationalization, such as the importance of informal consultation procedures or the domestic power configuration. While the related results are based on a limited number of observations and can therefore hardly be generalized, they are in line with previous studies in the field, in Switzerland and abroad. Finally, our quantitative analysis draws attention to the fact that next to cases of autonomous adaptation there also exist a number of other legislative acts that are indirectly internationalized. In contrast to cases of autonomous adaptation, which result from adaptation to legal changes at the international level, these acts are influenced by more diffuse forms of internationalization, 236 Internationalization and Domestic Politics such as increased economic integration, global technological transformations or ecological changes. Our results show that the characteristics of the decision-making processes (e.g. the resort to formal consultation mechanisms and the level of conflict) for these internal indirectly internationalized acts differ from those of ‘purely’ domestic acts, international agreements and autonomous adaptations. Given the increasingly pervasive consequences of internationalization, this category is likely to gain in importance in the future and thus merits further analysis.

Notes

1. This mechanism is central in the literature on the impact of globalization on competition among states, which argues that states tend to adapt their legislation to reduce undesirable outflows of internationally mobile factors – or to attract new inflows of the same kind (see, e.g., Scharpf, 1998). 2. The lack of a control case, that is of a case where the impact of the outside world is minimal, is a methodological bias of most of the studies on Europeanization (Radaelli, 2001). 3. On the definition and role of Swiss neutrality and on Switzerland’s unique position in the European and international contexts, see, e.g., Dupont and Sciarini, 2001; Kriesi, 1998; Kriesi et al., 1999, Sciarini et al., 2000. 4. See Sciarini et al., 2002 and Nicolet et al., 2003 for a detailed description of the data bank. The latter is now available at SIDOS (www.sidos.ch). 5. The level of internationalization of domestic legislative acts is measured on the basis of an assessment of these acts by Swiss politics experts on a scale ranging from 0 to 4, where 0 stands for acts for which Switzerland has a very wide room for internal manoeuvre and 4 for acts strongly influenced by the international environment (for an extensive definition, see Nicolet et al., 2003). The domestic acts with a level of internationalization of 0 serve as the base category in the statistical analyses. 6. Our two cases of internationalization are both instances of Europeanization. This illustrates the importance of Europeanization as a special occurrence of the more general internationalization phenomenon. More information on these three decision-making processes can be found in Sciarini et al., 2004. Note that our selection of case studies includes only one category of indirectly internationalized acts (the telecommunication reform, as a case of ‘autonomous adaptation’), but no case of the residual category of domestic acts that are highly internationalized. 7. In Switzerland, all federal laws adopted by the Swiss Parliament are submitted to the optional referendum: By collecting 50,000 signatures of citizens within 100 days against a law, a group may call for a final decision by the people. 8. A commission of experts, ‘as its name does not indicate, comprises not only specialists in the field but also representatives of interest groups, of the cantons and of the parties’ (Morand, 1987: 76, our translation). 9. A consultation procedure is launched either by the Federal Council (the Swiss government) or by a department of the administration. It usually consists of a written consultation in which all the main political parties, interested organizations, and the cantons, are invited to express their opinions with respect to a legislative project. Pascal Sciarini and Sarah Nicolet 237

The analysis of the results of the consultation procedure, undertaken by the appropriate department, paves the way for the final draft of the legislative act, which is then adopted by the Federal Council and transmitted to the Swiss Parliament. 10. This hypothesis is in line with Kriesi’s (1980: 611–12) finding that the consultative procedures implemented during the negotiation between Switzerland and the European Community (EC) on the free trade agreement in 1971 were highly informal and included a limited number of participants. 11. The degree of pre-parliamentary consultation is measured by adding the different forms of formal consultation procedures that have taken place in the pre- parliamentary phase. It ranges from 0 (no consultation) to 6 (six distinct forms of consultation; see Sciarini et al., 2002: 16). 12. This is also true for the other category of domestic acts that are highly internationalized. 13. In each case the administrative phase is by far the most important one. Note that the eleventh pension reform was repeatedly postponed during the parliamentary phase, and the reform proposal had not yet been discussed in the Council of State at the time of our inquiry (Spring 2002). As a result, many experts chose to base their evaluation on the phases that had already taken place. This presumably accounts for the non importance granted to the referendum phase in this specific case. 14. More recently, a similar argument has been developed in the literature on multilevel governance: To the extent that the incorporation in a system of multilevel governance increases the autonomy of state actors, this increase in autonomy equally bears the potential to ease political conflicts (Grande, 1996). 15. Given that Switzerland has a perfect bicameral system, it is worth including the second Chamber in the analysis as well as a measure of conflict between the two Chambers. 16. The figure presents the predicted probabilities of a successful launching of a facultative referendum calculated on the basis of the coefficients from a multivariate logistic regression. The effects of the following features of a legislative act were measured: Novelty of the decision, degree of importance, level of potential conflict, type of internationalization (international treaty, autonomous adaptation, or internationalized domestic legislative act), existence of a commission of experts, existence of a consultation procedure. The estimated coefficients for the model are the following: Y =−11.04 − 0.75 * New decision + 2.52 * Importance + 0.50 * Level of potential conflict − 1.00 * International treaty − 7.67 * Autonomous adaptation − 0.80 * Internationalized domestic act + 0.07 * Commission of experts − 0.39 * Consultation procedure (Sciarini et al., 2002: 24, table 7). 17. In our data set none of the cases of autonomous adaptation was attacked by referendum. 18. The question was the following: ‘With regard to the elaboration of this act, could you tell me with which organizations you had great divergence of views, respectively great convergence of views?’ The answers were coded in the following way: −1 for divergence, +1 for convergence and 0 for neither nor. 19. The average level of conflict is computed by adding the scores (−1, 0 or 1) of the convergence/divergence matrix. The resulting figure is then divided by the overall number of possible links. 20. The weakening of the Parliament is further reinforced by the strong ‘take-it-or- leave-it’ character of an agreement concluded at the international level that leaves hardly any room for amendments in the parliamentary phase. 238 Internationalization and Domestic Politics

21. The quantitative data bank provides no information in that respect. 22. The question that was asked to our interview partners was labelled as follows: ‘In your opinion, which are the three most important actors in the decision-making process on . . .?’ For comparative purposes, Table 9.2 presents a standardized measure of reputational power, namely the percentage of interview partners who mention a given actor as one of the three most important. 23. The collaboration network is based on the responses to a classic socio-metric question: ‘Here is a list of actors active in this decision-making process. Based on this list, could you please tell me with whom you collaborated closely?’ Betweenness centrality measures the percentage of actors that must transit through a specific actor in order to reach (i.e. to collaborate with) the other actors of the network. It thus indicates to what extent an actor plays the role of a ‘broker’ or ‘gatekeeper’ for the connection with other actors in the network. 24. See Börzel and Risse (2000) for a presentation of this argument with regard to the EU context. 25. For instance, the Swiss Business Federation (economiesuisse) is the only Swiss peak association that is directly represented in Brussels. 26. See Risse et al., 2001 for a development of this argument with respect to Europeanization. 27. Recent developments, however, suggest that the impact of Europeanization on Swiss legislation and decision making has weakened since the late 1990s (see, e.g., the popular rejection of the federal law on the electricity market or the failure of the unbundling of the last mile in the telecommunication sector; Fischer et al., 2003). 10 Reforming the Swiss Municipalities: Efficiency or Democracy? Andreas Ladner and Reto Steiner

Swiss local level reforms fit quite well into the international pattern. As in many other countries, municipalities introduced New Public Management, increased intermunicipal cooperation, amalgamated with other municipalities and tried to improve the functioning of the authorities and the decision-making process. Whereas intermunicipal cooperation and the merging of municipalities aimed at making service provision at local level more efficient and at increasing the quality of the services provided, New Public Management reforms tried to improve the municipalities’ capacity to act and aimed at moving closer to citizens by increasing customer orientation. Altogether, the reforms cannot be regarded as an attempt trying to downsize the state sector. However, strengthening the input-side of democracy by moving decisional power to the citizens was of lesser importance in Swiss municipalities than in other countries. This is mainly due to the far-reaching direct democratic rights of Swiss citizens.

1 Introduction and research questions

Local government reforms are currently on the agenda throughout the world (Caulfield and Larsen, 2002: 9). In Western Europe reforms began in the 1960s, but they have accelerated and spread more widely in recent years (Kersting and Vetter, 2003: 11). In the 1970s, when a first wave of reforms swept over Northern Europe, leading to a dramatic reduction in the number of municipalities, Swiss municipalities were left untouched. In the 1990s this situation changed considerably, as an increasing number of Swiss municipalities began to modify their institutional settings. New Public Management (NPM), reforms of local parliaments and executives, intermunicipal cooperation and even an amalgamation with other municipalities became major concerns in local politics. The reform activities affect different aspects of local government. Four groups of reforms can be distinguished (see, for example, Kersting and Vetter,

239 240 Reforming the Swiss Municipalities

2003): territorial and functional reforms, management reforms, changes towards more direct participation, and changes in local representative democracy: the former two measures are commonly classified as reforms of the local administration and the latter two as reforms of the political system. At first sight, political system reforms affect the input-oriented legitimacy of democracy (see Scharpf, 1999) by making local government and the decision-making process more democratic whereas reforms of the local administration and functional or territorial reforms are directed towards output legitimacy by attempting to make service delivery more efficient. However, this division is short-sighted. Territorial or administrative reforms are also linked to the question of democratic decision making, whereas reforms of the political system cannot ignore the question of efficiency when it comes to taking decisions. Or, in other words, NPM tries to make service delivery more efficient by enhancing transparency in the costs of local services, creating market-like competition, and increasing the operational autonomy of local bureaucrats. NPM furthermore aims to increase the responsiveness and accountability of the political authorities through customer surveys and more citizen participation and tries to improve the political steering facilities. Similarly, the amalgamation of municipalities not only offers more efficient service delivery through ‘economies of scale’, but also reduces spill-over effects, takes decisions on specific issues into a single territorial constituency, and makes them politically more contestable. Swiss municipalities differ from municipalities in many other countries in at least three key respects. First, although most of them are very small,1 they enjoy a privileged position within the political system, commanding a considerable amount of discretion and autonomy, especially where their financial resources are concerned.2 Since their activities do not fall directly into the realm of the nation-state but depend upon the cantons, there are considerable differences to be found within the country (see Ladner, 1991a).3 Secondly, Swiss citizens dispose of a broad set of direct democratic means of participation, which is particularly well developed at the local level (see Ladner, 2002). And thirdly, since Swiss municipalities were left untouched by the reforms of the 1970s, they are more or less in a position to start from scratch, having the advantage of taking into account the experiences of other countries. A first aim of this chapter is to show to what extent the reform activities of the Swiss municipalities fit into the international reform pattern. The main research question we consider, however, is whether they are increasing output-oriented legitimacy by making service delivery more efficient, given the far-reaching means of direct democratic participation offered to Swiss citizens. And if this is the case, by what means is this achieved? Such a finding could make a good argument for the path dependency of institutional change, making patterns of reforms proceed very differently in the individual countries depending upon the pre-existing conditions (see, for example, Pierson, 2000). Andreas Ladner and Reto Steiner 241

Our 1998 nationwide survey covering all 2,842 Swiss municipalities gives us a first indication of the spread of the different reform activities.4 Looking at the structural differences between reform municipalities and municipalities without any reform activities we also gain a first impression of the possible causes for these reforms. For a better understanding of the aims of the reform activities and whether they are really attempting to make service delivery more efficient, it is, however, crucial to look at the reform projects more closely. On the grounds of nearly 200 reform projects which were selected on the basis of the results of the first part of our research, it will then be possible to analyse the triggers of these reform activities more closely and to have a look at their intentions and goals.

2 Change and reforms at the local level

In recent years, Swiss municipalities have come under great pressure in various respects. Changes in Swiss society have increased not only the number of tasks, but also their complexity. Particularly affected are the social and cultural spheres and all environmental matters (Geser, Höpflinger, Ladner and Meuli, 1996). The growing political integration and ‘Politikverflechtung’ (Scharpf et al., 1977) reduced the autonomy of the municipalities and they increasingly became ‘agencies of implementation’ for the cantons and the federal government. Along with the overall economic recession at the beginning of the 1990s, the financial situation of the municipalities also worsened, in particular, the debt–equity ratio increased (Eidgenössisches Finanzdepartement, 2000). The response of the Swiss municipalities to these challenges was an increase in reform activities which mirrored almost perfectly the pattern of international reform activities and the categories mentioned by Kersting and Vetter (2003).5 The most widespread type of reform is the growth in intermunicipal cooperation. Two-thirds of municipalities have increased their level of cooperation (Table 10.1). For all other municipalities the degree of cooperation had remained unchanged.6 In 1998 at least half of all municipalities worked together with at least one other municipality in areas such as schools, medical care, care for the elderly, refuse disposal, water supply, sewage treatment, and the civil service. In areas such as support for the unemployed, the civil service, the fire brigade and medical care, cooperation has been particularly intensified within the last few years (Ladner et al., 2000). In a number of areas, however, more intensified cooperation still remains possible. Such areas include general administration (computer networks and facilities, accounting, and registration offices), provisions for asylum seekers, planning, construction permits, public buildings, environmental issues, private transport, the integration of immigrants, and local executives. Up to the time of writing, less than 20 per cent of municipalities have worked together in these areas. 242 Reforming the Swiss Municipalities

Table 10.1 Spread of reforms in Swiss municipalities

Reform activity Percentage of municipalities

Territorial and functional reforms Intensification of intermunicipal cooperation 63.0 Discussing amalgamations 18.2 Concrete plans for a merger of municipalities 7.9 Management reforms* Considering New Public Management reforms 34.7 First steps with New Public Management 25.3 Product definitions 4.2 Changes in local representative democracy** Executive reforms 34.6 Parliamentary reforms*** 3.9 Changes towards more direct participation Meetings to inform citizens or to integrate them into 63.0 the planning process (1994)**** Increasing the scope of initiatives and referenda (1994)**** 15.3 = = nmin 1,843, nmax 2,289, depending on the survey and the question

* NPM reforms affect the entire political-administrative system. A categorization is therefore only a limited reflection of reality. ** Similarly difficult is categorization in the case of executive and legislative reforms. Naturally, it is predominantly the political processes that are often affected (e.g., reorganization of competences); if, however, the size of the parliament or of the government is altered, then this is primarily a structural reform. *** Only about 18 per cent of the Swiss municipalities have a parliament. Among those municipalities having a parliament about 20 per cent had some reform activities to report. **** The results for this reform category stem from a similar survey conducted in 1994 (Geser et al., 1996).

For intermunicipal cooperation in general, it seems obvious that in larger territorial units a number of services have profited from ‘economies of scale’. But cooperation also raises questions of democratic decision making and control. How can decisions be taken within a federation of municipalities of differing sizes, if the principle of ‘one man one vote’ places smaller municipalities at a disadvantage? How can delegates on the boards of a federation of municipalities be controlled democratically and in the case of cooperation on the basis of private law, how are contracts to be formulated and property rights to be regulated? Amalgamation of municipalities was not a popular issue in Switzerland until recently. Territorial reforms carried out in most northern European countries in the 1970s had no influence in Switzerland, and there were never any serious attempts to reduce the number of municipalities.7 In the middle of the 1990s, however, amalgamations did become more widely discussed, especially in Andreas Ladner and Reto Steiner 243 cantons like Fribourg, Thurgau, Lucerne, Tessin and Graubünden, where, with the exception of Lucerne, the municipalities are particularly small. Considering the stability of municipal boundaries it is noteworthy that almost 20 per cent of the municipalities surveyed in 1998 claimed to have been discussing an amalgamation with one or more neighbouring municipalities. However, only 8 per cent of the municipalities, mainly from the cantons named above, have more precise plans and projects. A large-scale reform seems unlikely for the foreseeable future. To bring all municipalities to a minimal size of 3,000 inhabitants, as was recently suggested by authorities in the canton of Lucerne, would mean that 80 per cent of the municipalities would have to merge. It furthermore remains an open question whether there is anything like an optimal size for municipalities, and whether the traditional form of territorial units with a general responsibility for all local tasks will survive. However, it is likely that there will be some hundred amalgamations of rather small municipalities (out of 2,842) in the next twenty years (Steiner, 2002). In the 1990s the wave of New Public Management (NPM) also swept across Switzerland. Initial pilot projects were launched in municipalities within the cantons of Bern, Basel-Land and Zurich in the early 1990s (Haldemann and Schedler, 1995: 100). Since then, the number of NPM reform projects has risen considerably. Our survey of the cantons in 1998 showed that NPM reform projects were under way in 24 out of 26 Swiss cantons (Ladner and Steiner, 1998: 23), and at the end of 1998 the survey among the municipalities revealed that more than a third had already considered NPM reforms (Ladner et al., 2000: 128). Although it may not be said that NPM is an issue throughout Switzerland, NPM theory has been received favourably in practice. Approximately one quarter of the municipalities claim that they not only have considered NPM reforms but that they have even taken first steps towards implementing NPM measures. Among the municipalities claiming that they have taken first steps, only a small portion have already implemented core elements of NPM such as product definitions, global budgets, performance agreements, and contracts. These elements are essential in operating a municipal administration in a performance- and outcome-oriented fashion (Schedler and Proeller, 2000: 121–2; Steiner, 2000). More popular in Swiss municipalities are changes in the human resources area (elimination of the civil servant status, performance- related pay, and personnel development measures). This is related to the goal of aligning working conditions in the public sector with those in the private sector. Alongside human resource management, normative strategic management is also becoming increasingly important. This trend can be seen in the creation of mission statements and the delegation of operational tasks from the political to the administrative sphere. One-third of municipalities have undertaken changes in local representative democracy (e.g. changes in the size of the local government or a shift 244 Reforming the Swiss Municipalities in operational tasks among the legislative, executive, and administrative branches). As a general trend, many municipalities have tried to make decision making both easier and more efficient by shifting power from the citizens or the legislative to the executive body, and from the executive to the administration or to the different specialized commissions (for schools, planning, construction and so on), and by reducing the number of commissions or by reducing the size of all these bodies. In some cases they have extended the administration or transferred services and tasks to the private sector. Reforms of the political system have been rarer. There has been tendency to replace majority voting with proportional representation voting as far as the electoral system for the local executives is concerned, although there are also some examples of shifts from PR to majority voting. In addition, municipalities with rather large executive bodies have tended to reduce the number of seats. All these changes, however, may be considered as minor reforms, which are hardly likely to increase overall municipal performance. With regard to efforts to increase citizen participation in decision making, the situation is somewhat different. Swiss citizens have long enjoyed far-reaching means of influencing political decisions. This applies not only at the national level, but especially at the local level. Basically, there are two distinct systems: the bigger municipalities and many of those in the French-speaking part of Switzerland have a local parliament. Here the citizens have the possibility to launch initiatives to change the communal code and communal laws and to initiate optional referenda against projects and decisions of the parliament or the executive. Furthermore, in the case of expenditures above a certain limit, there are compulsory referendums. Smaller municipalities have a local assembly which is held between two and four times a year. Here, all citizens are entitled to vote on the proposals put forward by the executive. In the course of the assembly they also have the possibility of altering the content of these proposals. In most of the assembly municipalities, citizens also have the possibility to launch initiatives by collecting signatures, to seek a referendum against proposals and decisions of the executive, and sometimes even against decisions of the local assembly (Ladner, 1991b; Lafitte, 1987). Given the existing possibilities for citizen influence, the idea of introducing or increasing direct democracy is not an issue in Swiss municipal reforms. If there are changes, they are aiming at making compulsory (financial) referendums optional. This implies a reduction in the number of the referendums to be held. If no opposition against a project arises, no ballot has to be held. However, there have been attempts to increase the involvement of citizens in local politics, particularly with the aim of committing people to a project and preventing a failure in the final decision. In a survey conducted in 1994, almost two-thirds of the municipalities who responded claimed that they more frequently organize meetings to inform their citizens to try and integrate them in the planning process for municipal projects. In comparison, only around 15 per cent claimed that they have increased the scope of initiatives Andreas Ladner and Reto Steiner 245 and referendums. Additionally, New Public Management reforms also oblige authorities to find out whether their ‘customers’ agree with the quality of the goods and services they provide and increasingly push them to conduct surveys. Goal-oriented political management needs to know what goals are to be achieved. Ideally, these goals have to be defined jointly with the citizens and fixed with a mid-term perspective in a municipal development or legislative programme. In the light of these last results it would nevertheless be wrong to conclude that reform activities in Switzerland are primarily about giving the citizens more power in terms of binding decisions. These reforms are directed more at increasing the participation of the citizens at the very beginning of a project in order to ensure their support when it comes to the final decisions. Our survey of the municipal secretaries also provided some information on where and how reforms are being implemented. Context variables, such as ‘size of the municipality’, ‘language area’ and percentage of Catholics living in the municipality, performance variables, such as the municipality’s financial situation and the perceived performance thresholds together with political variables like the strength of different political parties can be expected to have an influence on the general reform activities and the most frequently occurring reforms such as the intensification of intermunicipal cooperation (IMC) during the past five years, executive reforms, legislative reforms, the existence of concrete municipality merger projects, and the introduction of NPM (Table 10.2).8 Most of the models explaining the different reform activities, however, are hardly satisfactory. Only in the case of NPM reforms and merger plans does the Pseudo-R Square exceed the value of 0.2. The most important variable in explaining the extent of reform activities is the size of a municipality. If we control for all of the other variables, size has a positive influence on executive and legislative reforms, on NPM reforms and, perhaps surprisingly, on the intensification of cooperation. It appears as if a second wave of cooperation is about to take place. Whereas the smaller municipalities had to cooperate in many fields (schools, infrastructural services) right from the beginning, it is now the bigger municipalities which seek economies of scale. The negative influence of size on amalgamation had to be expected. It was in the small municipalities that the merging projects started. A significant increase in population size, however, remains without any influence on the reform activities. The language area also has some influence. In the German-speaking municipalities, New Public Management reforms and reforms of the executive are more widespread. And, finally, the dominant religion does not remain completely without influence. In Catholic cantons the municipalities are more likely to merge but there are fewer executive und NPM reforms. Compared to the context variables the performance variables are of minor importance. Only the likelihood of merging plans is to some extent increased by budget deficits and the perceived performance threshold has a positive influence 246

** 0.005 − New Public Management reforms **

Legislative reforms at least. ** 0.005 n.s. 0.005 n.s. n.s. − − Executive reforms ** 0.437 0.294 0.795 1.759 0.795 0.437 0.294 0.168 n.s. n.s. n.s. n.s. n.s. n.s. n.s. − − 1871 18871871 1887 1789 Concrete Concrete merger plans

n.s. Intensification of cooperation ** * index

1.171 0.318 0.005 n.s. 0.009 3 1879 0.325 n.s. n.s. n.s. n.s. n.s. n.s. 0.325 n.s. n.s. n.s. 0.719 n.s. n.s. 0.434 n.s. 1.228 n.s. 0.719 n.s. 0.434 n.s. 0.143 0.007 0.241 0.047 0.094 0.263 0.094 0.047 0.241 0.143 0.007 n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. 0.025 n.s. 0.025 n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s.

Explaining different reform activities through characteristics of the municipality (linear and binary logistic regression logistic binary and (linear municipality of the characteristics through activities reform different Explaining : *: multivariate Linear regression;** binary logistic regression, listed coefficients all significant are ata cent 5 per level last three years three last within the local executive executive the local within local executive executive local Table 10.2 Reform Notes models, regression coefficients) coefficients) regression models, Context variables Context Inhabitants (2000) Increase inhabitants 1990–2000 Performance variables the within the accounts balancing upon Surplus German speaking municipalitiesGerman speaking Percentage of Catholics (2000) Percentage of members of Democrats Percentage the Social Receiving money from intermunicipal money transfersmoney intermunicipal from money Receiving n.s. Percentage of members of the Radicals within the within of members of the Radicals Percentage Adj. R Square/Nagelkerke R Square Political variables Political N 184 Perceived performance threshold (Index) Andreas Ladner and Reto Steiner 247 on the overall reform index. The influence of political variables such as the Social Democrats’ or the Radicals’ share in the executive only has an influence on executive and legislative reforms. If the Social Democrats are stronger there are more executive reforms, and if the Radicals are stronger there are fewer legislative reforms. It may be surprising that financial difficulties in the municipality do not, or only slightly, correlate with most of the reforms, so that breakdown or crisis approaches do not gain strong support from these results. The performance threshold index also points in this direction. If the municipalities report that they have reached performance thresholds at an especially high frequency, then one may not assume that they are particularly enthusiastic about reforms.

3 Reforms as intentional processes – triggers of reform projects

Reforms are a particular form of institutional change. In at least two respects they share distinct common features that make them analytically interesting: first, reforms are concrete processes or projects to change political institutions, and, secondly, reforms are goal-oriented. Consequently, reform processes are intentional processes. In addition to the general concept of institutional change, reforms have a beginning and an end, and they pursue explicit goals. The triggers of reforms give us a first idea about the reasons behind them. The explanations offered by reform literature make a difference between ‘intentional agents’, ‘mismatch dynamics’, and ‘external shocks and environmental pressures’ (Norgaard, 1996: 49). March and Olsen (1989: 58 ff.) more precisely identify six explanatory approaches, in order to interpret changes in organizations and institutions:

1. Variation and Selection: This corresponds to the model of evolution. Institutions evolve or change in order to meet the demands and requirements of a changing environment. The changes result from ‘surviving’ in the face of competition. 2. Problem Solving: Changes are due to certain decision-making criteria and to the expected effects, and choices are made between alternative solutions. This corresponds to the rational choice approach, which is found in decision-making theory, as well as in microeconomics and the behavioural sciences. 3. Experimental Learning: Reforms emerge from ‘trial and error’. Rules that have proven to be successful in an institution are maintained; unsuccessful rules, on the other hand, are abandoned. 4. Conflict: In this approach, changes are seen as a product of conflicts between individuals and groups with different interests. Confrontations, negotiations, and coalitions are the corresponding processes. The result depends upon the original preferences and the political weight of the agents. Changes occur if the participants or the controlled resources can be mobilized. 248 Reforming the Swiss Municipalities

5. Contagion: Reforms spread by means of emulation. The attractiveness of the new solutions of competitors tempts the agents to act and reform. 6. Turnover: Changes with respect to the participating agents and their capabilities, especially those caused by the recruitment of new agents with other opinions, capabilities, and goals, lead to institutional change. This model corresponds to the idea of regeneration.

These triggers are by no means mutually exclusive (e.g. March and Olsen, 1989: 59). A municipal merger can, for example, have various causes. A local government may have learned about the positive experiences of other municipalities and this may coincide with the election of a mayor who is more reform-oriented. The questions, however, remain the same. To what extent do reforms just happen and to what extent are they caused by agents? May reforms be considered deliberate institutional policy choices aware of what they are aiming at or are they simply attempts to improve the actual situation without knowing much about their chances of success? Or, more precisely focused on the causes and aims of reform processes, can they be seen as intentional changes to cope with deficiencies concerning efficiency and input legitimacy or are they to be seen as changes in the institutional setting taking up more or less assumed demands of a changing society? The answers of our respondents from the 199 case studies carried out in 2001 help to shed light on these questions. The results show a picture similar to the survey of all Swiss municipalities (see Ladner et al., 2000): reforms happen and reforms are made. According to the opinions of the municipalities, the causes for reforms can be attributed especially to general societal change (Figure 10.1).9 But the reforms are also the product of a careful evaluation of advantages and disadvantages. Individual agents distinguishing themselves politically (‘conflict’) and the testing of new solutions (‘experimental learning’) cannot be counted among the appropriate backgrounds for the emergence of reform projects. In at least some cases the new solutions were due to new persons (‘turnover’) and emulation (‘contagion’). If we consider the significance of the different models for the individual areas of reform, considerable similarities become apparent. In only a few cases does the evaluation of the individual types of reform deviate from the overall trend: part of the few clear-cut exceptions are the NPM projects which, in contrast to the other projects, are not products of a careful evaluation. Furthermore, it is striking that NPM projects are more likely to be more dependent than other projects on personnel changes in the governmental agencies. It is also striking that the merger of political municipalities occurs at an above-average rate, as a result of an evaluation of advantages and disadvantages, and that emulation is the most prominent cause of executive reforms. Andreas Ladner and Reto Steiner 249

1.5

1 Variation and Selection Problem Solving 0.5 Turnover Contagion 0

–0.5

–1 Conflict Experimental Learning –1.5

–2

Figure 10.1 Triggers of reform processes according to the assessments by the municipal secretaries (case studies) Response categories: Variation and selection: New solutions logically result from changes in society. Problem solving: Careful evaluation has shown that one is dealing with a better solution. Experimental Learning: One should attempt something different for a change; perhaps one could discover a better solution. Conflict: Various political actors try to distinguish themselves by offering new suggestions for solutions. Contagion: Other municipalities have had success with similar projects. Turnover: Changes in responsible personnel and their capacities lead to new solutions. 2 = fully applies, −2 does not apply at all, n = 199.

The analysis of the municipalities’ responses to the question of reform triggers, which were provided in an open dialogue, provides more details about the origin of the different reform types:

• NPM was often introduced by a parliamentary initiative. In some municipalities, the initiative came from local governments or higher-ranking officials, or the reforms were suggested by other municipalities or by the superordinate government level. One characteristic of NPM reforms is that the projects were seldom launched as a result of initiatives ‘from outside’ (superordinate government level, continuing education event). • In the cases of the executive reforms, the projects can be traced back to the difficulty of finding qualified personnel capable of reconciling political mandates with professional activities. In some cases, changes in 250 Reforming the Swiss Municipalities

the structure of the municipal institutions (e.g. separation of the municipal public works and changes in administrative procedures) were responsible for launching the project. In the case of parliamentary reforms the detailed responses did not exhibit any clear pattern. • The intensification of intermunicipal cooperation is, on the one hand, triggered by problems of recruiting personnel, and, on the other hand, by initiatives from the superordinate government level: In some cases, tasks were delegated to the municipality by the superordinate government levels; in other cases, the superordinate government level launched a reform project; and in a few cases, the subsidy practice was altered. The pressure to reform, which can be triggered by the canton, however, was clearly evident. • For the mergers of municipalities the important role of the canton as superordinate government level is evident, as had been the case with intermunicipal cooperation. Reforms were triggered either by new regulations in cantonal laws or by merger-development projects (for example, by organizing financial contributions). There were, however, also mergers resulting from political initiatives (emanating from involved citizens, associations, or the parliament). A long-standing cooperation with other municipalities, on the other hand, seems to have played an insignificant role. Thus, problems in the recruitment of personnel for political offices or costly individual objects (e.g. a school) appear to play a rather subordinate role.

Even if the results must be carefully interpreted because of the limited number of cases, the importance of the canton as a superordinate government level was clearly expressed. A second ‘concrete trigger’ of reforms is difficulties in recruitment. With the exception of NPM projects, they occur in all types of reforms. The difficulty with recruitment is frequently mentioned in the case of executive reforms (reduction in the size of local government) and in the case of intermunicipal cooperation (consolidation of the municipal administration). It is mentioned less frequently in the case of parliamentary reforms and rarely in the case of mergers. In contrast, internal municipal crises, such as, for example, abuse of power by individual members of the local governments or cases of financial misconduct, are much less frequently cited as concrete triggers of municipal reform projects. The case studies confirm another result from the nationwide survey of the municipalities. There is no strong connection between the financial situation of a municipality and the reform activities. More than half of our respondents deny a direct link between their reform project and the financial situation of their municipality. If there is a connection it is mainly in the fields of intermunicipal cooperation and the merging of municipalities. Also interesting to note is the case of the NPM reforms. It is not the financially badly off municipalities which start NPM projects. In 70 per cent of the projects such a direct link was denied. Andreas Ladner and Reto Steiner 251

The nature of the reforms is further revealed by the question of whether the reforms have been initiated from inside the political or administrative system or from the outside, i.e. from the citizens. In the case of the examined case studies, the initiative came primarily from the executive branches of the municipalities, followed by the mayors and the municipal secretaries (Table 10.3). These results correlate almost exactly with the data from the Switzerland-wide survey of all municipal secretaries. In contrast to the other reforms, NPM reforms are more frequently initiated by the local authorities and the administration. Executive and parliamentary reforms are primarily triggered by politicians, as are intermunicipal cooperation and the mergers of municipalities, although it is noticeable here that municipal secretaries and the higher-ranking administrative specialists contribute much less to the reforms, probably because they are politically very sensitive. A consideration of the triggers of reform projects and their promoters leads us to believe that there is a proactive element behind this form of institutional change. Reforms may be interpreted as a form of finding solutions for existing problems and the actors responsible have an idea of what

Table 10.3 Initiators of reforms according to reform type (proportions in percentages, multiple responses possible) (case studies)

Initiative NPM Executive Parliament Intermunicipal Mergers Total Cooperation

Mayor 36.4 40.0 30.0 45.1 38.3 38.1 Member of local 48.5 70.0 30.0 46.0 59.6 49.0 government (executive branch) Member of legislature 15.2 20.0 35.0 2.0 2.1 8.6 Municipal secretary 39.4 30.0 25.0 27.5 10.6 23.9 Upper-level 45.5 0.0 10.0 27.5 4.3 19.3 administrative specialist Mid-level administrative 6.1 0.0 5.0 5.9 0.0 3.0 specialist Low-level administrative 0.0 0.0 0.0 0.0 0.0 0.0 specialist Administrative 0.0 0.0 0.0 0.0 0.0 0.0 employee Citizenry 0.0 10.0 45.0 5.9 19.1 15.7 Canton 6.1 0.0 10.0 19.6 40.4 19.7 Media 3.0 0.0 0.0 2.0 2.1 1.5 Scientists 6.1 0.0 0.0 0.0 0.0 1.5 External consultants 3.0 10.0 0.0 5.9 10.6 6.6 n 33 10 20 51 47 198 252 Reforming the Swiss Municipalities they want to improve. Whether this is democracy or efficiency will be the crucial question addressed by the next section.

4 Goals: democracy or efficiency

To decide whether the reforms aim at increasing the input- or the output- oriented legitimacy, it is essential to look at the goals of the different reform projects. There is a wide range of such goals, reaching from participation, representation and responsiveness to efficiency and effectiveness. Referring to the so-called democratic dilemma (Dahl, 1994), there is supposed to be a trade-off between democracy and efficiency (e.g. Mouritzen, 1989): An increase in democracy leads to less efficiency and vice versa. In a long-term perspective, however, it is probably less a question of either democracy or efficiency but rather a question of finding a balance between the two. Legitimacy does not only stem from the output side of democracy but also from the democratic qualities of the decision-making process, telling the authorities what lies within the scope of the state and how this has to be done. In such a perspective, there is no efficiency without legitimacy. Reform activities during the 1990s were often accused of following the steps of neo-liberalism and of being biased towards more efficiency and effectiveness. Output-oriented legitimacy was supposed to be more important than democratic decision making. In Switzerland, where direct democracy offers far-reaching possibilities for citizen participation, such a trend would have been particularly salient. Among all of the reform projects the most commonly shared goal is the attempt to improve the quality of the services provided, followed by better possibilities to influence the development of the municipality (Figure 10.2). A reduction of service costs and speeding up decision making are also named as goals of the reform projects. Not mentioned were goals such as the reduction of services, an increase in quantitative terms, getting closer to the citizens, a higher rate of participation and strengthening democracy. Taking all reform projects together, it is therefore as expected, efficiency and effectiveness what all these reforms in Swiss municipalities are about. A better relationship between the authorities and the citizens and an improvement of democracy are not explicit aims of the reforms, in contrast to what happens in Germany with the ‘Bürgerkommune’ and the introduction of means of direct democracy. However, the reform projects are not simply attempts to save money or reduce the scope of the services provided by the state. These nine goals are partly interlinked and can be reduced to three groups by a factor analysis.10 First, there are goals referring directly to the input side of democracy – that is, attempting to make the state more responsive or to increase citizen participation. A second group of goals is related to the output side of democracy. They strive to improve services either in terms of quality or in terms of quantity. A third group of reform goals concentrates Andreas Ladner and Reto Steiner 253

1.5 Better services 1

Reduction of 0.5 costs Moving closer to Strengthening the citizens democracy 0 Speeding up Better More services development of decisions –0.5 the municipality Increase in participation –1

–1.5 Reduction in the services provided –2

Figure 10.2 Goals of the reform projects (case studies) = − = Notes: 2 fully applies, 2 does not apply at all, nmin 199. on the costs for the different services provided, on the municipalities’ competitiveness and their ability to decide on important issues rapidly. These reforms are also concerned with the output side of democracy, and even more so with the municipalities’ capacity to act. Different types of reforms have different goals (see Table 10.4). New Public Management reforms try to improve both the input side and the output side, at least in terms of improving the capacity to act. Executive and parliamentary reforms try to strengthen democracy, whereas cooperation and amalgamation are mainly concerned with offering better services. Closely connected to the idea of amalgamation is an improvement of the ability to act. Considering the fact that executive and parliamentary reforms are minor changes compared to cooperation and amalgamation, Swiss municipal

Table 10.4 The different types of reforms and their goals (factorscores, means)

Democracy Better services Capacity to act

New Public Management reforms 0.35 −0.04 0.60 Executive reforms 0.19 −0.62 −0.33 Legislative reforms 1.18 −0.50 −0.25 Intermunicipal cooperation −0.64 0.48 −0.47 Amalgamation of municipalities −0.42 0.28 0.32 254 Reforming the Swiss Municipalities reforms are still much more concerned with enhancing output legitimacy. New Public Management is in this respect a somewhat ambiguous concept, since it combines the ability to act with an increased customer orientation.

5Conclusions

As far as reform activities at the local level are concerned, Switzerland fits quite well into the international pattern. As in many other countries, New Public Management has been an important reform topic in recent years. There have been attempts to amalgamate municipalities and there has been an increase in intermunicipal cooperation. Finally, there have also been reforms aiming more directly at the political authorities and the decision- making process. All of these reforms may not be considered as a mere product of social change. Changes in society, a scarcity of resources, and more demanding citizens, were partly responsible for the reforms. But equally responsible were political actors and a better understanding of how to organize and manage municipalities more effectively. The reform projects of the 1990s – especially the intermunicipal cooperation and the merging of municipalities – aimed at making service provision at the local level more efficient and at increasing the quality of the services provided. New Public Management reform tried to increase the municipalities’ ability to act but also aimed at moving closer to citizens by increasing customer orientation. The reforms were not just about reducing services. In this respect they cannot be regarded as reforms trying to downsize the state sector, but much more as reforms trying to improve state work. However, strengthening the input side of democracy by moving decisional power to the citizens was also of lesser importance in Swiss municipalities than in other countries. This is mainly due to the far-reaching direct democratic rights of Swiss citizens. Despite obvious parallels regarding the reform activities in other countries, it is misleading to speak of a convergence of Swiss municipalities in the direction of a single international model. First, there is no such model in sight and, secondly, their smallness, their fiscal and political autonomy, and their direct democratic political systems, still distinguish them to a considerable degree from the municipalities in many other countries. Not even within Switzerland is there a predominant model for an ideal form of municipal organization to which all municipalities are converging. To some extent this can be seen as positive. By and large the same problems are approached with different and competing solutions. However, considering the enormous differences between the municipalities and the still increasing gap between the urban and rural areas, it remains an open question whether all municipalities can be legally treated in the same way in future. Andreas Ladner and Reto Steiner 255

Notes

1. In 2000 some 54 per cent of the ca. 2,900 municipalities had fewer than one thousand inhabitants. 2. Swiss municipalities generate 70 per cent of their income through their own financial means. No other European country has smaller transfers from the superordinate state government to the municipalities (Council of Europe 1997). 3. On the grounds of cantonal laws, municipalities choose their appropriate structure and administrative organization, levy taxes, and independently fulfil those tasks that do not lie within the jurisdiction of the cantons or the federal government (Linder, 1999: 156 ff.). 4. In autumn 1998, a 16-page questionnaire was sent to all Swiss municipal secretaries. A total of 2,465 municipalities responded, which corresponds to a response rate of 84.5 per cent. The municipal secretary assumes a top leadership role within the municipality. The function is comparable to the town clerk or the mayor in other countries. 5. The municipal secretaries had to note on a list containing 26 possible reform projects whether their municipality had introduced such reforms or not. The list was the result of several interviews with leading experts. 6. However, the idea that municipalities should work together to provide certain services more efficiently is by no means new. In the history of Swiss municipalities there have always been forms of cooperation, which is not astonishing if we consider how small many of them are. The traditional form of intermunicipal cooperation is an administrative union (‘Zweckverband’), an association under public law. In recent years, however, municipalities have increasingly cooperated on the grounds of private law, which offers them more flexibility. 7. The average Swiss municipality has only 2,501 inhabitants (Bundesamt für Statistik, 2001). Between 1848 and 2003, the number of political municipalities decreased only from 3,203 to 2,842. 8. The reform index counts the different reform activities and ranges from 0 to 17 with a mean of 6. The reform activities are coded as binary variables. 9. The six approaches of March and Olsen (1989) were rendered operational with the help of statements to which the municipalities were obliged to take a position. This was done by an assessment based on a scale from 1 to 5. Within the framework of an openly structured dialogue, the municipalities could moreover declare their position regarding what, in their opinion, were the respective triggers of reform. 10. The corresponding factor analysis (principal component, varimax) reduces the variance of the 9 variables to three factors with Eigenvalues of 2.8, 1.6 and 1.1). The three factors explain 62 per cent of the variance. 11 The Lost Dimension of Swiss Federalism: Democracy Problems of New Regionalism in Metropolitan Areas Daniel Kübler

Emerging in the 1990s, the concept of new regionalism focuses on the building of area-wide metropolitan governance through networks of voluntary cooperation rather than through institutional consolidation (‘old regionalism’). This chapter focuses on a yet under-investigated aspect of new regionalism, namely its effects on the relationship between citizens and the state. In the first, conceptual, section three hypotheses are developed on how new regionalism impacts on service satisfaction, government loyalty as well as democratic legitimacy. The second section tests these hypotheses using case study evidence and survey data from four Swiss metropolitan areas (Zurich, Lausanne, Lucerne, Lugano). The results suggest that ‘new regionalism’ has no significant effect on service satisfaction or government loyalty, but reduces democratic legitimacy. A closer look at intervening variables reveals that size of communes and institutional consolidation of metropolitan areas play a stronger role than the type of regionalism. Hence, efforts to increase area- wide governance capacity in Swiss metropolitan areas should pursue ‘old’ rather than ‘new regionalism’.

1 Introduction1

During the twentieth century, urban sprawl has resulted in the emergence of large metropolitan areas2 in Switzerland, stretching across a multitude of local jurisdictions. In other OECD countries, the reaction to growing disparities between functional and institutional urban territories has con- sisted in adapting the size of local governments through amalgamation, or by the creation of new institutional structures – such as the Communautés urbaines in France and Canada, the Umlandverbände in Germany or the (recently created) Greater London Authority in England. In contrast to this international trend, the institutional structure of Swiss metropolitan areas has been virtually unchanged since the early twentieth century. As a consequence, in Switzerland institutional fragmentation of metropolitan areas is

256 Daniel Kübler 257 high and there is no distinct political authority responsible for the common needs of the population living there. As Linder (1994: 77) has put it, metropolitan areas are the ‘lost dimension of Swiss federalism’. Hence, the area-wide scope in metropolitan policy making is achieved essentially through purpose-oriented intergovernmental agreements involving cities, their suburbs, as well as cantons, and sometimes even the Confederation. The aim of this chapter is to assess the Swiss way of policy making in metropolitan areas with respect to core elements of democratic governance. My argument is developed in two main sections. The first section develops the conceptual framework used for this analysis. Drawing on the long-running theoretical debate about the organization of governance in metropolitan areas, I develop three hypotheses relating to how various types of region- alisms in metropolitan areas impact on central features of democratic governance, namely service satisfaction, government loyalty, as well as democratic legitimacy. In the second section, I test these hypotheses using case-study evidence as well as survey data from empirical research on the four metropolitan areas of Zurich, Lausanne, Lucerne and Lugano. In the conclusion, I discuss the significance of the results with respect to the general debate over the best ways to achieve metropolitan governance, and briefly comment on policy implications for Switzerland.

2 Regionalism, old and new

The organization of governance in metropolitan areas is one of the most debated topics of urban social science. Since the acceleration of urban sprawl in the late nineteenth and early twentieth century, generations of urban economists, geographers, sociologists, public administration specialists and political scientists have racked their brains over the best way to achieve regional governance in institutionally fragmented metropolitan areas. This long-running debate has produced several intellectual traditions and schools of thought.3

2.1 Old regionalism and its critics The older metropolitan reform tradition views the existence of a large number of independent jurisdictions within a metropolitan area as the basic problem. Scholars whose work reflects this tradition argue that the institutional fragmentation of functionally integrated areas hampers the efficiency of public services because it leads to diseconomies of scale, as well as democratic citizen participation on area-wide issues. In combination with a somewhat Weberian trust in the rationality and planning capacities of large public bureaucracies, institutional consolidation is advocated for metropolitan areas, either by amalgamating existing jurisdictions, or through the creation of two-tier metropolitan authorities to which competences on area-wide matters are transferred. Based on this line of reasoning, reforms towards institutional consolidation have been undertaken in many OECD countries. 258 Democracy Problems of New Regionalism

Annexations of suburbs by core cities have occurred across the globe in the late nineteenth and the early twentieth centuries. The 1960s and 1970s – the ‘golden age of metropolitan government reform’ (Lefèvre, 2001: 136) – have seen the establishment of encompassing two-tier authorities in major metropolitan areas in North America (Toronto, Montreal, Vancouver in Canada, as well as Minneapolis–Saint Paul and Portland in the USA) and across Europe (London and six Metropolitan Counties in England, four Communautés urbaines in France, Rotterdam, as well as Barcelona). However, drawing on a general distrust against ‘big government’ a critical public choice perspective on metropolitan governance has developed from the 1960s onwards. Drawing on Tiebout’s (1956) argument of people ‘voting with their feet’, public choice scholars hold that the competition of local authorities for taxpayers leads to effective matching of urban public service demands and increases efficiency in the provision of these services. More- over, they argue that large jurisdictions result in a greater distance existing between citizens and public authorities. Hence, public choice scholars are convinced that the institutional fragmentation of metropolitan areas into a multitude of small and autonomous jurisdictions is beneficial. Local autonomy is decisive and should be hampered neither by institutional consolidation, nor by higher-level interventions, and scale economies should be realized only by purpose-oriented voluntary cooperation. This public choice position heavily undermined the intellectual foundations for consolidation reforms in metropolitan areas. Disillusionment with the feasibility of such reforms as well as the rather poor performance of newly established metropolitan governments brought consolidation reforms to a halt during the 1980s. Over the last fifty years, the scholarly and political debate on metropolitan governance has largely been dominated by the dispute between the metropolitan reform tradition and the public choice perspective. However, both of them can provide only limited guidance for debating issues of governance in contemporary metropolitan areas. On the one hand, institutional consolidation was not really a success story (Lefèvre, 1998): most reforms have failed, many experiences with metropolitan authorities have proved disappointing, and some places have even witnessed the abolition of metropolitan governments created beforehand (e.g. Rotterdam in 1985, London and the English Metropolitan Counties in 1986, Barcelona in 1987). On the other hand, the public choice notion of unbiased competition between autonomous local authorities is also far removed from the reality found in most metropolitan areas (Frey and Eichenberger, 2001).

2.2 New regionalism and the different routes towards metropolitan governance In the 1990s, empirical work on metropolitan governance in Europe and North America has fuelled the emergence of a third approach, nowadays Daniel Kübler 259 known as new regionalism (Frisken and Norris, 2001). New regionalism conveys the notion that strong metropolitan governance does not necessarily require institutional consolidation. Instead, it argues that governance capacity in metropolitan areas can also be achieved by voluntary cooperation that stabilizes networks of policy-relevant actors. Whereas the two traditional schools of thought argue that there is one best way to go (consolidation/ multi-tiered authorities versus fragmentation), new regionalists hold that several paths may lead to area-wide governance. New regionalists therefore do not express any preference per se for either institutional consolidation, multi-tiered authorities, voluntary intergovernmental cooperation, regional networks based on public–private partnerships, or competition between autonomous localities. However, it is evident that most of the authors in the wake of new regionalism have concentrated on intergovernmental agreements and regional networks involving public–private partnerships, basically because experience has shown that, on the one hand, amalgamation or metropolitan governments have little chance to happen and, on the other hand, free competition between localities is very likely to be distorted by territorial spillover (externalities) and unequal resources. Instead, the major thrust of metropolitan governance today seems to lie with heterogeneous conglomerates of actors based on purpose-oriented cooperation that involve localities, governmental and non-governmental agencies at various territorial levels, as well as private service providers.4 New regionalist thinking therefore emphasizes governance building through a collaborative, non-hierarchical process that involves participants from both the public and the private sectors who are committed to shared leadership in metropolitan problem solving (Hamilton, 2002: 404–5). Thus, there clearly is a parallel between new regionalism and the debate on the transformation of the state, ‘from government to governance’ (see Le Galès, 1995; Rhodes, 1996; Stoker, 1998; Kooiman, 2003), conveying the notion that public policies are produced by fluid, self-governing networks rather than only by elaborate public bureaucracies.5 In spite of an impressive variety of definitions used to describe this supposed shift ‘from government to governance’,6 the various discussions of this phenomenon refer to a relatively stable grid of analytical elements serving to classify forms of interactions in the organization of political power and policy making, including at least (i) actors and (ii) conventions of interaction.7 Concerning actors, ‘governance’ is broader than ‘government’, covering state agencies from various governmental levels, as well as non-state actors. This involves a blurring of boundaries both within the public sector, but also between the public, the private and the voluntary sectors. Concerning conventions of interaction, ‘governance’ relies, more than ‘government’, on interactions rooted in trust and structured by rules of the game that are agreed upon in a cooperative manner by participants independently from formal hierarchies. In addition, Savitch and Vogel (2000: 161–2) have argued that there is an 260 Democracy Problems of New Regionalism intergovernmental dimension to the discussion on the shift ‘from government to governance’:

Whereas government is vertical and firmly institutionalised, governance is horizontal and flexible. Whereas government is formal and directed from above, governance is informal and self-regulating. Whereas higher level government (e.g. states) connects to localities through demarcated procedures, lower level governance (e.g. interlocal agreements) is looser and less confined by boundaries. Government emphasises the centralising features of regionalism, whereas governance stresses the decentralising virtues of local cooperation.

A summary of these arguments is presented in Table 11.1. It expresses the main thrust of new regionalist thinking, namely that area-wide governance capacity can not only be achieved through strong institutional solutions advocated by ‘old regionalism’ (i.e. government), but also through ‘new’ ways

Table 11.1 Four routes towards metropolitan governance

‘Old’ regionalism ‘New’ regionalism

Consolidated local Multi-tiered Linked functions Complex network government government governance governance

Actors: Actors: Actors: Actors: State actors of State actors of State actors of State actors of one single local several local several local several local and/or authority authorities. authorities. supra-local State actors of Agencies from authorities. supra-local the private and/ Agencies from the authority or voluntary private and/or sector voluntary sector. Conventions of Conventions of Conventions of Conventions of interaction: interaction: interaction: interaction: Formal ‘vertical’ Formalised Formalized Ad-hoc ‘horizontal’ interactions ruled ‘vertical’ ‘horizontal’ interactions by legislation on interaction interactions according to GOVERNANCE GOVERNMENT local public ruled by ruled by self-determined administration. agreements autonomously rules of the game. on shared negotiated autonomy. agreements. Examples: Example: Example: Example: Amalgamated Metropolitan Interlocal service Overlapping local jurisdiction government agreements webs of voluntary co-operation

Source: Adapted from Savitch and Vogel (2000: 165). Daniel Kübler 261 of network-based cooperation beyond the state (i.e. governance). More precisely, Table 11.1 suggests to distinguish four routes through which metropolitan governance can be achieved. This is done by classifying cooperational agreements and intergovernmental relations according to two features on which the debate on the shift from government to governance tends to focus, namely (a) actors and (b) conventions of interaction.

2.3 Regionalism and democratic governance: three hypotheses As we have seen, new regionalism entails the idea that area-wide governance is achieved by reharnessing the relationships between existing institutions rather than by creating multi-purpose regional institutions. New regionalists tend to argue that heterogeneous networks with variable geometry can be (at least) as good and as efficient as fully fledged regional governments in the provision of area-wide public services. But whereas new regionalists have been quite confident in defining what constitutes efficiency, effectiveness and equity in metropolitan service delivery, they have yet failed to argue about the implications that network-based metropolitan problem solving might have for democratic politics. In western democracies, the legitimacy of the state is seen to rest upon procedures and practices that give good reasons to believe that public policies are ultimately the result of public deliberation among free and equal citizens (Habermas, 1992). This is the raison d’être for mechanisms assuring accountability and citizen control over state activities and public policies in the sense of an ‘overhead democracy’ (Redford, 1969), that has traditionally been organized within Rechtsstaat-type governmental institutions featuring a clearly defined territory. Again, there are clear parallels between new regionalist thinking and the work on the transition ‘from government to governance’, where this shift is viewed as a considerable ‘challenge for democratic accountability’ (Rhodes, 1996: 53). Hence, there is reason to argue that new regionalism based on de-territorialized policy networks impacts on democracy. Some scholars indeed argue that new regionalism contributes to a ‘recomposition of the political’ (Le Galès, 1998: 501), that it has a democratic deficit and/or lacks democratic quality (Heinz, 2000; Benz, 2001). Since the relationship between the citizen and the state has traditionally been framed by institutional territories, it is likely that a de-territorialization of policy making in the wake of new regionalism changes this relationship. Although this issue has been raised by many scholars, to date it lacks empirical assessment. Therefore, the goal of the remaining part of this chapter is to investigate these questions. More precisely, we propose to assess the impacts of various metropolitan governance-types on the relationship between citizens and the state by subsequently focusing on three operational elements of democratic governance: service satisfaction, loyalty towards government and overall legitimacy. 262 Democracy Problems of New Regionalism

2.3.1 Citizens and the state (1): a focus on metropolitan service users The relationship between citizens and the state has been conceptualized in various ways. Based on Lowery et al. (1992: 70), we can say that a first, ‘liberal’ view of citizenship emphasizes individual rights, and sees the res publica primarily as a sphere of production and consumption of public goods. The citizen is conceived in terms of his or her role in this market of public goods, i.e. as a user who consumes public services and pays for it in form of taxes. Thus, a first – liberal – method of measuring the relationship between the citizens and the state in metropolitan areas is to focus on citizens’ perception of metropolitan services. Drawing on the above discussion on the long-running debate on metropolitan governance, we can isolate clear statements on how metropolitan governance structures impact on service satisfaction. According to the metropolitan reform tradition, efficiency, effectiveness and equity of metropolitan services are hampered by a high number of small governmental entities, but favoured by institutional consolidation and strong government. The public choice perspective adopts the opposite stance. In contrast to these clear-cut ideas, new regionalists have no a priori preference for a particular type of governance in the name of service efficiency, effectiveness or equity. If this latter contention is correct, we would expect the type of regionalism not to be a significant predictor for satisfaction with metropolitan services. Hence, a first hypothesis can be formulated:

H1: There is no significant relationship between new regionalism and satisfaction with regard to metropolitan public services.

2.3.2 Citizens and the state (2): loyalty towards elected government In contrast to the liberal view, the ‘communitarian’ conception of citizenship starts from the ancient Greek understanding of citizenship as the prideful and active participation in the common struggle over common problems (Lowery et al., 1992: 70). Here, the res publica is seen as the arena in which people get involved to decide about the community, and the citizen is conceived as an active participant in this arena. However, given evident practical limitations for active participation – citizens cannot participate actively on all issues at all times – Lowery et al. argue that a communitarian conception of citizenship can also be extended to include a representative element, namely loyalty towards elected representatives: ‘[O]n those issues in which direct participation is not possible or is mediated through the selection of representatives, loyalty will be a valid indicator of citizenship’ (1992: 801). Thus, a second – communitarian – way of examining the relationship between citizens and the state in metropolitan areas can consist in focusing on citizens’ loyalty towards those elected to decide about the fate of the community. Daniel Kübler 263

Again, the long-running theoretical debate on metropolitan governance provides several hints as to the impacts of metropolitan governance structures on loyalty towards elected representatives. Metropolitan reformers would argue that consolidated settings with strong government favours loyalty towards elected representatives, whereas public choice scholars contend precisely the opposite – namely, that such loyalty is better in fragmented settings characterized by small and autonomous local jurisdictions. Concerning new regionalists, they start from the assumption that area-wide governance can be achieved by a re-harnessing of existing (local) institutions: it comes about via the active involvement of various different authorities and agencies. Hence, they argue that, in a new regionalist setting, local authorities can play a more active role in metropolitan policy making, and citizens credit this engagement in managing the common fate, whereby loyalty towards them increases. Thus, the second hypothesis to be tested is as follows:

H2: New regionalism increases citizens’ loyalty towards local authorities.

2.3.3 Citizens and the state (3): democratic legitimacy Besides citizenship, another way to examine the relationship between citizens and the state is to focus on legitimacy. According to Kaase and Newton (1995: 168), legitimacy in the sense of a ‘micro concept which concerns individual beliefs about macro features of the political system’ is generally seen as ‘consisting of supportive attitudes and favourable evaluations of the political regime, that is, the non-partisan institutions, procedures, norms, and values of the polity’. Legitimacy measured at the individual level expresses not only the adhesion of the citizens to the rules of the political game, but also the perception of political efficacy – that is, the possibility for citizens to sanction effectively those in charge in case of dissatisfaction. Legitimacy has repeatedly been operationalized by satisfaction with the functioning of democracy in a given national or local context (see Kaase and Newton, 1995; Fuchs and Klingemann, 1995). With respect to the question of how various types of regionalism in metropolitan areas influence the legitimacy of political institutions in these areas, the three above-mentioned schools of thought portend contrasting views. On the one hand, metropolitan reformers would argue that low fragmentation, large entities and strong government structures in metropolitan areas (‘old regionalism’) increase the legitimacy of the metropolitan political system, since governmental activity can effectively act upon metropolitan problems and thereby prove its relevance to the citizens. On the other hand, public choice advocates would argue in the opposite direction – namely, that large-scale metropolitan governments place citizens at a distance and thus reduce their feeling of political efficacy, thereby resulting in the legitimacy of the political system. New regionalists have not yet developed a consistent 264 Democracy Problems of New Regionalism argumentation on consequences of new regionalism on legitimacy. However, a central tenet of new regionalism consists of achieving area-wide policy making through the involvement of non-governmental private agencies. It can therefore be argued that new regionalism contributes to the blurring of boundaries between the state and the private sector and thereby reduces the significance of state institutions in the policy process. In a situation where the state becomes increasingly irrelevant, citizens would lose the faith in its capacity to effectively aggregate political demands and solve public problems. Hence, we can formulate a third hypothesis, namely

H3: New regionalism reduces democratic legitimacy.

3 New regionalism and democracy in Swiss metropolitan areas

As in some other counties, urbanization in Switzerland began in the second half of the nineteenth century, when the growth of cities intensified together with the process of industrialization. As of the early twentieth century, this urban growth made cities extend beyond traditional communal boundaries, giving rise to a new sociospatial phenomenon: the metropolitan area, i.e. a functionally integrated urban space stretching across a multitude of local governments. After the Second World War, urban sprawl was accentuated during the suburbanization phase, when metropolitan areas continued to grow in parallel to the economy. From the 1960s onwards, due to soaring land prices in central locations and thanks to the ‘democratization’ of the private car that made commuting easier, new housing was constructed even outside suburbia, since wealthy families were seeking to leave expensive and congested core cities for individual housing ‘out in the green’. The result of this currently ongoing trend was exacerbated urban sprawl and a dramatic extension of metropolitan areas. According to the 2000 population census, 73.3 per cent of the Swiss population live in cities or metropolitan areas, as compared to only 38.9 per cent in 1941. Thus, in spite of the widely publi- cized bucolic image as the home of Heidi, cheese and chocolate, Switzerland must be considered as a heavily urbanized country, with its metropolitan areas being the main motors of economic, social and cultural life.8 The question of how metropolitan areas can be governed is thus a significant issue in Switzerland, and significant attempts have been made to achieve such governance. They provide an excellent ground to empirically assess the general questions with respect to metropolitan governance developed above, and test the three derived hypotheses. We will do so by focusing on four major Swiss metropolitan areas – namely Zurich, Lausanne, Lucerne and Lugano. More precisely, we will first present their general context, then work out the ways in which area-wide governance is achieved in four major policy fields, and then finally test each of the three hypothesis using multivariate regression on survey data. Daniel Kübler 265

3.1 The general context: institutional fragmentation and economic disparities In spite of continued urban sprawl and the growing significance of metropolitan areas in Switzerland, their institutional structure has been virtually unchanged since the beginning of the twentieth century. As far as the communal level is concerned, the annexation of suburban communes by major cities has occurred between 1893 and 1934. Since 1934, such amalgamations have been very rare or have been insignificant in terms of population or surface. Indeed, most suburban communes were able to consolidate their economic basis, whereas core cities ran into financial difficulties due to the loss of wealthy taxpayers during the phases of sub- and peri-urbanization. Since communes have to agree to amalgamation and cannot simply be obliged to do so by higher state levels,9 this configuration is an obstacle to territorial reform in most metropolitan areas: rich suburban communes strongly oppose any step towards annexation by financially distressed core cities (Geser, 1999: 426). Only the extraordinarily wealthy city of Lugano managed, in 2004, to annex eight suburbs, thereby almost doubling its population. Similarly, the territory of cantons – the intermediate territorial level between communes and the Swiss Confederation – has not witnessed any major changes since the mid-nineteenth century. Cantons as the federate states are the constitutive entities of Swiss federalism and the formal hurdles to changing cantonal boundaries are so high that they are almost impossible to implement (Germann, 1999: 397). Finally, the creation of regional institutions (i.e. situated between the communes and the cantonal level) during the 1970s has been limited to peripheral mountainous regions in order to improve infrastructure for economic development. Hence, in contrast to most other Western European nations Switzerland has seen no significant reform of institutional territories in metropolitan areas during the twentieth century. As a consequence, the institutional fragmentation of metropolitan areas is high in Switzerland (see Table 11.2). And given the strong resistance against territorial reforms in metropolitan areas (Kübler et al., 2001), fragmentation is very likely to remain high in the years to come. This is also true for the four metropolitan areas under scrutiny here Table 11.2. Over the years the urbanization process has also led to social segregation within Swiss metropolitan areas. In the core cities, the exodus of wealthy families starting with sub-urbanisation and intensified with peri-urbanisation has left behind above-average proportions of economically disadvantaged segments of the population. In the suburban zone, industry and newly located service activities such as retail and shopping malls attracted low-qualified workers and employees. Because of low wages, the proportion of immigrants among them is high. In the peri-urban communes, the situation is completely 266 Democracy Problems of New Regionalism

Table 11.2 Institutional fragmentation of Swiss metropolitan areas (in 2000)

Metropolitan area Population Number of communes Herfindahl’s index of institutional consolidation1

Zurich 1,084,027 133 0.11932 Lausanne 311,441 70 0.17604 Lucerne 196,550 17 0.14746 Lugano 136,032 77 0.06635

Source: Swiss Statistical Office, population census 2000. Note: 1Herfindahl’s index (Herfindahl, 1950) was originally constructed to measure market concentration: it is defined as the sum of all squared market shares. Here, it is defined as the sum of squared communal shares of the total metropolitan population and used as a measure for institutional consolidation. The index ranges from 0 to 1: the higher the value the more consolidated is the institutional structure of a metropolitan area, the lower the value, the more fragmented the structure. (Thanks to Henry Bäck to whom I owe this suggestion of using Herfindahl’s index to measure consolidation-fragmentation of metropolitan areas.) different. Residents are mostly wealthy families who can afford individual housing.10 The fragmented institutional setting and social segregation in Switzerland has led to significant economic and fiscal disparities between communes within metropolitan areas. Due to the significant financial autonomy of Swiss communes,11 institutional fragmentation has induced spillovers into the distribution of costs and benefits, as a result of which surrounding communes free-ride on centrality infrastructures provided by the core cities. Together with social segregation, such uncompensated centrality charges have resulted, over the years, in significant fiscal disparities within Swiss metropolitan areas. Whereas core cities were obliged to raise their tax rates in order to equilibrate income and expenditures, suburban, peri-urban and especially rich communes could lower their tax rate since expenditures for urban public infrastructure or social services also tends to be low in such communes (see Table 11.3).

Table 11.3 Communal tax rates according to types of communes (indexed to the mean of all communal tax rates within the metropolitan area; mean = 100)

Metropolitan area Core cities Suburban communes Rich communes Peri-urban com.

Zurich 116 101 79 107 Lausanne 118 100 76 105 Lucerne 104 104 56 102 Lugano 90 100 89 103

Source: Fiscal offices of the cantons of Zurich, Vaud, Lucerne and Ticino, data for 2001. Daniel Kübler 267

3.2 Area-wide governance in Swiss metropolitan areas: case study evidence In spite of institutional fragmentation, social segregation and fiscal disparities, there is area-wide governance in Swiss metropolitan areas. However, because of the absence of institutional reforms, it is mainly based on horizontal, as well as vertical cooperation between existing institutional entities, sometimes with significant involvement of non-governmental actors. Such cooperation is mostly purpose-oriented, i.e. organized differently according to each policy field, and is found in a broad variety of forms. This is also true in the four metropolitan areas under scrutiny here, Zurich, Lausanne, Lucerne and Lugano, where area-wide governance was examined in the fields of water supply, public transport, social services for drug users, and cultural institutions.12

3.2.1 Water supply The organization of water supply in Switzerland lies in the competency of local authorities. Traditionally, water supply is organized by local public enterprises. In the wake of the deregulation of the energy market in Europe during the late 1990s, some local authorities have privatized their water departments, together with gas and electricity services, as part of a strategy to face future market competition in these fields. The recognition of a ‘metropolitan’ – i.e. supra-local – scope for water supply had occurred in the 1970s. The shift from local to supra-local management back then was mainly driven by the quest better to exploit scarce water resources in reaction to urban population growth. In all of the metropolitan areas under scrutiny here, water supply is organized on a supra-local scale at least since the 1970s and involves the core city and a certain number of surrounding communes. However, there are very different ways by which supra-local coordination is achieved. In the Lausanne and Zurich agglomerations, suburban localities purchase water from the communal water department of the core city, on the basis of long-term ‘pay-for-service’ contracts. The same applies to the Lucerne and Lugano agglomerations, with the only difference being that there, the core cities have privatized their water departments and transformed them into joint stock companies.

3.2.2 Public transport In the area of public transport, planning and management today occur on an area-wide basis in all four metropolitan areas. The supra-local scope started to be recognised during the 1980s and the corresponding mechanisms were established in the early 1990s, generally following the revision of cantonal legislation on public transport (Zurich: 1987; Lausanne: 1990; Lugano: 1994; Lucerne: 1996). In all places, the regionalization of public transport was simultaneous to a new emphasis on the development of public transport 268 Democracy Problems of New Regionalism as a response to problems induced by automobile commuter traffic, such as pollution, noise and congestion. In parallel, transport infrastructure was improved (especially in Zurich and Lausanne) and attractive area-wide season- tickets were introduced (in Zurich, Lucerne and Lugano). All coordination schemes in the field of public transport involve at least cantonal agencies and localities and thus appear as two-tiered bodies. A closer look, however, reveals a great variety of ways in which area-wide governance is achieved. The cases of Zurich and Lugano may be considered truly two-tiered since both cantonal agencies and localities are substantially involved in the definition of the service offer. However, whereas the Zurich scheme is more centralized and the main power lies with the canton, localities in the Lugano scheme have more discretion and influence through an inter-communal commission whose decisions are then generally followed by the canton. In the cases of Lucerne and Lausanne, the canton has delegated the definition and implementation of the service offer to a regional transport company, funded mainly by localities and other public agencies.

3.2.3 Social services for drug users Social services for drug users started to be provided in the mid-1980s, as a response to the spread of HIV/Aids through the sharing of needles and drug-related prostitution. At their inception, such services were implemented only by core cities. Suburbs were then quickly accused of externalizing costs and nuisances linked to the drug problem. In a process that was sometimes conflictual, the regional scope of the service offer for drug users came to be acknowledged from the late 1980s onwards. Regional mechanisms for service management and cost-sharing were established in all metropolitan areas covered by this study (Lugano: 1988; Zurich: 1992; Lausanne: 1992; Lucerne: 1994). Since in Switzerland, competencies for social services lie with the communal level, localities are the main actors in these regional drug service schemes. Yet the Zurich and Lucerne schemes also feature a (subsidiary) involvement of the canton. The various schemes also differ with respect to the involvement of civil-society actors, which are traditionally important players in social policy in Switzerland. In the Zurich scheme, non-governmental organizations participate as equal partners. In the Lugano scheme, non- governmental organizations are regularly associated as field experts. In Lucerne and Lausanne, only state agencies are involved in the regional coordination scheme.

3.2.4 Cultural institutions In Switzerland, cultural institutions are mainly supported by local authorities. During the 1990s, it has been recognized that cultural institutions are of increasing importance in a context of international urban competition, and that public expenditures in this realm should thus be considered as centrality charges. Nowadays, there is a large consensus that suburban localities should Daniel Kübler 269 participate in the financing of urban cultural institutions located in the core city, since they are beneficial for the economic development of the entire metropolitan area. Accordingly, various systems of area-wide financial compensation were set up. In Lausanne suburban localities set up a common pool out of which cultural institutions in the centre-city are co-sponsored on a voluntary basis (in 1988). A voluntary pool was also set up by suburban localities in the Lucerne area (in 1996). In the Zurich area, regional coordination in the cultural realm rests mainly upon a system of financial compensation, set up by cantonal legislation in a two-step process (in 1986 and 1999). In Lugano, there is no area-wide mechanism of coordination to date, basically because of the comfortable financial situation of the core-city in comparison to the surrounding communes.

3.2.5 Synthesis Table 11.4 suggests a classification of these area-wide governance schemes according to the typology of metropolitan governance developed above.13 There is a great variety of governance-types, across problem areas and across agglomerations. More precisely, Table 11.4 leaves us with 8 cases of ‘old regionalism’ (3 cases of ‘local government’ and 5 cases of ‘multi-tiered government’), and 8 cases of ‘new regionalism’ (5 cases of ‘linked functions governance’ and 3 cases of ‘complex network governance’) (see Table 11.4). For the purpose of the quantitative analysis hereafter, information contained in this qualitative classification is synthesized by an additive ‘metropolitan governance-index’ for each metropolitan area under scrutiny. The theoretical range of this index lies between the minimal value of 4 (consolidated local government in all four policy domains) and 16 (complex network governance in all four policy domains). Thus, the higher the index value, the more governance in the according metropolitan area corresponds to new regionalism. Applied to the cases under scrutiny here, the index value is 9 for the Zurich metropolitan area, 10 for Lucerne and Lugano, and 11 for Lausanne. Hence, while new regionalism is present in all the four metropolitan areas, it is least present in Zurich and most in Lausanne.

3.3 The impact of new regionalism upon democratic governance To test the three hypotheses about the impact of new regionalism on core elements of democratic governance, developed in the theoretical section above, we will perform multivariate regression on data collected with a representative telephone survey conducted in the four metropolitan areas under scrutiny.14

3.3.1 New regionalism and metropolitan service users As a first, liberal way of conceptualizing the relationship between citizens and the state, we suggested to focus on the effects of various types of metropolitan governance on service satisfaction. More precisely, drawing on new 270

4) 4) (‘Koordinations-gruppe (‘Koordinations-gruppe = Complex network governance (index value Drugs Forum Drogenhilfe’): Dezentrale involving the canton, the core city as well as governmental and non-governmental suburban service agencies.No public resources.No formal rules of interaction. Public transport (‘Transports publics de la région lausannoise SA’): Joint-stock canton, the by owned company 38 communes (including the core shareholders. private and city), transport of metropolitan Deficit offer covered by the canton (36%)and the communes (64%). Ruledown by statutes based on federalon legislation jointstock companies. New regionalism 3) = (‘Groupe (‘Groupe

(index value (index

Drugs “toxicomanie” de Lausanne Région’): Commission involving the core city suburbs. 26 and Subsidised by member bycommunes. Ruled agreements. negotiated 2) Linked functionsgovernance = (‘Zürcher (‘Zürcher (‘Steuerkraft- und und (‘Steuerkraft- regionalism Public transport Verkehrsverbund’): Public enterprise of involvement with canton the of all 171 communes (including the core city). Deficit of metropolitan transportation offer covered bythe communes the and (50%) canton legislation cantonal by Ruled (50%). on public transport. Culture Lastenausgleich’): Cantonal the involving system compensation canton (punctual subsidies), core the 169 and amount) total of (29% city amount). total of (71% communes Interactions ruled by cantonal legislation. Multi-tiered government (index value (index government Multi-tiered

(‘Services 1) 1) = Types of metropolitan governance found in the four metropolitan areas metropolitan the four found in governance Types of metropolitan Water supply (‘Wasserversorgung Zürich’): Public enterprise of the core suburbs 63 to water sells city; ona cost-effective basis. Submitted to local legislation public on administration. Local government (index (index government Local value Water supply industriels lausannois’): Public enterprise of the core city; sells water and to 51 services maintenance suburbs on cost-effective a basis. Submitted to local legislation public on administration. Zurich

Table 11.4Table Old Lausanne 271

(‘Ewl Wasser (‘Ewl Wasser (‘Fonds intercommunal Water supply Water company Joint-stock AG’): ownedthe by central city; sells waterand maintenance services to 7 communes suburban on a cost-effective basis. Ruled by own statutes based on federal legislation on join-stock companies. Public transport (‘Zweckverband Öffentlicher Luzern’): in Verkehr set body Purpose-oriented and communes 15 by up the canton. Deficit of metropolitan transportation offer coveredthe by canton communes the and (20%) statutes own by Ruled (80%). legislation based on cantonal on public transport and co-operation. intercommunal Culture institutionsde soutien aux culturelles’): Voluntary intercommunal convention involving 25 suburbs. Subsidizedmember by bycommunes. Ruled agreements. negotiated (‘Regionalkonferenz (‘Regionalkonferenz (‘Beitragsfonds für fördernde Culture Kultur’): Foundation supported by and 14 suburbs. the centre-city canton, the Subsidisedby canton (30%), the the centre by Ruled (10%). suburbs the and (60%) city agreements. negotiated Drugs Sozialhilfe’): Contract involving and 74 suburbs. the centre-city canton, the the and (50%), canton the by Subsidized by own Ruled (50%). communes member on legislation cantonal on based statutes intercommunal cooperation. Lucerne

272 4) 4) (‘Commissione (‘Commissione = Complex network governance (index value Drugs intercomunale tossicodipendenza’): Commission involving the core cityand 11 suburbs, as well organizations. as non-government No formal resources. public No rules of interaction. New regionalism 3) (‘Aziende = Water supply Water industriali luganesi SA’): Joint-stock companyowned theby core city; sellswater services maintenance and a on suburbs 17 to Ruled basis. cost-effective by own statutes based on federal legislationjoint on companies. stock (index value (index 2) Linked functionsgovernance = (‘Commissione regionalism Public transport regionale dei trasporti del luganese’): Regional commission on behalfof Involves communes 87 canton. the (including the core city). Deficit coveredthe by canton (50%)andmember communes (50%). Ruledby cantonal legislation publictransport. on Multi-tiered government (index value (index government Multi-tiered 1) 1) (‘Dicastero di di (‘Dicastero = (Continued) Lugano no is There cultura’): supra-local scope of cultural policy in Lugano. Cultural institutionsare subsidized entirely by the core city. Ruled by local legislation on public administration. Local government (index (index government Local value Lugano

Table 11.4Table Old Daniel Kübler 273 regionalist thinking, we hypothesized that there should be no significant relationship between the types of metropolitan governance and service satisfaction.15 Bivariate frequency analysis suggests that satisfaction with the four types of metropolitan services is generally high (data not shown). However, there is significant variation in service satisfaction not only between but also within the four metropolitan areas. On the one hand, the level of satisfaction is higher with respondents living in the two German-speaking metropolitan areas (i.e. Zurich and Lucerne) than with those who live either in Italian (i.e. Lugano) or the French-speaking (i.e. Lausanne) areas. On the other hand, within metropolitan areas, service satisfaction is higher with respondents living in core cities and rich communes than in suburban and periurban communes. For the multivariate analysis, there are several intervening variables whose influence on service satisfaction needs to be controlled for. First, at the individual level, there are socio-demographic variables (age, gender, education). At the contextual level, the result of the bivariate analysis indicates that we need to control for language region (German-speaking versus the rest) as well as type of commune (core-cities versus rich, suburban and peri-urban communes). In addition, given the long-running discussion between metropolitan reformers and public choice theorists on the positive or negative effect of institutional consolidation/fragmentation of metropolitan areas, as well as of the size of governmental units, we need to take these aspects into account, too. Similarly, public choice theorists would argue that the level of income tax also plays a role, since service satisfaction is always a function of a cost–benefit ratio. Thus, the communal tax rate (see Table 11.3) will also be taken into account (see Table 11.5). The results shown in Table 11.5 suggest that, in accordance with what was expected, the type of area-wide governance is not a significant predictor for

Table 11.5 Predictors of service satisfaction (OLS-regression)

Level of variables Variable in the model Standardized Significance (p) coefficients (Beta)

Individuals Age .155 .000 Gender (male) .019 .375 Education .010 .653 Commune Communal tax rate −.076 .003 Size (population) .090 .021 Core city .104 .001 Metropolitan areas German-speaking .203 .000 Consolidation index .118 .016 Governance-index −.092 .205

Dependent variable: ‘Service satisfaction’ (regression factor score). R2 = 14.9%. 274 Democracy Problems of New Regionalism service satisfaction. A further interesting result is the significant positive influence of institutional consolidation and the size of governmental entities: service satisfaction is significantly higher in larger communes and in less fragmented metropolitan areas. Thus, the analysis suggests that, in terms of service satisfaction, the path of institutional consolidation advocated by ’old’ regionalism brings about a better level of performance. In addition, the communal tax rate is shown to be a significant predictor for service satisfaction – a result that supports the validity of the public choice position according to which service satisfaction is not a mere question of service level, but needs to be considered in terms of cost–benefit ratio. Finally, language region and type of commune are also a significant predictors of service satisfaction, thereby confirming the findings of the bivariate analysis.

3.3.2 New regionalism and citizen loyalty towards elected local government In the theoretical section above, we argued that a citizen’s loyalty towards elected local government16 could be taken as a second indicator for the relationship between citizens and the state. Drawing on the new regionalism literature, we hypothesised that new regionalism has a positive effect on citizens’ loyalty towards elected local government. Bivariate descriptives suggest that, overall, citizens’ loyalty towards their elected local authorities is quite high – a result that is consistent with other national surveys on trust in government, where local authorities continuously score best compared with cantonal or national government. However, there are significant variations between and within metropolitan areas. In the two German-speaking metropolitan areas, local authorities are more positively perceived than in the two other ones. Within metropolitan areas, respondents living in the core city have less favourable views of their local representatives than respondents living in other communes (rich, suburban or periurban). Again, for the multivariate analysis, there are a number of intervening variables, both at the individual level and at the contextual level. First, there are socio-demographic variables to be controlled for (age, gender, education). Secondly, information on the involvement of local authorities in area-wide cooperation may not be assumed to be equally distributed, and those better informed are more likely to credit local representatives for their engagement. Thirdly, research in the realm of local political participation suggests that the size of the locality also has an effect on the relationship between citizens and their representatives (see Keating, 1995). Fourthly, the debate on the pros and cons of institutional fragmentation in metropolitan areas suggests that institutional consolidation may also have an impact upon this relationship. Fifthly, the results of the bivariate analysis again suggests that language region (German-speaking versus French- or Italian-speaking) as well as the type of commune (core cities versus rich, suburban and periurban communes) Daniel Kübler 275

Table 11.6 Predictors of loyalty towards local authorities (OLS-regression)

Level of variables Variable in the model Standardized Significance (p) coefficients (Beta)

Individuals Age .068 .002 Gender (male) .008 .691 Education −.053 .015 Informed on area-wide .130 .000 cooperation Communes Size (population) −.088 .012 Core city −.114 .000 Metropolitan areas German-speaking .254 .000 Consolidation index .053 .287 Governance-index −.068 .342

Dependent variable: ‘Loyalty towards local authorities’ (regression factor score). R2 = 12.9%. may also be significant predictors for citizen loyalty towards their local representatives (see Table 11.6). The results shown in Table 11.6 do not support this hypothesis: the type of metropolitan governance is not significantly associated with citizens’ loyalty towards their elected local authorities. There is thus little empirical support for the assumption that new regionalism has a positive influence upon the relationship between citizens and their local representatives in a positive way. Similarly, the results do not yield much support for the metropolitan reform position, according to which institutional consolidation influences this relationship. However, there is some evidence for the public choice argument according to which loyalty towards local representatives tends to be higher in small communes. Results on the other intervening variables suggest that the strongest effects on citizen’s loyalty towards their local authorities are exerted by the cultural-linguistic context (German speakers are more loyal to their local representatives), by the level of information on area-wide cooperation, and by the type of commune (residents of core cities are less loyal). Age and education are also intervening variables with a significant – though not very strong – effect on this relationship.

3.3.3 New regionalism and legitimacy Focusing on legitimacy as a third indicator for the relationship between citizens and the state, we hypothesized that new regionalism reduces democratic legitimacy – operationalized by satisfaction with the functioning of democracy at the local level. Bivariate descriptives again show that satisfaction with local democracy is generally high: a majority of respondents declare themselves to be highly or even very highly satisfied with the functioning of local democracy. Across 276 Democracy Problems of New Regionalism metropolitan areas, levels of satisfaction with local democracy is significantly higher in the two German-speaking metropolitan areas than in the other ones – a result consistent with national survey data showing that satisfaction with democracy tends to be lower in the French- and especially in the Italian-speaking parts of the country. Within metropolitan areas, satisfaction with local democracy tends to be lower in the core cities than in the surrounding rich, suburban or peri-urban communes. Again, a number of intervening variables are to be controlled for in the multivariate analysis. Socio-demographics (age, gender, education) are the usual suspects. Moreover, democracy theorists have pointed out that there are two distinctive components of legitimacy (see Dahl, 1994; Scharpf, 1999), namely output-legitimacy related to the effectiveness of state policies, and input-legitimacy related to the aggregation of citizen demands. In order to single out the importance of these two factors in overall legitimacy, we will include service satisfaction (as a proxy for output legitimacy) and loyalty towards local representatives (as a proxy for input legitimacy) as intervening variables. At the contextual level, the bivariate results suggest that language region (German-speaking versus Italian- and French-speaking metropolitan areas) and type of commune (core cities versus the rest) must be taken into account as intervening variables. Lastly, the debate between metropolitan reformers and public choice theorists also suggests that institutional consolidation or fragmentation, population size of the communes, as well as the communal tax rate could have a significant effect on perceived legitimacy of local government (Table 11.7). Results shown in Table 11.7 basically call for two types of comments. First, regarding the hypothesis to be tested here, new regionalism is a significant

Table 11.7 Predictors of satisfaction with local democracy (OLS-regression)

Level of variables Variable in the model Standardized Significance (p) coefficients (Beta)

Individuals Age −.005 .802 Gender (male) .044 .028 Education .024 .244 Service satisfaction .140 .000 Loyalty towards local .425 .000 representatives Communes Comunal tax rate −.042 .084 Communal population −.039 .294 Core city .045 .121 Metropolitan areas German-speaking −.207 .000 Consolidation index .237 .000 Governance-index −.229 .001

Dependent variable: ‘Satisfaction with the functioning of local democracy’. R2 = 25.2%. Daniel Kübler 277 predictor of satisfaction with local democracy, and the association is negative, i.e. satisfaction with local democracy is lower when there is new regionalism. Hence, the empirical data support the hypothesis put forward above: there is evidence that new regionalism does diminish democratic legitimacy. Secondly, however, in relation to the intervening variables controlled for in the above analysis new regionalism is not the strongest predictor for satisfaction with local democracy: institutional consolidation is at least as important as new regionalism. And there are individual-level variables which are stronger predictors than contextual ones. In this respect, it is particularly interesting to note that, for satisfaction with local democracy, the perception of local authorities is much more important than service satisfaction. In other words, the input dimension (politics) contributes much more to overall legitimacy than the output dimension (policies) – an interesting result in itself.

4Conclusion

In spite of the high institutional fragmentation of metropolitan areas, a broad range of area-wide governance mechanisms has emerged in policy fields that are today considered to be some of the most urgent metropolitan problems in Switzerland. Some of them are close to old regionalism, some to new regionalism, independently of the period in which they were founded. These results therefore challenge the contention that ‘new’ regionalism is something new and old regionalism is obsolete. However, as a concept to distinguish various routes towards area-wide governance, this distinction proved operational and useful for a systematic classification of mechanisms found in the four areas under scrutiny. Concerning the impacts of types of area-wide governance upon the relationship between the citizens and the state, the above analysis suggests that such impacts do exist. First, our data suggest that although network governance advocated by new regionalists does not influence satisfaction of service users, institutional consolidation and large governmental units advocated by ‘old’ regionalism still produces higher levels of service satisfaction. Secondly, we did not find any evidence that ‘new regionalism’ would improve the relationship between citizens and their local representatives. Thirdly, there is some evidence that new regionalism does threaten the legitimacy of the local political system, but this effect is rather weak. Taken together, our results suggest that the issues related to democratic politics in metropolitan areas depend upon a mixture of explanatory factors, where the chosen route towards area-wide governance is only one among various others. Among those other factors, there are ‘hard’ ones, such as the size of a commune, its location within the centre–periphery pattern of a metropolitan area (i.e. communal types), or the degree of institutional fragmentation/consolidation of a metropolitan area, which play 278 Democracy Problems of New Regionalism a much stronger role. Similarly, aspects pertaining to political culture, such as language regions (Kriesi, 1996), as well as the traditionally strong input- orientation of the Swiss local government system (Hesse and Sharpe, 1991) have stronger effects on the relation between the citizens and the state than the type of regionalism (‘old’ or ‘new’) found in particular metropolitan areas. Last but not least, our analysis has several policy implications for metropolitan governance in Switzerland. More particularly, it shows that the system of ‘old’ regionalism performs better in terms of service satisfaction and democratic legitimacy. This suggests that current efforts to change the institutional setting in Swiss metropolitan areas should pursue the ‘old’ routes towards area-wide governance – that is, institutional consolidation – rather than the ‘new regionalist’ ones.

Notes

1. Research for this article has been supported by the Swiss National Science Foundation through the ‘Demain la Suisse’-project Gouvernance métropolitaine et légitimité conducted at the Laboratoire de sociologie urbaine of the Swiss Federal Institute of Technology Lausanne between April 2000 and April 2003 (Grant number 5004-058522), as well as through a post-doctoral fellowship granted to the author between June 2003 and March 2004 (Grant number 8210-67632). The author thanks Hanspeter Kriesi for thorough and constructive remarks. 2. The Swiss Statistical Office speaks of ‘agglomerations’ (Agglomerationen/agglomérations/ agglomerati) to describe these functionally integrated urban spaces (Schuler 1994). For the purpose of this article, we will however use the term of metropolitan areas as a synonym. 3. For a portrayal of the arguments, as well as a review of the relevant work produced by scholars pertaining to the various schools of thought, see Ostrom (1972), Dente (1990), Keating (1995), Lefèvre (1998), Lowery (1999), Kübler (2003), Kübler and Heinelt (2004). 4. This impression is confirmed by empirical research in Europe (see Van den Berg et al., 1993; Bagnasco and Le Galès, 1997; Benz, 2001; Kübler and Wälti, 2001) and North America (see Downs, 1994; Rusk, 1995). 5. On this parallel between New regionalism and the government-governance distinction, see especially Lefèvre (1998), Le Galès (1998), Savitch and Vogel (2000), Benz (2001). 6. It is not the place here to discuss this debate. For a thorough overview, cf. Leresche (2001). 7. The following synthesis draws on Le Galès (1995), Rhodes (1996) and Stoker (1998, 2000). 8. This synthetic presentation of the metropolization process in Switzerland draws on Rossi (1995) Schuler (1999), Cunha and Schuler (2001). 9. The approval of amalgamation by the concerned communes is required in most cantons. Only in the cantons of Thurgau and Ticino can communes be compelled to disappear without their consent. But even in those two cantons, it is current Daniel Kübler 279

practice to hold a referendum in the communes concerned by amalgamation. So far, forced amalgamation happened only once. 10. The social segregation within metropolitan areas can been illustrated by using the Swiss Statistical Office’s typology of communes (Joye et al. 1988) permitting classification of communes within metropolitan areas into four types: core cities, suburban communes, periurban communes as well as rich communes (where proportions of wealthy inhabitants are particularly high). 11. Communes in Switzerland are fiscally autonomous. They raise communal taxes on income and property and they are competent to fix the communal tax rate. In average, 70% of communal income stems from taxes and fees. Only 18 per cent of communal income stem from transfer payments made by higher state levels – the lowest proportion found among all Western European countries (Steiner, 2002). 12. See the methodological appendix for the data base of these case studies. The rationale behind the selection of the four policy domains was to cover territorial network services (public transportation, water supply) as well as centrally located services (social services for drug users, cultural institutions). In addition, it covers highly politicized domains (public transportation, drug policy) and lowly politicised domains (water supply, cultural institutions). With respect to citizen perception of metropolitan policy and politics, previous research (Kübler and Wälti, 2001b) suggested that both of these factors (network versus centrality infrastructure, high versus low politicisation) were important contextual variables to be controlled for. 13. For a further development of this classification, see Kübler and Schwab (2004). 14. See the methodological appendix. 15. Service satisfaction was operationalized by four items in the questionnaire, asking about satisfaction with the quality of management of water supply, public transport, drug services and cultural institutions. Responses on these four items were transformed into a single indicator using factor analysis. 16. Respondents’ loyalty towards elected local government was measured by four items in the questionnaire, gauging the perception of responsiveness, management quality, trustworthiness and communication efforts. Factor analysis was then used to transform responses on these four items into one loyalty indicator.

Methodological appendix

Case-study material The case-study material included documents, archival records and analysis of secondary data, as well as data from 66 in-depth interviews, conducted between 1997 and 2002. In each metropolitan area, representatives of the main agencies and institutions involved in area-wide coordination schemes were interviewed, as well as elected officials and council politicians. Interview topics included: the coming about of the coordination scheme, decision making within the scheme, relationships among the participating actors, as well as the relevance of representative institutions and elected politicians with respect to the functioning of the coordination scheme. 280 Democracy Problems of New Regionalism

Table A11.1 List of interviews (qualitative studies)

Coordination schemes Zurich Lausanne Lugano Lucerne Total

General coordination 1 — 2 — 3 Water supply 4 1 3 1 9 Public transport 5 3 6 — 14 Services for drug users 12 8 6 2 28 Cultural institutions 4 5 3 — 12 Total 26 17 20 3 66

Telephone survey The telephone survey included 2,010 Swiss citizens between 18 and 74 years of age in the four metropolitan areas under scrutiny. The four subsamples are representative on a random-quota basis. The interviews were carried out by MIS Trend in Lausanne in April and May 2001. Interviews were conducted with CATI according to a standardized questionnaire (the three language versions of the questionnaire are published in Kübler et al. (2002)); mean duration of interviews was 27 minutes.

Table A11.2 Interviews telephone survey

Type of commune Zurich Lausanne Lugano Lucerne Total

Core city (%) 210 (41.9) 220 (43.1) 177 (35.4) 210 (42.2) 817 (40.6) Suburban 210 (41.9) 206 (40.3) 167 (33.4) 258 (51.8) 841 (41.8) communes (%) Rich communes (%) 49 (9.8) 45 (8.8) 30 (6.0) 12 (2.4) 136 (6.8) Peri-urban 32 (6.4) 40 (7.8) 126 (25.2) 18 (3.6) 216 (10.7) communes (%) Total (%) 501 (100) 511 (100) 500 (100) 498 (100) 2,010 (100) 12 Sustainable Fiscal Policy in a Federal System: Switzerland as an Example Lars P. Feld and Gebhard Kirchgässner

How a sustainable fiscal policy can be performed in a federal system is not only a Swiss problem but is also discussed in other federal countries like Germany and Austria, and in the European Union. Contrary to most other countries, the Swiss fiscal system is characterized by an extensive fiscal federalism with high fiscal autonomy at all governmental levels, by direct popular rights which include fiscal referenda at the cantonal and local levels, and by particular constitutional and/or statutory fiscal restraints in order to prevent excessive public debt. In this chapter the effects of these constitutional clauses on public finances are investigated. Using a panel of the 26 Swiss cantons from 1980 to 1998, we provide evidence that direct democracy leads to significantly lower expenditure and revenue. The fiscal constraint, on the other hand, significantly reduces budget deficits. Total, cantonal as well as local expenditure and revenue are the lower the higher the share of local expenditure is.

1 Introduction

Current policy debates on public finances across OECD countries focus on the question of how a sustainable fiscal policy can be obtained. The most pertinent discussion takes place in the European Union (EU) where the Stability and Growth Pact (SGP) as a follow-up to the fiscal convergence criteria of the Maastricht Treaty requires EU member states to keep budget deficits below 3 per cent of GDP and public debt below 60 per cent of GDP. Deviations from this general rule are only allowed for under specific circumstances such as severe economic downturns or extraordinary events like natural disasters.1 The governments of member states are held responsible to stick to the SGP which poses additional problems in those that are organized as federal states. In Austria, Germany and Spain, regional authorities have a certain fiscal autonomy such that deficits at the sub-federal level may go to the expense of the federal level. In Germany, this situation has led to a national stability pact. The discussion in the EU is not unique. In the United States, the Gramm–Rudman–Hollings (GRH) Act was enacted in 1985 in order to reduce

281 282 Sustainable Fiscal Policy in a Federal System federal public debt (Gramlich, 1990). Several US states have formal fiscal restraints with characteristics that strongly vary across the states. In 2001, Switzerland also introduced a debt brake at the federal level in order to keep the federal government from incurring excessive public debt. Contrary to nearly all other OECD countries, however, the Swiss and the US fiscal system have two special features: fiscal federalism, which is organized in a competitive way and gives the lower-level jurisdictions (cantons/states and local communities) the power to tax, and direct popular rights in political decision making which include fiscal referenda at the sub-federal levels. As in the US states, there are special constitutional and/or statutory restrictions in some Swiss cantons in order to induce sound finances. Moreover, large differences with respect to the institutional design between the cantons persist that even exceed those between US states, thus making Switzerland a unique laboratory where the effects of fiscal institutions can be studied. Since the 1990s, a number of empirical studies have considered the variety of different institutional designs of fiscal systems in the US and Switzerland. Bohn and Inman (1996), for example, extensively study the impact of balanced budget requirements on public finances of US states by investigating which specific design is most successful in restraining governments.2 O’Sullivan, Sexton and Sheffrin (1995, 1999) analyse how Proposition 13 has affected fiscal policies of US states and local jurisdictions.3 The effectiveness of fiscal restraints has also been investigated in the Swiss case. Using a panel of the 26 Swiss cantons and the years 1986 to 1997, Feld and Kirchgässner (2001a) show that cantons with such restrictions have significantly lower debts and deficits. Similar results are obtained by Schaltegger (2002) for a different time period. The cross-country and the US results on the impact of fiscal federalism on public finance are more ambiguous (Kirchgässner, 2002a, 2002b). According to the theoretical arguments by Brennan and Buchanan (1980), fiscal decentralization reduces the ability of governments to exploit tax bases because the latter have increased exit possibilities in a federal system. While several authors find evidence for this proposed effect of fiscal decentralization on the size of government, others do not. In a recent paper, Feld, Kirchgässner and Schaltegger (2003) present evidence for the Swiss cantons from 1980 to 1998 that fiscal decentralization decreases government revenue mainly because of an intense tax competition. Finally, there is a large body of evidence on the impact of referenda and initiatives on public finance. Matsusaka (2002) provides a comprehensive discussion about the impact of legislative initiatives on spending and revenue of US states and local jurisdictions, while Kiewiet and Szakaly (1996) present evidence on the influence of referenda on guaranteed debt of US states. These US studies show that direct democracy is associated with sounder public finances. The studies on Swiss cantons and local jurisdictions corroborate this conclusion (Feld and Kirchgässner, 1999, 2001a, 2001b; Feld and Matsusaka, 2003). Fiscal referenda induce lower spending and revenue. Lars P. Feld and Gebhard Kirchgässner 283

In this chapter, the effects of these three types of constitutional or statutory clauses on public finances of the 26 Swiss cantons are investigated again by looking at four important indicators of public finances: public spending, public revenue, tax revenue, and budget deficits for the period 1980 to 1998. We focus on the impact of all three classes of constitutional or statutory provisions in the same econometric model. This is done by estimating models in logarithmic and in absolute terms in order to assess the robustness of the estimates to the functional form of the underlying model. The results form the basis of an evaluation whether such restraints are desirable. The analysis deviates in several ways from the existing studies. Either they investigate a shorter time period, for example 1986 to 1998 (Feld and Kirchgässner, 2001a), or they only look at one indicator of public finance if they use data for the longer period 1980 to 1998 (e.g. Schaltegger, 2001 or Feld and Matsusaka, 2003 on spending; Feld, Kirchgässner and Schaltegger, 2003 on revenue), or they do not consider all three types of constitutional or statutory clauses.4 Our main results are that direct democracy leads to significantly lower expenditure, revenue, and tax revenue, but does not have any significant effect on the budget deficit. The fiscal constraint, on the other hand, significantly reduces budget deficits. Cantonal expenditure, revenue, and tax revenue is the lower the higher the share of local expenditure. Taking all results together, one may conclude that at least those cantons with ‘strong’ fiscal instruments have the institutional prerequisites to perform a sustainable fiscal policy. They can serve as examples for developing corresponding fiscal rules for other federal countries. The chapter is organized as follows: In section 2, we present the basic econometric model. In section 3, we discuss the empirical results. An evaluation of these results and some conclusions are provided in section 4.

2 An econometric model for fiscal policy

In order to consider fiscal policy as comprehensively as possible, we consider public expenditure (per capita), public revenue, tax revenue, and the budget deficit as dependent variables.5 The explanatory variables of interest are the institutional variables which represent the constitutional and legal structure of fiscal policy decisions in the Swiss cantons. The first and most important variable is the index of direct democracy as it is employed by Frey and Stutzer (2000, 2002) in various studies. Most cantons have some form of semi-direct democracy with a parliamentary system with legislators elected according to a system of proportional party representation. Today, only two rural cantons, Appenzell-Innerrhoden (AI) and Glarus (GL), still take political decisions in canton meetings (Landsgemeinde). In addition, the cantons have different institutions of political participation rights (Trechsel and Serdült, 1999; Feld and Matsusaka, 2003). Proposals can be initiated by the voter initiative, and new laws passed by the legislature are, to different degrees, subject to an optional or a mandatory popular referendum. Given the results by Peltzman 284 Sustainable Fiscal Policy in a Federal System

(1992) that voters are fiscally more conservative than representatives, we can expect that fiscal referenda restrict the spending capabilities of representatives. It should lead to lower spending, revenue, and deficits. As a second variable, an index of statutory fiscal restraints is included in the model. It ranges from zero in cantons without statutory fiscal restraints to three in the cantons with the most stringent restraints. While nearly all cantons have constitutional fiscal restraints that demand them to balance their budgets over time in one way or the other, during the time period considered in this chapter only five cantons, Appenzell Ausserrhoden, Fribourg, Graubünden, St Gallen and Solothurn, have statutory fiscal requirements.6 Those require the cantons to increase their tax rates if budget deficits increase above a deficit threshold. In Fribourg this requirement is specified such that local taxes are not covered by it. In St Gallen and Solothurn, there is an additional restriction on decreases of the tax rates in order to restrict deficit financing. The requirements are less restrictive in Appenzell and even lesser in Graubünden. The more restrictive the statutory fiscal constraints are, the sounder public finances should be, i.e. the lower public spending, and budget deficits should be. Since these requirements have a strong revenue bias forcing the cantons to avoid deficits by higher tax rates, the sign of this variable on public revenue is not determined a priori: It might lead to higher revenue. But if this is the case, the size of the coefficient in the revenue equation should be smaller than the one in the spending equation. The third institutional peculiarity of Switzerland is its strong extent of fiscal autonomy at the sub-federal level which establishes a system of competitive federalism. To analyse the impact of federalism on cantonal fiscal policy, two different variables are used: decentralization and tax competition. Decentralization is proxied by the ratio of local revenue to the aggregated state and local revenue. Tax competition is measured by the inverse of the weighted average of the competing cantons’ tax burden in the highest income tax bracket of a million Swiss francs annual taxable income. The competing cantons are all cantons except the one of consideration, weighted by the inverse of the distance (see Feld and Reulier, 2002, for a discussion of empirical studies). This variable indicates that the lower the average tax burden of the other cantons, the higher the pressure of tax competition on the cantonal and local tax authorities and the lower is tax revenue, total revenue and spending. A negative impact on government revenue and spending is thus expected for both fiscal federalism variables. The impact of tax competition on budget deficits is, however, indeterminate. We additionally include economic, demographic and political control variables. The economic and demographic variables are those usually employed in models of fiscal policy. The most important of these variables is the disposable income per capita. Generally, higher income is supposed to lead to higher spending and revenue. Higher spending results because citizens increase their demand for public services if their income increases. Higher revenue results Lars P. Feld and Gebhard Kirchgässner 285 because the revenue of the Swiss cantons is mainly derived from progressive personal income taxes. Whether lower or higher deficits occur due to higher income is not easy to determine a priori. On the one hand, higher income may be accompanied by a lower level of public deficits for liquidity reasons. On the other hand, sub-federal jurisdictions with higher incomes may have to contribute larger amounts to fiscal equalization systems and thus have an incentive to increase public deficits in order to reduce these contributions. In this case, higher income might be associated with higher deficits. In order to control for the impact of intergovernmental grants between jurisdictions, the model also contains federal unconditional grants per capita. In contrast to matching grants, unconditional grants enable cantons to allocate the funds according to their own priorities. A higher level of unconditional grants should lead to higher spending as well. In the literature on the flypaper effect, it is much discussed whether the availability of lump-sum grants increases public spending by more than the amount of these grants (Feld and Schaltegger, 2004). Unconditional grants may also be used to reduce spending from own public funds such that the increase of spending due to the grants is less than 100 per cent. In addition, a high amount of grants is related to a higher extent of bail-out by other jurisdictions. This might lead to lower incentives to use the resources economically. Therefore, it may – ceteris paribus – incur higher budget deficits. We also include a regional dummy variable that reflects the language differences among the Swiss cantons and takes on the value of one for cantons with a French- or Italian-speaking population. A quite common prejudice is that ‘Latin’ cantons and communities have stronger preferences for ‘public sector solutions’ of social problems and are thus inclined to have more ‘unsound’ public finances, i.e. higher spending, higher revenue, and higher deficits. Moreover, the model contains a political variable which follows the arguments of the partisan cycles models that left-wing parties generate unsound public finances. The share of left-wing parties in the government should have a positive impact on the level of public spending, public revenue, and budget deficits. Since the number of inhabitants can play a crucial role on the level of public expenditure, a population variable has to be included in the equation as well. However, the expected sign of this variable is ambiguous. On the one hand, more inhabitants will pay for public goods. This reduces cost per capita, and it should lead to lower public expenditure. On the other hand, because of economies of scale in their provision some public goods might only be provided in agglomerations. In this case, the overall level of public expenditure for the agglomeration might increase and – ceteris paribus – budget deficits might also rise. In order to disentangle both effects, the share of the urban population is additionally included in the model. Moreover, we control for the demographic structure of a canton by using the shares of the population older than 65 and younger than 20 years. Both variables may be interpreted as indicating the demand of these two 286 Sustainable Fiscal Policy in a Federal System particular population groups for public spending as well as their ability to generate public revenue. Finally, a dummy variable for the canton of Appenzell a.Rh. in 1996 is included. In this year, the cantonal revenue of that canton lies about 50 per cent above the ‘normal’ value, because this canton sold its ‘own’ cantonal bank to the Union Bank of Switzerland (UBS) which created a large additional revenue.7 Thus, for our empirical analysis we end up with the following model:

=α +α +α +α +α +α (1) y 0 1 Dem 2 Constr 3 Fed 4 Taxcomp 5 grants +α +α +α +α +α +α 6Ideol 7 inc 8 pop 9 Urban 10 Latin 11 Old +α +α +ε 12 Young 13 D(AR-96) . where the dependent variable X, stands for the following four fiscal variables (all in 1000 CHF per capita),

exp public expenditure, rev public revenue, tax tax revenue, deficit budget deficits.

The explanatory variables are:

Dem index of direct democracy, Constr statutory fiscal constraints which takes on values between zero for the cantons with no and three for those with the strongest statutory fiscal restraint, Fed share of local from the sum of cantonal and local spending, Taxcomp inverse of the weighted average of the competing cantons’ tax burden in the highest income tax bracket of 1 million SFr annual taxable income (logarithms), grants federal unconditional grants per capita (logarithms), Ideol ideological position of the cantonal government, inc disposable income per capita (logarithms), pop population (logarithms), Urban share of urban population, Latin dummy variable=1, for cantons with a French- or Italian-speaking population, Young share of population younger than 20, Old share of population older than 65, DAR96 dummy variable =1, for the canton of Appenzell Ausserrhoden in 1996, ε stochastic term.

The analysis uses annual data for the 26 cantons from 1980 to 1998, deflated to the year 1980. The empirical analysis is performed using a pooled Lars P. Feld and Gebhard Kirchgässner 287 cross-section time-series model. We follow Feld and Kirchgässner (2001a), who argue that despite the panel structure of the data the inclusion of fixed effects in the cross-section domain is inappropriate because the institutional variables vary only very little or remain constant over time in some cantons. Accordingly, cantonal intercepts do not make sense as the captured impact on fiscal outcomes is either solely driven by the time variation or in case of time-invariant variables, fixed effects are likely to hide the effect of institutional variables and render them insignificant. Cantonal dummies are, however, used as instruments in order to cope with possible endogeneity of the decentralization variable. Moreover, year dummies to circumvent time dependency are included, and the standard errors are corrected by a GMM method (Newey–West). With the exception of the deficit equation, the model is estimated in logs.

3 Empirical results

For total cantonal and local expenditure we obtain the following results:8

(2) exp = 0.829 − 0.039 Dem − 0.009 Constr− 0.504 Fed− 0.108 taxcomp (0.91) (−1.96) (−0.61) (−3.08) (−1.82) + 0.113 grants + 0.132 Ideol + 0.199 inc +0.002 pop +0.161 Urban (2.25) (1.49) (1.78) (0.11) (1.74) −0.009 Latin + 0.020 Old − 0.005 Young+ 0.089 D(AR-96)+ε (−0.15) (2.42) (−0.52) (2.43) R2 = 0.736, SER = 0.118, J.–B = 15.026.

The impact of the index for direct democracy is significant at the 10 per cent level, the impact of the decentralization variable even at the 1 per cent level; both lead to a reduction of public expenditure. Tax competition also reduces public expenditure, while unconditional grants increase them. The variable for the constraints also has a negative coefficient, but its impact is far from being significant. A left-wing orientation of the cantonal government increases (as expected) public expenditure, but this effect is not significant. Contrary to what is usually assumed, expenditures are not larger in those cantons where French or Italian is the dominant language. It might be objected against these results that, according to the result of the Jarque–Bera test, the estimated residuals are not at all normally distributed. This might impair the validity of the results. Due to the large number of observations this should not be a great problem. However, to consider this argument we re- estimated the model excluding three outliers. Then, the null hypothesis that the estimated residuals are normally distributed can no longer be rejected even at the 5 per cent significance level, while the results show only minor changes.9 The important question is, however, whether these impacts are not only statistically significant but also economically important. To address the economic significance, we calculated how large the difference in cantonal and 288 Sustainable Fiscal Policy in a Federal System local expenditure is between those cantons where the corresponding variables take on their maximum value in our sample and those cantons where they have their minimum. To calculate these partial effects we assumed that all other variables take on their mean, and we used the price level of the year 2000.10 The results are given in Table 12.1. It is shown that fiscal decentralization, direct democracy and tax competition have strong dampening effects, while unconditional grants lead – ceteris paribus – to considerably higher expenditure. Compared to this, the ideology of the government as well as the fiscal constraints have negligible effects. For the total revenue of the cantons and the local communities we got the following results:

(3) rev = 0.860 −0.035 Dem + 0.001 Constr − 0.537 Fed − 0.129 taxcomp (1.05) (−2.02) (0.04) (−3.44) (−2.35) + 0.153 grants + 0.113 Ideol + 0.171 inc+0.003 pop+ 0.114 Urban (3.21) (1.33) (1.67) (0.14) (1.37) −0.053 Latin + 0.014 Old − 0.011 Young + 0.349 D(AR-96) +ε (−0.95) (1.96) (−1.24) (11.05)

R2 = 0.744, SER = 0.108, J.−B = 34.040. With respect to both their significance and their quantitative impact, which is again demonstrated in Table A12.2, these results are close to those for total public expenditure. One major difference is that the variable for direct democracy is now significant at the 5 per cent level. However, if we eliminate some observations to ensure the normality of the estimated residuals, the coefficient of this variable is again only significantly different from zero at the 10 per cent level.11 Another difference is that fiscal constraints now have a positive impact, whereas their impact on expenditure was negative. Thus, both coefficients have the expected signs, but they are far from being statistically significant. Finally, the language variable has now a much larger effect, but it is also far from being significant. The results deviate more strongly from those of the expenditure equation if we only consider tax revenue:

(4) tax =−1.151 − 0.038 Dem+ 0.011 Constr − 0.190 Fed− 0.200 taxcomp (2.06) (−2.76) (1.47) (−2.17) (−5.84) − 0.068 grants + 0.027 Ideol + 0.364 inc − 0.022 pop + 0.388 Urban (−1.88) (0.61) (4.57) (−1.66) (6.23) −0.036 Latin − 0.006 Old − 0.030 Young + 0.145 D(AR-96) +ε (−1.02) (−1.15) (−5.53) (5.13) 2 R = 0.914, SER = 0.078, J.–B = 5.428.

The impact for fiscal federalism is still significant but, as the figures in Table A12 show, comparatively small. By contrast, the impact of direct 289 Deficit Deficit 22 CHF (0.21)22 CHF 95 CHF (0.89)95 CHF 424(4.01) CHF 329 CHF (3.11) le. In the case of case In the le. − − 237 CHF − − Taxes Taxes 30 CHF (0.55) − 478 CHF (8.61) 198 CHF (3.56) 529 CHF (5.00) 327 CHF (5.90) 8 CHF (0.07) − − − 860 CHF (15.49) 5,552 CHF − 1,452(26.16) CHF 198 CHF (1.87) − 552 CHF (5.34) 234 CHF (2.26) Public revenue Public − − 1,744 CHF (16.86) 1,468 CHF (14.19) 1,723 CHF (16.65) 10,345 CHF − − − 99 CHF (0.94) − 296 CHF (2.80) 19 CHF (0.18) 189 CHF (3.40) 279 CHF (2.63) − − 1,642 CHF (15.51) 1,502 CHF (14.19) 1,505 CHF (14.22) 2,002 CHF (19.35) 1,652 CHF (15.61) 10,581 CHF − − − Public expenditure Public Quantitative impacts of the explanatory variables (per capita) capita) (per variables explanatory of the impacts Quantitative : For public expenditure, revenue, and taxes, the numbers in parentheses are the percentage of the mean of the dependent variab dependent the of mean the of percentage the are in parentheses numbers the taxes, and revenue, expenditure, : For public French- and Italian-speaking population Italian-speaking and French- Unconditional grants government of the Ideology Mean of dependent the variable Fiscal constraints constraints Fiscal Table 12.1Table variables Explanatory Note Fiscal decentralization Fiscal Direct democracy Tax competition budgetdeficits,percentitin expenditure. is of amount The Francs in Swiss2000 in is prices. 290 Sustainable Fiscal Policy in a Federal System democracy is quantitatively much more important and strongly significant. As expected, tax competition has a strong dampening and effect on tax revenue. In addition, unconditional grants also have a negative effect. This speaks against fiscal illusion, but is an interesting contrast to the result for total expenditure (and revenues) which suggested the existence of such an illusion. Fiscal constraints have a positive but again insignificant impact on tax revenue. Again, neither the ideology of the government nor the language region have a sizeable impact. Finally, we consider the deficit equation. As there are positive and negative values of this variable, we cannot apply the logarithmic transformation and, therefore, use a linear specification. The results are as follows:

(5) deficit =−2.961 − 0.023 Dem − 0.110 Constr − 0.025 Fed + 0.726 taxcomp (−2.53) (−0.32) (−2.96) (−0.06) (1.41) − 0.001 grants + 0.111 Ideol + 0.014 inc+ 0.000 pop (−2.72) (0.56) (2.02) (0.63) + 0.591 Urban+ 0.529 Latin + 0.063 Old + 0.080 Young (2.25) (2.75) (3.59) (3.01) – 3.038 D(AR-96) +ε (−22.33)

2 R = 0.434, SER = 0.434, J.–B = 82.908.

As the figures in Table 12.1 show, these results are consistent with those of the expenditure and revenue equations. While direct democracy has significant dampening effects on both public expenditure and revenue, it has no significant impact on the deficit. The same holds for fiscal decentralization. On the other hand, fiscal constraints have – as they should – a significant negative effect on the deficit, while the dampening effect on the expenditure as well as the increasing effect on the revenue side are not statistically significant. Unconditional grants had positive effects on both, the expenditure as well as the revenue side. However, because the expenditure effect was smaller, the deficit is also significantly reduced. Tax competition has a positive impact on the deficit: the stronger and the more difficult it is, correspondingly, to raise taxes, the more people are willing to accept deficits. This was already the picture of the expenditure and revenue equations: the dampening effect on the expenditure side was much smaller, which, necessarily, results in a larger deficit. However, the effect on the deficit is neither statistically significant nor especially strong. Once again, the ideology of the government plays no role at all, while the language variable now has a significant positive impact: the French- and Italian-speaking population seems to be accepting a deficit in the public budget more willingly than the Lars P. Feld and Gebhard Kirchgässner 291

German-speaking Swiss, and the impact is not negligible: It is about twice the average deficit.

4 Putting the results into perspective

We have investigated the effects of fiscal restraints, competitive fiscal federalism and direct democracy on public spending, public revenue, tax revenue and budget deficits of the 26 Swiss cantons for the period 1980 to 1998, and the results are unambiguous. If we want to achieve fiscal sustainability, the message of these results is clear: Fiscal federalism as well as direct democracy can effectively restrain government activity. Of these two factors, the effect of fiscal federalism is statistically much more significant and also more robust with respect to changes in the specification of the equations. It is, moreover, backed up by the impact of tax competition which can only come into effect if there is fiscal decentralization not only on the expenditure but also on the revenue side of the public budget. However, neither fiscal federalism nor direct democracy seems to be able to prevent public deficits. In order to achieve this goal, additional institutional constraints are necessary, and the ‘debt breaks’ which have been introduced in several cantons seem to be well suited for this purpose. They are, however, unable to prevent public expenditure and revenue from increasing. Thus, to achieve fiscal sustainability we need all three institutions: fiscal federalism, direct democracy as well as institutional constraints. Concluding that these institutions actually serve as restrictions to fiscal policy does not necessarily imply, however, that they also serve the purposes of citizens. The effects of these fiscal instruments on the quality of the services which are provided are also interesting. While the pressure exerted by these fiscal instruments might increase the efficiency of the public bureaucracy, the quality of the services, might deteriorate. In the US, where most of the studies have been conducted, this is especially relevant in the case of educational expenditures. The empirical results with respect to the influence of formal fiscal restraints are, however, mixed and thus support the arguments by Besley and Smart (2003). In a first paper, Downes (1992) concludes that the performance of Californian students has improved between 1976 and 1985, i.e. comparing the situations before and after the passing of Proposition 13, which started the introduction of the most important fiscal constraints. In a second study, Downes, Dye and McGuire (1998) also find no significant reduction in students’ performance for about 1,500 students in Illinois, except in mathematics. This is contrary to the result of Figlio (1997) in a study of 5,600 students from 49 states for the periods 1987/88 and 1990/91. He finds evidence that students in states with tax or expenditure limitations had – ceteris paribus – a worse performance in several areas, including in core subjects like sciences and mathematics. 292 Sustainable Fiscal Policy in a Federal System

These results have been corroborated by Downes and Figlio (1997, 1999). Comparing the performance of students from the years 1972 (8,672 observations) and 1992 (6,054 observations), it is shown that students in states with tax or expenditure limitations had significantly worse results in mathematics, but not in English. The reason for this deterioration might be, as Figlio (1998) shows in a study for 305 schooling districts in Oregon and 296 schooling districts in Washington, that the interaction between teachers and students decreased after the introduction of the limitations and that the starting salaries of teachers have been reduced which – according to Figlio and Rueben (2001) – had the consequence that highly qualified teachers did not want to teach in districts with fiscal limitations. It is, moreover, highly questionable whether the production efficiency of the publicly provided services really increased. Figlio and O’Sullivan (1997) show for 5,150 US local communities in the period from 1975 to 1986 that the expenditure for public security, i.e. for police and fire brigades, decreased compared to expenditure for general administration. Similar evidence exists for the relation between teachers and administrative staff in schooling districts. The existing evidence on the efficiency of fiscal competition and decentralization is also inconclusive. Bergstrom, Roberts, Rubinfeld and Shapiro (1988) are the first to provide evidence that fiscal competition leads to an efficient provision of public education. Hoxby (2000) compares the relative efficiency of school services of jurisdictions depending upon the intensity of fiscal competition in which jurisdictions find themselves. She presents evidence that the performance of students per input unit is increased by fiscal competition, although it leads to significantly less spending per student. Similarly, Feld and Kirchgässner (1997) find no strong evidence against the hypothesis that fiscal competition leads to efficient public goods supply for Switzerland. Kirchgässner and Pommerehne (1996) and Feld (2000) report evidence on the impact of fiscal competition on income redistribution showing that the Swiss welfare state did not break down under the threat of fiscal competition. On the other hand, the international evidence on the impact of fiscal federalism on economic growth is very mixed (Feld, Zimmermann and Döring, 2003). The record is quite different and much more unambiguous in the case of direct democracy. The systematic empirical analyses of the impact of referenda and initiatives on economic policy for Switzerland and the US states provide strong support for the hypothesis that the Wicksellian (1896) connection rather exists in direct than in representative democracy such that the tax prices citizens pay for public goods and services are linked to the benefits the citizens obtain from these public goods and services.12 With respect to spending structure, the fiscal referendum mainly restricts welfare and administrative spending (Schaltegger, 2001; Vatter and Freitag, 2002). With respect to revenue structure, both the US states with initiatives and the Swiss cantons with a fiscal referendum rely more on user charges than Lars P. Feld and Gebhard Kirchgässner 293 on broad-based taxes.13 Moreover, Matsusaka (1995, 2002) and Feld, Schaltegger and Schnellenbach (2004) provide evidence that centralization of spending and revenue is reduced by the referendum. However, this does not necessarily mean that direct democracy is associated with an erosion of the social welfare state. As Feld, Fischer and Kirchgässner (2003) show, direct democratic Swiss cantons redistribute less income if the income gap between the highest- and lowest-income decile is relatively low and redistribute significantly more income if the income gap between the highest and lowest income decile is relatively high. Income redistribution appears to be more targeted and more effective in direct democratic cantons such that less funds are necessary to achieve redistributive goals. In addition, direct democracy also leads to a more efficient public sector. Pommerehne (1983) analysed costs and prices of local garbage collection in 103 Swiss cities in 1970. He found that average refuse collection costs (per household) were – ceteris paribus – lowest in cities with direct legislation and private garbage collection. For a panel of the Swiss cantons from 1970 to 1996, Barankay (2002) reports significantly lower infant mortality rates and a higher share of college degrees in more direct democratic cantons, suggesting that this indicates a higher quality of public goods in the cantons. Pommerehne and Weck-Hannemann (1996), Feld and Frey (2002) and Torgler (2002) show that in those Swiss cantons in which citizens have an impact on budgetary policy in direct legislation, tax morale is – ceteris paribus – higher compared to the cantons without such direct influence. If, however, the willingness to pay taxes is the higher the more satisfied citizens are with public services supplied, then these results are evidence for a higher satisfaction of citizens and, therefore, for greater efficiency of the provision of public services. Indeed, Frey and Stutzer (2000, 2002) present evidence that people in Switzerland perceive themselves as more satisfied with their life as a whole in direct democratic cantons keeping income levels and other controls constant. These studies lend support to the hypothesis that direct democratic systems are more efficient than representative democratic ones. A more efficient political system should also lead to better economic performance. Feld and Savioz (1997) study the relationship between budgetary referenda and the economic performance of Swiss cantons measured by GDP per employee. In a panel with annual data from 1984 to 1993 for the 26 Swiss cantons, they arrive at the conclusion that GDP per employee is – ceteris paribus – about 5 per cent higher in those cantons with budgetary referenda compared to cantons without those referenda. Again there is corroborating evidence by Kirchgässner, Feld and Savioz (1999), Freitag and Vatter (2000) for Switzerland and by Blomberg, Hess and Weerapana (2004) for the US states. All in all, the empirical evidence from the United States and Switzerland supports the hypothesis that (economic) policy outcomes in jurisdictions with referenda and initiatives are more closely oriented towards the Wicksellian connection of spending and tax prices. Fiscal restraints effectively reduce 294 Sustainable Fiscal Policy in a Federal System budget deficits, but, according to the existing evidence, do not necessarily increase the efficiency of public goods supply. Fiscal decentralization appears to increase the efficiency of public goods supply, but the long-run impact on economic growth is unclear. Still, fiscal decentralization is a precondition for direct democratic institutions to work properly in a federal state. On the other hand, fiscal restraints may serve as a means to reduce transaction costs, in particular in Swiss direct democracy if citizens do not want to veto government spending projects too frequently or want to bind themselves against future temptations for unsound fiscal policy. It looks like there is not much evidence against the Swiss mixture of fiscal institutions.

Notes

1. For a description of the SGP and the most recent cases of excessive deficits in the EU see http://europa.eu.int/comm/economy_finance/about/activities/sgp/edp_en.htm. Eijffinger and De Haan (2000) provide a good overview on the EU discussion. 2. Poterba (1997), Kirchgässner (2002a) and Schaltegger (2002) provide surveys about the effects of constitutional and/or statutory rules which are intended to reduce expenditure and/or the deficit. 3. Kirchgässner (2002a) summarizes the subsequent US studies about the impact of Proposition 13 on the quality of public goods and services, in particular the quality of public education. 4. Feld and Kirchgässner (2001a) leave out fiscal decentralization; Vatter and Freitag (2002) as well as Freitag, Vatter and Müller (2003) leave out fiscal restraints; Schaltegger (2002) leaves out fiscal federalism; and Feld and Matsusaka (2003) leave out fiscal decentralization and fiscal restraints. 5. The model we use is quite common in the study of fiscal policy, it corresponds, e.g., to the deficit and debt models of Roubini and Sachs (1989) and the spending and revenue models of Matsusaka (1995). 6. For a detailed description of these restraints see Stauffer (2001: 72). Since 1998, Lucerne (in 2001) as well as Berne and the Wallis (both in 2002) also introduced such regulations. 7. Because these variables might have an impact on government expenditure and revenue, we have to include these variables in order to get unbiased estimates for the coefficients of the other variables. – Due to space limitations, we restrict the discussion of our results to the interesting institutional and political variables. 8. The estimates have been performed with EViews, Version 4.1. The numbers in parentheses are the estimated t-statistics, based on the Newey–West autocorrelation- consistent standard errors. SER is the standard error of the regression, and J.–B. the value of the Jarque–Bera test for normality of the residuals. 9. We did the same with similar results for the revenue, tax and deficit equations. The results of these additional regressions are given in the Appendix. 10. Descriptive statistics of the political and institutional data are given in the Appendix. 11. For details see Table A12.2 in the Appendix. 12. See Breton (1996) for the argument and Pommerehne (1978) and Gerber (1996, 1999) for evidence showing that direct democracy leads to policy outcomes that are closer to citizens’ preferences. 13. See Matsusaka (1995) for the US states and Feld and Matsusaka (2003a) for the Swiss cantons. Lars P. Feld and Gebhard Kirchgässner 295

Appendix

Source of the data

• cantonal and local public expenditure per capita, • cantonal and local public revenue per capita, • cantonal and local tax revenue per capita. Source: Swiss Federal Finance Administration.

• disposable income per capita, • cantonal population, • share of population younger than 20, • share of population older than 65, • share of urban population, i.e. of people living in local communities with more than 10,000 inhabitants. Source: Swiss Federal Statistical Office.

• federal unconditional grants per capita, • tax burden in the highest income tax bracket of 1 million SFr annual taxable income, weighted with the inverse of the distances of the cantons’ capitals. Source: Own calculations on the basis of data of the Swiss Federal Finance Administration.

• ideological position of the cantonal government Source: Own calculations on the basis of data of the Swiss Federal Statistical Office.

• index of direct democracy Source: Own calculation of an index proposed by Stutzer and Frey (2000), using additional data from Trechsel and Serdült (1999).

• index of constitutional constraints Source: Own calculations, based on Stauffer (2001) and all monetary data have been deflated using the implicit GDP deflator with basis 2000 = 100.

Table A12.1 Descriptive statistics of the explanatory variables

Variables Mean Standard deviation Theoretical range Empirical Range

Fiscal decentralization1 0.331 0.109 0–1 0.152–0.462 Direct democracy 4.285 1.224 1–6 1.627–5.653 Fiscal constraints 0.298 0.820 0–3 0.000–3.000 Tax competition 0.237 0.079 0–1 0.101–0.373 Unconditional grants 267.065 106.443 0–... 223.49–1,318.68 Ideology of the government −0.100 0.185 −1–1 −0.600–0.400

Notes: The empirical range is calculated for the average values of the cantons over the total observation period, with the exception of the ideology of the government. For this variable the empirical range is calculated for the most left-wing and the most right-wing government. 1 The canton Basel-City is excluded from the empirical range, because its cantonal budget is nearly identical to the local budget. 296 .97) 23 3.05) − − ** ( 35* . ‘***’, ‘**’, ‘*’ or (*) 0.504 3.0 0.001* (2.56) 0.106** ( 0.017 (0.26) 3.053** (2.97) − − − − − 487 R is the standard error of the 5.68) 0.080** (3.29) 5.87) 0.605 (1.26) 1.68) 0.000 (0.95) 1.68) 0.000 2.84) − − − − 1.16) 0.498** (2.97) 1.16) 0.498** 1.44) − − 0.031*** ( 0.006 (1.24) 0.064*** (3.81) 0.039 ( 0.022(*) ( 0.201* (2.33) 0.041 (0.11) 0.011 ( 0.200*** ( 0.039** ( − − − − − − − − 485 1.66) 3.03) − 2.18) − 0.39) 0.54) − − − 4.059 1.982 4.471 4.059 1.982 0.093 0.077 0.390 0.093 0.077 0.017* ( 0.019 ( 0.008 ( 0.141** ( 0.300** (2.68) 0.026(*) ( − − − − − − 480 1.89) − 2.04) 0.74) 0.04) 0.01) 0.39) 0.012 (1.06) 0.39) 0.012 − − − − − 0.021** (2.62) 0.015* (2.53) 0.179* (1.97) 0.186** (2.60) 0.382*** (6.23) 0.564 (2.50) 0.758 0.810 0.916 0.827 (0.94) 0.967 (1.45) 1.198* (2.16) 0.119* (2.37) 0.158*** (3.60) 0.067(*) (1.85) 0.145(*) (1.74) 0.159* (2.32) 0.024 (0.054) 0.034 (0.19) 0.194(*) (1.78) 0.167(*) (1.82) 0.359*** (4.53) 0.012* (1.96) 0.007 ( 0.002 ( 0.000 ( 0.434** (2.81) 0.115* ( 0.006 ( 0.036(*) ( − − − − − − − Public expenditurePublic revenue Public revenue Tax Deficit 491 Expenditure, revenue, and deficit (per capita), 1980–1998 : The numbers in parentheses are the estimated t-statistics, based on the Newey–West autocorrelation-consistent standard errors standard autocorrelation-consistent Newey–West the on based t-statistics, estimated the are parentheses in numbers : The Italian-speaking population population Italian-speaking 2 R Dummy for Appenzell Ausserrhoden in 1996 0.085* (2.35) 0.331*** (11.16) 0.146*** (5.20) Share of young population population young of Share Share of old population Urbanization and for French- Dummy Ideology of the government Disposable income per capita Population Tax competition Tax competition Unconditional grants per capita Fiscal decentralization decentralization Fiscal Institutional constraints Institutional SER 0.113 J.–B. 3.815 Number of observationsNumber of Direct democracy Constant Table A12.2 variable Dependent Notes show that the estimated parameter is significantly different from zero at the 0.1, 1, 5, or 10 per cent level, respectively. SE respectively. level, cent per 10 or 5, 1, 0.1, the at zero from different significantly is parameter estimated the that show residuals. the of normality for Jarque–Bera-test the of value the J.–B. and regression, References

Abbott, A. and Hrycak, A. (1990) ‘Measuring Resemblance in Sequence Data: An Optimal Matching Analysis of Musicians’ Careers’, American Journal of Sociology, vol. 96, no. 1, 144–85. Acemoglu, D. (1998) ‘Why Do Technologies Complement Skills? Directed Technical Change and Wage Inequality’, Quarterly Journal of Economics, vol. 113, no. 4, 1055–89. Acemoglu, D. (2002) ‘Technical Change, Inequality and the Labour Market’, Journal of Economic Literature, vol. 40, no. 1, 7–72. Adams, P., Hurd, M.D., McFadden, D., Merrill, A. and Ribeiro, T. (2002) ‘Healthy, Wealthy, and Wise? Tests for Direct Causal Paths between Health and Socio- economic Status’, Journal of Econometrics, http://emlab.berkeley.edu/users/mjansson/ Courses/ECON242_SPRING02/McFadden.pdf. Ágh, A. (1999) ‘Europeanization of Policy-Making in East Central Europe: the Hungarian Approach to EU Accession’, Journal of European Public Policy, vol. 6, no. 5, 839–54. Aiken, L.H., Clarke, S.P., Sloane, D.M., Sochalski, J.A., Busse, R., Clarke, H., Giovannetti, P., Hunt, J., Rafferty, A.M. and Shamian, J. (2001) ‘Nurses’ Reports on Hospital Care in Five Countries’, Health Affairs, no. 20, 43–53. Aiken, L.H., Clarke, S.P., Sloane, D.M., Sochalski, J. and Silber, J.H. (2002) ‘Hospital Nurse Staffing and Patient Mortality, Nurse burnout, and Job Dissatisfaction’, Journal of the American Medical Association, no. 288, 1987–1993. Aisenbrey, S. (2000) Optimal Matching Analyses. Anwendungen in den Sozialwissenschaften, Opladen: Leske und Budrich. Aldous, J. (1977) ‘Family Interaction Patterns’, Annual Review of Sociology, no. 3, 105–35. Anderson, J.J. (2002) ‘Europeanization and the Transformation of the Democratic Polity, 1945–2000’, Journal of Common Market Studies, vol. 40, no. 5, 793–822. Andeweg, R. (1995) ‘The Reshaping of National Party Systems’, West European Politics, vol. 18, no. 3, 58–78. Antonucci, T.C. (1990) ‘Social Supports and Social Relationships’, in R.H. Binstock and L.K. George (eds), Handbook of Aging and the Social Sciences, San Diego: Academic Press, pp. 205–26. Aoki, M. (1986) ‘Horizontal vs. Vertical Information Structure of the Firm’, American Economic Review, vol. 76, no. 5, 971–83. Appelbaum, E. (2002) ‘Transformation of Work and Employment and New Insecurities’, in P. Auer and C. Daniel (eds), The Future of Work, Employment and Social Protection: The Search for New Securities in a World of Growing Uncertainties, Geneva: Inter- national Labor Organization (ILO), pp. 17–37. Arber, S. and Khlat, M. (2002) ‘Introduction to “Social and Economic Patterning of Women’s Health in a Changing World”’, Social Science & Medicine, vol. 54, pp. 643–7. Arbuckle, J.L. (1997) Amos. 3.62. Copyright 1994–1997, Chicago: SmallWaters Corporation. Arbuckle, J.L. and Wothke, W. (1999) Amos 4.0 User’s Guide. Chicago: SmallWaters Corporation. Argyle, M. (1999) ‘Causes and Correlates of Happiness’, in D. Kahneman, E. Diener and N. Schwarz (eds), Well-being: The Foundations of Hedonic Psychology, New York: Russell Sage Foundations, pp. 353–73.

297 298 References

Ariès, P. (1975) Geschichte der Kindheit, Munich and Vienna: Carl Hanser. Armingeon, K. (1998) ‘The Impact of Globalization on Swiss Policy Making: a Comment’, Swiss Political Science Review, vol. 4, no. 2, 104–11. Arnold, J. (1997) ‘The Psychology of Careers in Organizations’, in C.L. Cooper and I.T. Robertson (eds), International Review of Industrial and Organizational Psychology 1997, vol. 12, Chichester: Wiley, pp. 1–37. Arvanitis, S. (2003) Information Technology, Workplace Organization and the Demand for Labour of Different Skills: Firm-level Evidence for the Swiss Economy, KOF- Arbeitspapiere/Working Papers no. 83, Zurich. Arvanitis, S. (2004) ‘Information Technology, Workplace Organization, Human Capital and Firm Productivity: Evidence for the Swiss Economy’, in OECD (ed.), The Economic Impact of ICT – Measurement, Evidence and Policy Implications, Paris: OECD, pp. 183–211. Arvanitis, S., Bezzola, M., Donzé, L. and Hollenstein, H. (2000) Globalisierung, technischer Fortschritt und Qualifikationsstruktur der Schweizer Wirtschaft, Zurich: Federal Institute of Technology. Arvanitis, S., Bezzola, M., Donzé, L., Hollenstein, H. and Marmet, D. (2001), Die Inter- nationalisierung der Schweizer Wirtschaft. Ausmass, Motive, Auswirkungen, Zurich: vdf Hochschulverlag. Arvanitis, S. and Donzé, L. (2000) The Effects of Technological Change and Foreign Direct Investment on Labour Demand: An Empirical Analysis Based on Swiss Firm Data, Paper Presented at the 27th Annual Conference of the European Association for Research in Industrial Economics (EARIE), Lausanne, September, pp. 7–10. Arvanitis, S., Donzé, L., Hollenstein, H., Marmet, D. and Staib D. (2003) Technologischer und organisatorischer Wandel, Qualifikationserfordernisse und ihre Auswirkungen auf die Leistungsfähigkeit von Unternehmen, Zurich: Federal Institute of Technology. Arvanitis, S. and Hollenstein, H. (2002) ‘Die Wirtschaft in der Informationsgesellschaft’, in Bundesamt für Statistik, Gruppe für Wissenschaft und Forschung und Bundesamt für Kommunikation (Hrsg.), Informationsgesellschaft Schweiz: Standortbestimmung und Perspektiven, Neuchâtel, pp. 68–82. Arvanitis, S. and Marmet, D. (2001) Unternehmensgründungen in der schweizerischen Wirtschaft, Staatssekretariat für Wirtschaft, Bern. Asendorpf, J.B. and Wilpers, S. (1999) ‘KIT: Kontrolliertes Interaktions-Tagebuch zur Erfassung sozialer Interaktionen, Beziehungen und Persoenlichkeitseigenschaften’. ‘CID: Controlled Interaction Diary for the Assessment of Social Interactions, Relationships, and Personality Traits’, Diagnostica, vol. 45, no. 2, 82–94. Ashforth, B.E. and Saks, A.M. (1995) ‘Work-Role Transitions: a Longitudinal Examin- ation of the Nicholson Model’, Journal of Occupational and Organizational Psychology, no. 68, 157–75. Athey, S. and Stern, S. (1998) An Empirical Framework for Testing Theories about Complemen- tarity in Organizational Design, NBER Working Paper No. 6600, Cambridge, MA: NBER. Atkinson, A.B. (1998) Poverty in Europe, Oxford: Blackwell. Atkinson, A.B., Rainwater, L. and Smeeding, T.M. (1995) Income Distribution in OECD Countries: Evidence from the Luxembourg Income Study. Series Social Policy Studies No. 18, Paris: OECD. Autor, D.H., Katz, L.F. and Krueger, A.B. (1998) ‘Computing Inequality: How Computers Changed the Labour Market?’, Quarterly Journal of Economics, vol. 113, no. 4, 1169–1213. Autor, D.H., Levy, F. and Murane, R.J. (2003) ‘The Skill Content of Recent Techno- logical Change: An Empirical Exploration’, Quarterly Journal of Economics, vol. 118, no. 4, 1279–1333. References 299

Baethge, M. (2001a) ‘Abschied vom Industrialismus: Konturen einer neuen gesellschaftlichen Ordnung der Arbeit’, in M. Baethge and I. Wilkens (eds), Die grosse Hoffnung für das 21. Jahrhundert? Perspektiven und Strategien für die Entwicklung von Dienstleistungsbeschäftigung, Opladen: Leske & Budrich, pp. 23–69. Baethge, M. (2001b) ‘Qualifikationsentwicklung im Dienstleistungssektor’, in M. Baethge Martin and I. Wilkens (eds), Die grosse Hoffnung für das 21. Jahrhundert? Perspektiven und Strategien für die Entwicklung von Dienstleistungsbeschäftigung, Opladen: Leske + Budrich, pp. 85–106. Baethge, M. and Wilkens, I. (eds) (2001) Die grosse Hoffnung für das 21. Jahrhundert? Perspektiven und Strategien für die Entwicklung von Dienstleistungsbeschäftigung, Opladen: Leske+Budrich. Bagnasco, A. and Le Galès, P. (eds) (1997), Villes en Europe, Paris: La Découverte. Baillod, J. and Semmer, N. (1994) ‘Fluktuation und Berufsverläufe bei Computerfach- leuten’, Zeitschrift für Arbeits- und Organisationspsychologie, no. 38, 152–63. Barankay, I. (2002) Referenda, Citizens’ Initiatives and the Quality of Public Goods: Theory and Evidence, mimeo, University of Warwick. Bartley, M., Blane, D. and Smith, G.D. (eds) (1998) The Sociology of Health Inequalities, Oxford: Blackwell. Bartley, M., Sacker, A., Firth, D. and Fitzpatrick, R. (1999) ‘Social Position, Social Roles and Women’s Health in England: Changing Relationships 1984–1993’, Social Science & Medicine, vol. 48, pp. 99–115. Bartolini, S. (1999) ‘Political Representation in Loosely Bounded Territories. Between Europe and the Nation-State’, paper presented at the Multi-level party systems conference, ‘Europeanisation and the Reshaping of National Political Representation’, Florence, Italy. Bauer, T.K. and Bender, S. (2001) ‘Flexible Work-Systems and the Structure of Wages: Evidence from Matched Employer–Employee Data’, paper presented at the Annual Conference of the ‘Verein für Socialpolitik’, Magdeburg, 25–9 September. Bauman, Z. (1992) Intimations of Postmodernity, London: Routledge. Bauman, Z. (1998) Work, Consumerism and the New Poor, Buckingham, PA: Open University Press. Baumgartner, A.D. (2003) ‘Towards a New Regular Employment Relationship of Part-Time Work? Different Part-Time Employment Types in Switzerland’, Swiss Journal of Sociology, vol. 29, no. 1, 7–34. Beck, B. (2004) ‘A Special Case: A Survey of Switzerland’, The Economist, 14 February. Beck, U. (1986) Risikogesellschaft. Auf dem Weg in eine andere Moderne, Frankfurt/Main: Suhrkamp. Beck, U. [1986] (1992) Risk Society: Towards a New Modernity, London: Sage. Beck, U. and Beck-Gernsheim, E. (1994) Riskante Freiheiten. Individualisierung in modernen Gesellschaften, Frankfurt am Main: Suhrkamp. Beck, U., Giddens, A. and Lash, S. (1994) Reflexive Modernization: Politics, Tradition and Aesthetics in the Modern Social Order, Stanford: Stanford University Press. Becker, R. and Schömann, K. (1996) ‘Berufliche Weiterbildung und Einkom- mensdynamik’, Kölner Zeitschrift für Soziologie und Sozialpsychologie, vol. 48, no. 3, 426–61. Becker-Schmidt, R. (1987) ‘Die doppelte Vergesellschaftung – die doppelte Unter- drückung: Besonderheiten der Frauenforschung in den Sozialwissenschaften’ in Lilo Unterkircher and Ina Wagner (eds) Die andere Hälfte der Gesellschaft, Österreichischer Soziologentag 1985, Vienna: ÖGB-Verlag. Beehr, T.A. (1995) Psychological Stress in the Workplace, London: Routledge. 300 References

Beehr, T.A. and Glazer, S. (2001) ‘A Cultural Perspective of Social Support in Relation to Occupational Stress’, in P.L. Perrewé and D.C. Ganster (eds), Exploring Theoretical Mechanisms and Perspectives: Research in Occupational Stress and Well-Being, vol. 1, Amsterdam: Elsevier, pp. 97–142. Bell, D. (1973) The Coming of Post-Industrial Society: a Venture in Social Forecasting, New York: Basic Books. Bendel, S., Buchmann, M. and Sacchi, S. (1995) Personalsuche und Stelleninserate. Kurzbericht zur Firmenbefragung, Soziologisches Institut der Universität Zürich. Benz, A. (2001) ‘Vom Stadt-Umland-Verband zu “Regional Governance” in Stadtregio- nen’, Deutsche Zeitschrift für Kommunalwissenschaft, vol. 40, no. 2, 55–71. Berger, P.A., Steinmüller, P. and Sopp, P. (1993) ‘Differentiation of Life-Courses? Changing Patterns of Labour-Market Sequences in West Germany’, European Sociological Review, vol. 9, no. 1, 43–65. Bergstrom, T.C., Roberts, J.A., Rubinfeld, D.L. and Shapiro, P. (1988) ‘A Test for Efficiency in the Supply of Public Education’, Journal of Public Economics, vol. 35, 289–307. Berkman, L.F. (1995) ‘The Role of Social Relations in Health Promotion’, Psychosomatic Medicine, vol. 57, no. 3, 245–54. Berkman, L.F. and Glass, T. (2000) ‘Social Integration, Social Networks, Social Support and Health’, in L.F. Berkman and I. Kawachi (eds), Social Epidemiology, Oxford: Oxford University Press, pp. 137–73. Berkman, L.F. and Kawachi I. (2000a) ‘A Historical Framework for Social Epidemiology’, in L.F. Berkman and I. Kawachi (eds), Social Epidemiology, Oxford and New York: Oxford University Press, pp. 3–12. Berkman, L.F. and Kawachi, I. (2000b) Social Epidemiology, Oxford and New York: Oxford University Press. Berman, E., Bound, J. and Machin, S. (1998) ‘Implications of Skill-Biased Technological Change; International Evidence’, Quarterly Journal of Economics, vol. 113, no. 4, 1245–1279. Bernhardt, A. and Marcotte, D.E. (2000) ‘Is “Standard Employment” Still What It Used to Be?’ in Françoise Carré, Marianne A. Ferber, Lonnie Golden and Stephen A. Herzenberg (eds), Nonstandard Work: The Nature and Challenge of Changing Employment Arrangements, Champaign, IL: Industrial Relations Research Association (IRRA), pp. 21–40. Bernhardt, A., Morris, M., Handcock, M.S. and Scott, M.A. (2001) Divergent Paths: Economic Mobility in the New American Labor Market, New York: Russell Sage Foundation. Besley, T. and Smart M. (2003) ‘Fiscal Restraints and Voter Welfare’, mimeo, London School of Economics. Black, S.E. and Lynch, L.M. (2000) What’s Driving the New Economy: The Benefits of Workplace Innovation, NBER Working Paper No. 7479, Cambridge, MA: NBER. Blaxter, M. (1991) ‘Fifty Years On – Inequalities in Health’, Population Studies, vol. 45, Population Research in Britain: Supplement, pp. 69–94. Blickle, G. (1998) ‘Berufsorientierungen und Motive: Eine Längsschnittstudie’, Zeitschrift für Differentielle und Diagnostische Psychologie, no. 19, 164–78. Blomberg, S.B., Hess, G.D. and Weerapana, A. (2004) ‘The Impact of Voter Activity on Economic Activity’, European Journal of Political Economy, vol. 20, 207–26. Blossfeld, H.P. and Hakim, C. (eds) (1997) Between Equalization and Marginalization: Women Working Part-Time in Europe and the United States of America, Oxford: Oxford University Press. Bodenmann, G. (1995) ‘A Systemic-Transactional View of Stress and Coping in Couples’, Swiss Journal of Psychology, no. 54, 34–49. References 301

Bohn, H. and Inman, R.P. (1996) Balanced-Budget Rules and Public Deficits: Evidence from the U.S. States, Carnegie-Rochester Conference Series on Public Policy, no. 45, pp. 13–76. Böhnke, P. and Delhey, J. (1999) Poverty in a Multidimensional Perspective: Great Britain and Germany in Comparison. Discussion Paper FS III 99-413, Berlin: Social Science Research Center. Boje, T., Almqvist, P. and Almqvist, A.-L. (2000) ‘Citizenship, Family Policy, and Women’s Patterns of Employment’, in Thomas P. Boje and Arnlaug Leira (eds), Gender, Welfare State and the Market, London: Routledge. Bolger, N., Davis, A. and Rafaeli, E. (2003) ‘Diary Methods: Capturing Life as it is Lived’, Annual Review of Psychology, no. 54, 579–616. Bolin, K., Lindgren, B., Lindström, M. and Nystedt, P. (2003) ‘Investments in Social capital – Implications of Social Interactions for the Production of Health’, Social Science & Medicine, vol. 56, no. 12, 2379–2390. Born, C., Krüger, H., Lorenz-Meyer, D. (1996) Der unentdeckte Wandel. Annäherung an das Verhältnis von Struktur und Norm im weiblichen Lebenslauf, Berlin: Edition Sigma. Börzel, T.A. and Risse, T. (2000) ‘When Europe Hits Home: Europeanization and Domestic Changes’, European Integration Online Papers (EIoP) 4 (15). http://eiop.or.at/ eiop/texte/2000-015.htm. Bosma, H., Stansfeld, S.A. and Marmot, M.G. (1998) ‘Job Control, Personal Character- istics, and Heart Disease’, Journal of Occupational Health Psychology, no. 3, 402–9. Bott, E. (1971) Family and Social Network: Roles, Norms and External Relationships in Ordinary Urban Families, London: Tavistock. Boylan R.D. and Morgan, J.P. (1993) ‘The Effect of the Number of Diplomas on Their Value’, Sociology of Education, vol. 66, no. 3, 206–21. Braverman, H. (1974) Labor and Monopoly Capitalism: The Degradation of Work in the Twentieth Century, New York: Monthly Review Press. Brennan, G. and Buchanan, J.M. (1980) The Power to Tax: Analytical Foundations of a Fiscal Constitution, Cambridge: Cambridge University Press. Bresnahan, T.F. (1999) ‘Computerisation and Wage Dispersion: An Analytical Reinter- pretation’, Economic Journal, vol. 109, 390–415. Bresnahan, T.F., Brynjolfsson, E. and Hitt, L.M. (2002) ‘Information Technology, Workplace Organisation, and the Demand for Skilled Labour: Firm-level Evidence’, Quarterly Journal of Economics, vol. 112, no. 1, 339–76. Breton, A. (1996) Competitive Governments: An Economic Theory of Politics and Public Finance, Cambridge: Cambridge University Press. Brett, J.M. (1980) ‘The Effect of Job Transfer on Employees and Their Families’, in C.L. Cooper and R. Payne (eds), Current Concerns in Occupational Stress, Chichester: Wiley, pp. 99–136. Brody, L.R. and Hall, J.A. (2000) ‘Gender, Emotion, and Expression’, in M. Lewis and J.M. Haviland-Jones (eds), Handbook of Emotions, 2nd edn, New York: Guilford, pp. 338–49. Bruggemann, A. (1974) ‘Zur Unterscheidung verschiedener Formen von “Arbeitszu- friedenheit”’, Arbeit und Leistung, no. 28, pp. 281–4. Bryk, A.S. and Raudenbush, S.W. (1992) Hierarchical Linear Models: Applications and Data Analysis Methods, Thousand Oaks, CA: Sage Publications Inc. Buchmann, M. (1989) The Script of Life in Modern Societies: Entry into Adulthood in a Changing World, Chicago: University of Chicago Press. Buchmann, M. (2002) ‘Labour Market Entry and Beyond: Some Reflections on the Changing Structure of Work’, Education & Training, vol. 44, no. 4/5, 217–23. 302 References

Buchmann, M., König, M., Li, J.H. and Sacchi, S. (1999) Weiterbildung und Beschäfti- gungschancen, Rüegger, Zürich, Chur. Buchmann, M. and Sacchi S. (eds) (forthcoming), Wandel der betrieblichen Qualifika- tionsnachfrage. Der schweizerische Arbeitsmarkt im Spiegel von Stelleninseraten. Budowski, M., Hanson, T. and Suter, C. (1999) ‘Living Conditions of Single Mothers in Zurich, Switzerland and Urban United States’, in C. Mäder, C. Burton-Jeangros and M. Haour-Knipe (eds), Health, Medicine and Society: Contributions to the Sociology of Health, Zurich: Seismo, pp. 256–87. Budowski, M., Tillmann, R. and Bergman, M.M. (2002) ‘Poverty, Stratification and Gender in Switzerland’, Swiss Sociological Review, vol. 28, no. 2, 297–317. Buehler, C. (1990) ‘Adjustment’, In J. Touliatos and B.F. Perlmutter (eds), Handbook of Family Measurement Techniques, Thousand Oaks, CA: Sage Publications Inc, pp. 493–574. Bühler, E. (2001) Frauen- und Gleichstellungsatlas, Zurich: Seismo. Bühler, E.(2002) Atlas suisse des femmes et de l’égalité, Zurich: Seismo. Bundesamt für Statistik (1996) Die Schweizerische Arbeitskräfteerhebung (SAKE). Konzepte, methodische Grundlagen, praktische Ausführung, Bern: BFS. Bundesamt für Statistik (2001) Bilanz der Wohnbevölkerung in den Gemeinden der Schweiz. Statistik des jährlichen Bevölkerungsstandes (ESPOP) 2000. Neuchâtel: BFS. Bundesamt für Statistik (2003) Swiss Labor Force Survey (SLFS) 2002. Retrieved 25 March 2003, from the World Wide Web: www.statistik.admin.ch Burgess, E.W., Locke, H.J., and Thomes, M.M. (1960) The Family: From Institution to Companionship, New York: American Book. Burgess, J. and Watts, M. (1999) Labour Market Reform, the End of the Standard Employment Model and Workforce Polarisation in Australia, University of Newcastle, Australia. Burke, R.J. and Greenglass, E.R. (1999) ‘Work-Family Conflict, Spouse Support, and Nursing Staff Well-Being During Organizational Restructuring’, Journal of Occupational Health Psychology, no. 4, 327–36. Byrne, B.M. (2001) Structural Equation Modelling with AMOS: Basic Concepts, Applications, and Programming, Mahwah, NJ: Lawrence Erlbaum. Callan, T., Nolan, B. and Whelan, C.T. (1993) ‘Resources, Deprivation and the Measure- ment of Poverty’, Journal of Social Policy, vol. 22, no. 2, 141–72. Canada Statistics (1999) Information and Communications Technologies and Electronic Commerce, Survey, Ottawa: Canada Statistics. Capelli, P. and Carter, W. (2000) Computers, Work Organization and Wage Outcomes, NBER Working Paper No. 7987, Cambridge, MA: NBER. Caroli, E. (2001) ‘New Technologies, Organizational Change and the Skill Bias: What Do We Know?’, in P. Petit and L. Soete (eds), Technology and the Future of European Employment, Cheltenham, Edward Elgar, pp. 259–92. Caroli, E., Greenan, N. and Guellec, D. (2001) ‘Organizational Change and Skill Accumulation’, Industrial and Corporate Change, vol. 10, no. 2, 481–506. Caroli, E. and van Reenen, J. (1999) Skill Biased Organizational Change? Evidence from a Panel of British and French Establishments, Paris: Centre d’études prospectives d’économie mathematique appliquées a la planification (CEPREMAP). Caroli, E. and Van Reenen, J. (2001) ‘Skill Biased Organizational Change? Evidence from a Panel of British and French Establishments’, Quarterly Journal of Economics, vol. 116, no. 4, 1449–92. Carroll, G.R. (1984) ‘Organizational Ecology’, Annual Review of Sociology, vol. 10, 71–93. Carroll, G.R., Havemann H. and Swaminathan, A. (1990). ‘Karrieren in Organisationen. Eineökologische Perspektive’, in Mayer, Karl Ulrich (ed.), Lebensverläufe und sozialer Wandel. Opladen: Westdeutscher Verlag, pp. 146–78. References 303

Case, A., Lubotsky, D. and Paxon, C. (2001) Economic Status and Health in Childhood, Research Briefs Princeton, Woodrow Wilson School of Public and International Affairs, Princeton University. Castells, M. (1996) The Rise of the Network Society. The Information Age: Economy, Society and Culture, vol. 1, Oxford: Blackwell. Caulfield, J., and Larsen, H. (2002) Local Government at the Millenium, Opladen: Leske+Budrich. Chennels, L. and Van Reenen, J. (1999) Has Technology Hurt Less Skilled Workers? An Econometric Survey of the Effects of Technical Change on the Structure of Pays and Jobs, Working Paper No. 27, London: Institute of Fiscal Studies. Christoph, B. and Noll, H.-H. (2003) ‘Subjective Well-Being in the European Union during the 90s’, Social Indicators Research, vol. 64, 521–46. Chun, H. (2003) ‘Information Technology and the Demand for Educated Workers: Disentangling the Impacts of Adoption versus Use’, Review of Economics and Statistics, vol. 85, no. 1, 1–8. Clark, A.E. and Oswald, A.J. (1994) ‘Unhappiness and Unemployment’, Economic Journal, vol. 104, pp. 648–59. Cohen, P., Cohen, J., Aiken, L.S. and West, S.G. (1999) ‘The Problems of Units and the Circumstance for POMP’, Multivariate Behavioral Research, vol. 34, no. 3, 315–46. Comment, T., Hulliger, B. and Ries, A. (1996) Gewichtungsverfahren für die Schweizerische Arbeitskräfteerhebung (1991–1995), SAKE-News Nr. 4/96, Bern: BFS. Cornetz, W. (1991) ‘Ökonomische Aspekte des dienstleistungsorientierten Strukturwan- dels’, in W. Littek, U. Heisig and H.-D. Gondek (eds), Dienstleistungsarbeit. Struktur- veränderungen, Beschäftigungsbedingungen und Interessenlagen, Berlin: Edition Sigma, pp. 35–52. Council of Europe (1997) Local Finance in Europe, Strasbourg: Council of Europe Publishing. Cramer, U. (1990) ‘Probleme bei der Besetzung von offenen Stellen. Ergebnisse einer Betriebsbefragung’, Mitteilungen aus der Arbeitsmarkt- und Berufsforschung, vol. 23, no. 2, 246–54. Cummins, R.A. (2000) ‘Objective and Subjective Quality of Life: An Interactive Model’, Social Indicators Research, vol. 52, 55–72. Cunha, A. and Schuler, M. (2001) ‘Métropolisation, changement de régime d’urbanisation et fragmentation de l’espace: enjeux de la gouvernance des aglomérations en Suisse’, Revue suisse de science politique, vol. 7, no. 4, 121–4. Dahl, R.A. (1994) ‘A Democratic Dilemma: System Effectiveness Versus Citizen Participation’, Political Science Quarterly, no. 1, 23–34. Dahrendorf, R. (1964) ‘Recent Changes in the Class Structures of European Societies’, Daedalus, vol. 93, Winter, 225–70. Daniels, N., Kennedy, B.C. and Kawachi, I. (2000) Is Inequality Bad for Our Health?, Boston, MA: Beacon Press. Deeke, A. (1991) Informelle Beziehungen auf dem Arbeitsmarkt. Marktregulierung und Chancenverteilung durch Arbeitsvermittlung, Frankfurt: Campus. De Grip, A., Hoevenberg, J. and Willems, E. (1997) ‘Atypical Employment in the European Union’, International Labour Review, vol. 136, no. 1, 49–71. Delcher, A.L., Kasif, S., Fleischmann, R.D., Peterson, J., White, O. and Salzberg, S.L. (1999) ‘Alignment of Whole Genomes’, Nucleic Acid Research, vol. 27, no. 11, 2369–76. Dente, B. (1990) ‘Metropolitan Governance Reconsidered, or How to Avoid Errors of the Third Type’, Governance, vol. 3, no. 1, 55–74. 304 References

Deutsch, K.W. (1976) Die Schweiz als ein paradigmatischer Fall politischer Integration, Bern: Haupt. Dicken, P. (1992) Global Shift: The Internationalization of Economic Activity, London: Paul Chapman. Diener, E., Suh, E.M., Lucas, R.E. and Smith, H.L. (1999) ‘Subjective Well-Being: Three Decades of Progress’, Psychological Bulletin, vol. 125, no. 2, 276–302. Dietzen, A. and Kloas, P.-W. (1999) ‘Stellenanzeigenanalyse – eine effektive Methode zur Früherkennung des Qualifikationsbedarfs’, in L. Alex and H. Bau (eds), Wandel beruflicher Anforderungen. Der Beitrag des BIBB zum Aufbau eines Früherkennungssystems Qualifikationsentwicklung, Bielefeld: Bertelsmann, pp. 13–32. Dombois, R. (1999) ‘Der schwierige Abschied vom Normalarbeitsverhältnis’, Aus Politik und Zeitgeschichte, no. B 37/99, 13–20. Doms, M., Dunne, T. and Troske, K.R. (1997) ‘Workers, Wages and Technology’, Quarterly Journal of Economics, vol. 112, no. 1, 253–90. Donati, P. (1985) Famiglia e politiche sociali, Milan: Franco Angeli. Donzé, L. (2001) ‘L’imputation des données manquantes, la technique de l’imputation multiple, les conséquences sur l’analyse des données: l’enquête 1999 KOF/ETHZ sur l’innovation’, Schweizerische Zeitschrift für Volkswirtschaft und Statistik, vol. 137, no. 3, 301–17. Downes, T.A. (1992) ‘Evaluating the Impact of School Finance Reform on the Provision of Public Education: The California Case’, National Tax Journal, vol. 45, 405–19. Downes, T.A., Dye, R.F. and McGuire, T.J. (1998) ‘Do Limits Matter? Evidence on the Effects of Tax Limitations on Student Performance’, Journal of Urban Economics, vol. 43, 401–17. Downes, T.A. and Figlio, D.N. (1997) ‘School Finance Reform, Tax Limits, and Student Performance: Do Reforms Level Up or Dumb Down?’, mimeo, University of Florida, Gainsville. Downes, T.A. and Figlio, D.N. (1999) ‘Do Tax and Expenditure Limits Provide a Free Lunch? Evidence on the Link between Limits and Public Sector Service Quality’, National Tax Journal, vol. 52, 113–28. Downs, A. (1994) New Visions for Metropolitan America, Washington, DC: Brookings Institution. Drobnic, S. (2000) ‘The Effects of Children on Married and Lone Mothers’ Employment in the United States and (West) Germany’, European Sociological Review, vol. 16, no. 2, 137–57. Dupont, C. and Sciarini, P. (2001) ‘Switzerland and the European Integration Process: Engagement Without Marriage’, West European Politics, vol. 24, no. 2, 211–32. Edwards, P.N. (1995) ‘From “Impacts” to Social Processes: Computers in Society and Culture’, in S. Jasanoff, G.E. Markle, J.C. Peterson and T. Pinch (eds), Handbook of Science and Technology Studies, Thousand Oaks, CA: Sage, pp. 257–85. Eidgenössisches Finanzdepartement (2000) Schulden von Bund, Kantonen und Gemeinden; http://www.admin.ch/efv/finanzen/d/oefffina/oeffschu.htm. Eijffinger, S.C.W. and De Haan, J. (2000) European Monetary and Fiscal Policy, Oxford: Oxford University Press. Eisner, M., Graf, N. and Moser, P. (2003) Risikodiskurse. Die Dynamik öffentlicher Debatten über Umwelt- und Risikoprobleme in der Schweiz, Seismo, Zürich. Elder, G.H. (1995) ‘The Life Course Paradigm: Social Change and Individual Development’, in M. Phyllis, H.E. Glen and K. Lüscher (eds), Examining Lives in Context: Perspectives on the Ecology of Human Development, Washington, DC: APA Press, pp. 101–39. References 305

Elfering, A., Grebner, S., Semmer, N.K. and Gerber, H. (2002) ‘Time Control, Catecho- lamines and Back Pain among Young Nurses’, Scandinavian Journal of Work, Environment, Health, vol. 28, 386–93. Elfering, A., Semmer, N.K. and Kälin, W. (2000) ‘Stability and Change in Job Satisfaction at the Transition form Vocational Training into “Real Work”’, Swiss Journal of Psychology, vol. 59, 256–71. Elfering, A., Semmer, N.K., Kälin, W. and Monnerat, F. (2004a) ‘Core Self-evaluations and Job Satisfaction: Evidence from a Three-Wave Study’, Manuscript under review. Elfering, A., Semmer, N.K., Tschan, F., Kälin, W. and Bucher, A. (2004b) ‘First Years in Job: A Three-Wave Analysis of Work Experiences’, paper submitted for publication. Elstad, J.I. (1998) ‘The Psycho-Social Perspective on Social Inequalities in Health’, in M. Bartley, D. Blane and G.D. Smith (eds), The Sociology of Health Inequalities, Oxford: Blackwell Publishers, pp. 39–58. Epstein, S. (1998) ‘Personal Control from the Perspective of Cognitive-Experiential Self-Theory’, in M. Kofta, G. Weary and G. Sedek (eds), Personal Control in Action: Cognitive and Emotional Mechanisms, New York: Plenum, pp. 5–26. Erzberger, C. and Prein, G. (1997) ‘Optimal-Matching-Technik: Ein Analyseverfahren zur Vergleichbarkeit und Ordnung individuell differenter Lebensverläufe’, ZUMA- Nachrichten, vol. 21, no. 40, 52–80. Eschuk, G. (2003) New Forms of Work Organization, Skills and Training, Quebec: Human Resources Development Canada. Esping-Andersen, G. (1990) The Three Worlds of Welfare Capitalism, Princeton, NJ: Princeton University Press. Esping-Andersen, G. (1996) ‘Equality or Employment? The Interaction of Wages, Welfare States and Family Change’, Transfer, vol. 4, 615–34. European Foundation for the Improvement of Living and Working Conditions (1997) New Forms of Work Organization: Can Europe Realize its Potential? Results of a Survey of Direct Employee Participation in Europe, Luxembourg: Office for Official Publications of the European Communities. Evans, R.G., Barer, M.L. and Marmor T.R. (1994) Why Are Some People Healthy and Others Not? The Determinants of Health of Populations, New York: Aldine de Gruyter. Fahrenberg, J. (1994) ‘Ambulantes Assessment, Computerunterstütze Datenerfassung unter Alltagsbedingungen’, Diagnostica, vol. 40, 195–216. Falk, M. (2001) Diffusion of Information Technology, Internet Use and the Demand for Heterogeneous Labour, ZEW Discussion Paper No. 48, Mannheim. Falk, M. (2002) ‘Endogenous Organizational Change and the Expected Demand for Different Skill Groups’, Applied Economics Letters, vol. 9, 419–23. Falk, M. and Seim, K. (1999) ‘Worker’s Skill Level and Information Technology: Evidence from German Service Firms’, in L. Bellmann and V. Steiner (eds), Panelanalysen zu Lohnstruktur, Qualifikation und Beschäftigungsdynamik, Nuremberg: Institut für Arbeitsmarkt- und Berufsforschung (IAB), pp. 79–116. Falk, M. and Seim, K. (2001) ‘The Impact of Information Technology on High-Skilled Labour in Services: Evidence from Firm Level Panel Data’, Economics of Innovation and New Technology, vol. 10, no. 1, 289–323. Feij, J.A., Whitely, W.T., Peiró, J.M. and Taris, T.W. (1995) ‘The Development of Career-Enhancing Strategies and Content Innovation: A Longitudinal Study of New Workers’, Journal of Vocational Behavior, vol. 46, 231–256. Feld, L.P. (2000), ‘Tax Competition and Income Redistribution: an Empirical Analysis for Switzerland’, Public Choice, vol. 105, 125–64. 306 References

Feld, L.P., Fischer, J. and Kirchgässner, G. (2003) ‘The Effect of Direct Democracy on Income Redistribution: Evidence for Switzerland’, Unpublished Manuscript, Philipps-University Marburg. Feld, L.P. and Frey, B.S. (2002) The Tax Authority and the Taxpayer: An Exploratory Analysis, mimeo, University of St Gallen. Feld, L.P. and Kirchgässner, G. (1997) ‘Die Kapitalisierung von Steuern und öffentlichen Leistungen in den Mietzinsen: Eine empirische überprüfung der Tiebout-Hypothese für die Schweiz’, in H. Schmid and T. Slembeck (eds), Finanz- und Wirtschaftspolitik in Theorie und Praxis, Bern: Haupt, pp. 64–92. Feld, L.P. and Kirchgässner, G. (1999) ‘Public Debt and Budgetary Procedures: Top Down or Bottom Up? Some Evidence from Swiss Municipalities’, in J. Poterba and J. von Hagen (eds), Fiscal Institutions and Fiscal Performance, Chicago: Chicago University Press and NBER, pp. 151–79. Feld, L.P. and Kirchgässner G. (2001a) ‘The Political Economy of Direct Legislation: The Role of Direct Democracy in Local and Regional Decision-Making’, Economic Policy, vol. 16, no. 33, 329–67. Feld, L.P. and Kirchgässner G. (2001b) ‘Does Direct Democracy Reduce Public Debt? Evidence from Swiss Municipalities’, Public Choice, vol. 109, 347–70. Feld, L.P., Kirchgässner, G. and Schaltegger, C.A. (2003) Decentralized Taxation and the Size of Government: Evidence from Swiss State and Local Governments, CESifo Working Paper No. 1087. Feld, L.P. and Matsusaka, J.G. (2003a) ‘Budget Referenda and Government Spending: Evidence from Swiss Cantons’, Journal of Public Economics, vol. 87, 2703–24. Feld, L.P. and Matsusaka, J.G. (2003b) The Political Economy of Tax Structure: Some Panel Evidence for Swiss Cantons, mimeo, University of St Gallen. Feld, L.P. and Reulier, E. (2002) ‘Strategic Tax Competition in Switzerland: Evidence from a Panel of the Swiss Cantons’, Unpublished Manuscript, University of St Gallen. Feld, L.P. and Savioz, M.R. (1997) ‘Direct Democracy Matters for Economic Performance: An Empirical Investigation’, Kyklos, vol. 50, 507–38. Feld, L.P. and Schaltegger, C.A. (2004) ‘Voters as Hard Budget Constraints: On the Determination of Intergovernmental Grants’, forthcoming in: Public Choice. Feld, L.P., Schaltegger, C.A. and Schnellenbach, J. (2004) On Government Centralization and Budget Referenda: A Model and Evidence from Switzerland, mimeo, Philipps- University Marburg. Feld, L.P., Zimmermann, H. and Döring, T. (2003) ‘Föderalismus, Dez1entralität und Wirtschaftswachstum’, Vierteljahreshefte für Wirtschaftsforschung, vol. 72, 361–77. Fernandez, R.M. (2001) ‘Skill-Biased Technological Change and Wage Inequality: Evidence from a Plant Retooling’, American Journal of Sociology, vol. 107, no. 2, 273–320. Figlio, D.N. (1997) ‘Did the “Tax Revolt” Reduce School Performance?’, Journal of Public Economics, vol. 65, 245–69. Figlio, D.N. (1998) ‘Short-Term Effects of a 1990s-Era Property Tax Limit: Panel Evidence on Oregon’s Measure 5’, National Tax Journal, vol. 51, 55–70. Figlio, D.N. and O’Sullivan, A. (1997) ‘The Local Response to Tax Limitation Measures: Do Local Governments Manipulate Voters to Increase Revenues?’, Journal of Law and Economics, vol. 44, 233–56. Figlio, D.N. and Rueben, K.S. (2001) ‘Tax Limits and the Qualifications of New Teachers’, Journal of Public Economics, vol. 80, 49–61. Fincham, F.D. (2003) ‘Marital Conflict: Correlates, Structure, and Context’, Current Directions in Psychological Science, vol. 12, no. 1, 23–7. References 307

Fincham, F.D., Beach, S.R.H., Arias, I. and Brody, G.H. (1998) ‘Children’s Attributions in the Family: The Children’s Relationship Attribution Measure’, Journal of Family Psychology, vol. 12, no. 4, 481–93. Fischer, A. (2002) ‘Wirtschaftsbranche, Gewerkschaftsstärke und Interessengegensätze der Arbeitgeber’, Schweizerische Zeitschrift für Politikwissenschaft, vol. 8, no. 3/4, 85–100. Fischer, A. (2003) ‘Vetospieler und die Durchsetzbarkeit von Side-Payments’, Schweizerische Zeitschrift für Politikwissenschaft, vol. 9, no. 2, 27–58. Fischer, A., Nicolet, S. and Sciarini, P. (2002) ‘Europeanisation of Non-EU Countries: The case of Swiss Immigration Policy Towards the EU’, West European Politics, vol. 25, no. 3, 143–70. Fischer, A., Nicolet, S. and Sciarini, P. (2003) ‘La politique des télécommunications suisse: Entre pression internationale et résistance nationale’, Politiques et Management Public, vol. 21, no. 3, 1–20. Fourastié, J. (1952) Le Grand Espoir du XXe Siècle (édition définitive), Paris: Gallimard. Fox, J. (ed.) (1989) Health Inequalities in European Countries, Aldershot: Gower. Francois, J.-P., Favre, F. and Greenan, N. (1999) Organizational Changes in Industrial Firms and Computerization of Industrial Enterprises, Paris: OECD. Freitag, M. and Vatter, A. (2000) ‘Direkte Demokratie, Konkordanz und Wirtschaft- sleistung: Ein Vergleich der Schweizer Kantone’, Schweizerische Zeitschrift für Volkswirtschaft und Statistik, vol. 136, 579–606. Freitag, M., Vatter, A. and Müller, C. (2003) ‘Bremse oder Gaspedal? Eine empirische Untersuchung zur Wirkung der direkten Demokratie auf den Steuerstaat’, Politische Vierteljahresschrift, vol. 44, 348–69. Frese, M. and Semmer, N. (1991) ‘Stressfolgen in Abhängigkeit von Moderatorvariablen: Der Einfluss von Kontrolle und sozialer Unterstützung’, in S. Greif, E. Bamberg and N. Semmer (eds), Psychischer Stress am Arbeitsplatz, Göttingen: Hogrefe, pp. 135–53. Frey, B.S. and Eichenberger, R. (2001) ‘Metropolitan Governance for the Future: Functional Overlapping Competing Jurisdictions (FOCJ)’, Swiss Political Science Review, vol. 7, no. 3, 124–30. Frey, B.S. and Stutzer, A. (2000) ‘Happiness, Economy and Institutions’, Economic Journal, vol. 110, 918–38. Frey, B.S. and Stutzer A. (2002) Happiness and Economics, Princeton/Oxford: Princeton University Press. Frisken, F. and Norris, D.F. (2001) ‘Regionalism Reconsidered’, Journal of Urban Affairs, vol. 23, no. 5, 467–78. Fuchs, D. and Klingemann, H.-D. (1995) ‘Citizens and the State: a Changing Relationship?’, in H.-D. Klingemann and D. Fuchs (eds), Citizens and the State, Oxford: Oxford University Press, pp. 1–23. Fux, B. and Baumgartner, A.D. (2001) ‘Impact of Population Related Policies on Selected Living Arrangements: Comparative Analyses at the Regional Level in Belgium, the Netherlands, and Switzerland’, in E. Klijzing and M. Corijn (eds), Flagship Conference: Partnership and Fertility a Revolution?, vol. 2, A Selection of Contributed Papers, Geneva: UN-ECE. Gallagher, M., Millar, R., Hargie, O. and Ellis, R. (1992) ‘The Personal and Social Worries of Adolescents in Northern Ireland: Results of a Survey’, British Journal of Guidance and Counselling, vol. 20, 274–90. Garrett, G. and Lange, P. (1995) ‘Internationalization, Institutions, and Political Change’, International Organization, vol. 49, no. 4, 627–55. Gerber, E.R. (1996) ‘Legislative Response to the Threat of Initiatives’, American Journal of Political Science, vol. 40, 99–128. 308 References

Gerber, E.R. (1999) The Populist Paradox: Interest Group Influence and the Promise of Direct Legislation, Princeton, NJ: Princeton University Press. Gerlach, K., Hübler, O. and Meyer W. (eds) (1998) Ökonomische Analysen betrieblicher Strukturen und Entwicklungen. Das Hannoveraner Firmenpanel, Frankfurt/M: Campus. Gerlach, K. and Jirjahn, U. (1998) ‘Technischer Fortschritt, Arbeitsorganisation und Qualifikation: Eine empirische Analyse für das verarbeitende Gewerbe Niedersachsens’, Mitteilungen aus der Arbeitsmarkt- und Berufsforschung, vol. 31, no. 3, 426–37. Germann, R.E. (1994) Staatsrefom. Der Übergang zur Konkurrenzdemokratie, Berne: Haupt. Germann, R.E. (1996) Administration publique en Suisse. L’appareil étatique et le gouvernement. Berne: Haupt. Geser, H., Höpflinger, F., Ladner, A. and Meuli, U. (1996) Die Schweizer Gemeinden im Kräftefeld des gesellschaftlichen und politisch-administrativen Wandels. Zurich: Sofiolo- gisd??. Giddens, A. (1992) Modernity and Self-Identity, Cambridge: Polity Press. Giddens, A. (1994) The Transformation of Intimacy: Sexuality, Love and Eroticism in Modern Societies, Stanford, CA: Stanford University Press. Giesecke, J. and Gross M. (2003) ‘Temporary Employment: Chance or Risk?’ European Sociological Review, vol. 19, no. 2, 161–77. Glotz, P. and Schultze, R.-O. (2002) ‘Reform’, in D. Nohlen and R.-O. Schultze (eds), Lexikon der Politikwissenschaft. Theorien, Methoden, Begriffe, Munich: C.H. Beck, 789–95. Godaert, G., Sorbi, M., Peters, M., Dekkers, C. and Geenen R. (2001) ‘Ambulatory Monitoring of Diurnal Changes in Pain in Chronic Pain Disorder, Migraine and Rheumatoid Arthritis’, in J. Fahrenberg and M. Myrtek (eds), Progress in Ambulatory Assessment: Computer-assisted Psychological and Psychophysiological Methods in Monitoring and Field Studies, Kirkland, WA: Hogrefe & Huber Publishers, pp. 123–8. Goebel, H. and Cordes, P. (1990) ‘Circadian Variation of Pain Sensitivity in Pericranial Musculature’, Headache, vol. 30, no. 7, 418–22. Goldman, N. (2001) Social Inequalities in Health: Disentangling the Underlying Mechanisms, paper presented at the conference on Demography and Epidemiology: Frontiers in Population Health and Aging, Georgetown University, Washington, DC. Gordon, D., Davey-Smith, G., Dorling, D. and Shaw M. (eds) (2000) Inequalities in Health: The Evidence Presented to the Independent Inquiry into Inequalities in Health, Bristol: The Policy Press. Gordon, D. and Pantazis, C. (1997) ‘Measuring Poverty: Breadline Britain in the 1990s’, in D. Gordon and C. Pantazis (eds), Breadline Britain in the 1990s, Aldershot: Ashgate, pp. 5–47. Gosling, S.D., John, O.P., Craik, K.H. and Robings, R.W. (1998) ‘Do People Know How They Behave? Self-Reported Act Frequencies Compared with On-line Codings by Observers’, Journal of Personality and Social Psychology, vol. 74, 1337–49. Gossens, F. and Degelmann, A. (eds) (1952) Die Praxis der Mitarbeiter Auswahl. Bericht über eine Untersuchung, Munich. Gottman, J.M. (1994) What Predicts Divorce? The Relationship Between Marital Processes and Marital Outcomes, Hillsdale, NJ: Lawrence Erlbaum. Gottman, J.M. and Porterfield, A.L. (1981) ‘Communicative Competence in the Nonverbal Behavior of Married Couples’, Journal of Marriage and the Family, vol. 43, 817–24. Goux, D. and Maurin, E. (2000) ‘The Decline in Demand for Unskilled Labour: An Empirical Analysis Method and Its Application to France’, Review of Economics and Statistics, vol. 82, no. 4, 596–607. Gramlich, E.M. (1990) ‘U.S. Federal Budget Deficits and Gramm–Rudman–Hollings’, American Economic Review, Papers and Proceedings, No. 80.2, 75–80. References 309

Grande, E. (1996) ‘Das Paradox der Schwäche: Forschungspolitik und die Einflusslogik europäischer Politikverflechtung’, in M. Jachtenfuchs and B. Kohler-Koch (eds), Europäische Integration, Opladen: Leske & Budrich. Gravelle, H., Wildman, J. and Sutton, M. (2002) ‘Income, Income Inequality and Health: What Can we Learn from Aggregate Data?’, Social Science & Medicine, vol. 54, 577–89. Grebner S., Elfering, A., Semmer, N.K., Kaiser-Probst, C. and Schlapbach M.-L. (2004) ‘Stressful Situations at Work and in Private Life among Young Workers: An Event Sampling Approach’, Social Indicators Research, vol. 67, 11–49. Grebner, S., Semmer, N.K. and Elfering, A. (in press) ‘Working Conditions and Three Types of Well-Being: A Longitudinal Study with Self-report and Rating Data’, Journal of Occupational Health Psychology. Grebner, S., Semmer, N.K., Lo Faso, L., Gut, S., Kälin, W. and Elfering, A. (2003) ‘Working Conditions, Well-being, and Job-related Attitudes Among Call Centre Agents’, European Journal of Work and Organizational Psychology, vol. 12, 341–65. Green Cowles, M., Caporaso, J. and Risse, T. (eds) (2001) Transforming Europe. European- ization and Domestic Change, Ithaca: Cornell University Press. Greenan, N. (2003) ‘Organizational Change, Technology, Employment and Skills: An Empirical Study of French Manufacturing’, Cambridge Journal of Economics, vol. 27, 287–316. Greenan, N. and Guellec, D. (1994) ‘Co-ordination within the Firm and Endogenous Growth’, Industrial and Corporate Change, vol. 3, no. 1, 173–5. Greenberg, J. (1990) ‘Organizational Justice: Yesterday, Today, and Tomorrow’, Journal of Management, vol. 16, 399–432. Greenglass, E.R., Burke, R.J. and Moore, K.A. (2003) ‘Reactions to Increased Workload: Effects of Professional Efficacy of Nurses’, Applied Psychology: An International Review, vol. 52, 580–97. Grob, A., Little, T.D., Wanner, B., Wearing, A.J. and Euronet (1996) ‘Adolescents’ Well-being and Perceived Control Across 14 Sociocultural Contexts’, Journal of Personality and Social Psychology, vol. 71, 785–95. Grob, A., Lüthi, R., Kaiser, F.G., Flammer, A., Mackinnon, A. and Wearing, A.J. (1991) ‘Berner Fragebogen zum Wohlbefinden Jugendlicher (BFW)’, Diagnostica, vol. 37, 66–75. Grote, J. and Schmitter, P. (1999) ‘The Renaissance of National Corporatism: Unintended side – Effect of European Economic and Monetary Union or Calculated Response to the Absence of European Social Policy?’ transfer, no. 1–2, 34–63. Gschwandtner, U., Hersberger, J. and Rauchfleisch, U. (2000) Self-Monitoring of Mental State by Patients with Psychiatric Disorders. A Pilot Study, Research report, Basle: Psychiatrische Universitätspoliklinik PUP. Habermas, J. (1992) Faktizität und Geltung. Diskurstheorie des Rechts und des demokratischen Rechststaats, Frankfurt: Suhrkamp. Haldemann, T. and Schedler, K. (1995) ‘New Public Management-Reformen in der Schweiz – Aktuelle Projektübersicht und erster Vergleich’, in P. Hablützel, T. Haldemann, K. Schedler and K. Schwaar (eds) Umbruch in Politik und Verwaltung, Berne, Stuttgart, Vienna: Haupt, 99–127. Halleröd, B. (1995) ‘The Truly Poor: Indirect and Direct Consensual Measurement of Poverty in Sweden’, Journal of European Social Policy, vol. 5, no. 2, 111–29. Halleröd, B. (1998) ‘Poor Swedes, Poor Britons: A Comparative Analysis of Relative Deprivation’, in H.-J. Andress (ed.), Empirical Poverty Research in a Comparative Perspective, Aldershot: Ashgate, 283–329. 310 References

Hamermesh, D.S. (2001) ‘Demand for Labor’, in N.J. Smelser and P.B. Baltes (eds), International Encyclopedia of the Social and Behavioral Sciences, Amsterdam: Elsevier, pp. 3384–9. Hamilton, D. (2002) ‘Regimes and Regional Governance: the Case of Chicago’, Journal of Urban Affairs, vol. 24, no. 4, 403–23. Haskel, J.E. and Heden, Y. (1999) ‘Computers and the Demand for Skilled Labour: Industry- and Establishment-Level Panel Evidence for the UK’, Economic Journal, vol. 109, 69–79. Haskel, J.E. and Slaughter, M.J. (2002) ‘Does the Sector Bias of Skill-Biased Technical Change ExplainChanging Skill Premia?’, European Economic Review, vol. 46, 1757–83. Hauptmanns, P. (1995) Rationalisierung und Qualifikationsentwicklung. Eine empirische Analyse im deutschen Maschinenbau, Opladen: Leske+Budrich. Haverland, M. (2000) ‘National Adaptation to European Integration: The Importance of Institutional Veto Points’, Journal of Public Policy, vol. 20, no. 1, 83–103. Havighurst, R.J. (1972) Developmental Tasks and Education, New York: McKay. Heinz, W.R., Kelle, U., Witzel, A. and Zinn, J. (1998) ‘Vocational Training and Career Development in Germany – Results from a Longitudinal Study’, International Journal of Behavioral Development, vol. 22, 77–101. Heinz, W. (ed.) (2000) Stadt & Region: Kooperation oder Koordination? Ein internationaler Vergleich. Stuttgart: Kohlhammer. Held, T. and Levy, R. (1974) Die Stellung der Frau in Familie und Gesellschaft, Frauenfeld: Huber, (En français: Femme, famille et société. Enquête sociologique sur la situation en Suisse. Vevey: Delta, 1975). Helweg-Larsen, M., Kjøller, M. and Thoning, H. (2003) ‘Do Age and Social Relations Moderate the Relationship Between Self-Rated Health and Mortality among Adult Danes?’, Social Science & Medicine, vol. 57, no. 7, 1147–1341. Hendrick, S.S. (1988) ‘A Generic Measure of Relationship Satisfaction’, Journal of Marriage and the Family, vol. 50, 93–8. Herfindahl, O.C. (1950) Concentration in the US Steel Industry, Columbia: Columbia University Press. Héritier, A., Kerwer, D., Knill, C., Lehmkuhl, D., Teutsch, M. and Douillet, A.-C. (2001) Differential Europe – EU Impact on National Policymaking, Boulder: Rowman & Littlefield. Hertzman, C. (1999) ‘The Biological Embedding of Early Experience and Its Effects on Health in Adulthood’, Annals of the New York Academy of Sciences, vol. 896, 85–95. Hesse, J.J. and Sharpe, L.J. (1991) ‘Conclusions’, in J.J. Hesse (ed.), Local Government and Urban Affairs in an International Perspective, Nomos: Baden-Baden, 603–21. Hirst, P. and Thompson, G. (1996) Globalization in Question: the International Economy and the Possibilities of Governance, Cambridge: Polity Press. Hix, S. (1999) ‘Dimensions and Alignments in European Union Politics: Cognitive Constraints and Partisan Responses’, European Journal of Political Research, vol. 35, 69–106. Hoffmann, E. and Walwei, U. (1998) ‘Regular Employment Relationship: a Discontinued Model? Thoughts on an Explanatory Model for the Change in Employment Forms’, Mitteilungen aus der Arbeitsmarkt- und Berufsforschung, vol. 31, no. 3, 409–25. Holmes, J.G. and Murray, S.L. (1996) ‘Conflict in Close Relationships’, in E.T. Higgins and A.W. Kruglanski (eds), Social Psychology: Handbook of Basic Principles, New York: Guilford, pp. 622–54. Höpflinger, F., Charles, M. and Debrunner, A. (1991) Familienleben und Berufsarbeit. Zum Wechselverhältnis zweier Lebensbereiche, Zurich: Seismo. References 311

Hosmer, D.W. and Lemeshow, S. (1989) Applied Logistic Regression, New York: Wiley & Sons. House, J.S. and Umberson, D. (1988a) ‘Social Relationships and Health’, Science, vol. 241, 540–5. House, J.S., Umberson, D. and Landis K.R. (1988b) ‘Structures and Processes of Social Support’, Annual Review of Sociology, vol. 14, 293–318. Hoxby, C.M. (2000) ‘Does Competition Among Public Schools Benefit Students and Taxpayers?’, American Economic Review, vol. 90, 1209–1238. Hug, S. and Sciarini, P. (1995) ‘Switzerland – Still a Paradigmatic Case?’ In G. Schneider, P. Weitsman and T. Bernauer (eds), Towards a new Europe: Stops and Starts in Regional Integration, New York: Praeger/Greenwood, pp. 55–74. Hujer, R., Caliendo, M. and Radic, D. (2002) Skill Biased Technological and Organizational Change: Estimating a Mixed Simultaneous Equation Model Using the IAB Establishment Panel, IZA Discussion Paper No. 566, Bonn: IZA. Ickes, W. (1997) ‘Introduction’, in W. Ickes (ed.), Empathic Accuracy, New York: Guilford, pp. 1–16. Ickes, W. (2001) ‘Measuring Empathic Accuracy’, in J.A. Hall and F.J. Bernieri (eds), Interpersonal Sensitivity: Theory and Measurement, Mahwah, NJ: Lawrence Erlbaum, pp. 219–41. Ickes, W., Gesn, P.R. and Graham, T. (2000) ‘Gender Differences in Empathic Accuracy: Differential Ability or Differential Motivation?’, Personal Relationships, no. 7, 95–109. Ickes, W. and Simpson, J.A. (1997) ‘Managing Empathic Accuracy in Close Relationships’, in W. Ickes (ed.), Empathic Accuracy, New York: Guilford Press, pp. 218–50. Jaccard, James (2001) Interaction Effects in Logistic Regression, Beverly Hills: Sage. Jacobebbinghaus, P. and Zwick, T. (2001) New Technologies and the Demand for Medium Qualified Labour in Germany, ZEW Discussion Paper No. 12, Mannheim. Janzen, B.L. and Muhajarine, N. (2003) ‘Social Role Occupancy, Gender, Income Adequacy, Life Stage and Health: a Longitudinal Study of Employed Canadian Men and Women’, Social Science & Medicine, vol. 57, no. 8, 1491–1503. Jesuit, D. and Smeeding, T. (2002) ‘Poverty and Income Distribution’, Luxembourg Income Study Working Paper No. 293, Syracuse University, Maxwell School of Citizenship and Public Affairs, Syracuse. Johnson, G. (1997) ‘Changes in Earnings Inequality: The Role of Demand Shifts’, Journal of Economic Perspectives, vol. 11, no.1, 41–54. Journal of Marriage and the Family (2000), Issue 4. Joye, D., Schuler, M., Nef, R. and Bassand, M. (1988) Le système des communes suisses. Approche typologique du modèle centre-périphérie, Bern: Office fédéral de la statistique. Kaase, M. and Newton, K. (1995) Beliefs in Government, Oxford: Oxford University Press. Kaiser, L.C. (2001a) ‘Permanent Full-Time Job Still Dominant Form of Employment in Europe’, DIW Economic Bulletin, vol. 38, no. 4, 119–24. Kaiser, L.C. (2001b) Standard and Non-Standard Employment Patterns across Europe, EPAG Working Paper 25, University of Essex, Colchester. Kaiser, U. (1999) New Technologies and the Demand for Heterogeneous Labour: Firm-Level Evidence for the German Business-Related Services Sector, ZEW Discussion Paper No. 7, Mannheim. Kälin W. (2004), Coping – Moderator oder Mediator zwischen Stressoren und Befinden? (unveröffentlichte Dissertation), Philosophisch-historische Fakultät, Universität Bern. Kälin, W., Semmer, N.K., Elfering, A., Tschan, F., Dauwalder, J.-P., Heunert, S. and Crettaz von Roten, F. (2000) ‘Work Characteristics and Well-Being of Swiss Apprentices Entering the Labor Market’, Swiss Journal of Psychology, vol. 59, 272–90. 312 References

Kälin, W., Semmer, N.K., Elfering, A., Tschan, F., Grebner, S., Rochat, S. and Inversin, L. (2004) ‘Work Characteristics and Well-Being of Swiss Job Beginners: A Four-Wave Study’, manuscript submitted for publication. Kalleberg, A.L. (2000) ‘Nonstandard Employment Relations: Part-time, Temporary and Contract Work’, Annual Review of Sociology, vol. 26, 341–65. Kalleberg, A.L., Reskin, B.F. and Hudson, K. (2000) ‘Bad Jobs in America: Standard and Nonstandard Employment Relations and Job Quality in the United States’, American Sociological Review, vol. 65, no. 2, 256–78. Kantor, D. and Lehr, W. (1975) Inside the Family, San Francisco: Jossey-Bass. Kaplan, G.A. (1999) ‘Part III Summary: What is the Role of the Social Environment in Understanding Inequalities in Health?’, in N.E. Adler, M. Marmot, B.S. McEwen and J. Stewart (eds), Socioeconomic Status and Health in Industrial Nations: Social, Psychological, and Biological Pathways. Proceedings of the New York Academy of Sciences conference, May 10–12, New York Academy of Sciences, New York, vol. 896, pp. 116–19. Käppler, C. and Rieder, S. (2001) ‘Does the Retrospection Effect Hold as a Stable Phenomenon? First Results from a Trans-Cultural Self-Monitoring Study of Mood and Cognitive States in Brazil and Germany’, in J. Fahrenberg and M. Myrtek (eds), Progress in Ambulatory Assessment: Computer-Assisted Psychological and Psychophysio- logical Methods in Monitoring and Field Studies, Kirkland, WA: Hogrefe & Huber Publishers, pp. 113–22. Karney, B.R. and Bradbury, T.N. (1995) ‘The Longitudinal Course of Marital Quality and Stability: a Review of Theory, Method and Research’, Psychological Bulletin, vol. 118, 3–34. Karney, B.R. and Bradbury, T.N. (2000) ‘Attributions in Marriage: State or Trait? A Growth Curve Analysis’, Journal of Personality and Social Psychology, vol. 78, no. 2, 295–309. Katschnig, H. (1980) Sozialer Stress und psychische Erkrankung, Munich: Urban & Schwarzenberg. Katzenstein, P.J. (1984) Corporatism and Change: Austria, Switzerland, and the Politics of Industry, Ithaca and London: Cornell University Press. Katzenstein, P.J. (1985) Small States in World Market: Industrial Policy in Europe, Ithaca and London: Cornell University Press. Keating, M. (1995) ‘Size, Efficiency and Democracy: Consolidation, Fragmentation and Public Choice’, in D. Judge et al. (eds), Theories of Urban Politics, London: Sage, pp. 117–34. Kellerhals, J. and Montandon, C. (1991) Les stratégies éducatives des familles. Milieu social, dynamique familiale et éducation des pré-adolescents, Lausanne: Delachaux et Niestlé. Kellerhals, J., Perrin, J.-F., Steinauer-Cresson, G., Vonèche, L. and Wirth, G. (1982) Mariages au quotidien. Inégalités sociales, tensions culturelles et organisation familiale, Lausanne: P.-M. Favre. Kellerhals, J., Troutot, P.-Y. and Lazega, E. (1993) Microsociologie de la famille, 2nd edn, Paris: PUF. Kenny, D.A. and Acitelli L.K. (2001) ‘Accuracy and Bias in the Perception of the Partner in a Close Relationship’, Journal of Personality and Social Psychology, no. 80, 439–48. Keohane, R.O. and Milner, H.V (eds) (1996) Internationalization and Domestic Politics, Cambridge: Cambridge University Press. Kern, H. and Schumann, M. (1970) Industriearbeit und Arbeiterbewusstsein, Frankfurt: Europäische Verlagsanstalt. References 313

Kern, H. and Schumann, M. (1984) Das Ende der Arbeitsteilung? Rationalisierung in der industriellen Produktion, Munich: Beck. Kersting, N. and Vetter, A. (eds) (2003) Reforming Local Government in Europe: Closing the Gap between Democracy and Efficiency, Opladen: Leske+Budrich. Kiewiet, D.R. and Szakaly, K. (1996) ‘Constitutional Limitations on Borrowing: An Analysis of State Bonded Indebtedness’, Journal of Law, Economics and Organization, no. 12, 62–97. Kirchgässner, G. (2002a) ‘The Effects of Fiscal Institutions on Public Finance: A Survey of the Empirical Evidence’, in S.L. Winer and H. Shibata (eds), Political Economy and Public Finance: The Role of Political Economy in the Theory and Practice of Public Eco- nomics, Cheltenham: Edward Elgar, pp. 145–77. Kirchgässner, G. (2002b) ‘Föderalismus und Staatsquote’, in U. Wagschal and H. Rentsch (eds), Der Preis des Föderalismus, Zurich: Orell Füssli, pp. 71–91. Kirchgässner, G., Feld, L.P. and Savioz, M.R. (1999) Die direkte Demokratie: Modern, erfolgreich, entwicklungs- und exportfähig, Basle: Helbing and Lichtenhahn. Kirchgässner, G. and Pommerehne, W.W. (1996) ‘Tax Harmonization and Tax Competition in the European Union: Lessons from Switzerland’, Journal of Public Economics, vol. 60, 351–71. Klein, T. and Lauterbach, W. (1994) ‘How Does Women’s Level of Education Affect Family Formation and Labour Force Interruption? An Empirical Analysis of Micro- economic Theory’, Kölner Zeitschrift für Soziologie und Sozialpsychologie, vol. 46, 278–98. Klodt, H., Maurer, R. and Schimmelpfennig, A. (1996) Tertiarisierung in der deutschen Wirtschaft, Tübingen: J.C.B. Mohr. Knill, C. and Lehmkuhl, D. (1999) ‘How Europe Matters. Different Mechanisms of Europeanization’, European Integration online Papers (EIoP), no. 3, pp. 7. http://eiop. or.at/eiop/texte/1998-007a.htm. Kohli, M. (1985) ‘Die Institutionalisierung des Lebenslaufs’, Kölner Zeitschrift für Soziologie und Sozialpsychologie, vol. 37, no. 1, 1–29. Kohli, M. (1986) ‘Gesellschaftszeit und Lebenszeit, Der Lebenslauf im Strukturwandel der Moderne, Die Moderne – Kontinuitäten und Zäsuren’, Soziale Welt, numéro spécial, pp. 183–208. Kohli, M. (2003) ‘Der institutionalisierte Lebenslauf: ein Blick zurück und nach vorn, in J. Allmendinger (eds), Entstaatlichung und soziale Sicherheit. Verhandlungen des 31. Kongresses der DGS, Opladen: Leske+Budrich, pp. 525–45. Kohli, M., Rein, M., Guillemard, A.-M., and van Gunsteren, H. (eds) (1991) Time for Retirement: Comparative Studies of Early Exit from the Labor Force, Cambridge University Press. Kommission für Zukunftsfragen der Freistaaten Bayern und Sachsen (1998) Erwerb- stätigkeit und Arbeitslosigkeit in Deutschland. Entwicklung, Ursachen, Maßnahmen. Leitsätze, Zusammenfassung und Schlußfolgerungen der Teile I, II und III des Kommi- sionsberichts [sic], Munich: Olzog. Kooiman, J. (2003) Governing as Governance, London: Sage. Korpi, W. (2001) ‘Faces of Inequality: Gender, Class, and Patterns of Inequalities in Different Types of Welfare States’, Social Politics: International Studies in Gender, State and Society, vol. 7, no. 2, 127–91. Kress, U. (1998) ‘From the Regular Employment Relationship to the Flexibilisation of the Labour Market – a Literature Report’, Mitteilungen aus der Arbeitsmarkt- und Berufsforschung, vol. 31, no. 3, 488–505. Kriesi, H. (1980) Entscheidungsstrukturen und Entscheidungsprozesse in der Schweizer Politik, Frankfurt: Campus Verlag. 314 References

Kriesi, H. (ed.) (1996) Le clivage linguistique: problèmes de compréhension entre les communautés linguistiques en Suisse, Berne: Office fédéral de la statistique. Kriesi, H. (1998) Le système politique suisse, Paris: Economica. Kriesi, H., Armingeon, K., Siegrist, H. and Wimmer, A. (eds) (1999) Nation and National Identity: the European Experience in Perspective, Chur: Verlag Rüegger. Krueger, A.B. (1993) ‘How Computers Have Changed the Wage Structure: Evidence from Micro Data, 1984–1989’, Quarterly Journal of Economics, vol. 108, no. 1, 33–60. Krüger, H. (2001) ‘Social Change in Two Generations: Employment Patterns and Their Costs for Family Life’, in V.W. Marshall, W.R. Heinz, H. Krüger and A. Verma (eds), Restructuring Work and the Life Course, Toronto: University of Toronto Press, pp. 401–23. Krüger, H. and Levy, R. (2001) ‘Linking Life Courses, Work and the Family: Theorizing a Not So Visible Nexus Between Women and Men’, Canadian Journal of Sociology, vol. 26, no. 2, 145–66. Kübler, D. (2003) ‘“Metropolitan governance” oder: die unendliche Geschichte der Institutionenbildung in Stadtregionen’, Informationen zur Raumentwicklung, vol. 9, no. 8, 535–41. Kübler, D. and Heinelt, H. (2004) ‘Metropolitan Governance, Democracy and the Dynamics of Place’, in H. Heinelt and D. Kübler (eds) Metropolitan Governance: Capacity, Democracy and the Dynamics of Place, London: Routledge. Kübler, D. and Schwab, B. (2004) ‘New Regionalism in Five Swiss Metropolitan Areas. An Assessment of Inclusiveness, Deliberation and Democratic Accountability’, European Journal of Political Research, submitted. Kübler, D., Schwab, B., Joye, D. and Bassand, M. (2002) La métropole et le politique. Identité, services urbains et citoyennté dans quatre agglomérations en Suisse, Lausanne: LaSUR/EPFL. Kübler, D. and Wälti, S. (2001) ‘Drug-Policy Making in Metropolitan Areas: Urban Conflicts and Governance’, International Journal of Urban and Regional Research, vol. 25, no. 1, 34–54. Kugler, P. and Spycher, S. (1992) ‘Der Einfluss des Technologiewandels auf die Struktur der Arbeitsnachfrage in der Schweiz von 1950–1988’, Schweizerische Zeitschrift für Volkswirtschaft und Statistik, vol. 128, no. 4, 617–41. Künzler, J. and Walter, W. (1999) ‘Division of Household Labour. The Role of Welfare States, Gender Roles and Economic Resources’, paper, ESA Conference, Amsterdam. Kurer, O. and Pohl, R. (1999) ‘Globalisierung und die Misere der Gesellschaft’, in G. Schmidt and R. Trinczek (eds), Globalisierung. Ökonomische und soziale Herausforderungen am Ende des zwanzigsten Jahrhunderts (Soziale Welt, Sonderband 13), Baden-Baden: Nomos, pp. 41–54. Lachat, R. and Kriesi, H. (2003) ‘Globalization and the Transformation of the National Political Space’, paper presented at the Political Change and Globalization conference, in Rethymnon, Crete. Ladner, A. (1991a) Politische Gemeinden, kommunale Parteien und lokale Politik. Eine empirische Untersuchung in den Gemeinden der Schweiz, Zurich: Seismo. Ladner, A. (1991b) ‘Direkte Demokratie auf kommunaler Ebene – Die Beteiligung an Gemeindeversammlungen’, Schweizerisches Jahrbuch für Politische Wissenschaft, vol. 31: 63–86. Ladner, A. (2002) ‘Size and Direct Democracy at the Local Level: the Case of Switzerland’, Environment and Planning C: Government and Policy, vol. 20, no. 6, 813–28. References 315

Ladner, A., Arn, D., Friederich, U., Steiner, R. and Wichtermann, J. (2000) Gemein- dereformen zwischen Handlungsfähigkeit und Legitimation. Forschungsbericht des Schweizerischen Nationalfonds, Berne: IOP. Ladner, A. and Steiner, R. (1998) Gemeindereformen in den Schweizer Kantonen. Konzeptionelle Grundlagen und empirische Ergebnisse einer Kantonsbefragung. Arbeits- bericht Nr. 28 des IOP der Universität Bern, Berne: IOP. Lafitte, P. (1987) Les institutions de démocratie directe en suisse au niveau local, Lausanne: Cahiers de l’Idheap. Lahelma, E., Arber, S., Kivelä, K. and Roos, E. (2002) ‘Multiple Roles and Health among British and Finnish Women: the Influence of Socioeconomic Circumstances’, Social Science & Medicine, vol. 54, 727–40. Lang, T. (1994) ‘Berufliche Selbstselektion von Hochschulabsolventen und ihre Folgen auf Einstellungen zur Arbeit’, in L. von Rosenstiel, T. Lang and E. Sigl (eds), Fach- und Führungsnachwuchs finden und fördern, Stuttgart: Schäffer-Poeschel, pp. 202–20. Larson, R. and Richards, M. (1994) Divergent Realities: The Emotional Lives of Mothers, Fathers and Adolescents, New York: Basic Books. Laumann, E.O. and Knoke, D. (1987) The Organizational State: Social Choice in National Policy Domains, Madison, WI: The University of Wisconsin Press. Lavenex, S. and Uçarer, E.M. (eds) (2002) Migration and the Externalities of European Integration, Lanham, MD: Lexington Books. Law, I., Rozun, I. and Rozun, E. (2004) IZY-Builder. Retrieved, from the World Wide Web: http://www.izibuilder.com Layte, R., Maître, B., Nolan, B. and Whelan, C.T. (2000) Explaining Levels of Deprivation in the European Union, EPAG Working Paper, vol. 12, Dublin: The Economic and Social Research Institute. Lebo, R.B., Harrington, T.F. and Tillmann, R. (1995) ‘Work Values Similarities among Students from Six Countries’, Career Development Quarterly, vol. 43, 350–62. Lefèvre, C. (1998) ‘Metropolitan Government and Governance in Western Countries: A Critical Review’, International Journal of Urban and Regional Research, vol. 22, no. 1, 9–25. Lefèvre, C. (2001) ‘Metropolitan Goverment Reform in Europe: Current Trends and Challenges’, Swiss Political Science Review, vol 7, no. 3, 136–41. Le Galès, P. (1995) ‘Du gouvernement des villes à la gouvernance urbaine’, Revue française de science politique, vol. 45, no. 1, 57–95. Le Galès, P. (1998) ‘Regulations and Governance in European Cities’, International Journal of Urban and Regional Research, vol. 48, 482–506. Leibfried, S., Leisering, L., Buhr, P., Ludwig, M., Mädje, E., Olk, T., Voges, W. and Zwick, M. (1995) Zeit der Armut. Lebensläufe im Sozialstaat, Frankfurt: Suhrkamp. Leresche, J.-P. (2001) ‘Gouvernance et coordination des politiques publiques’, in J.-P. Leresche (ed.), Gouvernance locale, coopération et légitimité. Le cas suisse dans une perspective comparée, Paris: Pédone, pp. 29–65. Leu, R.E., Burri, S. and Priester, T. (1997) Lebensqualität und Armut in der Schweiz, Berne: Haupt. Lev, B. (2003) The Measurement of Firm-specific Organizational Capital, NBER Working Paper No. 9581, Cambridge, MA: NBER. Levy, R. (1977) Der Lebenslauf als Statusbiographie, Stuttgart: Enke. Levy, R., Widmer, E. and Kellerhals, J. (2002) ‘Modern Family or Modernized Family Traditionalism? Master Status and the Gender Order in Switzerland’, Electronic Journal of Sociology, vol. 6, no. 4. Levy, R., Joye, D., Guye, O. and Kaufmann, V. (1997) Tous égaux? De la stratification aux représentations, Zurich: Seismo. 316 References

Lind, J. and Møller, I.H. (eds) (1999) Inclusion and Exclusion: Unemployment and Non-standard Employment in Europe, Aldershot: Ashgate. Lindbeck, A. and Snower, D.J. (2000) ‘Multi-Task Learning and the Reorganization of Work: From Tayloristic to Holistic Organization’, Journal of Labour Economics, vol. 18, 353–76. Linder, W. (1994) Swiss Democracy: Possible Solutions to Conflict in Multi-Cultural Societies, New York: St Martin’s Press. Linder, W. (1999) Schweizerische Demokratie. Institutionen, Prozesse, Perspektiven, Berne: Haupt. Lipsmeier, G. (1999) ‘Die Bestimmung des notwendigen Lebensstandards – Einschät- zungs-unterschiede und Entscheidungsprobleme’, Zeitschrift für Soziologie, vol. 28, no. 4, 281–300. Littek, W., Heisig, U. and Gondek, H.-D. (eds) (1991) Dienstleistungsarbeit. Struktur- veränderungen, Beschäftigungsbedingungen und Interessenlagen, Berlin: Edition Sigma. Lombard, M., Snyder-Duch, J. and Bracken, C.C. (2002) ‘Content Analysis in Mass Communication: Assessment and Reporting of Intercoder Reliability’, Human Communication Research, vol. 28, 587–604. Louis, M.R. (1980) ‘Surprise and Sense Making: What Newcomers Experience in Entering Unfamiliar Organizational Settings’, Administrative Science Quarterly, vol. 25, 226–51. Lowery, D. (1999) ‘Answering the Public Choice Challenge: a Neoprogressive Research Agenda’, Governance, vol. 12, no. 1, 29–55. Lowery, D., DeHoog, H. and Lyons, W.E. (1992) ‘Citizenship in the Empowered Locality. An Elaboration, a Critique and a Partial Test’, Urban Affairs Quarterly, vol. 28, no. 1, 69–103. Luginbühl, M., Baumberger, T., Schwab, K. and Burger, H. (2002) Medientexte zwischen Autor und Publikum. Intertextualität in Presse, Radio und Fernsehen, Zurich: Seismo. Luhmann, N. (1982) Liebe als Passion: Zur Codierung von Intimität, Frankfurt: Suhrkamp. Luhmann, N. (2000) Soziale Systeme. Grundriss einer allgemeinen Theorie. Frankfurt am Main: Suhrkamp. Lynch, J. and Kaplan, G. (2000) ‘Socioeconomic Position’, in L.F. Berkman and I. Kawachi (eds), Social Epidemiology, New York: Oxford University Press, pp. 13–35. Mach, A. (1999) ‘L’articulation entre facteurs externes et internes comme clef d’analyse des réformes économiques et sociales’, in A. Mach (ed.), Globalisation, néo-libéralisme et politiques publiques dans la Suisse des années 1990, Seismo, Zürich. Mach, A. (2003) La Suisse entre internationalisation et changements politiques internes: Législation sur les cartels et relations industrielles dans les années 1990, Basle: Helbling & Lichtenhahn. Mach, A., Häusermann, S. and Papadopoulos, Y. (2003) ‘Economic Regulatory Reforms in Switzerland: Adjustments Without European Integration or how Rigidities Become Flexible’, Journal of European Public Policy, vol. 10, no. 2, 302–19. Machin, S. and Van Reenen, J. (1998) ‘Technology and Changes in Skill Structure: Evidence from Seven OECD Countries’, Quarterly Journal of Economics, vol. 113, no. 4, 1215–43. Machin, S. and van Reenen, J. (1998) Technology and Changes in the Skill Structure: Evidence from Seven OECD-Countries, London: The Institute for Fiscal Studies. Mack, J. and Lansley, S. (1985) Poor Britain, London: George Allen & Unwin. Macran, S., Clarke, L. and Joshi, H. (1996) ‘Women’s Health: Dimensions and Differentials’, Social Science & Medicine, vol. 42, no. 9, 1203–16. Mair, P. (2000) ‘The Limited Impact of Europe on National Party Systems’, West European Politics, vol. 23, no. 4, 27–51. References 317

March, J.G. and Olsen, J.P. (1989) Rediscovering Institutions: The Organizational Basis of Politics, New York: The Free Press. Marelli, B. (1985) Arbeitsvermittlung und Stellensuche in der Schweiz, Grüsch: Rüegger. Margraf, J., Taylor, C.B., Ehlers, A. and Roth, W.T. (1987) ‘Panic Attacks in the Natural Environment’, Journal of Nervous and Mental Disease, vol. 175, no. 9, 558–65. Marini, M.M. (1987) ‘Measuring the Process of Role Change During the Transition to Adulthood’, Social Science Research, vol. 16, 1–38. Marmot, M. (2000) ‘Multilevel Approaches to Understanding Social Determinants’, in L.F. Berkman and I. Kawachi (eds), Social Epidemiology, Oxford: Oxford University Press, pp. 349–67. Marmot, M. and Wilkinson, R. (eds) (2000) Social Determinants of Health, Oxford: Oxford University Press. Marmot, M. and Wilkinson, R.G. (2001) ‘Psychosocial and Material Pathways in the Relation between Income and Health: a Response to Lynch et al’, British Medical Journal (BMJ), vol. 322, pp. 1233–6. Matsusaka, J.G. (1995) ‘Fiscal Effects of the Voter Initiative: Evidence from the Last 30 Years’, Journal of Political Economy, vol. 103, 587–623. Matsusaka, J.G. (2002) For the Many or the Few: How the Initiative Process Changes American Government, Chicago: University of Chicago Press. Mayer, K.-U. (2000) ‘Arbeit und Wissen: Die Zukunft von Ausbildung und Beruf’, in J. Kocka and K. Offe (eds), Geschichte und Zukunft der Arbeit, Frankfurt a.M: Campus, pp. 383–409. Mayer, K.-U. (2001) ‘The Paradox of Global Social Change and National Path Depend- encies: Life Course Patterns in Advanced Societies’, in A.E. Woodward and M. Kohli (eds), Inclusions and Exclusions in European Societies, London: Routledge, pp. 89–110. McDonough, P. and Berglund, P. (2001) Histories of Poverty and Self-rated Health Trajectories, paper presented at the Annual Meeting of the British Sociological Association Medical Sociology Group Joint Meeting of Working Groups I and II of the European Science Foundation Scientific Program, London. Melchior, M., Berkman, L.F., Niedhammer, I., Chea, M. and Goldberg, M. (2003) ‘Social Relations and Self-Reported Health: a Prospective Analysis of the French Gazel Cohort’, Social Science & Medicine, vol. 56, no. 8, 1817–30. Mertens, D. (1974) ‘Schlüsselqualifikationen: Thesen zur Schulung für eine moderne Gesellschaft’, Mitteilungen aus Arbeitsmarkt und Berufsforschung, vol. 7, pp. 36–43. Michel G. (submitted), Daily Patterns of Symptom Reporting in Families with Adolescent Children. Mikula, G. and Freudenthaler, H. (2002) ‘Division of Tasks and Duties and the Perception of Injustice: The Case of Household Chores’/Verteilung von Aufgaben und Pflichten und Ungerechtigkeitswahrnehmung: Eine Analyse am Beispiel Haushaltsarbeit, Psychologische Beiträge, vol. 44, no. 4, 567–84. Milgrom, P. and Roberts, J. (1990) ‘The Economics of Modern Manufacturing’, American Economic Review, vol. 80, no. 3, 511–28. Miller, A.R., Treiman, D.J., Cain, P.S. and Roos, P.A. (eds) (1980) Work, Jobs and Occupations: A Critical Review of the Dictionary of Occupational Titles, Washington: National Academy Press. Mills, M. and Blossfeld, H.-P. (2003) ‘Globalization, Uncertainty and Changes in Early Life Course’, Zeitschrift für Erziehungswissenschaft, vol. 6, 188–218. Moen, P. and Han, S.-K. (2001) ‘Reframing Careers: Work, Family and Gender’, in V.W. Marshall, W.R. Heinz, H. Krüger and A. Verma (eds), Restructuring Work and the Life Course, Toronto: University of Toronto Press, pp. 424–45. 318 References

Moen, P. (ed.) (2003) It’s about Time: Couples and Careers. Ithaca: Cornell University Press. Morand, C.-A. (1987) ‘La formation et la mise en oeuvre du droit’, Pouvoir, vol. 43, 73–86. Moravcsik, A. (1994) Why the European Community Strengthens the State: Domestic Politics and International Cooperation, Working Paper 52, Cambridge, MA: Harvard University Press. Moravcsik, A. (1998a) ‘Does International Cooperation Strengthen National Executives? The Case of Monetary Policy in the European Union’, paper presented at the Third Project Workshop, Europeanization and Domestic Political Change, in The European University Institute, Robert Schuman Centre, Florence, Italy. Moravcsik, A. (1998b) The Choice for Europe: Social Purpose and State Power from Messina to Maastricht, Ithaca, NY: Cornell University Press. Moss, N.E. (2002) ‘Gender Equity and Socioeconomic Inequality: a Framework for the Patterning of Women’s Health’, Social Science & Medicine, vol. 54, 649–61. Moulin, S. (2003) Decline of the Permanent and Full-Time Employment Model and Gender Discrimination in France, Aix-en-Provence: Laboratoire d’Economie et de Sociologie du Travail (LEST). Mouritzen, P.E. (1989) ‘City Size and Citizen’s Satisfaction’, European Journal of Political Research, vol. 17, 661–88. Mückenberger, U. (1985) ‘Die Krise des Normalarbeitsverhältnisses’, Zeitschrift für Sozialreform, vol. 31, no. 7, 415–75. Müller, S.G. (2000) Bestimmungsgründe der prozyklischen Arbeitsproduktivität. Theorie und Empirie für das produzierende Gewerbe der Bundesrepublik Deutschland zwischen 1960 und 1994, Hamburg: Universität Hamburg. Müller, W. (2001) ‘Zum Verhältnis von Bildung und Beruf in Deutschland. Entko- ppelung oder zunehmende Strukturierung?’, in P.A. Berger and D. Konietzka (eds), Die Erwerbsgesellschaft. Neue Ungleichheiten und Unsicherheiten, Opladen: Leske & Budrich. Murray, S.L., Holmes, J.G., Bellavia, G., Griffin, D.W. and Dolderman, D. (2002) ‘Kindred Spirits? The Benefits of Egocentrism in Close Relationships’, Journal of Personality and Social Psychology, vol. 82, no. 4, 563–81. Murray, S.L., Holmes, J.G. and Griffin D.W. (1996) ‘The Benefits of Positive Illusions: Idealization and the Construction of Satisfaction in Close Relationships’, Journal of Personality and Social Psychology, vol. 70, 79–98. Mutz, G., Ludwig-Mayerhofer, W., Koenen, E.J., Eder, K. and Bonß, W. (1995) Diskon- tinuierliche Erwerbsverläufe. Analysen zur postindustriellen Arbeitslosigkeit, Opladen: Leske & Budrich. Myrdal, A. and Klein, V. (1956) Women’s Two Roles: House and Work, London: Routledge & Kegan Paul. Nahrstedt, W. (1992) Die Entstehung der Freizeit, Göttingen: Vandenhoeck & Ruprecht. Nave-Herz, R. (1997) Familie heute. Wandel der Familienstrukturen und Folgen für die Erziehung, Darmstadt: Primus. Neidhart, L. (1970) Plebiszit und pluralitäre Demokratie, Eine Analyse der Funktionen des schweizerischen Gesetzesreferendum, Berne: Francke. Nelson, D.L. and Burke, R.J. (2002) ‘A Framework for Examining Gender, Work Stress, and Health’, in D.L. Nelson and R.L. Burke (eds) Gender, Work Stress, and Health, Washington, DC: American Psychological Association, pp. 3–14. Nelson, D.L., Quick, J.C. and Eakin, M.E. (1988) ‘A Longitudinal Study of Newcomer Role Adjustment in US Organizations’, Work and Stress, no. 2, 239–53. References 319

Nelson, D.L., Quick, J.C., Eakin, M.E. and Matuszek, P.A.C. (1995) ‘Beyond Organizational Entry and Newcomer Stress: Building a Self-Reliant Workforce’, International Journal of Stress Management, no. 2, 1–14. Nicholson, N. (1984) ‘A Theory of Work Role Transitions’, Administrative Science Quarterly, no. 29, 172–91. Nicholson, N. and West, M.A. (1988) Managerial Job Change: Men and Women in Transition, Cambridge: Cambridge University Press. Nicolet, S., Sciarini, P. and Fischer, A. (2003) ‘In the Shadow of Direct Democracy: A Quantitative Analysis of Decision-Making Processes in Switzerland’, paper presented at the Drei-Länder-Tagung, at the University of Berne, Switzerland. Noller, P. (1984) Nonverbal Communication and Marital Interaction, Oxford: Pergamon Press. Norgaard, A.-S. (1996) ‘Rediscovering Reasonable Rationality in Institutional Analysis’, European Journal of Political Research, vol. 19, no. 1, 31–57. Nurmi, J.-E., Poole, M.E. and Seginer, R. (1995) ‘Tracks and Transitions – A Comparison of Adolescent Future-Oriented Goals, Explorations, and Commitments in Australia, Israel, and Finland’, International Union of Psychological Science, vol. 30, 356–75. O’Brien, G.E., Feather, N.T. and Kabanoff, B. (1994) ‘Quality of Activities and the Adjustment of Unemployed Youth’, Australian Journal of Psychology, vol. 46, 29–34. O’Brien Caughy, M., O’Campo, P.J. and Muntaner, C. (2003) ‘When Being Alone Might be Better: Neighborhood Poverty, Social Capital, and Child Mental Health’, Social Science & Medicine, vol. 57, no. 2, 227–37. OECD (1988) ‘Technology and Employment’, Employment Outlook, vol. 6, 173–93. OECD (1992) ‘Recent Labour Market Developments and Prospects’, Employment Out- look, vol. 10, 3–40. OECD (1994) ‘Job Gains and Job Losses in Firms’, Employment Outlook, vol. 12, 103–34. OECD (1996) Technology, Productivity and Job Creation (Vol.2: Analytical Report), Paris: OECD. OECD (1998) Technology, Productivity and Job Creation. Best Policy Practices, Paris: OECD. OECD (2000) ‘Employment in the Service Economy: A Reassessment’, Employment Outlook, vol. 18, 79–128. OECD (2004) OECD Economic Studies – Switzerland, Paris: OECD. Oesch, D. (2003) ‘Labour Market Trends and the Goldthorpe Class Schema: A Conceptual Reassessment’, Swiss Journal of Sociology, vol. 29, no. 2, 241–62. Olson, D.H. and McCubbin, H.I. (1989) Families: What Makes Them Work?, 2nd edn, Beverly Hills, CA: Sage. O’Neill, N. and O’Neill, G. (1972) Open Marriage, New York: Evans. O’Reilly J. (1996) ‘Labour Adjustments through Part-time Work’, in G. Schmid, J. O’Reilly and K. Schömann (eds) International Handbook of Labour Market Policy and Evaluation, Cheltenham: Edward Elgar. Organization for Economic Co-Operation and Develoment (1998) OECD Data on Skills: Employment by Industry and Occupation, STI Working Paper 1998/4, Paris: OECD. Organization for Economic Co-operation and Development (OECD) (1999) OECD Employment Outlook, Paris: OECD. Oschmiansky, H. and Schmid, G. (2000) Wandel der Erwerbsformen. Berlin und die Bundesrepublik im Vergleich, Discussion Paper FS I 00-204, Berlin: WZB. Osterland, M. (1990) ‘“Normalbiographie” und “Normalarbeitsverhältnis”’, in P.A. Berger and S. Hradil (eds) Lebenslagen, Lebensverläufe, Lebensstile, Göttingen: Otto Schwartz, pp. 351–62. 320 References

Ostrom, E. (1972) ‘Metropolitan Reform: Propositions Derived from Two Traditions’, Social Science Quarterly, vol. 53, 474–93. O’Sullivan, A., Sexton, T.A. and Sheffrin, S.M. (1995) Property Taxes and Tax Revolts: The Legacy of Proposition 13, Cambridge: Cambridge University Press. O’Sullivan, A., Sexton, T.A. and Sheffrin S.M. (1999) ‘Proposition 13: Unintended Effects and Feasible Reforms’, National Tax Journal, no. 52, 99–111. Parker, S. and Wall, T. (1998) Job and Work Design, London: Sage. Parsons, T. and Bales, R.F. (1955) Family: Socialization and Interaction Process, New York: The Free Press. Peltzman, S. (1992) ‘Voters as Fiscal Conservatives’, Quarterly Journal of Economics, vol. 107, 327–61. Pensola, T.H. and Martikainen, P. (2003) ‘Research Report: Cumulative Social Class and Mortality from Various Causes of Adult Men’, Journal of Epidemiology and Community Health, vol. 57, 745–51. Perrez, M., Berger, R. and Wilhelm, P. (1998) ‘Die Erfassung von Belastungserleben und Belastungsverarbeitung in der Familie: Self-Monitoring als neuer Ansatz’/ Assessment of stress and coping in the family: Self-monitoring as a new approach, Psychologie in Erziehung und Unterricht. Perrez, M. and Reicherts, M. (1996) ‘A Computer-Assisted Self-Monitoring Procedure for Assessing Stress-Related Behavior under Real Life Conditions’, in J. Fahrenberg and M. Myrtek (eds) Progress in Ambulatory Assessment: Computer-Assisted Psychological and Psychophysiological Methods in Monitoring and Field Studies, Kirkland, WA: Hogrefe & Huber Publishers, pp. 51–67. Perrez, M. and Wilhelm, P. (2000) ‘Control Psychology under the Control of Questionnaires? The Search for an Alternative Assessment Procedure’, in A. Grob (ed.) Control of Human Behavior, Mental Processes, and Consciousness: Essays in Honor of the 60th Birthday of August Flammer, Mahwah, NJ: Lawrence Erlbaum Associates, pp. 245–61. Perrez, M., Wilhelm, P. and Schoebi, D. (2000) ‘How to Assess Social Regulation of Stress and Emotions in Daily Family Life? A Computer-Assisted FAmily SElf-Monitoring System (FASEM-C)’, Clinical Psychology and Psychotherapy, vol. 7, 326–39. Perrez, M., Wilhelm, P., Schoebi, D. and Horner, M. (2001) ‘Simultaneous Com- puter-Assisted Assessment of Causal Attribution and Social Coping in Families’, in J. Fahrenberg and M. Myrtek (eds) Progress in Ambulatory Assessment: Computer-Assisted Psychological and Psychophysiological Methods in Monitoring and Field Studies, Kirkland, WA: Hogrefe & Huber Publishers, pp. 25–43. Perrig-Chiello, P. and Höpflinger, F. (2001) Zwischen den Generationen, Zurich: Seismo. Pfau-Effinger, B. (1999) Kultur, Wohlfahrtsstaat und Frauenarbeit im europäischen Vergleich, Leverkusen: Leske+Budrich. Pierson, P. (2000) ‘Increasing Returns, Path Dependency, and the Study of Politics’, American Political Science Review, vol. 94, no. 2, 251–67. Piva, M., Santarelli, E. and Vivarelli, M. (2003) The Sill Bias Effect of Technological and Organizational Change: Evidence and Policy Implications, IZA Discussion Paper No. 934, Bonn. Pommerehne, W.W. (1978) ‘Institutional Approaches to Public Expenditure: Empirical Evidence from Swiss Municipalities’, Journal of Public Economics, no. 9, 255–80. Pommerehne, W.W. (1983) ‘Private versus öffentliche Müllabfuhr – nochmals betrachtet’, Finanzarchiv, no. 41, pp. 466–75. Pommerehne, W.W., Hart, A. and Feld, L.P. (1997) ‘Steuerhinterziehung und ihre Kontrolle in unterschiedlichen politischen Systemen’, Homo oeconomicus, no. 14, 469–87. References 321

Pommerehne, W.W. and Weck-Hannemann, H. (1996) ‘Tax Rates, Tax Administration and Income Tax Evasion in Switzerland’, Public Choice, no. 88, 161–70. Poterba, J.M. (1997) ‘Do Budget Rules Work?’, in: A.J. Auerbach (ed.), Fiscal Policy: Lessons from Economic Research, Cambridge, MA: MIT Press, pp. 53–86. Putnam, R.D. (1988) ‘Diplomacy and Domestic Politics: The Logic of Two-Level Games’, International Organization, vol. 42, no. 3, 427–60. Quick, J.C. and Quick, J.D. (1984) Organizational Stress and Preventive Management. New York: McGraw-Hill. Radaelli, C.M. (2001) ‘The Domestic Impact of European Union Public Policy: Notes on Concepts, Methods, and the Challenge of Empirical Research’, Politique européenne, no. 5, 107–42. Radloff, L.S. (1977) ‘The CES-D Scale: A Self-Report Depression Scale for Research in the General Population’, Applied Psychological Measurement, no. 3, 385–401. Rasbash, J., Browne, W., Goldstein, H.Y.M., Plewis, I., Healy, M., Woodhouse, G. and Draper, D. (1999) A User’s Guide to Mlwin, retrieved from the World Wide Web: http://multilevel.ioe.ac.uk Redford, E. (1969) Democracy in the Administrative State, New York: Oxford University Press. Reicherts, M. and Pihet, S. (2000) ‘Job Newcomers Coping with Stressful Situations: A Micro-Analysis of Adequate Coping and Well-Being’, Swiss Journal of Psychology, no. 59, 303–16. Reiss, D. (1971) ‘Varieties of Consensual Experience’, Family Process, no. 10, 1–35. Reiss, D. (1981) The Family’s Construction of Reality, Cambridge, MA: Harvard University Press. Reitzle, M., Vondracek, F.W. and Silbereisen, R.K. (1998) ‘Timing of School-to-Work Transitions: A Developmental-Contextual Perspective’, International Journal of Behav- ioral Development, no. 22, 7–28. Reskin, B. and Padavic, I. (1994) Women and Men at Work, Thousand Oaks, CA: Pine Forge Press. Reyher, L., Spitznagel, E. and Kretschmer, G. (1990) ‘Das gesamtwirtschaftliche Stellenangebot’, Mitteilungen aus der Arbeitsmarkt- und Berufsforschung, vol. 23, no. 3, 347–72. Rhodes, M. (1998) ‘Globalization and the Welfare State: The Emergence of Competitive Corporatism’, Swiss Political Science Review, vol. 4, no. 1, 99–107. Rhodes, R.A.W. (1996) ‘The New Governance: Governing without Government’, Political Studies, vol. 44, no. 4, 652–67. Rimann, M., Udris, I. and Weiss, V. (2000) ‘Values in Transition from Apprenticeship to Occupational Work’, Swiss Journal of Psychology, vol. 59, 291–302. Ringen, S. (1988) ‘Direct and Indirect Measures of Poverty’, Journal of Social Policy, vol. 17, no. 3, 351–65. Risse, T., Cowles, M.G. and Caporaso, J. (2001) ‘Europeanization and Domestic Change: Introduction’, in M. Green Cowles, J. Caporaso and T. Risse (eds), Trans- forming Europe: Europeanization and Domestic Change, Ithaca: Cornell University Press. pp. 1–20. Rodden, J. (2003) Reviving Leviathan: Fiscal Federalism and the Growth of Government, mimeo, Department of Political Science, MIT, Cambridge. Rodgers, G. and Rodgers, J. (eds) (1989) Precarious Jobs in Labour Market Regulation: The Growth of Atypical Employment in Western Europe, Geneva: ILO. Rook, K.S. (1992) ‘Detrimental Aspects of Social Relationships: Taking Stock of an Emerging Literature’, in H.O.F. Veiel and U. Bauman (eds), The Meaning and 322 References

Measurment of Social Support, New York and Washington: Hemisphere Publishing Corporation, pp. 157–69. Rosenstiel, L. von and Nerdinger, F.W. (2000) ‘Die Münchner Wertestudien – Bestandsaufnahme und (vorläufiges) Resümee’, Psychologische Rundschau, no. 51, 146–57. Ross, N.A., Wolfson, M.C., Dunn, J.R., Berthelot, J.-M., Kaplan, G.A. and Lynch, J.W. (2000) ‘Relation Between Income Inequality and Mortality in Canada and in the United States: Cross-Sectional Assessment using Census Data and Vital Statistics’, BMJ, vol. 320, 898–902. Rossi, A.A. (1995) Concurrence territoriale et réseaux urbains. L’armature de la Suisse en transition, Zurich: Vdf. Rossi, G. (2003) ‘La variabilità delle forme familiari in Europa’, in G. Rossi (ed.), La Famiglia In Europa, Rome: Carocci, pp. 13–42. Roubini, N. and Sachs, J.D. (1989) ‘Political and Economic Determinants of Budget Deficits in the Industrial Democracies’, European Economic Review, vol. 33, 903–38. Roussel, L. (1980) ‘Mariages et divorces: contribution à une analyse systématique des modèles matrimoniaux’, Population, no. 6, 1025–40. Roy, K.-B. (1989) ‘Erosion des Normalarbeitsverhältnisses und die Konsequenzen für den Sozialstaat’, Sozialer Fortschritt, vol. 38, no. 11/12, 273–9. Rubin, D.B. (1987) Multiple Imputation for Non-Response in Surveys, New York: John Wiley. Rusk, D. (1995) Cities without Suburbs, 2nd edn, Washington, DC: Woodrow Wilson Center. Sacchi, S. (2003) ‘Diffusionsmuster von Computerarbeit in der Schweiz – Eine Forsc- hungsnotiz’, Schweizerische Zeitschrift für Soziologie, vol. 29, no. 3. Sacchi, S. and Kriesi, I. (forthcoming), Berufliche Arbeitsmarktsegmentation in der Schweiz, 1970–1990. Sacchi, S., Salvisberg, A. and Buchmann, M. (2004) ‘Long-Term Dynamics in Employers’ Demand for Skills’, Chapter 5 in this volume. Sacker, A., Bartley, M., Firth, D. and Fitzpatrick, R. (2001) ‘Dimensions of Social Inequality in the Health of Women in England: Occupational, Material and Behavioural Pathways’, Social Science & Medicine, vol. 52, no. 5, 763–81. Salvisberg, A., Klarer, U. and Sacchi, S. (2001) ‘Ungleicher Einsatz von Homepages in der Schweizer Wirtschaft’, Die Volkswirtschaft, vol. 74, no. 7, 54–60. Sanders, M., and ter Weel, B. (2000) Skill-Biased Technical Change: Theoretical Concepts, Empirical Problems and a Survey of the Evidence, DRUID Working Paper No. 8, Copen- hagen: DRUID. Savitch, H. and Vogel, R.K. (2000) ‘Paths to New Regionalism’, State and Local Government Review, vol. 32, no. 3, 158–68. Schaltegger, C.A. (2001) ‘The Effects of Federalism and Democracy on the Size of Government: Evidence from Swiss Sub-national Jurisdictions’, ifo Studien, no. 47, 145–62. Schaltegger, C.A. (2002) ‘Budgetregeln und ihre Wirkung auf die öffentlichen Haushalte: Empirische Ergebnisse aus den US-Bundesstaaten und den Schweizer Kantonen’, Schmollers Jahrbuch, no. 122, 369–413. Scharpf, F.W. (1997) Games Real Actors Play: Actor-Centered Institutionalism in Policy Research, Boulder, CO and Oxford: Westview Press. Scharpf, F.W. (1998) ‘Globalization: The Limitations on State Capacity’, Swiss Political Science Review, vol. 4, no. 1, 92–8. References 323

Scharpf, F.W. (1999) Governing in Europe: Effective and Democratic?, Oxford: Oxford University Press. Scharpf, F.W., Reissert, B. and Schnabel, F. (eds) (1977) Politikverflechtung II: Kritik und Berichte aus der Praxis, Kronberg: Athenäum Verlag. Schedler, K. and Proeller, I. (2000) New Public Management, Berne, Stuttgart and Vienna: Haupt. Scherer, K.R. (1988) ‘Criteria for Emotion-Antecedent Appraisal: A Review’, in V. Hamilton, G.H. Bower and N.H. Frijda (eds), Cognitive Persprctives on Emotion and Motivation, Dordrecht: Martinus Nijhoff, pp. 89–126. Schmidt, M.G. (1985) Der Schweizer Weg zur Vollbeschäftigung, Frankfurt: Campus. Schmidt, M.G. (1995) ‘Vollbeschäftigung und Arbeitslosigkeit in der Schweiz. Vom Sonderweg zum Normalfall’, Politische Vierteljahresschrift, vol. 36, no. 1, 35–48. Schmidt, J.U. (1984) ‘Stabilität der Interessen von Oberschülern nach Ende der Schulzeit’, Psychologie und Praxis, no. 28, 26–31. Schnell, R. (1997) Nonresponse in Bevölkerungsumfragen: Ausmass, Entwicklung und Ursachen, Opladen: Leske+Budrich. Schoebi, D. (2004) Konfliktregulation im Alltag von Familien. Konflikte in Familien als Prozesse sozialer Belastungsbewältigung, Berlin: Tenea. Schoon, I., Sacker, A. and Bartley, M. (2003) ‘Socio-Economic Adversity and Psychosocial Adjustment: a Developmental-Contextual Perspective’, Social Science & Medicine, vol. 57, no. 6, pp. 1001–15. Schuler, M. (1994) Die Raumgliederungen der Schweiz/Les niveaux géographiques de la Suisse, Berne: Bundesamt für Statistik. Schuler, M. (1999) Régionalisation et urbanisation: des concepts convergents?, Département d’Architecture, Ecole polytechnique fédérale de Lausanne, Lausanne. Sciarini, P. (1992) ‘La Suisse dans la négociation sur l’Espace économique européen: de la rupture à l’apprentissage’, Annuaire suisse de science politique, no. 32, 297–322. Sciarini, P. (1994) La Suisse face à la Communauté européenne et au GATT. Le cas-test de la politique agricole, Geneva: Georg. Sciarini, P., Dupont, C. and Hug, S. (2000) ‘Example, Exception or Both? Swiss National Identity in Perspective’, in L.-E. Cederman (ed.), Constructing Europe’s Identity: The external dimension, Boulder, CO: Lynne Rienner. Sciarini, P., Nicolet, S. and Fischer, A. (2002) ‘L’impact de l’internationalisation sur les processus de décision en Suisse’, Revue Suisse de Science Politique, vol. 8, no. 3/4, 1–34. Sciarini, P., Fischer, A. and Nicolet, S. (2004) ‘How Europe Hits Home: Evidence from the Swiss Case’, Journal of European Public Policy, vol. 11, no. 3, 353–78. Seeman, T.E. and Crimmins, E. (2001) ‘Social Environment Effects on Health and Aging: Integrating Epidemiologic and Demographic Approaches and Perspectives’, Annals of the New York Academy of Sciences, vol. 954, 88–117. Segalman, R. (1986) The Swiss Way of Welfare: Lessons for the Western World, New York: Praeger. Seifert, H. (ed.) (1993) Jenseits der Normalarbeitszeit. Perspektiven für eine bedürfnisgerechtere Arbeitszeitgestaltung, Cologne: Bund Verlag. Semmer, N., Baillod, J., Stadler, R. and Gail, K. (1996) ‘Fluktuation bei Computerfach- leuten: Eine follow-up Studie’, Zeitschrift für Arbeits- und Organisationspsychologie, vol. 40, 190–9. Semmer, N.K., Grebner, S. and Elfering, A. (2004) ‘Beyond Self-Report: Using Observational, Physiological, and Situation-Based Measures in Research on Occu- pational Stress’, in P.L. Perrewé and D.C. Ganster (eds), Research in Occupational Stress and Well-Being, Amsterdam: Elsevier, vol. 3, pp. 205–63. 324 References

Semmer, N.K. and Jacobshagen N. (2003) ‘Selbstwert und Wertschätzung als Themen der arbeitspsychologischen Stressforschung’, in K.-C. Hamborg and H. Holling (eds), Innovative Personal- und Organisationsentwicklung, Göttingen: Hogrefe, pp. 131–55. Semmer, N.K., McGrath, J.E. and Beehr, T.A. (in press) ‘Conceptual Issues in Research on Stress and Health’, in C.L. Cooper (ed.), Handbook of Stress and Health, 2nd edn, New York: CRC Press. Semmer, N. and Schallberger, U. (1996) ‘Selection, Socialization, and Mutual Adaptation: Resolving Discrepancies Between People and Work’, Applied Psychology: An Interna- tional Review, vol. 45, 263–88. Sen, A.K. (1981) Poverty and Famines: An Essay on Entitlement and Deprivation, Oxford: Clarendon Press. Sengenberger, W. (1987) Struktur und Funktionsweise von Arbeitsmärkten. Die Bun- desrepublik Deutschland im internationalen Vergleich, Frankfurt: Campus. Sennett, R. (1998) The Corrosion of Character: The Personal Consequences of Work in the New Capitalism, New York: W.W. Norton & Company. Shadbegian, R.J. (1999) ‘Fiscal Federalism, Collusion, and Government Size: Evidence from the States’, Public Finance Review, vol. 27, 262–81. Sheldon, G. (1996) Die berufliche Flexibilität im Spiegel der Zeit, Bern: Bundesamt für Statistik. Sheldon, G. (2003) ‘Die Auswirkungen der Ausländerbeschäftigung auf die Löhne und das Wirtschaftswachstum in der Schweiz’, in H.-R. Wicker, R. Fibbi and W. Haug (eds), Migration und die Schweiz, Zurich: Seismo, pp. 335–67. Sidjanski, D. (1966) ‘Les groupes de pression et la politique étrangère en Suisse’, Annuaire Suisse de Science Politique, vol. 6, 28–45. Siegrist, J. (2002) ‘Effort–Reward Imbalance at Work and Health’, in P.L. Perrewé and D.C. Ganster (eds), Research in Occupational Stress and Well-Being, Amsterdam: JAI, vol. 2, pp. 261–91. Simpson, B. (1998) Changing Families: An Ethnographic Approach to Divorce and Separation, Oxford and New York: Berg. Singer, B. and Ryff, C.D. (1999) ‘Hierarchies of Life Histories and Associated Health Risks’, Annals New York Academy of Sciences, vol. 896, 96–115. Singlemann, J. (1978) From Agriculture to Services: The Transformation of Industrial Employment, London: Sage. Smith, R.E., Leffingwell, T.R. and Ptacek, J.T. (1999) ‘Can People Remember How They Coped? Factors Associated with Discordance Between Same-Day and Retrospective Reports’, Journal of Personality and Social Psychology, vol. 76, no. 6, 1050–61. Smith, V. (1997) ‘New Forms of Work Organization’, Annual Review of Sociology, vol. 23, 315–39. SOEP Group (2001) ‘The German Socio-Economic Panel (GSOEP) after more than 15 years – Overview’, Quarterly Journal of Economic Research, vol. 70, no. 1, 7–14. Spenner, K. (1988) ‘Technological Change, Skill Requirements, and Education: The Case for Uncertainty’, in R.M. Cyert and D.C. Mowery (eds), The Impact of Technological Change on Employment and Economic Growth, Cambridge, MA: Ballinger, pp. 131–84. Stamm, H., Lamprecht, M. and Nef, R. (2003) Soziale Ungleichheit in der Schweiz. Struk- turen und Wahrnehmungen, Zurich: Seismo. Stauffer, T.P. (2001) Instrumente des Haushaltsausgleichs: Ökonomische Analyse und rechtli- che Umsetzung, Basle: Helbing und Lichtenhahn. Steinberg, J. (1988) ‘Imitation of Switzerland: Historical Reflections’, Government and Opposition, no. 23, 13–31. References 325

Steiner, R. (2000) ‘New Public Management in the Swiss Municipalities’, International Public Management Journal, vol. 3, no. 2, 169–89. Steiner, R. (2002) Interkommunale Zusammenarbeit und Gemeindezusammenschlüsse in der Schweiz, Erklärungsansätze, Umsetzungsmöglichkeiten und Erfolgsaussichten, Berne, Stuttgart and Vienna: Haupt. Stoker, G. (1998) ‘Cinq propositions pour une théorie de la gouvernance’, Revue internationale des sciences sociales, LX, vol. 1, pp. 19–30. Stoker, G. (2000) ‘Urban Political Science and the Challenge of Urban Governance’, in I. Pierre (ed.), Debating Governance, Oxford: Oxford University Press, pp. 91–109. Stone, A.A., Broderick, J.E., Porter, L.S. and Kaell, A.T. (1997) ‘The Experience of Rheu- matoid Arthritis Pain and Fatigue: Examining Momentary Reports and Correlates over One Week’, Arthritis Care & Research, vol. 10, no. 3, 185–93. Stone, A.A., Shiffman, S., Schwartz, J.E., Broderick, J.E. and Hufford, M.R. (2002) ‘Patient Non Compliance with Paper Diaries’, BMJ, vol. 324, pp. 1193–94. Streeck, W. and Schmitter, P. (1991) ‘From National Corporatism to Transnational Pluralism: Organised Interests in the Single European Market’, Politics and Society, vol. 19, no. 2, pp. 133–64. Streuli, E. (2003) ‘In Ermangelung finanzieller Ressourcen: Privatverschuldung in der Schweiz’, Swiss Journal of Sociology, vol. 29, no. 2, 293–317. Stutzer, A. and Frey, B.S. (2000) ‘Stärkere Volksrechte, zufriedenere Bürger, eine mikroökonometrische Untersuchung’, Swiss Political Science Review, vol. 6, 1–30. Suter, C. (ed.) (2000) Sozialbericht 2000, Zurich: Seismo. Suter, C. (ed.) (2004). Sozialbericht 2004, Zurich: Seismo. Suter, C., Iglesias, K. and Oegerle, T. (2003) ‘Soziale Beziehungen und Soziale Unterstüt- zung’, in E. Zimmermann and R. Tillman (eds), ‘Vivre en Suisse 1999–2000’. Une année dans la vie des ménages et familles en Suisse, Neuchâtel: Lang Verlag, pp. 131–57. Suter, C. and Pahud, C. (eds) (2000) Rapport social 2000, Zurich: Seismo. Suter, C. and Paris, D. (2002) ‘Inequality and Deprivation: Switzerland in a three Country Comparison’, Swiss Journal of Sociology, vol. 28, no. 2, 217–240. Suter, C., Renschler, I. and Joye, D. (eds) (2004) Rapport social 2004, Zurich: Seismo. Taris, T.W., Peeters, M.C.W., Le Blanc, P.M., Schreurs, P.J.G. and Schaufeli W.B. (2001) ‘From Inequity to Burnout: The Role of Job Stress’, Journal of Occupational Health Psychology, vol. 6, pp. 303–23. Terry, D.J. and Jimmieson, N.L. (1999) ‘Work Control and Employee Well-Being: A Decade Review’, in C.L. Cooper and I.T. Robertson (eds), International Review of Industrial and Organizational Psychology, vol. 14, Chichester: Wiley, pp. 95–148. Thoits, P. (1982) ‘Conceptual, Methodological, and Theoretical Problems in Studying Social Support as a Buffer Against Life Stress’, Journal of Health and Social Behavior, vol. 23, 145–59. Thoits, P. (1995) ‘Stress, Coping, and Social Support Processes: Where Are We? What Next?’, Journal of Health and Social Behavior, Extra Issue, 53–79. Thom, N., Blum A. and Zaugg R.J. (1996) Ausgewählte Trends der Personalmanagements in der Schweiz, Berne: Institut für Organisation und Personal der Universität Bern. Thom, N. and Zaugg, R. (1995) Personalbeschaffung und Personalauswahl in schweizerischen Unternehmungen, Berne: Institut für Organisation und Personal der Universität Bern. Tiebout, C.M. (1956) ‘A Pure Theory of Local Expenditures’, Journal of Political Economy, vol. 44, 416–24. Torgler, B. (2002) Tax Morale and Institutions, mimeo, University of Basel. Touraine, A. (1969) La société post-industrielle, Paris: Denoël. 326 References

Townsend, P. (1979) Poverty in the United Kingdom: A Survey of Household Resources and Standards of Living, Penguin Books, London. Traxler, F. and Schmitter, P. (1995) ‘The Emerging Euro-Polity and Organised Interests’, European Journal of International Relations, vol. 1, no. 2, 191–218. Trechsel, A. and Serdült, U. (1999) Kaleidoskop Volksrechte: Die Institutionen der direkten Demokratie in den schweizerischen Kantonen (1970–1996), Basle: Helbing and Lichtenhahn, Basel. Trinczek, R. (1999) ‘“Es gibt sie, es gibt sie nicht, es gibt sie, es . . . ” - Die Globalisierung der Wirtschaft im aktuellen sozialwissenschaftlichen Dsikurs’, in Schmidt G., Trinczek R. (eds.), Globalisierung. Ökonomische und soziale Herausforderungen am Ende des zwanzigsten Jahrhunderts (Soziale Welt, Sonderband 13), Nomos, Baden-Baden, pp. 55–75. Tschan, F., Semmer, N.K., Inversin, L. (2004), Work related and ‘private’ social interactions at work, Social Indicators Research, vol. 67, pp. 145–182. Ulich, E. (2001), Arbeitspsychologie (5. Aufl.), vdf, Zurich. Umberson, D. (1987), ‘Family Status and Health Behaviors: Social Control as a Dimension of Social Integration’, Journal of Health and Social Behavior, vol. 28, pp. 396–319. US Employment Service (1977), Dictionary of Occupational Titles, National Academy Press, Washington D.C. Van den Berg, L., Van Klink, A. and Van der Meer, J. (eds) (1993) Governing Metropolitan Regions, Aldershot: Avebury. Van den Bosch, K. (1998) ‘Perceptions of the Minimum Standard of Living in Belgium: Is There a Consensus?’, in H.-J. Andress (ed.), Empirical Poverty Research in a Comparative Perspective, Aldershot: Ashgate, pp. 135–66. Vatter, A. and Freitag, M. (2002) ‘Die Janusköpfigkeit von Verhandlungsdemokratien: Zur Wirkung von Konkordanz, direkter Demokratie und dezentraler Entschei- dungsstrukturen auf den öffentlichen Sektor der Schweizer Kantone’, Swiss Political Science Review, vol. 8, 53–80. Veit-Wilson, J. (1987) ‘Consensual Approaches to Poverty Lines and Social Security’, Journal of Social Policy, vol. 16, no. 2, 183–211. Vickery, G. and Wurzburg, G. (1998), The Challenge of Measuring and Evaluating Organizational Change in Enterprises, Paris: OECD. Wagner, A. (2000) ‘Krise des “Normalarbeitsverhältnisses”? Über eine konfuse Debatte und ihre politische Instrumentalisierung’, in C. Schäfer (ed.), Geringere Löhne – mehr Beschäftigung? Niedriglohn-Politik, Hamburg: VSA-Verlag, pp. 200–46. Wagner, G. (1998) ‘Teilzeitbeschäftigung zu Lasten der Vollzeiterwerbstätigkeit?’ DIW Wochenbericht, vol. 65, no. 44. Wallerstein, J.S. and Kelly, J.B. (1980) Surviving the Breakup: How Children and Parents Cope with Divorce, New York: Basic Books. Walters, V., McDonough, P. and Strohschein, L. (2002) ‘The Influence of Work, Household Structure, and Social, Personal and Material Resources on Gender Differences in Health: an Analysis of the 1994 Canadian National Population Health Survey’, Social Science & Medicine, vol. 54, 677–92. Walwei, U. and Werner, H. (1995) ‘An International Comparison of the Development of Part-time Employment: Causes, Labour Market Effects and Consequences’, Mitteilungen aus der Arbeitsmarkt- und Berufsforschung, vol. 28, no. 3, 365–82. Wang, T., Delahanty, D.L., Dougall, A.L. and Baum, A. (1998) ‘Responses of Natural Killer Cell Activity to Acute Laboratory Stressors in Healthy Men at Different Times of Day’, Health Psychology, vol. 17, no. 5, 428–35. Warnken, J. and Ronning, G. (1989) Bestimmungsgründe betrieblicher Faktorsubstitution, Berlin: de Gruyter. References 327

Wasserman, S. and Faust, K. (1994) Social Network Analysis: Methods and Applications, Cambridge: Cambridge University Press. Waterman, M.S. (1995) Introduction to Computational Biology: Maps, Sequences and Genomes, London: Chapman & Hall. Watson, D., Wiese, D., Vaidya, J. and Tellegen, A. (1999) ‘The Two General Activation Systems of Affect: Structural Findings, Evolutionary Considerations, and Psychobio- logical Evidence’, Journal of Personality and Social Psychology, vol. 76, no. 5, 820–38. Weiner, B. (1987) ‘The Social Psychology of Emotion: Applications of a Naive Psy- chology’, Journal of Social and Clinical Psychology, vol. 5, no. 4, 405–19. Wellmann, B., Yuk-lin Wong, R., Tindell, D. and Nazer, N. (1997) ‘A Decade of Network Change: Turnover, Persistence and Stability in Personal Communities’, Social Networks, vol. 19, 27–50. Werner, R. (1999) ‘Untersuchung der Stellenanzeigen mit Methoden der Text- und Inhaltsanalyse’, in A. Laszlo and H. Bau (eds), Wandel beruflicher Anforderungen. Der Beitrag des BIBB zum Aufbau eines Früherkennungssystems Qualifikationsentwicklung, Bielefeld: Bertelsmann, pp. 61–72. Wicki, M. (1998) Qualifikationsbedarf auf dem Arbeitsmarkt. Zusatzauswertung einer Firmenbefragung in den Regionen Biel-Seeland und Oberaargau, Bern: Koodinationsstelle für Weiterbildung der Universität Bern. Wicksell, K. (1896) Finanztheoretische Untersuchungen nebst Darstellung and Kritik des Steuerwesens Schwedens, Jena: Gustav Fischer. Widmer, E., Kellerhals, J. and Levy, R. (2003) Couples contemporains – Cohésion, régulation et conflits, Zurich: Seismo. Widmer, E.D., Kellerhals, J. and Levy, R. (2004a) ‘Types of Conjugal Networks, Conjugal Conflict and Conjugal Quality’, European Sociological Review, vol. 20, no. 1, 63–77. Widmer, E.D., Kellerhals, J. and Levy, R. (2004b) ‘Quelle pluralisation des relations familiales? Conflits, styles d’interactions conjugales et milieu social’, Revue française de Sociologie, vol. 45, no. 1, 37–67. Widmer, E.D., Levy, R., Hammer, R., Pollien, A. and Gauthier, J.-A. (2003) ‘Une analyse exploratoire des insertions professionnelles et familiales: trajectoires de couples résidant en Suisse’, Revue suisse de Sociologie, vol. 29, no. 1, 35–67. Wilhelm, P. (2001) ‘A Multilevel Approach to Analyze Ambulatory Assessment Data: An Examination of Family Members’ Emotional States in Daily Life’, in J. Fahrenberg and M. Myrtek (eds), Progress in Ambulatory Assessment: Computer-Assisted Psychological and Psychophysiological Methods in Monitoring and Field Studies, Kirkland WA: Hogrefe & Huber Publishers, pp. 173–89. Wilhelm, P. (2004) Empathie im Alltag von Paaren. Akkuratheit und Projektion bei der Einschätzung des Befindens des Partners, Bern: Huber. Wilhelm, P. and Perrez, M. (2004) ‘How is My Partner Feeling in Different Daily-Life Settings? Accuracy of Spouses’ Judgments about their Partners’ Feelings at Work and at Home’, Social Indicators Research, vol. 67, 183–246. Wilkinson, R. (1996) Unhealthy Societies: the Afflictions of Inequality, London: Routledge. Wilkinson, R.G., Kawachi, I. and Kennedy, B.C. (1998) ‘Mortality, the Social Environ- ment, Crime and Violence’, in M. Bartley, D. Blane and G.D. Smith (eds), The Sociology of Health Inequalities, Oxford: Blackwell, pp. 19–38. Willitts, M., Benzeval, M. and Stansfeld, S. (2004) ‘Research Report: Partnership History and Mental Health Over Time’, Journal of Epidemiology and Community Health, vol. 58, 53–8. Wood, A. (1994) North–South Trade, Employment and Inequality: Changing Fortunes in a Skill-Driven World, Oxford: Clarendon Press. 328 References

Wood, A. (1995) ‘How Trade Hurts Unskilled Workers’, Journal of Economic Perspectives, vol. 9, 57–80. World Bank (2001) World Development Indicators 2001, CD-Rom, Washington DC. Online at: World Bank: www.worldbank.org/data/countrydata/countrydata.html World Health Organisation, http://www.who.int/country/en/ WZB (Social Science Research Centre Berlin) (various years), EuroModule – Towards a European Welfare Survey, Codebook and Data Table, Berlin: WZB. Yen, I.H. and Syme, S.L. (1999) ‘The Social Environment and Health: a Discussion of the Epidemiologic Literature’, Annual Review of Public Health, vol. 20, 287–308. Zachert, U. (1988) ‘Die Zerstörung des Normalarbeitsverhältnisses’, Arbeit und Recht, vol. 36, no. 5, 129–37. Zimmermann, E. and Burton-Jeangros, C. (2004) ‘Changing Health in Switzerland – 1999–2000’, in E. Zimmermann and R. Tillmann (eds), Vivre en Suisse, 1999–2000, Bern: Peter Lang AG, pp. 331–55. Zürn, M. (1998) ‘Schwarz–Rot–Grün–Braun: Reaktionsweisen auf Denationalisierung’, in U. Beck (ed.), Politik der Globalisierung, Frankfurt: Suhrkamp, pp. 297–330. Index

Notes: f = figure; n = note; t = table; bold = extended discussion or heading emphasized in main text.

‘absolute necessities’ 13–14, 16 Appenzell Innerrhoden (AI) 283 Acemoglu, D. 160(n1–2) appreciation at work 167f, 168, 169f, Adams, P. et al. (2002) 100, 297 179, 180, 187–9 Hurd, M.D. 297 see also social support McFadden, D. 297 apprenticeship 111, 120, 164, 199 Merrill, A. 297 apprentices, trainees 195, Ribeiro, T. 297 209–17t adolescents 61–80 see also education/vocational ‘AEQUAS’ project (‘Work Experience and Arber, S. 87 Quality of Life in Switzerland’) Arbuckle, J.L. 223 163, 164 ‘arena of emotions’ (Larson and age 21, 22, 24, 27n, 30(n7), 31t, 34t, Richards) 61, 62 36t, 42, 66, 84, 87, 89f, 90, 91, 92t, Argyle, M. 11 97–100, 190–1, 194, 197–205, Ariès, P. 62 273–6, 276t Armingeon, K. 227, 314 0–20 years 286, 295, 296t Arn, D. 315 16–24 years 204 Arvanitis, S. xi, 5, 109, 125, 139, 140, 16–64 years 196f, 198f, 199, 202, 160(n3), 161(n7) 203f, 205, 209–10t, 212t, Arvanitis, S. et al. (2000) 106, 214t, 215t 127(n1), 298 18–74 years 280 Bezzola, M. 298 25–55 years 193, 199, 202, 204–5, Donzé, L. 298 210–11t, 213t, 215–17t Hollenstein, H. 298 55–64 years 199, 204 Arvanitis, S. et al. (2001) 106, 65+ years 286, 295, 296t 111, 298 ‘increasing age negatively affects Bezzola, M. 298 general health’ 94 Donzé, L. 298 see also birth cohorts Hollenstein, H. 298 Agglomerationen/agglomérations Marmet, D. 298 (agglomerations) see metropolitan Arvanitis, S. et al. (2003) 106, 298 areas Donzé, L. 298 Ágh, A. 222 Hollenstein, H. 298 Albania 82 Marmet, D. 298 AMOS programme (linear structural Staib, D. 298 equations) 90, 100 Asendorpf, J.B. 66 Anderson, J.J. 222 Athey, S. 149 Andeweg, R. 221 Atkinson, A.B. 10 anxiety 25, 27n, 32t, 35t, 37t Atkinson, A.B. et al. (1995) 10, Aoki, M. 138 30(n7), 298 Appelbaum, E. 192 Rainwater, L. 298 Appenzell Ausserrhoden (canton) 284, Smeeding, T.M. 298 286, 296t Atkinson scale 30(n7)

329 330 Index attitudes 169–70 Bender, S. 157t positive attitude towards life 169–70, Benz, A. 278(n4–5) 170f, 174–6, 178, 184, Berger, P.A. et al. (1993) 207(n3), 300 186f, 189 Sopp, P. 300 resigned attitude towards work 169–70, Steinmüller, P. 300 170f, 173–5, 178, 179, 179f, Berglund, P. 98 181f, 182, 185f, 186f, Bergman, M.M. 302 187, 187f Bergstrom, T.C. 292 Austria 12, 29(n2), 67, 111, 281 Berkman, L.F. 84, 85 ‘absolute necessities’ 17f, 17 Berman, E. et al. (1998) 137, factorial analysis for well-being 33t 160(n2), 300 GNP per capita PPP 12, 30(n3) Bound, J. 300 income equality 30(n4) Machin, S. 300 individual and societal well-being Bern (canton) 243, 294(n6) (and their determinants) 23–5, Bernhardt, A. 192 26–7t, 34–7t Bernhardt, A. et al. (2001) 207(n3), 300 measuring relative deprivation 15 Handcock, M.S. 300 relative deprivation 19–20, 20f, 29 Morris, M. 300 standard of living 18f, 18–19, 28 Scott, M.A. 300 Autor, D.H. 109, 127(n2) Besley, T. 291 Autor, D.H. et al. (1998) 160(n2), 298 ‘betweenness centrality’ 232f, Katz, L.F. 298 238(n23) Krueger, A.B. 298 Bezzola, M. 298 Autor, D.H. et al. (2003) 126, 137, 298 biographization (Kohli) 39 Levy, F. 298 birth cohorts 53, 54t, 55, Murane, R.J. 298 56, 60(n9) bivariate analysis 273, 274, 275 Bäck, H. 266n Black, S.E. 143 back pain 173 Blickle, G. 164 Baethge, M. 107, 109 Blomberg, S.B. 293 Bagnasco, A. 278(n4) Blossfeld, H.P. 188, 191 Baillod, J. 181, 323 Blum, A. 113 balanced budgets 282, 284, Bodenmann, G. 74 294(n2) Bohn, H. 282 bank clerks 164, 165t, 166, 172–6 Bonß, W. 318 banks/insurance (sector) 130t, 162t Börzel, T.A. 222, 238(n24) Barankay, I. 293 Bound, J. 300 Barcelona 258 Boylan, R.D. 127(n3) Bartholomé, K. xviii Bracken, C.C. 128(n6) Bartley, M. 322 Braverman, H. 107 Bartolini, S. 221 Brennan, G. 282 Basel City (canton) 295n Bresnahan, T.F. 137 Basel-Land (canton) 243 Bresnahan, T.F. et al. (2002) 137, Bassand, M. 314 155t, 301 Bauer, T.K. 157t Brynjolfsson, E. 301 Bauman, Z. 39 Hitt, L.M. 301 Baumgartner, A.D. 197 Breton, A. 294(n12) Bavaria 207(n1) Brody, L.R. 73 Beck, U. 22, 39, 191, 207(n4) Browne, W 321 Beck-Gernsheim, E. 39 Bruggemann, A. 170 Becker, R. 127(n3) Brussels [meaning ‘EU’] 238(n25) Beehr, T.A. 324 Bryk, A.S. 67 behaviour 62–3, 73 Brynjolfsson, E. 301 behaviour sampling 64–6 Buchanan, J.M. 282 Bell, D. 107, 109 Bucher, A. 305 Bellavia, G. 318 Buchmann, M. xi, 5, 108 Index 331

Buchmann, M. et al. (1999) 127, 302 CATI 88, 280 König, M. 302 CESifo Network xii, xiii Li, J.H. 302 changes in organizations and Sacchi, S. 302 institutions: six explanatory budget deficits 283–6, 289t, 290, approaches (March and 291, 294, 294(n5, n9), 296t Olsen) 255(n9) national 281, 294(n1) conflict 247, 248, 249f Budowski, M. xi, 5 contagion 248, 249f Budowski, M. et al. (2002) 83, 302 experimental learning 247, Bergman, M.M. 302 248, 249f Tillmann, R. 302 problem-solving 247, 249f Buehler, C. 62 turnover 248, 249f bureaucracy, Weberian trust in 257 variation and selection 247, 249f bureaucratization 39 Chennels, L. 160(n2) Bürgerkommune (Germany) 252 chi squares 54t, 101 Burgess, E.W. 40, 41, 59(n3) child care 178 Burgess, J. 192 children 30(n7), 40, 44, 53–6, 84, Burke, R.J. 176, 178 177, 193 Burton-Jeangros, C. 97 pre-school 80 business, export-oriented 232f see also adolescents differential empowerment 233–4, Christoph, B. 12, 23 235, 238(n24–6) chronologization (Kohli) 39 business-to-business services 110 Chun, H. 160(n2) business cycle 1, 128(n7) cities 293 business services (sector) 162t annexation of suburbs 258 Byrne, B.M. 223 see also core cities citizen loyalty towards elected local Caliendo, M. 311 government 256, 257, 261, California 291 262–3, 274–5, 276t, 278, Canada 194, 258 279(n16) Canada Statistics 161(n5) hypothesis: new regionalism increases cantonal banks 286 citizens’ loyalty towards local cantons 232f, 236(n8–9), 240, authorities 263 241, 251f, 265, 268, 270–2t, hypothesis ‘not supported’ 275 274, 281–6, 291–3, citizen participation 252, 257, 283 294(n13), 295 citizen satisfaction 293 ‘debt breaks’ 291 citizenry/citizens 251f importance 250 preferences 294(n12) ‘lower debts and deficits’ 282, 284 purposes 291 parliamentary system 283 citizens and the state 256, 261, 293 semi-direct democracy 283 democratic legitimacy 263–4 Capelli, P. 155t focus on metropolitan capital (factor) 234 service-users 262 capital mobility 206 loyalty towards elected capital use costs 143, 161(n8) government 262–3 capitalist societies, economic citizenship 274 organization 57 ‘communitarian’ conception 262 Caporaso, J. 309 ‘liberal’ view 262, 269 car/private car 16, 17f, 18f, 28, Clark, A.E. 11 264, 268 class, social 22, 25, 27n, 31t, 34–5t, cardiovascular disease 168 37t, 79, 91 Caroli, E. 138, 155t clothes 17f, 18f, 161t Caroli, E. et al. (2001) 156t, 302 cluster analysis 49t, 53, 58 Greenan, N. 302 Ward’s method 48 Guellec, D. 302 Cohen’s Kappa 128(n6) Carter, W. 155t cohesion (concept) 48, 49t, 53, 57 332 Index collaboration networks 231, 238(n23) fusion 48, 49t, 50, 59 collective labour agreements 190 internal orientation 48, 49t, 50 collinearities 100–1 limited number of styles 47–53, 55, Comment, T. et al. (1996) 195, 303 58–9, 60(n7–8, n10) Hulliger, B. 303 master status 48, 49t, 50, 56, 57, 59 Ries, A. 303 openness/closure 48, 49t, 50, 57, 79 commissions of ‘experts’ 224, 236(n8), relational roles 48, 49t, 50 237(n16) routinization 48, 49t, 50 Communautés urbaines 256, 258 sex-typing 48, 59 communes 265, 266t, 269, 270–2t, six dimensions 48 278–9(n9) see also couples (types) economic and fiscal disparities conjugal models 266, 267 ‘limited set’ hypothesis 41 fiscal autonomy 279(n11) conjugal problems 51, 52t, 60(n7–8) inter-communal commissions 268 conjugal quality 51–3, 57, 60(n11) inter-communal cooperation 271t conjugal relationships 5, 40 periurban 266, 266t, 273t, 274, 276, Consensual Deprivation Index (CDI) 279(n10), 280t (Mack and Lansley) 14, 16 rich 266, 266t, 273t, 274, 276, construction (sector) 130t, 135, 141, 279(n10), 280t 141t, 162t, 241, 244 size (population) 256, 273t, 275t, consultants 115, 132t, 251f 276–7 consultation 144, 237(n10), small 275 237(n16) suburban 266, 266t, 273t, 274, 276, formal 224–7, 234–6 279(n10), 280t frequency 226, 235 transfer payments from higher informal 224–7, 234–5, 237(n10) authority 279(n11) consumer demand 109, 110 types 277 consumption 10, 11, 12, 16, 28–9 typology 279(n10) see also poverty measures/direct see also municipalities Contemporary Switzerland: Revisiting the communication 279(n16) Special Case communications science 4 comparative perspective 1 commuting 264, 268 contributors xi–xvii, 4–6 complementarity 136, 145 labour market xviii, 1, 5, 103–217 ‘complementary system’ 138–9 outside observers 1 computer investment 137 political institutions xviii, 1, 5–6, computer work 112 219–96 diffusion 115, 118, 120, 122, way of life xviii, 1, 4–5, 7–101 123, 129(n12) see also Zukunft Schweiz – Demain la inability to handle 127 Suisse see also ICT; IT contracts computerization 127(n2) municipal 242 COMRES (Computer Recording ‘pay-for-service’ 267 System) 64 work 181, 195, 201, 206 conflict 66, 69f, 74, 79, 247, 248, 249f conventions of interaction 261 interaction behaviour 74–5 cooks 165t, 166, 172f, 172, 173f, 174f, interpersonal (intra- and 175f, 175, 176 extra-familial) 75 coping/coping behaviour 63, 64, 65t, conflict behaviour 65, 66, 80, 165 functional and dysfunctional 75–7 Cordes, P. 71 functionality hypotheses 67 core business 193, 206 conjugal functioning 57 core cities 258, 265–76, 279(n10), 280t conjugal interactions core countries 111 decisional power 49t, 50 cost-cutting 193, 206, 252, 253f functional roles 48, 49t, 50 Costa Rica 82 Index 333

Council of States decision-making process 221, 223, 234, (Switzerland) 237(n13) 235, 236, 236(n6), 239, 240, 244, couples 40, 42, 59, 62, 80 252, 254 general: cohesion and conflicts 41; consultation procedures 224, degree of autonomy 41 236–7(n9) types: associative 49t, 50, 51–5, 57; impact of internationalization on bastion 49t, 50, 51–5, 57; institutions 224–7, 236–7(n7–13) cocoon 49t, 50, 51–5, 57; impact of internationalization on level companionship 49t, 50, of conflict 227–30, 51–5, 79; parallel 49t, 237(n14–19) 50–1, 51–5 parliamentary 223, 225, 226t, 226, see also conjugal interactions 228, 237(n13, n20) Coutts, E. 190n pre-parliamentary 223, 224, 225, covariance 167n 226t, 234, 237(n11) Cramer, U. 114 pre-parliamentary (administrative credit [financial] 84 sub-phase) 226t, 226, Crettaz von Roten, F. 311 237(n13) crime 25, 27n, 32t, 35t, 37t pre-parliamentary (consultation Cuba 82, 85 sub-phase) 225–6, 226t cultural institutions 268–9, 270–2t, referendum 223, 225, 226t, 229, 279(n12, n15) 237(n13, n16–17) Cummins, R.A. 11 ‘referendum threat’ 224, 236(n7) Cunha, A. 278(n8) Deeke, A. 113 customer orientation 254 DeHoog, H. 316 customer surveys 240 democratic dilemma (Dahl) 252 democratic legitimacy 6, 227, 256, 257, Dahl, R.A. 276 261, 275–7, 278 Dahrendorf, R. 109 hypothesis: new regionalism reduces data 142 democratic legitimacy 264 cross-sectional 85, 90, 97, 106, hypothesis supported, ‘but not the 127(n1), 136, 145, 149, strongest predictor’ 277 155–8t, 159 input- or output-oriented longitudinal 85, 87, 88, 114, 155–7t, (Scharpf) 240, 252–4, 276–7 159, 165 demographic characteristics self-report 63, 64, 66, 75, 77–8, (variables) 31t 78–9, 83, 84 ‘denationalization’ (Zürn) 221 data deficiencies 14, 30(n7), 41, 60(n4), Denmark 59 63, 106, 108, 111, 127(n3), 128(n8), Dente, B. 278(n3) 136, 143, 159, 161(n6), 195, 201, depression/depressivity 53, 169, 206, 208(n6) 170f, 176 Dauwalder, J.-P. 311 deprivation 5, 23, 34t, 36t, 81, De Grip, A. et al. (1997) 191, 303 83, 91 Hoevenberg, J. 303 conceptualization and Willems, E. 303 measurement 10 De Haan, J. 294(n1) levels and distribution 12 decentralization 284, 287, 292 list of variables 32t decision-making 137, 144, 172, see also relative deprivation 279, 247 Deutsch, K.W. 1 delegation 135, 138, 140, 147t, 149, developed countries/North 160(n3) 151n, 159 developing countries/South 160(n3) democratic 242 ‘poor countries’ 82 institutions 222 developmental tasks (Havighurst) 163 international 221 Dicken, P. 110 internationalized processes 231 Dictionary of Occupational Titles 107–8 national 221 Diekmann, A. xi, 5 334 Index

Diener, E. et al. (1999) 11, 24, 304 continuous 127 Lucas, R.E. 304 dual system 126, 154, 159–60, 164 Smith, H.L. 304 impact of ‘Proposition 13’ 294(n3) Suh, E.M. 304 increased participation 199 Dietzen, A. 114 tertiary 140, 143, 159, 160 ‘differential empowerment’ hypothesis vocational 5, 111, 135, 141t, 141, (result of internationalization) 230, 143, 160, 163–6, 175, 181, 188 233–4, 235 educational level 21, 31t, 35t, 37t, 40, direct democracy 10, 223, 239, 240, 67, 87, 91, 94, 98, 128(n6), 166, 244, 252, 254, 282, 283, 286–94, 273–6, 276t 294(n12), 295, 295t, 296t men 55 leads to lower expenditure/ women 53, 54t, 55, 56 revenue 281 workplace organization 135–62 disability 22 educational opportunities 16 dishwashers 16, 17f, 18f, 19 efficiency 6, 240, 241, 252–4, 257, diversity issue 10 291, 293 division of labour 51 Eigenvalues 255(n10) gender 48, 49t, 50 Eijffinger, S.C.W. 294(n1) international 160(n3) electronic data exchange (EDI) 139 divorce 22, 51–3 electronic mail (e-mail) 139, 142 Dolderman, D. 318 electronic technicians 165t, 166, 171–6 Dombois, R. 192 Elfering, A. xii, 309, 311, 312, 323 Doms, M. et al. (1997) 137, 139, Elfering, A. et al. (2004b) 166, 305 160(n2), 304 Bucher, A. 305 Dunne, T. 304 Kälin, W. 305 Troske, K.R. 304 Semmer, N.K. 305 Donzé, L. 142, 160(n3), 298 Tschan, F. 305 Douillet, A.-C. 310 Elfering, A. et al. (2002) 173, 305 Downes, T.A. 291, 292 Gerber, H. 305 Downs, A. 278(n4) Grebner, S. 305 downsizing 206, 254 Semmer, N.K. 305 Draper, D. 321 Elfering, A. et al. (2000) 170, 305 Dunne, T. 304 Kälin, W. 305 Dupont, C. 236(n3), 323 Semmer, N.K. 305 Dye, R.F. 291 elites 223, 224, 228, 230 Ellis, R. 307 econometrics 106, 142, 161(n6, n8) Elstad, J.I. 83 economic emotional regulation in families with cycles 123 adolescents (new research approach) disparities 265–6, 278–9(n9–11) 61–80 downturn/recession 9, 28, 241, 281 approaches 79 growth 292 assessment of affective, cognitive and performance 293 behavioural phenomena 64–7 situation (variables) 32t controllability 78 structure 108, 110 discussion 78–80 economics 4 drop-outs 79 economies of scale 240, 242, 245, emotional states and system 258, 285 reporting 67–71 diseconomies of scale 257 empathic accuracy 71–4, 79 Economist, The 1 experience and behaviour economists 136, 143 sampling 64–6, 78 Eder, K. 318 limitations of study 79 education 2, 23, 24t, 25, 27n, 29, 33t, methodological aspects of traditional 47, 82, 83, 204–5, 292 measurements in family psychology basic 127 62–3 Berufslehre 159 methodology 61, 62–3, 64–6, 75, 78, 80 Index 335

modalities (functional and emotional state 70 dysfunctional) 74–7, 79 erosion of regular work 190–217 research questions 61, 63–4 full-time 55–6, 67, 190, 191, 194, results 67–78 195, 196f, 200n, 201, 207, samples 66–7, 79 208(n6, n8) 209–17t settings 77f German-language 112, 128(n4) statistical analyses 67 job-sharing 208(n8) emotional states 65t, 67–8 marginal 192, 195, 201, 202, 204, empathic accuracy 71–4 214t-17t empathic accuracy 71–4, 79 ‘other forms’ 5 direct 72 part-time 55–6, 67, 128(n7), 147t, indirect 72 150t, 152t, 153, 191, 192–3, literature 73 195, 197, 201, 202, 204, empathic projection 72, 73, 74 205, 207, 208(n8), empiricism 23, 28, 29, 39, 42, 75, 83, 209–17t 85, 105–9, 111, 112, 127(n2–3), part-time, low-income 206 135–7, 144, 160(n1–2), 163, permanent 209–17t 190, 192, 225, 258, 261, 264, satisfaction 31t 275, 277, 278(n4), 282, 284, standard 5, 190 286, 291–3, 295n temporary 197, 202, 209–17t literature survey (information temporary full-time 201, 204, 206 technology and demand for terminology 207(n2) employees) 154, 155–8t, types (Switzerland, 1991–2003) 161(n13) 209–11t employee competence 149 types (Switzerland, 1970–2000) employee voice 149, 143–4 212–13t employees (labour/workers) 140, 189, types (western Germany, 193 1985–2002) 214–17t autonomy 138 undemanding 127 education levels 135–62 employment status 21–2, 25, 27n, 31t, employment share 143, 146–8t, 35t, 37t 150–2t, 153–4, 156–8t, empowerment 161(n7) ‘differential’ 230 foreign 160(n4) export-oriented business over trade highly educated 135–6, 138, 143–54, unions 233–4, 238(n24–6) 159, 161(n11) state actors 230–3, 237–8(n20–3) individual responsibility 109 endogeneity/endogenization of low-educated 143, 144, 145, 146–54, organization 136, 149, 152t, 153, 159, 160, 161(n11) 161(n12) low-skilled 106, 160(n3) Engelhardt, H. xi low-skilled substituted by high-skilled engineers and technicians 115, 132t 136 environment middle-educated 143–5, 159, international 6 161(n11) natural 16, 23, 29, 31t, 33t skill requirements 139 social 23–5, 27n, 41, 82, 83, 84–5, skilled/highly-skilled 106, 136, 137, 88, 90 156–8t, 160(n3), 188 work 2, 189 see also workers/unskilled erosion of regular work (Switzerland and employer, change of 163, 166–7, 167n, Germany) 190–217 180–1, 182–4, 189 age-dependent rate of regular work employers 5, 111, 189, 206, 207 (Switzerland, 1970–2000) 200f employers’ organizations 233, 238(n25) development of regular work in employment 23, 30(n9), 39, 40, 54t, Germany 201–4, 208(n7) 58, 110, 113, 126, 127 development of regular work in advertised jobs by industry 129–30 Switzerland 194–201, atypical 192, 206, 207 207–8(n4–6) 336 Index erosion of regular work – continued external world disproportionate growth of atypical (outside family) 50 employment hypothesis 192, 193, 205 Fachhochschulen (technical and business end of regular work hypothesis 191–4, colleges) 143, 160 207(n1–3) factor analysis 252, 255(n10), explanations 197, 202 279(n15–16) labour-market flexibility hypothesis factor prices 139 194, 206 Falk, M. 106, 157t, 160(n2) literature 190, 191–2, 207(n3) family/families 3, 38, 40, 41, 50, 53, ‘occurred primarily in 1990s’ 199, 204 59(n1), 83, 100 parallelism instead of special paths with adolescents (new research 204–7, 208(n8) approach) 61–80 ratio of standard to non-standard ‘arena of emotions’ (Larson and employment 191 Richards) 61, 62 speed 193 companionship model 41, 59(n3) substitution and displacement domestic roles and power 41 hypothesis 192, 193, 205–6 everyday experience 5 supply and demand 206–7 external constraints 41, 59(n3) terminology 194, 201, 207(n2, n4) institution model 41 Esping-Andersen, G. 86 pluralization of forms 61 Euromodule (survey tool) 15, 17f, 18f, pre-industrial 61 20f, 21f, 22–3, 24t, 26–7t, see also conjugal interactions; 33–7t emotional regulation Europe 9, 12, 23, 40, 58, 62, 139, family cycle 205 236(n3), 255(n2), 258, 278(n4) family life 44, 57, 58 Central 5 pluralization since 1960s 40 Northern 239, 242 ‘Family Life and Professional Work: Southern 194 Conflict and Synergy’ (EU project) Western 5, 239, 265, 279(n11) xv, 79–80 European Community family psychology 62–3 free trade agreement with Switzerland Farago, P. xii (1971) 237(n10) farmers, growers, agriculturalists 115–18, European Economic Area (EEA) 5, 230 128(n8), 132t European Foundation (author) 160(n5) FASEM-C (family self-monitoring European Household Panel [survey] system) 64, 78–9 14, 22 applications in clinical European integration 222 psychology 77–8, 79 European Social Survey 3 information types and item European Sociological Association formats 65 (ESA) xiii Favre, F. 307 European Union 4, 5, 10, 222, 230, Feather, N.T. 319 235, 238(n24–5), 281 federal states ‘negative’ versus ‘positive’ fiscal policy (sustainable) integration 233 281–96 Stability and Growth Pact (SGP) 281, federalism 6, 10 294(n1) Feld, L.P. xii, 6, 282, 283, 284, Europeanization 222, 233, 236(n6), 287, 292, 293, 238(n26–7) 294(n4, n13), 320 ‘control case’ lacking 236(n2) Fernandez, R.M. 106 euroscepticism 5 Figlio, D.N. 291, 292 EViews (Version 4.1) 294(n8) financial situation 87, 93t examinations 164, 166, 171 firm size 110, 115, 116, Excel (spreadsheet) 77–8 128(n6, n10), 133, 142, experience 64–6, 114 144, 147t, 150t, 152t, experimental learning 247, 248, 249f 153, 154, 161(n6), 162t Index 337 firm-level evidence (Swiss economy) ‘not much evidence against Swiss 135–62 mixture of fiscal firms 107, 110, 112, 161(n13) institutions’ 294 all types 113 results in perspective 291–4 (n12–13) centralized/decentralized 139 three constitutional or statutory changing pattern of work clauses 282–3, 294(n4) organization 109, 112 Fischer, A. 319, 323 flexible organizational practices 142 Fischer, A. et al. (2002) 222, 307 foreign 152t, 153 Nicolet, S. 307 French and Italian 154 Sciarini, P. 307 high-technology 127 Fischer, A. et al. (2003) 238(n27), 307 joint-stock 267, 270–2t Nicolet, S. 307 low-technology 127 Sciarini, P. 307 manufacturing 137 Fischer, J. 293 organizational change 108, 112 Fitzpatrick, R. 322 panel studies 106, 127(n1) foreign direct investment 160(n3) performance 159 foreign policy 224 private 232f four ‘i’s (initiative, institutions, re-organization 206 information, ideas) 230 small and medium-sized enterprises Fourastié, J. 109 (SMEs) 111 France 59, 155–6t, 160(n2), 194 technological innovation 108, 112 François, J.-P. et al. (1999) Firth, D. 322 160(n5), 307 fiscal autonomy 284, 286, 287, 288, Favre, F. 307 289t, 290 Greenan, N. 307 fiscal competition 292 Franzen, A. xi fiscal constraints/restraints 282, 284, free-riding 266 286–91, 293–4, 294(n3, n6), 295, Freitag, M. 293, 294(n4) 295t, 296t Frey, B.S. 283, 293, 295 fiscal decentralization 282, 288–90, Fribourg (canton) 243, 284 294, 294(n3), 295t, 296t Friederich, U. 315 fiscal equalization systems 285 Fuchs, D. 263 fiscal federalism 282, 288, 291, furniture 16, 17f, 18f, 19 292, 294(n3) future plans 177, 178f competitive 291 fiscal policy (sustainable) in federal Gail, K. 323 system 6, 281–96 Gallagher, M. et al. (1992) 164, 307 aim of chapter 283 Ellis, R. 307 data sources 295 Hargie, O. 307 dependent variables 283 Millar, R. 307 econometric model 283–7, gardens or balconies 16, 17f, 18f 294(n5–7) Garrett, G. 221 economic significance 287–8 GDP 16, 281, 295 empirical results 287–91, gender/sex 21, 24, 27n, 31t, 39, 41, 294(n8–11) 56, 58, 59(n2), 76, 83, 87, 89f, 90, existing literature (empirical 92t, 94, 97–9, 101, 164, 167n, studies) 282 273–6, 276t expenditure, revenue, and deficit division of labour 48, 49t, 50 (per capita, 1980–98) 296t gender bias 59(n1) explanatory variables 286, 295t gender roles 177 four indicators (1980–98) 283 gender studies 2 important question 287 Gerber, E.R. 294(n12) index of constitutional Gerber, H. 305 constraints 295 Gerlach, J. 156t index of direct democracy 286, 295 German Institute for Economic Research introduction 281–3, 294(n1–4) (DIW) 192, 203n 338 Index

Germann, R.E. 228 GNP per capita 4, 12–13, 28, 30(n3) Germany [re-unified] 11, 12, 15, Goebel, H. 71 29(n2), 67, 111, 154, 156–7t, Goldman, N. 82 160(n2), 190, 252, 281 Goldstein, H.Y.M. 321 ‘absolute necessities’ 17f, 17 Goldthorpe classification 91, 92–3t, actual standard of living 18f, 19 94, 95t development of regular work 201–4, goods and services 16, 109, 160(n3), 245 208(n7) Gottman, J.M. 74 direct and indirect measures of Goux, D. 160(n2) poverty 29 governance erosion of regular work (Swiss democratic 261–4 comparison) 190–217 multilevel 237(n14) factorial analysis for well-being 33t government 193, 274 GNP per capita PPP 12, 30(n3) ‘government to governance’ 259–60, income equality 30(n4) 260t, 261, 278(n5–6) income quintiles 21f, 21 Graham, T. 73 individual and societal well-being (and Gramm–Rudman–Hollings (GRH) Act their determinants) 23–5, 26–7t, (1985) 281–2 34–7t grants (financial) 81n, 286, 287, 288, labour market law 206 289t, 295, 295t, 296t measuring relative deprivation 15 Graubünden (canton) 243, 284 micro-censuses 192 Greater London Authority 256 pluralization of life-course Grebner, S. xii, 305, 312, 323 hypothesis 58 Grebner, S. et al. (2003) 170, 309 relative deprivation, 19–22 Elfering, A. 309 standard of living 28 Gut, S. 309 structural changes in the labour Kälin, W. 309 market 190–217 Lo Faso, L. 309 working experience 5 Semmer, N.K. 309 Germany, East [DDR] 201 Green Cowles, M. et al. (2001) Germany, eastern [ex-DDR] 12, 23, 24t, 222, 309 25, 208(n7) Caporaso, J. 309 ‘absolute necessities’ 17 Risse. T. 309 actual standard of living 19 Greenan, N. 138–9, 156t, 302, 307 relative deprivation 20f Greenberg, J. 189 Germany, West [FRG pre-1990] 201 Griffin, D.W. 74, 318 Germany, western [Bundesländer, Gross, M. 191 ex-pre-1990-FRG] 23, 24t, Grote, J. 227 25, 201 Guellec, D. 138–9, 302 ‘absolute necessities’ 17 Gut, S. 309 actual standard of living 19 Guye, O. 315 relative deprivation 20f Gesellschaft Schweiz – Analyses sociales Hakim, C. 191 (2004) 4 Hall, J.A. 73 Gesn, P.R. 73 Halleröd, B. 14–15, 21 Giddens, A. 39, 59(n2) Handcock, M.S. 300 Giesecke, J. 191 Hänggi, Y. xii Gilbert, A. 81n Hargie, O. 307 Gini coefficients 30(n4) Harrington, T.F. 315 Glarus (GL) 283 Haskel, J.E. 136, 160(n2) Glass, T. 84 Hauptmanns, P. 106 globalization 191, 236(n1) Haverland, M. 222 competition among states 234 health 3, 12, 16, 23, 24t, 25, 27n, effects on national decision-making 29, 31t-3t, 35t, 37t, 165 systems 221 determinants 5 GMM method 287 factors 82 Index 339 health: impact of social inequalities Héritier, A. et al. (2001) 222, 310 81–101 Douillet, A.-C. 310 data 87–8 Kerwer, D. 310 degree of social inequality 83, 85 Knill, C. 310 discussion 97–100 Lehmkuhl, D 310 five pathways 83 Teutsch, M. 310 general health 86, 88, 89f, 90, Hess, G.D. 293 92–3t, 94–8 Heunert, S. 311, 312 health-related events 86–90, 92–8 ‘high-end jobs’ 111 household circumstances 83, 85, Hirst, P. 111 86, 87–90, 91, 92t, 93t, 94, Hitt, L.M. 301 95t, 96–7, 98, 101 Hix, S. 221 hypotheses 86–7 HIV/AIDS 268 implications for social policy 100 Hoevenberg, J. 303 introduction 81–7, 100(n1–2) Hoffmann, E. 192, 202, 204, 206 interviews (first) 91–5 holidays 17f, 17, 18f, 19 interviews (second and third) 95–7 Holland see Netherlands latent factors 90, 91 Hollenstein, H. 139, 298 literature 81, 82–3, 86, 100(n2) Holmes, J.G. 74, 318 material and financial home economics and cleaning circumstances 83, 85, 87, 88, personnel 115, 128(n8), 132t 89f, 91, 92t, 93t, 94, 96, 97, Horner, M. 320 98–9, 101 Hosmer, D.W. 134 measuring health and social hospital staff and health service environment 90–1 workers 115, 132t methods 87–91 ‘hot cognition’/‘hot emotions’ 64, 75 model (dynamic structural hotels (and allied trades) 129, 130t, equation) 88–90, 100–1 132t, 142, 162t psychological health 83, 84, House, J.S. et al. (1988b) 84, 311 86–7, 88, 89f, 91, 92–3t, Landis, K.R. 311 94–9 Umberson, D. 311 research questions (and answers) household composition 24, 32t, 34t, 82, 85, 97–100 36t, 83, 99, 100 results 91–7 see also households/type self-assessment of personal household income 23, 24t, 27n, situation 83, 84, 85–8, 90, 91, 30(n7), 31–4t, 36t 96–9 relationship with deprivation 20–1 social and cultural circumstances 83–4, see also income 85, 86–7, 89f, 91, 97–8 household work 177, 178, 207 social environment ‘does not explain households 11, 15, 16, 18, 30(n5), 88 much’ 98–9 Germany 201, 208(n7) social support 87, 89f, 91, one-person 94, 96, 99 93t, 94–5, 95t, spatial division from workplace 96–9, 101 207(n4) socio-demographic characteristics 87 type 40, 91, 101 time-lagged effects of type (in youth) 92t circumstances 83, 84–5, ‘housekeepers’ 87, 101 86, 88, 91, 92t, 94, housing 12, 17f, 18f, 23, 266 97, 98 ‘apartments’ 31t three dimensions 86, 100(n1) Hoxby, C.M. 292 health care costs (Switzerland) 81–2 Hudson, K. 312 Healy, M. 321 Hug, S. 228, 323 Heden, Y. 160(n2) Hujer, R. et al. (2002) 157t, 311 Heinelt, H. 278(n3), 279(n12) Caliendo, M. 311 Herfindahl’s index 266n Radic, D. 311 institutional consolidation 266t Hulliger, B. 303 340 Index human capital 138, 143, 155t, 159 information technology (IT) 108, 109, human resources 243, 250 118, 122, 126, 129(n17), 142 Hungary 28, 29(n2) ‘consultants, IT and media Hurd, M.D. 297 specialists’ 115, 132t diffusion 122, 125 Ickes, W. 73 see also computer work ideology 286, 287, 288, 289t, 290, information technology and demand for 295, 295t, 296t employees 135–62 Iglesias, K. xii, 5 comparative evidence and survey of Illinois 291 empirical literature 154, 155–8t, illness 170, 171f, 208(n5) 161(n13) income 10, 18, 105, 100, 127(n3) comparative studies 160(n2) per capita 16, 30(n7), 286, 296t complementarity hypothesis 138–9, disposable 286, 296t 153, 154, 155–8t, 159 permanent surplus 19 composition of data set 161–2t relationship with well-being 11–12 conceptual background 137–9 see also household income contribution to existing income equality/inequality 13, 30(n4), literature 136 85, 191 data 142, 160–1(n5–6) income groups, lowest 28 dependent variables 155–8t income quintiles 20, 21f, 29, 30(n8), descriptive analysis 139–42 34t, 36t determinants of employment shares of income redistribution 293 highly educated, middle educated, income tax 273 and low-educated employees individual rights 50 (1999) 146–8t individual and society (module) 2 future research 161(n9) individualization 22, 39, 53, hypotheses 136, 139, 145, 153, 159 60(n4), 191 implications for economic hypothesis rejected 57, 195, 197 policy 159–60 individuals 16, 247 literature 136–9, 144, 160(n2, n4), ‘forced to be happy’ 24 161(n12) industrial sector ‘winners’ 110 model specification 143–5, 161(n7–10) industries 125, 128(n6–7), organizational factors 149, 150–1t, 128–9(n10–11), 129(n13), 155–8t 142, 144, 161(n8) results of model estimations 145–54, all 113 161(n11–12) categorization 133 shortcomings of study 136, 159 changing structure 122, 126 skill demand (Swiss firms, IT (negative effect on job openings 1997–2000) 141t for unskilled workers) summary and conclusions 159–60 129(n17) technological factors 145, 150–1t, productivity increase and skill 155–8t, 159, 161(n12) demand 129(n18) initiatives 244–5, 282, 293 inequality 9, 10, 11, 29, 58 parliamentary 249 post-industrial societies 22 voter 223, 283 infant mortality rates 293 Inman, R.P. 282 information, asymmetrical 231 innovative activities 152t, 153 information and communications insertions in social structure 53–5, technology (ICT) 119f, 120, 60(n9–10) 121f, 124f, 135, 136, 137, poststructural hypothesis 53 138–45, 153, 154, 155–8t, sociological hypothesis 53 159, 160(n4) Institut de Sociologie website 27n, ICT input (ICT capital) 143 30(n8, 10, 12–13) new organizational practices 139–40 institutional change (module) 2 information and media society intensity of conflicts in country 32t, (module) 2 35t, 37t Index 341 intercoder reliability 128(n6) empirical tests (qualitative and interest groups 224, 231, 232f, quantitative techniques) 221, 236(n8–9), 247 223–4 interlocal agreements 260, 260t further research required 230, intermunicipal cooperation (IMC) 245 234, 236 see also communes; municipalities hypothesis 1 (direct international pressure 222, 225, 227 internationalization increases International Social Survey importance of informal rather Programme 3 than formal pre-parliamentary internationalization 110, 136, 160(n3) consultation) 224, 225, 227, ‘differential empowerment’ 237(n10) hypothesis 230, 233–4 hypothesis 2 (greater leeway for formal direct/indirect 221–36 consultation in case of indirect empowerment of export-oriented internationalization) 225, 227 business over trade unions 233–4, hypothesis 3 (reduction of 235, 238(n24–6) parliamentary conflict as empowerment of state actors 230–3, consequence of 237–8(n20–3) internationalization) impact on domestic power 228–30, 234 configuration 230–4, introduction 221–3, 236(n1–3) 237–8(n20–6) literature 221 impact on institutions of decision methodological bias (most existing making-process 224–7, literature) 236(n2) 236–7(n7–13) quantitative analysis 223, 228, 234, impact on level of conflict in 235, 238(n20) decision-making process 227–30, internationalization: indirect 237(n14–19) autonomous adaptation 222, 223, polity and politics implications 226–8, 229f, 231, 233, 235, 236, 221, 222 236(n6), 237(n16–17) transmission mechanisms 222 residual category 223, 236(n6) internationalization: case studies internet 4, 113, 135, 139, 143, 144, (Switzerland) 145, 146t, 148n free movement of persons with the EU extranet 139, 142 (direct internationalization, intranet 135, 139, 142, 143, 144, 1994–8) 223, 226t, 226, 145, 146t, 148n 230–4, 236(n6) interpersonal abilities 138 old-age pension, eleventh reform interpersonal behaviour 75, 76 (control case, 1995-present) 223, interviews 3, 42, 91–5, 201, 226t, 227, 230, 231, 232f, 207–8(n5), 225, 226, 237(n13) 255(n5), 279, 280t telecommunication policy reform face-to-face 223 (indirect internationalization, random-quota basis 280 1994–7) 223, 226t, 226, 230–4, telephone 88, 195, 269, 280, 280t 236(n6) see also questionnaires internationalization: domestic politics intra-firm processes 108 221–38 Inversin, L. 312, 326 case studies 223, 226, 228–31, 233–4, investment decisions 147t, 149 235, 236(n6) Ireland 4 conclusion 234–6, 238(n27) irritability 169, 170f, 176, 177f, contribution (threefold) of chapter to 178, 179 existing literature 222 language differences 180f, 182f data and methodology 223–4, 234, Italy 59, 158t 235, 236(n4–6) IZY–Builder 78 domestic consequences of internationalization 224–34, Jacobebbinghaus, P. 154 236–8(n7–26) Jacobshagen, N. 188, 189 342 Index

Jann, B. xii–xiii, 5 Heunert, S. 311 Jarque-Bera (J.-B.) test 287, 288, 290, Semmer, N.K. 311 294(n8), 296t Tschan, F. 311 Jirjahn, U. 156t Kälin, W. et al. (2004) 166, 312 job advertisements 105, 107, 111 Elfering, A. 312 cyclical fluctuation 113 Grebner, S. 312 firm size 115, 116, 128(n10) Inversin, L. 312 by industry 115, 129–30 Rochat, S. 312 placement costs 115, 116, 120, Semmer, N.K. 312 128(n9–10) Tschan, F. 312 representative survey 112–14, Kalleberg, A.L. 191 128(n4–7) Kalleberg, A.L. et al. (2000) 207, 312 samples 112, 128(n5) Hudson, K. 312 ‘silver bullets’ of personnel Reskin, B.F. 312 recruitment 113 Kaplan, G.A. 82 structure of industry 116 Katz, L.F. 298 survey year 123, 128–9(n10) Katzenstein, P.J. 227 target week 128(n5) Kaufmann, V. 315 type of vacancies 114 Kawachi, I. 85, 327 unskilled workers 131f Keating, M. 274, 278(n3) use by employers (changes over Kellerhals, J. xiii, 5, 42, 60(n5, n8, time) 125 n10–11), 195 job control 163, 167f, 168, 172–3, Kennedy, B.C. 327 174f, 174, 177, 179, 184, Keohane, R.O. 221 185f, 186, 189 Kern, H. 107 job deskilling 107 Kersting, N. 239, 241 job enrichment schemes 144 Kerwer, D. 310 job rotation 135, 138, 140, 142, 144, Khlat, M. 87 146t, 149, 151n Kiewiet, D.R. 282 job satisfaction 169–70, 171f, 177f, Kirchgässner, G. xiii, 6, 282, 283, 181–6, 189 287, 292, 293 job share 144 294(n2–4) job tasks 108, 127(n2–3) Klarer, U. 113 job-specific tasks 107, 108, 127(n2) Klingemann, H.-D. 263 job vacancies 114, 126, 129 Kloas, P.-W. 114 manual, low-skilled 128(n9) Klodt, H. 110 number 107 Knill, C. 222, 310 prospective incumbents Knoke, D. 223 (characteristics) 107, 112 Koenen, E.J. 318 Johnson, G. 136 Kohli, M. xiii, 39, 59(n2) Journal of Economic Perspectives standardisation theses 60(n4) (1997) 136 Kommission für Zukunftsfragen journals xi–xvi, 62 (Commission on Future Issues), Joye, D. 314, 315 Germany 191, 192, 207(n1) justice 62 König, M. 302 Kooiman, J. 259 Kaase, M. 263 Kress, U. 192, 193 Kabanoff, B. 319 Kretschmer, G. 113, 321 Kaiser, L.C. 192, 204 Kriesi, H. xiii, 221, 224, 236(n3), Kaiser, U. 160(n2) 237(n10), 278(n1) Kälin, W. xiii, 305, 309 Kriesi, H. et al. (1999) 236(n3), 314 Kälin, W. et al. (2000) 166, 168, 169, Armingeon, K. 314 170, 311 Siegrist, H. 314 Crettaz von Roten, F. 311 Wimmer, A. 314 Dauwalder, J.-P. 311 Kriesi, I. 111 Elfering, A. 311 Krueger, A.B. 160(n2), 298 Index 343

Krüger, H. 48 Steiner, R. 314 Kübler, D. xiv, 6, 278(n1, n3), Wichtermann, J. 314 279(n12–13) LAN/WAN (firm computer Kübler, D. et al. (2002) 280, 314 networks) 139 Bassand, M. 314 Landis, K.R. 311 Joye, D. 314 Lang, T. 164 Schwab, B. 314 Lange, P. 221 Kugler, P. 106, 125, 160(n4) language 10, 12, 42, 105, 107, 161(n5), Kurer, O. 111 164, 165t, 166, 167n, 178–80, 181–3f, 194, 245, 246f, 273–6, 276t, 278, labour see employees; workers 285–91, 296t labour (factor) 234 German-speaking Switzerland 112, labour (module) 2 128(n4) labour demand Lansley, S. 13–14, 16 Arvanitis’ hypothesis 136 Larson, R. 61, 62 observed shift 136 Lash, S. 39 overall 113 late modernity 38, 59(n3) shift 136, 137 Laumann, E.O. 223 shift towards highly educated Lausanne (metropolitan area) 256, 257, employees 140–2 264, 266t, 267–9, 270–1t, 273, see also skill demand 280t, 280 labour force (employed plus Lausanne: Swiss Federal Institute of unemployed) 85, 88, 195, 201, Technology 209–17t Laboratoire de sociologie urbaine participation 192–3, 204, 207(n1) 278(n1) labour market (Swiss) xviii, 1, 5, 40, 41, lavatory/bath 16, 17f, 18f 60(n9), 87, 91, 103–217 Lavenex, S. 222 domestic 110–11 law 270–2t erosion of regular work 190–217 cantonal 250, 255(n3), 267, 269, 283 firm-level evidence for Swiss communal 244 economy 135–62 legislation adapted under information technology (workplace international pressure 222, organization and demand for 236(n1) employees of different private 242, 255(n6) education levels) 135–62 public 255(n6) participation 195 see also internationalization regional 127(n1) Law, I. 78 restructuring 191–2 Le Blanc, P.M. 325 skill demand (long-term dynamics, Le Galès, P. 259, 278(n4–5, n7) 1950–2000) 105–34 Lebo, R.B. et al. (1995) 164, 315 structural changes (Switzerland and Harrington, T.F. 315 Germany) 190–217 Tillmann, R. 315 young adults entering the Lefèvre, C. 278(n3), 278(n5) workforce 163–89 Leffingwell, T.R. 62 labour market flexibility 144, 160, 193, Lehmkuhl, D. 222, 310 194, 206, 207 leisure (spare time/recreation) 3, 16, 69f labour market regulation 194, 206 Lemeshow, S. 134 labour market segmentation 105, 111, Leresche, J.-P. 278(n6) 115 Lev, B. 143 labour market structure 190 Levy, F. 109, 127(n2), 298 labour productivity 144 Levy, R. xiv, 5, 39, 42, 48, 60(n5, n8, Lachat, R. 221 n10–11), 195 Ladner, A. xiv, 6, 240 Levy, R. et al. (1997) 39, 315 Ladner, A. et al. (2000) 248, 314 Guye, O. 315 Arn, D. 314 Joye, D. 315 Friederich, U. 314 Kaufmann, V. 315 344 Index

Li, J.H. 302 Mair, P. 221 life courses 22, 41, 53, 57, 59(n2), management 138, 140, 156t, 191, 197 279(n16) sex-typed 39 management levels 144, 147t, 148n, transitions 21 149, 151n life cycle 205 manufacturing 118, 135, 137 life expectancy 16, 87 formal education of employees lifestyle 12, 22, 81, 84 (1999) 141, 141t Likert scale 144, 148n ‘niche’ medium-sized high-technology Lind, J. 191 industries 140 Linder, W. 257 March, J.G. 247–8, 249f, 255(n9) linear regression models 24, 27n, Marcotte, D.E. 192 30(n13), 34–7t marital conflict 74 ‘linked lives’ (Elder) 57 marital status 24, 27n, 31t, 34t, Lipsmeier, G. 14 36t, 83, 166 living at home 167 Marmet, D. 298 living conditions 11, 22, 83, 87, Marmot, M. 82 88, 97, 100 mathematics 291 objective 29 Matsusaka, J.G. 282, 283, 293, living with a partner 32t, 35t, 37t 294(n4–5, n13) Living in Switzerland Survey Matthey, R. 190n (1999–2020) 81n Maurer, R. 110 Lo Faso, L. 309 Maurin, E. 160(n2) local autonomy 258 mayors 251, 251f local government 6, 240 McDonough, P. 98 see also municipal reform McFadden, D. 297 locational attractiveness 222, 236(n1) McGrath, J.E. 324 logistic regressions 115 McGuire, T.J. 291 binary 246f media 2, 53, 56, 251f multinomial 54t, 55, 128(n8) newspapers 17f, 18f, 112 Lombard, M. 128(n6) membership of an association 25, 27n, London 258 32t, 35t, 37t Lowery, D. 278(n3) memory bias 63, 64, 70 Lowery, D. et al. (1992) 262, 316 men 39, 42, 50, 52t, 53, 55, 56, 70, DeHoog, H. 316 166, 176–8, 190, 195 Lyons, W.E. 316 age-dependent rate of regular work Lucas, R.E. 304 (Switzerland, 1970–2000) 200f Lucerne (canton) 243, 294(n6) decline in standard employment 201 Lucerne (metropolitan area) 256, 257, declining employment rate 204 264, 266t, 267–9, 271t, 273, 280t empathic accuracy 73 Ludwig-Mayerhofer, W. 318 employment (Switzerland) 196f, 198f Lugano 3, 256, 257, 264, 265, 266t, full-time employment 67 267–9, 272t, 273, 280t professional trajectories 197 Luhmann, N. 62 regular work 199 Luxembourg Income Study 3 regular work (Switzerland, Lynch, L.M. 143 1991–2000) 196f Lyons, W.E. 316 regular work (western Germany) 202, 203f Maastricht Treaty (1991–) 281 Merrill, A. 297 Mach, A. 231 method control 168 Mach, A. et al. (2003) 222, 316 metropolitan areas (agglomerations) Häusermann, S. 316 6, 285 Papadopoulos, Y. 316 democracy problems of new Machin, S. 125, 126, 139, 160(n2), regionalism 256–80 161(n8), 300 ‘golden age of reform’ (1960s, Mack, J. 13–14, 16 1970s) 258 Index 345

institutional consolidation/ multilevel ANOVA model 67, 68, 72 fragmentation 256–9, 265–6, multiple correspondence analysis 266t, 266n, 267, 273–8, (MCA) 51 278–9(n9–11) multiple imputations technique ‘lost dimension of Swiss federalism’ (Rubin) 142 (Linder) 257 multiple regressions 67, 188 networks 259 multivariate ‘no significant reform’ (twentieth analysis 273, 274 century) 265 linear regressions 223, 246f ‘purpose-oriented cooperation’ 259, logistic regressions 21, 25, 30(n8), 278(n4) 223, 237(n16) terminology 278(n2) regressions 264, 269 metropolitan areas (Swiss): area-wide statistical methods 42, 43, 60(n5) governance (four policy municipal reform 239–55 domains) 267–9, 270–2t, aims 241 279(n12–13) case studies 248, 250–1, 251f, 253f classification 270–2t, 279(n13) causes 241, 248 cultural institutions 268–9, 270–2t, change and reforms at local level 279(n12, n15), 280t 241–7, 255(n5–8) public transport 267–8, 270–2t, conclusions 254 279(n12, n15), 280t different types, different goals 253, 253f social services for drug users 268, executive, legislative 245, 246f, 270–2t, 279(n12), 280t 249–50, 253, 253t synthesis 268, 270–2t four groups of reforms 239–40 water supply 267, 270–2t, goals (democracy or efficiency) 252–4, 279(n12, n15), 280t 255(n10) Metropolitan Counties (England) 258 implementation (where and metropolitan governance index 269, how) 245 270–2t initiators 251f metropolitan reform tradition 257, institutional change: path 258, 262, 263, 273, 275, 276 dependency 240 Michel, G. xiv, 67 international comparison 240, 254 micro-census (Germany) 202 ‘reforms happen, reforms are microeconomics 247 made’ 248 Milgrom, P. 138 reforms as intentional processes – triggers Millar, R. 307 of reform projects 247–52, Mills, M. 188 255(n9) Milner, H.V. 221 research questions 239–41, 255(n1–4) MIS Trend 280 spread 242t Modernity and Self-Identity variables (context, performance, (Giddens) 59(n3) political) 245, 246f modernization 191 municipal reform: four groups ‘second wave’ 39 changes towards more direct modernization of family participation 240, 242t traditionalism 58, 59 changes in local representative Møller, I.H. 191 democracy 240, 242t, 243–4 Montreal 258 management 240, 242t Moravcsik, A. 224, 230 territorial and functional 240, 242t Morgan, J.P. 127(n3) municipal secretary 245, 251, 251f, Morris, M. 300 255(n4–5) Moulin, S. 192 municipalities Mouritzen, P.E. 252 administrative organization 255(n3) Mückenberger, U. 192 autonomy 241 Müller, C. 294(n4) budget deficits 245, 247 Müller, S.G. 128(n7) citizen participation (two Müller, W. 127(n3) systems) 244 346 Index municipalities – continued Nahrstedt, W. 62 cooperation (purpose-oriented) 267 nation-state 240 debt-equity ratio 241 national debt 281–2 distinguishing features (Switzerland nationality 89f, 90, 91, 92t, 94, 97, 98 versus foreign countries) 240, Nave-Herz, R. 61 255(n1) ‘necessities’ 14, 16 executive branch 251, 251f socially perceived 9 financial autonomy 240, 254, negotiation 41, 50, 58, 59(n3) 255(n2–3), 279(n11) negotiations, international 222, 224, financial situation 250 226, 230 local assembly (smaller neighbourhoods 23, 31t, 83 municipalities) 244 Nelson, D.L. 176, 178 nationwide survey (1998) 241, neo-liberalism 252 255(n4) Nerdinger, F.W. 164 number (1848–2003) 255(n7) Netherlands 59, 67 parliaments 242t, 242n, 244, network analysis 223 250, 251f networks performance threshold index informal 114 246f, 247 policy-relevant actors 259, 260t quality of life 29 New Public Management (NPM) 239, reform index 246f, 255(n8) 240, 242t, 243, 248–51, reform triggers 249f, 249–50 251f, 253–4 size 245 core elements 243 spill-over effects 240 Pseudo-R Square 245, 246f structural differences (reform versus new regionalism see regionalism (new) non-reform) 241 New Zealand 194 urban-rural gap 254 Newey–West autocorrelation-consistent ‘very small’ 240, 243, 255(n1, n7) standard errors 287, 294(n8), 296t see also communes Newton, K. 263 municipalities: amalgamation/ Nicolet, S. xiv, 6, 307, 323 mergers 239, 240, 242–3, Nicolet, S. et al. (2003) 236(n4–5), 319 245, 246f, 248, 250, 251, Fischer, A. 319 251f, 253, 253t, 254, Sciarini, P. 319 265, 278–9(n9) Noll, H.-H. 12, 23 municipalities: decision-making non-alignment/neutrality 5 process 239, 244, 254 non-employed/non-employment 195, municipalities: inter-municipal 197, 201, 202, 208(n6), 209–17t cooperation 239, 241, 242, 245, non-governmental organizations 268 246f, 250, 251f, 253, 253t, ‘non-necessities’ 14 254, 255(n6) non-state actors 224, 226, 227, 231, Muntaner, C. 319 232f, 259, 267 Murane, R.J. 109, 127(n2), 298 North America 258, 278(n4) Murray, S.L. et al. (2002) 73, 318 NPM see New Public Management Bellavia, G. 318 Nurmi, J.-E. et al. (1995) 164, 319 Dolderman, D. 318 Poole, M.E. 319 Griffin, D.W. 318 Seginer, R. 319 Holmes, J.G. 318 nurses 163, 165t, 166, 171–6, 189 Murray, S.L. et al. (1996) back pain 168 73, 74, 318 Griffin, D.W. 74, 318 O’Brien, G.E. et al. (1994) 164, 319 Holmes, J.G. 74, 318 Feather, N.T. 319 Mutz, G. et al. (1995) 207(n3), 318 Kabanoff, B. 319 Bonß, W. 318 O’Brien Caughy, M. et al. (2003) Eder, K. 318 84, 319 Koenen, E.J. 318 Muntaner, C. 319 Ludwig-Mayerhofer, W. 318 O’Campo, P.J. 319 Index 347

O’Campo, P.J. 319 pain 65t, 65–6 O’Reilly, J. 191 parents 66, 68, 70–2, 76, 167 O’Sullivan, A. 282, 292 social status 91, 94, 98 occupational categories 118 Paris, D. 30(n8), 83 occupational group 128(n6) parliaments 231, 237(n20) occupations 163, 165t partnerships (private life) 167 advertised jobs 117f, 119f, PDI see Proportional Deprivation Index 121f, 124f Peeters, M.C.W. 325 categories 115, 116, 128(n8), 132t Peltzman, S. 283–4 labels 115 perception 63, 84, 86, 279(n12, n16) OECD [Organisation for Economic performance-related pay 243 Co-operation and periurbanization 265 Development] xvii, 4 see also communes/periurban author 109, 113, 125, 137, 194 periphery 111 OECD countries 135, 139, 159, 256, Perrez, M. xii, xiv–xv, 5, 64, 66, 73 281, 282 Perrez, M. et al. (2001) 66, 320 shift towards more highly-educated Horner, M. 320 workers 137 Schoebi, D. 320 OECD scale 30(n7) Wilhelm, P. 320 Oesch, D. 91 Perrez, M. et al. (2000) 68, 75–6, 320 office and administrative staff 115–18, Schoebi, D. 320 119f, 129(n12), 132t Wilhelm, P. 320 old age 84, 208(n5) personal computers 16, 17f, 18f, 19, OLS [ordinary least squares] 133, 145, 78, 139, 143 146–8t, 150–1t, 161(n11), see also pocket/palmtop computers 273, 275t personal services (sector) 130t, 162t Olsen, J.P. 247–8, 249f, 255(n9) personal trajectories and conjugal optimal matching analysis interactions 5, 38–60 43, 60(n5) conclusion 55–9, 60(n11) organizational capital 139, 143, 144, conjugal interactions (limited number 149, 161(n10) of styles) 47–53, 55, 58–9, ‘work design’ component 149 60(n7–8, n10) organizational change, skill-biased 135 data and method 41–2 hypothesis 138 five theses 41, 56, 60(n4) hypothesis confirmed 153 further research required 59 related to technological change 138–9 insertions in social structure 53–5, organizational ecology 106, 110, 56, 60(n9–10) 115–22, 124f, 128(n8) personal trajectories 42–7, 53, 55, organizational factors 154, 161(n13) 58, 60(n5) organizational population 110, 118, professional and relational dimensions 120, 123, 125, 126, 129(n14), of life courses 38–41 160(n4) ‘restricted variation dependent organizational practices (new) 139–40 on elements of social organizations insertion’ (hypothesis convergence/divergence of view 229, confirmed) 56 237(n18–19) results 42–55 supranational 222 personnel recruitment 107, 113 ORGANS (organizational variable) 144–5, physical capital 139 149, 150–1t, 152t, 153, physical complaints/disorders 67–8, 161(n13) 69f, 78 endogeneity 149, 152t, 153 Pierson, P. 240 Oschmiansky, H. 192 Piva, M. et al. (2003) 158, 320 Osterland, M. 207(n3) Santarelli, E. 320 Ostrom, E. 278(n3) Vivarelli, M. 320 Oswald, A.J. 11 planning 241, 244, 257 outsourcing 110, 193, 206 Plewis, I. 321 348 Index pluralization of life-course poverty 81, 83 hypothesis 22, 38–60, 207(n3) comparative research 14–15 conclusion 55–9 conceptualization and data and method 41–2 measurement 10 professional and relational dimensions ‘consensual definition’ 14 of life courses 38–41 Council of Europe’s definition 10, rejected 56 14–15, 29(n1) results 42–55 definition 10 pocket/palmtop computers predictors 22 used in Fribourg self-observation structural 11 method 64, 66, 77, 78 time-lagged effects 84–5 Pohl, R. 111 poverty line 11, 20, 191 political institutions (Swiss) xviii, 1, 5–6, poverty measures 10, 219–96 direct (level of consumption or relative democracy problems of new deprivation) 10–11, 12, regionalism 256–80 13–15, 29 internationalization and domestic indirect (income available) 10, 12, politics 221–38 13, 22, 24–5, 28–9 reforming the municipalities 239–55 poverty reduction 100 sustainable fiscal policy in poverty research federal system cross-sectional studies 11 281–96 income measure 10, 11 political parties 224, 231, 232f, relative deprivation measure 10, 11 236(n8–9), 285 ‘two traditions’ 10–11 political power 259 power, reputational 231, 232f, 233 political science 2, 4 printing (sector) 161t political system, reform 240, 244 private households 130n politics 85, 234, 235, 236(n5), 247 private life 167 democratic 261 private sector 111, 243, 259, 260t, 264 domestic 228 private–public partnerships 259 international (Switzerland ‘legally privatization 267 sheltered from’) 222 problem-solving 247, 249f, 259, 261 national and international production 106, 110, 136 ‘increasingly entangled’ 221 productivity 4, 109, 110, 119f, 121f, Politikverflechtung 241 123, 124f, 128(n7) polity (decision-making process) 221, increase 118 234, 235 industry-specific 128(n7) pollution 268 technology-bound component of Pommerehne, W.W. 292, 293, growth 122 294(n12), 320 technology-induced change 112, 115 Poole, M.E. 319 profession, change of 163, 167n, 182, Poor Britain (Mack and Lansley, 184–6, 187f, 189, 207(n3) 1985) 13–14 ‘difficulties not resolved’ 186 population 196f, 197, 209–17t, well-being 186f 285–90, 296t work characteristics 185f cantonal 295 professional trajectories 5, 197 urban 285, 286, 295 eight categories 42 Portland (USA) 258 intentional leave of absence 42 Portugal 59 interruptions (ill-health) 42, 44, 47 post-industrialization 11, 22, 38, interruptions (negative) 43–7f 60(n4), 191 interruptions (positive) 43–7f post-materialistic values 28 interruptions (travel) 42, 44 postgraduate programmes 2 interruptions (unemployment) 42, postmodern theorists 39 44, 47 Poterba, J.M. 294(n2) periods of education/training 42, 43–7f Index 349

periods at home (homemaker) 42–7 public opinion 56 professional activity (full-time) 42–7 public revenue 283, 285, 286, 288–91, professional activity (part-time) 42–7 294(n5, n9), 295, 296t residual categories 44, 47 public safety/security 16, 23, 24t, 25, retirement 43–7f 29, 31t, 33t, 292 see also employer/change of; public sector 111, 243, 259, profession/change of 260t, 293 professional trajectories: by sex public services 257, 262, 284, 291 gender-neutral type (hypothesis public spending 283, 285–93, rejected) 56 294(n5), 295, 296t interdependence between male and public transport 28, 267–8, 270–2t, female trajectories 57 279(n12, n15) men’s trajectory type ‘dominant’ 43, purchasing power 109 43f, 47, 55, 58 ‘pure relationship’ (Giddens) 59(n3) men’s trajectory type ‘minority’ 43–4, Putnam, R.D. 231 44f, 55–6 multinomial logistic regressions 54t, 55 qualifications 108, 111, 116, plural but structured 42–7, 60(n5) 127(n3) women’s trajectory type ‘back to quality employment’ 46–7, 47f, 47, 54t, management 279(n16) 55, 56, 60(n9) public bureaucracy 291 women’s trajectory type ‘full-time public goods 293 worker’ 44, 45f, 47, 54t, public services 291 55, 56, 58 social relationships 165 women’s trajectory type ‘homemaker’ see also conjugal quality 44, 45f, 47, 53, 55, 56, 58 quality circles 138, 144 women’s trajectory type ‘part-time quality of life 6, 29, 163, 164 worker’ 46f, 46, 47, questionnaires 13, 14, 63, 65, 54t, 55, 56 66, 78, 160–1(n5), 165t, Proportional Deprivation Index (PDI) 202, 229, 237(n18), (Halleröd) 14–15, 19–20, 25, 279(n15–16) 30(n5–6) contradictory or non-plausible four European countries 20f answers 142 level of deprivation by quintiles of non-respondents 136, 165–6 equivalent income 20, 21f non-response analysis 142 weighting 15 ‘postal survey’ 136 Proposition 13 (USA) 282, 291, 294(n3) response rates 42, 142, psychiatry 78 166, 255(n4) psychology 2, 4, 40, 62–3, 64, 86–7 selectivity bias 142 clinical and health (application of see also interviews automated FASEM evaluation) 77–8 race 83 coping mechanisms 83, 84 Radic, D. 311 psychosomatic complaints 169, 170f, Radical Party (Switzerland) 247 176, 177f Rainwater, L. 298 Ptacek, J.T. 62 Rasbash, J. et al. (1999) 67, 321 public administration/ Browne, W. 321 bureaucracy 130t, 130n, Draper, D. 321 232f, 259, 291 Goldstein, H.Y.M. 321 public choice perspective 258, 262, Healy, M. 321 263, 273–6 Plewis, I. 321 public debt 281 Woodhouse, G. 321 public goods 293 rational choice 247 efficiency of supply 294 Raudenbush, S.W. 67 production and consumption 262 Rechtsstaat 261 350 Index referenda 245, 279(n9), 283 schools of thought 257, 278(n3) compulsory/mandatory 223, 244 spillovers 259, 266 fiscal 284, 292–3 see also citizen loyalty; democratic ‘induce lower spending and legitimacy; service satisfaction revenue’ 282 regionalism (old) 256, 260, 260t, optional 223, 236(n7) 263, 278 popular initiative 223 ‘better level of performance’ 274, 277 predicted probabilities of successful critics 257–8 launching 229f local government 269, 270–2t ‘threat’ 224, 229, 234, 236(n7) multi-tiered government 269, 270–2t regionalism (new) 256, 257, 258–61, regulation (concept) 48, 49t, 53, 57 262, 263–4, 278, 278(n4–7) ‘Regulation of Emotions in Families with complex network governance 269, Adolescents’ (Fribourg project) 270–2t 61, 63, 78 linked-functions governance 269, First project 70 270–2t Second project 66, 67–8, 72, 75, 76 regionalism (new): democracy in Third project 66, 68, 75, 76 metropolitan areas 264–77, Reicherts, M. 64 278–9(n8–16) Relationship Assessment Scale 76 area-wide governance: case-study relationship satisfaction 73–4, 76 evidence 267–9, 270–2t, relative deprivation (Townsend) 10, 11, 279(n12–13) 12, 19–22, 25, 29, 30(n7–8) context (institutional fragmentation concept 13–15 and economic disparities) 265–6, ‘enforced lack of socially perceived 278–9(n9–11) necessities’ (Mack and impact of new regionalism on Lansley) 14 democratic governance 269, measurement 15 273–7, 279(n14–16) relationship with household policy fields 264 income 20–1 regionalism (new): democracy relative deprivation and well-being 9–37 problems 256–80, 278–9(n1–16) chapter aims 9 aim of chapter 257 concept 13–15 case-study evidence 257, 267, concluding remarks 25, 28–9 279(n12), 279–80 consequences of relative deprivation conclusion 277–8 for well-being 22–5, 26–7t conventions of interaction 260t measuring relative deprivation 15, introduction 256–7, 278(n1–2) 30(n5–6) methodology 279(n12), 279–80 research questions 12 new regionalism and democracy in standard of living and relative metropolitan areas 264–77, deprivation 16–22 278–9(n8–16) theoretical and methodological policy implications 278 debates 9–12, 28–9 regionalism and democratic religion 245, 246f governance: three res publica 262 hypotheses 256, 257, 261–4, 269 residual categories 44, 47 hypothesis 1: no significant Reskin, B.F. 312 relationship between new retirement 39, 87, 94, 96, 97, 98 regionalism and service early 199, 204 satisfaction 262, 269, 273–4 pensions 191 hypothesis 2: new regionalism private pension plan 16, 17f, 18f, 19 increases citizens’ loyalty towards Reulier, E. 284 local authorities 263, 274–5 Reyher, L. et al. (1990) 113, 321 hypothesis 3: new regionalism reduces Kretschmer, G. 321 democratic legitimacy 264, 275–7 Spitznagel, E. 321 routes towards metropolitan Rhodes, M. 221 governance 258–61, 278(n4–7) Rhodes, R.A.W. 259, 278(n7) Index 351

Ribeiro, T. 297 Scharpf, F.W. 221, 236(n1), 240, 276 Richards, M. 61, 62 globalization 234 Ries, A. 303 Schaufeli, W.B. 325 Risse, T. 222, 238(n24), 309 Scherpenzeel, A. xv, 5 Risse, T. et al. (2001) 238(n26), 321 Schimmelpfennig, A. 110 Caporaso, J. 321 Schmid, G. 192 Cowles, M.G. 321 Schmidheiny, K. xi Roberts, J. 138 Schmidt, J.U. 164 Roberts, J.A. 292 Schmidt, M.G. 5 Rochat, S. 312 Schmitter, P. 227 Rodgers, G. 191 Schnellenbach, J. 293 Rodgers, J. 191 Schoebi, D. xv, 66, 79, 320 Ronning, G. 128(n7) scholarships 3 Rook, K.S. 84 Schömann, K. 127(n3) Rosenstiel, L. von 164 schools/schooling 39, 69f, 70, 241, 244, Rossi, A.A. 278(n8) 245, 250, 291–2 Rotterdam 258 Schreurs, P.J.G. 325 Roubini, N. 294(n5) Schuler, M. 278(n8) Roy, K.-B. 192 Schumann, M. 107 Rozun, E. 78 Schwab, B. 278(n13), 314 Rozun, I. 78 Sciarini, P. xv, 6, 224, 227, 228, 231, Rubin, D.B. 142 236(n3), 307, 319 Rubinfeld, D.L. 292 Sciarini, P. et al. (2000) 236(n3), 323 Rueben, K.S. 292 Dupont, C. 323 rules of game 259, 260t, 263 Hug, S. 323 Rusk, D. 278(n4) Sciarini, P. et al. (2002) 236(n4), Ryff, C.D. 84–5, 98 237(n11), 323 Fischer, A. 323 Sacchi, S. xv, 5, 111, 113, 160(n4), 302 Nicolet, S. 323 Sachs, J.D. 294(n5) Sciarini, P. et al. (2004) 222, 225, Sacker, A. et al. (2001) 86, 322 236(n6), 323 Bartley, M. 322 Fischer, A. 323 Firth, D. 322 Nicolet, S. 323 Fitzpatrick, R. 322 science and technology (module) 2 St Gallen (canton) 284 sciences/scientists 115, 130t, 132t, 140, Saint Paul-Minneapolis 258 251f, 291 salaries 147t, 150t, 152t, 155t, 190, Scott, M.A. 300 194–9, 200n, 203f, 203n, sectors (economic) 106, 109, 207(n2), 209–17t 161(n13) low 114 formal education of employees sales (export share) 152t, 153 (1999) 141, 141t salespeople 115–18, 132t, Segalman, R. 1 164–6, 172–6 Seginer, R. 319 Salvisberg, A. xv, 5, 113 Seifert, H. 207(n4) Sanders, M. 160(n2) Seim, K. 106, 160(n2) Santarelli, E. 320 Seismo (publishing house) 4 satisfaction see well-being self-development 40–1 savings [personal investments] 17f, 17, self-efficacy 176 18f, 19, 28 self-employed/self-employment 191, Savioz, M.R. 293 195, 197, 201, 202, 204, 208(n6, n8), Savitch, H. 259, 260t, 278(n5) 209–17t Saxony (Germany) 207(n1) self-esteem 77, 178, 180 Scandinavia 139 language differences 183f Schallberger, U. 183 Semmer, N.K. xv–xvi, 5, 181, Schaltegger, C.A. 282, 283, 293, 183, 188, 189, 305, 309, 294(n2, n4) 311, 312, 326 352 Index

Semmer, N.K. et al. (1996) 181, 183, 323 determinants 108–11 Baillod, J. 323 development of overall employment Stadler, R. 323 and advertised jobs by Gail, K. 323 industry 129–30 Semmer, N.K. et al. (2004) 166, 323 econometric studies 106 Elfering, A. 323 educational level (1950–2000) Grebner, S. 323 106, 132t Semmer, N.K. et al. (in press) 189, 324 hypotheses 112 Beehr, T.A. 324 industry and firm size 115, 116, 120 McGrath, J.E. 324 introduction 105–7, 127(n1) Sen, A.K. 13 job advertisements (empirical Sengenberger, W. 111 analyses) 106, 114–23, 124f, Sennett, R. 191 128–9(n8–18) sequentialization (Kohli) 39 job advertisements (representative SER (standard error) 287, 288, 290, survey) 112–14, 123, 128(n4–7) 294(n8), 296t lack of long-term, differentiated, Serdült, U. 295 empirical studies 106 service delivery 240, 241, 261 literature 106, 107, 108 service industries 140 manual and craft 108 service provision 258 methodology 123, 125 private 259 ‘missing link’ in existing service satisfaction 256, 257, 261, research 108–9 262, 269, 273–4, 276t, 277, occupational composition 278, 279(n15) (1950–2000) 106, 111–12, 114, hypothesis: no significant relationship 115–20, 122, 123, 132t between new regionalism and performance of job-specific service satisfaction 262 tasks 107, 108, 109, 127(n2) service sector 91, 108, 109, 135 personnel recruitment 107, 108, 109 formal education of employees (1999) policy implications 126–7 141, 141t qualitative shifts 105 ‘losers’ 110 research findings 125–6 services 118 shifts (two components) 110 improvement 252 skill levels 112, 115, 120–3, 124f, quality 253f 129(n15–18) Sexton, T.A. 282 Swiss context (specific aspects) 111–12 Shapiro, P. 292 theory 107–12, 127(n2–3) Sheffrin, S.M. 282 upgrading (general) 108, 125 Sheldon, G. 111, 160(n4) validating the indicator of skill SHP see Swiss Household Panel level 130, 131f siblings 66 skill level (job advertisements) Sidjanski, D. 224 no qualifications required 120, Siegrist, H. 314 121f, 122, 125–6, 127, Siegrist, J. 189 129(n15), 132t Simpson, J.A. 73 occupation-specific vocational Singer, B. 84–5, 98 qualifications required 120, 126, single parents 86, 100(n2) 129(n16), 132t skill demand (long-term dynamics, supply and demand 122, 129(n15) 1950–2000) 5, 105–34, 160(n4) tertiary-level qualifications case studies 106 required 120, 123, 124f, coding of industry and firm size 133 125–7, 132t computation of observed and skill polarization (Kern and model-based trends 133–4 Schumann) 107 conclusions 123, 125–7 skills 107, 113, 136, 139, 160(n4) credential inflation 108 demand equations 139 cyclical fluctuations 129(n10) formal 108 data validation (four findings) 113 supply 160(n1) Index 353 skill upgrading 111, 122, 138 social structure 38, 39, 41, 57, 98 Slaughter, M.J. 136 insertions 53–5, 56, 60(n9–10) Slovenia 12, 29(n2) social support 65t, 66, 78, 84, 87, 89f, ‘absolute necessities’ 17f, 17 91, 93t, 94–5, 96–9, 101, 163, 168, actual standard of living 18f, 18–19 172f, 173f, 174, 175, 177, 179, direct and indirect measures of 180, 189 poverty 29 see also appreciation at work factorial analysis for well-being 33t socialization of values 165 GNP per capita PPP 12–13, 30(n3) socio-demography 24, 43, 87, 273, income equality 30(n4) 273t, 274, 276, 276t income quintiles 21f, 21 Socio-Economic Panel (SOEP, individual and societal well-being (and Germany) 201, 202, their determinants) 23–5, 203, 208(n6–7) 26–7t, 34–7t Socio-Economic Panel Group measuring relative deprivation 15 (Germany) 201 relative deprivation 19–20, 20f, 29 socioeconomic status 83, 97, 100 standard of living 28 sociological studies/sociology 4, 38, small states 227 59(n1), 62 Smart, M. 291 SOEP see Socio-Economic Panel Smeeding, T.M. 298 (Germany) Smith, H.L. 304 Solothurn (canton) 284 Smith, R.E. 62 somatic states 63, 65t, 65, 71, 74, 79 Smith, V. 191 Sopp, P. 300 Snyder-Duch, J. 128(n6) Spain 28, 29(n2), 59, 281 social contacts 69f, 70 Spitznagel, E. 113, 321 social context 23, 85–6 spouses 42, 73 social and cultural circumstances 83–4 degree of autonomy 41 Social Democratic Party Spycher, S. 106, 125, 160(n4) (Switzerland) 247 Stadler, R. 323 social dumping 234 Staib, D. 298 social exclusion 9, 20, 29 standard of living 1, 9–15, 16–22, social inequalities 23, 24t, 25, 29, 31t, 32t, impact on health 81–101 33t, 82–3, 191 social inequality (module) 2 minimum 12, 14, 15, 17, 18, 22, 28 social insurance 190, 194, 195, 201, 206 perception and reality 16–19 social integration 84, 160 relative deprivation 19–22, 30(n7–8) social mobility 22 standardization of life courses ‘elevator effect’ 22 (hypothesis) 39–40 social networks 57, 83–4, 85, 87, 91, 98 state, the 111, 252, 256, 259, 264 social origins 81 legitimacy 261 social position 86–7, 97, 99, 100, 101 relationship with citizen 261 social psychology 62 state actors/executives 226, 232f, Social Science Research Centre, Berlin 235, 260t (WZB) 29(n2) autonomy 237(n14) social sciences 2–4, 82 empowerment 230–3, 237–8(n20–3) state-of-the-art (Switzerland) xviii, 4 federal 227 social segregation 265–6, 267, 279(n10) state sector social services for drug users 268, downsizing 254 270–2t, 279(n12, n15) states [countries] 236(n1), 260 social status 40, 41, 53, 57, 83 states [in federal systems] 282 parental 91, 94, 98 statistics 3–4, 113, 235 social stratification 11, 29, 40 Stauffer, T.P. 294(n6), 295 post-industrial societies 22 Steinberg, J. 1 ‘Social Stratification, Cohesion, and Steiner, R. xvi, 6, 315 Conflict in Contemporary Families’ Steinmüller, P. 300 (project) 38, 41–2 Stern, S. 149 354 Index

Stoker, G. 259, 278(n7) Swiss Labour Force Survey stress 64, 65t, 65, 67, 74, 80, 84, 163, (SLFS, 1991–) 194–5, 164, 165 196f, 197, 201, 204 due to conflict 75 response rate 195, 207–8(n5) life events 86 Swiss National Science Foundation xii, stress symptoms 169 xiv, 81n, 87, 135n, 190n, stressors 177 278(n1) private 169, 170, 187, 187f, 188 see also Zukunft Schweiz – Demain social 166–8 la Suisse task-related 166–8, 171, 172f, 172, Swiss Parliament 235 174, 184, 185f, 189 bicameral 228, 237(n15) work 186, 187 ‘domestic’ legislative acts 223, 225, structural equation models 81, 223 229f, 234, 236 structuralism 22 federal law 224, 236(n7), 236–7(n9) Stutzer, A. 283, 293, 295 legislative acts (1995–9) 223, suburbanization 264, 265 225, 229, 235, 236(n5), suburbs 268–9, 272t 237(n16) annexation by core cities 265 Switzerland 29(n2) Suh, E.M. 304 ‘absolute necessities’ 17f, 17 summer school 2–3 factorial analysis for well-being 33t supply and demand GNP per capita 12–13, 30(n3) highly-skilled workers 140 government departments 236–7(n9) skills and technological change 136, income equality 13, 30(n4) 160(n1) individual and societal well-being Suter, C. xvi, 4, 5, 30(n8), 83 (and their determinants) 23–5, Sweden 15, 21 26–7t, 34–7t Swiss Business Federation 238(n25) measuring relative deprivation 15 Swiss census data 113, 130, 131f, 195, neutrality 222, 236(n3) 197, 198n, 199, 200n, 201, 204, relative deprivation 19–20, 20f 205, 208(n6) ‘special circumstances’ 5 Swiss Confederation 265 standard of living 18f, 18–19, 28 Swiss economy 1 test case (impact of firm-level evidence 135–62 internationalization) 222 intra-sectoral adaptations 105 working experience 5 relative decline 4 see also Contemporary Switzerland; sectoral composition 105 labour market; political uncertainties 188 institutions; way of life Swiss Establishment Census Switzerland: federal companies 232f (1955–98) 130n Switzerland: Federal Council (Swiss Swiss exceptionalism 12 Government) 232f, 236–7(n9) Swiss Federal Financial Switzerland: federal government 241, Administration 295 255(n3) Swiss Federal Institute of Technology, Switzerland: Federal Office of Zürich (ETHZ) xi, xii, xv, 164 Communications (OFCOM) 231 ETHZ: Swiss Institute of Business Cycle symptom reporting 69f, 70, 71 Research 127(n1) Szakaly, K. 282 Swiss Federal Statistical Office (Bundesamt für Statistik) 88, 130n, t-statistics 294(n8) 194–5, 194, 196f, 198n, 200n, Taris, T.W. et al. (2001) 189, 325 279(n10), 295 Le Blanc, P.M. 325 Swiss federalism 265 Peeters, M.C.W. 325 lost dimension (democracy problems Schaufeli, W.B. 325 of new regionalism) 256–80 Schreurs, P.J.G. 325 Swiss Household Panel (SHP) Survey 3, task performance 149, 151n 22, 81, 81n, 87, 92–3t, 95t, 97 task requirements 111 Index 355 tax time 64–6, 78, 81, 88, 89f, 99, 101, 106, base 282 116, 120, 123, 125, 128(n7), 167n, burden 295 169, 171, 173, 176–7, 183–5, 202, competition 282, 284, 286, 288–91, 228, 283–4, 287 295t, 296t days of week 68, 68f, 69f morale 293 delayed effects of social revenue 283, 284, 286, 288–91, environment 83, 84–5 294(n5, n9), 296t weekends and weekdays 70 taxation 262, 282 time control 168 broad-based 293, 294(n13) Tobit estimates 161(n11) communal 266, 266t, 273, 273t, Torgler, B. 293 274, 276t, 279(n11) Toronto 258 cost–benefit ratio 273, 274 Touraine, A. 109 taxpayers, wealthy 265 Townsend, P. 13, 20, 83 Taylorism 118 trade unions 193, 232f, 233–4, 235, teachers 115, 130t, 132t, 292 238(n24–6) teamwork 109, 135, 138, 140, transaction costs 294 142, 144, 146t, 149, 151n, Transformation of Intimacy: 158n, 159 Sexuality, Love, and Eroticism technological change 118, 120, 126, in Modern Societies 128(n7), 159 (Giddens) 59(n3) accelerated by increased supply of transition (post-communism) skills 160(n1) 12, 201 impact on skill demand 112 transparency 160, 227 related to organizational transport/communications change 138–9 (sector) 129, 130t, 162t sector-biased 136, 145 treaties/international agreements 225, ‘skill-biased’ hypothesis 106, 109, 228, 236, 237(n20) 122, 125, 135, 136, 137, 159 229f, 237(n16) within-sector 136 direct internationalization 223 technological factors 154, 161(n13) ratification 228 technological innovation 108 Trechsel, A. 295 technology 115, 116, 146t Trinczek, R. 111 technology hypothesis 136 Troske, K.R. 304 TECHNS (technology variable) 144, trust 257, 259, 274, 279(n16) 145, 150–1t, 152t, 153, Tschan, F. xvi, 305, 311, 312 161(n12) Tschan, F. et al. (2004) 166, telecommunications 238(n27) 180, 326 telephones 16, 17f, 18f, 88, Inversin, L. 326 195, 224 Semmer, N.K. 326 television sets 16, 17f, 18, 18f, 28 Turkey 29(n2) ter Weel, B. 160(n2) two-level game (Putnam) 231 tertiarization 109–10, 120, 122, 2SLS estimates 149, 152t, 153 160(n4) tertiary-level qualifications Uçarer, E.M. 222 required 120, 132t Umberson, D. 311 Tessin (canton) 243 Umlandverbände (Germany) 256 Teutsch, M. 310 unemployment 5, 9, 21, 22, 87, 96, Thoits, P. 84 97–8, 101, 105, 113, 192, 201, 202, Thom, N. 113 204, 206, 207(n1), 209–17t Thompson, G. 111 fear of 175 Thurgau (canton) 243, 278–9(n9) insurance 127 Ticino (canton) 278–9(n9) Union Bank of Switzerland 286 Tiebout, C.M. 258 United Kingdom 15, 21, 194, 160(n2), Tillmann, R. 302, 315 154, 155t 356 Index

United Nations 5, 222, 235 Watzek, D. xvi ‘Blue Helmets’ Corps 230 way of life (Swiss) xviii, 1, 4–5, 7–101 United States of America 4, 59, emotional regulation in families with 139, 154, 155t, 160(n2), adolescents 61–80 194, 207, 281 health (impact of social health care expenditure 82 inequalities) 81–101 individual states 282, 291–3, ‘normalization’ 5, 10, 12 294(n13) personal trajectories and conjugal job advertisements 107 interactions 38–60 pluralization of life-course pluralization of life-course hypothesis 58 hypothesis 38–60 universities xi–xvii, 3, 4, 81n, 88, 143, relative deprivation and 164, 165t well-being 9–37 university degrees 120, 141t, 293 wealth 11, 29, 83 upskilling 107, 136 Weber, M. 39 urbanization 264, 278(n8), 296t websites 30(n8, 10, 12–13), 113, 139, 161(n5), 236(n4), 294(n1) vacancies 112, 130 Weck-Hannemann, H. 293 Van den Berg, L. et al. (1993) weekend effect 70, 71 278(n4), 326 Weerapana, A. 293 Van Klink, A. 326 welfare state 10, 12, 292, 293 Van der Meer, J. 326 comparative research 85–6 Van den Bosch, K. 18 well-being 5, 9–37, 63, 67, 70, 71, 76, Van Klink, A. 326 79, 81, 83, 84, 87, 168, 170f, 178–80, Van der Meer, J. 326 182, 185–9 Van Reenen, J. 125, 126, 139, 155t, consequences of relative 160(n2), 161(n8) deprivation 22–5, 26–7t, Vancouver 258 30(n9–13) variables factorial analysis 33t economic, demographic, and political individual 9, 23, 29, 30(n11), 31t, control 284–90 169–70 Vatter, A. 293, 294(n4) individual and interpersonal 74 Vetter, A. 239, 241 individual and societal (and their Vickery, G. 160(n5) determinants) 23–5, 26–7t, 29, violence 51, 62 30(n11–13), 34–7t Vivarelli, M. 320 influences from work, private life, and Vogel, R.K. 259, 260t, 278(n5) personality 187f voluntary sector 259, 260t list of variables 31t, 31–2t voting system 244 measurement 22–3, 30(n9–11) prediction 187–8 wages 143, 155t, 161(n8), 190, 194–9, relationship with income 11–12 200n, 203f, 203n, 207(n2), sex differences 176–8 209–17t societal 9, 23, 30(n11), 31t low 111, 234, 265 subjective 11–12, 22, 23, 28, 29, 53 Wagner, A. 192 work-related 183 Wagner, G. 192, 193, 204 Werner, H. 191 Wallis (canton) 294(n6) White procedure 148n, 151n Wälti, S. 278(n4) Wichtermann, J. 315 Walwei, U. 191, 192, 202, Wicki, M. 113 204, 206 Wicksell, K. 292, 293 Warnken, J. 128(n7) Widmer, E. xvi, 5, 79, 195, 207(n3) water supply 16, 162t, 267, 270–2t, Widmer, E.D. 42, 60(n5, n8, n10–11) 279(n12, n15) Wilhelm, P. xvi–xvii, 66, 67, Watkins, A. xviii 72, 73, 320 Watts, M. 192 Wilkinson, R.G. 82 Index 357

Wilkinson, R.G. et al. (1998) 82, 327 world market 5, 127 Kawachi, I. 327 world trade 160(n3) Kennedy, B.C. 327 Wothke, W. 223 Willems, E. 303 Wurzburg, G. 161(n5) Wilpers, S. 66 Wyder, D. xi Wimmer, A. 314 WZB (Social Science Research Centre, women 39, 42, 50, 52t, 53, 54t, 56–7, Berlin) 29(n2) 60(n9), 70, 176–8, 195 age-dependent rate of regular work young adults 5, 22 (Switzerland, 1970–2000) 200f employer, change of 182–4 empathic accuracy 73 entering workforce 163–89 employment 54t, 67, 91, 196f, 198f existing literature (point of departure first child 40 from) 164 labour force participation 191, 193, individual well-being 169–70 202, 204, 205, 207 lessons for employers 189 married 207 linguistic regions 178–80, 181–3f regular work (Switzerland) 196f, 199 occupations (differences regular work (western Germany) between) 171–6 202, 203f personal characteristics, private stability in standard circumstances, working employment 201 conditions 187–8, 189 see also professional trajectories profession, change of 184–6, 187f Wood, A. 136, 327–8 project overview 165t Woodhouse, G. 321 results 166–88 work 3, 38, 59(n1), 69f, 70, 100 sex differences in well-being 176–8 flexible 160 stability and change of working unpaid 195 conditions 167–9 unskilled 111 stayers versus leavers 182–4, 186 ‘Work Experience and Quality of Life in study 164–6 Switzerland’ (‘AEQUAS’ project) study: design and sample 165 163, 164–5, 165t study: sample bias and attrition 165–6 work organization 105, 106 sub-projects 165t changing pattern 109, 112 summary and conclusions 188–9 flexible 140 trajectories: changing jobs, changing workers, unskilled 115, 120–3, 125–7, professions 180–6, 187f 129(n17), 130, 131f, 156t, 157t, transition into work as time of 158t, 265 change 166–7 job market trends (observed and well-being (prediction) 187–8 expected) 121f worries about the future 170–1, 171f long-term decline in demand 122 youth 88, 89f, 98, 106, 296t working conditions 114, 150t, 153 financial disadvantage 92t, 94 wellbeing 163–89 social circumstances (health Working Conditions, Work Orientation, impact) 81 and Labour Market Participation (research grant) 190n Zachert, U. 192 working time 207(n4) Zarin-Nejadan, M. xvii flexible 144, 147t, 150t, 152t, 153 Zaugg, R.J. 113 workplace 107, 125, 127 Zimmermann, E. 97 computer-equipped 112 Zukunft Schweiz – Demain la Suisse reorganization 138 (research programme) xii–xviii, spatial division from 1–6, 135n household 207(n4) domains/modules 2 workplace organization 146–7t eight-year programme for demand for employees of different development of social educational levels 135–62 sciences 2–4 358 Index

Zukunft Schweiz – Demain la Suisse Zürich (canton) 243 (research programme) – continued Zürich (metropolitan area) 256, 257, Gouvernance métropolitaine et légitimitée 264, 266t, 267–9, 270t, project 278(n1) 273, 280t ‘problem-oriented basic research’ 2 Zürn, M. 221 Social Report 4 Zweckverband (administrative ‘structural measures’ 2 union) 255(n6) ‘two types of activity’ 2 Zwick, T. 154