Globalization, Value Change, and Generations
European Values Studies The European Values Studies is a series based on...
24 downloads
569 Views
1MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Globalization, Value Change, and Generations
European Values Studies The European Values Studies is a series based on a large-scale, cross-national and longitudinal research program. The program was initiated by the European Value Systems Study Group (EVSSG) in the late 1970s, at that time an informal grouping of academics. Now, it is carried on in the setting of a foundation, using the (abbreviated) name of the group (EVS). The study group surveyed basic social, cultural, political, moral, and religious values held by the populations of ten Western European countries, getting their work into the field by 1981. Researchers from other countries joined the project, which resulted in a 26-nations data set. In 1990 and 1999/2000, the study was replicated and extended to other countries. By now, all European countries are involved in one or more waves of the study, including those in Central and Eastern Europe. This series is based on the survey data collected in this project. For more information see: www.europeanvalues.nl.
VOLUME 10
Globalization, Value Change, and Generations A Cross-National and Intergenerational Perspective
Edited by
Peter Ester, Michael Braun, and Peter Mohler
BRILL LEIDEN • BOSTON 2006
This book is printed on acid-free paper. Library of Congress Cataloging-in-Publication Data Globalization, value change, and generations: a cross-national and intergenerational perspective / edited by Peter Ester, Michael Braun and Peter Mohler. p. cm. — (European values studies, ISSN 1568-5926; v. 10) Includes bibliographical references and index. ISBN-13: 978-90-04-15127-7 ISBN-10: 90-04-15127-3 (pbk.: alk. paper) 1. Social values—Europe. 2. Values—Europe. I. Ester, P. II. Braun, Michael. III. Mohler, Peter Ph., 1945– IV. Series: European values studies (Leiden, Netherlands); v. 10 HM681.G575 2006 303.3'72094—dc22 2006047470
ISSN 1568-5926 ISBN-13: 978 90 04 15127 7 ISBN-10: 90 04 15127 3 © Copyright 2006 by Koninklijke Brill NV, Leiden, The Netherlands. Koninklijke Brill NV incorporates the imprints Brill Academic Publishers, Martinus Nijhoff Publishers and VSP. All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission from the publisher. Authorization to photocopy items for internal or personal use is granted by Brill provided that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive, Suite 910, Danvers, MA 01923, USA. Fees are subject to change. printed in the netherlands
CONTENTS Figures and Tables .................................................................... vii Series Editors’ Preface ................................................................ xiii Preface and Outline of the Book .............................................. xvii Aknowledgements ........................................................................ xxi PART I
Chapter One Values and the Social Sciences: A Global World of Global Values? ...................................................... Peter Ester, Peter Mohler, and Henk Vinken
3
Chapter Two A Decline of Religious Values? ...................... Loek Halman and Thorleif Pettersson
31
Chapter Three Individualization of Family Values? ............ Jacqueline Scott and Michael Braun
61
Chapter Four Eroding Work Values? .................................... Peter Ester, Michael Braun, and Henk Vinken
89
Chapter Five The End of Traditional Political Values? ...... 115 Oddbjørn Knutsen Chapter Six Is the West Becoming More Tolerant? ............ 151 Nina Rother and Juan Díez Medrano Chapter Seven Is Volunteering Going Down? ...................... 179 Paul Dekker and Andries Van den Broek Chapter Eight Decreasing Desires for Income Inequality? ... 207 Stephen Fisher and Anthony Heath PART II
WHAT EXPLAINS DOMAIN-SPECIFIC VALUES? THE PERSPECTIVE ROLES OF GENERAL VALUE AND PERSONALITY TYPOLOGIES Chapter Nine Values, Personalities, and Generations .......... 233 Henk Vinken and Beatrice Rammstedt
vi
contents
Chapter Ten The Great Value Divide—Testing Hofstede’s Convergence Assumption .................................... 251 Peter Mohler, Beatrice Rammstedt, and Katrin Wohn Chapter Eleven The Power of Values, Personalities, and Generations ............................................................................ 273 Michael Braun and Henk Vinken Chapter Twelve Values, Value Change, and Generations: Grand Picture and Summarizing Perspective ...................... 303 Peter Ester, Michael Braun, and Peter Mohler Author Index .............................................................................. 315 Subject Index .............................................................................. 319 About the Authors ...................................................................... 325
FIGURES AND TABLES Figures 2.1
Results from a confirmatory factor analysis for a two factor model of religious commitment. Results for 13 countries from the EVS/WVS studies in 1981, 1990, and 1999/2000 ................................................................
3.1 3.2
Cross-national differences in pro-marriage attitudes .... Cross-national differences in pro-motherhood attitudes ............................................................................ 3.3 Cross-national differences in views about homosexuality .................................................................. 3.4 Cross-national differences in views about abortion ...... 3.5 Cross-national differences in views about divorce ........ 3.6 Cross-national differences in views about adultery ...... 3.7 Cross-national differences in sexual values .................... 3.8 Generational divide in pro-marriage attitudes .............. 3.9 Generational divide in pro-motherhood attitudes ........ 3.10 Generational divide in sexual values .............................. 5.1 5.2 5.3 5.4
5.5 5.6 5.7
6.1
Trends in equality versus freedom, based on opinion balance measure .............................................................. Trends in libertarian/authoritarian values ...................... Trends in materialist/post-materialist values according to the four-item battery .................................................. Cohort differences in emphasis of equality versus freedom. Differences between the war and the two other cohorts. Means for the three surveys .................. Cohort differences in libertarian/authoritarian values ................................................................................ Cohort differences in MPM values ................................ Mean value for 11 countries on the MPM index for different cohorts in each of the surveys ........................
44 70 71 73 73 74 74 75 78 78 81
126 129 132
134 138 140 140
General tolerance—development over time across countries ............................................................................ 165
viii 6.2 6.3 6.4 6.5 6.6 6.7
figures and tables Intolerance towards ‘social deviants’—development over time across countries .............................................. Intolerance towards foreigners—development over time across countries ................................................................ Intolerance towards political extremists—development over time across countries .............................................. Intolerance towards ‘social deviants’ by generation, wave, and country .......................................................... Intolerance towards foreigners by generation, wave, and country .............................................................................. Intolerance towards political extremists by generation, wave, and country ..........................................................
10.1 10.2 10.3 10.4
German values for Hofstede .......................................... Ideal Schwartz Value Circle .......................................... Schwartz Value Circle—Germany (West) ESS 2002 ... Schwartz Value Circle—ZUMA experiment 2004 (West only) ...................................................................... 10.5 Inglehart’s index ALLBUS 2004 & ZUMA 2004 experiment ........................................................................
166 167 168 172 172 173 256 257 258 258 259
Tables 2.1 2.2 2.3 2.4
2.5
2.6
3.1 3.2 3.3 3.4
Changes in belief-oriented religion in 13 countries ...... Changes in church-oriented religion in 13 countries ... Changes in over-all religious involvement in 13 countries ...................................................................... Religious involvement, moral strictness, and civic-society involvement in 1981, 1990, and 1999/2000 in three generations .................................... Changes in the levels of compartmentalization between religious involvement and two secular orientations among 13 countries ........................................................ Standard deviations for religious involvement in 1981 and 1999 ..........................................................................
45 46
Pro-marriage attitudes .................................................... Pro-motherhood attitudes ................................................ Sexual values .................................................................... Pro-marriage attitude by generation and wave ............
71 72 76 79
47
48
50 53
figures and tables 3.5 Pro-motherhood attitude by generation and wave ........ 3.6 Sexual values by generation and wave ............................ 3.7 (Logistic) Regression of importance of marriage, importance of children, and sexual values on country dummies, sex, marital status, employment status, religion, church attendance, study waves, and generations .......................................................................... 4.1 Fifteen EVS work values, coded on facets .................... 4.2 Extrinsic work values ........................................................ 4.3 Intrinsic work values ........................................................ 4.4 Extrinsic work values by generation and wave .............. 4.5 Intrinsic work values by generation and wave .............. 4.6 Regression of intrinsic and extrinsic work values on country dummies, sex, employment status, study waves, and generations ..................................................................
ix 80 82
84 97 99 101 103 105
109
5.1 5.2 5.3
Indicators for economic left-right values ........................ 122 Indicators for New Politics values .................................... 123 R2 and change in R2 from forced stepwise regressions with the value measures as dependent variables ............ 144
6.1 6.2 6.3 6.4
General tolerance .............................................................. 162 Intolerance towards specific groups ................................ 163 General tolerance by generation and wave .................... 170 Intolerance towards specific groups by generation and wave .................................................................................... 171 Regression of intolerance towards ‘social deviants’, foreigners and political extremists on country dummies, sex, marital status, employment status, religion, church attendance, waves, and generations .................................................................. 174/175
6.5
7.1 7.2 7.3 7.4 7.5
Volunteering 1981–1999: doing unpaid work for at least one of eight types of organizations ........................ Summary statistics on volunteering, 1981–1999 ............ Generational differences in volunteering and the extent to which they explain differences between years ............ Relation between volunteering and the importance attached to six life domains, 1999 .................................. Relation between volunteering and work values, 1999 (adjusted odds ratiosa) ........................................................
186 187 189 192 194
x
figures and tables
7.6
Leisure values: their valuations by the Dutch population aged 16 and over by three generations .......................... 196 Appendix Table 1 ‘Currently doing unpaid voluntary work’ for eight types of voluntary organizations/fields of activity ........................................................................ 202–203 Appendix Table 2 Volunteering in three population surveys ................................................................................ 204 Appendix Table 3 Relation between the importance attached to six life domains and volunteering, 1999 .................... 205 8.1 8.2 8.3 8.4 8.5
Gini coefficients ................................................................ Mean of the maximum income ratio by country and year .................................................................................... Standard deviations of the maximum income ratio by country and year .............................................................. Mean of the maximum income ratio by country, year, and cohort .......................................................................... Pooled and separate country regressions of maximum income ratio on cohort, age group, year, country, and year country interactions ..................................................
10.1 Means and standard deviations of the five BFI-10 scales .................................................................................. 10.2 Intercorrelations of the BFI-10 scales .............................. 10.3 Correlations between Hofstede’s and Schwartz’s value dimensions .......................................................................... 10.4 Correlations Schwartz’s dimensions—Inglehart Index .... 10.5 Correlations Hofstede—Big Five ...................................... 10.6 Correlations Big Five—Schwartz ...................................... Appendix Table 1 Value types and value items ....................
211 218 220 222
223
261 261 262 263 264 265 271
11.1 Unadjusted explained variances and significant variables .............................................................................. 288 11.2 Number of significant effects of Hofstede, Schwartz, Inglehart, personality types, and demographic variables .............................................................................. 289 11.3 Regression of extrinsic work values on general values, personality types, and demographic variables ................ 291 11.4 Logistic regression of marriage outdated institution on general values, personality types, and demographic variables ...................................................................... 292–293
figures and tables
xi
11.5 Regression of sexual values (Sexval) on general values, personality types, and demographic variables ........ 293–294 11.6 Regression of left-right-scale on general values, personality types, and demographic variables ................ 295 11.7 Regression of importance of God on general values, personality types, and demographic variables ........ 296–297 11.8 Total and additionally explained variance by general values, personality types, and demographic variables in regression of key domain-specific variables .................... 298
SERIES EDITORS’ PREFACE This book is the tenth volume in a series on European values published by Brill Academic Publishers. The main purpose of this series is to present and distribute the main findings from the European Values Study, a large-scale cross-national and longitudinal research project on fundamental values in Western societies. The project was initiated by Jan Kerkhofs of the Catholic University of Leuven (Belgium) and Ruud de Moor of Tilburg University (Netherlands). At the end of the 1970s, they established the European Value Systems Study Group (EVSSG ), aiming at empirically investigating the main fundamental value patterns of Europeans. They succeeded in conducting a large-scale survey in 1981 in all countries of the European Community (EC) plus Spain. The study aroused interest in many other European as well as non-European countries, and other researchers joined the project and used the original EVS questionnaire in their country. In this way, comparable surveys became available also outside Western Europe. Apart from a large series of books on the findings for individual countries, several cross-national comparative studies were published by, e.g., J. Stoetzel (Les Valeurs du Temps Présent: Une Enquête Européenne, Paris: Presses Universitaires de France, 1983), S. Harding, D. Phillips & M. Fogarty (Contrasting Values in Western Europe: Unity, Diversity Change, London: MacMillan, 1986), and, L. Halman, F. Heunks, R. de Moor & H. Zanders (Traditie, Secularisatie en Individualisering, Tilburg: Tilburg University Press, 1987). In order to investigate changes in values, a replication study was necessary. A second wave of surveys was fielded in 1990 again in all EC countries (minus Greece), as well as in Scandinavia, some Eastern European countries, and several countries outside Europe. Ronald Inglehart from the University of Michigan, Ann Arbor, has been very active and successful in getting the survey fielded in the early nineties in countries other than those participating in the European Values Study. This worldwide project is called the World Values Surveys (WVS ) and his group managed to conduct a wave of surveys in 1995–1996 in many countries inside and outside Europe. For results and more information on the EVS 1990 study, we can
xiv
series editors’ preface
refer to many publications, for instance, D. Barker, L. Halman & A. Vloet (1992) The European Values Study 1981–1990, London: Gordon Cook Foundation; S. Ashford & N. Timms (1992) What Europe thinks, Aldershot: Dartmouth; and, L. Halman & A. Vloet (1994) Measuring and Comparing Values in 16 Countries of the Western World, Tilburg: WORC. To monitor further changes in values and explore the dynamics in attitudes, beliefs, and values in Europe, a third wave of surveys has been conducted in almost all European countries in 1999/2000. A source book on this 1999/2000 EVS study has been published in 2001 (L. Halman (2001) The European Values Study: A Third Wave, Tilburg: EVS, WORC, Tilburg University). In collaboration with the World Values Surveys a source book was produced presenting the data of more than 80 countries all over the world (Inglehart, R., M. Bazáñez, J. Díez-Medrano, L. Halman & R. Luijkx (2004) Human Values and Beliefs, Mexico: Siglo XXI Editores). Cleaning, documentation, and distribution of the data from the European Values Study was done in collaboration with Hermann Duelmer, Evelyn Brislinger, and Wolfgang Zenk-Moeltgen at the Zentralarchiv für Empirische Sozialforschung (ZA) in Cologne. Helga van Gelder, Cor van der Meer, and Berry Feith from the Netherlands Institute for Scientific Information Services (NIWI) in Amsterdam also cleaned parts of the European Values Study. We are indebted to these people, and to several students-assistants from Tilburg University, who assisted in processing the data. We are grateful to the data archive of the European Values Study, the Zentralarchiv, and to the Research Institute of the Faculty of Social Sciences at Tilburg University for supporting the European Values Study project in numerous ways. The international data set is deposited at the Zentralarchiv in Cologne and can be obtained from there. For more information about the European Values Study and recent developments, see the EVS website: www.europeanvalues.nl and the website of the Zentralarchiv in Cologne: www.gesis.org/za. For information on the world values surveys we can refer to their website: www.worldvaluessurvey.org. This book is, as mentioned before, the tenth volume in this series. The first five volumes in this series were based on the 1990 and/or 1981 data, and were published by Tilburg University Press. The first volume was The Individualizing Society; Value Change in Europe and North America (1993; revised edition 1994), edited by Peter Ester, Loek
series editors’ preface
xv
Halman & Ruud de Moor. The second book Values in Western Societies (1995) was edited by Ruud de Moor, and the third book Political Value Change in Western Democracies (1996) was edited by Loek Halman & Neil Nevitte. The fourth volume From Cold War to Cold Peace (1997) was a comparison of Russian and European values. The authors were Peter Ester, Loek Halman, and Vladimir Rukavishnikov. This book has also been published in Russian in 1998. In 1999, a fifth volume was released on Religion in Secularizing Society, edited by Loek Halman and Ole Riis. A second printing of this book appeared in 2003 in the re-established series on the European Values Study now at Brill Academic Publishers. In the sixth volume, edited by Wil Arts, Jacques Hagenaars, and Loek Halman, The Cultural Diversity of European Unity (2003), Europe’s values are examined from an economic, political, social, and religious-moral point of view. The focus was not only on the current value patterns but also on the dynamics of value changes over time using the data from the three time points, and all contributions deal with the relationships between values and societal structural characteristics. The seventh volume dealt with Europe’s diversity and homogeneity at the turn of the millennium in terms of fundamental value orientations. This volume was titled European Values at the Turn of the Millennium (2004), edited by Wil Arts and Loek Halman, and it contained chapters in which authors from all quarters of Europe tried to identify patterns in the basic values and attitudes of Europeans round about the year 2000. They also offered explanations and interpretations of these patterns. The emphasis in this volume was on phenomena connected with civil society/citizenship, family and work, and religion and morality. In 2005, an Atlas of European Values was produced as the eighth volume in this series. This publication differs from the books published before in the sense that it presents the ideas, values, and beliefs of Europeans in for the general public more accessible form of graphs, charts, and maps instead of advanced statistics. The atlas provides answers to questions like Who are the Europeans? How do they think? What values do they hold dear? What binds them and what divides them? The Atlas combined data from the European Values Study, the World Values Surveys, and the Eurobarometer. It covers all European countries from Iceland to Turkey, from Portugal to Ukraine. The authors are Loek Halman, Ruud Luijkx, and Marga van Zundert (see also: www.atlasofeuropeanvalues.com). Tony Fahey,
xvi
series editors’ preface
Bernadette C. Hayes, and Richard Sinnott have written the ninth volume in the series. Their study on Conflict and Consensus. A study of values and attitudes in the Republic of Ireland and Northern Ireland presents a detailed comparison of cultural values and attitudes in the Republic of Ireland and Northern Ireland. It is based on survey data covering the period from the 1970s to 2003 but focusing especially on the European Values Study (EVS) as fielded in the two parts of Ireland in 1999–2000. The study confirms the deep divisions in identity and political allegiance that separate the Catholic and Protestant communities in Northern Ireland. But it also shows that, on many issues, Protestants and Catholics on the island of Ireland are culturally more similar to each other than to any other national population in Europe, including Britain. The current, tenth volume in this series, addresses a number of very interesting comparative themes: whether value convergence or divergence has occurred; whether traditional values are indeed on the decline as predicted in so many theories of modernization or not; and whether younger generations play the proverbial vanguard role in culture shifts. The editors Peter Ester, Michael Braun, and Peter Mohler, invited a number of scholars to address these issues with respect to the societal domains of their expertise, such as religion, morality, work, tolerance, family, equality, and volunteering. Further, they explicitly asked the authors, wherever possible, to compare the European findings with those for the United States because such a comparison enhances the scope of the analyses. The comparison enables the authors to assess whether the findings are quite similar on both sides of the Atlantic or whether they point to fundamental differences between Europe and the United States. Wil Arts & Loek Halman Gouda & Tilburg, May 2006
PREFACE AND OUTLINE OF THE BOOK Values are a hot topic in Europe, both in the social sciences and in the public and political debate. The heated discussion on what constitutes the cultural borders of the European Union clearly reflects the vitality of the value issue. Is Europe indeed a community of values—values that refer to a common past, a common present, and a common future—and if so, how united are Europeans in their fundamental values? Values seem to matter even or maybe particularly in a globalizing era. At the same time there is broad concern about the vanishing of certain traditional values such as the decline of religious values, the weakening of family values, the erosion of work values, and the waning of values that are at the core of civil society such as volunteering and other civic virtues. Moreover, there is widely voiced concern that due to the assumed European convergence of value systems, national identity—including nationally unique value patterns—will gradually dissolve. But with the decline of certain basic values, new values may come to the fore as brought about by new generations. In this sense, value changes may reflect societal turbulence or more gradual trend developments caused by the dynamic process of generation renewal. The private and public discourse on basic values—and their assumed “deterioration”—in contemporary culture and society is full of pitfalls, wishful thinking, overt speculation, and conceptual peculiarities, many of which will be addressed in this book. As this discourse is not an innocent debate, we feel that the social sciences need to contribute to a more balanced discussion regarding the development of fundamental social, cultural, religious, and political values in late modernity. As empirical social scientists we feel that our best contribution lies in the empirical comparative approach of the major changes that take place in various value domains. In this sense, social scientists can play a key role in keeping the ongoing debate on value changes within firm boundaries of empirically observable trends in a wide variety of basic values among the population at large. And if possible: with a fresh shot of sociological debunking. We strongly believe that the empirical social sciences are increasingly well equipped to play a major role in the current discussion
xviii
preface and outline of the book
with regard to what constitutes the main trends of basic values in modern European societies. With the availability of large-scale crossnational longitudinal data sets on value changes, we as social scientists are in a much better position to make a distinctive contribution to the debate. And this is exactly the aim of this book and the underlying research project. For this purpose, ZUMA launched a substantial comparative research project on value changes by bringing together leading European value survey researchers. This book is the main result of this cross-national team of value specialists. Analyzing and understanding major trends in basic values is a fascinating subject and we hope that some of this fascination is mirrored in the various contributions in this book. The structure of the book The first chapter by Peter Ester, Peter Mohler, and Henk Vinken sketches the issues that inspired ZUMA to launch this major European project on the values of Europeans and by introducing its main themes: the question of value convergence or value divergence, the issue of a possible decline in traditional values, and the potential vanguard role of young generations in embracing new values. The following chapters of Part I of the book contain a series of comparative empirical contributions in which these three themes are applied to specific value domains. Loek Halman and Thorleif Pettersson analyze whether the domain of religious values is witnessing a clear and generationbound trend towards (further) secularization and a decline of traditional religious values. Jacqueline Scott and Michael Braun explore the main European trends with respect to family values in order to see whether a general shift towards self-expressive values can be observed, particularly embodied by younger generations. In the following, Peter Ester, Michael Braun, and Henk Vinken report on comparative and intergenerational trends in European work values, especially within the context of a European Union that aims at becoming the most dynamic and competitive region in the world. Oddbjørn Knutsen addresses the issue of empirical developments in the domain of political values and investigates the vital question whether new generations in Europe embrace new political values. Next, Nina Rother and Juan Díez Medrano examine trends in general and specific tolerance underlying increasingly multicultural and globalizing European nations and describe its
preface and outline of the book
xix
main determinants and generational correlates. Paul Dekker and Andries Van den Broek look at the main European developments within the domain of volunteering as a crucial indicator of civil-society participation and highlight the role of younger generations. Stephen Fisher and Anthony Heath conclude the first part of the book by comparing values regarding social equality, and the distinct position of different generations on this important dimension of social stratification. Part II of the book presents the main findings of the experimental phase of our project. The aim of this novel part of our project was to empirically compare classic value typologies as developed in the social sciences with each other and with classic personality-trait studies as developed in psychology. By proceeding so we hope to trespass traditional borders both within and between disciplines. Value researchers and personality researchers may have more in common than they realize. Henk Vinken and Beatrice Rammstedt introduce the logic and challenges of comparing the various multi-level dimensions of the Big Three of value research (Hofstede, Schwartz, and Inglehart) and the Big Five dimensions of personality traits (extraversion, agreeableness, conscientiousness, neuroticism, and openness to experience), also by taking into account the generations as a relevant level of analysis. Peter Mohler, Beatrice Rammstedt, and Katrin Wohn analyze the interrelationships between the Big Three and Big Five dimensions based on data collected among a small but representative sample of the German population. Finally, Michael Braun and Henk Vinken address the relative explanatory power of the value and personality-traits dimensions on key variables from the previous chapters, especially focusing on religious, family, work, and political values. In the final and concluding chapter of this book, Peter Ester, Michael Braun, and Peter Mohler summarize and combine the main trends in Europe on value change and the role played by generational replacement and put these trends into a wider perspective. Based on these overall conclusions, and taking into consideration the main results of the methodological study presented in the second part of the book, they set the agenda of value research in the next decade. The way we worked During the project, the authors of the chapters—all experienced and internationally renowned value researchers—met a number of times
xx
preface and outline of the book
(in Mannheim at ZUMA) to discuss the rationale of the book, its underlying ideas and concepts, the methods to be used, the data to be analyzed, first results, and final chapters. We thoroughly discussed and criticized the various contributions, and in doing so also advanced and improved the individual chapters. As such this book is an outcome of an interactive series of meetings and discussions via the Internet, which we all enjoyed enormously. Peter Ester Michael Braun Peter Mohler
ACKNOWLEDGEMENTS We would like to thank ZUMA for its generous support of this European value research project. Of course, we also want to acknowledge the various scholars that contributed to the book. It was a very stimulating and challenging experience working together with such an excellent team of value specialists. We are grateful to Loek Halman and Ruud Luijkx of the European Values Study for their permission to use the EVS data. Special thanks goes to Margit Bäck (ZUMA) for organizing the various author meetings and for her administrative and secretarial assistance throughout the project. Last but not least we would like to thank Miriam Baumgärtner (ZUMA) for her invaluable assistance during the final stage of producing this book.
PART I
CHAPTER ONE
VALUES AND THE SOCIAL SCIENCES: A GLOBAL WORLD OF GLOBAL VALUES? Peter Ester, Peter Mohler, and Henk Vinken The booming business of value research The cross-national study of empirical value patterns and value trends is an increasingly advanced and expanding area of social scientific research. The greater availability of large-scale databases on comparative trends in values and value patterns is of course a major impetus behind this flourishing academic enterprise. But it is not merely the availability of data that makes the comparative study of values a booming business.1 Particularly the perception of a rapidly globalizing world evokes basic questions in the realm of human values. Does globalization of the political, economic, media, and cultural domain imply that cross-national differences in fundamental value orientations will vanish? Or does it mean that the accentuation of value disparity will grow in importance? Will values based in specific locales lose their share on the market for ideological meaning or will they attract larger market segments, maybe beyond their original locale? Will values systems converge or diverge? To what extent does McDonaldization in economic and cultural affairs relate to fundamental ideological and moral ‘tastes’? These are pressing issues that impact not only the domain of the mind, but also the domain of everyday life. In the world of business and work, for instance, the emergence and wide dispersal of a ‘Davos’-culture is noted (Berger & Huntington, 2002). Around the globe people (usually men) in this world wear the
1 We also witness a wide variety of commercial best sellers—however superficial— on value trends (e.g., Faith Popcorn, Clicking: 17 trends that drive your business & your life, HarperBusiness, 1997 or John Naisbitt, Megatrends 2000, Avon, 1991). Similarly, the Internet is packed with references to cultural and value-related trends needed to do business today and stay in business tomorrow (see. e.g., the website of the World Future Society at www.wfs.org).
4
peter ester, peter mohler, and henk vinken
same salary-men suits, talk the same type of transnational language (a particular kind of English), apply similar notions to negotiation and bargaining processes, share the same values on what it entails to be a successful businessperson, employer, and employee, and thus refer to a common set of attitudinal references and behavioral heuristics. Daily ‘breaking news’ in the international political realm, furthermore, frequently depict cultural conflicts over basic values, much in line with the analysis of Huntington (1996) who foresees a ‘clash of civilizations’ along a cultural divide between the value systems of Western Christianity and Islamic fundamentalism.2 Intra-societal ethnical clashes in Western societies—which increasingly transform themselves into multi-cultural societies—are also often related to contrasting value patterns. The discussion on the enlargement of the European Union is—after all—a discussion on what constitutes the values of Europe, on who its cultural insiders and outsiders are, on what comprises its cultural borders (cf. Arts et al., 2003). Likewise, the influx of migrant minorities has led to an open debate in many European countries on what comprises the cultural entity and identity of the individual nation-state. The debate on value diversity as an indicator of cultural diversity is central to the late-modern discourse on multiculturalism. Globalization, in short, fuels a certain reflexivity on the condition humaine as embedded in basic anthropocentric values, a reflexivity that relates to real-life decision-making as the above mentioned examples show. Whether cultures, including values, tend to converge—totally or partially—or diverge is a much-debated topic in the social sciences and in the political arena (cf. Ester et al., 1994). A rather traditional hypothesis might go that due to globalization processes, values tend to converge across societies. The perceived globalization of media, economics, politics, leisure activities, music, and sports, hence, aligns with globalization of values. Increased value exchanges and broader cultural transactions result from rising travel and migration flows as well as ever-faster information and communication technologies. As a consequence, core values become less nation-specific and national identity will increasingly be exchanged for a global identity. This
2 Compare the analysis of Barber (1995) who holds that the individual nation state is threatened by two powerful but opposed global forces: the bloodless McWorld politics of profit and a bloody Jihad politics of identity.
values and the social sciences
5
process is not welcomed by everyone. A further unification of Europe beyond economic cooperation, for instance, troubles many national politicians and political observers, as they seem to fear that individual countries will lose their ‘national’ identity, their ‘national’ habits, and their ‘national’ culture. In short, that an all-encompassing European convergence of value orientations may become reality. One could also argue that globalization does not per se entail the rise of one globally converged and unified value system. Globalization is a matter of transactions between global (transnational), national, and sub-national players that simultaneously both serve the role of ‘emitters’ and ‘receivers’ of cross-border culture (in its broadest sense, hence, including values). Not only the United States (usually the prime example both scholars and laymen provide of globalization), but many other countries produce and emit culture beyond their national borders.3 The response to these forces is not limited to convergence to some sort of footloose global culture. The responses vary from complete adaptation (as the above mentioned traditional hypothesis suggests) to out-right rejection of these forces. In between are the two more interesting responses: coexistence and synthesis (Berger & Huntington, 2002; Berger, 1997). Coexistence first refers to the compartmentalization of cultures, with global and local cultures existing side by side without any significant merger: e.g., the aforementioned Davos-culture look-alike transforms into a ‘traditional’ local when taking off his or her transnational business suit (and mind-set and values) after working hours. The synthesis response refers to a newly produced mix of global, universal culture with the indigenous culture, a mix that differs from both originals. In accordance with these coexistence and synthesis responses is the idea that globalization of culture is counterbalanced by the rediscovery of national, regional, and local cultures. More global exchange and transactions in terms of culture lead to real quests and actions that cater to the question on what it is that constitutes one’s unique local, regional or national identity. Concrete examples are found in the rising popularity of traditional theater, seasonal festivals, typical fashions, native language songs, and for instance local artisanal (‘slow’)
3 See e.g., Hannerz (1996) for some fine examples. An excellent topical study on how internationally circulating (American, European, and foremostly Japanese) culture is borrowed (in Taiwan) in a remarkably inventive way is by Hsiao (2002).
6
peter ester, peter mohler, and henk vinken
foods. Less comfortable examples include the rise in nationalism and in the above mentioned ethnic clashes, which in the last decades can be observed in large parts of Europe.4 We are dealing with the so-called globalization-localization ‘paradox of modern culture’ (Featherstone, 1990).5 Precisely because of globalization, an anxious search for national and local identity can be found. Globalization is definitely not seen as the development towards one global melting pot of cultures or a ‘spaghettization’ of value systems. Much more, when we limit ourselves to the cultural realm, globalization can be seen as a process that gives impetus to rising cultural transactions at global, national, and sub-national levels. The result might well be the co-existence of globally and nationally or even locally inspired cultures. Another result might be the synthesis of these cultures into something new, something that consists of strong elements of both ‘other’ and ‘own’ origin cultures merging to a point of non-distinction. The key point is that globalization is likely to fuel counter-processes that emphasize the vital importance of national (and local) cultures and communities for peoples’ everyday life. One may hypothesize that at the national level it is plausible that one’s own culture, including one’s own values, are even more strongly stressed on the waves of globalization. In other words, values of various countries might well grow apart and we have the contrary of convergence of values at hand: divergence of values. In this volume we will focus on value shifts in either the converging or diverging direction at the level of nations. In spite of the some-
4 Appadurai, a well-known scholar on the cultural dimensions of globalization, calls this phenomenon ‘culturalism’: the conscious mobilization of cultural differences in the service of larger national or transnational politics, more precisely involving struggles for stronger recognition from existing nation-states or various transnational bodies. Many of these culturalist movements are examples of the form cultural differences take in the era of globalization, says Appadurai (1998: 15–5–16). See also Waters (1995), Tomlison (1999), and Crane et al. (2002) for more on cultural globalization and concrete (policy) responses. 5 Featherstone argues that these global cultural flows are not merely the product of flows between nation-states but are embedded in global scale processes. He, therefore, believes it is misleading to conceive of global culture as necessarily involving the weakening of national/local cultures or entailing global homogenization. Rather, various diverse, rich, and popular local cultures are resisting and feeding global culture. The variety of these responses to the globalization process is viewed as evidence that there is little prospect of a unified global culture. Still, Featherstone, like may theorists in the field of cultural globalization, relies heavily on persuasion and anecdotal evidence instead of sound empirical and comparative data.
values and the social sciences
7
times high-pitched discussions on globalization, many social scientists, including classic-status scholars, cling to the nation-state as a relevant entity in cross-cultural research.6 As we will see further below, some of them sampled cross-cultural data from across the world and established cultural diversity that still aligns strongly with national borders. Now a major task of cross-cultural value research, we believe, is to determine whether the general populations in European nationstates differ and continue to differ in fundamental value preferences or show a trend towards value convergence. What are values? Before reflecting upon these issues further, it is necessary to define what we mean by ‘values’ and to illustrate the way values are and have been conceptualized in the social sciences. Values are seen here, in a Kantian sense, as ‘beliefs about the desirable’ (Kluckhohn, 1951). Values provide the individual, the group, and the society with fairly stable, similar orientations across nonsimilar, often incommensurable situations and thus add to social integration. Values are likely seen by both psychologists and sociologists as deeply rooted motivations or orientations acquired in early stages of life which guide and/or explain attitudes, norms, and opinions which, in turn, control human action (Spates, 1983). In this tradition of value theory, regularities in behavior are explained by cognitive or affective schemata which as such are based on more encompassing values.7 Values shape and justify behavior. Parsons (1939) speaks of values as moral beliefs to which people appeal for the ultimate rationales of action and claims that values are the most important elements of social life.8 More into
6 Schwartz (2004: 57), e.g., explicitly argues that there are hardly any good theoretical or empricial arguments to believe that nation-states are no longer relevant levels of analysis. Schwartz (2004: 56–57) attempted to address this issue empirically by comparing the cultural distance between samples within countries to the cultural distance between samples from different countries. He observed that the cultural distance between samples from different countries was almost always (183 out of 187 comparisons) greater than between samples from the same country. “The findings support the view that nations are meaningful cultural units” (Schwartz o.c. 57). 7 As Triandis (2004: 29) puts it—with reference to Kluckhohn (1954)—“Culture is to society what memory is to individuals”. 8 Notwithstanding this, in his seminal study on social structure Parsons (1939)
8
peter ester, peter mohler, and henk vinken
the direction of psychological reasoning, Rokeach (1968) claims that a value is seen to be “a disposition of a person just like an attitude, but more basic than an attitude, often underlying it”. Hofstede (1984) again gives another, but similar, definition of a value as “a broad tendency to prefer certain states of affairs over others”, and Schwartz (1992) more concretely defines values as “desirable states, objects, goals, or behaviors, transcending specific situations and applied as normative standards to judge and to choose among alternative modes of behavior”.9 An important merit of this last psychological definition is that, similar to Parsons and Luhmann, it distinguishes values from attitudes by pointing at their generalized nature: attitudes are people’s beliefs about specific objects or situations. In this sense attitudes take a lower place than values in a person’s hierarchy of beliefs. Again according to Schwartz (1999) values are trans-situational criteria or goals of the desirable, ordered by importance as guiding principles in life. Moreover, values (being ‘cultural ideals’) are always positive, whereas attitudes can be positive or negative.10 Today, most social scientists agree that a person’s values are inherent characteristics which can only be measured as latent constructs, i.e. they cannot be observed directly. However, they can be measured indirectly, i.e. in the way in which people evaluate states, activities or outcomes. Values, moreover, are seen as relatively stable or even durable: they do not change overnight. They thus differ from norms, which might cease to exist if not followed. There may be many instances of value negligence by a person or group. However, these instances might not harm the status of the value itself. In other words, values are counterfactual resistant or ‘enttäuschungsfest’ as Luhmann (1967) stated. Finally, individuals, groups or societies may differ in their hierarchical patterns of values (some values are more important than other values), which is the prime field of interest of comparative value research.
explicitly excluded values from his analysis, because he claimed that one can disentangle social structure and culture. This position was implicitly refuted by Niklas Luhmann in his discussion of modal generalizations (Luhmann, 1967). 9 Hofstede disagrees with Schwartz on values representing the ‘desirable’. Hofstede (1980, 2001) argues that values (should) reflect the ‘desired’. Seen as such values refer to more concrete situations (much like attitudes, see further) instead of abstract ideals that do have but limited impact on real behaviors. 10 However, as Mohler (1978) showed, empirically value expressions often come in the disguise of identifying what is not desirable in a society or in a group.
values and the social sciences
9
Most social scientists probably agree that one of the main cultural trends in the way values change in modern societies is summarized by the notion of ‘individualization’. Although this notion is somewhat of a catch-all concept (Schwartz, 2004), its core assumption is quite straightforward and transparent: values in Western societies have become detached from traditional institutions and authoritative forces (such as the church) and increasingly find their legitimation in personal choices and preferences (Ester et al., 1994). Not only the basis of value legitimation has changed (from institution to the individual), but also its substantial direction. Observers of the contemporary cultural juncture are quite divided on the societal merits of the process of individualization. One group, the cultural optimists, points at the fact that individualization has created room for expressing values such as personal growth, self-expression, creativity, equality, democracy, personal freedom, gender concerns, and an “ethic of commitment” (Yankelovich, 1981; Bellah et al., 1992; Inglehart, 1990, 1997). The second group, the cultural pessimists, stresses that individualization brought about a vast number of quite reprehensible developments such as consumerism, hedonism, privatism, and narcissism. Individualization in this perspective is regarded as a primarily egoistic set of attitudes and values leading to social dissolution, isolation, and the undermining of collective solidarity and citizenship in society (Lasch, 1979; Sennett, 1977). Individualization is said to generate a pluralization of values. People choose values à la carte which ultimately leads towards fragmentation of values systems. Values, according to this group, have become generalized, i.e. detached from concrete situations and degenerated into vague and noncommittal concepts (Zijderveld, 1988).11 A third—less morally outspoken—group, which we will call cultural economists, propose the argument of the interdependency between societal economic development and increased individual wealth on the one side and value orientations on the other. They see individualization as an inevitable consequence of the dramatic increase of individual’s resources since 1945 in Western countries, which is a direct consequence of the disappearance of famine and general poverty in these societies (Mohler, 1978, 1989). They regard contemporary individualistic values
11 Ester and Halman (1994; 1995) found little evidence for widespread ‘consumeristic’ attitudes.
10
peter ester, peter mohler, and henk vinken
as an offspring from European aristocratic values, such as social and political engagement, taking responsibility, being learned, or following own strategies. Individualization is also regarded as another means of modern societies to reduce complexity and is thus seen as an inevitable ingredient of modern, wealthy, and highly differentiated societies.12 Further on in this chapter we will come back to this issue of concern over value development in (post)modern societies as expressed by a number of leading social scientists. But first we have to go into some more detail regarding the conceptual understanding of values and value changes. The ‘Big Three’ Hofstede, Schwartz, and Inglehart on comparative values and value change The comparative study of value convergence and divergence is a dynamic and inspiring, but also a very heterogeneous terrain of study. Its various research lines and emerging research traditions are characterized by typical and typifying theories, concepts, methods, and methodologies, and even by its own gurus and followers. Vinken et al. (2004) distinguish between three dominant perspectives: the postmodernist, the particularist, and the dimensionalist perspective.13 For the ‘postmodernist’, cultural values are hybrid, fluid, and permanently changing; world views lack coherence and are fundamentally ambiguous. Values are multiplicit, complex, and dynamic (Edgar & Sedgwick, 2002). Their breadth of perspective is almost endless and virtually all embracing. Values in the postmodern analysis are free floating, and not restricted by time, space, tradition or structure. Individuals are solipsists, who continually play with values, and who permanently re-invent culture (Van Binsbergen, 1999). Values are not embedded in structured, well-defined, homogeneous, uniform, and dimensional cultural patterns. The very idea of patterned
12 Of course, this conclusion is much debated by cross-cultural researchers, including for instance Hofstede. A cultural factor is at play, as not all complex, advanced, and wealthy societies are equally individualized, e.g., societies in Asia are and have remained comparatively collectivist inspite of modernization. 13 Actually these distinctions are made with respect to the slightly broader domain of cultural studies.
values and the social sciences
11
values is seen as a denial of ambivalence, as deterministic, as a reification of cultural realities (Rosaldo, 1993). Apart from an up tempo rhythm of staccato adjectives, the main contribution of postmodernist accounts of the concept of values is its emphasis on the hybrid, ambivalent, and multi-dimensional nature of values that may prevent theoretical oversimplifications (Vinken et al., 2004). For conventional survey analysis of comparative empirical trends in values, the postmodern approach is, however, of limited relevance and applicability. The ‘particularist’ view of values, by contrast, supports the notion of values as structured patterns but focuses on specific value areas such as religious values, family values, political values or work values (e.g., Ester et al., 1994; Arts et al., 2003). The main interest is in (trends in) these value specifics and not so much in its overarching mega-structure(s). The particularist theoretical and empirical preoccupation with values is inspired by detecting observable developments in various value domains and by exploring the major determinants of these developments. The analysis itself, though, may very well be directed by more integrative theories such as modernization theory. Often, particularist value research is based on comparative empirical (trend) studies (e.g., Ester et al., 1997; Halman, 1991; Halman & Nevitte, 1996; Halman & Riis, 1999).14 With some criticism we note the tendency of our scientific communities to neglect the accumulated evidence, data, and progress in theory in particular value areas. The ‘dimensionalist’ analyst of contemporary culture is explicitly interested in the systemic whole of values that overrides specific values in specific value domains. According to Vinken et al. (2004: 8): “The dimensionalist study aims at finding the ultimate, most frugal, and yet most meaningful basic set of axes with which to explain the broad range of attitudes, beliefs, life styles and the diversity of practices among large populations and/or organizations across societies. The very focus is on empirically validating the existence of a unifying, universal [. . .] pattern, that regardless of social differentiation, displays homogeneity, is broadly shared, and has the power to shape people’s identities, attitudes, and all other aspects of their culture”. In this section we will discuss the dimensionalist approach to studying values of three eminent value researchers: Geert Hofstede, Shalom Schwartz, and Ronald Inglehart, all of whom will play a role
14
Many studies using the EVS-data sets are based on particularist approaches.
12
peter ester, peter mohler, and henk vinken
throughout this book, and a particularly prominent one in the experimental part of this volume. The three authors all have established value dimension models that each are perceived as all-encompassing by their respective inventors and that compete in explaining values, concrete attitudes and behaviors in a wide range of separate domains. We will present the key developments in these domains in the subsequent chapters. The issue of competing explanatory dimensionalist models will be important in especially the experimental part of this volume that addresses the explanatory ‘value’ of either dimensionalist model. This necessitates a further introduction of the Big Three of cross-cultural value studies.15 For Hofstede, culture—and values as core indicators of culture— represents the uniqueness of a group, a collective or a nation, which he frames as ‘software of the mind’ (cf. Hofstede, 1980, 2001). He is internationally best known for his pioneering work in comparing national cultures on a basic set of five fundamental value dimensions. These dimensions include: power distance (the extent to which unequal distribution of power is expected and accepted), uncertainty avoidance (the extent to which people feel threatened by uncertain or unknown situations), individualism (loose ties between members of a culture or society, or vice versa: strong in-group cohesion, protection, and loyalty), masculinity (distinct gender roles, stressing males’ assertiveness and material success), and long-term orientation (favoring future-oriented virtues). Although the original data were collected (well) over thirty years ago—among over 100,000 IBM employees16— Hofstede claims that his value dimensions are relatively stable over time and thus are still valid value constructs today. International developments such as globalization, so his argument goes, will likely affect national cultures but leave the inter-cultural distances untouched.17 Schwartz (1992, 1994, 2004) distinguishes seven prototypical cultural value orientations on which societies and cultures can be positioned.18 These orientations reflect the way societies and cultures 15 We will also compare with fundamental personality traits (the ‘Big Five’) in the experimental part of this volume. The ‘Big Five’ model is another main competitor in explaining attitudes and behaviors. See the chapter by Vinken and Rammstedt in this volume for an extensive introduction of the Big Five. 16 Data were gathered in the 1960s and 1970s among samples of IBM-employees (total number over 115,000) in more than 50 countries. 17 See Vinken et al. (2004: 20–21) for a critique of this argument. 18 Note that Schwartz prefers to use the concept of value orientations instead of values (cf. Schwartz, 2004).
values and the social sciences
13
respond to basic issues in regulating human activity.19 For Schwartz, too, values are carriers of culture that can be positioned in a multidimensional space.20 These seven value dimensions include: harmony (unity with nature, world at peace), embeddedness (social order, obedience, respect for tradition), hierarchy (authority, humbleness), mastery (ambition, daringness), affective autonomy (pleasure, exciting life), intellectual autonomy (broadmindedness, curiosity), and egalitarianism (social justice, equality). These prototypic value dimensions are inferred from averaging the value priorities of individuals in matched samples from each society.21 As Schwartz indicates, his theory of cultural value orientations has been validated with data from about 200 samples from more than 65 nations, mostly teachers and pupils but recently also general population surveys (Schwartz, 2004). Inglehart is our third key author in the domain of comparative (cross-national) value research. He, too, published widely on trends in modern values and is best known for his forecast of a ‘silent revolution’ in Western democracies in which materialist values (law and order, fighting rising prices) gradually but increasingly lose support to postmaterialist values (such as self-development, democracy, tolerance; Inglehart, 1977, 1990, 1997). His thesis attracted much attention (particularly in Europe), both in scholarly and popular circles (Vinken et al., 2004). Two assumptions underlie his cultural forecast: a) commodities that are scarce meet the greatest subjective value (Maslow’s scarcity hypothesis) and b) basic values reflect the economic conditions that dominated one’s pre-adult formative years, the period in which one’s prime world outlooks are formed (socialization hypothesis). In combination these hypotheses predict that in current (relatively
19 These cultural value dimensions specify three bipolar dimensions of culture that mirror alternative solutions to three fundamental societal problems: embeddedness versus autonomy, hierarchy versus egalitarianism, and mastery versus harmony. 20 Says Schwartz: “I view culture as the rich complex of meanings, beliefs, practices, symbols, and values prevalent among people in society. The prevailing values emphases in a society may be the most central feature of culture [. . .]. These value emphases express shared conceptions of what is good and desirable in the culture, the cultural ideals” (Schwartz, 2004: 43). 21 The individual value constructs are believed to be recognized cross-culturally and cross-nationally. The constructs enclose: power (status, control, dominance), achievement (success and competence), hedonism (pleasure, sensuality), stimulation (excitement, challenge), self-direction (independence), universalism (understanding, protection, tolerance), benevolence (welfare preservation), tradition (commitment to traditional customs and ideas).
14
peter ester, peter mohler, and henk vinken
prosperous and secure) socio-economic conditions a higher proportion of younger generations than of older generations will prioritize non-material goals over material gains. These generations grew up in affluent times and consequently its material needs are generally well provided for, leaving ample room for developing postmaterialist values.22 Inglehart’s more recent publications take a somewhat wider perspective (Inglehart, 1997; Inglehart & Baker, 2000). Two dimensions are discerned by which cultures worldwide can be located: survival versus well-being values and traditional versus secular-rational (authority) values. Postmaterialist values are still the core of the first factor but augmented by beliefs about life satisfaction and general social trust.23 The development towards global modernization—now framed as a two-dimensional process—covers not only a shift from survival to well-being values but also a change from traditional to secular-rational orientations towards authority. Inglehart’s trend analysis is based on data (1981–2002) from the European Values Study (EVS) and the World Values Survey (WVS), surveys that were conducted in over 80 societies.24 A logical question that follows is: what are the theoretical and methodological similarities and dissimilarities between these three ‘doyens’ of comparative value research, what unites and what divides them? The first and most indicative similarity is that all three authors emphasize that values are the nucleus of any culture and the key concept in doing cross-cultural analyses. From a social science perspective, the three of them claim, values are the most revealing, essential, and enduring part of what constitutes a culture (and hence also a society). In studying cultures, their main focus is on the ‘etic’ (general) instead of the ‘emic’ (specific) nature of cultures (Vinken et al., 2004: 17). Next, they share a theoretical conception in which pre-adult socialization is crucial in understanding value patterns: values are learned dispositions—they are not ‘innate’. Explaining cultural change, furthermore, is not a major asset of both Hofstede’s and Schwartz’ value models: they primarily focus on stability and
22 “[. . .] Liberation from threats to material existence reduces the centrality of survival-focused values and gives higher priority to freedom of choice” (Inglehart & Oyserman, 2004: 86). 23 See Inglehart & Baker (2000: 24). 24 Inglehart is keen on mentioning that these 80 societies contain almost 85% of the world’s population (Inglehart & Oyserman, 2004: 76, note 1).
values and the social sciences
15
continuity. Inglehart’s dimensionalist work is more explicit on determinants and correlates of value change, which he attributes to generational replacement and a wide variety of broad societal developments (e.g., rising welfare, educational levels, and technological advancement), which typically are part of mainstream modernization theories. Methodologically, all three value researchers use large population samples and aggregate individual responses. It is only Inglehart who from the onset has relied on representative samples of the general public. Hofstede and Schwartz tested their hypotheses first with matched samples of specific groups (IBM employees, teachers, and students) and only later extended their database to general publics. Their value items are plotted in a multi-dimensional space, though Hofstede’s and Schwartz’ value measures—contrary to Inglehart’s— are not hierarchically ordered. In their labeling of dimensions, there is, not surprisingly, some overlap between the three value typologies as also recently indicated by the authors themselves (Hofstede, 2004; Inglehart & Oyserman, 2004; Schwartz, 2004).25 Schwartz points at the linkage of his autonomy/embeddedness dimension with Hofstede’s individualism/collectivism, as well as between egalitarianism/hierarchy and power distance. Calculating correlations between the Hofstede and Schwartz dimensions tend to support these conceptual similarities though differences remain rather substantial (Schwartz, 2004: 51–53).26 Schwartz also found some support for overlap between his autonomy/embeddedness dimension and Inglehart’s traditional/secular dimension but little empirical overlap with the egalitarianism/hierarchy dimension. Furthermore, survival/self-expression tends to correlate positively with autonomy/embeddedness and with the egalitarianism/hierarchy dimension.27 Likewise, Inglehart shows the linkages between the three clusters of individualism/collectivism (Hofstede), autonomy/embeddedness (Schwartz), and survival/self-
25 See for more and detailed comparisons the chapter by Vinken and Rammstedt. Here we note that there is also a tendency of reification in their writings: the labels for computed scales of factors out of a sudden turn into theoretical concepts. A dimension called autonomy is treated as ‘real’ autonomy. 26 Correlations regard the arraying of nations on the various Hofstede and Schwartz dimensions. 27 In a final analysis (see next note), Schwartz included the three value typologies and found that they “identify such similar cultural regions around the world” (Schwartz, 2004: 71).
peter ester, peter mohler, and henk vinken
16
expression values (Inglehart). He concludes that: “High levels of Individualism go with high levels of Autonomy and high levels of Self-expression values. Hofstede’s, Schwartz’ [. . .] and Inglehart’s measures all tap cross-cultural variation in the same basic aspect of human psychology—the drive toward broader human choice. Societies that rank high on self-expression tend to emphasize individual autonomy and the quality of life rather than economic and physical security” (Inglehart & Oyserman, 2004: 82).28 Concerned social scientists—Moral under- and overtones Labels for principal component factors or dimensions of a multidimensional scaling of data such as individualism, autonomy, tolerance, self-development, materialism or masculinity, are not—of course—only subjects of valid and reliable scale construction in professional cross-cultural empirical value research. They refer to theoretical concepts, which have been debated for a long time in the history of the social sciences. In fact, the concern of social scientists with the negative societal side effects of the assumed decline of (pro-social) values is as old as the social sciences itself (Dekker et al., 2003; Ester & Vinken, 2003).29 Some topics of concern were already briefly introduced above. Both Karl Marx, Max Weber, and Emile Durkheim—founding fathers of the social sciences—were clearly troubled by what they saw as the drawbacks of the processes of modernization and industrialization: the loss of solidarity, the weakening of Gemeinschaft orientations, the spread of feelings of anomie and impersonalization, and the overemphasis of instrumental rationality. Though rooted in quite different paradigmatic approaches, these pioneering nineteenth century social scientists all shared a basic notion of the erosion of community values as a consequence of the emerging modern industrial society. Yet, this notion has entered the proclaimed next era of post-
28
Schwartz attempted to multi-dimensionally scale the three value models (using aggregated data) and observed (at least) three combined value dimensions: desirable degree of independence of the person from in-groups versus embeddedness in these groups, equal versus hierarchical allocation of resources, roles, rights, and obligations and assertively using or changing the social and natural environment in the active pursuit of goals versus maintaining harmony (Schwartz, 2004: 72). 29 Several points made in this section have been adopted from Ester & Vinken (2003).
values and the social sciences
17
modern and postindustrial society. What is more, the disciplinary discomfort by social scientists with the weakening of fundamental values did not soften (Paxton, 1999). On the contrary: it looks as if the preoccupation with the assumed decline of values that shape “the good society” develops into a growing unifying theme in an otherwise quite dispersed disciplinary approach and status of the social sciences (Misztal, 1996). The number of studies by social scientists that question the value basis of contemporary culture and society is accelerating, often framing the process of value erosion in rather normative terms. Analysts, however, put somewhat different emphases in accounting for the crumbling of social values. Sandel (1996) points at the erosion of the moral fabric of the triad of family, community, and nation, and speculates that people are individually and collectively losing control of forces that govern their lives—leading to “fragmented and storyless selves”. Wolfe (1989) is particularly concerned with the prevailing value dominance of the market calculus in the moral and social realm: “the logic of buying and selling”. Bellah et al. (1992) frame the decline of social values in terms of a severe malfunctioning of basic social institutions and depleted resources to deal with them. Berger & Neuhaus (1996) also point at the weakening of social institutions—such as the family and the neighborhood—which traditionally mediated between the private life of the individual and the megastructures of public life. Fukuyama (1995) relates the loss of social capital to declining social trust, i.e. through the advancement of the economic model of the rational self-interested individual we have come to forget that human behavior is also deeply embedded in social life and directed by fundamental values and norms of sociability. The preoccupation with the decline of social values—or perhaps better: with the decline of behavior based on pro-social values—is especially vibrant in circles of civil society analysts and communitarian theorists.30 Widespread feelings of social mistrust, citizens turning away from prime institutions and political authorities, and engaging less in informal interactions are seen as indicators of the decline of the traditional civic ethic.31 Leading contributor to communitarian
30
E.g., Etzioni (1996, 2001). See Ester & Vinken (2003) for a fundamental critique of these and similar outdated indicators of civic decay. 31
18
peter ester, peter mohler, and henk vinken
thought Amitai Etzioni (2001) stresses that what present society needs is a strong moral voice, speaking for and from a set of shared core values, that guides community members to pro-social behavior. The general idea in the contemporary concern about what constitutes the good life, the good citizen, and the good society is that prime social institutions such as the family, the school, the neighborhood, the church, civic organizations—the ‘little schools of citizenship’ according to de Tocqueville—no longer fulfil essential psychological and symbolic functions in providing people with basic values and meaning structures.32 These traditionally mediating institutions between the individual and society have lost their significance and led to a fraying of the social fabric, de-unification of collective moral standards, and fragmentation of individual pursuits. On the societal side, the enlargement of scale, the professionalization and bureaucratization of the welfare state, as well as the ‘thickening’ of the state, have disintegrated the societal midfield, so many civil society analysts hold (e.g., Zijderveld, 1979). Social capital as embedded in primary social institutions has inflated and the postmodern citizen becomes detached, non-affiliated, and feels no personal loyalty to the wider community. Malcontent is said to be the prevailing public mood. “[N]umerous signs—unstable voting patterns, a return to religious orthodoxy, increases in antisocial behavior, opposition to scientific and technological advance, a withdrawal from public issues into private worlds, and the rise of irrationality—indicate, for reasons both sound and unsound, a feeling of discontent with progress” (Wolfe, 2000: 51). The privatization of values and norms, according to these social critics, no longer bound by common moral principles, has brought about a society where each individual is involved in private matters and is no longer committed to the public cause. The individualization of life courses and life styles has disintegrated former overarching systems of value coherence and meaning supply. The individual in postmodern society and postmodern culture is primarily pursuing private goals detached from broader collective community interests.33
32
See Nisbet (1971). There is a remarkable and for empirical social science alarming overlap between these pessimist perspectives, especially reflected in concerns on the robustness of civil society engagement (that is, engagement as measured with instruments dated from the early 1970s) and the radical views of hyper-individualization voiced by 33
values and the social sciences
19
As indicated, we witness a substantial number of studies in citizens’ civic virtues, political values, moral codes, volunteering, informal bonding, and civic engagement which accumulated in Robert Putnam’s both much praised and criticized book Bowling alone (2000). Putnam’s main message is that in the last quarter of the previous century, Americans have become increasingly disconnected from their families, friends, neighbors, communities, social institutions, and public life; in short: American communities are confronted with a serious and painful loss of social capital. Using a wide variety of, rather conventional, indicators, Putnam shows that Americans are less and less partaking in the political game, are less involved in religious and secular social activities, their civic participation is going down, they attend fewer informal social gatherings, and their social connections have substantially thinned. Americans trust their fellow Americans less, feel that honesty and morality is weakening, and increasingly mistrust institutions and political authorities. “Our growing socialcapital deficit threatens educational performance, safe neighborhoods, equitable tax collection, democratic responsiveness, everyday honesty, and even our health and happiness” (Putnam, 2000: 368).34 The post-war boomers, he argues, have never been able to renew the social capital that was so vigorously accumulated by pre-war generations.35 In fact, much of the civic decay is blamed on the younger generations. Says Putnam (2000: 255) “It is as though the post-war generations were exposed to some anti-civic X-ray that permanently and increasingly rendered them less likely to connect with the community”. Values underlying community disengagement and civic decay, it seems, are almost a defining feature of younger generations. This rather harsh judgment raises the question which role new generations, in fact, play in bringing about value change in modern society.36 This takes us to generation theory.
postmodernists in cultural studies (for whom reality itself seems an obsolete concept that no longer requires empirical analysis beyond armchair theorizing). 34 Putnam concludes that “[W]e desperately need an era of civic inventiveness to create a renewed set of institutions and channels for a reinvigorated civic life that will fit the way we have come to live” (Putnam, 2000: 401). 35 For recent studies on social capital see e.g., Norris (1999), Skocpol & Fiorina (1999), Van Deth et al. (1999), Putnam (2000), Dekker & Uslaner (2001), Edwards et al. (2001). 36 It also raises the question, whether the Putnams of this world have been part of the anti-civic X-ray machine themselves. Moreover, how these authors find their
20
peter ester, peter mohler, and henk vinken Generational replacement and value change
A classic approach to the explanation of cultural change is generation theory. In its most simple form, generation theory stipulates that changes in cultural orientations and behavior are a consequence of generational replacement through which new generations emerge who adhere to values, norms, ideas, and lifestyles that differ from older generations. Mannheim (1928–1929) was among the first sociologists who developed a more systematic theory as part of his broader ‘sociology of knowledge’ of the nature and effects of generational replacement in terms of cultural change. His theory inspired a firm empirical research program and tradition with respect to generational replacement and its relationships with cultural change (cf. Diepstraten et al., 1998).37 According to Mannheim, the mere biological fact of succession of birth-cohorts is the basis of both similarities and dissimilarities between generations but social events and developments during the formative years give momentum to the shaping of generations with distinctive characteristics. Individuals who are born at about the same time within the same socio-cultural space are called a generation location. But being part of the same generation location does not necessarily imply perceived feelings of alliance and solidarity. In its present definition a generation location is at best a generation in spe. Subjective generational identification or generation as an actuality will only occur when individuals socialized within a specific socio-cultural space are being exposed to the same ‘Schicksale’, i.e. when they experience the same major social, cultural, economic, and political events during their formative period such as wars, national crises, as well as periods of economic and cultural recession or flourishing. In short, they share a common destiny. Exposure to major societal events during one’s formative years is of crucial importance as value orientations, political beliefs, and social and cultural attitudes are being formed in this period with lasting effects over the life-course. Generation units are located within generations as an actu-
baselines and benchmarks for what is good and what is bad in societies, or what is essential for social cohesion and what not, is, positively said, vague. 37 His theory can be considered a major step forward vis-à-vis primarily positivistic and biological (Comte, 1880) or romantic-historical (Dilthey, 1875) conceptualizations of generations and cultural innovation that were popular in the nineteenth century. See also Diepstraten et al. (1999a+b).
values and the social sciences
21
ality, as they are formal or informal networks or subcultures in which styles, value orientations, and social and political attitudes of a generation are reflected and embodied based on a distinctive and crystallized generational self-perception as is the case, for example, for youth subcultures. Particularly in times of intensive social dynamics and changes generation units are likely to emerge. Thus, central to Mannheim’s understanding of cultural change are the effects of the ‘Zeitgeist’ on value orientations of cohorts during their formative years.38 In this sense, Inglehart’s value change theory is an application of Mannheim’s generational perspective: the structural shift from materialist to postmaterialist values, or in his more recent formulation: from survival to well-being values and from traditional to secular values is primarily attributed to (gradual) cohort replacement and differential socialization. Older generations were socialized in periods characterized by famine, warfare, and economic insecurity and therefore hold materialist values such as physical security and safety. Younger generations (in the US and the Western part of the EU), however, were raised in relatively peaceful, famine-absent, prosperous times and therefore can give more priority to postmaterialist values such as having a say in politics, freedom of speech, quality of life, self-expression, and self-esteem. Value change, in short, is attributed to generational differences in formative dynamics, life chances, and cohort-peer socialization. Numerous other researchers have studied the relationship between generational renewal and cultural change (e.g., Ashford & Timms, 1992; De Moor, 1995; Dekker & Ester, 1993, 1995a+b; Ester et al., 1994; Halman, 1991; Van den Broek, 1996; Van Deth & Scarbrough, 1995). The basic idea underlying such studies is a classic one: old generations with distinct cultural value preferences are replaced by new generations, which bring in their own preferences. Again, whether there is only a shift in preferences or actually new values (or at least new meaning of values) is a question, which cannot be solved by argument but by empirical
38 Mannheim used the term ‘Zeitgeist’ with much care as he especially criticized romantic-historical writers of using this concept in the 19th century in an un-sociological way, referring to some mysterious, if not supra-natural force that, footloose of sociological reality, inspired entire generations. It is interesting to note that Hofstede (e.g., 2001) uses the same term as a label for period effects (without mentioning these effects as such), effects that have an impact on members of all generations at a specific moment in time.
22
peter ester, peter mohler, and henk vinken
evidence only (Mohler, 1978).39 Typically, these value studies are embedded in more general modernization studies on intergenerational differences in life chances, frequently inspired by a sociology of stratification perspective. A key notion of this perspective is that post-war modern society features a fundamental shift from life chances based on ascription to life chances based on achievement, formerly open only to higher classes, now open to almost every member of the respective societies (Mohler, 1989). This shift is first and foremost the result of an enormous increase in higher education that took place in all Western societies. This line of research has produced a similar abundance of studies, particularly in the domain of intergenerational mobility and educational attainment (early studies include e.g., Blau & Duncan, 1967; Erikson & Goldthorpe, 1992; Goldthorpe, 1980; Heath, 1981; Lipset & Bendix, 1959; Shavit & Blossfeld, 1993). Studying value change from the perspective of generational replacement, we believe, is both a major challenge and a much-needed assignment for comparative value research. Ideally, this should be done from a Mannheimian ‘generation units’ analysis, i.e. socially well-defined groups of individuals that function as core change agents for their generations (Vinken et al., 2004: 23). But often such an analysis needs to move beyond the conventional approach and methods of cross-national (or cross-cultural) values research.40 For the time being, it suffices that comparative empirical value research at least systematically studies the relationship between generational renewal (as an abstract mechanism) and the process of value change. A major goal of this book is to consistently address this issue of intergenerational value change and thus to contribute to the ongoing debate
39 There is a striking lack, though, of studies that address the crucial question how generations themselves subjectively evaluate intergenerational differences in life chances and dominant values and norms. This is striking because the very core of Mannheimian generation theory states that generations are united by a sense of subjective belonging, i.e. they identify with their generation, because they share a common history (see Diepstraten et al., 1998; 1999a+b for an extensive elaboration of this crucial issue). 40 Mainstream generational values and life-chances research typically focuses on objective indicators. As such these studies can be characterized as ‘generational accountancy’. We need to know more, however, about how generational identification is related to attributing possible intergenerational differences to unique youth experiences in one’s formative years (cf. Diepstraten et al., 1998, 1999a+b).
values and the social sciences
23
on how generations differ in their world views and in their basic underlying cultural, economic, political, and moral values. The main issues and research questions to be addressed In this introductory chapter we have outlined a number of issues related to globalization, value change, and generations that will be studied from a comparative empirical perspective in the following chapters of this book. These issues are related to three themes in particular: value convergence, the farewell to traditional values, and the vanguard role of younger generations. In summary: Does globalization of the political, economic, and cultural domain result in a process of convergence in basic value trends in Europe or will national value differences remain or even become stronger? Does the modernization of European societies result in systematic decline of support for traditional values (including civic virtues)? Or is Europe embracing new self-expressive values? Do we observe consistent intergenerational differences in value agreement in Europe, in the sense that younger generations are most likely to proclaim new, non-traditional values (and/or are the forerunners of declining support for traditional values)? In this book we want to explore these three themes for European societies, in comparison with American society. Though our basic interest is in assessing empirical trends in Europe, we believe that such an approach gains in interpretation power when results are being compared with a major non-European society, i.e. the United States. Such a comparison also enables to assess whether observations as expressed by e.g., Putnam are exemplary of American exceptionalism or points at broader social and cultural phenomena. The three themes will be analyzed for a wide variety of value domains, including religion, family and morality, work, politics, multiculturalism and tolerance, volunteering, as well as equality. The European Values Study as the main data source Given the fact that we want to investigate the three leading questions for quite a number of value domains, it is of vital importance that the same data are being analyzed. Moreover, the nature of the questions implies that we need longitudinal, comparative, and European
24
peter ester, peter mohler, and henk vinken
value data—which furthermore have to be compared with American data on value trends. This set of data requirements leads to the use of the European Values Study (EVS) data as the prime source of data analysis.41 In fact, the choice is quite simple and obvious: the EVS is the only data source that meets all the requirements simultaneously. Other comparative data sets, such as the International Social Survey Program (ISSP) or the Eurobarometers, are either subject-rotating surveys or include only a limited and changing number of values.42 Three waves of EVS are currently available (1981, 1990, and 1999/2000), spanning a period of two decades of value developments in Europe.43 In the first (1981) module 12 European countries participated (plus Canada and the United States), in the second (1990) module the questionnaire was fielded in 31 European countries (plus Canada and the United States), and 33 European countries (plus the United States) took part in the third (1999/2000) module.44 Formally, both Canada and the United States participate as members of the World Values Survey (WVS) project (headed by Ron Inglehart of the University of Michigan).45 WVS is the non-European addendum to EVS.46 We will—for the most part—restrict the analyses reported in this book to the European countries that participated in all three EVS waves (plus the United States). EVS measures quite a number of fundamental values, with respect to, e.g., religiosity, civic morality, moral permissiveness, sexuality, tolerance, trust, politics and political participation, instrumental and expressive work
41
With one exception: the chapter by Fisher and Heath on equality values. See for details: www.issp.org and http://europa.eu.int/comm/public_opinion/index_en.htm 43 See for a detailed overview of the three EVS modules, the EVS organization, the participating countries, the organization of the fieldwork, the questionnaires, the availability of the data, and the publications of the EVS project: www.europeanvalues.nl. Data are documented and can be obtained by the Central Archive for Empirical Social Research of the University of Cologne (http://www.gesis.org/en/ data_service/topics/50–0–CD-ROM/index.htm). Since the start of the EVS-project, its secretariat has been hosted by the department of Sociology of Tilburg University. 44 See for recent publications: Ester et al. (1994, 1997), De Moor (1995), Arts et al. (2003), Halman & Riis (1999). 45 See: www.worldvaluessurvey.org 46 See for the combined 1999/2002 data set: R. Inglehart, M. Basanez, J. DiezMedrano, L. Halman & R. Luijkx (Eds). Human beliefs and values. A cross cultural source book based on the 1999–2002 values survey. Mexico: SiglioXXIEditores. 42
values and the social sciences
25
values, educational and family values, parent/child relations, solidarity, and postmaterialism. The single value measurement instruments will be introduced in the chapters concerned. Finally, regarding the measurement of (comparative) intergenerational value differences, the authors of this book agreed to use a basic three-cohort typology: the war generation (born before 1940), the baby boom generation (born between 1940 and 1960), and the baby bust generation (born after 1960).47 A little experimentation on value measurement As shown above, the present intellectual debate on basic human values is characterized by a strong moral and political overtone framed within a cultural perspective on what defines Western civilization. If one looks at the vast literature on values, it becomes immediately clear that the value concept is rather vague. Hundreds of different definitions can be detected (Kmieciak, 1976; Halman, 1991).48 Do values make a difference at all or are they just generalized, somewhat abstract goals that people just adhere to without any marked behavioral impact? Most people will support values such as justice, honesty, solidarity, peace etc., but what is the wider relevance of such support? An increasing body of scientists (and laymen) is using one particular model of personality traits (the so-called Big Five of personality structure) to explain attitudes and behaviors. Do we need
47 Such a simple typology is quite suited for cross-national analysis (cf. Diepstraten et al., 1998). There are arguments to believe that in most countries these three birth cohorts each had quite similar formative experiences. The first underwent the severe experiences of war, destruction, and reconstruction, the second one experienced the political and cultural upheaval in the late 1960s (or, at least in hindsight successfully framed their experiences in this way), the third one concretely felt worldwide crises and reform in geopolitical (e.g., nuclear threat and fall of the Berlin Wall), environmental (degradation and the rise of counterbalance policies), and socio-economic terms (huge youth unemployment and new opportunities in the ICT-driven economy). See for more detail also Ester et al. (2002). 48 Moreover, though not studied in this book, the relationship between values, norms, attitudes, and behavior is blurred and theoretically and empirically unsatisfying. Do values mediate the relationships between norms, attitudes, and behavior or do they have an autonomous influence? Is this true for some values but not for others? Do we need values to explain behavior, which ultimately is the main mission of social scientists? Or might other basic drivers of attitudes and conduct do the trick?
26
peter ester, peter mohler, and henk vinken
values if we know personality traits, a straightforward question would go? But there is more. In this chapter we zoomed in on the value models of the Big Three: Hofstede, Schwartz, and Inglehart. We do not know the competitive power of these models in explaining beliefs, norms, attitudes, and behavior. Is one model better than another? Particularly on these issues we need a lot of further experimentation, both on theoretical and methodological grounds. Fortunately, we were able to do some experimentation in the course of our value research project: we included the three value models in one tailormade survey module (fielded in Germany). The survey results enable us to study the micro-correlations of the three value models—to the best of our knowledge—for the first time.49 The survey attempts to relate these classic value models to one another and to personality trait models as developed in psychology. In doing so we attempt to assess the relative explanatory power of value typologies as well as of personality typologies. Moreover: in doing so we trespass traditional disciplinary borders between the social sciences and psychology. We feel that combining these very different perspectives and typologies may strongly add to our understanding of current cultural changes. References Appadurai, A. 1998. Modernity at Large. Cultural Dimensions of Globalization. Minneapolis & London: University of Minnesota Press. Arts, W.A., J. Hagenaars & L. Halman (eds.) 2003. The Cultural Diversity of European Unity. Findings, Explanations and Reflections from the European Values Study. Leiden & Boston: Brill. Ashford, S. & N. Timms 1992. What Europe Thinks. A Study of Western European Values. Dartmouth: Aldershot. Barber, B.R. 1995. Jihad vs. McWorld. New York: Random House. Bellah, R.N., R. Madesen, W.M. Sullivan, A. Swidler & S.M. Tipton 1992. The Good Society. New York: Vintage Books. Berger, P.L. 1997. Four faces of global culture. National Internet, 49, 23–29. Berger, P.L. & R.J. Neuhaus 1996. To Empower People. Washington, D.C.: American Enterprise Institute. Berger, P.L. & S.P. Huntington (eds.) 2002. Many Globalizations. Cultural Diversity in the Contemporary World. Oxford: Oxford University Press
49
Attempts have been made, however, to e.g., relate personality traits to the Hofstede value dimensions (see Hofstede & McCrae, 2004).
values and the social sciences
27
Blau, P.M. & O.D. Duncan 1967. The American Occupational Structure. New York: Wiley. Comte, A. 1839 [1969]. Cinquante-unième leçon. Lois fundamentales de la dynamique sociale, ou théorie générale du progrès naturel de l’humanité. Cours de philosophie positive. Bruxelles: Culture et Civilisation. Crane, D., N. Kawashima & K. Kawasaki 2002. Global Culture. Media, Arts, Policy, and Globalization. London & New York: Routledge. Dekker, P. 2003. Social capital of individuals. In P. Selle & S. Prakash (eds.), Investigating Social Capital. London: Sage. Dekker, P. & P. Ester 1993. Social and Political Attitudes in Dutch Society. Rijswijk, Den Haag: SCP/VUGA. —— 1995a. Political attitudes in a generational perspective. The Netherlands, 1970–1992. Acta Politica, 30, 57–74. —— 1995b. De empirische waarde van generatietypologiën. Repliek. Acta Politica, 30, 355–359. Dekker, P., P. Ester & H. Vinken 2003. Civil society, social trust and democratic involvement. In W. Arts, J. Hagenaars & L. Halman (eds.), The Cultural Diversity of European Unity. Findings, Explanations and Reflections from the European Values Study. Leiden & Boston: Brill. Dekker, P. & E. Uslaner (eds.) 2001. Social Capital and Participation in Everyday Life. London: Routledge. De Moor, R.A. (ed.) 1995. Values in Western Societies. Tilburg: Tilburg University Press. Diepstraten, I., P. Ester & H. Vinken 1998. Mijn generatie. Zelfdbeelden, jeugdervaringen en lotgevallen van generaties in de twintigste eeuw. Tilburg: Syntax Publishers. —— 1999a. Over generaties gesproken. Alter- en egobeelden van generaties in Nederland. Amsterdams Sociologisch Tijdschrift, 26, 344–370. —— 1999b. Talkin’ bout my generation. The Netherlands’ Journal of Social Sciences, 35, 91–109. Dilthey, W. 1875. Über das Studium der Geschichte der Wissenschaften vom Menschen der Gesellschaft und dem Staat (Printed in Ges. Schr. Bd. V). Edgar, A. & P. Sedgwick 2002. Key Concepts in Cultural Theory. London & New York: Routledge. Edwards, B., M.W. Foley & M. Diani 2001 (eds.). Beyond Tocqueville. Hannover: Tufts University Press. Erikson, R. & J.H. Goldthorpe 1992. The Constant Flux: A Study of Class Mobility in Industrial Societies. Oxford: Clarendon Press. Ester, P. & L. Halman 1994. Empirical trends in religious and moral beliefs in Western Europe. International Journal of Sociology, 24, 81–110. —— 1995. Modernization and the nature of individualism. A cross-cultural and empirical perspective. Sociale Wetenschappen, 38, 28–53. Ester, P. & H. Vinken 2003. Debating civil society. On the fear for civic decline and hope for the Internet alternative. International Sociology, 18, 659–680. Ester, P., L. Halman & R. De Moor 1994. The Individualizing Society. Value Change in Europe and North America. Tilburg: Tilburg University Press. Ester, P., L. Halman & V. Rukavishnikov 1997. From Cold War to Cold Peace. A Comparative Study of Russian and Western Political Cultures. Leiden & Boston: Brill. Ester, P., H. Vinken & I. Diepstraten 2002. Reminiscences of an extreme century. Intergenerational differences in time heuristics: Dutch people’s collective memories of the 20th century. Time Society, 11, 39–66. Etzioni, A. 1996. The New Golden Rule. Community and Morality in a Democratic Society. New York: Basic Books. ——. 2001. The Monochrome Society. Princeton: Princeton University Press. Featherstone, M. (ed.) 1990. Global Culture. Nationalism, Globalization and Modernity. London: Sage.
28
peter ester, peter mohler, and henk vinken
Fukuyama, F. 1995. Trust. London: Penguin Books. Goldthorpe, J.H. 1980. Social Mobility and Class Structure in Modern Britain. Oxford: Clarendon Press. Halman, L. 1991. Waarden in de Westerse wereld. Tilburg: Tilburg University Press. Halman, L. & N. Nevitte 1996. Political Value Change in Western Democracies. Integration, Values, Identification, and Participation. Leiden & Boston: Brill. Halman, L. & O. Riis 1999. Religion in a Secularizing Society. The European’s Religion at the End of the 20th Century. Leiden & Boston: Brill. Hannerz, U. 1996. Transnational Connections. Culture, People, Places. London & New York: Routledge. Heath, A. 1981. Social Mobility. Glasgow: Fontana. Hofstede, G. 1980. Culture’s Consequences. Newbury Park: Sage Publications. ——. 1984. Culture’s Consequences: International Differences in Work-related Values. Beverly Hills: Sage. ——. 2001. Culture’s Consequences (second revised edition). Thousand Oaks: Sage Publications. ——. 2004. Epi-dialogue. In H. Vinken et al. (eds.), Comparing Cultures. Dimensions of Culture in a Comparative Perspective. Leiden & Boston: Brill. Hofstede, G. & R.R. McRae 2004. Personality and culture revisited: linking traits and dimensions of culture. Cross-Cultural Research, 38, 52–88. Hsiao, H.-H.M. 2002. Coexistence and synthesis. Cultural globalization and localization in contemporary Taiwan. In P.L. Berger & S.P. Huntington (eds.), Many Globalizations. Cultural Diversity in the Contemporary World. Oxford: Oxford University Press. Huntington, S.P. 1996. The Clash of Civilizations and the Remaking of World Order. New York: Simon and Schuster. Inglehart, R. 1977. The Silent Revolution. Changing Values and Political Styles Among Western Publics. Princeton: Princeton University Press. ——. 1990. Culture Shift in Advanced Industrial Society. Princeton: Princeton University Press. ——. 1997. Modernization and Postmodernization. Princeton: Princeton University Press. Inglehart, R. & W.E. Baker 2000. Modernization, cultural change and the persistence of traditional values. American Sociological Review, 65, 19–51. Inglehart, R. & D. Oyserman 2004. Individualism, autonomy, self-expression. The human development syndrome. In H. Vinken et al. (eds.), Comparing Cultures. Dimensions of Culture in a Comparative Perspective. Leiden & Boston: Brill. Kluckhohn, C.K. 1951. Values and value orientations in the theory of action. In T. Parsons & E. A. Shils (eds.), Toward a General Theory of Action. Cambridge, MA: Harvard University Press. Kmieciak, P. 1976. Wertstruktur und Wertewandel in der Bundesrepublik Deutschland. Göttingen: Otto Schwartz. Lasch, C. 1979. The Culture of Narcissism: American Life in an Age of Dimishing Expectations. New York: Warner. Lipset, S.M. & R. Bendix 1959. Social Mobility in Industrial Society. Berkeley: University of California Press. Luhmann, N. 1967. Soziologie als Theorie sozialer Systeme. Kölner Zeitschrift für Soziologie und Sozialpsychologie, 9, 615–644. Misztal, B.A. 1996. Trust in Modern Societies. Cambridge: Polity Press. Mohler, P.Ph. 1978. Abitur 1917–1971—Reflektionen des Verhältnisses zwischen Individuum und kollektiver Macht in Abituraufsätzen. Frankfurt/M.: Peter Lang. ——. 1989. Wertkonflikt oder Wertdiffusion? Ein Vergleich von Ergebnissen aus Bevölkerungsumfragen und einer Inhaltsanalyse von Leitartikeln der FAZ. Kölner Zeitschrift für Soziologie und Sozialpsychologie, 41, 95–122. Nisbet, R.A. 1971. The Social Bond. An Introduction to the Study of Society. New York: Knopf.
values and the social sciences
29
Norris, P. (ed.) 1999. Critical Citizens. Oxford: Oxford University Press. Parsons, T. 1939. The Structure of Social Action. New York: Free Press. Paxton, P. 1999. Is social capital declining in the United States? American Journal of Sociology, 105, 88–127. Putnam, R.D. 2000. Bowling Alone. New York: Simon & Schuster. Rokeach, M. 1968. Beliefs, Attitudes, and Values. San Francisco: Jossey-Bass. Rosaldo, R. 1993. Culture and Truth. The Remaking of Social Analysis. Boston: Beacon Press. Sandel, M.J. 1996. Democracy’s Discontent. Harvard: Harvard University Press. Schwartz, S.H. 1992. Universals in the content and structure of values: theoretical advances and empirical tests in 20 countries. Advances in Experimental Social Psychology, 25, 1–65. ——. 1994. Beyond individualism/collectivism. New cultural dimensions of values. In U. Kim et al. (eds.), Individualism and Collectivism. Thousand Oaks: Sage Publications. ——. 1999. Cultural value differences. Some implications for work. Applied Psychology. An International Review, 48, 23–47. ——. 2004. Mapping and interpreting cultural differences around the world. In H. Vinken et al. (eds.), Comparing Cultures. Dimensions of Culture in a Comparative Perspective. Leiden & Boston: Brill. Sennett, R. 1977. The Fall of Public Man. New York: Alfred A. Knopf. Shavit, Y. & H.-P. Blossfeld (eds.) 1993. Persistent Inequality: Changing Educational Stratification in Thirteen Countries. Boulder: Westview Press. Skocpol, T. & M.P. Fiorina (eds.) 1999. Civic Engagement in American Democracy. Washington, D.C.: Brookings Institution Press; New York: Russell Sage Foundation. Spates, J.L. 1983. The sociology of values. American Review of Sociology, 9, 27–49. Tomlison, J. 1999. Globalization and Culture. Cambridge: Polity. Triandis, H. 2004. Dimensions of culture beyond Hofstede. In H. Vinken et al. (eds.), Comparing Cultures. Dimensions of Culture in a Comparative Perspective. Leiden & Boston: Brill. Van Binsbergen, W. 1999. Culturen bestaan niet. Inaugural address. Rotterdam: Erasmus University. Van den Broek, A. 1996. Politics and Generations. Cohort Replacement and Generation Formation in Political Culture in the Netherlands. Tilburg: Tilburg University Press. Van Deth, J.W., M. Maraffi, K. Newton & P.F. Whiteley (eds.) 1999. Social Capital and European Democracy. London: Routledge. Van Deth, J.W. & E. Scarbrough (eds.) 1995. The Impact of Values (Beliefs in government, Volume 4). Oxford: Oxford University Press. Vinken, H., J. Soeters & P. Ester (eds.) 2004. Comparing Cultures. Dimensions of Culture in a Comparative Perspective. Leiden & Boston: Brill. Waters, M. 1995. Globalization. London: Routledge. Wolfe, A. 1989. Whose Keeper? Berkeley: University of California Press. Yankelovich, D. 1981. New Rules. New York: Random House. Zijderveld, A.C. 1979. Het ethos van de verzorgingsstaat. Sociale Wetenschappen, 22, 179–203. ——. 1988. De culturele factor. ’s-Gravenhage: Vuga.
CHAPTER TWO
A DECLINE OF RELIGIOUS VALUES? Loek Halman and Thorleif Pettersson Introduction Both religion in general and churches and church leaders in particular are often said to have lost much of their former prominent positions, especially in Western Europe. The empirical evidence is primarily found in declining levels of church attendance and shrinking rates of church membership. Whether there are corresponding declines in religious beliefs has not been so extensively researched. Among others, Grace Davie has noted the discrepancy between religious beliefs and practices. According to her, the European situation is best characterized as ‘believing without belonging’, indicating that the “marked fall-off in religious attendance (especially in the Protestant North) has not resulted, yet, in a parallel abdication of religious belief ” (Davie, 2002: 5). Thus, instead of speaking of secularized Europe, Grace Davie prefers to portray Europe as more un-churched than un-believing. Sociologists of religion usually interpret secularization as a consequence of modernization. “Put simply, the idea has been that the relation between religion and modernity is inverse—the more of the former, the less of the latter” (Berger, 2001: 443). However, this assumption has become increasingly disputed. One of the arguments concerns the unexpected religious trends in the United States. Although this country is one of the most modernized countries in the world, Americans have remained highly religious, not only in their beliefs but also with regard to their church participation. The vitality of American religion is said to challenge the assumption of secularization as an inherent part of modernization (Finke & Stark, 1992; Greeley, 1989). Furthermore, some have also questioned the downward religious trends in Europe, claiming that there is “no demonstrable long-term decline in European religious participation” (Stark & Finke, 2000: 62, italics in the original). Such findings have led to the conclusion that the secularization doctrine ought to be buried beneath
32
loek halman and thorleif pettersson
tons of contrary facts and that it is time to whisper a final ‘Requiescat in pace’ to the secularization paradigm at its burial on the graveyard of failed theories (Stark & Finke, 2000: 33, 79). Proponents of the opposite view have in turn argued that such a dismissal of the secularization thesis is based on a theory of human behavior that is incompatible with the historical record. They have also suggested that the rejection of the secularization doctrine must be once and for all crushed by a “stake through the vampire’s chest” (Bruce, 1999: 2). Secularization theory has undoubtedly become a hot and highly debated issue, and the last word on this has obviously not yet been said! In this chapter, we will analyze religious change in Western Europe and the United States, both with regard to religious beliefs and church involvement. Our analyses will be confined to the last two decades of the 20th century. This is, of course, a very short period in time. Our findings can therefore not be unequivocally related to secularization theories of long-term religious decline. The reason is simply that the assumption of a long-term religious decline need not be incompatible with short-term religious revivals and upheavals. It should, however, also be noted that the previous two to three decades were very interesting from a secularization theory point of view. The developments into the late-modern knowledge and information society have accelerated during this period, and the ever-increasing globalization is said to have had a far-reaching impact, especially on more traditional cultural patterns. According to key assumptions in secularization theory, these developments are likely to bring about religious decline, both in religious practices and beliefs. Based on previous research we expect to find different trajectories of such changes. A comparison between Europe and the United States will yield an empirical test of the assumption that religious decline is primarily taking place in Europe, in contrast to “the surprising vitality and extreme pluralism of denominational forms of salvation religion in America” (Casanova, 2001: 426). Thus, we will investigate if Western Europe as compared to the United States is the (only) region “in which the old secularization theory continues to be empirically tenable” (Berger, 2001: 446). This is an intriguing issue, not the least since secular tendencies have been recently observed in the United States as well, suggesting that this country might become slightly closer to the levels of religion in Western Europe (Norris & Inglehart, 2004: 95). Furthermore, that Europe is likely to
a decline of religious values?
33
show lower levels of religious involvement than the United States does not necessarily imply that Europe should be regarded as homogeneous in this regard. It is common practice to split Europe in a North-South divide: the North being less religious and/or less active in church activities than the South. These differences are often attributed to the theological differences between Northern Protestantism and Southern Catholicism, and also to the differences in religious fractionalism and the relations between church and state (Martin, 1978a+b). Thus, there are many reasons to assume that the processes of religious change have developed differentially in different parts of Europe. There are often fluctuations in the religious decline, “depending upon the particular socio-cultural complex within which individual secularization occurs: the state ideology; the link between churches and nationalism or regional nationalism; the differences in denominations; the freedom of churches to organize religious socialization and their access and exposure to the mass media; the position of the intelligentsia etc.” (Dobbelaere, 2002: 153). Thus, even from a secularization point of view, one should not expect the religious decline to follow the same monotonous downward trend in each and every corner of the (Western) world. In the next section, we focus on the theoretical aspects of secularization. We will argue that since the social significance of religion has declined, religious beliefs and practices have declined as well, although these declines may well be of different formats in different parts of Europe. We will also argue that the impact of people’s religious values on their social orientations will diminish, although in different patterns in different social contexts. Thus, the trajectory of religious change in a given country is likely to depend on countryspecific internal circumstances as well. A key question asks whether or not the changes that have occurred drive European populations towards greater homogeneity or heterogeneity in religious matters. In other words, we will investigate whether religious values are diverging or converging in Europe. We will also argue that the religious changes are usually located within specific social strata, as for instance in different age groups. It is often suggested that cultural change and generational renewal are closely linked. Since there are obvious generational differences in formative dynamics, life chances, and socialization experiences, different age groups tend to develop distinct preferences, world views, and values. The gradual replacement of the older generations by the
34
loek halman and thorleif pettersson
younger with their distinct preferences and values will therefore eventuate in broader cultural value changes at the societal level. In the following section we will describe our empirical dataset and present our measurements and analytical strategies, and the results of our analyses. In a concluding section, we will summarize and evaluate our findings. Secularization Most secularization theorists agree that the differentiation between religious and secular institutions is the core dimension of secularization. Most also agree that this differentiation has increased over time, at least in the Western world, and that specialized roles and institutions have developed in order to handle specific tasks which were previously carried out by religious institutions (Wilson, 1996). As a consequence, religious institutions have lost much of their previous social functions, and they are no longer the main providers of education, health care, social welfare, etc. When secularization is understood as the differentiation between the religious and the secular institutions, this is just a specific instance of general functional differentiation. In this sense, secularization can be described as the repercussion of this general development on the religious subsystem (Dobbelaere, 1995: 1; 2002: 166). Paradoxically, due to this kind of secularization, the religious subsystem has become more exclusively occupied with ‘pure religion’; free from involvement in for example education and social care. It is an open question whether the differentiation between religion and the secular at the societal level is followed by a similar differentiation at the mental level of religious laypeople. The latter kind of ‘differentiation’ has been discussed as ‘secularization-in-mind’ and the compartmentalization between people’s religious orientations and their views on secular issues (Dobbelaere, 2002). Compartmentalization can be thought of as the psychological parallel to macro-level differentiation between religion and the secular, and as ‘differentiationin-mind’. Obviously, differentiation between religion and the secular at the macro level may drive individuals to isolate and compartmentalize their religious orientations from their views on secular spheres of life such as politics, gender relations, bio-ethical issues, etc. However, this need not be the case (Casanova, 1994; Dobbelaere,
a decline of religious values?
35
2002). Even if the religious institutions have become differentiated from secular society, religious laypeople can still want religion to have a social impact, and they can still base their views on secular issues on their religious convictions. Therefore, a key question concerns whether people’s preferences for a religious impact on various secular matters have declined or not. A related key question asks whether the actual relations between a religious commitment and views on various secular issues have weakened or not. If most secularization theorists agree on the differentiation thesis at the societal level, there is more disagreement about the consequences of the macro-level differentiation at the individual level. Some assume that people’s religious involvement has declined over time. Memberships rates in churches and denominations are said to have decreased and the adherence to religious belief systems to have diminished. The religious decline is said to be especially apparent in Western Europe (Casanova, 1994; Acquaviva, 1979). Whether a corresponding decline is also experienced in the United States is much more debated (see e.g., Stark & Finke (2000) and Norris & Inglehart (2004) for two partly contrasting views on this). Others assume that the differentiation between religion and the secular leads to the privatization of religion. As the religious and secular institutions become more differentiated, religion is assumed to have remained significant to personal and private matters (Turner, 1991: 9) and to be assigned to the home understood as “the sphere of love, intimacy, subjectivity, sentimentality, emotions, irrationality, morality, spirituality . . . and religion as well as morality became simply matters of individual, private taste” (Casanova, 1994: 22). However, despite these different understandings, secularization is generally understood as the process in which religion gradually loses its former encompassing role. In the words of Brian Wilson, secularization is a process “by which religious thinking, practices, and institutions lose their significance for the operation of the social system” (Wilson, 1998: 49). As a result of differentiation, people are increasingly bound to participate in different societal spheres, each governed by its own set of values. The growing independence of these spheres from religion is likely to challenge any overarching claims which religion may pose on both society and individuals (Dobbelaere, 2002: 27). Since the contemporary autonomous individual gradually has become the main agent in the shaping of values, attitudes, and beliefs, people are said to choose the beliefs they
36
loek halman and thorleif pettersson
prefer, and that in this regard the teachings of the churches have become of less importance. In these developments, there is “a shift away from the traditional churches, and growing numbers of people define and practice their religiosity in non-traditional, individualized and institutionally loose ways” (Berger, 2001: 447). Secularization theories link the religious decline to the modernization of society. The more modern a society becomes, the more people are assumed to turn away from religion. Both modernization in general and secularization in particular are assumed to be triggered by economic development, and according to the secularization paradigm, the modernization processes of economic growth, industrialization, urbanization, and rationalization cause individuals to become less religious, both in terms of religious practices and beliefs. Inglehart (1997; Norris & Inglehart, 2004) argues that the religious decline can be attributed to increasing levels of economic and physical security, produced by the establishment of the modern welfare state and increasing levels of material wealth. In economically less advanced societies, religion provides people with existential certainties and the assurance of salvation (Inglehart, 1997: 80). In economically more advanced societies, existential security is assured and the need for the religious reassurance has diminished. As a result, participation and confidence in religious institutions have declined but not necessarily religious beliefs. According to Inglehart, the latter is unlikely to happen, not only because “spiritual concerns will always be part of the human condition” but also and mainly because the “continuing decline in traditional religious beliefs . . . is linked with a growing concern for the meaning and purpose of life” (Inglehart, 1997: 80). In secure circumstances, that is, in economically well advanced and wealthy societies, people are not so much concerned with existential security but they are looking around for other things which can improve their quality of life, for instance the satisfaction of their spiritual needs. “The need for meaning becomes more salient at high levels of existential security so that, even in rich countries, although church attendance is declining, spiritual concerns more broadly are not disappearing” (Norris & Inglehart, 2004: 75). Whether such spiritual concerns in the broad sense should be theoretically associated with religion in the same way as church attendance and adherence to traditional church dogmas is of course quite another matter. The different levels of secularization are often regarded as a con-
a decline of religious values?
37
sequence of differences in religious traditions (Therborn, 1995: 274; cf., Davie, 2000), for instance the theological differences between Catholicism and Protestantism. The “seeds of individualism were manifest much earlier in Protestantism. In contrast to Catholics, Protestants are personally responsible before God in religious matters, and the church has a lesser role as mediator between the believer and God. The Catholic Church . . . imposes a more collective identity upon its faithful” ( Jagodzinski & Dobbelaere, 1995: 81). Yet another factor which may affect the levels of religious involvement relates to both past and present church-state relations (Davie, 2000; Wallis & Bruce, 1992). Such relations are said to constitute “the parameters within which religious life on a national level takes place” (Davie, 2000: 15). In some countries, the churches have for whatever historical and political reasons remained in a strong position, for instance because the church has “provided a central focus of cultural identity in opposition to an imperialistic neighbor which sought to impose an alien set of cultural values and identities” (Wallis & Bruce, 1992: 16), or because religion served “as carrier of nationalism” (Casanova, 2003: 25). In other countries, such a role of the churches is modest if existent at all. For instance, the national churches in the Scandinavian countries are not generally expected to fulfil such functions (Riis, 1994; Gustafsson & Pettersson, 2000). Recently, American sociologists of religion have developed a new theoretical perspective which assumes that the degree of secularization is affected by the structure of the ‘religious economy’ of a society. In order to explain the religious vitality in the multi-religious North-American society, they argue that competition between different ‘religious firms’ results in a more attractive religious supply which in turn yields substantially higher levels of religiosity and church affiliation (Finke & Stark, 1992; Stark & Iannaccone, 1994; Stark & Finke, 2000). In homogenous cultures on the other side, religious competition is assumed to be weaker, yielding a less attractive religious supply. However, this view is highly debated, particularly by European sociologists of religion. One reason for this is that in Europe, a free, open, and competitive religious market has not (yet) developed. In most European countries, the dominating churches have often been supported by the state, and if religious competition has been at all possible, it was state-regulated competition anyway (Bruce, 1999: 46–47). Further, the options for individual religious choices are often restricted by national, ethnic or social demarcations (Bruce, 1999:
38
loek halman and thorleif pettersson
78). Hence, the religious situation in Europe differs from the United States where religious firms compete for customers and where the religious supply is not affected by state regulations. In Europe, religious pluralism is also assumed to have undermined the taken-for granted nature of religion and thereby to have fostered individual secularization (Dobbelaere, 2002: 194). Thus, in the European setting, religious pluralism has been expected to cause lower levels of religious participation. Indeed, analyses of the recent data from the European Values Study have shown aggregated measures of religious pluralism to correlate negatively with national averages for religiosity and church attendance (Halman & Draulans, 2004: 313). However, other studies have found evident empirical support for the vitalizing effect of religious pluralism in the European context (Hamberg & Pettersson, 1994; 2002; Pettersson & Hamberg, 1997a+b; Pettersson, 2000). These studies have argued that the positive effects of religious pluralism are principally to be expected at the local level. The reason is simple. Individual choices to participate or not in a given religious activity are primarily affected by the locally available options for religious participation. And consequently, when church attendance was analyzed at the local level, it was positively related to pluralism among the locally available churches and denominations. Furthermore, religious competition can also take place between different local congregations within one and the same church or denomination. It might therefore be misleading to assume that, for instance, the Catholic Church in Spain or the Protestant Church of Sweden is homogenous with regard to the local supplies of religious services. Such arguments suggest that it may be unwise to decide a priori that religious pluralism will not have any positive effects in the European context. That Europeans tend to be less inclined to go to church or to rely on church leaders than Americans does not necessarily imply that their religious beliefs have decreased. Yves Lambert has noted that in countries where large parts of the population are un-churched, the majority consider themselves to be ‘religious’ (Lambert, 2004; cf., Davie, 2002). According to Lambert, this indicates that Europe is at a turning point of a religious revival, even among young and un-churched people. In addition to a trend towards religious decline, Lambert has also observed a contrary trend towards an autonomous spirituality and a vague religiosity, ”which is detached from Christianity and is also characterized by an increased belief in the after-life”
a decline of religious values?
39
(Lambert, 2003: 72). Therefore, although the numbers of those who declare themselves as ‘unchurched’ have grown, this need not be interpreted as an inevitable sign of declining levels of religious beliefs (see also Lambert, 2003, 2004). Also today, those who declare themselves to be atheists or non-religious are comparatively few in European countries. And even if the adherence to traditional religious beliefs and dogmas has diminished, there has also been an increased quest for new spiritualities. Thus, religious beliefs may have become more diversified. The growing participation in the new spiritualities can of course also be seen as an effect of a broadened religious supply of new ways to manifest a religious interest. Empirical studies have consistently found religious beliefs and practices to be positively correlated with age. This pattern has lead many researchers to understand religious decline as the result of gradual population replacements, where the older generations are successively replaced by the younger. According to the Mannheimian generational perspective, successive birth cohorts or generations are differentiated from each other by processes of social change (the formative periods of the successive birth cohorts differ from each other). Such differences between birth cohorts tend to bring about social change (the younger birth cohorts with their specific cultural orientations gradually replace the older birth cohorts). Because the younger generations are raised and socialized in more secure and more secular circumstances, they tend to be less religious than the older generations. Since it is also assumed that the level of individual religious involvement tends to remain stable over the individual life cycle, the gradual disappearance of the older generations with their higher levels of religious involvement by younger less religious generations would result in a progressively more secular society. However, in contrast to this view, it has also been noted that when people grow older and mature they may also become more religious. The reason would simply be that they are increasingly faced with existential matters and the meaning of life (Stark & Bainbridge, 1987). Thus, as people get older, their religious involvement would increase. As a result, young people might return to the views of their parents as they grow older (Wuthnow, 1976 referred to by Dobbelaere, 2002: 152) and ageing may boost religious activity and religious beliefs (Sherkat & Ellison, 1999: 368). It should, however, be noted that this view need not contradict the assumption that religious decline is brought about by generational replacements. Whether this is the
40
loek halman and thorleif pettersson
case or not would depend on the size of the generational differences in religious involvement and the strength of the effects of ageing. In our empirical analyses, we will elaborate on these issues. Research questions Based on the above theoretical discussion, this chapter will focus on two basic research questions which relate to the two main dimensions of secularization theory: the declining levels of religious involvement and the diminishing impact of religion on secular society. The two research questions will be applied to individual-level data from 12 Western European countries and the United States, respectively. The first research question asks how the levels of religious beliefs and church-oriented religion have developed between 1981, 1990, and 1999/2000. The second question asks how the levels of compartmentalization between religious commitment and a set of secular orientations have developed during the same period. The theoretical discussion leads us to expect that the levels of religious beliefs and church-oriented religion have declined and that the compartmentalization between religious commitment and various secular orientations has increased. We also expect that the gradual replacement of the war generation by the boom and bust generations help explain these processes of change. Furthermore, we expect the processes of religious change to differ, not only between the United States and Western Europe, but also between different European countries. With regard to the transatlantic comparison, we expect the changes to be more pronounced on the European side, and also that the levels of religious involvement will be lower there. With regard to the intra-European differences, we expect to find differences between the Northern Protestant and the Southern Catholic parts of Europe. The levels of religious involvement are expected to be higher and the levels of religious compartmentalization to be lower in the South as compared to the North. And finally, since we see no compelling reasons to assume that the Nordic countries will catch up the higher levels of religiosity in Southern Europe, we do not assume that Europeans have become more homogeneous in religious matters.
a decline of religious values?
41
Data, measurements, and analytical strategy Data Since all chapters in this book use the same set of EVS/WVS data from 1981, 1990, and 1999/2000, we need not describe our data set more in detail. Our analyses are confined to the following 13 countries: Belgium, Denmark, France, Iceland, Ireland, Italy, Netherlands, Spain, Sweden, Great Britain, West Germany, Northern Ireland, and the United States of America. It should be noted that the data for Germany only cover the regions which belonged to the former West Germany. The reason is simply that the 1981 data is restricted to this part of today’s Germany. Measurements Our measure of religious beliefs is based on three indicators: The first concerns the belief in a personal God. This is tapped by a simple additive scale which ranges between 0 and 3. Those who say that they believe in God get one point, and those who have an understanding of God as a ‘personal God’ get two additional points.1 Those who believe in God and understand God as a kind of impersonal force get a score of 2, while those who believe in God but are uncertain about the nature of God get a score of 1. Those who don’t believe in God at all get a score of 0. The second indicator taps the subjective importance of God in the respondents’ lives. The responses were given on a 10-point scale, ranging from ‘not at all important’ to ‘very important’. The third indicator taps eschatological beliefs in the hereafter, and is based on three questions on whether one believes in a life after death, in heaven, and hell, respectively. Those who believe in each of these receive a score of 3, while those who don’t believe in any get a score of 0. For the measurement of church-oriented religion, we will likewise rely on three indicators. One concerns church attendance. This is tapped by a question on how often one attends church, apart from weddings, funerals, and other religious life-cycle ceremonies. The responses are given on a 7-point scale, ranging from ‘never’ to ‘several times
1 Due to missing data, we rely on an estimate for the belief in a personal God for the United States.
42
loek halman and thorleif pettersson
a week’. The second indicator taps views on church adequacy. This indicator summarizes the answers to three questions on whether one finds that the churches give adequate answers in relation to man’s moral problems, family problems, and spiritual needs. Those who have given a positive answer to each of these get a score of 3, while those who have given negative answers to each of them receive a score of 0. The third indicator concerns confidence in the church. The responses are given on a 4-point response scale, ranging from ‘no confidence at all’ to ‘a great deal of confidence’. In order to investigate whether the religious compartmentalization has increased or not, we will analyze the relations between our measures of religious involvement on the one hand and two secular orientations on the other. The first of the secular orientations concerns moral strictness. This is tapped by two items on public economic morality (how justified one finds it to cheat on taxes and to claim social benefits one is not entitled to, respectively, and two items on bioethical morality (how justified one finds euthanasia and suicide). Those who find each of these four moral behaviors to be absolutely unjustified get a score of 4, while those who do not totally oppose any of them receive a score of 0. Thus, the stricter the respondents are on these kinds of moral issues, the higher their scores for moral strictness. As a second measure of secular orientations, we will investigate attitudes towards the open and tolerant civic society. This is tapped by five questions which ask whether one would accept having an immigrant or a person of another race as neighbor, whether one can trust other people, whether one finds it important to teach children tolerance and respect for others, and, finally, whether one has been involved in at least two out of three kinds of social activism (signed a petition, participated in a boycott or a legal demonstration). Those who have given positive answers to each of these five questions get a score of 5, while those who have given negative answers to all five receive a score of 0. Finally, in order to analyze value change as a result of generational replacement, we will compare the differences between three different generations. These are the war generation (born before 1940), the baby boomers (born between 1940 and 1960), and finally the bust generation (born after 1960). These generations are described more in detail in other parts of this volume.
a decline of religious values?
43
Results Belief-oriented and church-oriented religion In order to investigate whether the six indicators for religious commitment which were described above can be used to tap the two separate dimensions of belief-oriented and church-oriented religion, respectively, we have performed a set of confirmatory factor analyses. The results of these are presented in Figure 2.1. The results show that the 6 indicators are related as expected to the two kinds of religious commitment. However, it should also be noted that the two latent dimensions are very strongly correlated (.90). The correlation is so strong that one may even ask if it were not better to analyze the six indicators as belonging to one and the same latent dimension of religious involvement. However, tests which are not presented in detail here demonstrate that such a one-dimensional model shows a poorer goodness of fit than the two-dimensional model. Since the distinction between the two dimensions is theoretically important, and since the possibly different developments for these two dimensions have often been discussed, we have therefore decided to treat them as two different, but highly correlated dimensions of religious involvement. For the sake of convenience we have relied on factor analyses to obtain scores for each of the two religious orientations (one factor analysis for the indicators for belief-oriented religion and one for the indicators for church-oriented religion). In these analyses, the data are weighted to give 1.000 respondents for each country and each year. The factor-scores for each dimension have then been transformed to get a mean of 100 and a standard deviation of 10.2 In this way country differences on the two dimensions can be easily detected. Changing levels of religious involvement Table 2.1 reports the country scores for the belief-oriented religious orientation. The countries are ranked from highest to lowest on their overall scores for the entire period. The United States are, quite as expected, ranked first, followed by Ireland, Northern Ireland, and Italy. The countries with the lowest scores are not very surprisingly
2
The formula was: (factor score × 10) + 100.
44
loek halman and thorleif pettersson
Agfi = .978; Rmsea = .056; n = 51,000 Test of a similar two-factor structure at three points in time for 13 fountries: Chi-Square/df = 7.41; AGFI = .961; Rmsea = .011, p. = 1.000
Figure 2.1: Results from a confirmatory factor analysis for a two factor model of religious commitment. Results for 13 countries from the EVS/WVS studies in 1981, 1990, and 1999/2000
the Netherlands, France, Denmark, and Sweden. The differences between the high- and low-scoring countries are substantial. When one, however, looks at the changes which have taken place over the period, these appear to be modest. This suggests that the belief-oriented religion had changed slowly during the previous decades. Since this period had seen several major socio-economic events and changes (see above on this), one may conclude that people’s religious beliefs are comparatively little affected by such factors, at least not in a short-term perspective. Table 2.1 also reports whether the changes that have occurred are statistically significant for the entire period. In order to stay away from too many details, we have avoided reporting separate analyses for the two sub-periods. A reason for this is that our theoretical perspective seems less relevant for the more narrow time periods. And yet, even if the entire time period of about two decades demon-
a decline of religious values?
45
Table 2.1: Changes in belief-oriented religion in 13 countries Wave
Change over entire period?
1981
1990
1999
United States Europe total Europe and the US
108.4 99.4 100.0
107.8 99.4 100.1
107.7 99.2 99.9
Decline* No change No change
European countries: Ireland Northern Ireland Italy Iceland Spain Great Britain Germany Belgium Netherlands France Denmark Sweden
107.5 106.8 101.1 100.4 101.9 99.3 97.1 98.8 97.5 95.0 93.9 93.7
107.6 107.6 103.8 101.6 100.9 98.9 97.2 96.6 96.7 95.1 93.6 93.1
106.0 105.5 105.2 101.8 98.2 97.7 98.3 97.0 96.9 95.1 94.4 94.5
Decline** Decline** Increase*** Increase*** Decline*** Decline*** Increase** Decline*** No change No change No change Increase*
*** p < .001; ** p < .01; * p < .05. Results according to t-tests.
strates rather diversified changes, these are by and large fairly minor. For the United States, there is a small but significant decline, whereas Western Europe as a whole shows no significant changes. However, when one looks at the separate European countries, about half (5 out of 12) show declining levels. These countries are Ireland, Northern Ireland, Spain, Great Britain, and Belgium. With the exception of Italy and Iceland, it should be noted that it are the countries with the highest levels of religious beliefs which have experienced declines. In contrast to the countries with religious declines, three countries show significant increases. These countries are Italy, Iceland, and Germany. It is difficult to see any common characteristics among these three countries which could explain their increases. Table 2.2 reports the results for the church-oriented religion. For this dimension, the patterns of change seem to be less diversified. Instead, one can observe a fairly general trend towards declining levels for most of the countries, the United States included. Denmark and Italy are the only exceptions, showing significant increases. As
46
loek halman and thorleif pettersson Table 2.2: Changes in church-oriented religion in 13 countries Wave
Change over entire period?
1981
1990
1999
United States Europe total Europe and the US
107.0 100.1 100.6
106.2 99.5 100.0
106.1 98.7 99.3
Decline* Decline*** Decline***
European countries: Ireland Northern Ireland Italy Spain Germany Iceland Belgium Great Britain France Netherlands Denmark Sweden
108.8 106.0 101.8 101.2 99.5 99.2 100.5 97.8 97.9 97.5 95.4 95.0
107.5 107.1 103.5 100.0 98.5 98.8 97.8 97.7 97.0 96.2 95.5 94.1
102.7 104.3 104.4 98.5 99.3 98.5 96.8 96.0 96.0 95.9 96.6 95.6
Decline*** Decline*** Increase*** Decline*** No change Decline* Decline*** Decline*** Decline*** Decline*** Increase*** No change
*** p < .001; ** p < .01; * p < .05. Results according to t-tests.
for the levels of church-oriented religion, the United States and Ireland show the highest scores, followed by Northern Ireland and Italy. Again, France, Netherlands, Denmark, and Sweden show the lowest levels. This ranking of the countries is similar to the ranking of the belief-oriented religion (Spearman’s rho .96; p < .001). This is also to be expected from the strong correlation between the two religious dimensions. Table 2.3 reports the results for the over-all religious involvement, tapping each of the six indicators described above. As expected from the two previous tables, Table 2.3 demonstrates a fairly general trend towards declining levels of religious commitment. Both the United States and Western Europe as a whole have experienced such losses. Interesting enough, the two countries with the lowest over-all levels, Denmark and Sweden, are together with Italy the only countries which demonstrate slightly increasing levels. The ranking of the 13 countries with regard to the levels of belieforiented and church-oriented religion shows no relationship to the
a decline of religious values?
47
Table 2.3: Changes in over-all religious involvement in 13 countries Wave
Change over entire period?
1981
1990
1999
United States Europe total Europe and the US
108.3 99.6 100.2
107.6 99.5 100.1
107.5 99.0 99.7
Decline* Decline*** Decline***
European countries: Ireland Northern Ireland Italy Spain Iceland Belgium Great Britain Germany Netherlands France Denmark Sweden
108.6 106.9 101.5 101.6 99.7 99.6 98.4 97.9 97.2 95.8 94.1 93.8
108.2 108.0 104.1 100.6 100.3 97.0 98.2 97.7 96.2 95.8 94.0 93.1
104.9 105.4 105.3 98.3 100.3 96.7 96.7 98.7 96.1 95.2 95.1 94.6
Decline*** Decline*** Increase*** Decline*** No change Decline*** Decline*** No change Decline* No change Increase** Increase**
*** p < .001; ** p < .01; * p < .05. Results according to t-tests.
degree of religious pluralism. The measures for the latter are taken from a recent investigation of different kinds of cultural fractionalization (Alesina, 2003). The non-correlation between aggregate measures of religious pluralism and religious commitment is expected from our previous discussion on this. It should, however, be noted that we have only analyzed a small number of fairly similar Western countries. Therefore, corresponding analyses based on a greater number of countries might show different results. Patterns of change in different generations Table 2.4 presents separate developments for the war generation, the boom generation, and the bust generation. It should be noted that the differences for one and the same of these generations at different points in time are difficult to interpret. The reason is that the generations cover age intervals of different sizes at different points in time. For example, the bust generation covers only four age-groups in 1981 (those in the age of 18, 19, 20, and 21, respectively), while
48
loek halman and thorleif pettersson
Table 2.4: Religious involvement, moral strictness, and civic-society involvement in 1981, 1990, and 1999/2000 in three generations Wave 1981
1990
1999
Diff 1981–1999?
War Boom Bust Total
102.8 97.6 97.3 100.2
103.3 99.0 96.8 100.1
102.1 99.4 97.6 99.7
Decline*** Increase*** No change Decline***
War Boom Bust Total
103.3 100.0 99.5 101.6
101.8 98.0 98.2 99.6
100.8 98.6 97.2 98.9
Decline*** Decline*** Decline*** Decline***
War Boom Bust Total
96.4 97.6 96.5 96.8
100.8 102.1 100.6 101.2
101.3 103.0 101.4 102.0
Increase*** Increase*** Increase*** Increase***
Religious involvement Generation
Moral strictness Generation
Civic society Generation
*** p < .001; ** p < .01; * p < .05.
in 1990, this generation covers 12 age-groups (those aged between 18 and 30). In 1999, this generation covers 21 age-groups (those aged between 18 and 39). However, disregarding such difficulties, Table 2.4 demonstrates some rather interesting patterns of cultural change, the two secular orientations included. With regard to the latter two, each generation appears to follow the same general trend (decline for the moral strictness, increase for the civic society involvement). In contrast, the patterns for religious involvement appear to be different for the three generations. Even if the war generation demonstrates a (minor) decline for the entire period, their pattern by and large also suggests some stability in their religious orientations. The bust generation also appears to be stable throughout the period, while the boom generation shows successive increases. It is difficult to find an easy explanation for these generational differences. One explanation can be that at least parts of the boom generation had their formative years during the war and/or the difficult imme-
a decline of religious values?
49
diate post-war times, and that existential security therefore would be a more pertinent issue to them. Another explanation might be that the boom generation has experienced more ‘critical’ life-cycle passages than the other generations during the period under study (got married, have had children, have entered a professional career, are approaching retirement, etc.). Anyhow, Table 2.4 demonstrates that religious change seems to follow partly different patterns as compared to the secular orientations, and that in this regard the generational divides appear to be more crucial. Separate analyses for the single countries might shed further light on this. However, in such analyses the various generations would be rather limited in size, and the results would therefore be more difficult to interpret. In the lack of further empirical data on this, we tentatively conclude that the religious orientation in contrast to the secular value dimensions is more dependent on experiences during primary socialization and that it therefore reflects the conditions that prevailed during each generation’s formative years. Once these religious values have become internalized, they would then tend to be more resistant to various social influences. This finding is in line with the Mannheimian perspective as this was described in chapter one. Changing levels of religious compartmentalization Our second research question concerns whether the levels for religious compartmentalization have changed or not. We have defined religious compartmentalization as the psychological or mental counterpart of religious differentiation at the societal level, and as ‘differentiation in mind’. We have already mentioned that we will investigate the levels of compartmentalization from the correlations between the religious involvement and the two secular orientations (moral strictness and civic society involvement). Table 2.5 reports the results of these analyses. For each country and each of the secular orientations, we have made one multiple regression analysis and one partial correlation analysis. The multiple regression analysis demonstrates whether the religious involvement (the combined religious beliefs and the church-oriented religion) adds anything to the amount of explained variance in addition to the socio-economic background variables of gender, age, education, income, and participation in the workforce. If so, the religious involvement would not be fully
loek halman and thorleif pettersson
50
Table 2.5: Changes in the levels of compartmentalization between religious involvement and two secular orientations among 13 countries Moral strictness
United States 12 European countries3 12 separate European countries: Belgium Denmark France Germany Iceland Ireland Italy Netherlands Northern Ireland Spain Sweden Great Britain
Civic society involvement
1981
1999
1981
1999
.06*** .26*** .06*** .27***
.10*** .33*** .03*** .20***
.01 –.07 .01*** –.05***
.00 –.06 .00 –.02*
.05*** .23*** .05*** .25*** .07*** .28*** .08*** .30*** .06*** .26*** .10*** .34*** .06*** .24*** .10*** .33*** .06*** .26*** .10*** .32*** .03*** .19*** .04*** .22***
.02*** .14*** .02*** .13*** .03*** .18*** .02*** .13*** .03*** .20*** .04*** .21*** .08*** .30*** .05*** .24*** .05*** .24*** .07*** .28*** .01*** .12*** .04*** .21***
.00 .02 .01 –.03 .00 –.02 .02 –.05 .02*** –.12*** .01 .01 .01 –.06 .01* –.09* .02** –.16*** .00 –.05 .00 –.06 .00 –.05
.00 .01 .00 –.04 .00 –.04 .00 –.01 .01** –.09** .00 –.01 .00 –.02 .00 –.04 .00 –.03 .00 –.01 .00 .02 .00 .05
Results from multiple regression analyses (first row: increase in R2 after inclusion of religious involvement as a second block after socio-economic variables) and partial correlation analyses (second row: partial correlation coefficient after controls for socio-economic variables); *** p < .001; ** p < .01; * p < .05.
3
Including dummy controls for 11 countries.
a decline of religious values?
51
differentiated from the secular orientations. The results from these analyses on the 1981 and the 1999 data would then indicate if the religious involvement has become more compartmentalized over time. Basically, the results from the partial correlation analyses report the same thing, that is whether the religious orientation has any relation to the secular orientations, disregarding the effects of the socioeconomic background variables. However, the partial correlations also demonstrate whether the religious orientation exerts a positive or a negative influence on the secular orientations. Table 2.5 provides rich information and may appear difficult to interpret. However, on a close inspection some important features immediately stand out. If one first compares the United States and the 12 European countries as a block, the level of compartmentalization seems to have increased in Europe; both with regard to moral strictness and attitudes towards the civic society (the coefficients were higher in 1981 than in 1999). In the United States, the level of compartmentalization has remained unchanged, or even become somewhat weaker in the case of moral strictness. In this way, the importance of the religious factor appears to have been weakened in Europe, while it has remained or even grown somewhat stronger in the United States. A second feature which is evident from the comparison between the United States and Western Europe is that the religious involvement on both sides of the Atlantic Ocean is far more related to the moral dimension than to the civic society orientation. This finding can be related to the assumption about the privatization of religion, and that religion has become less important for public and political matters. Even in the highly religious United States, there is no significant relation between the religious involvement and the civic society orientation. Furthermore, among the Western European countries, the relationships that were weak already in 1981 have become even weaker. This is the case both for Iceland, Northern Ireland, and the Netherlands, the only Western European countries where there was a significant relationship between the religious involvement and the civic society orientation in 1981. And when one examines the absolute values of the various coefficients and disregards whether these are significant or not, the over-all impression is the same for all of the Western European countries. There is hardly any substantial relation between people’s religious involvement and their orientations towards the civic society.
52
loek halman and thorleif pettersson
A similar tendency of increase in religious compartmentalization can be noted in the case of moral strictness. Even if the relationships in each of the 12 Western European countries are equally significant in both 1981 and 1999, there seems to be a more or less systematic decrease in almost all of the countries between 1981 and 1999. If the moral orientation appeared to become somewhat stronger related to the religious involvement in the United States, the developments seem to go in the opposite direction in Western Europe, Italy excluded. For instance, the average partial correlation decreased from .27 to .20 between 1981 and 1999, and no less than 11 out of the 12 countries show lower absolute values in 1999 as compared to 1981. If one then looks at the average increases in the squared multiple correlation coefficients after inclusion of the religious variable, this diminished from .07 to .03. Therefore, even if the religious involvement continued to be related to the moral orientation, the relations appear to have become weakened in Western Europe. The over-all impression is that the compartmentalization between religion and the two secular orientations have increased in Europe, whereas they have remained the same or even become somewhat reduced in the United States. In short, the impact of the religious involvement on people’s secular orientations seems to have weakened in Western Europe, while it has remained the same in the United States. Religious convergence or divergence Whether the European countries have diverged or converged in religious matters can be investigated at two levels. First, one can look at individual-level data for the West-European countries and investigate whether people have become more heterogeneous or more homogeneous in their religious involvement. In order to investigate this, we have simply compared the standard deviation for the religious involvement in 1981 and 1999, using the Levene test for equality of variances. A decrease in the standard deviations indicates that people have become more homogenous, while an increase indicates that they have become more heterogeneous. It is a standard in the sociology of religion to assume that people have become more heterogeneous in their religious preferences. Therefore, one would generally expect increasing standard deviations. However, since the standard deviations are also dependent on the means, we have also as a control calculated another measure by dividing the variances
a decline of religious values?
53
Table 2.6: Standard deviations for religious involvement in 1981 and 1999 1981
1999
Difference 1981–1999?
United States Western Europe Europe and the US
7.8 10.0 10.1
7.6 9.6 9.7
No change Decrease*** Decrease***
Western European countries: Belgium Denmark France Germany Iceland Ireland Italy Netherlands North Ireland Spain Sweden Great Britain
9.1 8.2 9.5 9.7 7.2 8.3 9.6 10.2 9.1 9.8 8.3 9.0
9.2 7.0 9.1 9.7 7.3 8.0 8.6 9.0 9.5 9.4 7.6 9.6
No change Decrease*** No change No change No change No change Decrease*** Decrease*** No change No change Decrease*** Increase***
by the respective means. These checks did not, however, alter the conclusions from our analyses. The results in Table 2.6 show that the standard deviation did not change between 1981 and 1999 in the United States. In Western Europe, there was however, a significant decrease, but the magnitude of this does not appear to be especially large. Looking at the separate Western European countries, one finds that a weak majority of 7 out of the 12 countries showed no changes. The level of heterogeneity would thus have remained unchanged. However, in three countries (Denmark, Italy, and the Netherlands), there were more substantial decreases, indicating that people had become more homogenous in their religious involvement. In this regard, it should be remembered that Denmark and Italy were among the exceptions, in that they also had experienced increasing levels of religious involvement. These increasing levels can then be assumed to have contributed to an increasing religious homogeneity. Looking at individual-level data, it can however, be concluded that in Western Europe as a whole, people have become slightly more heterogeneous in their religious involvement. At this level, Western Europe does not seem to have converged and become more homogeneous. However, this issue can also be investigated at the aggregated
54
loek halman and thorleif pettersson
macro level. In order to do so, we have simply analyzed the national means for religious involvement in 1981 and 1999. The results show that the national differences in religious involvement which were found in 1981 had become slightly reduced in 1999. Even if the changes were not great, this indicates a certain tendency towards macro-level religious convergence. In this regard, it might be recalled that some of the countries which scored lowest in religious involvement had witnessed increases between 1981 and 1999, while several of the other countries had seen decreases. Thus, these tendencies have given a minor trend towards convergence in the national levels of religious involvement. Conclusions The current debate on secularization is dominated by two opposite views. On the one hand, some insist that the secularization doctrine is mistaken and should be abandoned. That the United States as one of the most modern societies in the world continues to show very high levels of religious involvement is seen as a fundamental argument against the secularization theory. Secularization is therefore often seen as a European phenomenon, and Europe is seen as the only exceptional case where the secularization doctrine holds. European sociologists of religion on their turn are confronted with declining levels of religious involvement, particularly in the case of church-oriented religion. They have therefore insisted that the declining levels are linked with modernization and societal development. In the discussion of these changes, many have referred to the distinction between church-oriented and belief-oriented religion. The European religious declines are often said to concern the former, but not the latter. In contrast to this view, we have demonstrated that these two sides of religious involvement are highly correlated and that the distinction between them is difficult to uphold in empirical analyses. From an empirical point of view, it can therefore be questioned whether the theoretical distinction makes much sense. However, there are also some counter-arguments in this. The religious changes which we have found were most visible in the case of church-oriented religion. We also found that the two dimensions of religious involvement are somewhat differently related to social background variables. Nevertheless, one important finding from our
a decline of religious values?
55
analyses is that church-oriented and belief-oriented religious involvement tend to ‘go together’ and that they are difficult to separate in empirical research. This finding is in line with research which has demonstrated that there is a kind of over-arching normative-religious value dimension, which includes diverse indicators and very closely related sub-dimensions (Hagenaars et al., 2003). With regard to religious change, we have investigated one leveloriented and one structure-oriented research question. The former concerns whether the levels of religious involvement have changed over the previous decades. Contrary to what is often said, we found that the levels have declined in the United States, not dramatically, but anyhow. This finding undoubtedly casts new light on the often repeated conclusion that the contemporary United States has not witnessed any religious declines. According to our analyses, both the United States and the Western European countries have experienced diminishing levels of religious involvement, particularly with regard to the church-oriented side of religion. Therefore, the (minor) declines that have occurred in the United States have not brought the United States closer to the European levels. In terms of the levels of people’s religious involvement, there is still a marked transatlantic divide, although it must also be emphasized that some of the European countries do not lag far behind the American levels. The transatlantic divide is much more visible over the Northern part of the Atlantic sea. With regard to our structure-oriented analyses, we have investigated whether the religious involvement had become more disassociated from people’s views on secular issues. In this regard, we introduced the concept of religious compartmentalization as the psychological counterpart to religious differentiation at the societal level. Our analyses showed that religious orientations were more important for people’s moral orientations than for their relations to the civic society. We argued that this reflects that religion is becoming increasingly privatized. Furthermore, our analyses have also demonstrated that, especially in Western Europe, the religious compartmentalization had increased, both with regard to moral orientations and views on the civic society. In Western Europe, the trend has been towards a religious involvement with fewer consequences for people’s secular attitudes. In short, religion appears to have become increasingly impotent as a social force in Western Europe. In this regard, the United States seems to be an exception. There, the level
56
loek halman and thorleif pettersson
of compartmentalization between religious and moral orientations appeared to be more stable. In the United States, religion continues to drive towards moral strictness, whereas in Western Europe this drive seems to have lost momentum. It is tempting to relate the American pattern to the rise of the ‘moral majority’. We also found that the religious involvement was particularly weakly related to people’s civic orientations. This finding is especially interesting, since the indicators which we have used to measure civic orientations (tolerance, horizontal social trust, and social activism) are regularly used to tap social capital, something which is usually regarded as an important cultural resource for social, political, and economical development. Not the least since the publication of Robert Putnam’s seminal study on the importance of the civic tradition, social capital and the civic society are seen as important factors contributing to the ‘good community’. Our results indicate that in these regards, religion is not ‘on our side’. Even if this tendency has been reported by previous investigations (see e.g., Halman & Pettersson, 2003), our findings are certainly of interest for those who are concerned with religion as an agent for social change. By and large, our results indicate that religion is not an especially fruitful ground for positive social development. During the previous decades, our results suggest that the ground has become even more sterile. In an attempt to explain the religious decline, we argued that because religious orientations are primarily internalized by primary socialization processes and that they therefore tend to be resistant to change in later stages in life, religious decline at the societal level would occur by the gradual disappearance of the older and more religious generations. In contrast to this assumption, we found that the ‘baby boom’ generation increased their religious involvement during the previous decades. In this regard, our analyses seem to confirm Lambert’s thesis of a religious revival and a turn towards new spiritualities among young people. As an explanation for such a development, we argued that younger people appear to be in search of resources which can improve their quality of life, for instance the satisfaction of their spiritual concerns. When further increases in material wealth begin to give diminishing returns, the religious ground may appear to be a more fruitful ground for other rewards. Speculatively, the developments for the boom generation can also be related to liberating effects of the changing social climate and the
a decline of religious values?
57
youth revolution during the 1960s. As an effect of this, the former monolithic religious structures seemed to open up in favor of a more pluralistic religious sector. In accordance with the supply-side theory of religious involvement, this might then have contributed to the growing religious interest of the baby boomers. In conclusion, our findings have demonstrated that religion continues to be a fascinating area for study. The debates on religion are often heated, and the assumptions often simplified and exaggerated. We have demonstrated that religious processes are more differentiated than is often assumed, and that (parts of ) the Atlantic religious divide may not be as dramatic as is often believed. We have also demonstrated that Western Europe should not be seen as a monolithic block in religious matters, but that this region offers a rich variety of religious profiles, both with regard to the processes of change and the actual levels of religious involvement. To consider Europe as one homogeneous and integrated cultural zone with regard to secularization and processes of religious change would undoubtedly be mistaken. Even if we have found some general trend towards a religious decline, Europe is far from united with regard to religion and secularization. Current theories of secularization appear too broad and too general to do justice to the particular developments which we have observed for different countries. Current views on European secularization are in need of nuances, and the sharp decline which so often is assumed to occur there is difficult to find. But at the same time, it has been even more difficult to find vivid examples of religion as a potent force for social development and as a rich source for moral values. Religious involvement seems to navigate cautiously between on the one side Scylla with her tempting invitation to become a strong and challenging social force and on the other side Charybdis with her firm warnings about a sharp religious decline. References Acquaviva, S.S. 1979. The Decline of the Sacred in Industrial Society. Oxford: Blackwell. Berger, P.L. 2001. Reflections on the sociology of religion today. Sociology of Religion, 62, 443–454. Alesina, A. et al. 2003. Fractionalization. Journal of Economic Growth, 8, 155–194. Bruce, S. 1999. Choice and Religion. A Critique of Rational Choice Theory. Oxford, New York, Toronto: Oxford University Press. Casanova, J. 1994. Public Religions in the Modern World. Chicago: University of Chicago Press.
58
loek halman and thorleif pettersson
——. 2001. Religion, the new millennium, and globalization. Sociology of Religion, 62, 415–441. ——. 2003. Beyond European and American exceptionalism: towards a global perspective. In G. Davie, P. Heelas & L. Woodhead (eds.), Predicting Religion. Christian, Secular and Alternative Futures. Aldershot: Ashgate. Davie, G. 2000. Religion in Modern Europe. Oxford: Oxford University Press. ——. 2002. Europe: The Exceptional Case. London: Darton, Longman and Todd Ltd. Dobbelaere, K. 1995. Religion in Europe and North America. In R. De Moor (ed.), Values in Western Societies. Tilburg: Tilburg University Press. ——. 2002. Secularization: An Analysis at Three Levels. Bern, Brussels: Publishing Group Peter Lang. Finke, R. & R. Stark 1992. The Churching of America, 1776–1990: Winners and Losers in Our Religious Economy. New Brunswick: Rutgers University Press. Greeley, A. 1989. Religious Change in America. Cambridge MA: Harvard University Press. Gustafsson, G. & T. Pettersson 2000. Folkkyrkor och religiös pluralism—den nordiska religiösa modellen. Stockholm: Verbum. Hagenaars, J., L. Halman & G. Moors 2003. Exploring Europe’s basic value map. In W. Arts, J. Hagenaars & L. Halman (eds.), The Cultural Diversity of European Unity. Leiden & Boston: Brill. Halman, L. & V. Draulans 2004. Religious beliefs and practices in contemporary Europe. In W. Arts & L. Halman (eds.), European Values at the Turn of the Millennium. Leiden & Boston: Brill. Halman, L. & T. Pettersson 2003. Religion and social capital revisited. In L. Halman & O. Riis (eds.), Religion in secularizing society. Leiden & Boston: Brill. Hamberg, E. & T. Pettersson 1994. The religious market: denominational competition and religious participation in contemporary Sweden. Journal for the Scientific Study of Religion, 33, 205–216. Hamberg, E. & T. Pettersson 2002. Religious markets: supply, demand, and rational choices. In T. Jelen (ed.), Sacred Markets, Sacred Canopies: Essays on Religious Markets and Religious Pluralism. Lanham, MD: Rowman-Littlefield. Inglehart, R. 1997. Modernization and Postmodernization. Princeton: Princeton University Press. Jagodzinski, W. & K. Dobbelaere 1995. Secularization and church religiosity. In J.W. Van Deth & E. Scarbrough (eds.), The Impact of Values. Oxford: Oxford University Press. Lambert, Y. 2003. New Christianity, indifference and diffused spirituality. In H. McLeod & W. Ustorf (eds.), The Decline of Christendom in Western Europe, 1750– 2000. Cambridge (UK): Cambridge University Press. ——. 2004. A turning point in religious evolutions in Europe. Journal of Contemporary Religion, 19, 25–45. Martin, D. 1978a. A General Theory of Secularization. Oxford: Basil Blackwell. ——. 1978b. The religious condition of Europe. In S. Giner & M. Scotford Archer (eds.), Contemporary Europe. Social Structures and Cultural Patterns. London: Routledge & P. Kegan. Norris, P. & R. Inglehart 2004. Sacred and Secular. Religion and Politics Worldwide. Cambridge: Cambridge University Press. Pettersson, T. 2000. Många eller få? In M. Skog (ed.), Det religiösa Sverige. Örebro: Libris. Pettersson, T. & E.M. Hamberg 1997a. Söndrade vi stå, enade vi falla—en studie av mångfald och medlemstal bland svenska frikyrkor 1974–1995. In M. Skog (ed.), En salig blandning. Svenska trossamfund i religionssociologisk och kyrkorättslig belysning. Uppsala: Svenska kyrkans forskningsråd, Tro & Tanke. —— 1997b. Denominational pluralism and church membership in contemporary
a decline of religious values?
59
Sweden. A longitudinal study of the period 1974–1995. Journal of Empirical Theology, 10, 61–78. Riis, O. 1994. Patterns of secularization in Scandinavia. In T. Pettersson & O. Riis (eds.), Scandinavian Values. Uppsala: Acta Universitatis Upsaliensis. Sherkat, D.E. & C.G. Ellison 1999. Recent developments and current controversies in the sociology of religion. Annual Review of Sociology, 25, 363–394 Stark, R. & W. Bainbridge 1988. A Theory of Religion. New York, Toronto: Lang. Stark, R. & R. Finke 2000. Acts of Faith: Explaining the Human Side of Religion. Berkeley: University of California Press. Stark, R. & L. Iannaccone 1994. A supply-side interpretation of the ‘secularization’ of Europe. Journal for the Scientific Study of Religion, 33, 230–252. Therborn, G. 1995. European Modernity and Beyond. London: Sage. Turner, B. 1991. Religion and Social Theory. London: Sage. Wallis, R. & S. Bruce 1992. Secularization: the orthodox model. In S. Bruce (ed.), Religion and Modernization. Sociologists and Historians Debate the Secularization Thesis, Oxford, New York, Toronto: Oxford University Press. Wilson, B. 1996. Religious toleration, pluralism and privatization. In P. Repstad (ed.), Religion and Modernity. Modes of Co-existence. Oslo: Scandinavian University Press. ——. 1998. The secularization thesis: critcisms and rebuttals. In R. Laermans, B. Wilson & J. Billiet (eds.), Secularization and Social Integration. Leuven: Leuven University Press.
CHAPTER THREE
INDIVIDUALIZATION OF FAMILY VALUES? Jacqueline Scott and Michael Braun Family diversity It has been claimed that the family as an institution has become individualized in contemporary European society (Beck & BeckGernsheim, 2004). What does this mean? Is it true? And does it matter? The claim is an important one. The argument is that there has been a qualitative shift in terms of the understanding of family. From the 1960s onwards, in the era that some social theorists have termed the ‘second modernity’, the very notion of a ‘normal’ family has become redundant. But what characterized a ‘normal’ family in previous eras? Is the notion that there was ever consensus about what constitutes a ‘normal family’ a myth? These are quite complex questions. However, we do not need to go into the more contentious areas of historical debates about the characteristics of family life in times past, to concede that the idea of a ‘normal family’ consisting of an adult heterosexual married couple and their own children, with gender-specific roles that accentuated male work and female caring responsibilities has enjoyed considerable prominence in modern European societies. The claim that the concept of a normal family has become redundant is not to say that heterosexual, parent-child families with traditional gender roles have vanished. Rather, it is claimed that this particular family type now co-exists with a diverse range of living arrangements. This diversification of lifestyles and values means that perceptions of what is ‘normal’ are becoming relativized. It is certainly the case that the latter part of the twentieth century has seen marked changes in family life, such as the normalization of cohabitation and the increased acceptability of divorce, nevertheless, it is unclear whether the family individualization thesis holds up to empirical scrutiny. Are family values becoming increasingly relativized? Are traditional absolutes being over-turned? Or do our theories need to be more
62
jacqueline scott and michael braun
refined so that they discriminate more carefully between family values that show remarkable continuity and those that appear to be in flux? Some earlier work looking at cross-national attitudes concerning gender relations and family choices suggested that the rhetoric concerning the demise of traditional family values was exaggerated (Scott, 1999). However, that conclusion was based on an analysis of the International Social Survey Program’s module on family and gender roles, which has a relatively limited time span. In this chapter, we use data from the European Values Study to examine comparable measures across nine nations from 1981 through to 1999/2000. These eight European countries (France, Germany, Great Britain, Ireland, Italy, the Netherlands, Spain, Sweden) are chosen to ensure a good national spread of welfare regime types, religious traditions, gender norms, and family policies. We also include in our analysis, for comparison purposes, the United States, which is often seen as the nation which led the advance in changes in individualism and questioning of traditional gender roles. We explore how values have changed over time and across nations. Two aspects of family values are examined: (1) perceptions about the importance of core family relationships concerning marriage and motherhood, and (2) beliefs about the rights and wrongs of specific family-related or sexual behaviors. Our analysis tests the assertion that family values are increasingly relativized or individualized. On the basis of the theoretical and empirical literature to date, we hypothesize that boundaries of acceptable family and sexual behaviors have shifted and diversified. However, we also expect to find considerable continuity and consensus concerning the importance of core family relationships. We focus particularly on how cross-national differences have changed over time and how far change is related to inter-generational differences in family values. Two major social trends are of relevance to the generational divide. First, the transformation of gender roles, associated with women’s increasing participation in the labor force means that traditional ideologies of marriage and motherhood may be increasingly rejected by more recent generations. Second, the secularization process, which has gathered pace for more recent generations, is likely to have eroded the basis of many of the moral absolutes associated with family issues and behaviors including divorce, adultery, homosexuality, and abortion. Some have argued that the demographic shifts across Europe asso-
individualization of family values?
63
ciated with the increased emphasis on individualization makes a crossnational convergence in family values likely. For example, Castells (1997: 221ff ) argues that, although there is no predetermined directionality in history, the diversification of families and the demise of traditional family values go hand-in-hand with the end of patriarchalism. We call this ‘convergence in individualism’. An alternative possibility, however, is that there is a relatively stable pattern of family values across Europe, with some countries embracing individualistic values, whereas in other countries traditional family values remain strong. We call this ‘divergence in traditionalism’. The two scenarios have different implications for the likely harmonization of European family policies. Theoretical framework and literature review The general pattern of change in household and family structures in Western European families has seen marriage rates fall, divorce and cohabitation on the increase, and a marked drop in overall fertility, despite the increased number of children born out of wedlock. Such trends are used as evidence that the importance of family life is declining, with dire consequences for societal integration, and specifically for children. For example the American sociologist David Popenoe suggests that it is individualism, in its modern meaning of ‘emphasis on self-fulfilment’, and ‘pursuit of individual rather than collective interests’ that places the values of familism and individualism in opposition. He stresses that “in the current era, individualism has had a remarkable run. Women especially have been able to achieve self-fulfilment apart from the bourgeois family to a degree never before thought possible” (Popenoe, 1988: 305). This theme of the changing role of women, both in terms of individual autonomy and in terms of female emancipation has been a core part of the conceptualization of the ‘second demographic revolution’. This refers to the changing demographic patterns since the 1960s, including high divorce, decreasing fertility, increased cohabitation, and delayed marriage (Lesthaeghe, 1995). These demographic changes are intertwined with changing ideologies concerning the importance of marriage and motherhood. There have also been important sociolegal changes in the family domain that reflect and encourage new moral stances regarding family related issues and behaviors.
64
jacqueline scott and michael braun
Marriage. Since the late 1960s and early 1970s marriage rates in most European countries have declined. Younger generations of Europeans have been marrying less and, among those who marry, the trend has been to do so at older ages and over a wider range of ages than was common among the preceding generations (Kiernan, 2000). In broad outline the decline in marriage rates began in Sweden and Denmark in the late 1960s, then spread through most of Western Europe in the early part of the 1970s, and became evident in Southern European countries (e.g., Spain and Italy) around the mid 1970s. The decline since the 1980s has continued but at a slower rate. Yet this decline in marriage is somewhat misleading because it is partly driven by the rise in cohabitation. There is little evidence to suggest that partnerships are less popular, although it is clear that the number of partners over a life time has increased with the decline in partnership stability. It is also the case that many cohabiting couples eventually marry, especially when children are involved. In Sweden and Denmark family law has come to be applied to married and cohabiting couples in the same way. Yet other countries have far more conservative family policies, particularly Ireland where divorce has only recently been allowed (Family Law Divorce Act, 1996). Motherhood. There has been widespread concern in Europe about falling fertility. More than half a century ago, Schumpeter suggested that “as soon as men and women . . . acquire the habit of weighing the individual advantages and disadvantages of any prospective course of action . . . they cannot fail to become aware of the heavy personal sacrifices that family ties and especially parenthood entail under modern conditions” (1988/1942: 501f ). Although the baby boom is long over, it may be a mistake to think that the transition to a relatively low value on childbearing has reached its final equilibrium for low-fertility societies. As Bumpass suggests, despite the great inertia of pro-natalist forces, sub-replacement fertility is the norm in Western industrial societies, and the processes underlying the longterm decline seem far from exhausted (Bumpass, 1990). The costs of having children in terms of lost opportunities for employment, earnings, and career advancement affect women more than men. Many countries have taken active steps to help promote the compatibility of family and work, particularly through the enhancement of childcare. Such changes are not only concerned with gender equality issues, but also motivated by economic concerns that
individualization of family values?
65
worsening dependency ratios, due to falling fertility and increased longevity, are unsustainable. Much of the new rhetoric of European policy is not so much concerned with family-work-balance but with the co-employability of men and women or, more specifically, fathers and mothers. Many countries now view maternal employment as crucial for addressing the impending welfare crisis, associated with an ageing population. Despite such Realpolitik, family friendly employment practices vary enormously from country to country, and even where child-care provision is relatively good and opportunities for part-time work abound, women with children tend to face a decrease in pay and career prospects, whereas having children incurs no such penalties for men. In this chapter, one issue we investigate is how far the ideology of motherhood is being challenged. Do country differences regarding support for motherhood go hand-in-hand with more traditional family values, as Popenoe’s notion of familism vs. individualism would suggest? It may be, however, that pro-natalism is unrelated to traditional family values. For example, in France, a country that is not renowned for puritant sexual morality, there are marked economic incentives for having more than one child. From this brief review, we derive our first hypothesis: the ideology of marriage and motherhood will be increasingly challenged throughout Europe and the United States, because labor market changes that enhance women’s autonomy heighten the costs of marriage and motherhood. We further hypothesize that national differences regarding motherhood will reflect national differences in pro-natalist policies rather than differences in traditional morality. Post-Traditional Morality. The weakening of taboos concerning issues related to family and sexuality has become central to depictions of contemporary ‘risk’ society, where the individual is faced with a far greater range of acceptable choice. Beck suggests, that as modernization proceeds, the decisions and constraints to decide multiply in all fields of social action, but especially with regards to sexuality and the family. He writes: “With a bit of exaggeration one could say: ‘anything goes’ . . . Marriage can be subtracted from sexuality, and that in turn from parenthood; parenthood can be multiplied by divorce, and the whole things can be divided by living together or apart” (Beck, 1992: 116). In this chapter we examine people’s moral judgement about four family issues: divorce, adultery, homosexuality, and abortion.
66
jacqueline scott and michael braun
Divorce has, in some of the countries, only recently become legal. There are still huge differences between countries concerning the conditions under which divorce is possible and the consequences (economic, legal, and social) that follow for husband, wife, and child(ren). The Catholic church is still uncompromising regarding the indissolubility of marriage. Interestingly, negative consequences of divorce can sometimes be circumvented if people cohabit rather than marry. Anecdotal evidence suggests that only non-divorced applicants are eligible for university chairs endowed by the German Catholic church. Applicants who have never married are thus in an advantaged position, as cohabitation split-ups do not have the same stigma. Adultery has long been grounds for divorce. In England, ‘innocent’ parties can still get a fast track divorce by citing adultery. (Interestingly, adultery precludes same-sex unfaithfulness which serves as ‘unreasonable behavior’). As many countries move towards no-fault divorce then it might seem that adultery will no longer be viewed with the same disapprobation. However, in an era of AIDS and increased prevalence of STDs there may be strong moral sanctions against unfaithfulness. Indeed, if marriage is viewed as increasingly optional then fidelity may be viewed as a crucial expression of commitment, for those who choose to ‘tie the knot’. The very different European policy responses towards same sex partnerships suggest that there will be marked country differences concerning the acceptability of homosexuality. Public opinion in the Netherlands is likely to be particular liberal in this respect. At the beginning of 1998, the Netherlands instituted a formal registration of partnerships for both heterosexual and homosexual couples which made legally registered cohabitation functionally equivalent to marriage, except that cohabiting couples do not have the right to adopt. France has also adopted a Nordic approach with the introduction of Civil Solidarity Pacts (PACS) which allow both homosexual and heterosexual couples to enter legal agreements which confer many of the same rights as marriage. Abortion is one of the most contentious of family policies, and it seems likely that the very different European policies will be linked to very marked country differences in the way abortion is viewed. Abortion has, in some countries, only recently become legal (and remains illegal in Ireland), and there are still considerable differences between countries in the way abortion laws are implemented.
individualization of family values?
67
The Catholic Church remains uncompromising in its oppositional stance. Religion and religiosity. The distinction between Catholicism and Protestantism in terms of the relative emphasis on conformity and individual autonomy is likely to be particularly important in regard to issues of family morality. Clearly, organized churches hold far more sway in some countries than in others, and one major source of variation in cross-national attitudes is likely to be the extent to which the Catholic Church has maintained its hold as a primary source of moral authority. Similarly religiosity is likely to be an important factor in within-country differences. With respect to post-traditional morality we hypothesize that there will be a general move towards greater liberalism as modernization, including greater secularization, proceeds. We further hypothesize that values will be broadly in line with legislative initiatives and, thus, expect Ireland to be unusually traditional in its rejection of a post-traditional stance and the Netherlands unusually liberal, particularly regarding homosexuality. National differences. Thus far we have made occasional reference to individual nation differences in family policy. But is there a more appropriate way of grouping countries in terms of their family orientation? One possible typology would be to use the welfare-state model proposed by Esping-Andersen (1990) and its modifications and extensions especially by Gornick and her collaborators (Gornick, 1999; Gornick, Meyers, and Ross, 1997, 1998). Esping-Andersen (1990) has suggested three types of capitalist welfare states on the basis of their socio-political orientations and labor market structures: the social-democratic type is represented by the Scandinavian countries, the liberal by the Anglo-Saxon, and the conservative-corporatist type by the continental European countries. Socio-political orientations include gender differences, structural conditions such as the availability of non-familial childcare, and beliefs about whether policy should maximize individual or family interests. In the countries of the social-democratic type, policy aims at a leveling of gender differences in the public domain and in families (Hantrais and Letablier, 1996). Both inequalities of opportunity and of outcomes were targeted by restricting the working of market mechanisms and mitigating their consequences. A high degree of female labor-force participation was facilitated by the provision of a comprehensive childcare system. In liberal welfare states market mechanisms
68
jacqueline scott and michael braun
are much less constrained, and the understanding of equality is confined to equality of opportunity. Family is regarded as a private matter, and the state does not intend to prescribe a certain form of living; it is ‘liberal’ in this sense, too. The philosophical foundation of liberalism is respect for the liberty and autonomy of the individual and, thus, “. . . a liberal state ought not to ram basic values down people’s throats” (Goodin et al., 1999: 40). In the corporatist regime typical gender differences are stabilized by measures which make a deviation from the ‘breadwinner-husband-and-homemaker-wife’ ideal unattractive or even prohibitive, e.g., the generous regulations of maternity and parental leave instead of the provision of a comprehensive system of public childcare (Gauthier, 1996; Gornick, 1999). While it would certainly make our analytical task easier if we could distinguish nations by regime types, there are at least three reasons to doubt whether this typology will serve us well. First, the distinguishing features of the regimes have never been unambiguously present. For example, in nearly all of the Western welfare states the ideal model of ‘breadwinner-husband-and-homemaker-wife’ families was dominant until relatively recently (Lewis, 1992). Second, family policy shows considerable heterogeneity within the Scandinavian, liberal, and corporatist regimes (Gornick, Meyers, and Ross, 1997, 1998; Lewis, 1992; Pfau-Effinger, 2001). In addition, change in family policy is ongoing. Third, some countries are very difficult to classify with regard to their welfare regime. The Netherlands is a case in point, viewed as a corporatist country by Esping-Andersen (1999), and a social-democratic country according to Goodin et al. (1999). Indeed Bahle and Pfenning (2000: 2) conclude: “The Netherlands fall in between three ‘worlds’, combining liberal British traditions of non-intervention, Scandinavian modes of universal social rights, and a traditional social division of labor similar to Germany”. In the light of such concerns, we do not expect any typology to serve us well in analyzing family change. It seems likely that any attempt to group countries would serve to mask important national differences. The following questions guide our analysis: how do attitudes concerning marriage, motherhood, and the morality of family behaviors differ across Europe and the United States of America, and have attitudes converged over time? Are there generational differences regarding family values and, if so, are they equally prevalent across different nations? In addition to these basic questions we also inves-
individualization of family values?
69
tigate how gender and religion relate to family values and attitudinal change across nations and time. Data and measures We use data from the three waves of the European Values Study (1981, 1990, and 1999/2000). We have selected countries for which we have information for all three points of time: France, West Germany, Great Britain, the Netherlands, Ireland, Italy, Spain, and Sweden. We also include for comparison the United States. Our items were selected on the basis of theoretical considerations, but are also based on the results of factor analyses and multidimensional scaling of a much larger set of pertinent items (not shown). The dependent variables were all recoded so that higher values show pro-marriage, pro-motherhood, and pro-traditional values (i.e. viewing as wrong divorce, adultery, homosexuality, and abortion). Pro-marriage is a single item: “Do you agree or disagree with the following statement? Marriage is an outdated institution”. We assigned the following values to the responses: ‘disagree’ = 2 and ‘agree’ = 1. Pro-motherhood is also a single question: “Do you think that a woman has to have children in order to be fulfilled or is this not necessary?” We assigned the following values to the responses: ‘needs children’ = 2 and ‘not necessary’ = 1. Sexual values were measured as an average of the valid values of four items: attitudes towards homosexuality, abortion, divorce, and adultery. Overall reliability is .78. These items were part of a larger battery of presumably deviant behavior. The introductory question was: “Please tell me for each of the following statements whether you think it can always be justified, never be justified, or something in between, using this card.” A scale from 1 through 10 was presented with only the endpoints 1 (= never) and 10 (= always) labeled. The range of the resulting scale is from 1 to 10. We reversed this coding so that high values mean more traditional attitudes. The independent variables are presented as we discuss our findings.
jacqueline scott and michael braun
70
Findings Changes across time Pro-marriage and pro-motherhood. As we would expect there are marked differences between the countries regarding people’s views as to whether marriage is an outdated institution (Figure 3.1), and whether women need children to be fulfilled (Figure 3.2). The data do not support our hypothesis that the ideologies of marriage and motherhood have increasingly come under challenge. The data do however lend support to our hypothesis that country stance regarding promotherhood has more to do with pro-natalist policies than with traditional morality. In Figure 3.1, we can see that the Americans are most supportive of marriage, and the French the least. However, the differences between countries are relatively slight, most values being in the range from 1.7 to 1.9. By contrast, Figure 3.2 shows that countries differ across the full range of the scale, with France being by far the most supportive of motherhood. Plot of Means Marriage outdated institution 2.0
Traditionality
1.9
1.8
France Ireland Italy Netherlands Spain Sweden Great Britain United States West Germany
1.7
1.6 1981
1990 Wave
1999
Figure 3.1: Cross-national differences in pro-marriage attitudes
individualization of family values?
71
Plot of Means Woman needs children 1,9 1,8
Traditionality
1,7 1,6 1,5 France Ireland Italy Netherlands Spain Sweden Great Britain United States Germany
1,4 1,3 1,2 1,1 1,0 1981
1990
1999
Wave
Figure 3.2: Cross-national differences in pro-motherhood attitudes Table 3.1: Pro-marriage attitudes Means
Diff.
Diff.
Std. dev.
1981
1990 1999 99–81 99–90 1981 1990
1999
France (West) Germany Great Britain Ireland Italy Netherlands Spain Sweden United States
1.68 1.83 1.85 1.87 1.77 1.84 1.75 1.85 1.91
1.71 1.85 1.82 1.90 1.86 1.79 1.85 1.86 1.92
1.65 1.80 1.73 1.80 1.83 1.75 1.79 1.80 1.90
–0.02 –0.03 –0.12 –0.08 0.06 –0.09 0.04 –0.05 –0.01
–0.06 –0.05 –0.08 –0.11 –0.03 –0.04 –0.06 –0.06 –0.02
0.47 0.38 0.36 0.34 0.42 0.37 0.44 0.36 0.29
0.46 0.36 0.39 0.30 0.35 0.41 0.36 0.35 0.27
0.48 0.40 0.44 0.40 0.38 0.44 0.41 0.40 0.30
Total
1.82
1.85
1.78
–0.03
–0.06 0.39
0.36
0.41
Mean all countries Std. dev. all countries
1.82
1.84
1.78
0.07
0.06
0.07
72
jacqueline scott and michael braun Table 3.2: Pro-motherhood attitudes Means
Diff.
Diff.
Std. dev.
1981
1990 1999 99–81 99–90 1981 1990
1999
France West Germany Great Britain Ireland Italy Netherlands Spain Sweden United States
1.71 1.36 1.22 1.25 1.53 1.11 1.48 1.16 1.18
1.75 1.44 1.22 1.26 1.64 1.11 1.50 1.21 1.20
1.68 1.50 1.21 1.17 1.56 1.07 1.48 1.24 1.15
–0.04 0.14 –0.01 –0.09 0.04 –0.04 0.001 0.07 –0.03
Total
1,34
1.39
1,39
0.05
Mean all countries Std. dev. all countries
1.33
1.37
1.34
0.20
0.22
0.22
–0.07 0.06 –0.01 –0.09 –0.08 –0.04 –0.02 0.03 –0.06
0.45 0.48 0.41 0.44 0.50 0.31 0.50 0.37 0.38
0.44 0.50 0.42 0.44 0.48 0.31 0.50 0.41 0.40
0.47 0.50 0.41 0.37 0.50 0.26 0.50 0.43 0.35
–0.01 0.47
0.49
0.49
In Tables 3.1 and 3.2 we can see the mean scores and standard deviations per country for the three waves of the EVS study. As the mean scores for marriage are from a scale that ranges from 1 to 2, it can be seen that most countries still view marriage as a relevant (not outdated) institution. In most countries (see Figure 3.1), promarriage positions rose slightly from 1981 to a high in 1990, then fell during the 1990s, ending with the lowest support for marriage in 1999/2000. Great Britain and the Netherlands are the only countries to show a decline across the three waves. Italy and Spain also differ from other countries in being more pro-marriage at the end of the millennium than two decades earlier. The pro-motherhood means (also based on a range from 1 to 2) across countries cover the entire scale range (see Table 3.2). What is remarkable is how little variation there has been in two decades. Sexual values. The component questions tapping whether certain behaviors can ever be justified show markedly different patterns according to which issue is at stake. The questions concern homosexuality (Figure 3.3), abortion (Figure 3.4), divorce (Figure 3.5), and adultery (Figure 3.6). The rank order of justifiability is fairly constant across countries, with adultery being seen as the least justifiable,
individualization of family values?
73
Plot of Means Homosexuality 10
Traditionality
9 8 7 6
France Ireland Italy Netherlands Spain Sweden Great Britain United States West Germany
5 4 3 2 1981
1990 Wave
1999
Figure 3.3: Cross-national differences in views about homosexuality Plot of Means Abortion 10
9
Traditionality
8
7
6
5
4
3 1981
1990
1999
France Ireland Italy Netherlands Spain Sweden Great Britain United States West Germany
Wave
Figure 3.4: Cross-national differences in views about abortion
jacqueline scott and michael braun
74
Plot of Means Divorce 9 8
Traditionality
7 6 France Ireland Italy Netherlands Spain Sweden Great Britain United States West Germany
5 4 3 2 1981
1990
1999
Wave
Figure 3.5: Cross-national differences in views about divorce Plot of Means Adultery 10
Traditionality
9
8 France Ireland Italy Netherlands Spain Sweden Great Britain United States West Germany
7
6 1981
1990
1999
Wave
Figure 3.6: Cross-national differences in views about adultery
individualization of family values?
75
Plot of Means SEXVAL 9,0 8,5 8,0
Traditionality
7,5 7,0 6,5 France Ireland Italy Netherlands Spain Sweden Great Britain United States West Germany
6,0 5,5 5,0 4,5 4,0 1981
1990
1999
Wave
Figure 3.7: Cross-national differences in sexual values
and divorce the most justifiable. The one exception is the Netherlands where homosexuality is seen as more justifiable than divorce. For most countries, there has been the expected over-time shift in the more permissive direction, with respect to abortion, homosexuality, and divorce. However, views on adultery run counter to the general trend and, in all countries, there has either been negligible shift or a shift in the more conservative direction. Although attitudes to adultery have become somewhat less liberal over time, attitudes have been relatively conservative in all years, and on a 1–10 scale where always justifiable is 1 and never justifiable is 10, mean approval roughly lies between 7 and 9—with Ireland the most conservative, and France the most liberal. The liberal position of France further reinforces our earlier claim that the pro-natalist position of France is not related to a traditional stance regarding family morality. We defined our sexual values index (Figure 3.7) as an average of the values for abortion, adultery, divorce, and homosexuality. Table 3.3 shows the mean scores and standard deviations per country for the three waves of the EVS study. Over all countries,
76
jacqueline scott and michael braun
attitudes become less traditional from 1981 to 1990 and, in particular, from 1990 to 1999. However, there are clear country differences in the pattern of change. In 1981, Ireland, the United States, and Spain are most traditional, whereas France, the Netherlands, and Sweden are least traditional. In 1990, Ireland and the United States are still the most traditional countries but Spain has become less traditional, and is now on an equal level with Britain and other countries. On the non-traditional side there is little change but for the dramatic further decline in traditionality of the Netherlands. In 1999, Ireland remains alone in its position as the most traditional country while the United States has also blended into the middle field. Spain, has become one of the less traditional countries and, in 1999, is on the same level with France. Sweden, which did not show any change between 1981 and 1990, has experienced a dramatic decrease in traditionality, surpassing the Netherlands as the least traditional country. While from 1981 to 1999 most countries have become less traditional, change in France, Italy, Great Britain, and West Germany has only been modest. Table 3.3: Sexual values Means
Diff.
Diff.
Std. dev.
1981
1990 1999 99–81 99–90 1981 1990
1999
France West Germany Great Britain Ireland Italy Netherlands Spain Sweden United States
6.43 7.01 7.08 8.56 7.14 6.51 7.68 6.47 7.79
6.43 6.67 7.17 8.13 6.78 5.60 7.05 6.44 7.55
5.87 6.45 6.58 7.71 6.81 5.33 5.83 4.67 5.98
Total
7.29
6.92
6.14
Mean all countries Std. dev. all countries
7.18
6.87
6.14
.71
.73
.88
–.56 –.57 –.50 –.85 –.33 –1.18 –1.84 –1.80 –1.80
–.57 –.22 –.59 –.42 .03 –.27 –1.22 –1.77 –1.57
2.26 2.26 2.04 1.71 2.33 2.40 2.23 2.17 1.86
2.15 2.21 1.76 1.67 2.25 2.14 2.26 2.06 1.89
2.17 2.36 2.06 1.94 2.17 2.02 2.55 1.92 2.38
2.23
2.17
2.35
individualization of family values?
77
The bottom row in Table 3.3 presents the standard deviations of the means of the 9 countries and tells us most directly whether we are dealing with divergence or convergence between countries concerning traditionality with regard to moral values. Hardly any change towards divergence or convergence took place in the 1980s (the deviation score developed from 0.71 to 0.73). In the 1990s, however, there is a modest trend towards country divergence (from 0.73 to 0.88). The penultimate row shows that the change in overall country mean has been more marked in the 1990s than in the 1980s. This relative speed of change across the two decades was driven by the changes in permissiveness towards homosexuality which increased markedly in the 1990s. The Netherlands led this trend, with the main shift to permissiveness coming a decade earlier, in the 1980s. A generational divide? We argued that younger generations in particular will adopt a less traditional stance towards family values. Is the notion of a generational divide supported regarding attitudes towards marriage, motherhood, and sexual values? Our analysis is in two parts. First, we show the generational divides for mean scores across the three waves of the EVS study. Second, we investigate further whether a generational divide holds when controlling for other variables that might be relevant to attitudinal change, including gender, marital status, and religion. A three-type generational categorization is used here, differentiating the War Generation, the Baby Boom Generation, and the Baby Bust Generation. The war generation refers to those born before 1940, the boom generation to those born between 1940 and 1960, and the bust generation to those born after 1960. In Figures 3.8 and 3.9, we see the generational divide in Europe and the United States for attitudes concerning marriage and motherhood. The country differences that we saw in Figures 3.1 and 3.2 remain largely the same, and the general trend of the generational differences remains constant across the years, with the war generation the most pro-motherhood and pro-marriage, and the baby bust generation the least. There are, however, a few exceptions. For example, among the French the lack of support for marriage seems to have bottomed out after the baby boom generation and, in 1999, there is little difference in pro-natalist sentiments between the boom and bust generations of Sweden and the Netherlands.
jacqueline scott and michael braun
78
Plot of Means Marriage outdated institution 2,00 1,95 1,90
Traditionality
1,85 1,80 France Ireland Italy Netherlands Spain Sweden Great Britain United States West Germany
1,75 1,70 1,65 1,60 1,55 War
Boom Bust 1981
War
Boom Bust 1990
War
Boom Bust 1999
Figure 3.8: Generational divide in pro-marriage attitudes Plot of Means Woman needs children 2,0 1,9 1,8
Traditionality
1,7 1,6 1,5
France Ireland Italy Netherlands Spain Sweden Great Britain United States West Germany
1,4 1,3 1,2 1,1 1,0 War Boom Bust 1981
War Boom Bust 1990
War Boom Bust 1999
Figure 3.9: Generational divide in pro-motherhood attitudes
1.75 1.82
Mean 13 countries 1.87 1.78
Yes
Yes/war Yes/war Yes/war Yes/war Yes/war Yes/war Yes No Yes
Diff. 1.66 1.83 1.80 1.91 1.83 1.76 1.85 1.84 1.92
1.89 1.82
1.90 1.83
1.78 1.94 1.87 1.92 1.88 1.87 1.94 1.88 1.93
1.71 1.85 1.82 1.90 1.86 1.79 1.85 1.86 1.92
1.79 1.84
1.79 1.85
1.68 1.73 1.74 1.86 1.87 1.75 1.76 1.86 1.89
EVS wave 1990 War Boom Bust Total
Yes
Yes/war Yes Yes/war Yes/war No Yes/war Yes No Yes/war
Diff. 1.63 1.80 1.78 1.80 1.81 1.72 1.82 1.78 1.92
1.85 1.79
1.86 1.78
1.74 1.89 1.82 1.86 1.89 1.82 1.89 1.83 1.95
1.65 1.80 1.73 1.80 1.83 1.75 1.79 1.80 1.90
1.74 1.78
1.74 1.78
1.63 1.71 1.66 1.74 1.81 1.73 1.71 1.80 1.87
EVS wave 1999 War Boom Bust Total
Yes
Yes/war Yes Yes Yes/war Yes/war No Yes No Yes
Diff.
War = war generation, born before 1940; Boom = baby boom generation, born 1940–1959; Bust = baby bust generation, born after 1959 Diff. = does the youngest generation (baby bust) differ from the older generations? yes = differs from both older generations; yes/war = differs from war generation but not from baby boom generation; yes/boom = differs from baby boom generation but not from war generation; no = no difference between generations; differences are analyzed with one-way anova (post hoc multiple comparisons. Tukey HSD test)
1.75 1.82
1.68 1.83 1.85 1.87 1.77 1.84 1.75 1.85 1.91
1.87 1.78
1.63 1.75 1.81 1.82 1.72 1.81 1.59 1.81 1.85
Total
1.61 1.75 1.84 1.85 1.74 1.81 1.68 1.82 1.90
1.77 1.91 1.88 1.92 1.81 1.90 1.83 1.89 1.93
France West Germany Great Britain Ireland Italy Netherlands Spain Sweden United States
EVS wave 1981 War Boom Bust Total
Table 3.4: Pro-marriage attitudes by generation and wave
individualization of family values? 79
1.27 1.33
Mean 13 countries 1.40 1.28
1.84 1.59 1.33 1.34 1.79 1.23 1.64 1.28 1.24
1.71 1.37 1.16 1.25 1.59 1.07 1.47 1.20 1.17
1.48 1.33
1.74 1.44 1.22 1.26 1.64 1.11 1.50 1.21 1.20
1.29 1.37
1.33 1.39
1.68 1.28 1.12 1.18 1.56 1.03 1.39 1.14 1.19
EVS wave 1990 War Boom Bust Total
Yes/war 1.50 1.35
Yes/war Yes/war No Yes/war Yes/war Yes/war Yes/war Yes/war No
Diff. 1.78 1.64 1.30 1.32 1.70 1.18 1.66 1.31 1.15
1.69 1.49 1.20 1.15 1.55 1.05 1.50 1.21 1.14
1.45 1.33
1.68 1.50 1.21 1.17 1.56 1.07 1.48 1.24 1.15
1.28 1.34
1.32 1.39
1.62 1.37 1.17 1.09 1.49 1.03 1.37 1.24 1.15
EVS wave 1999 War Boom Bust Total
Yes/war 1.54 1.37
Yes/war Yes Yes/war Yes/war Yes/war Yes/war Yes Yes/war No
Diff.
Yes
Yes Yes Yes/war Yes/war Yes Yes/war Yes No No
Diff.
War = war generation, born before 1940; Boom = baby boom generation, born 1940–1960; Bust = baby bust generation, born after 1960 Diff. = does the youngest generation (baby bust) differ from the older generations? yes = differs from both older generations; yes/war = differs from war generation but not from baby boom generation; yes/boom = differs from baby boom generation but not from war generation; no = no difference between generations; differences are analyzed with one-way anova (post hoc multiple comparisons, Tukey HSD test)
1.27 1.34
1.71 1.36 1.22 1.25 1.53 1.11 1.48 1.16 1.18
1.41 1.29
1.65 1.27 1.22 1.16 1.48 1.05 1.35 1.07 1.16
Total
1.68 1.29 1.18 1.23 1.46 1.05 1.42 1.11 1.15
1.77 1.45 1.26 1.31 1.60 1.20 1.56 1.21 1.21
France West Germany Great Britain Ireland Italy Netherlands Spain Sweden United States
EVS wave 1981 War Boom Bust Total
Table 3.5: Pro-motherhood attitudes by generation and wave
80 jacqueline scott and michael braun
individualization of family values?
81
Plot of Means SEXVAL 10 9
Traditionality
8 7 France Ireland Italy Netherlands Spain Sweden Great Britain United States West Germany
6 5
4 3 War
Boom Bust 1981
War
Boom Bust 1990
War
Boom Bust 1999
Figure 3.10: Generational divide in sexual values
Tables 3.4 and 3.5 show mean differences across generations and time, and indicate whether and which generational differences are significant. From Table 3.4 we can see that in the Netherlands and Sweden, there is no difference between generations in the most recent survey wave regarding marriage. Table 3.5 shows that there is no difference in 1999 between generations in both Sweden and the United States concerning the desirability of motherhood. The general patterns however that can be inferred from Figure 3.8 holds and, for the most part, generational differences are pronounced, particularly between the pre-war and more recent generations. Data from the three waves of EVS on sexual values present clear generational divides (Figure 3.10). In 1981, the average values across the nine countries shows the youngest generation, born after 1960, is only different from the war generation; whereas in both 1990 and 1999 it differs also from the bust generation in adopting distinctively less traditional values. The positioning of the different generations in the single countries shows a rather complex picture. In Germany, Italy, and Spain the initial generational divide between the bust and war generations is extended to a three-way generational divide by the end of the millennium. Ireland and Great Britain (with the exception of 1990) show
6.44 7.18
Mean 13 countries 7.94 6.65
7.31 7.67 7.76 8.81 7.71 6.70 8.17 6.83 7.88
6.02 6.26 6.83 7.89 6.48 5.09 6.97 6.24 7.25
7.65 6.56
6.44 6.67 7.17 8.13 6.78 5.60 7.05 6.44 7.54
6.32 6.87
6.24 6.91
5.91 5.64 6.60 7.52 6.24 5.12 6.06 6.29 7.45
EVS wave 1990 War Boom Bust Total
Yes/war 7.74 6.62
Yes/war Yes/war Yes Yes Yes/war Yes/war Yes/war Yes/war Yes/war
Diff.
Yes
Yes/war Yes Yes/war Yes Yes/war Yes/war Yes Yes/war Yes/war
Diff. 5.71 6.20 6.61 7.65 6.76 4.99 5.90 4.58 5.93
7.15 6.04
7.29 6.03
6.78 7.70 7.48 8.96 7.87 6.32 7.36 5.50 6.34
5.87 6.45 6.58 7.71 6.81 5.33 5.83 4.67 5.98
5.61 6.14
5.56 6.14
5.51 5.50 6.12 6.93 6.19 5.13 4.88 4.36 5.89
EVS wave 1999 War Boom Bust Total
Yes
Yes/war Yes Yes Yes Yes Yes/war Yes Yes/war No
Diff.
War = war generation, born before 1940; Boom = baby boom generation, born 1940–1960; Bust = baby bust generation, born after 1960 Diff. = does the youngest generation (baby bust) differ from the older generations? yes = differs from both older generations; yes/war = differs from war generation but not from baby boom generation; yes/boom = differs from baby boom generation but not from war generation; no = no difference between generations; differences are analyzed with one-way anova (post hoc multiple comparisons, Tukey HSD test)
6.62 7.29
6.43 7.01 7.08 8.56 7.14 6.51 7.68 6.47 7.79
8.02 6.75
5.69 6.01 6.28 7.73 6.52 5.97 6.64 5.60 7.51
Total
5.76 6.30 6.69 8.21 6.55 5.87 6.94 6.04 7.49
7.42 7.89 7.81 9.25 7.92 7.52 8.51 6.92 8.18
France West Germany Great Britain Ireland Italy Netherlands Spain Sweden United States
EVS wave 1981 War Boom Bust Total
Table 3.6: Sexual values by generation and wave
82 jacqueline scott and michael braun
individualization of family values?
83
the full generational divide throughout. In France, the Netherlands, and Sweden the divide is only between bust and boom generations, and in the United States, by 1999, the generational distinctions have disappeared. Not surprisingly, the generational divide for sexual values is most marked in Spain and Ireland, two Catholic countries which experienced dramatic social change in the last two decades (Table 3.6). Thus one might argue that over the last two decades of the 20th century generational distinctiveness has grown in Europe with a younger generation increasingly divergent from older ones in terms of its relative permissiveness regarding various family and sexual behaviors, while in the United States generational divides have become blurred. The relative impact of generational divides To what extent do these generational divides hold when other relevant variables such as marital status, religiosity, and, of course, gender are taken into account? Americans are more traditional than all Europeans with respect to marriage, and only the Irish are more traditional than Americans regarding sexual values (Table 3.7). However, with respect to motherhood only the Dutch are less traditional than the Americans. Among Europeans, the Dutch and the Swedish show particularly permissive values regarding sexual behaviors. Women tend to be less traditional regarding sexual values and motherhood but more pro-marriage than men. The never married are much less traditional than the married and widowed respondents. Marriage and employment are both indicative of life-cycle stages, and interestingly, the sizes of effects both for pro-motherhood and sexual values are reduced markedly when generation is introduced in the second model (columns 2, 4 and 6). Religion plays a relatively minor role, with the non-religious being markedly less traditional regarding sexual values than are Catholics. However, church attendance has one of the strongest effects, with those who attend church infrequently holding far less traditional values than frequent church attenders on both marriage and sexual values. The year entries for 1990 and 1999 serve as proxies of a period effect: in both years sexual values have become less traditional than in 1981, with the 1999 effect being the stronger of the two. In the second model, when generational effects are introduced,
84
jacqueline scott and michael braun
the period effects remain nearly as strong. For attitudes favoring motherhood the period effects show even a slight increase. This indicates that change is not simply reflecting shifts in generational membership. Table 3.7: (Logistic) Regression of importance of marriage, importance of children, and sexual values on country dummies, sex, marital status, employment status, religion, church attendance, study waves, and generations Pro-marriage France Great Britain West Germany Italy Spain Netherlands Sweden Ireland sex cohabitating divorced or separated never married employed no religion Protestant other religion church rarely or never church occasionally church once a month wave 1990 wave 1999 war generation boom generation Adjusted R square N
3.31* –1.06* –.63* –.61* –.77* –.78* –.62* –.30* –.96* .07* –1.37* –1.09* –.77* –.05 –.17* .30* .29* –1.50* –.89* –.53* .29* –.08* .10 36931
3.31* –1.07* –.64* –.61* –.77* –.78* –.61* –.31* –.95* .09* –1.40* –1.07* –.78* .02 –.15* .29* .29* –1.47* –.87* –.52* .28* –.19* .08* –.18* .11 36931
Pro-motherhood –.71* –1.06* 2.67* 2.67* .41* .39* 1.35* 1.33* 1.87* 1.87* 1.49* 1.49* –.66* –.66* .47* .43* .11 .13 –.37* –.33* –.72* –.61* –.21* –.19* –.65* –.50* –.43* –.31* –.29* –.26* –.11* –.13* –.08 –.06 –.58* –.51* –.27* –.21* –.17* –.13* .16* .21* –.01 .08* .41* .01* .19 .20 35609
35609
Sexual values –.03* .04* –.02* –.04* –.03* –.11* –.07* .04* –.05* –.06* –.08* –.16* –.11* –.09* .01 .00 –.39* –.24* –.12* –.09* –.22* .28
–.04* .03* –.03* –.04* –.03* –.11* –.08* .05* –.05* –.06* –.07* –.11* –.06* –.09* .00 .01 –.36* –.22* –.11* –.08* –.19* –.15* –.15* .30
38912 38912
Baseline or omitted categories are the United States, male, married (including widowed), not employed, Catholics, weekly church attendance, wave 1981, bust generation; * = significant on the .05 level.
individualization of family values?
85
Summary and conclusions Has the family individualization thesis held up to empirical scrutiny? Are family values becoming increasingly relativized? Are traditional absolutes being overturned? Our analysis of two decades of attitudinal change in Europe and America suggests that the individualization thesis is overstated. Marriage is considered by most countries a relevant (not outdated) institution, and there has been remarkably little country change in two decades. Pro-motherhood sentiments also remain fairly stable across countries and over time. The notion of relativized values commands greatest support when we examine sexual values, where traditional values have been in decline. Yet even regarding sexual permissiveness there is little support for the claim that in family values ‘anything goes’. While the Swedes and Dutch are becoming increasingly tolerant with respect to homosexuality, most countries still show moral concerns about same-sex relations. With the exception of Sweden, attitudes towards abortion have changed very little. It is true that attitudes towards divorce have become markedly more liberal across the decades, but this liberality does not extend to adultery where, if anything, disapprobation has increased. Thus our analysis does not convey much support for the idea of a ‘convergence on individualism’. Instead we see considerable diversity across the Western world in the extent to which traditional family values are coming under challenge. Careful scrutiny of the evidence concerning attitudinal data across two decades and nine countries provides little comfort for those who claim that new individualism has brought an almost limitless range of acceptable choice. Of course, it could be the case that whereas marriage and motherhood were once the traditional choice, the same choice now results from individual preference. However, such a logic means that there is no scope for using empirical analysis to test the individualism thesis: change or no change would both corroborate individualism. Moreover, prominent social theorists, such as Beck and Castells, do claim that post-modernity and the end of patriarchy are associated with the demise of traditional values. Contrary to such predictions, our crossnational analysis of attitudes show that traditional values still garner considerable support. Yet it is important not to swing to the other extreme and overstate the continuities in traditional values. There is evidence of a
86
jacqueline scott and michael braun
generational divide in Europe and the United States, with the war generation being rather distinctive in its championing of traditional morality. This applies right across the spectrum of family values considered here in terms of marriage, motherhood, and sexual values. It is necessary however to get the importance of the generational divide into perspective. In the family value domain, it is secularism that provides the strongest challenge to traditional views. Unless secularism trends are reversed, traditional family values are likely to continue to be challenged. One interesting issue, that we can only speculate on, is whether a secular basis for supporting family orthodoxy will emerge. We have argued that there are sound political and economic reasons for boosting pro-natalist values. There are equally sound economic reasons for promoting fidelity, given the huge health and social costs that are incurred through the pan-epidemics of STDs and AIDS. How different would family morality be, if based on secular interest rather than religious or moral constraints? We suspect that the answer is ‘not very’: traditional family values may prove surprisingly resilient. References Bahle, T. & A. Pfenning 2000. Introduction. In A. Pfenning & T. Bahle (eds.), Families and Family Policies in Europe. Comparative Perspectives. Frankfurt: Peter Lang. Beck, U. 1992. The Risk Society: Towards a New Modernity. London: Sage. Beck, U. & E. Beck-Gernsheim 2004. Families in a runaway world. In J. Scott, J. Treas. & M. Richards (eds.), Blackwell Companion to Sociology of Families. Oxford: Blackwell. Bumpass, L. 1990. What’s happening to the family? Interactions between demographic and institutional change. Demography, 27, 483–98. Castells, M. 1997. The end of patriarchalism. In Power of Identity, vol. 2. The Information Age, Economy, Society and Culture. Oxford: Blackwell. Esping-Andersen, G. 1990. The Three Worlds of Welfare Capitalism. Cambridge: Polity Press. ——. 1999. Social Foundations of Postindustrial Economies. Oxford: Oxford University Press. Gauthier, A.H. 1996. The State and the Family. A Comparative Analysis of Family Policies in Industrialized Countries. Oxford: Clarendon Press. Goodin, R.E., B. Headey, R. Muffels & H.J. Dirven 1999. The Real Worlds of Welfare Capitalism. Cambridge: Cambridge University Press. Gornick, J.C. 1999. Gender equality in the labour market: women’s employment and earnings. In D. Sainsbury (ed.), Gender and Welfare State Regimes. Oxford: Oxford University Press. Gornick, J.C., M.K. Meyers & K.E. Ross 1997. Supporting the employment of mothers: policy variation across fourteen welfare states. Journal of European Social Policy, 7, 45–70.
individualization of family values?
87
——. 1998. Public policies and the employment of mothers: a cross-national study. Social Science Quarterly, 79, 35–54. Hantrais, L. & M.-T. Letablier 1996. Families and Family Policies in Europe. London: Longman. Kiernan, K. 2000. European perspectives on union formations. In L. Waite, C. Bachrach, M. Hindin, E. Thomson & A. Thornton (eds.), Ties that Bind: Perspectives on Marriage and Cohabitation. New York: Aldine de Gruyter. Lesthaeghe, R. 1995. The second demographic transition in western countries: an interpretation. In K. Mason & A.-M. Jenson (eds.), Gender and Family Change in Industrialized Countries. Oxford: Clarendon Press. Lewis, J. 1992. Gender and the development of welfare regimes. Journal of European Social Policy, 2, 159–173. Popenoe, D. 1988. Disturbing the Nest: Family Change and Decline in Modern Societies. New York: Aldine de Gruyter. Pfau-Effinger, B. 2001. Wandel wohlfahrtsstaatlicher Geschlechterpolitiken im soziokulturellen Kontext. In B. Heintz (ed.), Geschlechtersoziologie. Kölner Zeitschrift für Soziologie und Sozialpsychologie, Sonderheft, 41, 487–511. Scott, J. 1999. Family change: revolution or backlash in attitudes? In S. McRae (ed.), Changing Britain, Families and Households in the 1990s. Oxford: OUP. Schumpeter, J.A. 1988 (1942). Decomposition. Population and Development Review, 14, 499–506. (Originally published 1942 in Capitalism, Socialism and Democracy. New York: Harper Row).
CHAPTER FOUR
ERODING WORK VALUES? Peter Ester, Michael Braun, and Henk Vinken More ambitious work values for Europe1 Flexibility and employability are key words in describing the main features of the modern European service and information-based economy (Ester et al., 2001). The modern worker is expected to permanently maintain, secure, and of course practice his or her ‘hard’ and ‘soft’ skills, human capital, and added-value for the organization. European employees are nowadays presumed to be highly flexible workers who can be employed at multiple positions within the organization and for whom continuous competence development is selfevident. Severe international competition, rapid globalization of the European economy, habitual innovation necessity, dynamic diffusion of ICT-applications, the shortening of products’ life cycles, the never ending need for customer and service orientations, integral quality care, and the vital importance of sustainable professionalization, all imply that more and more is expected from the average European worker (Ester & Vinken, 2001). Workers need to be willing to adapt, adjust, and be mobile, to be creative and innovative, and to develop multiple ‘portable’ skills. Internal and external mobility are seen as meaningful distinctive individual ambitions. Extensive employability is expected with respect to changing organizational goals and contexts, working times, and working locations. Investment in keeping one’s knowledge and competences up-to-date is seen as a crucial personal drive. The modern employee becomes an entrepreneurial employee. In summary: flexible and employable workers with work values aimed at flexibility and employability are the vital assets of modern European companies operating in a highly competitive, rapidly changing, and dynamic global environment. 1 Many thanks to Ingwer Borg (ZUMA, Mannheim) and Amelia Román (OSA, Institute for Labour Studies, Tilburg University) for their very helpful comments on an earlier draft of this chapter.
90
peter ester, michael braun, and henk vinken
At the 2000 Lisbon European Council meeting, the European Union redefined and sharpened its basic future policy goals and ambitions: by 2010 Europe wants to be “the most competitive and dynamic knowledge-based economy in the world”, even aiming at surpassing the United States. The Council particularly stressed the importance of active aging, lifelong learning, employability, and flexibility. In view of the rapid aging of the European community and the short average working hours, Europeans are prompted to work longer and harder. Europe wants to renovate its basic economic structures and innovation potentials and aims at leading the global economic field.2 This radical change towards a post-industrial knowledge society implies a fundamental paradigm shift, both in terms of economic structures, financial incentives, and work values: higher productivity, more R&D investments, but also more ‘can-do’ work values. These values assume a new mind set of the European employee: in a post-industrial, service-oriented European knowledge economy, mere extrinsic or conventional work values are no longer sufficient to do the job. These values have to be augmented if not replaced by intrinsic work values which stress continuous personal development (in line with the life-long learning debate), the use of initiative and taking responsibility (in line with the entrepreneurial emphases), and the aim to challenge one’s abilities (Ester et al., 2001). This means that the cultural factor as embedded in fundamental work values becomes a prominent factor on the European labor market.3 It also means that Europe as an intended globally competitive economy will have to make sure that its workforce is willing to embrace intrinsic work values. If Europe will stick to conventional extrinsic work values it is likely to face a slowdown in innovation and competitiveness and serious problems of sustainability of its social security system.4 Being in the front row of innovation and competi2 Recently, the EU’s operational targets have been revised and relaxed somewhat, particularly due to disappointing medium term results. See: http:// europa.eu.int/rapid/pressReleasesAction.do?reference=MEMO/05/123&format=HTM L&aged=0&language=en&guiLanguage=en 3 In the words of the Presidency Conclusions of the Lisbon European Council “a European framework should define the new basic skills to be provided through lifelong learning: IT skills, foreign languages, technological culture, entrepreneurship and social skills”. 4 Extrinsic work values are also referred to as traditional, conventional or instrumental work values as synonyms, and intrinsic work values as modern, unconventional or expressive work values.
eroding work values?
91
tiveness, what is more, being the champion in these fields is seen as the only strategy to (financially) uphold a system that in the near future does not unnecessarily exclude others such as the less welleducated or younger generations. Refraining from action now, now that global competition is on the rise and is coming from more directions than ever (not only the United States or Japan, but today also China), would seriously endanger an inclusive system, so it is believed. The response to this globalization-driven challenge is not defensive but offensive as well as very demanding. Official EU policy claims that Europeans simply have to work more, harder, and longer, that they have to be more creative, competitive, and flexible, and— though more implicitly—abstain from a traditional nine-to-five mentality. In short and straightforward vocabulary: Europe has to beat the United States in the economic rat race. But has it the mind set to do so? Does it embrace the necessary intrinsic work values? What in fact are the work values of Europeans and what trends—if any— can be observed? Here we arrive at the main issue to be addressed in this chapter: do European societies—compared to the United States—differ in fundamental work values?5 Which societies in Europe stand out from other societies in terms of a distinct profile of work values? Moreover, the issue comes to the fore—in line with the general scope of this book—whether young European generations in particular are embracing new work values. The main reason here is that young people are believed to be the first to adopt and shape work values underlying the future knowledge economy (Diepstraten et al., 1998; Vinken et al., 2002). Vice versa, from them it is expected that they in particular embrace the highly necessary intrinsic work values. They form if not the present, then certainly the future core on which the European model is and will be floating. Debating the sustainability of the European model in a global economy is speculating on younger generations an sich. The question then becomes important, even crucial, whether European (and American) young generations are supporting the ‘proper’ work values, those that are conducive to the knowledge economy. But before answering these
5 Europe’s economic competitiveness does not, of course, solely depend on the basic work values of its work force but also on, for example, organizational factors (e.g., performance schemes, training facilities), physical infrastructure, available resources, innovation investment patterns, productivity, ICT-applications, and leadership.
92
peter ester, michael braun, and henk vinken
questions we need to address the issue of work values somewhat further by pointing at several definitional and methodological topics. Work values Many attempts have been made by social scientists to map different values and different value structures across cultures (Roe & Ester, 1999; Super & ”verko, 1995; Vinken et al., 2004). As indicated in the first chapter, Hofstede (2001), for instance, points at basic values such as power distance, individualism versus collectivism, femininity versus masculinity, and uncertainty tolerance versus uncertainty avoidance. Schwartz (1994) differentiates values in terms of autonomy versus embeddedness, egalitarianism versus hierarchy, and harmony versus mastery. Work values are specific expressions of general values in the work setting and can be “ordered by their importance as guiding principles for evaluating work outcomes and settings, and for choosing among different work alternatives” (Ros et al., 1999: 54). In this sense work values are more specific than general basic values. It is important to note that in modern society work values are typically considered as salient, basic, and influential.6 The importance of the work role in many cultures makes work values into core values that take a cardinal position in the overall pattern of values (Roe & Ester, 1999).7 A classic distinction in the domain of work values is the one between extrinsic and intrinsic work values (Borg, 1986; Borg & Braun, 1996; Elizur et al., 1991; England & Ruiz Quintanilla, 1994; Van den Elzen, 2002; Zanders, 1987, 1993). This distinction, as we will see, directly relates to the issue of whether Western societies are indeed favoring work values that reinforce flexibility, employability, self-development, and achievement. Extrinsic work values can be defined as conventional or traditional values which prioritize security over other aspects of work, i.e. work is primarily seen as necessary for providing one’s livelihood, and which
6 This is clearly demonstrated by research on ‘work centrality’ as carried out in the context of the ‘Meaning of Work Study’ (England, 1991) as well as the ‘Work Importance Study’ (Super & ”verko, 1995). 7 The authors propose a generic work-values model at three levels (society, group/organization, individual) with three assumed (vertical and horizontal) links: general values, work values, and work activities (Roe & Ester, 1999).
eroding work values?
93
underline the importance of material job features (or outcomes) such as good pay, comfortable working times and vacation arrangements, protection, and the absence of work stress. Intrinsic work values are defined as values that emphasize non-material or postmaterialist job characteristics such as the possibility of personal development, achievement, autonomy, having a say in the work organization, being able to take initiative, and having an interesting, responsible, and challenging job. Intrinsic work values reflect “[. . .] openness to change— the pursuit of autonomy, interest, growth, and creativity in work. Extrinsic work values express conservation of values; job security and income provide workers with the requirements needed for general security and maintenance of order in their lives” (Ros et al., 1999: 55). Employees stressing extrinsic work values attach a different meaning to the importance of work. Extrinsic work values focus on ‘the security-of-the-taken-for-grantedness’, intrinsic work values stress ‘the ethic of self-development’. It has to be added, though, that different meanings do not per definition imply that these work values are mutually exclusive or unrelated.8 Employees may very well support both extrinsic and intrinsic work values.9 “The emerging emphasis on expressive work values does not mean that instrumental values are rejected or denied. On the contrary, such qualities are, to large publics, still highly relevant” (Halman, 1999: 41). Nevertheless, it can be argued that particularly in a service-oriented network economy that emphasizes flexibility and employability, the wide diffusion of intrinsic work values is a sociological and cultural advantage.10 The more the workforce in a nation supports intrinsic work values, so our argument goes, the better it is mentally equipped to be employable, competitive, and innovative in a globalizing economy, and to
8 Van den Elzen (2002: 75) found a Pearson correlation of .37 between extrinsic and intrinsic work values (based on the 1990 module of the European Values Study). 9 In their study of Dutch people’s expectations of future developments in the domain of work, Ester & Vinken (2001) observed a continuous emphasis on both extrinsic and intrinsic work values. Studies calculating the correlations of ipsative scores did however observe some trade-offs: people who value extrinsic work values very highly, tend to somewhat downgrade intrinsic work values (Borg & Braun, 1996; Borg & Brief, 1994). 10 Ros et al. (1999), for instance, found that intrinsic work values are positively correlated with openness to change, whereas extrinsic work values are negatively correlated.
94
peter ester, michael braun, and henk vinken
adjust to and profit from a flexible (‘transitional’) labor market. The ‘new employee’ will embrace intrinsic work values, whereas the ‘traditional employee’ will rather exclusively support extrinsic work values. As noted, we aim to assess whether or not Europeans are more in favor of the work values underlying the concept of the new employee, whether or not there is a trend in this direction, and also whether or not especially young generations are more supportive of these values. The next paragraphs deal with these issues. Data and measurement instruments As in the other chapters of this book, this chapter will use the three waves (1981, 1990, 1999/2000) of the European Values Study as its main data source. EVS contains both longitudinal measures of extrinsic and intrinsic work values as well as some other work-related attitudes—such as work centrality—for both Europe and the United States (cf. Arts & Halman, 2004; Arts et al., 2003; De Witte et al., 2004; Ester et al., 2004). The EVS is certainly not an ideal data set on work orientations in the sense meant in this chapter. Particularly the fact that neither ratings nor systematic rankings were used is a negative case in point. But its strengths clearly are the many thousands Europeans (from many countries) that expressed their values in the domain of work in response to an identical set of items, spanning a period of two decades. However limited the data might be from a methodological perspective, the number of respondents, the number of countries, the number of waves, the length of the time period concerned, and the replication of the same work values, makes the European Values Study a prime (if not only) data source for analyzing our research questions, albeit at a quite general level.11 In all three waves, subjects were given a set of fifteen different work features that (roughly) correspond with the classic typology of extrinsic and intrinsic work values.12 The survey question was as follows:
11 The International Social Survey Program (ISSP, see Braun, 1994) started monitoring work values in 1989 and to date has replicated this study only once in 1997. The next replication is scheduled for 2005/06 in most of the participating countries which means that comparable data will only be available by 2007. 12 The 1999/2000 module included two additional work features: ‘good physical working conditions’ and ‘to have time off at the weekends’.
eroding work values?
95
“Here are some aspects of a job that people say are important. Please look at them and tell me which ones you personally think are important in a job?”13 The selected work values included: ‘good pay’, ‘pleasant people to work with’, ‘not too much pressure’, ‘good job security’, ‘good chances for promotion’, ‘a job respected by people in general’, ‘good hours’, ‘an opportunity to use initiative’, ‘a useful job for society’, ‘generous holidays’, ‘meeting people’, ‘job in which you feel you can achieve something’, ‘a responsible job’, ‘a job that is interesting’, and ‘a job that meets one’s abilities’. Typical extrinsic work values refer to good pay, generous holidays, and not too much pressure, whereas intrinsic work values signify the possibility of taking initiative, a job that has societal relevance, and that leads to personal development.14 The typology of extrinsic and intrinsic work values is a very basic one and has been adjusted or varied upon by numerous work psychologists and work sociologists. A particularly useful classification or ‘facet’ was proposed by Alderfer (1972). He suggests classifying work values according to the need to which they relate, using a facet with three elements: existence needs, relatedness needs, and growth needs. Similar proposals have been made by Maslow (1954), Herzberg et al. (1958), Elizur (1984), and Rosenberg (1957). A second type of
13 Unfortunately, the EVS work-values measurements are not based on ratings and so assessing how similar two value items are is plagued with all the problems of computing binary similarity measures. However, if one employs factor analyses with these dichotomous items one would arrive at the same grouping of items as with the highly similar ISSP items which employ a 5-point rating scale (for ISSP results see Braun & Borg, 2004). In the following analyses we do not make use of factor analysis in constructing factor scores, but use simple sum scores for the two dimensions of intrinsic and extrinsic values, each of them based on five items. Items pertaining to a third dimension, that of social work values, are excluded. See the list of items below. 14 Some of the EVS work items are rather multi-interpretable. For instance: what is ‘good pay’? Good in relation to what? Is good pay similar to ‘high pay’? Also: the item ‘not too much pressure’ is quite confusing. Not too much for what? Could anybody seek ‘too much pressure’ in a job? The counter argument, of course, is that the wordings simply reflect the way people in everyday life tend to talk about job features. Finally, work features such as pleasant people to work with, a respected job, a job useful for society, and meeting people are irrelevant or ambiguous work values in this respect but seem to focus at the saliency of good social relations at the work place (items 2, 6, 9, 11). Finally, good chances for promotion (item 5) can be interpreted both in the framework of intrinsic and extrinsic work values, as promotion usually brings with it both a higher salary and more autonomy and responsibility, and is therefore extremely ambiguous.
96
peter ester, michael braun, and henk vinken
facet is also occasionally used in work values research. It builds on Katz & Kahn (1978) who distinguish work outcomes that result out of a person’s individual performance or out of belonging to a particular reference group, job grade, or organization. Levy and Guttman (1981), Elizur (1984), Borg (1986), and others have proposed classification systems that reflect this idea. Both types of facets are outlined in Table 4.1, using the EVS work value items as an example. The next step, now, is to conceptually relate the fifteen EVS work-value items as classified in Table 4.1 to the work orientations underlying the economic goals Europe—the policy context of our study—is seeking to accomplish in the near future. This is not an easy task as several work values cannot be unequivocally assigned to be either (strongly) pro or (strongly) con advancing these goals (or are simply irrelevant). We will select the most clear and non-ambiguous cases only. We do realize some of these indicators are arbitrary and far from perfect but as a first attempt they are adequate. The last column of Table 4.1 evaluates these indicators. First of all, it has to be emphasized that the EVS job-values question primarily taps subjects’ beliefs about what the important job features are. Work values—as mentioned in Table 4.1—that are not directly on a par with or contrary to the new European work ethic seem to be values that first and foremost highlight individual concerns such as the saliency of good pay, not too much pressure, job security, good hours, and generous holidays (items 1, 3, 4, 7, and 10). These mirror rather traditional, protective, and extrinsic work values—values that were typically characteristic for the industrial era. Such work values primarily reflect the saliency of one’s individual concerns, which do not per se include the broader organizational goals. Selecting, on the other hand, job characteristics such as opportunity to use initiative, achievement possibilities, responsibility, and features such as an interesting job that meets one’s abilities, at the upper-side of one’s work-values hierarchy, points at values that combine individual concerns and organizational goals (items 8, 12, 13, 14, and 15). Prioritizing such values reveals an employee’s ethic to want to make a difference, to be mobile, to be taking control of challenges, and the like, an ethic that thrives in a stimulating, challenging, and societally relevant context. In a nutshell, we base our indices on the following EVS items:
eroding work values?
97
Table 4.1: Fifteen EVS work values, coded on facets Item H M A E R L B
Work value
Criterion Selected for
1 2 3 4 5 6 7 8 9 10 11 12
h m h h m m h m h h h m
c s c c a s c p s c s p
extrinsic — extrinsic extrinsic — — extrinsic intrinsic — extrinsic — intrinsic
13 14 15
m a g k i g 3 m a g k i g 3 m a g k i n 3
Good pay Pleasant people to work with Not too much pressure Good job security Good chances for promotion Job that is respected Good hours Opportunity to use initiative Job useful for society Generous holidays Meeting people Job in which you can achieve something A responsible job Interesting job Job that meets one’s abilities
p p p
intrinsic intrinsic intrinsic
s b p s r r s a b p b a
e e e e g r e g r e r g
i a i i k a i k a i a a
e s e e e s e i s e s i
i n n n i n n g n n n n
1 4 3 2 1 1 4 3 4 4 4 2
H(erzberg) = {h = hygiene, m = motivator}; M(aslow) = {p = physiological, s = security, b = belongingness, r = recognition, a = self-actualization}; A(lderfer) = {e = existence, r = relations, g = growth}; E(lizur) = {i = instrumental-material, k = cognitive, a = affective-social}; R(osenberg) = {e = extrinsic, i = intrinsic, s = social}; L(evy-Guttman) = {i = dependent of individual performance, g = depends on group performance, n = not performance-dependent}; B(org-Elizur) = {1 = fully dependent on individual performance, 2 = more dependent on individual performance than on system, 3 = equally dependent on individual performance and on system, 4 = dependent on system only}. Criterion (pro or con values advancing a more competitive, innovative, flexible, and employable European work force) = {p = pro, c = con, i = irrelevant, a = ambiguous, s = social}.
Extrinsic work values: – good pay – not too much pressure – good job security – good hours – generous holidays Intrinsic work values: – an opportunity to use initiative – a job in which you feel you can achieve something – a responsible job – a job that is interesting – a job that meets one’s abilities
98
peter ester, michael braun, and henk vinken
We use these items to compute summary scores for extrinsic and intrinsic work values, thus avoiding the inherent problems—outlined above—connected to using factor-analytical procedures with these dichotomous items.15 Extrinsic work values Table 4.2 provides an overarching view of a diverse set of results. It shows the mean scores and standard deviations per country for the three waves of the EVS study. This allows us to draw conclusions on the absolute scores of extrinsic work-values adherence in a given country. As the scores are means on a scale ranging from 0 to 1, the mean scores indicate how many of the included items are on average endorsed in a given country (e.g., a mean country score of .43 equals a public support for 43% of the corresponding items) and compare these figures with the ones in other countries. The data shows that in 1999 in Italy, Northern Ireland, Ireland, and the United States the support for extrinsic work values is quite strong with more than 60% of the items endorsed on average. Lowest support is observed in Denmark and France. Great Britain, (West) Germany, and Iceland indicate a 50% adherence with extrinsic work values. All in all, we may conclude that at the brink of the third millennium, extrinsic work values still have the support of considerable population proportions in Europe and the United States. The longitudinal nature of the data also allows us to assess the trends in support of these values over time per country and to compare these trends within a given country with the same trend in other countries. We can see that, over all countries, in the last twenty years of the 20th century the popularity of extrinsic work values increased slightly from 43% in 1981, 45% in 1990 to 50% in 1999. In France, Spain, and the United States the rise over the last 20 years was modest. A much sharper rise is notable in Great Britain, Italy, the two Irelands, and Iceland, and most of these are also countries with a higher than average level of support for extrinsic work values at the end of the previous century. There has been, on the
15
Cronbach’s alpha is .67 for extrinsic and .72 for intrinsic work values.
eroding work values?
99
Table 4.2: Extrinsic work values Extrinsic values
1981
Means Diff. Diff. Std. Dev. 1990 1999 99–81 99–90 1981 1990 1999
Belgium Denmark France (West) Germany Great Britain Northern Ireland Iceland Ireland Italy Netherlands Spain Sweden United States
.43 .39 .31 .56 .40 .45 .41 .43 .38 .40 .52 .45 .56
.40 .34 .27 .51 .42 .45 .52 .47 .44 .46 .50 .57 .56
.43 .33 .36 .51 .54 .66 .51 .62 .64 .40 .59 .42 .61
Total
.45
.46
.51
Mean all countries .43 Std. dev. all .07 countries
.45 .08
.50 .11
.00 –.06 .05 –.04 .14 .21 .10 .20 .26 –.01 .06 –.03 .05
.03 –.01 .09 .00 .11 .21 –.01 .15 .20 –.06 .08 –.15 .05
.32 .30 .25 .32 .31 .30 .28 .30 .27 .32 .33 .29 .29
.30 .27 .24 .31 .29 .29 .28 .30 .29 .32 .31 .32 .29
.32 .26 .27 .30 .30 .30 .29 .32 .30 .30 .30 .31 .28
.31
.31
.31
contrary, a slight decline of extrinsic work values in Denmark, Germany, and Sweden, and hardly any change in Belgium and the Netherlands. The pattern of change over time is not simple and straightforward. In Ireland and Italy change is linear with a substantial rise in the adherence to extrinsic work values between 1980 and 1990 and even more so between 1990 and 1999. If we look at Great Britain and Northern Ireland, this rise is typical only for the last decade of the 20th century, a period of rapid reform in these countries. In Iceland, on the contrary, the increase in support for extrinsic work values was most marked in the 1980s. After that decade there has been no increase. The Netherlands is comparable to Sweden: the level of support rose in the 1980s and declined again in the 1990s. France and Belgium show a very modest decrease in the 1980s and an increase in the 1990s, and Denmark and Germany show a modest decline in the support for extrinsic work values in the earlier decade and no change thereafter. We may conclude that at the end of the 20th century levels of support for extrinsic work values, with some exceptions, are quite comparable in Europe, but
100
peter ester, michael braun, and henk vinken
that the developmental pathways to these more or less equal levels reflect diversity.16 This conclusion is supported by the last figures in Table 4.2, presenting the standard deviations of the 13 countries, part of all three waves of EVS. These standard deviations refer to the country level (i.e. the standard deviation of the country means; the other standard deviations refer to data at the individual level within countries or, as the totals show, in all countries). These figures tell us most directly whether or not we are dealing with divergence or convergence between countries concerning value support. As can be observed, hardly any change towards overall divergence or convergence took place in the 1980s (deviation score developed from 0.07 to 0.08). In the 1990s, there is a modest trend towards divergence in extrinsic work values (from 0.08 to 0.11). This trend is probably mostly due to the more explicit changes taking place in Great Britain, Italy, and the Irish countries. The trend is so modest that for most of Europe we can conclude that there is neither divergence nor convergence of extrinsic work values. Intrinsic work values We now turn to trends in intrinsic work values. Intrinsic work values were described above as values that emphasize non-material or postmaterialist job characteristics such as the possibility of personal development, achievement, and autonomy, having a say in the work organization, being able to take initiative, and having an interesting, responsible, and challenging job. These values express an openness to change, the pursuit of autonomy, growth, creative and interesting work. In the EVS study we could clearly replicate this dimension of intrinsic work values. For all countries taken together, intrinsic work values have increased more clearly in support in the last two decades, in particular between 1980 and 1990 (Table 4.3). On average 57% of the respective work values are endorsed at the turn of the millennium. Country differences can be observed as far as the level and development of support for intrinsic work values go. In most countries, among them Belgium,
16
Cf. De Witte et al. (2004) for a similar conclusion.
eroding work values?
101
Table 4.3: Intrinsic work values Intrinsic values
1981
Means Diff. Diff. Std. Dev. 1990 1999 99–81 99–90 1981 1990 1999
Belgium Denmark France West Germany Great Britain Northern Ireland Iceland Ireland Italy Netherlands Spain Sweden United States
.35 .48 .37 .64 .55 .43 .49 .44 .39 .38 .41 .48 .64
.42 .52 .47 .64 .52 .49 .64 .54 .50 .57 .39 .72 .62
.50 .54 .52 .57 .49 .60 .65 .64 .69 .53 .48 .58 .69
Total
.48
.52
.57
Mean all countries .45 Std. dev. all .09 countries
.54 .09
.57 .07
.16 .06 .15 –.07 –.06 .17 .16 .20 .29 .15 .07 .10 .05
.08 .02 .05 –.06 –.03 .11 .01 .10 .19 –.05 .10 –.14 .07
.35 .34 .30 .32 .32 .33 .29 .32 .30 .33 .35 .33 .34
.34 .30 .31 .32 .32 .30 .30 .32 .32 .33 .34 .30 .33
.34 .31 .33 .33 .32 .36 .32 .34 .32 .32 .35 .29 .28
.34
.34
.34
France, Italy, the Irelands, Iceland, and the Netherlands, intrinsic work values have become much more prominent at the end of the 20th century. This development, but much less explicit, can also be found in Denmark, Spain, Sweden, and the United States. In Great Britain and Germany, on the contrary, support for intrinsic work values has declined somewhat. In most of the countries where support for intrinsic work values has risen, change has been more or less continuous over the last two decades. In Iceland, however, change occurred in the first decade, only. In the Netherlands and Sweden, the number of adherents of intrinsic work values has risen between 1980 and 1990, but has dropped again between 1990 and 2000. After a stable 1980–1990 period the proportions of West Germans adhering to these values has dropped as well in the 1990–2000 period. Finally, in Spain and the United States, support fell down slightly during the first period and rose again in the second period above the 1980 level. One might argue that at the end of the 20th century people from Iceland, Italy, Ireland, and the United States are the publics most favorable of intrinsic work values. Hardly any
102
peter ester, michael braun, and henk vinken
convergence can be discerned (see the modest change in the overall standard deviations of .09 in 1981 and 1990 to .07 in 1999). There are country differences but they are neither really diminishing nor increasing (cf. De Witte et al., 2004). A generational divide? We argued that young Europeans in particular are prompted to shape the future knowledge economy (Vinken et al., 2002). In this section we will analyze generational divides on the two main workvalue dimensions. We will do so in two steps. First of all, generational divides at the level of mean scores are depicted and analyzed for the three waves of the EVS study. In a next step we further examine generational differences and provide some detailed insights in the generational case by comparing other variables that might be relevant, for instance having work or not, and gender. Generation studies across Europe argue that cohorts whose members have, in their formative years, experienced the Second World War in the 1940s and its aftermath in the 1950s (war generation), the cultural and political upheaval of the 1960s and 1970s (baby boom generation), and the severe economic crisis and pervasive political shifts of the 1980s and 1990s (baby bust generation) form distinctive generations (see, e.g., Arber & Attias-Donfut, 2000; Diepstraten et al., 1999; Edmunds & Turner, 2002; Kohli & Szydlik, 2000; Van den Broek 1996). This three-type generational categorization is used here, as it is in the other chapters of this part of the book. Generational data of the three EVS waves on extrinsic work values, at first glance, present hardly any serious generational divide (Table 4.4). For Europe as a whole, the youngest generation, born in 1960 or later, does not seem very distinctive from older generations in terms of their support for extrinsic work values. In 1981, it is the middle generation, the boom generation, that is more different from the oldest in terms of support for extrinsic values (but only slightly so), while in 1990 and 1999 the middle generation is inbetween, with (modestly) increasing differences both between the war and the bust and the boom and the bust generation over the waves. The positioning of the three generations in the single countries does not show a general trend, but it is useful to highlight some observations. In 1981, the youngest generation, the baby bust gen-
.46 .46
.45 .44
W
.45 .43
.41 .27 .25 .46 .40 .42 .54 .48 .45 .44 .50 .58 .55 .46 .45
.39 .36 .28 .53 .44 .48 .48 .45 .45 .45 .49 .54 .56 .48 .47
.42 .40 .30 .57 .45 .45 .55 .48 .44 .50 .52 .59 .58 .46 .45
.40 .34 .27 .51 .42 .45 .52 .47 .44 .46 .50 .57 .56 Y
N W W W Y N N N N N N N N .49 .47
.42 .30 .32 .48 .45 .61 .47 .64 .63 .34 .55 .40 .55
War
.51 .50
.43 .34 .34 .51 .57 .66 .51 .64 .62 .40 .59 .42 .61 .53 .52
.44 .34 .41 .55 .56 .69 .53 .60 .66 .42 .60 .43 .62 .51 .50
.43 .33 .36 .51 .54 .66 .51 .62 .64 .40 .59 .42 .61
Y
N N Y W W W N N N W W N W
EVS wave 1999 Boom Bust Total Diff.
Diff. = does the youngest generation (baby bust) differ from the older generations? y = yes, differs from both older generations; w = differs from war generation, but not from baby boom generation; b = differs from baby boom generation, but not from war generation; n = no difference between generations; differences are analyzed with one-way anova (post-hoc multiple comparisons, Tukey HSD test)
War = war generation, born before 1940; Boom = baby boom generation, born 1940–1960; Bust = baby bust generation, born after 1960
.46 .46
N W W N N N N N W N N W N
.44 .42
.43 .39 .31 .56 .40 .45 .41 .43 .38 .40 .52 .45 .56
Total Mean 12 countries
.45 .45 .34 .58 .40 .51 .41 .41 .33 .42 .56 .52 .55
.41 .35 .28 .53 .38 .40 .39 .42 .41 .38 .51 .41 .55
Belgium Denmark France West Germany Great Britain Northern Ireland Iceland Ireland Italy Netherlands Spain Sweden United States
.46 .42 .33 .58 .42 .49 .43 .44 .36 .41 .53 .47 .57
EVS wave 1981 EVS wave 1990 War Boom Bust Total Diff. War Boom Bust Total Diff.
Extrinsic values %
Table 4.4: Extrinsic work values by generation and wave
eroding work values? 103
104
peter ester, michael braun, and henk vinken
eration, is significantly different in extrinsic work-values adherence compared to the oldest generation (war generation). In Denmark, France, and Sweden its members put more emphasis on these values.17 In Italy, however, it is the other way round. The boom generation, in these countries, is partly located in the middle of the other two, and partly closer to the bust generation. In all other countries in that year, generations support extrinsic work values to roughly the same extent. Looking at 1990, the Swedish and Italian generational differentiations of 1981 have blurred, but the one in Denmark has even expanded. Also, both younger generations in West Germany and Great Britain now have a somewhat more distinctive profile, putting more emphasis on extrinsic work values than the oldest generation does. The 1999 wave shows again a different pattern. In France, the only country indicating generational variations for all three waves, the bust generation differs from both the war generation and the boom generation. In Germany and Great Britain, the differentiation which emerged in 1990 between the bust and the war generation persists. In Northern Ireland, the Netherlands, Spain, and the United States differences in extrinsic work values between both of these extreme generations emerge for the first time, while the differences observed for Denmark for both prior waves have blurred. In conclusion one might argue that over the last two decades of the 20th century generational distinctiveness has (somewhat) grown in Europe with a younger generation increasingly divergent from older ones in terms of its higher support for extrinsic work values. Do we find this pattern as well when we turn to intrinsic work values? Table 4.5 shows the details. Over all countries, the bust generation is pitted against the war generation only, in 1981, but against both older generations for the more recent waves. In 1981, we observe significant divides across generations in Denmark, Ireland, the Netherlands, and Spain only. In all of these countries, the bust generation is different from the war generation, but not the boom generation. In 1990, the popularity of intrinsic work values seems to be on a general rise. The older generations catch up with the young. Nevertheless, the num-
17 In Northern Ireland the difference appears to be impressive, too, but not statistically significant due to the lower number of cases in this country.
.35 .48 .37 .64 .55 .43 .49 .44 .39 .38 .41 .48 .64 .48 .45
N W N N N N N W N W W N N W
.39 .42 .44 .58 .50 .47 .58 .52 .45 .53 .35 .72 .61 .49 .50
.42 .55 .49 .67 .54 .48 .65 .51 .50 .55 .38 .72 .63 .53 .54
.47 .61 .49 .67 .54 .52 .68 .60 .53 .66 .43 .73 .61 .56 .58
.42 .52 .47 .63 .52 .49 .64 .54 .50 .57 .39 .72 .62 .52 .54
Y Y N W N N W Y W Y Y N N Y
.50 .44 .46 .54 .45 .54 .59 .61 .66 .45 .44 .54 .75 .53 .52
War .47 .55 .50 .57 .52 .61 .65 .67 .66 .49 .48 .60 .69 .57 .56
.53 .59 .57 .60 .50 .62 .67 .64 .72 .60 .51 .59 .66 .60 .60
.50 .54 .52 .57 .49 .60 .65 .64 .69 .53 .48 .58 .69 .57 .57
B W Y N N W W N Y Y W N W Y
EVS wave 1999 Boom Bust Total Diff.
Diff. = does the youngest generation (baby bust) differ from the older generations? y = yes, differs from both older generations; w = differs from war generation but not from baby boom generation; b = differs from baby boom generation but not from war generation; n = no difference between generations; differences are analyzed with one-way anova (post-hoc multiple comparisons, Tukey HSD test)
War = war generation, born before 1940; Boom = baby boom generation, born 1940–1960; Bust = baby bust generation, born after 1960
.39 .56 .38 .67 .57 .36 .47 .51 .39 .44 .48 .54 .64 .51 .49
.32 .41 .35 .61 .51 .43 .45 .40 .39 .33 .38 .45 .59 .44 .42
Belgium Denmark France West Germany Great Britain Northern Ireland Iceland Ireland Italy Netherlands Spain Sweden United States Total Mean 12 countries
.37 .53 .38 .66 .58 .45 .53 .46 .40 .40 .44 .52 .68 .50 .48
EVS wave 1981 EVS wave 1990 War Boom Bust Total Diff. War Boom Bust Total Diff.
Intrinsic values %
Table 4.5: Intrinsic work values by generation and wave
eroding work values? 105
106
peter ester, michael braun, and henk vinken
ber of countries showing generational distinctions increases, and variations also emerge between the bust and the boom generation. In detail: differences between the bust and only the war generation emerge in three countries that did not show significant divides in 1981: West Germany, Iceland, and Italy. The boom generation is partly in the middle between the war and the bust generation and partly close to the latter. All countries that in 1981 showed differences between the bust and the war generation, i.e. Denmark, Ireland, the Netherlands, and Spain, have extended these differences to the boom generation. The same pattern applies to Belgium. In 1999, support for intrinsic values has risen again, and the differences between generations decline in some countries, while new divides emerge in others. The bust generation now differs from both the war and the boom generation in France (which showed no generational divide both in 1981 and 1990), Italy, and the Netherlands. The fullest generational divide therefore is found in these three countries at the end of the 20th century, in other countries the youngest generation conflicts with particular generations only. In Denmark, Northern Ireland, Iceland, Spain, and the United States the bust generation is pitted against the war generation, while in Belgium the bust generation is significantly different from the boom generation only. Generational differences in the United States, however, are very special indeed. While in both 1981 and 1990, this country did not show any generational divide, in 1999 we find the war generation to put most emphasis on intrinsic values, while all over Europe (when generational divides are found) it is the youngest generation who does so. In 1999, there are four European countries which do not show any generational divide with regard to intrinsic values: Great Britain and Sweden did not show generational contrasts in 1981 and 1990 either; in West Germany the bust generation distinguished itself from the war generation in 1990, and Ireland showed generational differences both in 1981 and 1990. We can conclude that a generational divide concerning intrinsic work values is eminent with the youngest generations in most European countries over-emphasizing these values as compared to older generations throughout these two decades. With the exception of some countries showing disappearing distinctions and others that do not show any distinctions, the generational divide persists over time in most European countries. The relatively strongest generational difference is found in France, Italy, and the Netherlands. In the United States,
eroding work values?
107
on the contrary, where there were no generational distinctions in 1981 and 1990, in 1999 it is the war generation, surprisingly, that is the true supporter of intrinsic work values. The relative impact of generational divides It is important to further examine the generational divides and provide some detailed insights in the generational case by comparing with other variables that might be relevant, for instance having work or not, and of course, gender. One might argue that the findings mentioned above do not allow for final conclusions regarding the extent to which the value diversities observed are truly generational. It is clear that at a single moment in time it is hard if not impossible to discern generational effects from age or life-cycle effects, and, moreover, to identify period effects taking place at that moment in time as well. Only after identifying these three effects it is possible to draw firm conclusions on the existence of generational divides in Europe. This is why we have conducted a final analysis aimed at addressing the generational effect that might be at work. The analyses include a basic proxy for life-cycle effects, being the employment status, and other controls.18 One might argue that having a job is an outcome of age progress in the life cycle and can thus be regarded as a proxy of life-cycle effects (see De Graaf & De Graaf, 1988). We have included the research year as a proxy for period effects, in this case we compare the state of affairs at the 1990 and 1999 periods with the 1981 period. In a first step we enter the life-cycle and period effects and some additional controls, after which we include the three generations discussed above as a three-type
18 The key issue is to technically discern age from cohort effects in time series analysis. Incorporating age as a variable in the analysis and cohorts—also based on age—as well, leads to serious problems of multicollinearity. More substantially: the aim is to assess effects of being in a certain phase in the life cycle and not being in another (age effects) and to discerns these effects from having had specific and determining formative experiences in these life-course phases (cohort effects). Age itself is of course only a rather crude indicator for these age effects. A better indicator is one that taps into the transitions taking place in the life course. One of those transitions (among many others) includes becoming part of the work force (versus not being part yet or not being anymore). See for more details and attempts to overcome the identification problem Van den Broek (1996) and Diepstraten et al. (1999).
108
peter ester, michael braun, and henk vinken
differentiation that allows us to address the distinction of both the war and the baby boom generation as compared to the baby bust generation (reference group). Findings are shown in Table 4.6. Extrinsic work values are especially favored by the Americans, with all European countries, except Northern Ireland, showing significantly lower support. Among the Europeans, the French, British, Dutch, Belgians, and Danish display particularly low adherence. Women tend to endorse extrinsic work values less than men, and both the unemployed and the self-employed less than people in fullor part-time employment. Religion does not play a role in this context. Endorsement for extrinsic values has slightly, but significantly, risen from the early 1980s to the early 1990s, and more strongly during the last decade of the last millennium. After adding the generation dummies, the effect for the 1990 period diminishes. All other effects remain significant. The war generation and, to a lower extent also the baby boom generation are less in favor of extrinsic work values than the bust generation, regardless of their employment or any other status considered here. In short, living at the turn of the millennium, being American, being male, being employed and being a young generation member is conducive to instilling extrinsic work values in a person. Intrinsic work values are again especially favored by the Americans, this time with no exception among the European countries. Among the Europeans, the French, Spanish, Dutch, Belgians, and Danish show particularly low support. Women now also adhere to intrinsic work values less than men. Intrinsic work values are found significantly more among employed persons and, in contrast to the corresponding finding with regard to extrinsic values, also among the selfemployed. Religion does play a role in this context, with Protestants stressing intrinsic values more than other denominations. There are also strong effects both of the years 1990 and 1999, proxies of a period effect: in both years intrinsic work values are more stressed than in 1981. The relative size of the effects suggests that the increase of intrinsic work values is gradual and continuing. If the higher level of emphasis on intrinsic work values in 1999 is generational, then the period effects (waves 1990 and 1999) should diminish if not disappear in favor of generation membership. This does not happen when we introduce generations in the equation. The period effects remain nearly as strong and as significant even after adding gener-
eroding work values?
109
Table 4.6: Regression of intrinsic and extrinsic work values on country dummies, sex, employment status, study waves, and generations Extrinsic values France Great Britain West Germany Italy Spain Netherlands Belgium Denmark Sweden Ireland Northern Ireland Iceland women employed self-employed Protestant W1990 W2000 war generation boom generation Adjusted R square N
–.21* –.10* –.04* –.07* –.03* –.12* –.16* –.16* –.08* –.05* –.01 –.07* –.02* .04* –.05* .00 .02* .11*
Intrinsic values
6.8%
–.21* –.10* –.04* –.07* –.03* –.12* –.16* –.17* –.08* –.05* –.01 –.07* –.02* .03* –.05* .01 .01 .09* –.05* –.02* 6.9%
–.14* –.09* –.02* –.09* –.21* –.11* –.19* –.10* –.05* –.07* –.06* –.06* –.04* .07* .03* .03* .08* .16* 7.1%
–.14* –.10* –.02* –.09* –.22* –.11* –.19* –.10* –.05* –.07* –.06* –.06* –.04* .05* .03* .04* .07* .14* –.09* –.04* 7.5%
52447
52447
52431
52431
Baseline categories: United States, men, non-employed, non-Protestant religions, wave 1981, bust generation; * = significant on the .05 level.
ation effects.19 The generation effects itself are relatively strong with the youngest generation (across countries and time and regardless of gender, employment or religious status) clearly more in favor of intrinsic work values. In short, living at the turn of the millennium, being American, being male, employed or self-employed, being a Protestant, and being a member of the younger generation is conducive for the adherence to intrinsic work values.
19 Only the employment variables diminish somewhat in strength when we include the generation dummies, indicating that having employment is only partly a feature of the younger generation itself.
110
peter ester, michael braun, and henk vinken Conclusions
In the discourse on the desired direction of European modernization, counterbalancing the increased competition from the United States and Asia in a globalizing world, a work force is called for that no longer merely embraces conventional extrinsic work values but that is inspired by intrinsic work values as well. In this chapter we focused on analyzing trends in these work values. In drawing the main conclusions of our analyses of the EVS data we concentrate on the key value shifts taking place in Europe, compared to Europe’s main competitor: the United States, on the question of convergence or divergence, and finally on generational divides. Extrinsic work values are slightly more popular in Europe at the end of the 20th century compared to the early 1980s. Support for intrinsic work values has also risen—even more clearly so than support for extrinsic work values—in the last ten years of the previous century. There is no strong evidence of convergence or divergence regarding extrinsic and intrinsic work values in Europe. Diversity in Europe is strong. Young people in particular are assumed to take a leading role in advancing the knowledge society. Do younger European generations have work values that support taking such a role? It seems that generational divides have somewhat grown over the last two decades of the 20th century with the younger generations more strongly supporting both extrinsic and intrinsic work values than the war generation (born before 1940) does. These results indicate that for Europe to become the most dynamic and competitive economy, there is still a world to win. The comforting message is that the rise in intrinsic work values is sharpest in Europe (also compared to the United States) and that the youngest generation in most of Europe seems to be preparing for this shift, at least in their cultural framing: they are the ones strongly emphasizing intrinsic work values. It is interesting to note that they not only emphasize intrinsic, but also extrinsic work values more. Framed positively one could argue that probably this generation is well aware of the increasingly louder demands to invest in life-long personal development, to take more initiative by themselves, and to be personally more responsible for success or failure, but at the same time counteracts with demanding well-formatted and personally geared conditions at the work place. This trade-off logic, which in a sense
eroding work values?
111
is entrepreneurial and as such in line with the expectations for the modern employee as we stated in the introduction, was explicitly established in earlier research among Dutch younger generations (see Ester & Vinken, 2001) and could well apply to other young generations across Europe and the United States. Importantly, while the Americans are still topping the Europeans with regard to both extrinsic and intrinsic work values, the younger generations in the United States do not tend to stress intrinsic work values more (in the early 1990s) or even clearly less (at the end of the millennium) than the older generations in this country. The small increase in the endorsement of intrinsic work values in the United States is mainly driven by a sharp increase of war generation Americans supporting these values. It is still unclear whether the United States can be regarded as a forerunner of worldwide trends in this respect too: might it be that the economic future of a country will reside less in the innovation potential of the young, but in that of the older generations, which will outnumber the young ones in the long run anyhow, and on whose life-long learning and substantially postponed age of retirement the competitiveness of nations might reside in the future? A fascinating hypothesis that would radically challenge and reframe current fashionable policy thoughts on the relationship between generational renewal and economic innovation. References Alderfer, C.P. 1972. Existence, Relatedness, and Growth. Human Needs in Organizational Settings. New York: Free Press. Arber, S. & C. Attias-Donfut 2000. The Myth of Generational Conflict: the Family and State in Ageing Societies. London: Routledge. Arts, W.A. & L. Halman (eds.) 2004. European Values at the Turn of the Millennium. Leiden & Boston: Brill. Arts, W.A., J. Hagenaars & L. Halman (eds.) 2003. The Cultural Diversity of European Unity. Findings, explanations and reflections from the European Values Study. Leiden & Boston: Brill. Borg, I. 1986. A cross-cultural replication on Elizur’s facets of work values. Multivariate Behavioral Research, 21, 401–410. Borg, I. & M. Braun 1996. Work values in East and West Germany: different weights but identical structures. Journal of Organizational Behavior, 17, 541–555. Borg, I. & A.P. Brief 1994. Correlations and dependencies among work values. In I. Borg & S. Dolan (eds.), Proceedings of the Fourth International Conference on Work and Organizational Values. Barcelona. Braun, M. 1994. The International Social Survey Program (ISSP). In P. Flora et al. (eds.), Social Statistics and Social Reporting in and for Europe. Bonn: Informationszentrum Sozialwissenschaften.
112
peter ester, michael braun, and henk vinken
Braun, M. & I. Borg 2004. Berufswerte im zeitlichen und im Ost-West-Vergleich. In R. Schmitt-Beck, M. Wasmer & A. Koch (eds.), Sozialer und politischer Wandel in Deutschland. Wiesbaden: VS Verlag für Sozialwissenschaften. De Graaf, N.D. & P.M. De Graaf 1988. De samenhang tussen leeftijd en postmaterialistische waardenpatronen. Een inhoudelijke uiteenlegging in socialisatieen levensloopeffecten. Sociologische Gids, 35, 397–417. De Witte, H., L. Halman & J. Gelissen 2004. European work orientations at the end of the twentieth century. In W.A. Arts & L. Halman (eds.), European Values at the Turn of the Millennium. Leiden & Boston: Brill. Diepstraten, I., P. Ester & H. Vinken 1998. Mijn generatie. Zelfbeelden, jeugdervaringen en lotgevallen van generaties in de twintigste eeuw. Tilburg: Syntax Publishers. Diepstraten, I., P. Ester & H. Vinken 1999. Talkin’ bout my generation. The Netherlands’ Journal of Social Sciences, 35, 91–109. Edmunds, J. & B.S. Turner 2002. Generations, Culture and Society. Buckingham: Open University Press. Elizur, D. 1984. Facets of work values. Journal of Applied Psychology, 69, 379–389. Elizur, D., I. Borg, R. Hunt & I. Magyari-Beck 1991. The structure of work values: a cross-cultural comparison. Journal of Organizational Behavior, 12, 21–38. England, G.W. 1991. The meaning of working in the USA: Recent changes. The European Work and Organizational Psychologist, 1, 111–124. England, G.W. & S. Ruiz Quintanilla 1994. Work meanings: their structure and stability. Paper presented at the conference “Values and Work”, WORC, Tilburg University. Ester, P. & H. Vinken 2001. Een dubbel vooruitzicht. Doembeelden en droombeelden van arbeid, zorg en vrijd tijd in de 21e eeuw. Bussum: Coutinho. Ester, P., R. Muffels, & J. Schippers (eds.) 2001. Flexibilisering, organisatie en employability. Bussum: Coutinho. Ester, P., A. Román, & H. Vinken 2004. Work values on the old continent. Crossnational and intergenerational trends in work orientations in 12 European countries: 1980–2000. Paper prepared for the 99th Annual Meeting of the American Sociological Association. Public Sociologies. San Francisco, August 14–17, 2004. Halman, L. 1999. Europeans and their work orientations. WORC: paper 99.04.005/1. Tilburg: Tilburg University. Herzberg, F., B. Mausner & B. Snyderman 1958. The Motivation to Work. New York: Wiley. Hofstede, G. 2001. Culture’s Consequences. Comparing values, behaviors, institutions and organizations across nations. Thousand Oaks: Sage Publications. Katz, D. & R. Kahn 1978. The Social Psychology of Organizations. New York: Wiley. Kohli, M. & M. Szydlik (Hrsg.) 2000. Generationen in Familie und Gesellschaft. Opladen: Leske + Budrich. Levy, S. & L. Guttman 1981. Structure and level for rewards and allocation criteria in several life areas. In I. Borg (ed.), Multidimensional Data Representations: When and Why. Ann Arbor, MI: Mathesis Press. Maslow, A. 1954. Motivation and Personality. New York: Harper und Row. Roe, R.A. & P. Ester 1999. Values and Work. Special Issue of Applied Psychology: an International Review (Vol. 48). Ros, M., S.H. Schwartz & S. Surkiss 1999. Basic individual values, work values, and the meaning of work. In R.A. Roe & P. Ester, Values and Work. Special Issue of Applied Psychology: An International Review (Vol. 48). Rosenberg, M.J. 1957. Occupations and Values. Glencoe, III.: The Free Press. Schwartz, S.H. 1994. Beyond individualism/collectivism. In U. Kim et al. (eds.), Individualism and Collectivism. Thousand Oaks: Sage Publications. Super, D.E. & B. ”verko (eds.) 1995. Life Roles, Values, and Careers. International findings of the Work Importance Study. San Francisco: Jossey-Bass, Inc.
eroding work values?
113
Van den Broek, A. 1996. Politics and Generations. Cohort replacement and generation formation in political culture in the Netherlands. Tilburg: Tilburg University Press. Van den Elzen, A. 2002. Zekerheid of zelfontplooiing? Een internationaal vergelijkend onderzoek naar arbeids- en gezinswaarden. Serie: Sociaal-culturele ontwikkelingen in Nederland. Nijmegen: Katholieke Universiteit Nijmegen. Vinken, H., J. Soeters & P. Ester (eds.) 2004. Comparing Cultures. Dimensions of Culture in a Comparative Perspective. Leiden & Boston: Brill. Vinken, H., P. Ester, H. Dekkers & L. Van Dun 2002. Aan ons de toekomst. Toekomstverwachtingen van jongeren in Nederland. Assen: Van Gorcum. Zanders, H. 1987. Opvattingen over arbeid in Nederland en Europa. In: L. Halman, F. Heunks, R. de Moor & H. Zanders. Traditie, secularisering en individualisering. Een studie naar de waarden van de Nederlanders in een Europese context. Tilburg: Tilburg University Press. Zanders, H. 1993. Changing work values in Europe and North America. Continents and occupations compared. In P. Ester, L. Halman & R. de Moor (eds.), The Individualizing Society. Value change in Europe and North-America. Tilburg: Tilburg University Press.
CHAPTER FIVE
THE END OF TRADITIONAL POLITICAL VALUES? Oddbjørn Knutsen Introduction In European democracies, the most important value conflicts have originated in the most crucial socio-political cleavages. These cleavages are outlined in a developmental perspective in the famous work of Seymour Martin Lipset and Stein Rokkan (1967). The major equivalent value conflicts can be linked to these cleavages in the following way: Of the conflicts fostered by the National Revolution,1 those related to religious versus more secular values created the most uniform divisions in West European party systems. Christian values focus on the importance of Christian morals and principles in society and politics, and on traditional moral guidelines in school and society in general. The Industrial Revolution gave rise to economic interest conflicts that were anchored in hierarchical socio-structural variables. The most important political value orientations that have emerged from the Industrial Revolution were economic left-right values or left-right materialist values. These value orientations are economic in nature, and they refer in particular to the role of government in creating more economic equality in society versus the need for economic incentives and efficiency. These value orientations incorporate value conflicts related to control, power, and the degree of distribution of resources in the production sphere, and include workers’ control and
1
Lipset & Rokkan associated the National Revolution with cultural standardization on the model of a nation-wide code with standard education and assimilation of minorities. The major conflicts, which arose from the National Revolution, were the conflict between the central nation-building culture and the increasing resistance of the ethnically, linguistically, and religiously distinct subject population in the provinces and the peripheries (the center-periphery conflict); and the conflict between the centralizing, standardizing, and mobilizing nation state and the historically established corporate privileges of the church (the religious/secular conflict).
116
oddbjørn knutsen
state regulation of the economy versus private enterprise, private property, and the market economy; economic and social equality versus the need for differentiated rewards for stimulating effort (Inglehart, 1984: 25; Knutsen, 1995a; Lafferty & Knutsen, 1984). The moral value dimension and economic left-right values are often referred to as ‘Old Politics’ because they capture the essence of the traditional lines of conflict in industrial society. In contrast, ‘New Politics’ refers to value conflicts emerging from post-industrial society. There are different ways of conceptualizing the new politics value orientations. According to Ronald Inglehart, value conflicts related to materialist/postmaterialist value orientations (hereafter MPM) reflect the new politics conflict dimension. Inglehart argues that ‘new’ postmaterialist values are deeply rooted and stand in opposition to more traditional materialist values. He identifies a ‘silent revolution’, in which a gradual value change along the materialist/postmaterialist dimension takes place. This involves a shift from a preoccupation with physical sustenance and safety values towards a greater emphasis on belonging, self-expression, and quality-of-life values. The spread of postmaterialist values is basically explained by greater economic and physical security in society and then by generational replacement, the growth of the new middle class, and the spread of higher education (Inglehart, 1977, 1990). Scott Flanagan has emphasized in a series of articles that Inglehart’s conceptualization of value change combines two dimensions: a materialist/non-materialist dimension and a libertarian/authoritarian dimension (Flanagan, 1987; Flanagan & Lee, 1988, 2003). The overarching concept that integrates libertarian values is self-actualization, and the central value orientations within the notion of libertarian values are autonomy, openness, and self-betterment. The authoritarian value orientation “designates a broader cluster of values, which, along with concerns for security and order, includes respect for authority, discipline and dutifulness, patriotism and intolerance for minorities, conformity to customs, and support for traditional religious and moral values” (Flanagan, 1987: 1305). The libertarian/authoritarian value orientations are also a central component in Herbert Kitschelt’s (1994, 1995) important work on changes in the party systems of Western democracies and has also been found to be an important political value dimension in Cees Middendorp’s analyses of the Dutch political culture (Middendorp, 1978, 1991, 1993). Libertarian/authoritarian values are also related to the important conflict regarding
the end of traditional political values?
117
immigration and immigration policy in Western democracies (Knutsen & Kumlin, 2005).2 On the basis of the discussion above, I have identified the central political value dimensions, which are discussed in political science and political sociology. Religious/secular values are analyzed in the chapter by Halman and Pettersson. Therefore, the focus here will be on different aspects of the economic left-right orientations and on two sets of value orientations associated with New Politics, MPM, and libertarian/authoritarian values. The research questions in this chapter are as follows: (1) What are the trends in political value orientations over time, and how do political values vary between countries? (2) Is there a trend towards cross-national convergence in priorities of political values? (3) Are there generation differences in political values and are there smaller cross-national differences within the post-war cohort than within the war cohort? Hypotheses, perspectives and previous findings Introduction In this section, I will discuss some general perspectives on the distributions of values over time and in a comparative setting, and expectations of generation differences. I will not test the various perspectives in a strict way in the subsequent empirical analysis, but evaluate the findings according to these perspectives in a more general way. Economic left/right values Trends over time. The late 1970s and 1980s also saw a revival of conservative and liberal ideas, movements, and increased support for non-socialist political parties, which challenged the post-war Keynesian consensus on the welfare state, state intervention, and citizenship rights in the social domain—economic leftist values which had been implemented in most West European politics. This right-wing ideology
2 Another way of conceptualizing ‘New Politics’ is represented by environmental versus economic growth values. Indicators for such values are only available in EVS II and EVS III and are therefore dropped from the analysis in this chapter.
118
oddbjørn knutsen
underscored the superiority of market mechanisms as promoting both economic prosperity and the maximization of individual freedom. It also claimed that state intervention did not work, and accorded a minimal role to the state in the operation of the economy and the social order (King, 1987; Levitas, 1986). In many countries, nonsocialist governments with a neo-liberal agenda came to power.3 The value change, which underpinned these rightist trends was a move towards economic rightist values, which could be documented in many countries. However, during the 1980s, there was a trend towards leftist values and in many countries, the mass public’s values were as leftist as they had been before the shift towards the right around 1980 (Knutsen, 1995a: 173–182). In the 1990s, many social democratic governments came to power but with more centrist programs than in earlier time periods. I do not have any clear expectations about how the economic left-right values changed from the early 1980s to around 2000. There are two different perspectives on comparative differences in socio-political value orientations incorporated in the economic leftright orientations. The first one is that values will vary according to the welfare regime which exists in the given country. The point of departure is the well-known distinction between the universal or social-democratic, the conservative or Christian democratic, and the liberal or residual welfare state. Welfare-state attitudes are—according to this perspective—formed by the character of the institutional regimes of social policy. Universal welfare states are underpinned by strong popular support for the welfare state because most social groups are included as recipients. This is not the case with the liberal welfare state, where the recipients are the lower social classes. Support for these hypotheses is rather mixed (Arts & Gelissen, 2001; Bean & Papadakis, 1998; Svallfors, 1997). Arts & Gelissen (2001: 292–297) find, for example, that the Southern Mediterranean welfare states (Greece, Italy, Spain, and Portugal) show the strongest support for welfare solidarity, followed by the countries in the socialdemocratic regime. The Southern Mediterranean welfare states have the highest level of support for the equality justice principle, whereas
3 Of the countries in this study, this applied to Belgium, Denmark, West Germany, and the Netherlands, in addition to the most famous cases of Great Britain and the United States.
the end of traditional political values?
119
the social democratic countries have the lowest.4 The various welfare regimes are also expected to create different social cleavages and conflicts according to this perspective (Esping-Andersen, 1990: Chap. 8; Iversen & Wren 1998; Svallfors, 1997: 284–287). The second perspective has a different point of departure but predicts that leftist orientations will be most supported in the least advanced Western societies, where the welfare state and the regulation of the economy is small, and where social inequalities are fairly large. People in these countries want change in a more equalitarian direction, while the mass public in more advanced welfare states live in a different economic environment, and the need for further development in leftist policies is not appreciated. In Ronald Inglehart’s conceptualization there is a diminishing marginal utility or return of classical economic leftist policies, such as economic regulation, income equality, and welfare-state efforts in advanced welfare states. In rich countries, which have developed advanced welfare states (with a high level of (progressive) taxation), further redistribution and state intervention seems less urgent and the costs of further government intervention impinge on a larger share of a considerable part of the population. Consequently, the support for further economic policies of the left becomes weak and indeed public resistance to such courses becomes strong (Inglehart, 1990: Chap. 8, 1999: 256–65). This view is supported by different comparative studies of attitudes towards income equality (Kelly & Evans, 1993: 112–115; Roller, 1995), attitudes towards state intervention (Borre & Viegas, 1995: 239–242), and general economic left-right orientations (Inglehart, 1990: 253–257, 1997: 260–65; Knutsen 1995a: 176–179). Generation differences. There are no clear predictions about how cohort will influence economic left-right orientations. These conflicting value
4 Arts and Gelissen (2001: 285–286) employ a five-category division of welfare states of which four are relevant here. Their division is similar to Esping-Andersen’s but they use a fourth category which is called a Southern Mediterranean welfare state that resembles the conservative type (see also Esping-Andersen, 1999: Chap. 5). Arts and Gelissen (2001: 290) also classify countries into the various categories. The countries included in this study, which grouped in the liberal welfare type, are the United States and—as hybrid cases—Great Britain (and Northern Ireland) and Ireland. West Germany and France are grouped into the conservative type. Denmark, Sweden, and—as hybrid cases—Belgium and the Netherlands are classified as socialdemocratic welfare states, while Spain and Italy are grouped as Southern Mediterranean welfare states. This classification is the basis for the discussion in this chapter.
120
oddbjørn knutsen
orientations were anchored in the industrial society and neither leftist nor rightist values are new. One hypothesis is that the generations, which grew up in the 1920s and 1930s (the war cohort) experienced social conflicts, poverty and failure of the market, and economic liberalism, and can therefore be expected to be the strongest purveyors of a leftist materialist value orientation. In the United States, these generations have been the strongest supporters of the so-called New Deal alignment (Nie, Verba & Petrocik 1976: 85–95; Beck 1974) and in many West European countries, the ‘economic depression’ generations have been strong adherents of Social Democracy and the political value orientation, which Social Democracy traditionally have been associated with, i.e. central leftist economic values. This has been documented in the Scandinavian countries in particular (Esping-Andersen, 1985: Chap. 4; Knutsen, 2003: Chap. 8 and 9). Another hypothesis is that the boom cohort is characterized by more leftward values than other cohorts. The radicalism of the younger cohorts in the 1960s and 1970s can be expected to incorporate a traditional leftward value orientation as well as a postmaterialist value orientation. This might have influenced this cohort permanently. By contrast, the bust cohort experienced the increased focus on economic liberalism which challenged the post-war consensus in formative years. This cohort might then be more inclined to support rightist economic values than the other cohorts. It is, however, difficult to have any expectations regarding whether the war or the boom cohorts should be the most leftist. Previous empirical comparative studies found only weak correlations between economic left-right values and age (Knutsen, 1995a: 182–188; Aalberg, 2003: 58–64). It is difficult to have any expectations about comparative variations regarding cohort differences. One hypothesis is that in welldeveloped welfare states (the universal and perhaps also the conservative type), the bust cohort in particular would support economic liberalist or rightist values since leftist welfare-state values represent the established order. In the less-developed welfare states in Southern Europe, such differences will not be pronounced, and the younger cohorts might even be more leftist than the war cohort. New Politics values Trends over time. According to New Politics theories, the ‘new’ libertarian and postmaterialist values should generally increase due to
the end of traditional political values?
121
increased economic prosperity and economic and physical security in advanced industrial democracies, and due to generational replacement because the younger post-war cohort had their formative experiences in a more secure environment as a result of prosperity, the absence of war experiences, and a safer physical environment. Period effects might, however, cause these new values to decline in periods with, for example, economic decline, higher unemployment, and more unstable and insecure labor markets (Inglehart, 1981; Abramson & Inglehart, 1995: Chap. 2–4). Cross-national variations in support of these values should be influenced by the same factors. The new values are assumed to be most widespread in countries with a high level of prosperity and perhaps also well-developed welfare states, where people are economically more secure than in less developed welfare states (Kitschelt, 1988). Generation differences. The main perspective in Inglehart’s thorough analysis of cohort differences in MPM value orientations is that these differences persist and contribute to long-term value change through generation replacement even though there are period effects, which induce the mass publics to become more materialist in given time periods due to short-termed effects caused by deteriorated economic circumstances. The scarcity hypothesis implies that short-termed effects may also induce all cohorts to emphasize postmaterialist values when economic conditions are good. The cohort differences are explained by different economic and physical security during the formative years of the various cohorts. This socialization hypothesis predicts a watershed between the post-war and the pre-war (war) cohorts in value priorities because the war cohort and the post-war cohorts have such different experiences in formative years regarding economic security (economic scarcity versus economic prosperity) and physical security (war/absence of war) (Inglehart, 1977: Chap. 2, 3, 1981, 1990: Chap. 2; Abramson & Inglehart, 1995: Chap. 2–4). Libertarian/authoritarian values have not been studied this systematically within a similar framework, but age differences are among the most consistent socio-demographic findings from available studies.5
5
Libertarian/authoritarian values are found to be fairly strongly correlated with age, see Middendorp, 1991: 158–175; Flanagan & Lee, 1988.
122
oddbjørn knutsen Indicators and index constructions
Political values are not represented with many indicators in the various European Values Studies (hereafter EVS).6 In particular in the first wave, few political values were included. I have only included indicators which were available in all three waves of EVS.7 Table 5.1 shows the two available indicators for the economic left-right orientations. Table 5.1: Indicators for economic left-right values Equality/freedom Which of the two statements comes closest to your own opinion? A. I find that both freedom and equality are important. But if I were to choose one or the other, I would consider personal freedom more important; that is, everyone can live in freedom and develop without hindrance. B. Certainly both freedom and equality are important. But if I were to choose one or the other, I would consider equality more important; that is, that nobody is underprivileged and that social class differences are not so strong. Response alternatives were: 1 Agree with statement A. 2 Agree with statement B. 3 Neither. Pay equality Imagine two secretaries of the same age, doing practically the same job. One finds out that the other earns more a week more than she does. The better paid secretary, however, is quicker, more efficient and more reliable at her job. In your opinion is it fair or not fair than one secretary is paid more than the other?
6 In order to cover the various orientations discussed in the previous section, some of the indicators I use do not belong clearly to the political domain, but they comprise a dimension which is highly politically relevant. This applies to the libertarian/authoritarian indicators as well as to the pay equality indicators for economic left-right orientations. 7 Additional economic left-right indicators gathering views on business management were available in EVS I and EVS II, while indicators tapping attitudes towards economic equality, nationalization/privatization, and the state’s responsibility for the welfare of the citizens were available in EVS II and EVS III.
the end of traditional political values?
123
The two items on the economic left-right orientations8 are fairly different and it is therefore difficult to construct a single measure on the basis of them. The question on Equality/freedom comprises a central value conflict regarding the industrial left-right conflict. The leftist argument for equality stresses the value of economic and social equality, while the rightist argument for freedom stresses personal freedom and does not mention economic freedom. This might be considered inconsistent, but the argument in favor of freedom is frequently that economic freedom and personal freedom are strongly linked to each other. The second item covers an important type of pay equality, namely whether persons who have the same job should get the same reward economically or whether economic rewards should represent diffferentiations according to efforts. It should be emphasized that this type of economic differentiation is different from the issue of overall economic inequalities in society, which is based on unequal economic rewards regarding different occupations and educational attainments. Table 5.2 shows the items included in the measures of New Politics orientations. Table 5.2: Indicators for New Politics values Authoritarian/libertarian values The point of departure for constructing the index for libertarian/authoritarian values was the items used by Flanagan and Lee (1988) in their analysis of the first wave of the European Values Survey (1981–83). These were all replicated in the second and third wave. In Flanagan and Lee’s factor analysis (Flanagan and Lee 1988: 23–24, Table 2) they find support for the notion that nine personal values load on the libertarian/authoritarian dimension. I did not include two of these in my analysis because I could not see that they theoretically tap the libertarian/authoritarian dimension. The statements in the questionnaire for the remaining seven items are listed below. 1) People have different ideas about following instructions at work. Some say that one should follow instructions of one’s superiors even when one
8 The left-right self-placement variable, which is available in all three EVS waves, does not assess value orientations and is not included here. It represents an identification with the symbols ‘left’ and ‘right’. Research has shown that it can be associated with different value orientations over time and across countries (Knutsen, 1995b), and it is also causally a consequence of value orientation and party choice (Inglehart & Klingemann, 1976; Knutsen, 1997).
124
oddbjørn knutsen
Table 5.2 (cont.) does not fully agree with them. Others say that one should follow one’s superior’s instructions only when one is convinced that they are right. Which of these two opinions do you agree with? 2) Which of these statements do you tend to agree with? A. Regardless of what the qualities and faults of one’s parents are, one must always love and respect them B. One does not have the duty to respect and love parents who have not earned it by their behavior and attitudes 3–6) Here is a list of qualities which children can be encouraged to learn at home. Which, if any, do you consider to be especially important? Please choose up to five. (11 qualities were mentioned in 1990 and 1999/2000. 17 qualities were mentioned in 1981). Good manners Independence Imagination Obedience 7) Here is a list of various changes in our way of life that might take place in the near future. Please tell me for each one, if it were to happen, whether you think it might be a good thing, a bad thing, or don’t you mind? (seven changes were mentioned) C. Greater respect for authority Materialist/postmaterialist value orientations There is a lot of discussion about what the aims of this country should be for the next ten years. On this card are listed some of the goals which different people would prefer. (SHOW CARD). Would you please say which one of these you, yourself, consider most important? What is your second choice? Maintaining law and order in the country (M) Giving the people more to say in important governmental decisions (PM) Fighting rising prices (M) Protecting freedom of speech (PM)
The libertarian/authoritarian items are similar to those used by Flanagan and Lee, although I did not include items related to religious/secular values as these authors do. These values were validated elsewhere (Flanagan & Lee, 1988, 2003: 237–243; Knutsen, 1996: 303–305).
the end of traditional political values?
125
There is a problem with including the four items assessing the qualities which children should be encouraged to learn at home (items 3–6 in the table). In all waves, the respondents should emphasize only five qualities, but there were 17 alternatives in 1981, compared with 11 in the 1990 and 1999/2000 waves. All four alternatives concerning libertarian/authoritarian values were included also in 1990 and 1999/2000, and the other alternatives were exactly the same in these surveys. The stated change implies nevertheless that the likelihood of choosing the remaining 11 alternatives increases, and I find a clear increase in the portion that chooses the relevant values from 1981 to 1990 in all countries. This problem is dealt with in two ways: The index construction implies that two values (the libertarian) are coded 1 while the two authoritarian values are coded 0. The increased likelihood to choose the values included in the index then counterbalance each other: If there is a decrease of 10 percent in choosing one of the authoritarian values and a similar decrease in a libertarian value, there will be no value change according to the index, since the two changes cancel each other out. The other way is to construct an alternative index based on the remaining three items. Results based on this index will be briefly referred to in the text, but the seven-item index remains the main index in the empirical analysis in this chapter. Finally, MPM values are gathered by the four-item value batteries developed by Ronald Inglehart. Only this item battery was asked in all EVS surveys. Most of the research on materialist/postmaterialist values is done on the basis of this four-item battery. The two New Politics measures are constructed as unweighted additive scales with values from 0 to 10, where libertarian and postmaterialist values are assigned the highest scores. Trends in political values Introduction In this section, I will examine trends over time in the various value orientations in a comparative setting. I will also investigate if there is a trend towards cross-national convergence in the value orientations. In the following, I will start with the indicators for economic left-right values and then examine the new politics orientations.
126
oddbjørn knutsen Economic left-right values
Equality versus freedom. For the item on equality versus freedom, I rely on an ‘opinion balance’ measure. The percentage that chose freedom is subtracted from the percentage that chose equality.9 A minus value implies therefore that freedom is chosen by a larger portion of the sample than equality. Since the equality versus freedom item was not asked in the United States in 1999–2000, means and standard deviations are calculated based on the remaining 12 countries. 20,0 10,0 0,0 1981 –10,0 –20,0 –30,0 –40,0
1990
2000
Belgium Denmark France Iceland Ireland Italy Netherlnds Spain Sweden Great Britain USA West Germany Northern Ireland Mean 12 countries
–50,0 –60,0
Figure 5.1: Trends in equality versus freedom based on opinion balance measure
In 1981, the distributions clustered in three groups. In the first group—comprising Britain, United States, and Northern Ireland— freedom was clearly emphasized more strongly than equality. In the second group, freedom was still more strongly emphasized but not so pronounced as in the first group (Belgium, Denmark, France, the Netherlands, and Sweden), and finally, in the third group of coun-
9 The ‘neither’ category comprises less than 10% in most of the surveys, and is considered as a neutral category. It is not set to missing when the opinion balance measure is calculated.
the end of traditional political values?
127
tries, freedom and equality were fairly evenly emphasized among the mass public (Iceland, Ireland, Italy, Spain, and West Germany). According to the mean of the 12 countries, there was a small and gradual long-term increase in the emphasis of equality over time. This change was however only 3 percentage points from 1981 to 1999–2000,10 and freedom was still emphasized more strongly than equality around 2000 according to the average for the 12 countries. The greatest change towards a stronger emphasis on equality was found in Northern Ireland (33 percentage points) and Belgium (26 pp.), and then Great Britain, followed by France and Iceland (11–19 pp.). For the three latter countries, the main shift towards equality took place during the 1980s, whereas in Northern Ireland, the shift ocurred during the 1990s, while the change was of a more gradual nature in Belgium. The largest change in the opposite direction took place in Germany (39 pp.) followed by Denmark (23 pp.), while there were considerable smaller shifts towards stronger emphasis on freedom in Spain and Sweden (9–11 pp.). There was a dramatic shift away from equality during the 1980s in West Germany, and most of the shift in Spain and Sweden also took place during the 1980s. The shift in Denmark was more gradual. In Italy, the Netherlands, and Ireland, there were fairly stable patterns, although somewhat stronger emphasis on equality. We note a curvilinear pattern for Ireland, where equality was strongly emphasized in the 1990s, but less so around 2000. Around 2000, freedom was most strongly emphasized in Denmark and West Germany, the two countries where the shift towards freedom had been strongest. Freedom was also strongly emphasized in Sweden, Great Britain, and the Netherlands. In the United States, freedom was strongly emphasized both in 1981 and 1990. The emphasis of freedom in Denmark in 1999–2000 approached the American level in 1990. Although there were considerable changes regarding the relative ranking of the various countries over time in the priority of equality versus freedom, there was no convergence between the countries.
10 The change can be illustrated by the mean distributions of the 12 countries. On average, 56% of the mass public emphasizes freedom in 1981 and 1990. This declines to 54% in 1999–2000. The emphasis of equality increases from 37% in 1981 to 39% in 1990 and 40% in 2000.
128
oddbjørn knutsen
In fact, the standard deviation for the opinion balances increased slightly, from 15.4 in 1981 to 15.6 in 1990 and 16.6 in 1999–2000. This is reasonable, given that we observed changes in opposite directions—both towards freedom and towards equality in different countries. Pay equality. As to the item of pay equality, there is a clear trend towards greater acceptance of pay inequality over time (not shown in tables). There was generally a great acceptance of the principle that efforts should be rewarded economically even for the same type of jobs. According to the average of the 13 countries, this rightist principle was supported by 64% in 1981, 75% in 1990, and 77% in 1999–2000. There were fairly large variations between countries. In 1981, 80% in the United States and 68–70% in West Germany, Spain, and Great Britain accepted this form of economic inequality compared with 48% in Italy and 59–60% in Denmark, Ireland, and Sweden, which were the countries where pay inequality was least accepted. The shift towards accepting larger pay inequality was largest in some of the countries, where inequality was least accepted in the early 1980s: Italy (increase of 30 percentage points), Denmark, Sweden, and Iceland (17–23 pp.). There were smaller increases in the other countries apart from Spain, where a small decline (4 pp.) in the acceptance of pay inequality was detected. Around 2000, the acceptance of pay inequality for the same job was still strongest in the United States (91%) and West Germany (87%) and then Iceland and Denmark (82–86%), while the smallest acceptance was found in Spain and Ireland (65–66%). According to the standard deviations for the distributions of the 13 countries, there was no long-term convergence. There was a clear convergence from 1981 to 1990, but then the spread between the countries increased to the same level as in the early 1980s. From 1981 to 1990, there was an increase in the acceptance of pay inequality in all countries and 70–79% of the mass publics accepted the principle of pay inequality in all countries apart from 3, but from 1990 to 1999–2000 the trend towards greater acceptance of pay inequality continued in some countries (the Nordic countries and United States), while there was a trend towards greater acceptance of the principle of pay equality in some other countries (first and foremost Spain, Ireland, and Britain). This contributed to a larger spread in cross-national value priorities from 1990 to around 2000.
the end of traditional political values?
129
New Politics values I will first examine value change on libertarian/authoritarian values, then materialist/postmaterialist values according to Inglehart’s value battery. Libertarian/authoritarian values. There were fairly large cross-national differences in libertarian/authoritarian values. In the early 1980s libertarian values were strongest emphasized in West Germany and Denmark, and then in Sweden and France, followed by the Netherlands and Ireland. The mass public in Northern Ireland and then in Ireland and the United States had the most authoritarian values, while Spain, Italy, Belgium, and Great Britain comprised a middle group between the two first and the last one. According to the average mean for the 13 countries, there was only a slight change from authoritarian to libertarian values. Most of the change took place from 1990 to 1999/2000.11 There was,
6,00
5,00
4,00
Belgium Denmark France Iceland Ireland Italy Netherlnds Spain Sweden Great Britain USA West Germany Northern Ireland Mean 12 countries
3,00
2,00
1,00
0,00 1981
1990
1999–2000
Figure 5.2: Trends in libertarian/authoritarian values
11 The mean score for all 13 countries increased from 3.92 in 1981 to 3.96 in 1990 and 4.13 in 2000.
130
oddbjørn knutsen
however, considerably more change in several countries than the average score indicates. There was a trend towards more libertarian values in most countries, in fact in 9 countries. The largest change towards more libertarian values took place in Sweden (1.09), and then in Northern Ireland and the United States (around 0.65), and there was considerably smaller change in the other countries. In France, there was a great change towards more authoritarian values (1.05). This change took place mainly during the 1980s. There was also a smaller change in the same direction in Spain (0.34), while an authoritarian trend in West Germany was very small. Figure 5.2 reports how these changes alter the grouping of countries in 1999–2000: The three countries, which had the most authoritarian mass publics in the early 1980s, approached the group of countries, which comprised Spain, Italy, Belgium, and Great Britain in 1999/2000. France also belonged to this group due to the authoritarian trend, and in fact only in Ireland and Northern Ireland, is there a more authoritarian mass public in 1999/2000. The countries with the most libertarian mass public remained otherwise the same. There was no trend towards convergence, although there was considerable value change in the countries with the most authoritarian mass public in the early 1980s.12 The alternative three-item index (which does not include the items regarding qualities children should learn at home) shows very similar results. We find the same grouping of countries in the early 1980s, based on the ranking of the nations, and very similar trends over time. The trends towards more libertarian values in the group with the most authoritarian values in the early 1980s was very pronounced also for this alternative index, and the same applied to the authoritarian trend in France. The average value change was also fairly small. Materialist/postmaterialist values. Value change along the MPM value orientations have been studied intensively by Ronald Inglehart, who has used the Eurobarometers to study change over time in 8 of the countries included in this analysis based on the four-item value battery (Inglehart, 1977, 1990; Abramson & Inglehart, 1995). Inglehart emphasizes that there are clear long-term changes towards post materialist values, but also that there are short-termed changes due
12
The standard deviation was 0.79 in 1981, 0.84 in 1990 and 0.83 in 1999/2000.
the end of traditional political values?
131
to changing economic circumstances. The four-item battery is particularly sensitive to inflation rate due to the ‘fighting rising prices’ alternative included as one of the four political values. Unemployment rates also have some short-termed impact on value priorities. In Inglehart’s formulations, the scarcity hypothesis implies short-termed period effects caused by the economic circumstances of the respective time, while the socialization hypothesis implies cohort effects caused by different economic and physical security during the formative years of different cohorts (Abramson & Inglehart, 1995: Chap. 3 and 4). Inglehart’s Eurobarometer data allows him to study the change according to the four-item battery from the early 1970s. Due to the economic recession of the mid-1970s and early 1980s, there was no trend towards postmaterialist values from the early 1970s to around 1980. During the 1980s there was however considerable value change caused both by better economic environmental factors and generational replacement (Abramson & Inglehart, 1995: Chap. 3 and 4). Figure 5.3 shows how MPM values changed in the 13 countries according to the EVS data set. The mean for all countries is based on only 11 countries since I lack data for the United States (for 1981) and Britain (for 1999–2000). There was considerable change towards postmaterialism from the early 1980s to 1990 (from 3.70 to 4.73), and then stability according to the figure based on the average for the 11 countries. Indeed, there was a small decline, from 4.73 to 4.60 from 1990 to 1999/2000. In 1981, postmaterialist values were clearly strongest in Denmark, and then in a number of countries. The mass public in Northern Ireland, Spain, and Italy was most materialist, and the Italians and Irish were also clearly more materialist than average. The change towards postmaterialism from 1981 to 1990 was large in most countries. In all of the countries, where postmaterialist values were least emphasized in 1981, there was a large change. The largest changes took place in these countries and in the Netherlands, which was the country where postmaterialist values were most widespread in 1990. Three countries followed an opposite pattern with a trend towards materialism, namely Denmark and—of a smaller magnitude—France and Iceland. From 1990 to 1999/2000, there was a small average trend towards materialism, and this trend was most pronounced in West Germany and the Netherlands, and then in Ireland, Northern Ireland, and
132
oddbjørn knutsen
7,00
6,00
5,00 Belgium Denmark France Iceland Ireland Italy Netherlnds Spain Sweden Great Britain USA West Germany Northern Ireland Mean 12 countries
4,00
3,00
2,00
1,00
0,00
1981
1990
1999–2000
Figure 5.3: Trends in materialist/postmaterialist values according to the four-item battery
Spain, while there was a clear trend towards postmaterialism in Italy and United States, and a small trend in Denmark and Sweden. The long-term value change towards postmaterialism from 1981 to 1999–2000 was largest in Italy (2.48), followed by Belgium, Spain, and Northern Ireland (around 1.60), four of the five countries with the most materialist population in the early 1980s. A shift towards materialist orientations could be observed in two countries, Denmark and France (around 0.50). There was a clear convergence in the cross-national value priorities from 1981 to 1990, and then even stronger convergence during the 1990s. The large change towards post-materialism in the countries, where materialist values were most pronounced in 1981, is the main explanation. Generation differences in value orientations Introduction The main research questions in this section are: 1. Are there cohort differences in political value orientations and are there cross-national differences in this respect? Do these cohort
the end of traditional political values?
133
differences change over time? Here I will examine in particular the countries, where we detected considerable value change in the previous section. 2. When I take the younger cohorts as an indication of the new pattern of value priorities and the war cohort as an indication of an old pattern, is there a trend towards convergence or divergence in value orientations between countries? Do the post-war cohorts have more similar value priorities than the war cohort in a cross-national context? I have data for the three cohorts for 13 countries for 3 time periods, corresponding to the three waves of the EVS. It is rather complicated to present all these data for the various sets of value orientations. My research strategy is as follows: I first rely on a figure for the averages for the three periods for studying the value orientation of each cohort as a point of departure for describing the main pattern of each country. I will use the war cohort as my baseline cohort and compare the two other cohorts with the value orientation for this cohort. This facilitates the comparison of the cohort differences between countries, given that the overall priorities of values vary between countries. I will then describe changes of the cohort differences over time in the text. Economic left-right values Equality/freedom. On the equality/freedom item, there were on average fairly small cohort differences but the differences had an interesting form. The boom cohort emphasized equality stronger than the war cohort, while the bust cohort emphasized freedom even somewhat more strongly than the war cohort. The means for all countries for all periods were –18, –15, and –20 for the war, boom, and bust cohorts, respectively. Differences were small but in the expected direction according to the hypothesis about differences between the two post-war cohorts. Figure 5.4 shows the cohort differences in the various countries with the war cohort as a baseline. As for the other corresponding figures below, the countries are ranked (from left to right) according to the cohort differences between the war and the bust cohort. Cohort differences vary largely between countries. In some countries, the bust (and the boom) cohort emphasized equality more strongly than the war cohort, while the opposite occurred in other countries.
oddbjørn knutsen
134 25,0 20,0 15,0 10,0
Boom Bust
5,0 0,0 nd
–5,0 rela I
n er th –10,0 r o N
in
SA a
re
G
ita
r tB
U
ce
d
lan
Ire
an
Fr
ly
Ita
y
m
iu
lg Be
tG
es
W
k
an
ar
m
er
D
m en
n
de
e Sw
ds
ain
Sp
lan
et
r he
d
lan
e Ic
N
–15,0 –20,0 –25,0
Figure 5.4: Cohort differences in emphasis of equality versus freedom. Differences between the war and the two other cohorts. Means for the three surveys
In several countries, the boom cohort emphasized equality more strongly than the bust cohort. This was particularly pronounced in the case of Denmark, where the boom cohort emphasized equality more strongly than the war cohort, while it was the opposite in the bust cohort. Similar patterns of much smaller magnitudes are found in France, Italy, and Belgium, where there were no differences between the war cohort and the bust cohort, whereas the boom cohort emphasized equality more strongly than the other cohorts. In all the countries where both the boom and the bust cohort supported freedom more than the war cohort, there were fairly large differences between the two post-war cohorts. The emphasis on freedom was larger in the bust cohort. In fact, when the war and the boom cohort are compared, cohort differences in Denmark were second largest after Northern Ireland, and the largest cohort differences between the bust and the boom cohorts were found in Denmark, Iceland (16–17 percentage points), Spain, and in the Netherlands (8–11 pp.). In these countries, the stronger emphasis on freedom in the bust cohort was particularly large. It was in the liberal welfare states, not in the welfare states in Southern Europe, that the younger cohorts were most inclined to
the end of traditional political values?
135
emphasize equality, while in the well-developed universal welfare states the younger cohorts emphasized freedom. Spain showed, however, the same pattern as in the last group. There was no trend towards smaller cohort differences over time according to the average of the 13 countries. Cohort differences were smallest in 1990, regaining strength to the same level as in 1981 in 1999–2000, and the particular pattern from 1981 was the same in 1999/2000 with the boom cohort emphasizing equality more than the other cohorts and the bust cohort emphasizing freedom more than the two other cohorts. The most interesting aspect in regard to the examination of cohort changes over time is the degree to which periods effects interact with cohort effects, or if period effects alone can explain changes over time in value orientations. I will focus here upon the countries where most change over time was observed in the former section. Among the countries, where equality was emphasized more strongly over time, period effects were pronounced, but the cohorts moved towards the left direction, with somewhat varying speed. There was no clear pattern for all of the countries investigated. In Britain, there were only small cohort differences in 1981, and the trend towards stronger emphasis on equality was strong in the two post-war cohorts while the stronger emphasis of freedom in the war cohort persisted over time. In Northern Ireland, the drift towards equality was strong in all cohorts, but strongest in the boom cohort. In Belgium and Iceland, equality was more strongly emphasized in all cohorts over time, but changes were smaller in the bust cohort compared to the others. Finally, in France, the period effect appeared to be somewhat smaller in the boom cohort compared with the two other cohorts. Among the countries where freedom was more strongly emphasized over time, this shift was much stronger in the two post-war cohorts in Denmark and Sweden, while there was insignificant change in the war cohort. In Denmark, there appeared much stronger emphasis on equality in the boom cohort in 1981, which diminished over time. In 1999–2000, the war and the boom cohort had a very similar emphasis of freedom versus equality, while the bust cohort was considerably more in favor of freedom, compared to the other two cohorts. This is totally opposite to the pattern in Spain, where change was concentrated in the war cohort, which emphasized freedom considerably more over time, while there was a fairly stable pattern in the two other cohorts. The strong trend towards stronger emphasis
136
oddbjørn knutsen
on freedom in West Germany was mainly a period effect. The trend towards freedom was strong in all cohorts and the fairly small cohort differences persisted over time. According to the average standard deviation for all countries and surveys, there was a trend towards convergence in the post-war cohorts, compared with the war cohort. This trend was most strongly in 1981 and declined considerably in 1999–2000. In some of the countries where freedom was strongest emphasized in the war cohorts (United States, Britain, and Northern Ireland), the post-war cohort had a stronger tendency to emphasize equality. This contributes to smaller cross-national differences in the post-war cohorts. Pay equality. We found above that pay inequality was more accepted over time in almost all countries. However, it was not the case that the post-war cohorts were the vanguard in accepting this value. According to the average score for all countries, it was the war cohort that was most inclined to support pay inequality (not shown in any figure). Differences were not large: 74% in the war cohort compared with 72% and 68% in the boom and the bust cohorts, respectively. It was thus the bust cohort that was most leftist in this value orientation. According to the mean for the various countries, cohort differences vary from 11–12 percentage points in Britain and Northern Ireland, 5–9 percentage points in Denmark, France, Sweden, West Germany, Spain, United States, and the Netherlands, and less than 5 in Belgium, Iceland, Ireland, and Italy. The period effect was towards accepting pay inequality, and this effect interacts with cohort in an interesting way. This can be illustrated by examining the means of the various cohorts in the three time points for which I have data. In 1981, pay inequality was accepted by 69% in the war cohort, 61% in the boom cohort, and 56% in the bust cohort. These differences vanished over time and there was no significant difference between the cohorts in 1999–2000. Pay inequality was accepted by 77–78% in all three cohorts. The younger cohorts were then more strongly influenced by the period effect towards accepting greater pay inequality compared with the war cohort. While on average the change was 9 percentage points in the war cohort, it was 16 and 20 percentage points in the boom and bust cohorts, respectively. This pattern is asserted in most of the countries. Fairly strong cohort differences in 1981 decreased in 1990 and thereafter disappeared in 1999–2000. In Italy and Iceland, where a strong increase
the end of traditional political values?
137
in support of pay inequality over time could be recorded, there was even a reversed development concerning the differences between the cohorts. While the younger cohorts were most likely to support pay equality in 1981, they were least inclined to support pay equality in 1999–2000. Only in Great Britain, was there no clear interaction between cohort and period effects: there was a fairly small trend of similar magnitude towards accepting pay inequality in all cohorts, and the post-war (in particular the bust cohort) remained the most reluctant even in 1999–2000. There was no trend towards crossnational convergence in the youngest cohorts. New Politics values Libertarian/authoritarian values. Figure 5.5 shows the average cohort differences based on the three waves with the war cohort as a baseline. The two post-war cohorts have more libertarian values than the war cohort in all countries but the differences between the cohorts vary somewhat as the figure shows. The largest differences between the war and the bust cohorts were found in West Germany (2.11) and Spain (1.68), and the smallest one was found in the United States (0.47). The boom cohort had in most countries somewhat less libertarian values than the bust cohort but the differences were small. The main difference was between the war and the two post-war cohorts. This can be illustrated by the means of the various cohorts based on all surveys in all countries: It was 3.36 for the war cohort, 4.27 for the boom, and 4.46 for the bust cohort. There were some countries, where the two post-war cohorts had fairly similar average value orientations, but the boom cohort was not significantly more libertarian than the bust cohort in any of the countries. These patterns were impressively stable over time. There was no tendency towards smaller cohort differences as those found for the economic left-right orientations. This can again first be illustrated by the means for all surveys: The differences between the value priorities of the war and the bust cohorts were 1.08 in 1981, 1.16 in 1990, and 1.08 in 1999–2000. There were correspondingly small changes for the boom cohort, although there was a small decrease in the difference to the war cohort, from 0.95 and 0.97 in 1981 and 1990 to 0.83 in 1999/2000. Examining the cohort differences in the various countries over time, there is again an impressing stability. Only 3 changes were larger than 0.50 from 1981 to 1999–2000
oddbjørn knutsen
138 2,50
2,00
Boom Bust
1,50
1,00
0,50
SA U
ela Ic
lan
nd
d
in Ire n er rth
ce
ita Br
en
an Fr
ed Sw
um
k ar
lgi Be
ly
nm De
lan er
Ita
ds
d lan Ire
th Ne
ain Sp
No
W
es
tG
er
m
an
y
0,00
Figure 5.5: Cohort differences in libertarian/authoritarian values (differences between the war and the two other generations based on means for the three surveys)
when the war and the two post-war cohorts were compared: in Denmark, the bust cohort became relatively more authoritarian compared with the war cohort (differences declined from 1.45 to 0.69). In Spain and Italy, the boom cohort became relatively more authoritarian over time (differences declined from 1.42 to 0.71 and from 1.05 to 0.55, respectively). For some countries, long-term value change along the libertarian/authoritarian value dimension was found above. Can these changes be explained by cohort differences and generational replacement? The large shift towards authoritarian values in France was basically a period effect, which influenced all cohorts fairly similarly. Cohort differences persisted and were fairly large (see Figure 5.5) and if there had not been any generational replacement, the drift towards authoritarianism would have been significantly larger. In Spain, there was also a shift towards more authoritarian values, and there was also a clear period effect but the shift was largest in the post-war cohorts and in the boom cohort in particular as I have indicated. Cohort differences remained nevertheless large in Spain also in 1999–2000.
the end of traditional political values?
139
Differences between the war and the bust cohort remained second largest in all surveys after West Germany, where the cohort differences remained largest in all points in time. The shift towards libertarian values in Sweden, Northern Ireland, and in the United States were mainly caused by period effects. In the United States and partly also in Northern Ireland, cohort differences were fairly small and remained so over time. In Sweden, there was also a fairly strong period effect but the libertarian shift in all cohorts was considerably smaller than the total shift towards libertarian values. Given the fairly strong cohort differences in the Swedish case, generational replacement can explain some of the shift towards more libertarian values in the Swedish population. There was no tendency towards cross-national convergence when I compared the war cohort with the post-war cohorts. In fact, the opposite was the case. Cross-national differences were larger in the two post-war cohorts than in the war cohort and remained so for all three surveys. Materialist/postmaterialist values. Figure 5.6 shows the cohort differences with the war cohort as a baseline. The post-war cohorts were more postmaterialist than the war cohort in all countries according to the average scores, and there were large cross-national differences. Cohort differences were largest in Spain, West Germany, and Italy, smallest in the United States followed by Iceland and Denmark. There was a watershed between the war and the two post-war cohorts, and the bust cohort was more postmaterialist than the boom cohort on average. The average cohort differences between the war and the two other cohorts were 1.20 and 1.47 for the boom and the bust cohorts, respectively. The bust cohort was then on average more inclined to support postmaterialist values than the boom cohort, and from Figure 5.6 it becomes evident that this applied to most countries. Only in Iceland was the boom cohort more supportive of postmaterialist values while there were insignificant differences in Denmark, Sweden, the Netherlands, and the United States. Do these differences persist, become smaller or vanish over time? A main pattern is found by examining the average cohort differences in 1981, 1990, and 1999–2000. The average score on the index for each cohort and the differences between the cohorts in each wave are shown in Figure 5.7, based on the 11 countries for which data is available for all time points.
oddbjørn knutsen
140 2,50
2,00
Boom Bust
1,50
1,00
0,50
er
SA U
nd ela
k Ic
ar nm
rth
De
lan
d
ce n
Ire
en
an Fr
Sw
ed
um
d
in ita
lgi Be
Br
lan
ds lan er
th
Ire
ly Ita Ne
y an
No
W
es
tG
er
m
Sp
ain
0,00
Figure 5.6: Cohort differences in MPM values (differences between the war and the two other cohorts according to the mean for the three surveys) 6,00 5,61 5,16
5,02 4,81
5,00
4,00
4,37 4,23
3,83
3,00 2,99
3,83
War Boom Bust
2,00
1,00
0,00 1981
1990
1999–2000
Figure 5.7: Mean value for 11 countries on the MPM index for different cohorts in each of the surveys
the end of traditional political values?
141
Although there was a strong value change towards postmaterialism in the war cohort from 1981 to 1990, there were even larger changes in the two post-war cohorts, in particular in the bust cohort. The period shift towards postmaterialist values from 1981 to 1990 was stronger in the post-war cohorts and in particular in the bust cohort. From 1990 to 1999–2000 there was a significant decline in the cohort differences. Both post-war cohorts became more materialist in an absolute sense and compared with the war cohort which had a constant value profile according to the average mean. This implies that there was a significant decline in the cohort differences from 1990 to 1999/2000, and indeed there was also a decline in these differences from 1981 to 1999/2000. There were nevertheless fairly large cohort differences as well in 1990–2000, although somewhat smaller than in 1981. Due to the larger change towards materialism from 1990 to 1999/2000 within the two post-war cohorts, the largest value shift from 1981 to 1999/2000 appeared on average to take place in the war cohort. Examining the trends in the various countries, similar changes as those shown for the mean are found in most of the countries, although the shift towards postmaterialism from 1981 to 1990 and the shift towards materialism from 1990 to 1999/2000 varied between the countries. The shift towards postmaterialist values in the bust cohort (compared with the war cohort) was 0.40–0.85 in all countries apart from Belgium, West Germany, Ireland, Britain, and Denmark from 1981 to 1990. In these latter countries, the changes were small. From 1990 to 1999–2000, there is a relative shift towards materialist values (compared with the war cohort) in all countries, but the strength varies from 0.70–0.90 in West Germany, Iceland, Sweden, and Spain, 0.50–0.69 in France, Ireland, and United States, and 0.30–0.49 in Northern Ireland, Belgium, and Denmark. From 1981 to 1999–2000, the decline in cohort differences (bust compared to war cohorts) was particularly large in Ireland (decline 1.46), the Netherlands (0.72), Denmark (0.52), and West Germany (0.33), while they increased significantly only in Sweden, Spain, and France (0.30–0.45). I already mentioned above the shifts towards postmaterialism in the war cohort. These shifts were significant and even large in the countries with the largest value shift towards postmaterialism from 1981 to 2000: Italy (2.34), Belgium (1.67), Spain, Northern Ireland, Ireland, the Netherlands (1.20–1.40), and Sweden (0.82). In all these countries, there were strong period effects, which induced the war
142
oddbjørn knutsen
cohort to emphasize postmaterialist values. In several of these countries, cohort differences decreased from 1981 to 1999–2000, indicating that value change was even larger in the war cohort than in the other cohorts. Despite these variations in cross-national differences in cohort changes, the cohort differences in 1999–2000 followed strongly the pattern from Figure 5.6. They were largest in Spain, West Germany, and Italy, and then in France, Sweden, and Belgium, and smallest in United States, Iceland, and Ireland. Although there are clear cohort differences in MPM values indicating that the post-war cohorts were more inclined to support postmaterialist values, there was no trend towards convergence apparent when the war and the two other cohorts were compared. The crossnational differences were even larger in the two post-war cohorts. Multivariate analysis Value orientations are to some extent anchored in social structure. A major difference between Old Politics values, and economic leftright orientations in particular, and New Politics values, is that the former have been anchored in the hierarchical structural variables in industrial society. Economic leftist orientations are most strongly supported by workers and those with lower education and incomes, while the opposite is the case for those who support rightist values (Knutsen, 1995a: 182–190; Svallfors, 1997: 292–293). The emergence of new politics turns the old order upside down in the sense that radical and change-oriented postmaterialist values tend to be most strongly supported by the higher educated strata and those who belong to the service class (Inglehart, 1990: 259, 277–279). This is supported by the EVS data.13 Leftist economic values were most strongly supported by the lower educated strata and the working class, while rightist values were most strongly supported by employers and the higher level of the service class. Libertarian and postmaterialist values gained stronger support from the service class and the more highly educated strata than from workers, farmers, and employers and those with less education.
13 In this paragraph I rely on bivariate correlations based on all surveys for all countries included in this study.
the end of traditional political values?
143
I conducted multiple regressions with the various measures for value orientations as dependent variables. Since I already described the main patterns for the relationship between the various sociostructural variables and Old and New Politics values orientations above, I will therefore simply present the R2 and how it changes when additional variables are included in forced stepwise regressions. I will only report the regression coefficients for the time of the survey (a proxy for the period effect) and for the various cohorts. The socio-structural variables are included after country in the following order, according to causal considerations: 1) gender, 2) religious denominations, 3) education, 4) social class,14 5) time of the survey (which indicates period effects), and finally 6) cohort. The results are reported in Table 5.3.15 I will first comment on the variables that are included before time and cohort in the analyses, and then focus upon the impact of these latter variables. Country differences are largest for libertarian/ authoritarian values. Somewhat surprisingly, the total explanatory power of country and social structure (gender, religious denomination, education, and social class) is largest for the two New Politicsorientations and very small for the indicators of economic left-right orientations. This applies in particular to the hierarchical status variables education and social class, where the explanatory power is considerably larger for the New Politics orientations. There is also an important difference regarding which of the status variables is most influential. Education has larger explanatory power for the New Politics orientations, while social class dominates regarding economic left-right orientation. Religious denomination has considerable explanatory power for libertarian/authoritarian values and for MPM.
14
I used a version of the Erikson/Goldthorpe class scheme, which was possible to construct on the basis of the data of all three surveys. The variable comprises seven class categories and one category for those without an occupation (and a spouse without a present occupation). 15 The equality/freedom variable is coded as follows in this analysis: 0) freedom, 5) neither and 10) equality. The pay equality variable is coded 0) for acceptance of reward and 10) for no acceptance. I also ran logistic regressions for these two dependent variables since they have few values (equality/freedom is dichotomized in this analysis). The Cox and Snell pseudo R2 is nearly exactly the same as R2 in Table 5.3, while Nagerkerke’s R2 is somewhat higher for the full model (0.054 and 0.078, for pay equality and equality/freedom, respectively). The relative importance of the explanatory power of each variable is exactly identical in these stepwise logistic regressions, compared with the OLS-model reported in Table 5.3.
–0.116 0.147
0.191 –0.116
45878
b coefficients 1990 1999/2000
Boom Bust
N
0.002 0.001 0.001 0.004 0.000 0.001
R2 ch.
48071
0.046 0.068
–0.124 –0.124
0.015 0.017 0.017 0.026 0.036 0.050 0.053
R2 0.002 0.000 0.009 0.010 0.014 0.003
R2 ch.
B. Pay equality
R2 ch. means change in R2 from previous step.
0.028 0.030 0.031 0.032 0.036 0.036 0.037
Country Gender Religious den. Education Social class Time Cohort
R2
A. Equality/freedom
51313
0.659 0.732
–0.187 –0.304
0.096 0.097 0.131 0.189 0.195 0.197 0.215 0.001 0.034 0.058 0.006 0.002 0.018
C. Libertarian/ authoritarian values R2 R2 ch.
50447
0.800 0.949
1,381 0.099
0.027 0.032 0.051 0.085 0.095 0.124 0.135
0.005 0.019 0.034 0.010 0.029 0.011
D. Materialist/ postmaterialist values (four-item battery) R2 R2 ch.
Table 5.3: R2 and change in R2 from forced stepwise regressions with the value measures as dependent variables
144 oddbjørn knutsen
the end of traditional political values?
145
A closer examination shows that there is a division between all religious denominations and those without any religious affiliations, that is important. The latter group is considerably more supportive of libertarian and postmaterialist values than all the former affiliated groups (mainly Protestants and Catholics). The impact of period is largest for the four-item MPM index. We know from previous research that this index is sensitive to period effects caused by given items in that battery as I discussed above. The multivariate analyses confirm that cohort effects are largest for the New Politics orientations when prior variables are controlled for. Only for these orientations do considerable cohort effects exist when the data are pooled in this way. This is in accordance with the analyses in the previous section. The impact of cohort on economic left-right orientations is small, also compared with the impact of other social structural variables. Even for libertarian/authoritarian and MPM orientations, the impact of cohorts is secondary compared to education and religious denomination. Finally, the total explanatory power of the whole model is larger for the New Politics orientations, indicating that both country differences in value priorities and anchorage in social structure are larger for New Politics orientations than for Old Politics orientations in advanced industrial democracies. In particular libertarian/authoritarian values are both fairly strongly anchored in social structure (including age/cohort) and differently supported in the various countries. Conclusions The two indicators for economic left-right values do not provide us with very clear trends and comparative patterns. The decline in support for the pay equality principle indicates that there were some right-wing trends regarding the orientations of the mass public in Western democracies. On the other hand, the same trend towards the right regarding the important principles of equality versus freedom is not found. Freedom gained strongest support from nearly all countries in the early 1980s, and did so around 2000 as well, but there was a small average trend towards equality. The cross-national variations of support for economic left-right values did not conform clearly to any of the expected patterns discussed in the section about hypotheses and perspectives. There were
146
oddbjørn knutsen
nevertheless some patterns that can be linked to the expectations about strongest support for the leftist values in the least advanced welfare states, and a few patterns which indicated most strongly support for leftist values in the universal welfare states: In the early 1980s, the principle of equality was most strongly emphasized in the Southern European countries (Italy and Spain) and pay equality was strongest emphasized in Italy. Some of these patterns persist over time, while others diminish. For example, the principle of pay equality was not supported stronger in Italy in 1990 and 1999/2000 than in other countries, and the emphasis of equality (versus freedom) declined strongly in Spain but not in Italy. The countries, which belong to the universal welfare states (Denmark and Sweden, partly Belgium and the Netherlands) did not show any clear leftist orientation. In Denmark and Sweden (and in West Germany), there was a strong trend towards stronger emphasis on freedom. The mass public in Denmark and Sweden supported pay equality fairly strongly in a comparative setting in 1981, but there was a strong rightist trend away from pay equality in these countries. Cohort differences regarding left-right orientations were generally small and they varied between the countries even directionally. This was most pronounced for the equality/freedom item, where the younger cohorts in the liberal welfare states emphasized equality more strongly than the war cohort, while the opposite was true for some of the universal welfare states. According to this measure, the boom cohort was somewhat more leftist and the bust cohort somewhat more rightist than the war cohort. With regard to the pay equality indicator, the younger cohorts (both boom and bust) were more leftist than the war cohort, but the cohort differences became smaller over time. This can be explained as a combination of a period and a cohort effect but also as a life-cycle effect. In sum, there is no clear trend towards leftist or rightist economic values over time. There are some patterns, which indicate that leftist values are strongest in the southern European countries, but values vary little across welfare regimes. Neither economic leftist nor economic rightist values are new. Cohort differences regarding these values are small and the younger generations do not seem to be vanguards for leftist or rightist values. The new politics value orientations showed both patterns of similarity and difference. A main pattern of similarity was the trends over time. There was a clear trend towards both libertarian and
the end of traditional political values?
147
postmaterialist values over time.16 On the other hand, postmaterialist values gained stronger support from 1981 to 1990, while the largest increase in support for libertarian values took place during the 1990s. The cross-national variations in value priorities were considerably more consistent for the new politics orientations than for economic left-right orientations. It is the affluent societies with well-developed welfare states that had the most postmaterialist and libertarian population, first and foremost Denmark, Sweden, and the Netherlands, and partly also West Germany, Iceland, and the United States. The most traditional or materialist orientations were found in Northern Ireland, France, Spain, and Ireland, and then in Belgium, Italy, and Great Britain. A major deviation from this pattern was the fairly strong support for both authoritarian and postmaterialist values in the United States. Value change towards the ‘new’ postmaterialist values varied between the countries as described above. A particular opposite trend was found in France, where the population became much less supportive for all these new values. This was most pronounced regarding authoritarian values. Cohort differences were fairly large on the new politics orientations. The post-war cohorts were really vanguards regarding these new values. Cohort differences were large and persistent for libertarian/authoritarian values. There was a small decline in cohort differences for materialist/postmaterialist orientations, probably due to a period effect causing a trend away from the new values. This trend was particularly strong in the post-war cohorts. In sum, the spread of postmaterial and libertarian values is caused by generational replacement, but period effects are also important, and these period effects are strongly caused by the level of economic and physical security. These factors will determine the destiny of these new values in the future. The research question on convergence or divergence of countries regarding value priorities was approached both by analyzing crossnational variations over time, and by comparing variations within the post-war and the war cohorts. The results of these analyses were
16 This is modified in the multivariate analysis for libertarian/authoritarian values, where both periods (1990 and 1999/2000) have slightly negative effects.
148
oddbjørn knutsen
fairly mixed. The analyses of variation over time showed trends towards convergence of pay equality and MPM, while the analyses of cohort differences showed convergence in the post-war cohorts only for equality/freedom. The conclusion is that there was no trend towards convergence on most of the measures of political values. In the political cultures of western democracies, the old politics and new politics orientations studied in this chapter are the central politically-relevant orientations that influence political attitudes, identities, and behavior. This chapter focused upon the trends in these orientations over time. Another very important research topic is to study the impact of these orientations on political attitudes, identities, and behavior over time. Do the new politics orientations increasingly supplant the impact of old politics or do these orientations only supplement the dominant impact of old politics? References Aalberg, T. 2003. Achieving Justice. Comparative Public Opinion on Income Distribution. Leiden: Brill. Abramson, P.R. & R. Inglehart 1995. Value Change in Global Perspective. Ann Arbor: The University of Michigan Press. Arts, W. & J. Gelissen 2001. Welfare states, solidarity and justice principles: does the type really matter? Acta Sociologica, 44, 283–299. Bean, C. & E. Papadakis 1998. A comparison of mass attitudes towards the welfare state in different institutional regimes, 1985–1990. International Journal of Public Opinion Research, 10, 211–236. Beck, P.A. 1974. A socialization theory of partisan realignment. In R. Niemi et al. (eds.), The Politics of Future Citizens. San Francisco: Jossey-Bass. Borre, O. & J.M. Viegas 1995. Government intervention in the economy. In O. Borre & E. Scarbrough (eds.), The Scope of Government. Volume 3 in “Beliefs in Government”. European Science Foundation. Oxford: Oxford University Press. Esping-Andersen, G. 1985. Politics Against Markets—The Social Democratic Road to Power. Princeton, NJ: Princeton University Press. ——. 1990. The Three Worlds of Welfare Capitalism. Oxford: Polity Press. ——. 1999. Social Foundations of Postindustrial Economies. Oxford: Oxford University Press. Flanagan, S.C. 1987. Value changes in industrial societies. American Political Science Review, 81, 1303–1319. Flanagan, S.C. & A.-R. Lee 1988. Explaining Value Change and its Political Implications in Eleven Advanced Industrial Democracies. Paper presented at the 14th World Congress of the International Political Science Association. Washington, August 28 – September 1. ——. 2003. The new politics, culture wars, and the authoritarian-libertarian value change in advanced industrial democracies. Comparative Political Studies, 36, 235–270. Inglehart, R. 1977. The Silent Revolution. Changing Values and Political Styles Among Western Publics. Princeton, NJ: Princeton University Press. ——. 1981. Post-materialism in an environment of insecurity. American Political Science Review, 74, 880–900.
the end of traditional political values?
149
——. 1984. The changing structure of political cleavages in Western society. In R.J. Dalton, S.C. Flanagan & P.A. Beck (eds.), Electoral Change in Advanced Industrial Democracies. Princeton, NJ: Princeton University Press. ——. 1990. Cultural Shift in Advanced Industrial Society. Princeton, NJ: Princeton University Press. ——. 1997. Modernization and Postmodernization. Cultural, Economic and Political Change in 43 societies. Princeton, NJ: Princeton University Press. Inglehart, R. & H.-D. Klingemann 1976. Party identification, ideological preference, and the left-right dimension among Western mass public. In I. Budge, I. Crewe & D. Farlie (eds.), Party Identification and Beyond. London: John Wiley. Iversen, T. & A. Wren 1998. Equality, employment, and budgetary restraint. The trilemma of the service economy. World Politics, 50, 507–546. Kelly, J. &. M.D.R. Evans 1993. The legitimation of inequality: occupation earnings in nine nations. American Journal of Sociology, 99, 75–125. King, D.S. 1987. The New Right. Politics, Markets and Citizenship. Basingstoke: Macmillan Education Ltd. Kitschelt, H. 1988. Left-libertarian parties: explaining innovation in competitive party systems. World Politics XL, 194–234. ——. 1994. The Transformation of European Social Democracy. Cambridge: Cambridge University Press. ——. 1995. The Radical Right in Western Europe. A Comparative Analysis. Ann Arbor: The University of Michigan Press. Knutsen, O. 1995a. Left-right materialist value orientations. In J.W. van Deth & E. Scarbrough (eds.), The Impact of Values. Volume 4 in “Beliefs in Government”. European Science Foundation. Oxford: Oxford University Press. ——. 1995b. Value orientations, political conflicts and left-right identification—a comparative study. European Journal of Political Research, 28, 63–93. ——. 1996. Value orientations and party choice: a comparative study of the relationship between five value orientations and voting intention in thirteen West European countries. In O.W. Gabriel & J.W. Falter (eds.), Wahlen und politische Einstellungen in westlichen Demokratien. Frankfurt/M.: Peter Lang. ——. 1997. The partisan and the value-based components of left-right self-placement: a comparative study. International Political Science Review, 18, 191–225. ——. 2003. Generations, Age Groups and Voting Behaviour in the Scandinavian Countries—a Comparative Study. Research Report No. 4. Oslo: Department of Political Science. Knutsen, O. & S. Kumlin 2005. Value orientations and party choice. In J. Thomassen (ed.), The European Voter. Oxford: Oxford University Press. Lafferty, W. & O. Knutsen 1984. Leftist and rightist ideology in a social democratic state: an analysis of Norway in the midst of the conservative resurgence. British Journal of Political Science, 14, 345–367. Levitas, R. (ed.) 1986. The Ideology of the New Right. Oxford: Polity Press. Lipset, S.M. & S. Rokkan 1967. Cleavage structure, party systems, and voter alignments: an introduction. In S.M. Lipset & S. Rokkan (eds.), Party Systems and Voter Alignments. New York: The Free Press. Middendorp, C.P. 1978. Progressiveness and Conservatism. The Fundamental Dimensions of Ideological Controversy and Their Relationship to Social Class. Berlin, New York: Mouton/ Walter De Gruyter. ——. 1991. Ideology in Dutch Politics. The Democratic System Reconsidered 1970–85. Assen & Maastricht: Van Gorcum. ——. 1993. Authoritarianism: Personality and ideology. European Journal of Political Research, 24, 211–228. Nie, N.H., S. Verba & J.R. Petrocik 1976. The Changing American Voter. Enl. edition. Cambridge: Harvard University Press.
150
oddbjørn knutsen
Roller, E. 1995. The welfare state: the equality dimension. In O. Borre & E. Scarbrough (eds.), The Scope of Government. Volume 3 in “Beliefs in Government”. European Science Foundation. Oxford: Oxford University Press. Svallfors, S. 1997. Worlds of welfare and attitudes to redistribution: a comparison of eight western nations. European Sociological Review, 13, 283–304.
CHAPTER SIX
IS THE WEST BECOMING MORE TOLERANT? Nina Rother and Juan Díez Medrano Conceptualizing and Explaining Tolerance Never before has there been in the Western world so much information available on other cultures and other societies. Rising levels of education among the population, the media revolution, and globalization are some of the broader processes that have contributed to this. The media are also responsible for our knowing more than people ever did about the diversity of condition and lifestyles in our own societies. One question we address in this chapter is what impact have greater levels of information and knowledge about other groups on values, attitudes, and behavior towards groups different from one’s own? Have they led to more tolerance towards these groups or have levels of tolerance, on the contrary, remained unchanged by these processes? Does globalization in the political, economic, and cultural domain result in a process of convergence of tolerance in the Western world or have national value differences remained the same or become even stronger? Theories of social change have always posited that the younger generations play a crucial role in value change. Is this also true when it comes to tolerance? Have the younger generations been more influenced by greater exposure to otherness than have other groups in society and thus led the way towards greater tolerance? Are younger generations consistently more tolerant than older ones? Before answering these questions, we have to take a closer look at the concept of tolerance. The concept of tolerance is a tricky one to define, and especially so when one treats it as a continuous variable rather than as a dichotomy. The Latin meaning of the word is the acceptance of something or someone we do not agree with. Tolerance results from self-discipline and carries with it the meaning of suffering. Other uses of this concept are discernable over history. In medieval times, for instance, a pragmatic-rational understanding of this concept developed, related to the idea of accommodating
152
nina rother and juan díez medrano
enemies that one cannot defeat but at a great cost. The Enlightenment witnessed the development of a liberal conception of tolerance, understood as acceptance of and respect for those who are different, in the sense of live and let live. Tolerance came also to be understood as a human right and as both a virtue and a duty in religious, social, and political life. Finally, a contemporary interpretation of the term associates to tolerance attributes such as solidarity and active engagement in improving the living conditions of those who are different (see Fetscher, 1990). In this chapter, we borrow from the definitions of tolerance above and understand it as the acceptance of behavior and thought one does not like or with which one does not agree (see Sullivan, Piereson & Marcus, 1979). The roots of this dislike may be diverse, although some authors have emphasized the role of perceived differences in attitudes between ego and alter (Rokeach, 1960). While the range of disliked but tolerated behavior and thinking is central in the determination of a person’s degree of tolerance, it is not its sole dimension. To accept the practice of prostitution, but only in clearly bounded geographical areas, is not the same as accepting prostitutes performing their profession next door. To accept certain disliked practices, but only on special occasions, is not the same as accepting them all year long. Time and place must enter the definition of tolerance. The concept of tolerance must thus refer to the readiness to accept disliked behavior and thought at all times and in all places. Understood as a continuous variable, tolerance is a matter of degree, which varies with the range of disliked behavior and thought that are accepted as legitimate and with the time and spatial restrictions one would like to impose on the expression of particular forms of behavior and thought. The multidimensional character of the concept of tolerance requires that one assigns weights to range, spatial constraints, and time constraints when determining at the abstract level how to rank individuals in terms of their degree of tolerance. We cannot think of theoretical grounds, however, that would justify assigning specific weights. Another definitional problem originates in the continuous character of the variable ‘dislike’. It does not reflect the same degree of tolerance to accept behavior that one mildly dislikes than as it does to accept behavior that one dislikes strongly. In order to rank-order individuals based on their degree of tolerance one would thus have to weight differently behaviors and thoughts depending on the degree of dislike individuals profess towards them.
is the west becoming more tolerant?
153
The concept of tolerance thus remains unspecified at the definitional level. It is easier to characterize the extremes of the distribution of this variable than to fill out the middle. This definitional problem has implications at the measurement level. If one wants to empirically measure the individuals’ level of tolerance, one can follow two routes that do not necessarily lead to the same results (see Peffley, Kingge & Hurwitz, 2001). One of them is to rely on self-assessments of tolerance. An appropriately worded question would ensure a similar understanding of the question by everybody through clearly spelling out how we, researchers, interpret the concept of tolerance and letting respondents mentally browse through their list of liked and disliked groups in order to answer our question. The problem with this approach would be that the answers would suffer from a social-desirability bias that would complicate cross-time and crosssociety comparisons to the extent that different societies and, most importantly, those groups with more resources and greater societal influence in these societies, value tolerance differently.1 In order to somewhat reduce this social-desirability bias one can assess the individuals’ tolerance levels indirectly, by asking them to rate the importance they assign to inculcating tolerance, as defined here, among the population. The second route one can follow in order to measure tolerance takes into account the fact that tolerance is more easily conceptualized at the extremes than at the middle of the distribution. One could ask people to express the degree to which they accept behavior and/or the expression of thoughts that they particularly dislike—or groups incurring in that behavior or those thoughts—in their close vicinity and at any time. Most empirical approaches to the problem of tolerance have followed one of the two routes outlined above, but only imperfectly, which explains why there is so much controversy about the obtained results. Thus, there is little agreement to date, for instance, on whether tolerance levels in the United States have increased or decreased in recent decades (see e.g., Mondak & Sanders (2003) for a review of this literature). This chapter is no exception to previous work’s reliance on imperfect measures. We first use a general and indirect measure
1 This effect is compounded by the more educated groups’ greater awareness of what is ‘politically correct’ at a given moment (e.g., Jackman, 1978 and Jackman & Muha, 1984).
154
nina rother and juan díez medrano
of tolerance, which is the respondents’ ranking of tolerance within a list of qualities that one could inculcate in a child. Although this question is not a perfect indicator of tolerance, it has two virtues: first, by focusing attention on qualities desired in a child rather than on one’s personality, the answers to this question will be less affected by social-desirability factors than would a direct question referring to the respondent’s tolerance; second, this question’s advantage lies in the fact that to the extent that individuals generally understand that acceptance of disliked groups is an essential dimension of the concept of tolerance, the answers will not be affected by the degree to which respondents like or dislike particular groups. The second indicator of tolerance we rely on in this chapter concerns the respondents’ unwillingness to accept members of a series of groups as neighbors. These groups include political extremists, foreigners, people with a criminal record, people of a different race, heavy drinkers, emotionally unstable people, and people with large families. This question’s virtue is that of focusing mainly on groups that suffer from prejudice and rejection in all the countries considered here (the exception being perhaps people with large families). The problem with it is that one lacks information on the relative degree of dislike towards these groups in different societies and among different groups within each of the countries included in the dataset. It will thus be impossible to disentangle whether differences between the answers reflect contrasts in the degree to which these groups are liked or disliked and contrasts in the degree of tolerance (on this topic, see Kuklinski, Riggle, Ottati, Schwartz & Wyer, 1991; Sorrentino & Higgins, 1986). This chapter focuses on trends in tolerance in different countries and in different groups within these countries. The main questions we try to answer are the extent to which processes of convergence are taking place and which role generations play within these processes. In order to formulate hypotheses with regard to both questions, one first needs to understand the causal factors underlying tolerance. At the most immediate level, one can list two factors that determine tolerance in general as well as tolerance towards particular groups. These two factors are the familiarity one has with them, and the degree of control individuals have over their lives. We understand familiarity as the degree to which individuals are aware of the existence of groups of people who behave or think differently and know these groups’ customs and ways of thinking. This familiarity may be
is the west becoming more tolerant?
155
rooted in direct contact (e.g., in the neighborhood, in the labor market) or, at the more cognitive level, in exposure through books and/or other media. We understand sense of control over one’s life in a material, a physical, and a symbolic dimension (e.g., self-image). We assume that individuals who are better off economically, who live in safe neighborhoods, and who enjoy high status in a society experience greater control over their lives than other groups in society. These two variables are related to each other and in some cases interact in influencing people’s degree of tolerance. Familiarity, for instance, interacts with control in influencing how much one tolerates a particular group. Familiarity with members of another group combined with lack of control may lead to less tolerance towards members of this other group, whereas familiarity combined with control may lead to more tolerance towards them. Familiarity and sense of control are in turn influenced by contextual factors. Here we will only focus on democracy, economic development, and globalization. Democracy can impact on levels of tolerance by facilitating the diffusion of frames that emphasize the values of democracy, freedom, and tolerance (see Martin & Westie, 1959; Simmons, 1965; on frames and political cues, see Zaller, 1992, and Nelson, Clawson & Oxley, 1997). Socialization in democracy also allows individuals to develop democratic values and to become familiar with a more diverse set of groups than socialization in nondemocratic societies (Kuklinski, Riggle, Ottati, Schwartz & Wyer, 1991). The longer one lives in a democracy, the longer one is exposed to these frames and socialization factors. Consequently, democracy impacts general levels of tolerance among the population directly and through the development of a greater familiarity with otherness in the population. It is possible, however, that a democratic context may lead to the development of less tolerance for political groups that threaten democracy (e.g., far right and far left political organizations). Economic development can also contribute to greater tolerance. This results from greater standards of living in the population, that foster a higher degree of physical and material security or sense of control, and from increasing average levels of education in the population, which enhances cognitive familiarity with otherness (e.g., through reading, through travel) (on the effects of education, see Bobo & Licari, 1989; Glock, Wuthnow, Piliavin & Spencer, 1975). Familiarity and sense of control are therefore the socio-psychological mechanisms
156
nina rother and juan díez medrano
that mediate the impact of economic development on levels of tolerance. Finally, globalization is the third macro-sociological variable that can have a decisive impact on levels of tolerance. Contrary to the effect of the other processes, however, globalization’s effect on tolerance is more ambiguous. On the one hand, globalization at the level of communication leads to increased virtual familiarity with populations around the world. At the same time, however, economic globalization, because of increased international competition, is leading to a major socio-economic reorganization of advanced economies, to increasing unemployment, increasing insecurity, and to rising inequality. We suspect that the net effect of this has been a loss of sense of control by large segments of the population, which can manifest itself in less tolerance. Furthermore, globalization has motivated a sudden immigration wave to advanced economies, just as these economies are facing difficulties in coping with competition by emerging economies in the East. This immigration has suddenly put autochthonous populations in contact with foreign groups. Let’s just think of a country like Spain that in less than fourteen years has moved from being an extremely homogeneous country to the current situation where about 8% of the population is foreign. In this short period, many segments of the population, namely the working classes (in terms of education, the least cognitively prepared groups), have not had time to develop familiarity with the incoming population while, simultaneously, immigrants threaten the sense of control of many groups in society. While democratic institutions have tried to counteract these forces that lead to less tolerance towards foreigners, they most likely have been more successful among the better-off segments of the population (employers of immigrants rather than competitors both in the labor and residential markets, and the more cognitively prepared for diversity) than among the worse-off ones. At any rate, the literature shows that individual-level factors are less important than aggregate-level ones in explaining prejudice towards immigrants. That is, people’s sense of control when faced with immigrants depends less on their individual characteristics than on comparisons between the group to which they feel they belong and the group or groups that they see as ‘foreign’. Thus, Quillian (1995) has demonstrated that in the European Union levels of prejudice towards immigrants depend on a country’s level of economic development and the percentage of foreign-born population. Similarly,
is the west becoming more tolerant?
157
Olzak (1995) has shown from a dynamic perspective that ethnic conflict is positively associated with immigration rates and rates of business failure. Finally, Díez Medrano (1994) has shown that the propensity to vote along ethnic lines among dominant ethnic groups increases with the percentage of immigrants in a community and with the degree to which immigrants compete with the dominant ethnic group in the labor market. The discussion above does not lead to the formulation of clear hypotheses about trends in general tolerance and in tolerance towards particular groups. This is because there are at the moment countervailing forces, some of them pushing towards greater tolerance and others pushing towards lesser tolerance. Generally speaking, we expect increases in general tolerance and convergence between countries because of increases in the number of democracies in the West, in the length of time Western populations have lived under democracy, in general development levels, and because democratic culture and globalization have most likely eroded tolerance only towards very specific groups. Trends in tolerance towards specific groups, on the other hand, should vary depending on the groups one is studying. The development of democratic values in stable democracies would lead to the expectation that tolerance towards political extremists has eroded over the years. It is unclear, however, whether this trend has implied convergence or not. It is conceivable, for instance, that the declining economic and political performance of Western democracies has somewhat eroded the faith in democracy and thus led to increased support for political extremists in those countries that have been hit hardest by the process of globalization. If this is the case, the prospects for convergence in tolerance of these groups between and within countries are not very bright. The analytical framework for the explanation of general and specific tolerance presented above also leads to the prediction of a net decline in tolerance towards foreigners. This is because the 1980–2000 period has seen a large influx of immigrants into Western countries. Western populations have thus been shaken in their sense of control (e.g., economic competition) and have not had time yet to become familiar with the cultures of the new immigrants. Finally, the analytical framework described above leads to the expectation that tolerance towards criminals has followed trends in crime rates in the 1980–2000 period, since crime rates are related to the population’s feeling of physical security. Neither the democratic tradition nor economic
158
nina rother and juan díez medrano
development imply that one should accept criminal behavior. Tolerating criminals makes sense only if one perceives that certain forms of behavior or thought are unjustifiably made illegal. Since authoritarian regimes are more repressive of behavior and thought than are democracies, one would therefore expect democracy to contribute to greater trust in criminal law and justice and to lower levels of tolerance towards criminals. Meanwhile, rising crime rates in the West (see International Learning and Information Network for Crime Prevention and Community Safety, 2004), coupled with the joint processes of globalization-induced immigration, and the growing criminalization of the immigrant population can have led to less tolerance of this group. The two processes combined may indeed compound the physical insecurity that individuals experience when they think of criminals and thus lead to people tolerate them less. Thus far, we have formulated hypotheses concerning trends in general and specific tolerance for the population in general. Based on the causal mechanisms discussed above, one would expect cohort effects to complement period effects. Since each new generation since World War II has been socialized in a more democratic and wealthier society, we would expect younger generations to be more tolerant than older generations. The baby bust generation has shared in these changes but is also the first one to grow up in a world experiencing the communications revolution and the impact of economic globalization. This generation has been more exposed to cultural diversity than any generation before. At the same time, however, its sense of economic security and, by implication, sense of control, have been shaken by growing unemployment, lessened job security, lowered prospects for upward social mobility, and growing inequalities. It is thus difficult to make predictions about this generation’s levels of tolerance relative to the preceding ones. We would expect, however, that the uncertainties created by economic globalization in the baby bust generation express themselves in targeted intolerance toward groups seen as responsible for these uncertainties. In sum, we predict that general tolerance is greater among each succeeding generation, that tolerance towards social deviants, such as criminals and political extremists, has declined in each succeeding generation, and that, countering a generational trend towards increasing tolerance, tolerance towards foreigners is lower among the baby bust generation than among the baby boom generation.
is the west becoming more tolerant?
159
Methods and data To address changes of tolerance in the Western world, we use the three waves of the European Value Studies (EVS) as in nearly all the other chapters of this book. These data refer to three years (1981, 1990, and 1999/2000), spanning two decades, and encompass a large number of Western countries, including the United States. One focus of this chapter is to contrast changes in tolerance in European societies and in the United States. Consequently, only those Western countries with data available for all three waves will be included. These countries are: Belgium, Denmark, France, Ireland, Italy, the Netherlands, Spain, Sweden, Great Britain, West Germany, and the United States. Although the United States did not participate in the 1999/2000 EVS, comparisons are made possible with the World Value Studies (WVS) dataset of 2000. One shortcoming of the EVS/WVS is that only a small number of questions related to tolerance were asked in all three waves. To address the importance of general tolerance in comparison to other values, the EVS/WVS includes a ranking question that asks respondents to choose among the items in a list at most five qualities which children should be encouraged to learn at home. The list contains values like ‘good manners’, ‘independence’, ‘hard work’, ‘feeling of responsibility’, ‘imagination’, ‘religious faith’, ‘unselfishness’, ‘obedience’, and—most important for this chapter—the item ‘tolerance and respect for other people’.2 To analyze tolerance towards specific groups at the extremes of the distribution, a second item battery is included in the EVS/WVS, the so-called ‘neighbor-question’. This question asks the respondent to again sort out of a list those groups that the person would not want to have as a neighbor. Since the lists of groups of people used in the different waves were not identical across the waves and countries,3
2 Unfortunately, the advantage of this question, namely that it sorts out the most important values, so that we know how much importance respondents attach to tolerance relative to other values, also entails a disadvantage: the other items of the list may influence the selection of the tolerance item. This is problematic when the list changes over time, as it did from 1981 to 1990, with the number of items dropping from 17 to 11. 3 The number of neighbor groups also changed over the three waves. But since the groups that can be selected is not restricted, the impact of the context of the list is less dramatic than in the former question.
160
nina rother and juan díez medrano
only items asked in all three waves can be compared. The eight groups included are ‘people with a criminal record’, ‘people with a different race’, ‘heavy drinkers’, ‘emotionally unstable people’, ‘immigrants/foreign workers’, ‘left wing extremists’, ‘right wing extremists’, and ‘people with large families’. In order to facilitate the analyses and to obtain a clearer picture of change in levels of tolerance of specific neighbors, clusters of neighbor groups have been created.4 Factor analyses with target rotation showed that a 3-dimensional model describes the data best. The first dimension of tolerance can be labeled as tolerance towards ‘social deviants’, it comprises the group of people with a criminal record, heavy drinkers, and emotionally unstable people. The second dimension, tolerance towards foreigners, includes people of a different race, immigrants/foreign workers, and large families. Lastly, the third dimension of tolerance, tolerance towards political extremists, consists of the two items referring to left and right wing extremists. For a better interpretation, we computed additive indicators for these three dimensions instead of factor scores. Each of the indicators ranges from 0 to 1, where 0 indicates tolerance and 1 stands for intolerance. At the individual level, the indicator measures the percentage of items in a particular dimension chosen by the respondents. At the country-level, the mean scores of the indicators denote how many of the items were chosen on average in a given country, e.g., a mean score of .15 indicates that on average 15% of the corresponding items were chosen. In sum, higher values for these additive indicators reflect lower levels of tolerance of particular clusters of groups. Findings The availability of three EVS/WVS waves allows us to address three issues: changes in levels of tolerance, convergence in levels of tolerance, and generational contrasts in tolerance. We are interested in answering questions such as whether the modernization of Western world societies results in a systematic decline of support for toler-
4
By combining different neighbor groups into broader categories, we reduce somewhat the effect that differences in how much people like certain groups have on country comparisons.
is the west becoming more tolerant?
161
ance, whether globalization of the political, economic, and cultural domains results in convergence in levels of tolerance in the Western world, and whether younger generations are systematically more tolerant than are older ones. We follow a two-step approach to answer these questions: first, we analyze the item that measures general levels of tolerance; second, we examine tolerance of specific categories of groups. A general decline of tolerance? Do the simultaneous phenomena of democracy, economic development, and globalization result in systematic decline in tolerance? As discussed above, we expect a general increase in tolerance together with contrasting trends in tolerance towards ‘social deviants’, foreigners, and political extremists. Table 6.1 gives an overview of trends in general tolerance for all eleven countries taken together as well as for each country treated separately. As expected, the percentage of respondents who indicate that tolerance is among the five most important qualities that children should be encouraged to learn at home, has continuously risen, from 54.0% in 1981, to 77.6% in 1990, and up to 82.6% in 1999/2000.5 Rising levels of general tolerance do not necessarily translate into greater tolerance of particular groups of individuals. Table 6.2 displays general and country trends in means and standard deviations for the tolerance indicators that refer to three particular groups of potential neighbors. The higher the tolerance scores in this table, the lower the tolerance level. As expected, tolerance of ‘social deviants’, ‘foreigners’, and ‘political extremists’ has declined in the 1980–2000 period, as expressed in higher scores for the specific group tolerance indicator. First, tolerance towards ‘social deviants’ shows a strong decline from 1981 to 1990, followed by a smaller increase from 1990 to 1999/2000. Since the standard deviations increased only slightly, one can speak of an actual
5 It could be argued that the increase of tolerance from 1981 to 1990 is only due to a reduction of the number of items in the questionnaire. However, since the lists used in 1990 and 1999/2000 are identical, we think that the first increase is not only due to this reduction and that there is indeed a trend towards greater tolerance.
162
nina rother and juan díez medrano Table 6.1: General tolerance
West Germany Italy Spain Belgium Ireland Denmark France Netherlands Great Britain Sweden United States
1981
Rank
1990
Rank
1999/ 2000
Rank
43.1% 43.3% 44.2% 45.2% 56.0% 58.4% 58.9% 59.4% 61.9% 71.1% 52.4%
1 2 3 4 5 6 7 8 9 10 4a
76.9% 67.0% 75.0% 67.5% 76.4% 80.9% 78.3% 88.5% 79.6% 90.8% 72.4%
5 1 3 2 4 8 6 9 7 10 2a
71.5% 75.0% 82.4% 84.6% 76.6% 87.3% 84.7% 91.5% 83.0% 92.3% 79.8%
1 2 4 6 3 8 7 9 5 10 3a
Mean all countries 54.0% Std. dev. all countries 9.15
77.6% 7.48
82.6% 6.56
trend towards more intolerance of this group. Tolerance towards foreigners also diminished from 1981 to 1990, but less so than did tolerance of ‘social deviants’, and then did not change between 1990 and 1999/2000. The increase one observes in the standard deviation for this indicator means that divergence rather than convergence in tolerance of foreigners took place in this period. Finally, we observe that tolerance towards political extremists has tended to decline over the years. Western populations are now less tolerant towards this group than they were twenty years ago, especially with respect to right wing extremists. The standard deviations have also risen slightly, which indicates a trend towards divergence in tolerance of political extremists in the Western world. A comparison of levels of tolerance towards the three groups of neighbors shows that tolerance towards foreigners is and has always been highest. But whereas in 1981 and 1990, the level of tolerance towards political extremists was significantly higher than tolerance towards ‘social deviants’, tolerance of these two classes of people has converged over the years. As the discussion in the introduction led us to expect, we find contradictory trends in levels of tolerance. General tolerance, probably the best of our two measures, has increased over the years. Meanwhile, tolerance towards specific groups has generally declined. Tolerance towards
Diff. 99–81
1981
Foreigners 1990 1999/ 2000 Diff. 99–81
1981
Political extremists 1990 1999 Diff. 2000/ 99–81
0.25 0.18 0.21 0.33 0.36 0.28 0.31 0.28 0.38 0.40 0.49
0.34 0.24 0.29 0.39 0.45 0.36 0.34 0.32 0.39 0.41 0.51
0.31 0.27 0.30 0.40 0.42 0.39 0.32 0.28 0.46 0.36 0.54
0.06 0.09 0.09 0.07 0.06 0.11 0.01 0.00 0.08 –0.04 0.05
0.10 0.06 0.05 0.05 0.07 0.10 0.05 0.05 0.11 0.13 0.08
0.16 0.07 0.10 0.05 0.13 0.08 0.08 0.07 0.10 0.12 0.09
0.13 0.08 0.10 0.12 0.15 0.06 0.10 0.03 0.13 0.06 0.09
0.03 0.02 0.05 0.07 0.08 –0.04 0.05 –0.02 0.02 –0.07 0.01
0.18 0.05 0.11 0.19 0.37 0.35 0.25 0.19 0.24 0.46 0.26
0.37 0.06 0.28 0.25 0.31 0.50 0.26 0.27 0.30 0.56 0.30
0.41 0.15 0.34 0.33 0.29 0.59 0.27 0.36 0.29 0.64 n.a.
0.23 0.10 0.23 0.14 –0.08 0.24 0.02 0.17 0.05 0.18 n.a.
.33 (.33) .37 (.33) .36 (.34) .03 (.01) .07 (.18) .10 (.23) .10 (.25) .03 (.07) .25 (.39) .33 (.43) .36 (.42) .11 (.03)
Social Deviants 1990 1999/ 2000
High values indicate intolerance, low values tolerance.
Belgium Denmark France Ireland Italy Netherlands Spain Sweden Great Britain West Germany United States
All countries
1981
Table 6.2: Intolerance towards specific groups (means and standard deviations in brackets)
is the west becoming more tolerant? 163
164
nina rother and juan díez medrano
foreigners is and has always been highest but declined from 1981 to 1990 and then remained stable. Tolerance towards ‘social deviants’ increased slightly and then considerably decreased. Finally, tolerance towards political extremists was relatively high in 1981 but has declined quite strongly since. Convergence of tolerance in the Western world? We now focus on whether the last two decades have witnessed a process of convergence or divergence in levels of tolerance in the Western world. First, we focus on general levels of tolerance. The results of the statistical analysis, displayed in the upper part of Table 6.1, show a pattern which is very similar to that observed in the lowe;r part of the table. Tolerance was lowest in 1981 for all countries and rose in all countries except in West Germany in the 1990–1999/2000 period. The increase in the 1990–1999/2000 period, however, was not as steep as in the earlier one. A cross-country comparison shows that Denmark, France, the Netherlands, Sweden, and Great Britain (the latter with the exception of 1999/2000) are the countries that attach the highest importance to tolerance. West Germany, Italy, and Spain are among the countries that attach less importance to tolerance relative to other values (but still show very high values of tolerance). The position of the United States relative to all European countries leads us to assume that no divergence has taken place. The ranking of each of the countries in terms of general tolerance remains quite stable over time. Similar patterns in all countries are an indicator of convergence but one can only speak of ‘real’ convergence if we see that the scores on general tolerance for the different countries converge to a particular value. The results presented in Figure 6.1 show that convergence has indeed taken place. There still remains a range of almost 20 percent between the most tolerant and intolerant country, but the trend towards convergence is inescapable. One can also see that only few lines are crossing, which means that the positions of single countries in comparison to the other countries have remained stable. The means and between-country standard deviations reported in the last two lines of Table 6.1 further support the hypothesis of a gradual convergence.
is the west becoming more tolerant?
165
Means General Tolerance
0,9
0,8
0,7
0,6
0,5
0,4 1981
1990 WAVE
1999/2000
Belgium Denmark France Ireland Italy Netherlands Spain Sweden Great Britain USA West Germany
Figure 6.1: General tolerance—development over time across countries
Although there is convergence among Western countries with regard to general tolerance, there is no evidence that tolerance towards specific groups has also converged. Table 6.2 provides data concerning trends in tolerance for the three types of groups that we are examining in this chapter. The same data are presented in a more interpretable fashion in figure form. With regard to tolerance towards ‘social deviants’, the pattern that emerges in Figure 6.2 is rather clear. From 1981 to 1990 tolerance decreased in all countries. From 1990 to 1999/2000, the countries split and moved in two different directions. Whereas in Belgium, Italy, Spain, Sweden, and West Germany tolerance levels increased again, they continued to decrease in Denmark, France, Ireland, the Netherlands, and Great Britain. When we focus on the 1980–2000 period as a whole, however, we see that tolerance of ‘social deviants’ declined slightly in all countries except for West Germany. The positions of the different countries nevertheless remained quite stable, with Denmark the most tolerant European country and Italy, Great Britain, and West Germany the most intolerant European countries.
166
nina rother and juan díez medrano
0,55
Mean Index ‘Social Deviants’
0,50 0,45 0,40 Belgium Denmark France Ireland Italy Netherlands Spain Sweden Great Britain USA West Germany
0,35 0,30 0,25 0,20 0,15 1981
1990 WAVE
1999/2000
Figure 6.2: Intolerance towards ‘social deviants’—development over time across countries
The question of whether there is convergence of European countries towards a similar level of tolerance of ‘social deviants’ is harder to answer. Although they still differ considerably in 1999/2000, the range and the between-country standard deviations have become somewhat smaller, which one can interpret as a trend towards convergence. The United States is the only exception to this trend, for it displays a much greater level of intolerance of ‘social deviants’ than do the other countries. One could then conclude that only European convergence is taking place. The second indicator, tolerance towards foreigners, shows no simple and straightforward pattern of change. Intolerance has certainly increased in most countries examined here. The only exceptions are West Germany, Sweden, and the Netherlands. The crossing lines in Figure 6.3 reveal, however, that the pace of this trend differs between countries, with a steady decrease of tolerance in Spain, Denmark, and Italy, a steady increase in West Germany and the Netherlands, a combination of decrease in the 1980s and increase in the 1990s in Belgium, France, Sweden, and the United States and a combination of increase in the 1980s and decrease in the 1990s in Ireland and Great Britain. Resulting from these contrasting patterns of change the relative positions of the countries have not remained stable.
is the west becoming more tolerant?
167
0,16
Mean Index ‘Social Foreigners
0,14 0,12 0,10 Belgium Denmark France Ireland Italy Netherlands Spain Sweden Great Britain USA West Germany
0,08 0,06 0,04 0,02 1981
1990 WAVE
1999/2000
Figure 6.3: Intolerance towards foreigners—development over time across countries
Trends in the pooled-sample standard deviation in tolerance of foreigners show that Western populations have not become more alike with respect to their tolerance of foreigners. Figure 6.3 shows that this divergence has resulted in part from a slight increase in betweencountry variance. As discussed in the first section of the chapter, a possible reason for this divergence in tolerance of foreigners may be that globalization, in the form of immigration from poor countries, has had a different impact in different countries and at different times. A slightly different picture of change in tolerance levels emerges when analyzing tolerance towards political extremists, as in Figure 6.4. Intolerance has increased in all countries, except for Italy, between 1981 and 2000. The relative positions of different countries have remained stable, with West Germany and the Netherlands as the most intolerant countries and Denmark the most tolerant one. When one focuses on all countries taken together, the range for the level of tolerance and the between-country standard deviations increased slightly in the 1980–2000 period. If one ignores the three ‘extreme’ countries (West Germany, the Netherlands, and Denmark), however, one can conclude that a clear convergence has taken place for the
nina rother and juan díez medrano
168
Mean Index Political Extremists
0,6 0,5 0,4 Belgium Denmark France Ireland Italy Netherlands Spain Sweden Great Britain USA West Germany
0,3 0,2 0,1 0,0 1981
1990 WAVE
1999/2000
Figure 6.4: Intolerance towards political extremists—development over time across countries
remaining European countries between 1981 and 1999/2000. Although data for the United States are unavailable for the 1990–1999/2000 period, trends in the 1980s fit in this trend towards convergence. One can thus conclude that overall trends in tolerance of different groups are consistent with our expectations that intolerance would increase for all three groups examined here. As expected, trends in tolerance of each of these groups follow their own logic, so that one cannot speak of overall convergence or divergence in tolerance. These findings are consistent with our hypothesis that tolerance of specific groups results from the interplay of democracy, economic development, and globalization. Further research, with more countries and a greater number of time points, should make it possible to test the relative impact of democracy, levels of economic development, and globalization on these trends. Are younger generations more tolerant? In this last section, we investigate contrasts between generations in levels of tolerance. Are there consistent intergenerational differences in tolerance levels throughout the Western world, with younger gen-
is the west becoming more tolerant?
169
erations consistently being more tolerant than older generations? Going back to the first part of this chapter, we may expect this to be more the case with respect to general levels of tolerance but not necessarily with respect to tolerance of specific groups. To investigate generational contrasts in tolerance, we distinguish three cohorts: the war generation (born before 1940), the baby boom generation (born between 1940 and 1960), and the baby bust generation (born after 1960). These three groups are distinct from each other with regard to their experiences in their most formative years. The war generation experienced the Second World War as well as its aftermath. The baby boom generation grew up at a time of cultural and political change in the 1960s and 1970s. Lastly, the baby bust generation has grown up amidst a serious economic crisis and significant structural and political developments. Statistical analysis of the EVS data shows that between 1981 and 2000 overall tolerance has increased for all cohorts, whereas tolerance of social deviants, foreigners, and political extremists has decreased across all cohorts. Contrary to what we expected, however, general tolerance levels are not highest in the youngest generation—the baby busters— but rather in the middle generation, the baby boom generation. The upper part of Table 6.3 shows that, especially in 1981 and 1990, the respondents born between 1940 and 1960 chose tolerance as an important education goal more often than did respondents born before 1940 or after 1960. Only in 1999/2000 did the baby bust generation reach the baby boom generation’s level of tolerance. The findings above apply to almost all of the Western world countries. Whereas in Great Britain and the United States the war generation seems to be as tolerant or even more tolerant than the baby boom and especially the baby bust generation, the opposite is true for Spain. Here, the youngest generation is the most tolerant one, a circumstance that one can explain with reference to the socializing impact of Franco’s dictatorship on the older generations. Although the greater level of tolerance found among the baby boom generation than among the baby bust generations is not what one would have expected, it is not entirely out of line with our theoretical discussion in the introduction. As we suggest there, it is possible that the growing insecurity brought by this initial phase of the globalization process may have countered the trend to tolerance promoted by the democratic tradition and rising general levels of education among the population.
170
nina rother and juan díez medrano Table 6.3: General tolerance by generation and wave
Total
war generation baby boom generation baby bust generation Belgium war generation baby boom generation baby bust generation Denmark war generation baby boom generation baby bust generation France war generation baby boom generation baby bust generation Ireland war generation baby boom generation baby bust generation Italy war generation baby boom generation baby bust generation Netherlands war generation baby boom generation baby bust generation Spain war generation baby boom generation baby bust generation Sweden war generation baby boom generation baby bust generation Great Britain war generation baby boom generation baby bust generation West Germany war generation baby boom generation baby bust generation United States war generation baby boom generation baby bust generation
1981
1990
1999/2000
50.2% 56.3% 50.2% 43.8% 47.8% 40.5% 54.2% 60.8% 62.1% 59.8% 58.8% 57.3% 53.3% 61.5% 49.0% 38.3% 46.8% 45.0% 52.3% 65.6% 60.0% 44.7% 44.9% 41.2% 68.5% 74.6% 69.4% 64.0% 63.9% 50.0% 36.0% 50.5% 49.6% 50.6% 55.7% 46.6%
73.5% 78.3% 74.4% 66.3% 69.0% 67.6% 77.9% 85.2% 77.7% 75.6% 79.1% 80.6% 74.7% 78.1% 76.1% 62.7% 73.2% 65.0% 83.9% 93.0% 86.7% 72.5% 76.4% 76.6% 89.0% 94.0% 88.1% 79.0% 81.9% 77.1% 74.0% 79.2% 77.9% 74.1% 74.3% 64.1%
78.0% 83.7% 82.9% 81.5% 87.3% 84.2% 77.0% 90.7% 89.3% 84.2% 84.7% 85.3% 71.3% 79.3% 78.8% 70.7% 75.6% 77.4% 91.0% 89.0% 93.6% 75.5% 79.8% 83.1% 91.7% 93.0% 91.9% 84.5% 85.7% 82.8% 63.3% 79.7% 73.1% 85.1% 80.3% 76.6%
Table 6.4 again demonstrates that one should not expect similar findings when examining overall measures of tolerance and tolerance towards specific groups. When we focus on the latter, we find that regardless of the type of social group being considered, the youngest gen-
is the west becoming more tolerant?
171
eration appears to be the more tolerant one in every wave. This completely contradicts our expectations, for we expected that the younger generations would be the standard-bearers for the trends observed in the population as a whole. Since tolerance for all groups has declined in the past two decades, we would have expected the younger generations to have been less tolerant of these groups than have the older ones. This prediction applied especially to tolerance of foreigners, since it is the youngest generation that has experienced most the employment crisis and declining social mobility opportunities of the last couple of decades. Table 6.4: Intolerance towards specific groups by generation and wave
1981
war generation baby boom gen. baby bust gen. 1990 war generation baby boom gen. baby bust gen. 1999/ war gen. 2000 baby boom gen. baby bust gen.
Mean Mean Mean Mean Mean Mean Mean Mean Mean
(sd.) (sd.) (sd.) (sd.) (sd.) (sd.) (sd.) (sd.) (sd.)
Social deviants
Foreigners
Political extremists
.37 .30 .28 .43 .35 .34 .42 .36 .33
.10 .06 .05 .14 .08 .08 .15 .10 .08
.29 .22 .18 .38 .32 .28 .41 .36 .33
(.34) (.32) (.30) (.35) (.33) (.32) (.36) (.34) (.33)
(.21) (.16) (.15) (.27) (.22) (.21) (.29) (.24) (.22)
(.41) (.37) (.34) (.45) (.42) (.40) (.45) (.42) (.41)
Not only are the mean scores smallest for the youngest generation but also the standard deviations. It is often said that polarization of (traditional) values is high among younger people. But the opposite appears to be true with respect to tolerance of various social groups: younger people are more homogeneous with respect to tolerance than are members of older generations. Apart from a few exceptions, the observation that younger generations are more tolerant of various social groups than are older ones holds at the country level, as shown in Figures 6.5, 6.6, and 6.7. Sweden is one such exception. Here, the baby bust generation is less tolerant than the baby boom generation over all waves and across all three neighbor groups. Tolerance of social deviants and foreigners is also higher among the baby boomers in Denmark and France in 1981 and 1990. In these three countries, we do not observe inconsistency between general tolerance and tolerance of specific social groups.
nina rother and juan díez medrano
172 0,65
Mean Index ‘Social Deviants’
0,55
0,45 Belgium Denmark France Ireland Italy Netherlands Spain Sweden Great Britain USA West Germany
0,35
0,25
0,15
War
Boom Bust 1981
War
Boom Bust 1990
War Boom Bust 1999/2000
Figure 6.5: Intolerance towards ‘social deviants’ by generation, wave, and country 0,215 0,190
Mean Index Foreigners
0,165 0,140 0,115
Belgium Denmark France Ireland Italy Netherlands Spain Sweden Great Britain USA West Germany
0,090 0,065 0,040 0,015
War
Boom Bust War 1981
Boom Bust War Boom Bust 1990 1999/2000
Figure 6.6: Intolerance towards foreigners by generation, wave, and country
is the west becoming more tolerant?
173
0,7
Mean Index Political Extremists
0,6
0,5
0,4 Belgium Denmark France Ireland Italy Netherlands Spain Sweden Great Britain USA West Germany
0,3
0,2
0,1
0,0
War
Boom Bust War 1981
Boom Bust War Boom Bust 1990 1999/2000
Figure 6.7: Intolerance towards political extremists by generation, wave, and country
The question we address now is whether these generational differences remain when other independent variables such as gender, marital status, and religiosity are taken into account.6 The results of the three regression analyses with tolerance of specific groups as dependent variables depicted in Table 6.5 show that the effects of the country dummies are stronger than the effects of the background variables. Although significant due to the huge number of cases, the standardized regression coefficients for sex, marital status, employment status, religion, and church attendance are quite small. Consequently, the explained variance does not even reach 10% in any of the three regression analyses. As a proxy for period effects we included the year of the survey in the analysis, with the 1981 wave as the baseline category in the first model. A significant time effect can be seen in all three regression
6 Unfortunately, the impact of education and socio-economic status on the level of tolerance cannot be investigated because these variables have not been measured consistently across all three waves.
174
nina rother and juan díez medrano
analyses: both in 1990 and in 1999/2000 intolerance is greater than in 1981. If these effects resulted from a generational effect, we would expect them to diminish or even disappear when the generation dummy variables are included in the analysis, as we do in our second model. As expected, we find that the youngest generation is the most tolerant one, especially in comparison to the war generation. The inclusion of generations does not make period effects disappear, however. Instead, they even increase slightly, which reflects that trends in tolerance towards specific groups do not simply reflect generational replacement. Table 6.5: Regression of intolerance towards ‘social deviants’, foreigners and political extremists on country dummies, sex, marital status, employment status, religion, church attendance, waves, and generations (beta-coefficients) Social deviants Belgium Denmark Ireland Italy Netherlands Spain Sweden Great Britain West Germany United States sex divorced, widowed or separated never married never child employed no religion Protestant other religion church rarely or never church occasionally church once a month wave 1990 wave 1999
Foreigners .07* –.02* –.02* .05* .00 –.02* –.03* .03* .03* .01
.02* –.04* .06* .16* .05* .05* .01 .10* .09* .21*
.02* –.04* .06* .12* .05* .05* .01 .10* .09* .21*
.02* .02*
.03* .01
–.04* –.03* .04* .02*
–.08* –.07* .00 –.01*
–.05* –.01* –.03* –.03* .01* –.01* –.02* –.01 –.07*
–.04* –.01 –.01 –.02* .01 –.01* –.01 .00 –.06*
–.03* –.02* .00 .02* –.06* –.03* .00 .00 .01 .00 –.02* –.02* .00 .01* –.01 .00 –.01 .01
–.04* –.02* –.01* –.01* .02* .01 –.02* –.01 –.05*
–.02* –.01 .01 –.01* .02* .01* –.01* –.00 –.03*
.06* .06*
.07* .08*
.08* .12*
.09* .14*
.04* .05*
.07* –.02* –.01 .06* .00) –.01 –.03* .03* .03* .01*
Political extremists
.05* .07*
.05* –.12* –.02* .03* .12* –.01 –.01 –.00 .19* n.a.
.05* –.12* –.02* .03* .12* –.01 –.01* –.00 .19* n.a.
is the west becoming more tolerant?
175
Table 6.5 (cont.)
war generation boom generation Adjusted R square N
.068
.07* –.01*
.10* .02*
.072
.021 .027
.08* .02* .089
.093
48365 48365 48365 48365 45957 45957
Baseline or omitted categories are: France, male, currently having partner, not employed, Catholics, weekly church attendance, wave 1981, baby bust generation. * = significant on .05-level. High values indicate intolerance.
Conclusion Tolerance is a well-established value in modern Western societies. The literature suggests that democracy and high levels of economic development are the main factors underlying this. Our findings for trends in overall tolerance show that convergence is taking place across Western societies and that the process is still on-going, with younger generations being more tolerant than older ones. Our chapter also demonstrates that general trends in tolerance do not necessarily correspond to trends in tolerance of specific groups. Thus, while we observe a trend towards greater overall tolerance in Western societies, the opposite is true when we focus on social deviants, political extremists, and foreigners. There is no inconsistency here: tolerance levels result from multivariate processes of which democracy and levels of economic development are just some of the most significant underlying causes. The literature offers other examples where general attitudes do not correspond to specific ones.7 Other findings further suggest the presence of additional factors that we have not been able to test with the data at hand. Contrary to expectations, for instance, we have found that the baby boom generation shows greater levels of general tolerance than does the baby bust one. Also, and countering overall trends in tolerance toward specific groups, younger generations are more tolerant of these groups than
7 For instance, Bobo (1983) and Bobo-Kluegel (1993) argue that general attitudes towards equality among the US population do not correspond to attitudes towards race-targeted policies aimed at reducing racial inequality.
176
nina rother and juan díez medrano
are older ones. While one can to some extent account for the former unexpected finding through invoking the ‘revolutionary’ character of the baby boom generation, the latter unexpected findings are more puzzling. It may be, for instance, that age effects enter the picture here, with the findings simply telling us that young people are more tolerant of marginalized groups than are older people. The data at our disposal do not allow us, however, to disentangle cohort, age, and period effects in a satisfactory way. This chapter has first developed an analytical framework for the explanation of tolerance. This framework highlights the structural impact of democracy, economic development, and globalization as it is mediated by people’s sense of familiarity with particular groups and the citizens’ sense of control over their environment. Based on this model, we formulated hypotheses concerning trends in tolerance, as measured by two different constructs, general tolerance and tolerance of specific groups. Our expectations to find an increase in general tolerance as well as convergence in the Western world as resulting from the simultaneous phenomena of democracy, economic development, and globalization have been endorsed. General tolerance has increased over time and convergence across the Western world has taken place. Our expectations to find declining levels of tolerance towards social deviants, political extremists, and foreigners, possibly resulting from higher crime rates, a longer democratic tradition, and large foreign inflows respectively have also been confirmed. As expected, we have found that, consistent with the uneven trends in physical and economic security in different countries and with different immigration rates and immigration timing, one cannot speak of a clear convergence or divergence process in tolerance of specific groups in the Western world. Finally, our empirical findings with respect to generational contrasts are somewhat at odds with our predictions and call for further research. Beyond analyzing the degree of fit between findings and theoretical expectations, one can also examine the empirical findings described above from the perspective of recent discussions about the development of a growing value divide between the United States and Europe. Our study shows that, when examined in terms of tolerance levels, this divide does not exist. Generally speaking, convergence, not divergence, is the case here.
is the west becoming more tolerant?
177
In sum, one should move away from simplistic assumptions concerning linear trends towards greater tolerance in the West. Tolerance is not as high as one might have expected, nor is there an increase for all types of tolerance and all countries. We would like to posit that globalization in its multiple forms has come to disturb what may have been thought of as a steady trend in levels of tolerance throughout the Western world, although previous research in the United States alone already warned us against developing a very optimistic outlook. If globalization continues to weaken the people’s sense of control over their lives and economic development, even a decrease of tolerance in the future is conceivable. References Bobo, L. 1983. Whites’ opposition to busing: symbolic racism or realistic group conflict? Journal of Personality and Group Psychology, 45, 1196–1210. Bobo, L. & J. Kluegel 1993. Opposition to race targetting: self interest, stratification ideology, or racial attitudes? American Sociological Review, 58, 443–464. Bobo, L. & F. Licari 1989. Testing the effects of cognitive sophistication and target group affect. The Public Opinion Quarterly, 53, 285–308. Díez Medrano, J. 1994. The effects of ethnic competition and ethnic segregation on political mobilization in the Basque Country, 1988. American Sociological Review, 59, 873–889. Fetscher, I. 1990. Toleranz. Von der Unentbehrlichkeit einer kleinen Tugend für die Demokratie. Historische Rückblicke und aktuelle Probleme. Stuttgart: Radius-Verlag. Glock, Ch., R. Wuthnow, J. Piliavin & M. Spencer 1975. Adolescent Prejudice. New York: Harper and Row. International Learning and Information Network for Crime Prevention and Community Safety 2004. Crime prevention information and news. In Preventing Crime and Creating Safer Communities. University of the West of England. Web: http://environment.uwe.ac.uk/commsafe/eusor3.asp Jackman, M. 1978. General and applied tolerance: does education increase commitment to racial integration? American Journal of Political Science, 22, 302–324. Jackman, M. & M. Muha 1984. Education and intergroup attitudes: moral enlightenment, superficial democratic commitment, or ideological refinement? American Sociological Review, 49, 751–769. Kuklinski, J., E. Riggle, V. Ottati, N. Schwartz & R. Wyer 1991. The cognitive and affective basis of political tolerance judgments. American Journal of Political Science, 35, 1–27. Martin, J. & F. Westie 1959. The tolerant personality. American Sociological Review, 24, 521–528. Mondak, J.J. & M. Sanders 2003. Tolerance and intolerance, 1976–1998. American Journal of Political Science, 47, 492–502. Nelson, T., R. Clawson & Z. Oxley 1997. Media framing of a civil liberties conflict and its effects on tolerance. American Political Science Review, 91, 567–584. Olzak, S. 1995. The Dynamics of Ethnic Competition and Conflict. Stanford: Stanford University Press. Peffley, M., P. Knigge & J. Hurwitz 2001. A multiple values model of political tolerance. Political Research Quarterly, 54, 379–406.
178
nina rother and juan díez medrano
Quillian, L. 1995. Prejudice as a response to perceived group threat: population composition and anti-immigrant and racial prejudice in Europe. American Sociological Review, 60, 586–611. Rokeach, M. 1960. The Open and the Closed Mind. New York: Basic Books. Simmons, J.L. 1965. Tolerance of divergent attitudes. Social Forces, 43, 347–352. Sorrentino, R. & E. Higgins 1986. Motivation and cognition: warming to synergism. In R. Sorrentino & E. Higgins (eds.), The Handbook of Motivation and Cognition. Foundations of Social Behavior. New York: Guilford. Sullivan, J., J. Piereson & G. Marcus 1979. An alternative conceptualization of political tolerance: illusory increases, 1950s–1970s. American Political Science Review, 73, 781–794. Zaller, J. 1992. The Nature and Origins of Mass Opinion. Cambridge, UK: Cambridge University Press.
CHAPTER SEVEN
IS VOLUNTEERING GOING DOWN? Paul Dekker and Andries Van den Broek Introduction: Volunteering as a reflection of pro-social values Volunteering has attracted widespread public and political attention in recent years. In many countries it was debated how to preserve and encourage volunteering. A number of parliamentary and government commissions studied ways to stimulate voluntary activities among various groups, such as the young, the elderly, working parents, and immigrants.1 The United Nations Year of Volunteers in 2001 inspired all kinds of national and local manifestations, discussions, and policy initiatives, often attracting considerable media exposure.2 All over the world, volunteering is seen as highly important for the proper functioning of society. In fact, volunteering is considered to be among the key components of civil society, because it is held to generate social cohesion and societal self-regulation as well as to strengthen political democracy by developing individual citizenship and organizing countervailing powers. Issues such as these have gained momentum in recent years, especially since Putnam’s publications in the 1990s on civic community and democracy in Italy and on the decline of social capital in the United States (Putnam, 1993, 2000; cf. Dekker & Van den Broek, 2005). Volunteering has come to serve as an indicator of the (negative) trends in the literature on the (assumed) decline of civic community, crumbling of civil
1 Examples are the German Parliamentary Committee on Civic Engagement (Enquete-Kommission 2002) and the British Russell Commission on the engagement of new young volunteers (http://www.russellcommission.org/). 2 For an overview of activities and initiatives in and since the UN year, see http://www.worldvolunteerweb.org/browse/countries.html. For a comparative summary of volunteering and projects, partly related to the UN year, in 21 countries in 2001, see Govaart et al. (2001).
180
paul dekker and andries van den broek
society, and erosion of social capital but also as a possible instrument for recovery. Addressing volunteering suggests something quite deviant, as volunteering concerns behavior rather than values. The latter is indeed partly the case but is that really so deviant and so different? As in all other chapters, we use results from population surveys. In a way, the only difference is that part of our data does not consist of reported opinions but of reported behavior. Yet, we hold, behavior too may be interpreted to reflect values. There are, of course, a few more steps from values to behavior than there are from values to opinions. This, however, is a matter of degree rather than of principle. In between values and behavior, the restrictions of the available opportunity structure have to be taken into account: not all people who value volunteering positively may have had the incentive or opportunity to act on them, because they were not invited or lacked the time to do so. Besides, we do not focus on volunteering solely but connect it to values in other life domains, thus integrating this chapter into this book more tightly than might appear to be the case at first glance. Had a repeated measurement of values on volunteering been available, we would have surely used that as the focus of our contribution. Regrettably, such a measurement is not available in the EVS/WVS data. With an eye to the importance attached to volunteering as one of the key concepts within the notions of civil society and public virtue, about which much of the dispute on the quality of contemporary society is concentrated, this set-back in terms of available data should not, we think, stop us from trying to make the data ‘talk’ about volunteering. Generational succession has been argued by various observers (e.g., Putnam, 2000) to propel the decline of civil society and public virtue. The general nature of the processes identified as driving forces behind the drop in volunteering suggest some degree of international applicability and thus of convergence. Therefore, the three issues central to this book are akin to the core of the debate on volunteering: is there a trend toward less volunteering, does that trend unfold internationally, and is generational succession at work in propelling this trend? In definitions of volunteering or voluntary work, four common elements can be found: it is non-obligatory, it is carried out for the benefit of others (society as a whole, or a specific organization), it is unpaid, and it takes place in an organized context (Dingle, 2001: 9).
is volunteering going down?
181
The criterion of being non-obligatory is essential but often not evident. There are forms of voluntary work that can hardly be refused, such as community services to enable students to obtain credits for their study or to gain the necessary experience to find a job, or voluntary work to help reintegrate the long-term unemployed on the labor market, possibly with a financial incentive. The criterion of being unpaid is also not entirely straightforward. From indisputable reimbursements of expenses and generally accepted small material tributes of appreciation, it is only a small step to the acceptance of payments below the market value and a reformulation of the criterion of the work being unpaid to the work not being undertaken primarily for financial gain. The requirement of an organizational setting is probably the most controversial criterion. Many definitions do not include such a norm, and ‘unorganized’ informal volunteering is sometimes explicitly acknowledged.3 Why do people volunteer? Some explanations emphasize chances and circumstances, the basic idea being that people tend to do things because of where they are, not who they are (Parker, 2000). Parents are supposed to do some voluntary work for the school or sports club of their child, there is some pressing problem in the neighborhood and if you do not do anything, nobody else will, etc. Other explanations focus on differences between volunteers and non-volunteers, referring to differences in personality (some people are simply more helpful, active, and generous than others; cf. Bales, 1996) or more specifically to their socialization, cultural background, and also their deeply held convictions or values (Dekker & Halman, 2003). A good explanation of specific phenomena of voluntary action and work combines elements of these approaches. In a general analysis of numbers of volunteers and of differences between volunteers and non-volunteers it makes sense to focus on volunteering as an expression of pro-social values.4
3
Some definitions are more restricted, for instance with the requirement that a ‘public good’ is produced (probably excluding voluntary work for one’s own church or sports club) or that there is no connection with the volunteer’s paid work (Wilson, 2000; Dingle, 2001; Govaart et al., 2001). 4 Turning the relationships around, volunteering may also influence an individual’s values, and the level of volunteering may also have an impact on the values in a society. Whereas there are negative effects at the individual level—such as negative experiences of being exploited—positive effects are probably dominant. Positive
182
paul dekker and andries van den broek
How different are the values of volunteers? In a review, Smith (1994) concludes that studies seeking to explain volunteering with values find only weak correlations. Wilson (2000: 219) concludes from a number of studies that “overall, the relation between values and volunteering is weak and inconsistent”. However, in an extensive analysis of a long list of values and beliefs in a national survey of 2,300 Canadians, Reed and Selbee (2003) do find a limited number of remarkable differences with respect to certain values and ideas. They describe the characteristic value profile of volunteers as “a syndrome of generosity mixed with civic engagement and concern for the common good” (Reed & Selbee, 2003: 103). Values are probably more important for understanding the kind of volunteering people do than for predicting whether people volunteer. Kearney (2001) suggests that the kind of voluntary work people are engaged in depends to a considerable extent on values. Values can manifest themselves in motivations to volunteer. Altruism, solidarity, reciprocity, beneficence, justice are values/motives often related to volunteering (Dekker & Halman, 2003). Value changes or culture shifts will have consequences for volunteering. Individualization is probably the most important value change currently affecting volunteering. In his ‘Acts of compassion’, with the revealing subtitle ‘Caring for others and helping ourselves’, Wuthnow (1991: 22) concludes that “. . . being intensely committed to self-realization and material pleasure did not seem to be incompatible with doing volunteer work. . . . people who were the most individualistic were also the most likely to value doing things to help others.” A similar conclusion is also drawn in research on (post-) modernized and individualized ‘new volunteers’ (Hustinx, 2001). What might change, however, is the way people volunteer. Putnam (2000: 127 ff.) notes a steadily declining participation by Americans in recent decades in ‘community projects’, whereas they report doing more voluntary work. Apparently, people see voluntary work “as providing personal rather than community service”. The growing amount of voluntary work is increasingly being mobilized outside the con-
individual experiences of cooperation and gratefulness and the broader acknowledgement of the worth of volunteering strengthen pro-social values. Thus volunteering and values may reinforce each other, as they will be mutually dependent on each other (cf. the ‘virtuous circles’ of involvement and values in Putnam’s (1993) study of civic community in Italy).
is volunteering going down?
183
texts of clubs, churches, and other associations. Volunteering, according to Wuthnow, Putnam, and others, is more and more about ‘getting things done’ by individuals. The social aspects recede into the background and cooperation comes to depend on the task to be carried out and is no longer the core of the activity. In the following sections, we explain what indicators we use in analyzing volunteering. Then, we describe trends in volunteering, assess whether international trends point in the same direction, perhaps even are converging to similar levels of volunteering across the board (a between-countries comparison), and assess whether a generational pattern in the distribution of volunteering can be discerned (a within-countries comparison). In doing so, the questions central to this book are answered. To connect the issue of volunteering more closely to the theme of values, we then explore if and how volunteering is related to values in a number of life domains. Our rationale and plan for doing so are explained in that later section, first we turn to the analysis of the three questions on trends, convergence, and generations central to this book. Data and indicators The main data sources used here are the European Values Studies (EVS) of 1981, 1990, and 1999, complemented by 2000 data for the United States (and partly for Spain) with data from the World Values Studies (WVS). We confine our analyses to ten nations in Europe and the United States, for reasons of comparability through time.5 Dichotomies were constructed as indicators of our central concepts.6 Volunteering was measured in the surveys by the following question: “Please look carefully at the following list of voluntary organizations and activities and say which, if any, do you belong to?; . . . and which, if any, are you currently doing unpaid voluntary work 5
See Curtis et al. (1992) for an analysis of the 1981 data, Curtis et al. (2001) for the 1990 data, and Hodgkinson (2003) and Inglehart (2003) for the 1999 data. These authors all focus on patterns of membership or volunteering and on explanations of differences between countries. 6 In some cases variables are dichotomies, in other cases distributions are so skewed that dichotomies are the best representation. For the remaining indicators we prefer dichotomies to ordinal measurements for reasons of simplicity, in particular for a convenient presentation of interaction effects.
184
paul dekker and andries van den broek
for?”. Regrettably, those lists differ somewhat between the years (Dekker & Van den Broek, 2005). In this analysis, we use the largest set of organizations (and ‘activities’) common to all years of measurement and countries. Participation in organizations differs enormously, either between countries or within a country with the passing of time. Huge differences between countries may well, of course, correctly represent ‘real life’ differences in volunteering, but may also have to do with item-wording carrying different connotations in different countries. Huge fluctuations in a country between years make one suspicious about possible changes in the perceptions of what constitutes volunteering. As to the latter, suspiciously huge fluctuations occurred in volunteering in social and youth work associations in Great Britain (6% and 3% in 1981, 5% and 3% in 1990, and 13% and 15% in 1999; see Appendix Table 1). The formalization of parental assistance in the sports clubs of their children, for instance, might lead to higher numbers of volunteers. In the case of volunteering in religious associations in Sweden (9% in 1981, 3% in 1990, and 23% in 1999), rewording the survey question appeared to be responsible for the sudden growth in 2000.7 We decided to keep Sweden in the comparison and to skip religious volunteering in all countries. Several other fields and organizations were left out because of ‘slight’ alterations in item wording (see Dekker & Van den Broek (2005) for the full list). Our measurement of volunteering in this chapter refers to volunteering in one or more organizations in the following fields: social welfare, cultural activities, labor unions, political parties, human rights, environment and conservation, animal rights, professional associations, and youth work (see Appendix Table 1). Our strict selection of common fields in the three EVS/WVS waves in all countries leaves out some major fields of voluntary work, such as sports and religion. This seriously detracts from the richness of the EVS/WVS data and may in part explain the fact that the nation figures used in this chapter differ from those from other surveys. In Appendix Table 2, our 1999 figures are compared with those from similar measurements in the 1998/9 survey
7 In 1981 and 1990, the Swedes were asked about volunteering in the ‘free churches’ (not belonging to the Lutheran state church), in 2000, the year the state church was abolished, they were asked in the same way as the respondents in other countries, about a ‘church or religious organization’.
is volunteering going down?
185
of the International Social Survey Program (ISSP) and the 2002/3 survey of the European Social Survey (ESS). The levels of volunteering measured as well as the differences between countries both vary markedly from one survey to the other. There are constants, such as high levels of volunteering in Sweden and the Netherlands versus low levels in Spain, and Italy, but correlations of merely 0.6 between the three measurements of levels of volunteering in ten cases are worrying. Besides the different selections of organizations, the variations probably have to do with variations in item wording and with differences in the context in which the item was raised. In the ISSP questionnaire, for instance, the item follows questions on charity. Moreover, some of the sharp fluctuations within the EVS/WVS data raise some suspicions too. More work has to be done here, but for the time being the EVS/WVS data, after selections as made by us, are the only source of longitudinal international data on volunteering available. With respect to the issue of convergence, nations are used as the unit of analysis. When looking at the effects of generational succession, three generations are demarcated along the lines used throughout this volume, i.e. distinguishing a war generation (born before 1940), a baby boom generation (born between 1940 and 1960), and a baby bust generation (born after 1960). When connecting volunteering to values held with regard to other domains in life, we first use the answers to a general question, the opening question of the questionnaire, asking about the importance of the following six domains: work, family, friends and acquaintances, leisure time, politics, and religion. We then look closer at the relation of volunteering to work values and to leisure values. 1981–1999: On trends, convergence, and generations Trends Contrary to what might be expected on the basis of gloomy accounts of the descent of civil society and civic virtues, the EVS/WVS data do not point to a general decrease in the level of volunteering (Table 7.1). Instead, the general picture is one of rising levels of volunteering (Dekker & Van den Broek, 2005), as summarized in the line depicting the trend in average country percentages (giving each country an equal weight). That average was at the same level in 1990 as it was in 1981, but increased by about a third in 1999.
186
paul dekker and andries van den broek
Table 7.1: Volunteering 1981–1999: doing unpaid work for at least one of eight types of organizations (in %) 1981
1990
1999
Trend
Sweden Denmark Germany (West) Netherlands Belgium France Italy Spain Ireland Great Britain United States
19 16 17 19 18 15 14 18 18 16 19
22 13 14 22 19 16 13 7 17 13 25
33 21 8 32 22 14 17 9 15 36 47
+ + – + 0 0 0 – 0 + +
Average of country percentages
17
16
23
+
Results weighted for standard socio-demographic variables (see Appendix Table 1)
Beyond the overall stability in the average of country percentages between 1981 and 1990, comparing the various country percentages of people reporting volunteering reveals various fluctuations. Yet, as there were roughly equal increases and decreases, there was no shift across the board. Between 1990 and 1999, larger shifts occurred, increases taking the upper hand over decreases. The increases in the levels of volunteering in six countries outweigh the four countries where the level of volunteering in 1999 did not differ from that in 1981 (Belgium, France, Italy, and Ireland) and the two countries ((former) West Germany and Spain) where the ranks of volunteers became thinner. It is striking that the five countries where volunteering was on the rise are predominantly Protestant, while most of the six countries where volunteering remained stable or declined are Catholic. Convergence If increases in volunteering had taken place in countries where volunteering was low at the outset, and if it had slipped back in countries where it was at an initial high, those combined trends would have added up to a convergence. The trend, however, is in the reverse direction of divergence. Country averages moved away from being clustered around a common level of volunteering to a more diverse patterning with levels of volunteering lying further apart. This
is volunteering going down?
187
Table 7.2: Summary statistics on volunteering, 1981–1999 1981 Standard deviation of country percentages Difference between highest and lowest percentage
1.7 5
1990
1999
Trend
5.2
12.4
+
39
+
18
weighted results
can be seen from country scores in Table 7.1 and is summarized in Table 7.2. The standard deviation of national percentages was relatively small in 1981, all countries reporting 14% to 19% of volunteers. Differences grew by 1990: the standard deviation almost tripled, as did the difference between highest and lowest country scores. By 1999, both the standard deviation and the difference between extremes doubled once more. Apparently, then, country levels of volunteering are not influenced predominantly by global developments. In contrast, national differences and national explanations seem to be of growing importance. Generations Among the mechanisms that propel change in values, attitudes, and behavior is the continuous succession of generations. We call it a mechanism, as it not so much explains the origin of possible change, but reveals a way in which such change spreads. If change is spurred by generation replacement, the task still is to explain why generations came to differ from one another in the first place. Mannheim (1928/29) deserves credit for having given the first systematic description of this mechanism. The basic idea is that people carry the lifelong mark of their formative experiences. If more recent birth cohorts acquire a certain outlook on life, that outlook will gain weight in society at large because more recent cohorts gradually replace their predecessors. Among the better known contemporary applications of that mechanism are notions that a silent revolution (Inglehart, 1977) or a collapse of community (Putnam, 2000) are propelled by the succession of generations. Taking one’s clues from pessimistic readings of present-day developments as offered, among others by Lane (2000) and Putnam (2000), one would have thought that our task here would be to determine
188
paul dekker and andries van den broek
to what extent generation replacement contributes to the fall in volunteering. Given the fact that such a decline did not at all show up in the EVS/WVS data, our task now becomes to determine what role, if any, generation replacement plays in what did in fact happen. The first step to be taken is to determine whether generations differ in the extent to which their members volunteer, the second step is to assess whether and how such differences have an impact on overall changes over the years. The first was achieved by comparing generations, adjusted for both age and period (year of measurement). Especially controlling for age is important, as otherwise it is not clear whether differences between groups defined by year of birth refer to generational characteristics or reflect differences in age (i.e. life cycle). Secondly, to get an idea of the impact of generational replacement, the differences between years are compared both without and with controlling for generation and age (Table 7.3). The unadjusted differences are the same as those reported earlier, this time presented in terms of odds ratio’s (the greater the deviation from one, the greater the effect) rather than as percentages. The adjusted differences give the effect of year adjusted for belonging to a certain generation and being of a certain age. Since populations as such hardly differ in average age, those adjusted differences are in effect adjusted mainly for the effect of differences in the generational structure of populations in various years. If changes between years were largely caused by generations, the adjusted effects of year should at least be a lot smaller than the unadjusted effects. What do the data tell us? Do generations differ? And do generational differences have an impact? As shown in the first columns of Table 7.3, the most recent ‘bust generation’ differs in the extent to which its members are involved in volunteering especially from its immediate predecessor, the ‘baby boom generation’, in eight of the eleven countries investigated. The differences with the earlier ‘war generation’ are markedly fewer, though in the same direction. The direction of those differences is in line with expectations as phrased by Putnam (2000) and others: the members of the latest generation reveal lower levels of volunteering. This appears to confirm the notion that civil society and civic virtues, if measured in terms of volunteering, are less well off in the hands of the latest generation. With generational differences in the ‘expected direction’, i.e. the latest generation volunteering less, the mechanism that the baby bust
is volunteering going down?
189
generation gradually replaces the war generation could indeed have been the driving force behind the expected decline in volunteering. Since, however, as seen above, such a decline in volunteering did not take place, the picture is a radically different one. Whereas the mechanism of generation replacement inserted a modest downward influence on volunteering, levels of volunteering rose nonetheless (as shown again in the third and fourth columns of Table 7.3). Controlling for generation and age hardly alters the trends over the years (the fifth and sixth columns of Table 7.3). Where some difference catches the eye, it is of the kind that adjusted differences between the years are somewhat larger than the unadjusted differences. This is the case in Denmark, the Netherlands, Great Britain, and the United States (note that this observation is based on an eye-ball test, no tests of significance of differences between adjusted and unadjusted effects of year were conducted). This leads to two observations. First, the ‘period effect’ towards an increase in volunteering was of much more importance than the reverse influence inserted by the process of generation replacement. Second, without the wheels of generation replacement slowing it down a little, an even larger increase in volunteering would have resulted. Table 7.3: Generational differences in volunteering and the extent to which they explain differences between years (odds ratios) Generation (reference category = bust) adjusted for year and age
Sweden Denmark Germany (West) The Netherlands Belgium France Italy Spain Ireland Great Britain United States
Year (reference year = 1981) unadjusted
adjusted for generation and age
War
Baby boom
1990
1999
1990
1999
1.47 2.30** 1.62 2.72*** 1.13 2.64** 1.39 1.12 2.23** 1.84* 1.29
1.33 1.98*** 1.37* 2.00*** 1.18 2.10*** 1.48** 1.18 1.70** 1.86*** 1.49***
1.22 0.81 0.83* 1.16 1.11 1.02 0.91 0.34*** 0.96 0.82 1.42***
2.08*** 1.37** 0.46*** 1.94*** 1.31** 0.87 1.21* 0.48*** 0.82 2.91*** 3.68***
1.31* 0.93 0.89 1.40** 1.13 1.21 0.98 0.35*** 1.12 0.92 1.46***
2.36*** 1.85*** 0.54*** 2.76*** 1.39** 1.19 1.41** 0.51*** 1.09 3.66*** 4.11***
Significance: * < .05, ** < .01, *** < .001
190
paul dekker and andries van den broek
One has to bear in mind that it takes an enormous generation effect to either counterbalance or explain a period effect of some weight. In the nine year period between 1990 and 1999, the mechanism of a later generation replacing an earlier one applies to roughly oneseventh of the total adult population. To give an example, all other things being equal, for a 5% increase in volunteering to have been brought about by generation replacement, the percentage of volunteers in the youngest generation had to be 35 percentage points above that of the generation that sadly passed away. Similarly, it takes substantial generational differences for the mechanism of generation replacement to be able to counterbalance a mild period effect. As illustrated by this simple calculation, the mechanism of generation replacement can at best bring about piecemeal short-term change at the aggregate level. Only in the long run and in the absence of period effects can it be decisive (cf. Van den Broek, 1999). Volunteering and values in various life domains One need not be on the lookout for too long to come across worried accounts of where society is heading. Community is getting lost, as are civic virtues. Volunteering is one aspect of the alleged decline of qualities one fears will be dearly missed. A common thread in this line of thinking is that of a tension between volunteering and various life domains. The more people focus on their career, their family, and their leisure time, the less they are focused on volunteering. Not necessarily because they oppose volunteering but rather, and possibly more deeply disturbing for those who value volunteering, subconsciously: volunteering may simply no longer be part of the ‘rationale’ of people’s lives and may consequently evaporate. After having answered the three questions central to this book above, we now turn to exploring how volunteering relates to evaluations regarding certain life domains. This explorative part of our contribution consists of two steps. First, the relation between volunteering and the importance attached to a number of life domains will be assessed. This analysis addresses the importance of such domains rather than the values held in those domains. Secondly, to make up for this, volunteering is related to substantive values in the domains of work and leisure. The reason for focusing on these two life domains is that especially the importance attached to these two
is volunteering going down?
191
might be at odds with volunteering. The more people invest in work and leisure, and identify themselves in terms of work and leisure, a first and admittedly rough hypothesis runs, the less inclined they might be to volunteer. Reversely, as will also be addressed below, volunteering may be related to how work and leisure are evaluated and experienced rather than to how important they are held to be. First, however, the more general inroad via the importance of various life domains is taken. The importance of life domains Since 1990, the opening question of the EVS/WVS questionnaire asks about the importance of a number of life domains. As 1981 is missing and as we wish to explore possible correlations rather than trends, we here focus on the 1999 measurements only. The following six life domains were asked about in all countries: work, family, friends and acquaintances, leisure time, politics, and religion. The results of this explorative analysis seem to allow for certain general, though tentative, observations (Table 7.4). Across the board, whether one volunteers turns out to be related only to the importance one attaches to leisure.8 A higher appreciation of leisure time goes hand in hand with a higher propensity to be involved in volunteering. This holds in all countries but the Netherlands.9 On the other hand, the importance of work, friends and acquaintances, and politics are hardly related to volunteering at all. The importance of work is related to volunteering only once: a negative relationship in Spain. It is unclear here whether what is being referred to relates to having a job in the first place or to the quality of work
8 Note that the phrasing of the question makes it uncertain whether respondents are supposed to think of their amount of leisure time or of the way they spend their leisure time. 9 However, a similar analysis with data from the first wave of the European Social Survey yields quite different results. Here ‘leisure time’ shows a positive relationship with volunteering in Sweden and Ireland, but a negative relationship in Germany and Italy! The life domains are exactly the same in both surveys, but the measurement of their importance differs. As shown in Appendix Table 2 the measurements of volunteering differ as well. Further analyses are needed to find out why there are different results in different countries and differences between both surveys.
192
paul dekker and andries van den broek
(note: the extrinsic and intrinsic values of work are related to volunteering in the next subsection). The importance of friends is related negatively to volunteering in one country (Germany). The absence of any positive relationship with politics raises doubts about the (Neo-)Tocquevillian ideas about the politicizing experiences of social involvement.10 Table 7.4: Relation between volunteering and the importance attached to six life domains, 1999 (adjusted odds ratios)
Sweden Denmark Germany (West) The Netherlands Belgium France Italy Spain Ireland Great Britain United States
Work
Family
Friends and Leisure acquaintances time
Politics Religion
1.02 0.99 1.09 1.10 0.93 1.10 0.95 0.85* 1.23 1.10 0.99
1.14 1.04 1.26 1.22** 1.29*** 1.16 1.19* 1.23* 1.09 0.99 1.29***
0.99 0.94 0.64*** 0.94 1.03 0.90 1.10 1.05 1.20 1.17 0.96
1.03 1.03 1.15 0.99 0.98 1.00 1.01 1.03 1.10 1.06 0.99
1.21** 1.54*** 1.43** 1.14 1.47*** 1.33*** 1.33*** 1.51*** 1.32** 1.28** 1.23***
1.31*** 1.25** 1.18 1.19* 1.07 1.18* 0.89 0.98 1.04 1.68*** 1.13*
Statistical effects of 4-points assessments ‘not at all important’ → ‘very important’, adjusted for effects of the other five assessments (z-scores; all assessment scales are standardized per country to adjust for differences in general reply tendencies); significance: * < .05, ** < .01, *** < .001.
A middle position is taken by the life domains of the family (related to volunteering in five countries) and of religion (related in seven countries). In all these cases, attaching more importance to the family and/or religion goes together with a higher propensity to volunteer. As regards the family, this finding might be related to the fact that volunteering is often related to activities in the family, such as
10 See Putnam (1993, 2000) for empirical evidence and a theoretical account of a positive relationship between voluntary social involvement and political involvement in the tradition of De Tocqueville, and see Eliasoph (1998) and Lichterman (2005) for evidence of a negative relationship in American volunteers’ groups and the social mechanisms of ‘avoiding politics’ behind it. The overall statistical result of contradictory mechanisms and of common backgrounds of various forms of voluntary involvement are positive, but rather weak relationships between volunteering and political participation in Western Europe and the United States (Van Deth, 1997; Dekker, 2000).
is volunteering going down?
193
volunteering for sports clubs and leisure activities of children. As regards religion, much earlier research has shown strong relationships between belonging to a church and doing voluntary work. Opinions differ, however, as to how this relationship should be interpreted. Some authors have stressed values. Wuthnow (1991) and Wilson & Janoski (1995), for example, see the strength and content of religious beliefs as a strong motivating factor. Others have stressed the importance of church attendance. They believe ties with networks of other church members to be the decisive variable, not the content of religious beliefs (Putnam, 2000; De Hart, 2001; Bekkers, 2004; De Hart & Dekker, 2005).11 Work values This subsection builds on the data presented in the chapter on work values by Ester, Braun, and Vinken above. Table 7.5 presents the statistical relationships of volunteering with two outspoken representatives of instrumental work values and two representatives of intrinsic work values are related to volunteering. Though by no means true for all four variables in all countries, the outcomes suggest some, albeit not very strong, relation between volunteering and work values along the lines that people who view their work instrumentally are less inclined to volunteer, while people who value their work for intrinsic reasons are more inclined to do so. The aspects ‘good pay’ and ‘an opportunity to use initiative’ most clearly fit this pattern. The fact that these two work values have countervailing ‘effects’ on volunteering goes some way to explaining why the importance of work was unrelated to volunteering in Table 7.4.
11 Of course, this opposition of religious beliefs versus participation is somewhat artificial. Church attendance might be more important than religious affiliation as such or the support for statements about divinity or religious commandments, church attendance is also more important for volunteering than attending schools, workplaces, and leisure organizations and being involved in their social networks. Evidently, altruistic messages that are delivered from the pulpit, collecting money for the poor, institutional involvement in improving living conditions, etc. make a difference.
194
paul dekker and andries van den broek
Table 7.5: Relation between volunteering and work values, 1999 (adjusted odds ratios)
Sweden Denmark Germany (West) The Netherlands Belgium France Italy Spain Ireland Great Britain United States
Good pay
Good hours
A job that is interesting
Opportunity to use initiative
0.68** 0.95 0.56* 0.87 0.64** 0.53*** 0.81 0.80 0.56* 1.03 0.79
0.71* 0.88 1.33 1.13 1.00 0.68* 0.89 0.87 0.80 0.69* 0.84
0.93 1.01 0.99 0.88 1.14 1.57* 1.40 1.67** 1.33 0.97 1.43*
1.77*** 1.22 1.37 1.35 1.98*** 1.39 1.45** 2.24*** 1.81* 1.65** 1.11
Statistical effects of indicating aspects of a job as being important for oneself (dichotomies), adjusted for the effects of the other aspects; significance: * < .05, ** < .01, *** < .001; weighted results
Leisure values It is surprising if not strange that the EVS/WVS and other questionnaires generally devote much attention to values in the domains of work, religion, and politics, while, despite their richness, lacking an effort to measure values regarding leisure or consumption. This is surprising for various reasons. For one, the weekly hours of leisure time (not even taking holidays into account) far exceed both weekly working hours as well as the time spent on housekeeping chores in the course of a week, at least in the Netherlands (Van den Broek & Breedveld, 2004). Religion, politics, and child-rearing, other domains often asked about in surveys, also involve much less time than leisure. Although, of course, the amount of time spent on an activity does not directly indicate its importance, it is strange that leisure, second to sleep only in terms of the time involved, is so out of focus. Secondly, leisure is of growing importance as a marker of identity and well-being. Many sociologists claim that leisure has come to rival, if not surpass, work as the central organizing point of identity and well-being (Bocock, 1993; Featherstone, 1991; Giddens, 1991). This is illustrated by a study of personal advertisements in Dutch newspapers in the 1950s, 1970s, and 1990s, which reveals a major shift in the characteristics that people use to position themselves on the ‘relation’ market in the Netherlands. In the 1950s people chose
is volunteering going down?
195
to define themselves along financial, professional, and religious lines. By the 1970s more personal achievements such as level and direction of education were added. However, in the 1990s the content of personal advertisements had changed markedly, people presenting themselves by referring to their leisure activities (Zeegers, 1998). Finally, evidence from within the EVS/WVS survey itself points out the importance of leisure. As was brought to light by answers to the EVS/WVS question used in Table 7.4 above, leisure in the Netherlands has surpassed work as of perceived importance for life, and is more important than religion and politics (SCP, 2004). An increased focus on one’s leisure might well be detrimental to volunteering, as suggested above. Yet, on the relation between leisure values and volunteering, no international data are available. As a result, we here have to rely on a study in one country in one year: the Netherlands in 2004. In order to get an angle on the criteria which people use to value their leisure-time activities, respondents were asked to react to a number of possible evaluations. Lacking examples for doing so, the EVS/WVS set of items about work (see Table 7.5) was transcribed from the domain of work to that of leisure. An attempt to reduce the 13 items thus acquired to a smaller set of dimensions (through factor analysis) resulted in four such dimensions, which were used to structure Table 7.6. Of course, such an exercise only yields dimensions potentially included in the items analyzed, so if an aspect of leisure was overlooked in transcribing the items from the domain of work, that aspect does not show up as one of the dimensions either. Three of the four most highly valued aspects of leisure constitute the dimension of freedom: not being under pressure, doing interesting things, and being able to do what one feels like doing. These aspects of leisure find support within all generations (or age groups, see note in Table 7.6). The dimension of development is the second most supported one: being able to develop yourself, doing something you are good at, achieving something, and learning something new. This dimension clearly is related to generation, being valued more highly in the more recent generations, the earlier generations thinking it less important. The social dimension—meeting nice people, doing something useful, taking initiative—is considered about equally important as the dimension of development. This social dimension finds roughly equal support among the different generations.
196
paul dekker and andries van den broek
The hedonism dimension—taking it easy, shopping, not having to take account of others—is supported least of all. This holds for each generation. At the item level, the earlier generation attaches less importance to doing nothing, the recent generations favoring more strongly not having to take account of others. That freedom and development show up among the strongly supported leisure values seems to point at an ethos of individualism and self-realization in the leisure domain. This may be at odds with volunteering. About as highly valued as the development dimension is the social dimension, however, while the passive hedonism dimension is by far the least valued. This, in turn, seems conducive to volunteering. Table 7.6: Leisure values: their valuations by the Dutch population aged 16 and over by three generations (mean scores, 0 = very unimportant, 100 = very important) and their relationships with volunteering
Freedom dimension Not being under pressure Doing interesting things Being able to do what one feels like doing Development dimension Being able to develop yourself Doing something you are good at Achieving something Learning something new Social dimension Meeting nice people Taking initiative Doing something useful for society Hedonism dimension Shopping Taking it easy Not having to take account of others
All
Generations war boom
bust
Effects on volunteering unadjusted adjusted
75 72 76
74 67 77
76 73 76
74 73 77
0.97 1.00 0.86**
0.96 1.00 0,91
70
57
68
75
1.00
0.89
73
71
72
74
1.05
1.06
61 66
47 53
57 64
68 71
0.86** 0.97
0.84** 0.97
69 65 64
68 59 64
68 64 66
69 67 63
1.13** 1.10* 1.84***
1.08 1.06 1.92***
49 48 43
52 42 48
47 47 44
50 51 41
0.79*** 0.76*** 0.78***
0.79*** 0.82*** 0.94
Odds ratios, both unadjusted and adjusted for the effects of the other valuations; the same generations as in Table 7.3 are distinguished, but it should be noted that based on a one-shot survey, it is by definition undecided whether differences by year of birth should be interpreted in terms of generations (i.e. referring to socializing experiences) or in terms of age groups (i.e. referring to life-cycle stages). Source: SCP Future survey 2004, weighted results
is volunteering going down?
197
The relation between leisure values and volunteering can be assessed directly by determining the effect of support for those values on the (self-reported) involvement in volunteering (right columns in table). Other than might be expected, volunteering hardly is positively related to the social dimension of valuing leisure, at least not after controlling for background variables. Only the ‘doing something useful for society’-item is, hardly surprising, clearly related to volunteering. There is only little difference in the level of support for this item across the three generations. The freedom and development dimensions are largely unrelated to volunteering. This runs counter to the interpretation that, at an intra-individual level, an individualist and self-realization oriented ethos corrodes volunteering. That higher levels of support for these dimensions do not coincide with lower levels of volunteering, is in line with the earlier observation that individualists are not lagging behind in terms of volunteering (Wuthnow, 1991). Only the passive and consumerist hedonism dimension clearly is negatively related to volunteering, but this is the least highly valued of the leisure values. Summary, discussion, and conclusion This chapter set out to provide two things: an answer to the central questions in this volume with respect to volunteering, and an exploration of the relation between volunteering and values in other life domains. With respect to the former objective, analyzing of the EVS/WVS data (although not unproblematic themselves) allows for clear answers to the three central questions of this book. Was there a trend? Yes there was, mainly towards higher percentages of populations reporting volunteering. Not finding a general trend toward less volunteering goes against gloomy accounts of a general development in the Western world to a more individualistic ethos at the expense of volunteering. Was there either divergence or convergence? Yes, there clearly was divergence, away from a situation in which countries clustered together in a limited range of percentages of volunteers towards a situation where countries differ markedly from one another in this respect. This increases the burden for future research in this field, as explaining this requires going into national differences, which in
198
paul dekker and andries van den broek
turn requires detailed nation-specific information. More precisely, the future challenge for more detailed comparative analysis is to explain why, starting from a more or less compatible level, volunteering increased in Sweden, Denmark, the Netherlands, Great Britain, and the United States, decreased in Germany and Spain, and remained stable in France, Belgium, Italy, and Ireland. Did generational differences play a role in this trend? No, they did not. In several countries, the latest generation counted fewer volunteers than the generation it replaced, but this only caused a minor slowing down of the overall increase in volunteering. Remember that the initial expectation was that the replacement of earlier by later generations would help to explain a decrease in volunteering. The picture is a very different one. The difference is not that comparisons of generations do not point to lower levels of volunteering among later generations, since the baby bust generation indeed displays lower levels of volunteering. The big difference is that there was no decrease in volunteering, despite the wheels of generation replacement working in that direction. Rather than spurring a decrease in volunteering, generation replacement slowed down an increase in volunteering. That increase would have been bigger, had not later generations been less prone to volunteer than earlier generations. Our second objective was exploring if and how volunteering is related to values held in other life domains. This exploration was attempted in part because, we hold, the (future) position of volunteering is better understood by not only studying volunteering itself, but by widening the perspective, through interpreting volunteering as one of the many choices made within a wider array of choices in a number of other life domains, such as work, leisure, and family life. The results of our exploration of the relation between volunteering and values in other life domains are all but conclusive, which, we think, is in the nature of an exploration. The EVS/WVS data point at a clear relation between volunteering and the importance attached to leisure, other life domains being partly (family, religion) or hardly (work, friends, politics) related. A somewhat more coherent, though far from definite, picture arises from a closer inspection of the relation between volunteering and work values. No relation was found between volunteering and the importance of work. Probing further revealed contrasting relations
is volunteering going down?
199
of volunteering with intrinsic work values on the one hand and with extrinsic work values on the other. This suggests that it is not the importance of work as such but rather what one values in work that is related to volunteering. In half of the countries studied, the more people support the intrinsic work value of being able to use one’s initiative in one’s work, the more they volunteer. Likewise, support for instrumental values such as good pay and good hours is negatively related to volunteering. As regards leisure, data from the Netherlands suggest four leisure dimensions, of which the much valued freedom and development dimensions are largely unrelated to volunteering, and the social dimension partly. Only hedonism, the least valued dimension, clearly is negatively related to volunteering. A possible future increase in the importance attached to freedom and development, the values central to the ethos of individualism and self-realization, appears not to affect volunteering. A possible future threat for volunteering lies in reduced support for doing something useful for society (related positively to volunteering) and in increased support for the hedonism dimension (related negatively to volunteering). This being a one-shot case-study, such general long-term assertions on how trends in leisure values might affect volunteering can be made with due reservations only. Hardly original, this leads to a plea for more research into these themes. All in all, the future of volunteering does not look bleak. Despite the fact that generational replacement inserts some influence towards less volunteering, the general trend points at increases rather than decreases in the percentages of volunteers. In a way, the future of research into volunteering does not look bleak either. Many of the issues raised in this chapter led to further questions yet to be answered. The increase in volunteering challenges worried accounts about where society is heading. Why these accounts apparently are misguided is as yet to be better understood. The divergence in the extent of volunteering in various nations calls for an explanation based on more nation-specific information. The exploration of how volunteering relates to values held in various spheres of life left ample room for future research too. Both volunteering and research into volunteering are, it seems, here to stay.
200
paul dekker and andries van den broek References
Bales, K. 1996. Measuring the propensity to volunteer. Social Policy Administration, 30, 206–226. Bekkers, R. 2004. Giving and Volunteering in the Netherlands (PhD. Thesis). Utrecht NL: Utrecht University Bocock, R. 1993. Consumption. London: Routledge. Curtis, J.E., E.G. Grabb & D.E. Baer 1992. Voluntary association membership in fifteen countries. American Sociological Review, 57, 139–152. Curtis, J.E., D.E. Baer & E.G. Grabb 2001. Nations of joiners. American Sociological Review, 66, 783–805. De Hart, J. 2001. Religion and volunteering in the Netherlands. In P. Dekker & E.M. Uslaner (eds.), Social capital and participation in everyday life. London: Routledge. De Hart, J. & P. Dekker 2005. Churches as voluntary associations. In S. Roßteutscher (ed.), Democracy and the Role of Associations. London: Routledge. Dekker, P. 2000. Social Participation and Political Involvement. Paper for the 18. IPSA World Congress, Quebec City, 1–5 August. Dekker, P. & L. Halman 2003. Volunteering and values. In P. Dekker & L. Halman (eds.), The Values of Volunteering. New York: Kluwer Academic. Dekker, P. & A. Van den Broek 2005. Involvement in voluntary associations in North America and Western Europe. Journal of Civil Society, 1, 45–59. Dingle, A. (ed.) 2001. Measuring Volunteering. Washington, DC: Independent Sector and United Nations Volunteers. Eliasoph, N. 1998. Avoiding Politics. Cambridge: Cambridge University Press. Enquete-Kommission 2002. Bürgerschaftliches Engagement. Opladen: Leske + Budrich. Featherstone, M. 1991. Consumer Culture and Postmodernism. London: Sage. Giddens, A. 1991. Modernity and Self-Identity. Cambridge: Polity Press. Govaart, M., H.J. Van Daal, A. Münz & J. Keesom 2001. Volunteering Worldwide. Utrecht: NIZW. Hodgkinson, V.A. 2003. Volunteering in global perspective. In P. Dekker & L. Halman (eds.), The Values of Volunteering. New York: Kluwer Academic. Hustinx, L. 2001. Individualism and new styles of youth volunteering. Voluntary Action, 3, 57–76. Inglehart, R. 1977. The Silent Revolution. Princeton: Princeton University Press. ——. 2003. Modernization and volunteering. In P. Dekker & L. Halman (eds.), The Values of Volunteering. New York: Kluwer Academic. Kearney, J. 2001. The values and basic principles of volunteering. Voluntary Action, 3, 63–86. Lane, R.E. 2000. The Loss of Happiness in Market Democracies. New Haven: Yale University Press. Lichterman, P. 2005. Elusive Togetherness. Princeton: Princeton University Press. Mannheim, K. 1952 [1928/29]. The problem of generations. In K. Mannheim; edited by P. Kecskemeti, Essays on the Sociology of Knowledge. New York: Oxford University Press. Parker, I. 2000. Obedience. Granta, 71, 99–125. Putnam, R.D. (with R. Leonardi & R.Y. Nanetti) 1993. Making Democracy Work. Princeton: Princeton University Press. ——. 2000. Bowling Alone. New York: Simon & Schuster. Reed, P. & L.K. Selbee 2003. Do people who volunteer have a distinctive ethos? In P. Dekker & L. Halman (eds.), The Values of Volunteering. New York: Kluwer Academic. SCP 2004. In het zicht van de toekomst. The Hague: Social and Cultural Planning Office.
is volunteering going down?
201
Smith, D. 1994. Determinants of voluntary association participation and volunteering. Nonprofit and Voluntary Sector Quarterly, 23, 243–263. Van den Broek, A. 1999. Does differential cohort socialization matter? Political Psychology, 20, 501–523. Van den Broek, A. & K. Breedveld 2004. Trends in Time. The Hague: Social and Cultural Planning Office. Van Deth, J.W. (ed.) 1997. Private Groups and Public Life. London: Routledge. Wilson, J. 2000. Volunteering. Annual Review of Sociology, 26, 215–240. Wilson, J. & T. Janoski 1995. The contribution of religion to volunteer work. Sociology of Religion, 56, 137–152. Wuthnow, R. 1991. Acts of Compassion. Princeton: Princeton University Press. Zeegers, W. 1998. Op zoek naar een kanjer. Amsterdam: Bert Bakker.
social welfare services cultural activities labor unions political parties human rights environment, conservation, animal rights professional associations youth work at least one of these
social welfare services cultural activities labor unions political parties human rights environment, conservation, animal rights professional associations youth work at least one of these 3 7 22
4 3 19
8 9 2 2 3 3 2 5 22
9 5 1 2 1 1 2 5 19
Netherlands
3 3 6 4 3 2
Sweden 1990
2 4 3 2 2 4
1981
3 5 32
9 16 2 3 4 2
4 5 33
9 11 10 4 4 4
1999
2 3 18
4 6 1 2 0 2
5 5 16
2 1 3 2 1 0
1981
2 5 19
6 7 2 2 3 3
Belgium
3 3 13
2 4 3 2 1 1
Denmark 1990
3 5 22
6 9 2 3 4 3
4 5 21
4 5 4 3 1 2
1999
2 4 14
3 3 3 2 0 1
2 2 17
6 2 2 4 1 1
3 3 14
5 5 2 2 1 2
France
2 2 14
3 4 2 3 1 1
1 1 12
4 5 1 1 1 1
1 2 8
2 3 0 1 0 1
Germany (West) 1981 1990 1999
Appendix Table 1: ‘Currently doing unpaid voluntary work’ for eight types of voluntary organizations/fields of activity (percentages of the population of 18 years old and over)
Appendices
202 paul dekker and andries van den broek
1 3 13
1 1 14
5 3 1 2 1 2 2 3 13
6 1 1 1 1 1 1 6 16
Great Britain
3 3 3 3 1 1
3 3 4 4 1 1
8 15 36
13 3 2 1 4 8
3 3 17
5 6 2 2 2 2
5 8 19
7 4 1 3 2 2
2 3 18
8 4 2 2 1 1
5 10 25
6 10 2 5 1 3
United States
1 1 7
2 2 1 1 1 1
Spain
11 22 47
14 20 3 7 3 8
1 1 9
2 3 1 1 1 1
Source: European Values Studies 1981, 1990, and 1999 and World values Study 1999; weighted results
social welfare services cultural activities labor unions political parties human rights environment, conservation, animal rights professional associations youth work at least one of these
social welfare services cultural activities labor unions political parties human rights environment, conservation, animal rights professional associations youth work at least one of these
Italy
1 7 18
7 2 2 2 1 1 1 5 17
7 4 1 2 1 1
Ireland
3 5 15
4 4 2 2 2 1
is volunteering going down? 203
33 21 8 32 22 14 17 9 15 36 47
49 33 29 51 – 58 24 22 33 37 59
ISSP 1998/9c 35 28 26 29 23 19 5 7 16 23 –
ESS 2002/3d
b
Correlations between the percentages of ten countries: EVS/WVS—ISSP .63, EVS/WVS—ESS .58 and ISSP—ESS .59. See 7.1 and Appendix 1. c International Social Survey Program: Has done ‘unpaid work, of service or benefit to other people or the community and not only to one’s family or personal friends’ in the past 12 months in at least one of the areas ‘political activities’, ‘charitable activities’, ‘religious and church-related activities’ and ‘any other kind of voluntary activities’; 18+ population. See: http://www.gesis.org/en/data_service/issp/data/1998_Religion_II.htm d European Social Survey: indicated to have ‘Done voluntary (unpaid) work for such an organization’ for at least one of 12 types of organizations (including sports and church and ‘other voluntary organizations’); 16+ population; Germany includes the east, UK instead of GB. See: http://www.europeansocialsurvey.org/
a
Sweden Denmark Germany (West) Netherlands Belgium France Italy Spain Ireland Great Britain United States
EVS/WVS 1999b
Appendix Table 2: Volunteering in three population surveys a
204 paul dekker and andries van den broek
0.98 1.04 0.99 1.12* 1.15** 1.09
1.02 0.94 1.10 1.02 1.10 1.14*
dk 1.05 1.02 1.15** 0.90* 1.20*** 1.28***
de 1.00 0.99 1.21** 0.93 1.10* 1.20***
nl 0.93 0.87* 1.28** 1.15 1.35*** 0.94
be 1.10 1.04 1.26** 0.97 1.21** 1.00
fr 1.05 0.88 2.08*** 0.70* 1.03 1.13
It 0.93 0.89 1.18 0.83 1.41** 0.77*
es 1.07 1.33** 0.85* 1.29** 1.06 1.00
ie 0.99 0.85** 1.21** 1.09 1.15* 1.39***
uk
a Statistical effects of per country standardized scores of 11–points assessment ‘extremely unimportant’ to ‘extremely important’, adjusted for effects of the other four assessments; significance: * < .05, ** < .01, *** <001 Source: European Social Survey 2002/3; weighted results
Work Family Friends Leisure time Politics Religion
se
Appendix Table 3: Relation between the importance attached to six life domains and volunteering, 1999 (adjusted odds ratios a)
is volunteering going down? 205
CHAPTER EIGHT
DECREASING DESIRES FOR INCOME INEQUALITY? Stephen Fisher and Anthony Heath Introduction Equality is sometimes referred to as if it were a single value linked to the belief that everyone is born equal. But just as people are more equal in some respects than others, there are differences in the kinds of equality people value (Sen, 1982). Equal treatment by the law is different from equality of opportunity or outcome in the economic domain. Equality of opportunity refers to a situation in which people are allowed to compete freely on a fair basis for jobs and resources, whilst equality of outcome exists when people have the same levels of economic well-being. These two types of equality are difficult to achieve at the same time because equality of opportunity usually generates inequalities of outcome. To ensure equal outcomes resources must be redistributed from the rich to the poor regardless of the source of inequality, i.e. even if the inequality is generated by fair competition. Those who prefer equality of opportunity to equality of outcome tend to do so because they value the freedom for autonomous individuals to pursue their own conception of the good (Dworkin, 1981). Conversely, equality of outcome is sometimes preferred because the distribution of resources that results from the free exchange of goods and services tends not to reflect differences in effort but differences in talents and capabilities that are the product of luck rather than choice. In themselves, the ideals of equality of opportunity and outcome are not well defined with clear policy implications. There is considerable debate as to what can coherently be required to achieve each kind of equality (Walzer, 1983). Those who profess to value equality of opportunity sometimes accept that some redistribution of resources is necessary to compensate for variation in talents and capabilities (or even differences in the wealth people are born with)
208
stephen fisher and anthony heath
in order to ensure fair competition. Meanwhile, proponents of equality of outcome differ in the extent to which they promote wealth, income or some other measure of economic well-being as the outcome that should be equalized. Although in many ways inequalities in wealth are far more important than those in income, equalizing wealth at regular intervals would create incentives for people to consume their wealth quickly to entitle them to further transfers of resources in the future. Partly for this reason, income equality is the most common concern of those who value equality of economic outcomes and it is therefore the focus of this chapter. This book seeks to answer questions about the nature of value change: whether there has been a rise or decline, convergence or divergence, and in so far as there has been any change in traditional values, to what extent are younger generations responsible? For these questions to be meaningful in the context of this chapter we need to consider whether income egalitarianism can adequately be described as a traditional value in the same way that, for example, a duty towards civic engagement or a belief in the sanctity of marriage have been in other chapters. If Western societies are more individualistic and there is less social solidarity than there used to be, it may be argued that concerns about the poverty and inequality are traditional values. However, it is far from clear how widespread such concerns were fifty or a hundred years ago. Moreover, in so far as people favor greater equality, it may be because they have something to gain themselves from redistribution of wealth. So we cannot characterize egalitarianism as necessarily either an altruistic or traditional value. However, it is still more difficult to argue the opposite: large differences in wealth have definitely not traditionally been seen as an ideal. Whilst some conservatives may see inequality as natural or inevitable (Freeden, 1996), it is rarely if ever seen as a positive virtue to be enhanced. So although income egalitarianism cannot straightforwardly be considered a traditional value, this is the most appropriate way to recast the central questions of the book for the subject matter of this chapter. We will therefore address the following three questions: 1. Has there been a general decline in egalitarianism in Western societies? 2. Have Western societies converged or diverged in their egalitarianism? 3. Are younger generations less egalitarian?
decreasing desires for income inequality?
209
To answer these questions, we will use data from the six countries (Australia, Austria, Great Britain, Hungary,1 the United States and West Germany) in the International Social Survey Program (ISSP) Social Inequality surveys of 1987, 1992, and 1999 (UK Data Archive, 1989, 1994, 2001). The reasons for this choice of dataset are given in the data and methodology section. First, however, we consider what processes may lead to changes in levels of income egalitarianism in these countries. Hypotheses regarding the sources of value change Different people within a society at a particular point in time will vary in the extent to which they value greater income equality often according to their own education, income, wealth, experience of the labor market, and other factors (Austen, 1999; Kelley & Zagorski, 2005). Societies are likely therefore to differ from each other and change over time in their preference for equality partly according to the distributions of these individual-level factors within the society. However, Austen (1999) shows that standardizing normative attitudes towards inequality for demographic differences between countries makes little difference to cross-national variation. It is similarly unlikely that changes in demographic structure will make a substantial difference to the levels of egalitarianism within countries over just twelve years (1987–1999). However, if they do, we should expect similar changes in each country as Western societies have followed similar trajectories in terms of aging, expanding service sector, and declining manual working class (Crouch, 1999). Some authors suggest different cultures approach the issue of income egalitarianism differently. For example, Robinson Bell (1978) argue that people accept inequality in Britain because they believe income differences are the just outcome of a fair society, whilst in the United States many believe that they will achieve financial success and this belief in the American dream is negatively correlated with egalitarianism. Without denying that there may be important 1
Although Hungary is not a West European country and not considered elsewhere in the book, it is included here because it is useful for comparative purposes and to add a further test of robustness to any general claims.
210
stephen fisher and anthony heath
cultural differences in the way people think about the distribution of income, it is hard to be sure whether views about meritocracy and opportunity precede or act as post-hoc justification for judgments on how much inequality is acceptable. Even if we could confidently assume the former and were able to conclude that egalitarianism has changed because the culture has changed we would be left with a further problem of explaining why the culture has changed. More practically, we do not have relevant measures of cultural change that are comparable over time and between countries that we can analyze together with our data on acceptable income differentials. Another possibility is that the level of egalitarianism is influenced by the nature of political and social institutions, especially the welfare state. Indeed, Svallfors (1997) shows how social-democratic countries are less tolerant of large income differentials than those with radical, conservative or especially liberal welfare-state regimes. In so far as the welfare regimes might affect social values it is likely to be through a process whereby people come to accept and support the circumstances they are familiar with. This may be due to either conservative temperaments or ‘what is ought to be’ mindsets. Notice, however, that any influence of change in the nature of the welfare state on egalitarianism should operate through inequality; people are unlikely to change their views on income equality as a direct result of change in the welfare regime if the change in the welfare regime has no effect on the distribution of income. How the welfare state affects the level of inequality has been discussed at length elsewhere (e.g., Crouch, 1999; Esping-Andersen, 1990), and it is clear that the character of the welfare state is not the only factor affecting inequality. So in so far as changes in egalitarianism reflect changes in actual income inequality, they do not necessarily reflect changes in the welfare regime. The level of income inequality in each of our six countries is given in Table 8.1 below.2 Whilst inequality in 1999 was higher than in 2 There are a variety of measures of income inequality. We have chosen the Gini coefficient for net (after tax and transfer payments) income as the most commonly used. There are numerous sources of data and estimates of Gini coefficients for different countries at different points of time, some of which cannot be compared either between countries or over time within countries (Atkinson and Brandolini, 2001). We have chosen to use the Luxembourg Income Study (LIS) estimates which are relatively sparse in terms of the availability of estimates for each of the years
decreasing desires for income inequality?
211
1987 in all six countries, some of the differences are quite small. The two most considerable increases in inequality were in Hungary, presumably as a result of the transition from a command to market economy, and Britain due to the implementation of Margaret Thatcher’s economic policies which had by 1987 already produced a substantial increase in inequality. If value change follows the same pattern of change as actual income inequality we can expect the six countries to be no more, and generally less, egalitarian in 1999 than they were in 1987. Table 8.1: Gini coefficients Year Country Australia Austria Great Britain Hungary United States West Germany
1987
1992
1999
99–87
29 23 30 24 34 27
30 28 34 28 34 27
31 26 35 30 37 28
+2 +3 +5 +6 +3 +1
Note: Table contains the Gini coefficients for net income for the nearest available year3 from the Luxembourg Income Study (2005), with the exception of West Germany and Hungary for 1987 where the data are drawn from Atkinson (2003) and Atkinson and Micklewright (1992: 137) respectively.
Despite differences in the rate of increase in inequality, there is no sign of convergence in the Gini coefficients in Table 8.1. The rank order of the countries in 1999 is the same as it was in 1987, except that Hungary overtook West Germany. The range across the six countries is 11 for both 1987 and 1999, and the standard deviation in 1999, at 4.2, is practically the same as the 4.1 it was in 1987. Egalitarianism may be responsive to levels of inequality but also inequality could be partly the product of values in society. The Gini coefficients in Table 8.1 reflect the effects of monetary transfers to various groups, particularly those on low or without other income. of interest, but the LIS does cover our six countries and the estimates have been constructed with comparability across time and space as key criteria. 3 The cases where the nearest estimate was more than one year away were as follows. Australia: 1985 for 1987, 1989 for 1992, and 1994 for 1999, Austria: 1994 for 1992. Whilst the discrepancies in the Australian case look unsatisfactory, the tiny change over the period suggests that the impression given in Table 8.1 is not misleading.
212
stephen fisher and anthony heath
How much the state redistributes in this way is a political decision influenced by public opinion. It may also be that employers set wages according to their own and/or society’s conceptions of fair income differentials (Atkinson, 1999a) which in turn could be the result of union pressure or greater productivity of employees who feel they are working in an environment where wages are fair. Using mainly cross-sectional data, Weakliem et al. (2002) focus on the influence of public opinion on social inequality but they also find evidence for some effect in the opposite direction. Even if there is causal influence in both directions, there are important questions when analyzing change over time as to whether the effects are lagged, and if so by how much. Preferences for greater equality may not react to the current level of inequality but to recent experience. Similarly, income inequality may not respond immediately to a change in public opinion. So even if there is a strong relationship cross-nationally between egalitarianism and inequality, it may not be visible in the pattern of change over time, particularly over a short period. A further complication is that for people to respond to changes in inequality they must perceive those changes accurately. Headey (1991) and Osberg & Smeeding (2005) show that perceived and normative distributions of incomes and income differentials are actually very similar to each other, while both are vastly different from actual income distributions and differentials. However, it may still be that people perceive changes in inequality relatively accurately and respond to these despite maintaining a consistently high degree of error in their perceptions as to the general level of inequality. But this hypothesis cannot be tested with existing data. The relationship between normative and perceived income differentials in cross-sectional survey data is likely to be severely biased as a result of differences between respondents in terms of the kind of person they think of for each occupation that is asked about (Kelley & Zagorski, 2005). So to properly assess whether values are influenced by actual inequality via perceptions in the presence of massive misperception, it would be necessary to have good panel data to see whether individual perceptions and values change in response to changes in inequality over time. One theory of value change that provides a basis for a hypothesis is Inglehart’s (1990, 1997) theory of post materialism. How this theory relates to egalitarianism is not straightforward. Since the object of the value (the income distribution) is material one might expect
decreasing desires for income inequality?
213
materialists to be more concerned about income inequality than postmaterialists. But the level of concern does not tell us about the direction of concern. Inglehart (1990: 252) argues that in contemporary Western societies income inequality is sufficiently low that there is little demand for greater income equality based on materialistic selfinterest, instead such demands are based on appeals to justice, social solidarity, and other non-material motivations. However, he presents no evidence for this and the claim contradicts research which shows that egalitarianism varies between individuals precisely as one would expect based on economic self-interest: higher social classes are less egalitarian than poorer people (e.g., Kelley & Evans, 1993; Svallfors, 1997). Moreover, Inglehart (1990: 276) also shows that preferences for reductions in income inequality are more highly correlated with ‘old left’ values than ‘new left’ post materialist values. Alternatively, it is perhaps more plausible that income egalitarianism is neither straightforwardly materialist nor post materialist, but its character depends on the wealth of the person holding the value. A strong preference for greater income equality when held by a relatively wealthy person is an altruistic value that would be congruent with other values held by post materialists including concerns for human rights and the environment, but the same does not hold for relatively poor people. Whilst egalitarianism is neither exclusively a materialist nor a postmaterialist value, Inglehart’s theory of value change through cohort replacement may still apply to egalitarianism. Those who grew up in times of economic deprivation could perhaps be expected to value equality more than those who have only experienced prosperous times. This may be because they are more easily offended by conspicuous and extravagant consumption and perhaps less likely to see inequality as necessary for economic growth. If this is right we should expect older cohorts to be more egalitarian than younger ones, and so societies should become less egalitarian as older generations are replaced by younger more inegalitarian ones. Alternatively it may be that those who grew up in times of relative hardship believe that equality is an expensive goal which should not be promoted if it is likely to damage economic growth. In which case societies should have become more egalitarian as a result of cohort replacement in recent years. Given these competing hypotheses, we cannot generate clear expectations as to the direction of change at the national level before looking at the levels of egalitarianism among different cohorts.
214
stephen fisher and anthony heath
By contrast, the hypothesis that egalitarianism follows actual inequality suggests monotone decreasing egalitarianism over the period for the six countries in this study.4 Whilst this accords with one interpretation of the cohort-replacement hypothesis, the Gini coefficients in Table 8.1 suggest a particular pattern of different rates of change in different countries which cohort-replacement theory does not. This may help us discriminate between the two hypotheses, but the plausibility of the cohort-replacement hypothesis rests fundamentally on the analysis of cohort differences over time. With regard to the second of our research questions, we have seen nothing in our discussion of the possible mechanisms of change to suggest that there should be either convergence or divergence among our countries over this period. So we must remain ambivalent on this subject. There are various reasons why we might expect the egalitarianism of younger generations (the subject of our third research question) to differ from older ones. This may be that younger generations are less egalitarian because they have grown up in more prosperous times but, as discussed above, the same reason can be used to suggest that younger generations may be more egalitarian, depending on how you think socialization during a period of prosperity affects values towards equality. There are, however, other reasons to expect younger generations to differ from older ones, not because of their cohort but because of their age. Young adults are relatively poor and they have had little experience of big income differentials across their own age group. They may also have a tendency to be more idealistic. This notion of age effects challenges the view that egalitarianism is a value which, along with other values, is developed primarily through childhood socialization and resistant to serious change in later life. The possibility that there are both cohort and age or life-cycle effects at play means that we will have to pay some attention to whether one or other predominates. If age or life-cycle effects are dominant, then younger generations are likely to appear more egalitarian but cohort differences could go in either direction or none. So for different reasons our expectations regarding our three research questions remain open. Before we can answer the questions
4 Austria should be an exception to this rule since inequality rises between 1987 and 1992 and drops again by 1999. However, the relevant questions were not asked in the 1992 Austrian ISSP survey, so this exception cannot be tested.
decreasing desires for income inequality?
215
we must first address the methodological issue of how egalitarian values can be measured. Data and methodology The aim of studying values regarding income inequality both between countries and across time requires that we use survey measures that are not related to the conditions in a respondent’s country at a given time. Also, substantively it is helpful to focus on normative preferences as to how equal society ought to be rather than perceptions of inequality, reasons for inequality, moral evaluations of the injustice of inequality or policy preferences for combating inequality. The European Values Studies ask questions about meritocracy, the trade-off between freedom and equality, and whether the level of inequality is too high or low. Unfortunately, the first two are not measures of values towards income inequality per se and the last is too specific to the time and place the question was asked to be a comparable measure across countries and between surveys. The ISSP Social Inequality surveys of 1987, 1992, and 1999 have the primary virtue of providing a set of questions on normative incomes that avoid many of the major problems with cross-national and temporal comparability. These surveys are also valuable because they have been conducted over a period of twelve years in six countries with different levels and rates of change in inequality, as seen in Table 8.1. Between them the countries also span four different welfare regime types, with the United States and, to a lesser extent, the United Kingdom representing the liberal type, Australia as an example of radical welfare-state arrangements, Austria and West Germany in the conservative/corporatist camp, and the pre-1989 state-socialist Hungary (Esping-Andersen, 1990; Castles & Mitchell, 1992). By asking what respondents think people in different occupations ought to be paid, the ISSP surveys provide data which, after taking ratios, yield measures of normative income differentials. The 1999 wording for the relevant questions was as follows. Next, what do you think people in these jobs ought to be paid—how much do you think they should earn each year before taxes, regardless of what they actually get . . . Please write in how much they SHOULD earn each year First, about how much do you think a
216
stephen fisher and anthony heath
skilled worker in a factory should to earn? . . . . . A doctor in general practice? . . . . . . . . . . . . The chairman of a large national corporation? . . . . An unskilled worker in a factory? . . . . . . . . . . A cabinet minister in the [NATIONAL] government? . . . There were other occupations included in each of the 1987, 1992, and 1999 versions of the question but the ones listed above are those which were asked in all three waves. Kelley & Evans (1993) use the ratio of the responses for high-status occupations relative to lowstatus occupations as their principal measure of legitimate pay differentials. They were able to do this in part because they were using just the 1992 survey which included a greater range of both high- and low-status occupations. Whilst almost all respondents in each country agree that unskilled manual workers should be the lowest paid of the five occupations, there is disagreement as to which should be paid the most. In response to this situation our measure of legitimate pay differentials will be the ratio of the largest amount the respondent said that any of the five occupations should earn relative to the income the respondent thought that an unskilled manual worker should earn. We call this the ‘maximum income ratio’ for the respondent.5 The choice as to which particular ratio to calculate between high- and low-status occupations makes little difference to the main substantive conclusions (Kelley & Evans, 1993; Kelley & Zagorski, 2005). This question was asked according to local norms for describing salaries as pre/post tax and yearly/monthly/weekly. The issue of time frame is not a problem since we are looking at ratios of responses. The tax and benefit structure may be more important but as Kelley & Evans (1993) point out, marginal tax rates are relatively flat and unlikely to make much difference to the conclusions, even regarding between-country differences.
5
Some surveys set upper limits for the possible responses and the use and levels of these limits are not consistent, neither over time nor between countries. Whilst this means the data on normative maximum incomes are truncated, there were never more than 4% of respondents who choose a salary at or above the upper limit, and the patterns of change in the frequency with which the upper limit was reached are not ones which match the changes in the maximum income ratio. So we can be confident that differential truncation has had no substantial impact on our conclusions.
decreasing desires for income inequality?
217
Osberg & Smeeding (2005) rightly point out that the ratio of earnings of the rich to poor occupations is only one aspect of inequality. It also matters how many people are in each kind of occupation, but it is not clear which aspect matters most to people. Whilst there is an interesting question in the 1999 ISSP survey regarding the ideal structure of society in terms of the relative sizes of different income groups, we cannot look for change over time in attitudes on this issue. Results We will address each of our three research questions in turn. 1. Has there been a general decline in support for equality in Western societies? Table 8.2 below shows the mean of the maximum income ratios for 1987, 1992, and 1999 for the countries available in both the 1987 and 1999 waves. Overall, after treating each country as if it had equal population size, the mean rose from 6.8 to 7.4, implying a decline in support for equality. However, the pattern of change varies between countries. In Australia, Britain, and Hungary people were more comfortable with large income differentials in 1999 than they were in 1987, whereas the Americans, Austrians, and West Germans became less comfortable. The three countries, which saw a net increase over the period from 1987 to 1999, saw increases between both 1987 and 1992, and from 1992 to 1999, although the latter in Britain is not statistically significant. The reverse is not true of the two countries that experienced a net decline and for which we have 1992 data. West Germany saw no change between 1987 and 1992, whilst the United States experienced a rise between 1987 and 1992 followed by a larger drop from 1992 to 1999. The most substantial change over the period was, unsurprisingly in light of the transition to a market economy, in Hungary. The average Hungarian respondent in 1987 thought that the top profession should not be paid more than 3.3 times more than an unskilled factory worker. In 1999 the comparable figure was 8.0. It should be noted that Kelley and Zagorski (2005, Table 2, p. 338) found that although Hungary followed a similar trajectory to other East European former Communist countries, the rate of increase and the final level of inegalitarianism
218
stephen fisher and anthony heath
was greater. Overall then, even though the average change across the countries is in the direction of less egalitarianism, the differences between the trajectories is such that we cannot say there has been a general decline in support for equality. Table 8.2: Mean of the maximum income ratio by country and year Country
1987
Australia Austria Great Britain Hungary United States West Germany
3.8 7.6 7.9 3.3 10.5 7.8
Overall
Year 1992 5.1 8.9 5.1 12.5 7.8
6.8
1999 5.9 6.2 9.2 8.0 8.7 6.9 7.4
Note: Sample sizes for the ISSP surveys vary between 804 and 2606.
This conclusion is perhaps surprising in light of the relatively consistent monotone increases in actual inequality in Table 8.1. Even looking at just the change between 1987 and 1999, there is only a very weak positive association between increasing inequality and inegalitarianism, and this is mainly driven by the large increases in both variables in Hungary. It is interesting to note that the 1987 to 1999 changes in values have produced a stronger positive cross-national relationship between inequality and egalitarianism. In 1987, Austrians had relatively inegalitarian values given their more egalitarian income distribution and Britain was relatively egalitarian in its attitudes despite not being so in practice. By 1999, values in both countries had changed to bring them into line with the general pattern whereby more equal countries have more egalitarian people. Only Australia appears to be a significant exception to the rule, in having the most egalitarian values but average levels of actual inequality. Whilst this suggests the possibility of some kind of error-correction process in which people come to accept the level of equality in their society after a period of time, not all the changes fit this pattern and there are too few countries and time points to test this possibility seriously.
decreasing desires for income inequality?
219
2. Have Western societies converged or diverged in their egalitarianism? Table 8.2 above shows that our six countries appear to have converged in their attitudes towards income differentials. The standard deviation of the country means was 2.7 in 1987 but 1.4 by 1999. Similarly the range has dropped over this period from 7.2 to 3.3. This is only partly the result of Hungary moving from a communist to a market society and simultaneously changing from being the most egalitarian of our six to one of the least. If we exclude Hungary, there is still a drop in the standard deviation from 2.7 to 1.5 and in the range from 6.7 to 3.3 amongst the other five countries. However, the difference in the standard deviations between 1987 and 1999 is not statistically significant, whether or not Hungary is excluded.6 Perhaps most tellingly, the standard deviation amongst the five countries for which we have data for 1987, 1992, and 1999 does not follow a linear trend but is largest in 1992. Further evidence that the convergence in Table 8.2 is not underpinned by sociological mechanisms can be found by looking at levels of fragmentation within countries. Fragmentation is a process of entropy whereby people differ more in their values. If countries are converging in their values, then we should expect individuals within countries to be converging on the same international norm and thus countries would become less fragmented or at least not more so. Fragmentation can be measured using the variance of the maximum income ratios within each country. Table 8.3 shows us that the standard deviation of the maximum income ratio increased between 1987 and 1999 in Australia, Britain, and Hungary pointing to increased fragmentation, and decreased in West Germany, the United States, and Austria, suggesting greater cohesion. But the pattern of change was not consistently in the same direction for two of the five countries for which we have data for all three waves. By and large the patterns in Table 8.3 are the same as those in Table 8.2. When egalitarianism goes up, fragmentation goes down. In part this may be the somewhat inevitable result of the statistical relationship between
6 What statistical significance might mean in this context is unclear but its absence provides an indication that the change is not a strong and clear one. There may not be any sociological mechanisms driving the convergence. Instead the apparent convergence may simply be the result of random variation.
220
stephen fisher and anthony heath
the mean and variance of the distribution of the maximum income ratio. When people are on average more willing to tolerate large income differentials there is more room for disagreement between people. This accords with the Kelley & Evans (1993) finding that disagreement on how much different occupational groups should be paid is concentrated at the top end so “the main issue is not the welfare of the poor but rather the legitimacy (or lack thereof ) of high incomes.” (p. 110) Table 8.3: Standard deviations of the maximum income ratio by country and year Country
1987
Australia Austria Great Britain Hungary United States West Germany
1.6 6.5 10.0 1.8 15.7 9.5
Overall
Year 1992 3.6 9.4 4.4 18.4 8.0
9.2
1999 3.9 4.9 11.3 10.6 10.0 5.1 8.2
Note: Sample sizes for the ISSP surveys vary between 804 and 2606.
Across the six countries, giving each one equal weight, the level of fragmentation measured at the individual level has declined (from 9.2 to 8.2) as it has when measured between countries. However, the individual-level fragmentation mainly reflects the general decline in egalitarianism, not a process of convergence within countries leading to one between countries. There is only a very weak link between the change in fragmentation within a country and the extent to which it contributed to the convergence between countries. Since some countries have become more fragmented and others less so, the convergence between countries does not seem to have been the product of a systematic process. 3. Are younger generations less egalitarian? Egalitarianism has gone up in some countries and down in others and although this is not something we would predict from a cohortreplacement hypothesis, it also doesn’t preclude the possibilities either that younger generations are different from older ones or that there is a process of cohort replacement affecting change in egalitarianism.
decreasing desires for income inequality?
221
For this, we need to analyze value change in more detail. Following the convention in this book, we consider whether there are differences between the war time generation (born before 1940), the baby boomers (born between 1940 and 1960), and the baby bust generation (born after 1960). Table 8.4 below shows that younger generations are systematically (and statistically significantly) more egalitarian than older generations in the five Western countries. The exception is Hungary in 1999, where the generational differences are not significant. There are also significant period differences which follow a similar pattern for each generation within each country. This means that values relating to income inequality are not stable within cohorts and that change over time cannot be fully explained by a process of cohort replacement. It seems, however, that cohort replacement has, if anything, been driving the countries (with the possible exception of Hungary) to become more egalitarian, not less. Kelley & Evans (1993) and Robinson & Bell (1978) suggest two explanations for generational differences: a generation gap as a result of socialization in more recent egalitarian times (cohort effects); and a life-cycle effect whereby the difference in attitude reflects the difference in pay even after controlling for class (age effects). Kelley & Zagorski (2005) show (using data that overlaps with that used here) that there are intergenerational differences even after controlling for income, class, status, and education, and they attribute these to age and life-cycle effects rather than cohort differences but do not provide details for the basis of this attribution. Separating age, period, and cohort effects completely is impossible due to the identification problem (see for example, Glenn, 1976). However, given our definition of cohorts, it is possible to construct equivalent age groups and consider them simultaneously with cohort and period effects. In 1987, the war cohort was aged 48 or older, the boom cohort was between 28 and 47, and the bust cohort was between 16 and 27. In 1992 and 1999, there is progressively less overlap between age groups and cohorts. Operationalizing age and cohort in this symmetrical way allows an even-handed comparison of the relative importance of age groups and cohorts in accounting for differences in egalitarianism. Table 8.5 presents a series of regression analyses with age, period (year), and cohort dummy variables. The first six columns give the results of separate regressions for each country. These allow us to see whether there appear to be similar processes operating in each country before proceeding to a pooled analysis. In Australia, Britain,
6.3
3.7 7.2 6.8 3.4 9.9 7.6
1987 Boom
5.5
3.6 7.0 5.4 3.0 7.8 5.5
Bust
7.6
7.4 5.4 12.7 8.0
11.9 5.5 15.5 9.4 9.4
5.0
1992 Boom
5.8
War
Note: Sample sizes for the ISSP surveys vary between 804 and 2606.
7.9
4.1 8.2 10.5 3.4 13.2 8.8
Australia Austria Great Britain Hungary United States West Germany
Total
War
Country
6.2
7 4.1 9.1 5.9
4.1
Bust
Table 8.4: Mean of the maximum income ratio by country, year, and cohort
8.5
6.6 7.1 12.3 7.2 10.8 8.0
War
7.7
6.0 6.5 8.7 9.2 8.9 6.9
1999 Boom
6.6
5.4 5.4 7.5 7.4 7.7 6.2
Bust
8.5
5.4 7.8 11.5 5 13.7 8.8
War
7.2
4.9 6.9 7.6 5.8 10.6 7.5
Overall Boom
6.2
4.6 5.9 6.8 5.5 8.2 5.9
Bust
222 stephen fisher and anthony heath
decreasing desires for income inequality?
223
and the United States, the war generation is distinctly less egalitarian, whilst in West Germany the youngest cohort is noticeably more egalitarian. In Austria, the coefficients of both age and cohort groups are all statistically insignificant. Although there are variations in the p-values and the age effects are never statistically significant, the age group and cohort coefficients in all five Western countries consistently follow the same pattern: younger age groups are increasingly egalitarian and so are the more recent cohorts. This is in marked contrast with the pattern in Hungary where the middle aged are distinctly more egalitarian than those both older and Table 8.5: Pooled and separate country regressions of maximum income ratio on cohort (war and bust), age group (age 48 plus and 16–27), year, country, and year*country interactions AUS AUT GBR HUN Coef. Coef. Coef. Coef. Cohort (base = bust) Boom War Age (base = 16–27) Age 28–47 Age 48 plus Year (base = 1987) 1992 1999 Constant AUS FRG US AUT AUS*1992 FRG*1992 US*1992 AUS*1999 FRG*1999 US*1999 AUT*1999 N
US Coef.
FRG Pooled Coef. Coef.
0.3 0.7*
0.3 0.3
0.4 3.7*
1.4* –0.1
1.1 3.4*
1.1* 1.7*
0.7* 2.0*
0.1 0.4
0.2 1.2
0.6 1.4
–0.9 0.4
1.5 2.2
0.4 1.2
0.5 1.2*
1.2* 2.1* 3.3*
–1.4* 6.7*
1.0* 1.6* 5.6*
1.8* 4.7* 3.0*
2.0* –1.3* 7.6*
0.3 –0.5 5.9*
1.0* 1.4* 6.2* –4.0* –0.3 2.7* –0.4 0.1 –0.7 1.0 0.8 –1.9* –2.9* –2.5*
4907 1799
2572
4536
3399
3954
16631
Note: Pooled analysis excludes Hungary and is weighted to avoid differences in sample size across countries and years. Coef. is the raw regression coefficient. * by a coefficient denotes statistical significance at 5% level, but in a variable name it denotes multiplication to create an interaction term.
224
stephen fisher and anthony heath
younger than them, while the baby boom cohort (which corresponds to the middle age group in 1987) is less egalitarian than both the preceding and following cohorts. Moreover, inspection of Table 8.4 and further analysis shows that this pattern is not stable over time. In light of the very different pattern of age and cohort differences in Hungary, the final column of Table 8.5 presents a pooled analysis of just the five Western countries, including the year*country interaction terms. This suggests the presence of both age group and cohort effects of which the cohort effects seem to be stronger, most likely because of the particularly strong cohort effects in Britain and the United States. Further analysis (not shown), including controls for sex, marital status, years of schooling, formal educational attainment, religion, trade union membership, and employment sector, makes little difference to the magnitude and significance of the age group and cohort effects, in either the pooled or separate country regressions. Given that both the cohorts and age groups are operationalized in a relatively crude manner with just three categories it may be that the coefficients of the cohorts partially reflect age effects. A more stringent test of the presence of cohort effects would be to control for a curvilinear effect of age. In fact, including even just a linear age term in any of the regressions in Table 8.5 is sufficient for the coefficients of the cohort dummies to shrink substantially and become clearly insignificant. So whilst there is a case for the presence (excepting Hungary) of a cohort effect (as defined by the war-boom-bust division) when pitted against an equivalent age group effect, a more stringent test controlling for a linear age effect shows that there is insufficient evidence that there are systematic differences between the war, baby boom, and baby bust cohorts in the extent to which they value low income differentials. Conclusion The answers to our three key questions are all to some extent negative. Although there was less egalitarianism in 1999 than in 1987, there were different patterns and directions of change in each country. Levels of egalitarianism across the six countries in this study were more similar in 1999 than in 1987, but this convergence was not consistent, nor was it founded in a process of decreasing frag-
decreasing desires for income inequality?
225
mentation within each country. Finally, although there are intergenerational differences, they are not in the direction suggested by our research question. Younger generations are not less but more egalitarian than older generation, but whether the differences are mainly cohort or age based is not clear. Despite increases in inequality in all six countries over the period we did not observe the same consistency of change in values. Instead there was a remarkable variety of different patterns of change especially given that some of the countries are relatively similar. Whilst the countries with more equal income distributions tended to have higher levels of egalitarianism, there was no sign that change in one of these was accompanied by change in the other. It may be that there is in fact a relationship between egalitarianism and inequality but there are several reasons why we may not have found one. First, maybe there is some kind of lag between change in inequality and change in values, or vice versa. Unfortunately, Gini coefficients are not available for all the countries in precisely the years we are studying. However, further investigation of historical data on inequality provides little indication of any lagged effects of changes in inequality on values. Second, we have been comparing a measure of the income differentials that people are willing to tolerate when earnings dispersion is just a part of inequality as measured by the Gini coefficient (Atkinson, 1997). Moreover, the ratio of the earnings of the top decile to the bottom decile, although positively correlated with the Gini, does not necessarily result in the same pattern of change, and the direction of change can also be different at different ends of the income spectrum (Atkinson, 1999b). Third, the relationship between inequality and values may be mediated by perceptions of inequality. As discussed earlier, it is impossible to test this adequately without good panel data. However, if the patterns of value change were the product of a change in the perceived level of inequality, we would then need to explain why changes in perceived inequality bear little correspondence with changes in actual inequality. Whilst it is not hard to understand public perceptions of inequality being at odds with reality, it is difficult to see why the nature of the mismatch should change in different ways in different countries. So it seems unlikely that changes in perceptions of inequality would provide an adequate explanation for the pattern of value change we have observed. The negative conclusion regarding convergence is, naturally, restricted to the set of countries that we have studied. With an equal number
226
stephen fisher and anthony heath
but different set of countries and slightly different measures Stamm et al. (2002) concluded that country differences have not only failed to converge but increased over time. Meanwhile, Kelley & Zagorski (2005) show that levels of egalitarianism in eastern European countries have on average become more similar to those in Western countries since the late 1980s. So if we had data on more East European countries, the case in favor of convergence would be stronger. However, even though East and West Europe have become more similar, the process of convergence within the West is weak and, as Kelley & Zagorski (2005) show, non-existent in the East. These findings, coupled with increasing fragmentation in some countries, show that countries are not all headed towards the same level of egalitarianism. Younger people are more egalitarian than older people in Western countries, perhaps both because of ageing or life-cycle effects, and because people who were born more recently have experienced different social conditions which lead them to be more egalitarian. However, the idea that the war, baby boom, and baby bust generations are distinct in their levels of egalitarianism is not substantiated if we allow for a linear ageing effect whereby people become a bit more tolerant of inequality with each year older that they become. Since young people are more egalitarian, as Western societies age they have a tendency to become less egalitarian. However, this process accounts for very little of the slight decline in egalitarianism overall and practically nothing of the more substantial differences between countries in their patterns of change. References Atkinson, A.B. 1997. Bringing income distribution in from the cold. The Economic Journal, 107, 297–321. ——. 1999a. Is Rising Income Inequality Inevitable? A Critique of the Transatlantic Consensus. Helsinki: United Nations University, World Institute for Development Economics Research. ——. 1999b. The distribution of income in the UK and OECD countries in the twentieth century. Oxford Review of Economic Policy, 15, 56–75. ——. 2003. Income Inequality in OECD Countries: Data and Explanations. CESIFO Working paper 881. Atkinson, A.B. & A. Brandolini 2001. Promise and pitfalls in the use of “secondary” data-sets: income inequality in OECD countries as a case study. Journal of Economic Literature, 39, 771–799. Atkinson, A.B. & J. Micklewright 1992. Economic Transformation in Eastern Europe and the Distribution of Income. Cambridge: Cambridge University Press.
decreasing desires for income inequality?
227
Austen, S. 1999. Norms of inequality. Journal of Economic Issues, 33, 435–442. Castles, F.G. & D. Mitchell 1992. Identifying welfare state regimes: the links between politics, instruments and outcomes. Governance, 5, 1–26. Crouch, C. 1999. Social Change in Western Europe. Oxford: Oxford University Press. Dworkin, R. 1981. What is equality? I. equality of welfare. Philosophy and Public Affairs, 10, 185–246. Esping-Andersen, G. 1990. The Three Worlds of Welfare Capitalism. Cambridge: Polity Press. Freeden, M. 1996. Ideologies and Political Theory: a Conceptual Approach. Oxford: Clarendon Press. Glenn, N.D. 1976. Cohort analysts’ futile quest: statistical attempts to separate age, period and cohort effects. American Sociological Review, 41, 900–904. Headey, B. 1991. Distributive justice and occupational incomes: perceptions of justice determine perceptions of fact. The British Journal of Sociology, 42, 581–596. Inglehart, R. 1990. Culture Shift in Advanced Industrial Society. Princeton: Princeton University Press. ——. 1997. Modernization and Postmodernization: Cultural, Economic, and Political Change in 43 Societies. Chichester: Princeton University Press. Kelley, J. & M.D.R. Evans 1993. The legitimation of inequality: occupational earnings in nine nations. American Journal of Sociology, 99, 75–125. Kelley, J. & K. Zagorski 2005. Economic change and the legitimation of inequality: the transition from socialism to the free market in Central-East Europe. Research in Social Stratification and Mobility, 22, 321–366. Luxembourg Income Study 2005. Income Inequality Measures. Osberg, L. & T. Smeeding 2005. “Fair” Inequality? An International Comparison of Attitudes to Pay Differentials. . Robinson, R.V. & W. Bell 1978. Equality, success, and social justice in England and the United States. American Sociological Review, 43, 125–143. Sen, A. 1982. Equality of what? In Choice Welfare and Measurement. Oxford: Basil Blackwell. Svallfors, S. 1997. Worlds of welfare and attitudes to redistribution: a comparison of eight western nations. European Sociological Review, 13, 283–304. Stamm, H., M. Lamprecht & R. Nef 2002. Between the Market and the State: a Longitudinal Comparative Study of Inequality Perceptions in Five European Countries. Paper presented at the World Congress of Sociology, Brisbane, July 7–13. UK Data Archive 1989. ISSP Social Inequality I. Electronic Data File SN 2702. UK Data Archive 1994. ISSP Social Inequality II. Electronic Data File SN 3498. UK Data Archive 2001. ISSP Social Inequality III. Electronic Data File SN 4821. Walzer, M. 1983. Spheres of Justice. Oxford: Robertson. Weakliem, D.L., R. Andersen & A. Heath 2002. The Directing Power? A Comparative Study of Public Opinion and Income Distribution. CREST working paper 98. .
PART II