The Moral Economy of Class: Class and Attitudes in Comparative Perspective 0804752850, 9780804752855

Citation preview

THE

MORAL

ECONOMY

OF

CLASS

EDITORS

David B. Grusky s TA FoRo u N 1 v ER s 1 T Y Paula England s TAN F o R o u N 1 v ER s 1 T Y

EDITORIAL BOARD

Hans-Peter Blossfeld Mary C. Brinton Thomas DiPrete Michael Hout Andrew Walder Mary Waters

OTHER BOOKS IN THE SERIES

Studies in Social Inequality The Global Dynamics of Racial and Ethnic Mobilization By Susan Olzak

2006 Mobility and Inequality: Frontiers of Research in Sociology and Economics Edited by Stephen L. Morgan, David B. Grusky, and Gary S. Fields

2006 Poverty and Inequality Edited by David B. Grusky and Ravi Kanbur

2006

On the Edge of Commitment: Educational Attainment and Race in the United States By Stephen L. Morgan

2005 Occupational Ghettos: The Worldwide Segregation of Women and Men By Maria Charles and David B. Grusky

2004 Home Ownership and Social Inequality in Comparative Perspective Edited by Karin Kurz and Hans-Peter Blossfeld

2004 Reconfigurations of Class and Gender Edited by Janeen Baxter and Mark Western

2001 Women's Working Lives in East Asia Edited by Mary C. Brinton

2001 The Classless Society By Paul W. Kingston

2000

THE MORAL ECONOMY

Class and Attitudes in Comparative Perspective

Stefan Svallfors

S T ANFOR D UN I V ERSITY PRESS ST A NF O RD , CALIFO R N I A

2 00 6

Stanford University Press Stanford, California ©2006 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying and recording, or in any information storage or retrieval system without the prior written permission of Stanford University Press. Permission to use the following materials from previously published articles is gratefully acknowledged: Parts of Chapter 5 appeared as Stefan Svallfors. 2004. Class, Attitudes and the Welfare State: Sweden in Comparative Perspective. Social Policy & Administration 38 (2): 119-38. Chapter 7 is a reyised version of Stefan Svallfors. 2005. Class and Conformism: A Comparison of Four Western Countries. European Societies Vol 7 (2): 255-86.

Library of Congress Cataloging-in-Publication Data Svallfors, Stefan. The moral economy of class: class and attitudes in comparative perspective I Stefan Svallfors. p. cm. Includes bibliographical references and index. ISBN 0-8047-5285-0 (cloth: alk. paper) 1. Social classes. 2. Social classes- Cross-cultural studies. 3. Attitude (Psychology). 4. Social perception. 5. Public opinion. I. Title. HT609.S86 2006 2006002208 305.5-dc22 Typeset by G&S Book Services in 10/14 Sabon Original Printing 2006 Last figure below indicates year of this printing: 15 14 13 12 11 10 09 08 07 06

For Hugo For Signe

The past is never dead. It's not even past. William Faulkner

CONTENTS

List of Tables and Figures Acknowledgments xv

xz

CHAPTER ONE

The Moral Economy of Class

1

CHAPTER TWO

Class and Social Attitudes: A Framework for Analysis

9

CHAPTER THREE

Work: In the Sweat of Their Brows?

31

CHAPTER FOUR

The Market: Whosoever Hath Shall to Him Be Given?

52

CHAPTER FIVE

The Welfare State: My Brother's Keeper?

77

CHAPTER SIX

Family and Gender: Flesh of My Flesh?

103

CHAPTER SEVEN

Class and Conformism: The Dark Side of the Working Class?

121

CHAPTER EIGHT

Microclasses, Household Classes, Marginal Classes?

144

IX

x

Contents

CHAPTER NINE

Class and the Attitude Problem APPENDIX

A

Databases APPENDIX

162

178

B

The Class Variable and Occupational Codes APPENDIX

183

C

Construction of the Attitude Indexes Used in Chapter 5 APPENDIX

D

Analytical Methods Notes 193 References 205 Index 219

189

186

TABLES AND FIGURES

Tables

2.1. 3.1. 3.2. 3.3. 3.4. 3.5. 3.6. 3.7. 3.8. 4.1. 4.2. 4.3. 4.4. 4.5. 5 .1. 5.2. 5.3. 5 .4.

Attitudinal domains 20 Employment commitment in Sweden, Britain, Germany, and the u.s. 35 Job characteristics in Sweden, Britain, Germany, and the U.S. 37 Job satisfaction in Sweden, Britain, Germany, and the U.S. 41 Organizational commitment in Sweden, Britain, Germany, and the U.S. 43 Social relations at the workplace in Sweden, Britain, Germany, and the U.S. 45 Job satisfaction 47 Employment commitment index 48 Organizational commitment index 49 Legitimate income differences in Sweden, Britain, Germany, and the U.S. 58 Class differences in attitudes to income redistribution in Sweden, Britain, Germany, and the U.S. 61 Class differences in explanations for inequality in Sweden, Britain, Germany, and the U.S. 64 Class differences in attitudes to what money should buy in Sweden, Britain, Germany, and the U.S. 69 Correlations between indexes among unskilled workers and service class I in Sweden and the U.S. 72 Class differences in attitudes to state intervention in Sweden, Britain, Germany, and the U.S. 82 Government intervention index 85 Welfare spending index 86 Attitudes to welfare policies 90 XI

x11

Tables and Figures

5.5. 5.6. 6.1.

Spending index 94 Spending index 97 Class differences in attitudes to women's and men's paid work and household work in Sweden, Britain, Germany, and the U.S. 110 Gender equality index 115 Does more than fair share of household work 117 Attitudes toward sexual behavior 130 Attitudes toward infidelity and homosexuality 133 Attitudes toward civil liberties and rights 136 Attitudes toward civil rights and liberties 138 Index values in different SEI categories 147 Index values in different classes 149 Individual class, household class, and attitudes 153 Index values in difference classes and marginalized groups Class differences at spending index 158 Index values in different classes within the marginal group Election manifestos in Sweden 174 Achieved samples and response rates (percent) in the ISSP surveys 180 Samples and response rates in the Swedish Welfare State Surveys 181 Classes in Sweden, Britain, Germany, and the U.S. 185

6.2. 6.3. 7.1. 7.2. 7.3. 7.4. 8.1. 8.2. 8.3. 8.4. 8.5. 8.6. 9 .1. A.l. A.2. B.l. Figures

2.1. 2.2. 3.1. 3.2. 3.3. 4.1. 4.2. 4.3. 4.4. 4.5. 5 .1. 5.2. 5.3. 5.4. 6.1.

Interests and norms as mechanisms 21 Political institutions and political articulation 26 Employment commitment index 36 What people have and what they lack in their jobs Organizational commitment index 44 Pay-spread index 60 Income-redistribution index 62 Conflict index 66 Incentives index 67 Welfare purchase index 70 Government intervention index 84 Spending index 87 Finance index 88 Service index 89 Gender equality index 112

39

158 159

Tables and Figures

7.1a. 7.1b. 7.2a. 7.2b. 7.3a. 7.3b. 7.4a. 7.4b. 9.1.

Attitudes toward infidelity 132 Attitudes toward homosexuality 132 Attitudes toward infidelity 135 Attitudes toward homosexuality 135 Attitudes toward following the law without exception 137 Attitudes toward miscarriages of justice 137 Attitudes toward following the law without exception 140 Attitudes toward miscarriages of justice 140 Institutions, class stratification, and class articulation 167

xm

ACKNOWLEDGMENTS

Despite the seclusion that is such a common feature of writing, research is a collective product. Many institutions and individuals have helped in the making of this book. The Swedish Council for Working Life and Social Research (FAS) financed the Class and Attitudes in a Comparative Perspective project (FAS 2001-1005). FAS is also one of the financiers, along with the Bank of Sweden Tercentenary Foundation, of the Political Sociology of the Welfare State, a project of which my book forms a part (FAS 2002-0811 and K2002-0058, respectively). The Tercentenary Foundation also provided considerable funding to the collection of the data on which my analyses have been based (]1998-0043). The Swedish Research Council also contributed to the production of this book by awarding me a "long-term support for leading scholars" grant. I would also like to thank all the anonymous people who have taken time to answer the underlying questionnaires; the tens of thousands of individuals in Sweden, Britain, Germany, and the United States who in so doing have helped us to understand a little more about the societies in which they live. Without their help, this kind of research would be impossible. A third word of thanks goes to all those who have cast a thorough and critical eye over my drafts or assisted me with the day-to-day drudgery of research work. Any remaining shortcomings can only be blamed on the author's indolence and incompetence; better collaborators could no researcher have wished for, and better readers can no manuscript have. My closest collaborator over the years has been Jonas Edlund. Not only did Jonas pore over what probably seemed like an unending stream of chapter drafts and papers and comment wisely on their content, he also made two especially important contributions. First, not once have I been able to XV

xv1

Acknowledgments

present him with an SPSS problem that has taken him longer than 2 minutes to solve. He is a troubleshooter of the highest caliber, something I am not alone in being able to testify. Second, he makes an extremely pleasant traveling companion, an essential quality to possess in the constant stream of departure terminals, flights, and conference rooms that is the unhappy lot of the international researcher. After more than a decade of collaboration it is only far too easy to take him for granted, but he deserves much more. For the Political Sociology of the Welfare State research program I had the privilege to work with Jonas and a rare constellation of creative, enthusiastic, and positively critical people: Mikael Hjerm, Staffan Kumlin, Maria Oskarson, Maria Petterson, and in the final phases, Linda Berg as well as, earlier on, Emma Olsson. This is the kind of group that when given the choice between spending more time on holiday in Florence or discussing (yet again) my book, they would plump for the latter, a perplexing prioritization for which I am eternally grateful. Their comments have been extremely valuable and we always have genuine fun at our project meetings (which possibly explains the choices they make!). I was also bemused by Knud Knudsen's giving up a Stavanger spring for an Umea winter in order to comment on my manuscript. Knud's ability to deliver astute criticism in such a way as to make both critic and author appear charming and intelligent is rare indeed. He is thus living proof that a sharp mind is not always irreconcilable with a friendly demeanor. Robert Erikson, Kerstin Jacobsson, and Jens Rydgren took time to read the entire draft and come up with a number of valuable comments. If I have not followed Robert's good advice and detailed commentaries then it is probably because, despite my greatest efforts, I was not always able to decipher his idiosyncratic handwriting. The list of names of those who on different occasions have commented on particular chapters (in either the Swedish or English version) is a long one: Rune Aberg, Christina Axelsson, Richard Breen, Rickard Danell, Nils Eriksson, John Goldthorpe, David Grusky, Asa Gustafson, Rolf A. Gustafsson, Bjorn Hallerod, Mike Hout, Calle Hult, Lena Karlsson, Walter Korpi, Olle Lundberg, Mikael Nordenmark, Charlott Nyman, Bo Rothstein, Ingrid Schild, Michael Shalev, Al Simkus, and Eva Sundstrom. Some of them have no doubt forgotten that they once gave their opinions on my work; others I have probably forgotten, for which I apologize.

Acknowledgments

xvii

The book was translated by Neal Betteridge and the author. The book is dedicated to my children, Hugo and Signe. The Big Book of Statistics, as Hugo teasingly calls it, is done. Even though neither of you contributed in the slightest way to its production, it remains your book nonetheless. Not as an excuse for my absences, but with thanks for all the time we spend together. Without your love, humor, and intelligence my life would be infinitely more impoverished.

THE

MORAL

ECONOMY

OF

CLASS

CHAPTER

ONE

The Moral Economy of Class

This is a book about classes. Its main focus is not on how different classes live but on how classes differ in how they look at society. What are their opinions on distribution and justice? What do they think about work and the family, about the market and politics, about rights and morality? Do class patternings differ between countries, and, if so, why? Do the patternings in class opinion change over time, and, if they do, what is the reason for this change? Why are there class differences in opinions in the first place? The book is based on the notion that there is a "moral economy" in society, in which the mutual rights and obligations of the governing and the governed are collected and condensed (Svallfors 1996b). The term "moral economy" was coined by the English historian E. P. Thompson in an analysis of the recurrent bread riots of 18th-century England (Thompson 1971). Thompson found that it was seldom the starvation and the fury that were responsible for them. Instead, it was an idea that the governing powers had broken the unspoken contract, the idea of the rights to which all citizens were entitled, that was firmly rooted in society. The term "moral economy" has been rediscovered in political sociological theory recently (Svallfors 1996b: 17-20; Mau 2003: Chap. 3 ). In different contexts, researchers have considered it useful for complementing a pure selfinterest perspective on preferences and attitudes. These authors argue that people's notions of social institutions are guided by normative ideas of reciprocity, obligation, and responsibility, which cannot be reduced to a mere question of who is the winner or loser in different processes of distribution. In forging this moral economy, the role of public policies and political institutions is paramount, in that they "influence the ways individuals understand 1

2

The Moral Economy of Class

their rights and responsibilities as members of a political community" (Mettler and Soss 2004: 61 ). The moral economy of present-day societies may therefore, to a large extent, be seen as resulting from normative feedback effects of public policies and formal institutions. A normative feedback mechanism is present where public policies and institutions provide citizens with a sense of not only what their material interests are but the desirable state of affairs. In other words, one could say that the welfare capitalist countries of today are profoundly influenced by a modern form of moral economy, centering on notions of the justice and legitimacy of social structures. The moral economy is stratified insofar as its contents may vary between society's different groups and strata. For the purposes of this book, we could talk about a "moral economy of class," in which the focus is directed on ideas held by different classes on the reasonableness and fairness of a certain distribution of resources or the ways certain societal institutions operate. To many, talking about class in the 21st century could seem a little outdated. Didn't classes belong to the old industrial society? What can class analysis add to our understanding of a society where increasingly fewer people actually manufacture things? Is it not more interesting to study people's lifestyles, sexualities, or capricious cyber-selves? Isn't all this talk about classes quite simply hopelessly passe? We will be returning to these questions in the concluding chapter. Meanwhile, however, it should be remembered that much of the talk about the dissolution of class on close examination seems to be just this: empty talk, more often than not rooted in a desire to conjure up a picture of a society free from fundamental conflict. The conclusion that class has played out its role would therefore seem a little hasty and based more on hopes and fears than on any systematic research. If what people did to earn a living was no longer important for shaping their lives or the way they look at the world, it would represent a radical societal transformation; any claim that such a change is actually taking place would seem, however, grossly exaggerated. Complaints about exaggerated ideas of society's radically new disposition do not of course carry any implication that the class society should necessarily be static, that its structure and effect on other social phenomena should remain constant or similar over time. On the contrary, it is important to understand how dynamic the class situation is, both on an individual level (with people crossing class barriers during their working lives and often attaining class positions different from those of their parents) and on a so-

The Moral Economy of Class

3

ciallevel (where entire industries and professions die out to be replaced by others). However, in the midst of all this dynamism, the relative pros and cons of particular classes tend to be reproduced, as do the attitudes, values, and opinions to which they give rise. Nor does identifying class as one of the keys to understanding society mean that one believes that classes and their representatives are the driving forces of history, that "the history of all hitherto existing society is the history of class struggles" (Marx and Engels [1848] 1967: 79). The class analysis presented here is based on different, more modest, assumptions. In a nutshell, my claim is that people's position in the social distribution of labor, or more specifically the employment relations that their jobs entail, has an immense impact on other aspects of their lives. This is why classes and their political expression are an important feature of the processes of change in modern society. However, this is not to claim that the conflicts between the categories discernible in this respect, or between the political representatives who become their self-appointed spokespersons, are the driving forces of history. There are several reasons why it now seems urgent to study the relationship between class and social attitudes. One is that for the past few decades we have been witnessing a paradoxical coexistence of recommodification and dearticulation in the class arena. In this context, the term "recommodification" denotes processes that, each in their own way, make the population more dependent on the market for its welfare (Breen 1997). This differs of course from country to country and region to region, but a few key aspects of this trend can be summarized in point form nonetheless: • a stronger dominance of global capital and financial markets at the expense of national regulation and legislation • more precarious employment relationships, owing both to increased unemployment and to more irregular and temporary forms of employment • retrenchment of welfare state programs, making them less effective in countering market inequalities • strengthened links between positions in the labor market and other resources, such as social insurance coverage and benefits, housing standards, and access to social services • a decline of institutions for collective wage bargaining • greater income inequality • stronger roles for "internal markets" (e.g., "buy-sell systems" and the decentralization of budget responsibility) and "outsourcing" in the private and public sectors

4

The Moral Economy of Class

These are tendencies that to a greater or lesser extent can be found effectively throughout the Western world. Nonetheless, national variations have been considerable and the Western political economies have shown almost no signs of convergence on institutional organization and distributive results over the past few decades (Kitschelt 1999; Scharpf and Schmidt 2000; Pierson 2001). This is attributable to the way developments toward "more market" have been most manifest in the countries that originally had the strongest market liberal orientation, such as the United States and Britain. In other countries, including Germany, the changes have been much less marked. Just what the consequences of this recommodification will have for class is a matter of dispute. As will be seen in Chapter 2, some commentators argue that it will lead to a state of "individualized inequality," whereby inequality is made more pronounced yet less conformant to class or other such boundaries. Others argue that these developments will lead to a reinforcement of class differences. A parallel, and in some sense paradoxical, development to recommodification processes is the contemporaneous dearticulation of the class content in political discourse. The decline of classic social democracy, as witnessed in its gradual dissolution into the social liberalism of "the third way" and "Die neue mitte" in the major western European social democratic parties, makes the class articulation from the left less evident than before. The marginalization of the trade unions in public discourse and policy, compounded in many places by their weakening membership and coverage, also contributes to the dearticulation of class in politics. In addition, the collapse of the neoliberal project in the United Kingdom, and the subsequent reorientation of parties of the right elsewhere, has deprived us of any real class articulation from the political right as well. It should be emphasized that, as in the case of recommodification, the dearticulation of class is a highly uneven development among the Western countries. Some countries, such as the United States and Canada, never had a strong social democratic party or trade union movement to begin with. In other countries, such as Australia and New Zealand, the labor parties were originally the ones to espouse most clearly various neoliberal policies. In Europe, the extent to which the class content of politics is dearticulated varies enormously but is apparently a common trend nonetheless. The dearticulation of class can be partially understood as a gradual ideological adaptation to the difficulties of pursuing redistributive policies.

The Moral Economy of Class

5

Many of the instruments that have been used to pursue such policies, such as extensive taxes, social insurance and welfare services, or collective wage negotiations, have become less feasible. This is put down to factors of both an international (e.g., greater capital mobility) and national (e.g., the opposition of employers and groups with strong market positions to redistribution or demographic changes that frustrate the financing of welfare policies) nature. These problems of actually pursuing policies with a clear class content have led to a gradual dissolution of class articulation also in public rhetoric and discourse. Political actors on the left may see little point in articulating a class content that they do not believe can be realized in practical policies; hence the adaptation to the demands of the global economy clearly visible in the various "third-way" policies espoused by European social democracy over the past few decades. Meanwhile, political actors on the right, together with the mainstream private media, have little interest in articulating· their own policy when the dominant trend anyway seems to be toward a gradual kowtowing to the dictates of the market. One of the principal questions this book poses is what the combined effect of recommodification and dearticulation will be on attitude patterns. One possibility is that the relationship between class and attitude will gradually strengthen. If this happens, it would suggest that the microsociological experiences that people have are more important than the different mediated representations with which they are fed. Another possibility is that the relationship between class and attitude weakens. This would suggest either that the microsociological experiences will no longer follow class boundaries so closely or that a political discourse in which class is bereft of meaning makes the experiences increasingly difficult to interpret as class-based. A third possibility is that certain types of attitudes, such as those that concern specific attitudes to different social issues, will be more clearly structured on class lines, while other attitudes, such as social identity or political sympathies, will become less class determined. Another important reason for focusing on class and its impact on attitudes and values at this point in time is that working life and social structures have undergone considerable changes over the past decades. New forms of organization and leadership at work, new demarcation lines in society (e.g., between immigrants and nationals or between different family types and life-cycle phases), and new modes of political articulation and mobilization are some of the factors that come into play here. A perhaps

6

The Moral Economy of Class

especially pertinent development is the increased female labor market participation and correlated changes in gender patterns in the family and household. These factors are of such a kind that in certain respects they cut straight across the established class boundaries and can be expected to introduce greater variation within the classes. If the classes tend to become increasingly internally heterogeneous, we could also expect the class structuring of attitudes to be all the more diffuse. The questions are, in a way that public debaters and commentators do not always realize, deeply empirical and cannot be settled on a conceptual level or with reference to casual observations of social debate and cultural life. They can be properly examined only through systematic, comparative, extensive research of the kind that this book tries to emulate. This book draws on two interconnected but still largely separate bodies of research. One is the comparative study of social attitudes and values, where several large-scale comparative datasets now exist, allowing comparisons across space and over time (for some recent discussions, see Kuechler 1987, 1998; Svallfors 1996a; Jowell 1998). The other is class analysis, where several decades of concerted research from various theoretical perspectives has resulted in an immense literature on the associations between class and various social conditions and processes, such as social mobility, educational attainment, living conditions, voting, etc. (for some important recent examples, see Erikson and Goldthorpe 1992a; Marshall et al. 1997; Evans 1999). The wealth of data and the already considerable amount of research regarding comparative social attitudes is still growing rapidly. The latest bibliography of research that uses data from the International Study Programme contains almost 2,000 items (Smith 2005). The dearth of data that still characterized this research area at the end of the 1980s is now a deluge. Although this is in itself evidence of the dynamism and sharp growth of the research environments, it has yet to be paralleled by a similar increase in qualified analyses and interpretations. As Jowell lamented, "Analysts of cross-national data frequently abandon offering explanations and interpretations in favour of league tables of distributions showing merely 'gee whiz' national differences" (Jowell1998: 168). While class analysis is generally far better equipped with theories and explications, there are problems even here. As pointed out by Breen and

The Moral Economy of Class

7

Rottman (1995), class analysis has succeeded in demonstrating the persistent association between class and a range of social outcomes; yet it has been far less successful in explaining exactly why such associations occur. Explaining what Breen and Rottman call "class conscious outcomes," of which social attitudes are considered a central example, appears to these authors to pose a particularly formidable challenge to future class analysis, "one of linking a material basis (differential rewards) to differential forms of consciousness" (Breen and Rottman 1995: 466). Despite some promising attempts (e.g., Evans 1993a), it is clear that little progress has been made in explaining just why social attitudes are linked to class. One of the aims of this book is to contribute to such an enterprise, with a particular focus on class patternings in Sweden. One possibly obvious purpose is, however, to show what needs to be explained, i.e., how class differences in attitude within and between countries manifest themselves in different attitude domains. As will be made clear in the next chapter, there is every reason to adopt a strategy of national comparison in order to analyze class differences in attitudes. There are naturally many ways of doing this, but the approach I have decided to take is to strategically choose a select cluster of countries instead of as large a group as possible. This is in order to be able to include a wealth of information about the countries' institutional structures and the contexts within which people form their attitudes, which is often essential to giving meaningful interpretations of differences and similarities in attitudes (Jowell1998). Class differences in attitudes are compared in four Western countries: Britain, Germany, 1 Sweden, and the United States. As will be seen in the empirical chapters, these countries represent the span of institutional and political-economic configuration that exists in the West. The United States but also to an increasing extent Britain are examples of liberal regimes in which the processes of distribution are heavily dominated by different kinds of markets. Sweden and Germany, on the other hand, are examples of market economies in which politics and institutions are based on, and apply, relatively radical modifications of market distributions. These intercountry institutional and political discrepancies will be examined in more detail in each empirical chapter. The reason for choosing these four countries is that they represent the variation we find within the advanced capitalist countries while being fundamentally similar enough not to preclude meaningful

8

The Moral Economy of Class

comparisons. An even more pragmatic reason for this selection is that the comparative attitude data available is comprehensive and of acceptable quality. In the next chapter we will be discussing how class is defined in this book. What is it that distinguishes the groups we call "classes" from other categorizations? The attitude dimensions on which the empirical analysis focuses will also be defined, and there will be a discussion on the conceivable mechanisms that link class and attitudes. I will also present my methodological strategy and give a brief sketch of the data upon which the analyses are based (a more thorough account is provided in Appendix A). Chapters 3 through 7 contain comparative analyses. The first four of these chapters deal, in order, with class differences in attitude regarding work and production, market and distribution, state and redistribution, and family and reproduction. Chapter 7 contains an analysis of "class and conformism," highlighting attitudes on morality and justice. Throughout, all five chapters compare attitudes found in Britain, Germany, Sweden, and the United States. Chapter 8 tests some of the questions and areas of special inquiry that characterize contemporary class analysis on Swedish data. Here we use a more refined breakdown of class than applied in the preceding chapters to see if any interesting variations can be found in attitudes within the class categories used in the comparison of the countries. We also examine whether it is the class position of the individual or the household that is the most appropriate categorization factor and whether any specific attitudes can be discerned in marginalized groups on the labor and consumer markets. The closing chapter ties up the themes and arguments of the preceding chapters in a discussion of class and social attitudes in a comparative perspective. Are some countries and some social attitudes more class structured than others? What possible explanations are there for the differences and similarities we find? Is class declining in importance, or are class differences becoming more sharply defined? Or do they remain as strong as before but in different manifestations? In brief, what is the present nature of the moral economy of class and in what direction is it heading?

CHAPTER

TWO

Class and Social Attitudes: A Framework for Analysis

What is a "class"? What is it that constitutes a "class"? What is specific about the groups of people we call "classes" compared to other possible groupings? From the vast amount of literature on classes in contemporary society, a number of different perspectives may be distilled. The line taken in this book and the one that I will be defending as the most rewarding is to see class as comprising the employment relations characteristic of different positions in the division of labor (see the discussion and example in Erikson and Goldthorpe 1992a: Chap. 3; Marshall 1997; Goldthorpe 2000b). According to this perspective, classes are aggregations of positions within labor markets and production units and can be operationalized through a combination of professional title and employment relations. First then, class positions are demarcated according to ownership of the means of production. Employers and the self-employed control their means of production, employees do not. Among the large and very heterogeneous group of employees, the most important demarcator is between those who are on a relatively short-term, reward-for-input type of contract (e.g., piece or hourly wages) and those who are on a more diffuse, long-term, prospective type of contract. The first of these types of contract is typical of the working class; the second, of the "service class" (Erikson and Goldthorpe 1992a: Chap. 2). Occupations and tasks differ along two principal dimensions, which form the basis for the different types of employment contract (Goldthorpe 2000b ). The first is the extent to which they require specific education-based knowledge and skills; the second is the ease with which they can be monitored by the employer. The less specific the human assets involved, and the 9

10

Class and Social Attitudes

easier the tasks at hand are to monitor, the greater the likelihood that a pure labor contract will be used. Where work is hard to monitor and involves quite specific education-acquired skills, it is likely that a service contract will be used. Mixed forms of work (e.g., that not demanding specialist skills but hard to monitor) tend to be regulated by intermediate forms of employment contract, something that characterizes the "routine nonmanual" groups. This class schema has been the subject of numerous evaluations of its validity, in terms of both how effectively it defines positions with different employment relations and the extent to which such categories embody differences in social status over and above the working situation as such. In general, such evaluations tend to confirm the validity of the class schema; that is, the classes thus defined tend to differ with respect to employment relations (Evans 1992, 1996; Evans and Mills 1998a, 1998b, 2000) as well as to how the holders of these positions mutually differ in terms of various social outcomes (Erikson and Goldthorpe 1992a; Svallfors 1995, 1997, 2003a; Marshall1997; Marshall et al. 1997; Evans 1999). The reasons for choosing just this method of defining and indicating class can be further clarified by contrasting it with other ways of theorizing and operationalizing class (Erikson and Goldthorpe 1992a: 28-35; Goldthorpe and Marshall1992: 382-87). One such way is the ranking of class on a "status scale," on which the "prestige" or "status" of an individual is expressed as some combination of typical income levels, education levels, and social standing that comprise the particular ranking. The problem with such status scales seems to be that occupations found at similar levels on the scales often do not have much else in common other than this status rank and may indeed hold quite disparate locations within the social division of labor (Erikson and Goldthorpe 1992a: 30). Since class analysis should aim at capturing the social relations inherent in work and employment, and their implications for other social circumstances, such heterogeneity is unfortunate. Therefore, in the context of studying the relationships between class relations and social attitudes-as this book aspires to do-it would seem essential to work within a class schema that groups occupational positions with common characteristics beyond an artificial ranking on a status scale. One further unfortunate implication of the status attainment perspective is that it tends to treat all class differences as vertical differences along a onedimensional scale, ignoring nonhierarchical qualitative differences or class differences that cannot be reduced to a common scale.

Class and Social Attitudes

11

A third problem with status scales is that they combine education, income, and occupation in the determination of ranking, which makes it difficult to ascertain which one of them is the active component of the links uncovered. The choice between categorized class models and status scales therefore cannot be decided by testing which categorization "explains most" of the variation in various dependent variables. Regardless of whether the status scale can "explain" much of the variation in an independent variable, the links between work and employment relations on the one hand and social attitudes on the other are in danger of remaining diffuse, just like the mechanisms through which such links are maintained. The emphasis on employment relations also stands in stark contrast to more communitarian approaches to class analysis. Communitarian approaches are those that see classes as being defined by community formation and common worldviews; that is, they tend to adopt what is sometimes called a Gemeinschaftlich view on class (Thompson 1963; Holton and Turner 1989; Beck 1992: Chap. 3 ). Classes in this definition are aggregates of people who share location and worldviews and who tend to see themselves as a collective. If one chooses to define classes in such a way, then one has also to admit that classes have rarely been of any major societal significance. As Gordon Marshall points out, yesterday's communitarian, homogeneous, solidaristic classes seem to be more the product of fanciful conceptions than of any real social relations (Marshall1988: 202-7). Today it would appear almost impossible to detect traces of such strong "class communes." The fact that such homogeneous, communitarian classes do not exist has been taken by some researchers as confirmation that we should no longer even be talking in terms of class (e.g., Kingston 2000). However, it is an undue limitation of the use of the class concept to restrict it to such communitarian groupings. It would appear more fruitful to view communitarian relations as one (possible but far from likely) outcome of class relations rather than integrate such relations into the actual definition of class. One method of defining class that combines the weakness of the status scale and the communitarian approach is the "Cambridge scale" (Prandy 1990; for a critical analysis, see Evans 1998). The "Cambridge scale" is based on the pattern of friendships between people in different occupations. What should be seen as a result of class structuring-the roots of friendships in the division of labor-here constitute what class actually is. It is not clear

12

Class and Social Attitudes

the extent to which this scale reflects what we refer to here as employment relations. The many strong correlations between the "Cambridge scale" and other social outcomes are therefore rendered relatively uninteresting as we have no idea what the scale is measuring (apart from friendships, which can hardly be seen as a measure of class). A more persuasive version of the communitarian perspective is found in the writings of John Scott (Morris and Scott 1996; Scott 1996. See also Devine 1998; Bihagen 2000). Scott tries to steer a middle way between the "minimalist" conception of class as grounded exclusively in employment relations and the outright communitarian perspective. According to Scott, we need to distinguish between "economic classes," which are mere aggregations of typical economic situations, and "social classes," which are formed only when patterns of mobility, household formation, and social networks tend to cluster around these typical economic situations. According to Scott, the minimalist conception of classes produces only "nominally defined economic categories" (Scott 1996: 215). For social classes to exist, patterns of social interaction and mobility must follow from the aggregations of economic situations. The difference between the "minimalist" definition of class and the perspective of Scott and his followers might at first glance appear to reflect mere semantics: when should we call something "class"? But there is in fact a more fundamental issue at stake: the question of what to include in the very definition of class and what to treat as the factors that class might well influence. The minimalist conception of class regards the patterning of social interaction and life chances as regards mobility as phenomena that may be affected by the class positions and trajectories in which people are placed. Scott's perspective, on the other hand, builds such phenomena into the definition of what constitutes a class. I have chosen to adopt the "minimalistic" conception of class in this book since I consider it pointless to overload the definition of class itself. A more fruitful line of inquiry would be to have classes defined by the employment relations that typify different positions and leave as an empirical question whether class shapes other phenomena, such as social mobility and patterns of interaction. Class analysis would therefore be of little interest (1) if occupations and working-life positions were no longer structured in a way that can meaningfully be conceptualized as class differences; that is, if employment relations

Class and Social Attitudes

13

themselves dissolved, or (2) if class relations no longer affected other aspects of social life; that is, if power resources, life chances, or other social phenomena were no longer structured by class relations. As discussed in the previous chapter and as will be shown throughout the book, there is no empirical evidence to support these scenarios, and class continues to profoundly affect central aspects of social and political life. Even after deciding on such a "minimalist" definition of class, a number of difficult questions remain. One is how to demarcate classes, or, more precisely, decide which and how many classes we should distinguish between. Focusing on the nature of the employment contract yields three distinctly different class positions: self-employed, service class, and laborer. However, it is much harder to specify what characterizes the intermediate group between these last two, i.e., service workers in relatively subordinate positions (or routine nonmanuals). It is hard to identify a clear rationale for what is constitutive of this group of employees, although much research shows it is distinctively different from both the working class and the service class in terms of resources, patterns of social interaction and mobility, and attitudes and voting behavior (Marshall 1997: Chap. 3; Evans 1999; Bihagen 2000). The existence of qualitative boundaries between classes seems also to be confirmed by the validation tests of the class model referred to previously. The leading proponents of the "minimalist" class perspective tend at times to adopt a purely nominalistic stance on this issue, advocating that the number of classes should be "as many as it proves empirically useful to distinguish for the analytical purposes in hand" (Erikson and Goldthorpe 1992a: 46, n. 14). Sensible though this approach may be, it undermines the argument against using status scales and other such continuous methods of measurement. Even if the exact boundaries between classes can be blurred, a class definition (as opposed to a status scale) must be able to reveal qualitative differences between the "cores" of different classes. Taking a purely nominalistic line therefore risks undermining the entire theoretical structure. Another question relates to whether the individual or the family constitutes the most appropriate unit for class analysis. Erikson and Goldthorpe are fierce proponents of the idea that the family and household should be treated as class unitary (Erikson and Goldthorpe 1992a: Chap. 7). They argue that the individual experience of class is not confined to work or the workplace but carries over into individuals' lives more generally, especially as regards their economic security, stability, and prospects. When individuals live together as

14

Class and Social Attitudes

a family or household, their class experiences are mutually dependent (even if the household's resources are not always distributed equally). For this reason it is problematic to imagine class divisions within one and the same household. Erikson and Goldthorpe argue therefore that the class position of individual family members should be decided by a "dominance" method, whereby the class position of the family is determined by the individual with the strongest labor market attachment in terms of work time or work position and that all individuals in this family should thereafter be assigned this class position (Erikson and Goldthorpe 1992a: 265-66). In general, the arguments and evidence for this position are strong if not totally persuasive (see further Chapter 8). For example, married women's voting habits and class identity are to a greater degree determined by the class position of their husbands than by their own individual class (Erikson and Goldthorpe 1992b; Zipp and Plutzer 1996; Plutzer and Zipp 2001; Edlund 2003). However, in terms of the issues raised in this book, it may well be the case that the nature of the relations between class and attitudes depends on attitude area. For example, it could be assumed that attitudes to work are more strongly related to the individual's work and his or her immediate occupational situation, while attitudes to taxes, say, are more influenced by the class position of the family. What have so far been usually taken as dependent variables in the individual vs. family debate are class identity and political sympathies. These are factors that, more than pure attitudes, can be assumed to be the object of a joint, perhaps even negotiated, decision and the subject of family discussion. It would therefore seem advisable to let the issue of the most appropriate unit for class analysis be mainly an empirical one. In the empirical analyses that follow, however, I have been restricted by available data, which has made it hard to determine the class positions of husband or wife or cohabitant with any degree of confidence in many of the comparative datasets used. Instead, these comparative analyses are largely based on the individual's (most recent) profession as a determiner of class (see Appendix B). Having said this, a comprehensive dataset from Sweden that is analyzed in Chapter 8 makes it possible to compare on several points the results from class models based on individual occupation with those based on the dominance method. As can be seen, Goldthorpe's class model is operationalized through the classification of individuals on the basis of their employment status

Class and Social Attitudes

15

(self-employed vs. employees) and occupation. The occupation then becomes an approximation of the typical employment relations such a position embodies. An alternative would be to consult individuals directly about the demands, autonomy, reward systems, etc. that characterize their professional positions, and then, on the basis of this, sort them into classes. The well-known neo-Marxist class models developed by Erik Olin Wright (1979, 1985) are an example of this. 1 In principle, this approach would be preferable: instead of taking a circuitous route via occupational classification, we could characterize positions directly on the basis of their employment relations. It has, however, transpired that operationalizations deriving from direct questions are problematic and sometimes give rise to curious classifications (see the discussion and example in Marshall1988: Chap. 3). This is probably attributable to the fact that it can be difficult for interviewees to correctly or accurately describe their jobs in terms of the content that typifies themand to differentiate between this and their own individual qualifications. On the other hand, most people know what designation their jobs have. It would thus appear that a class definition based on occupational labels is to be preferred after all. For purely practical reasons, this is also the only viable option here, as none of the datasets analyzed contains enough information to define class in a Wright-inspired manner. A completely different objection to having occupation as an intermediate stop on the way to a definition of class is put forward by David Grusky and his colleagues, who question what the most appropriate level of aggregation is in the class analysis. They contend that class analyses have most often been pursued at an inappropriate level of aggregation (Grusky and S0rensen 1998; Grusky and Weeden 2001). According to Grusky, interests and identities are usually formed at the occupational level, rather than at the aggregated level. The class analysis should therefore take a step down in level of aggregation and focus on occupation as the principal factor of structure. It is only in rare cases, of which Sweden is the prime example, that classes are genuinely found at the level of aggregation of Goldthorpe's class schema. In the case of Sweden, one finds union organizations that have successfully created a general working class and a service class, in which occupations play a less relevant role than class. In most other cases, the classes are formed on a lower level of aggregation, where individual occupations rather than "large" classes are the most relevant units.

16

Class and Social Attitudes

The adoption of a more disaggregated approach to class analysis is certainly warranted in some respects, not least because it promises to enable us to better grasp the mechanisms through which class exerts its influence on social ·phenomena. The more disaggregated approach as advocated by Grusky and his colleagues nevertheless still remains at a highly programmatic level. Their claims that an aggregated class analysis, in which occupational titles are proxies for more fundamental aspects of employment relations rather than actual criteria for class stratification, is basically flawed seem exaggerated. Empirical research, as suggested previously, indicates that the class schema is quite valid and that the classes that are thus distilled really do differ from one another in quite critical respects. It seems that Grusky and his collaborators concede far too much ground to the end-ofclass thesis. It is also a matter of how far disaggregation can be taken. By way of example, the Swedish occupational schema NYK83 comprises some 1,500 individual occupations. It possibly goes without saying that a class analysis with 1,500 microclasses is unmanageable, necessitating always some form of aggregation. So far, however, Grusky and his colleagues have produced no principles or proposals as to what this aggregation is actually meant to look like. Even a schema containing only a few hundred occupational categories would require enormous amounts of data to be feasible. Moreover, the analytical complexity of the disaggregated models would probably quickly exceed our ability to interpret them in any meaningful way. While such strategies may certainly be developed, it remains an open question whether these models really will give us a better understanding of class and its dynamics than those that are currently in use. It therefore seems premature at this stage to abandon the class perspective outlined above. Another issue that should be discussed is the question of static vs. dynamic aspects of class, or class position vs. class mobility, in the formation of social attitudes (Turner 1992). This question may be broken down into several partially interrelated subissues. One concerns the effect of intergenerational social mobility on social attitudes: How does the family's class position during childhood affect present-day social attitudes? Another relates to the effect of mobility itself on social attitudes: Does the very experience of class mobility have a net influence on social attitudes once a person's current

Class and Social Attitudes

17

and childhood class positions are taken into account? A third is related to the effect of actual or anticipated working-life mobility: To what extent does one's working-life history, or the anticipated mobility along an occupational ladder or to or from self-employment, affect one's attitudes? 2 The existence of substantial social mobility introduces a large amount of heterogeneity in terms of mobility experiences into classes. As pointed out by Swift (2000: 669), class positions may themselves be thought of as "bundles of opportunities" and mobility as opportunities to access such positions. The total extent of class inequality is then the product of inequality between classes in terms of the opportunities they offer and inequality in terms of the chances of achieving such positions. Incumbents of identical present-day class positions may therefore have substantially different interests and hold quite different worldviews and social attitudes because of their actual or anticipated mobility. It is therefore important to take such mobility into account when analyzing and explaining variations in social attitudes. How do the tendencies toward recommodification as discussed in Chapter 1 affect the class structure? Although this is not entirely clear, two possible scenarios have been put forward. Aage S0rensen (2000) envisions a situation in which class relations more or less dissolve, as the rewards inherent in employment contracts approach a pure market distribution. Employment contracts thus become more like "spot market" contracts, in which the reward is equal to the marginal utility. "Structural support" for the workers in the form of collective bargaining agreements, internallabor markets, or general above-market returns on assets declines in importance. According to S0rensen, such a development is likely to increase inequality at the individual level at the same time as the class structure becomes less differentiated in terms of the distinct properties of different kinds of positions. 3 Goldthorpe (2000a) questions whether tendencies associated with the restructuring of labor markets and production units will actually produce the kind of "neoclassical" market outcome envisaged by S0rensen. Because of both the logic of production systems and the continuing ability of employees to make use of various kinds of collective action to affect their own circumstances, Goldthorpe would regard a relatively complex class structure as being an abiding, integral feature of the advanced industrial societies. There is no reason to expect that class conditions will become more diffuse or even dismantled.

18

Class and Social Attitudes

According to Goldthorpe, there is reason to believe that class divisions, especially those between the upper service level and other groups, will more likely be sharpened than rendered more diffuse. One outcome of this will be clearer differences and demarcations within the large and expanding group of nonmanual workers. While the lower-level nonmanual employees are likely to experience the downside of marketization-decreased employment security and increased work pressure-the opposite is the case for higherlevel nonmanuals and managers. They are likely to experience more stimulating work tasks and higher incomes from salaries as compensation for any increase in insecurity. The latter is furthermore unlikely to be overwhelming, since employers in "downsized" or "de-layered" organizations will depend increasingly on their higher-level employees. It is therefore likely, contend Goldthorpe and others, that the specificity of the service class will increase rather than decrease with time (Breen 1997; Goldthorpe 2000b: 224-27). Sorensen and Goldthorpe seem to agree, however, that groups of workers with weaker market positions are likely to suffer most in terms of security and living standards. The elimination of "rents" in the labor market, or a higher degree of pure market distribution, is detrimental to the well-being of those with a weaker market position, since "[r]ents are required in modern society to provide decent standards of living for the poorest part of the population" (Sorensen 2000: 1553; see also Breen 1997).

CLASS AND SOCIAL ATTITUDES: SCOPE, PATTERNS, MECHANISMS

After having discussed the best way to conceptualize and indicate class for our purposes, we shall now turn to the other half of the relationship between class and social attitudes. What is an "attitude"? And in what sense are attitudes "social" ? "Attitudes" are here understood as normative beliefs and opinions about particular social objects. 4 They differ from "values" in that they are tied to specific existing objects. They also differ from "positive" beliefs, which are beliefs about the actual state of affairs. While the class patterning of fundamental values or positive beliefs are interesting subjects for analysis in their own right (see, e.g., Breen 1999), they are excluded from the current definition of social attitudes. Attitudes are of interest here inasmuch as they concern opinions on the ideal nature of human interaction or social order in

Class and Social Attitudes

19

different respects. Strictly private attitudes and preferences concerning one's own private life or personal sympathies and antipathies are irrelevant to the purposes of this book. The answers to specific questions that constitute most of this book's dataset should be seen as indicators of underlying attitudes. For this reason, and for more methodological reasons as well (see Appendix D), the relationship between class and attitudes is usually described in terms of class differences on general-attitude indexes. Indexes summarize information from a number of individual questions and can therefore, if adequately designed, be seen as an approximation of underlying attitudes that are difficult or impossible to quantify. In analyzing the relationship between class and social attitudes, we need to delineate the scope of the attitudes we intend to study by asking what attitudes class might be expected to influence. Are any set of attitudes of interest for class analysis? If not, how can we delimit the set of relevant attitudes? There is, of course, nothing to stop a class analyst from surveying any set of social attitudes and describing how they correlate with class positions. This might well be of purely descriptive interest. But in order to attempt further explanation of the patterns we find, we need to develop a theoretically guided delimitation of the scope of attitudes. It would seem, then, that the wisest strategy would be to focus on attitudinal domains that for some sensible a priori reason may be assumed to have some kind of link (however tenuous) to class relations. What domains could these be? First, attitudes toward work itself and its rewards and costs must obviously rank high. Work orientations and commitment, attitudes toward the workplace and work environment, and views on wage bargaining and income distribution are all attitudinal issues that emanate directly from employment relations, making the class and attitude patterns and mechanisms we find in these areas interesting to analyze. Second, attitudes toward the redistribution of resources and life chances as effected through political intervention of various kinds are naturally also of interest. Those less endowed with crucial resources by virtue of their employment and market position, and who are more exposed to risks, are likely to differ in their views from those enjoying more fortunate positions in the primary distribution of resources and life chances. Third, issues related to reproduction and the family are also, in a derived sense, highly pertinent to class. Patterns of family formation and the

20

Class and Social Attitudes TABLE

2.1

Attitudinal domains Activity

Production

Distribution

Redistribution

Reproduction

Arena

Work

Market

State

Family

household division of labor could be expected to be influenced by the employment relations in which people find themselves; hence attitudes toward such matters might well be expected to be linked to class. In sum, we could think of the relevant attitudinal domains for class analysis as emanating from the activities and arenas summarized in Table 2.1. When comparing the attitudes in different classes, we should be interested in mapping the patterns of differences and similarities that occur. The ranking and relative differences between classes along the various attitudinal dimensions are of interest here. We should not necessarily expect the rankings and distances to be the same across different attitudinal dimensions. It is also important to study the dispersion and demarcation lines within classes. In this respect, class circumstances must not be the sole object of class analysis. Other factors such as gender, life cycle or generation, and religious affiliation need to be included in the analysis in order to study variation within classes. Nowhere is it written in stone that class is the only or even the most important factor behind attitudinal variation. But the main focus in a class analysis will remain the significance of class circumstances, with other factors brought in to explain within-class variation. Since the focus of this book is to elucidate class patterns across countries, comparatively little attention is put on the analysis of within-class variations. After mapping the class patterns, the next step in the analysis is to elucidate the mechanisms linking class and social attitudes (for a general discussion of social mechanisms, see Hedstrom and Swedberg 1998). What mechanisms have been at work to produce the patterns we observe? In general, it seems the two strongest candidates for explanatory mechanisms behind class and attitude are class-dependent interests and class-specific norms (Figure 2.1). The concept of "class interests" has a long and tortuous history within social science (Bertilsson and Eyerman 1979; Svallfors 1989: 24-29). Here I will circumvent the whole discussion of "subjective vs. objective interests," which seems to lead nowhere, and instead follow Erik Olin Wright (1985:

Class and Social Attitudes

21

[ Attitudes )

~ Figure 2.1.

Interests and Norms as

Mechanisms 28) in defining interests as "access to resources necessary to accomplish various kinds of goals or objectives." The matrix of risks and resources, in which actors find themselves through their employment relations, constrains and enables them in various ways. They obviously have an interest in increasing their resources and limiting their risks, regardless of what other goals they may pursue. Different class interests are then one obvious source of variation in social attitudes. A model based purely on interest would assume that members of different classes form their attitudes exclusively on the basis of conceivable objects of self-gratification (either immediate or deferred). In that case, one could also assume that classes that enjoy stronger positions on different markets and greater influence at work would differ from classes with weaker positions on the market and greater subservience at work. The former group would then presumably look more favorably upon the market's distribution mechanisms and upon their own work and the organization in which it is performed, and they would take a more skeptical view of the redistribution of what the market had once distributed. The second main source of variation in social attitudes between classes is that of class-specific norms. Such norms are to some extent based on interests, but they are modified by patterns of social interaction and communication. Norms do not change automatically as a result of changes in power resources and interests-if that were the case, there would be little point in proposing normative explanations at all-but are resilient to change. At an aggregated level, norms can be said to reflect the power struggles of yesterday as much as those of today. Furthermore, norms are nonconsequential in their underpinnings. They are not means to an end; they are ends in themselves (Elster 1989). Norms of fairness and distribution seem to be the most salient in terms of explaining class differences in social attitudes. Such norms can make distributions that have no direct benefit to oneself also seem legitimate. Other

22

Class and Social Attitudes

class-based norms of the kind that have no immediate connection with (re)distribution are also conceivable . The degree to which classes are shaped by superiority or inferiority at work or insecurity and exposure to competition on the labor market might well affect attitudes about normality and deviance. Those who carry out orders at work rather than decide on their own labors and who find themselves in a vulnerable position on the market can be expected to develop more rigid, narrow notions of how to live and thus exhibit greater conformism and moral conservatism. As explanatory mechanisms, interests and norms are not easily distinguished. Interest-based and norm-based explanations often overlap, since individuals and groups more readily cultivate and defend those norms from which they stand to gain. Struggles of interest are often fought using normative arguments, but if these norms carried no force independent of interests, there would be no point using them (Elster 1989). It is important to understand that the links between class, interests, norms, and social attitudes are of a probabilistic kind (cf. Goldthorpe 2000d). Class positions do not determine interests or norms, nor do interests and norms determine social attitudes. There is a large degree of contingency and randomness in both these relationships. However, class positions do tend to make some interests and norms more probable than others, and interests and norms do tend to make some social attitudes more likely than others.

POLITICAL INSTITUTIONS AND POLITICAL ARTICULATION

The mechanisms through which the magnitude and patterning of class differences are determined are modified by political institutions and by political articulation. The need to include the impact of political institutions in explanations of political outcomes and processes has been highlighted within institutionalist theories. 5 In both neoclassical economics and many sociological theories, we find an "institution-free world, where autonomous agents pursue their pre-defined interests" (Rothstein 1996: 14 7). This view needs to be replaced by a perspective that acknowledges that institutions to some extent "create or socially construct the actors' identities, belongings, definitions of reality and shared meanings" (Rothstein 1996: 147). Institutions are here understood as "the rules of the game" (North 1990: 3 -5) or, to use a stricter definition, "the formal rules, compliance procedures, and standard operating practices that structure the relationship between

Class and Social Attitudes

23

individuals in various units of the polity and economy" (Hall 1992: 96;

cf. Levi 1990). Institutions are hence something more than the constitution and the law, in that they also include more informal rules ("standard operating practices"). Still, it seems necessary to resist widening the definition of institutions so far as to include any set of structured interaction, thereby losing the possibility of distinguishing between the importance of formal political institutions and other social facts (Rothstein 1996: 145). The definition of institutions should therefore include only politically decided objects and their implementation. By this definition, institutions are thus systems of formal rules and procedures, manifested in phenomena such as social security systems, election systems, and family law. Social facts, on the other hand, such as family interactions, class structures, and norms, are excluded. Even based on this fairly narrow definition, it is clear that the presence and impact of institutions are immense in modern societies. In relation to class analysis, institutions are therefore important in the forming of class interests and class identities, and hence in the class patterning of social attitudes. There are a number of ways in which institutions work to this effect. First, institutions modify the structure of rewards and costs inherent in employment contracts through, for example, welfare state intervention or labor market legislation. This may be achieved either directly, through keeping employment contracts within certain legal limits, or indirectly, through welfare state benefits, services, and taxes. Second, institutions structure possibilities and incentives. Competition, recruitment, and social mobility are structured by political institutions, which therefore impinge on the incentives of social actors. While most institutional theories would concur that institutions affect rewards, costs, possibilities, and incentives, a third way in which institutions affect attitudes and actions is often disregarded in pure rational choice approaches to institutions 6 : institutions also affect perceptions and norms (Rothstein 1996, 1998). This is because institutions (1) affect the visibility of social phenomena, (2) affect what is considered politically possible to achieve, and (3) embody, and hence create, notions about what is fair and just (Rothstein 1998). Institutions provide "normality"; that is, they suggest to people what "the normal state of affairs" is and what is deviant or even unacceptable. A focus on institutions is necessitated by the need to get a grasp on what is historically variable and contingent in the relationship between class and social attitudes. Yet there is a risk of overemphasizing this. As argued by Pon-

24

Class and Social Attitudes

tusson (1995), fundamental structural properties of societies, such as the control and organization of production, are sometimes neglected in institutionalist analyses. Class relations, as embodied in employment relations, to a large extent constitute an extrainstitutional factor; they are relations that are modified but not created by political institutions. We could argue therefore that the relation between class and social attitudes is similarly modified by political institutions, but the bulk of research shows that fundamental patterns in attitudes do not differ widely between quite divergent institutional environments. A further source of historical contingency in the relationship between class and social attitudes is political articulation. To "articulate" something is to link it to something else, in this case class to politics. Political articulation, as conducted by political parties and other organized interests and often played out in the mass media, is by its very nature an attempt to change social attitudes and political preferences. We can distinguish here between two aspects of political articulation and class: one is a purely rhetorical appeal in which voters or other political subjects are appealed to. By presenting itself as a "white-collar" or "blue-collar" party or as a party that defends the interests of "small- and medium-size enterprises," for example, a political party can play a part in the elevation of class identity at the expense of other possible identities. The other way class can be politically articulated is to make class an aspect of political reform. A reform justified and designed to be particularly beneficial to laborers will, for instance, be discussed in terms of class and class will dominate the various fronts of conflict the reform encompasses. By using various forms of "symbolic politics" in this way, political actors hope to change or even create perceptions and attitudes (Edelman 1964, 1971, 1977, 1988; Svallfors 1996b: Chap. 3). The importance of political articulation and political entrepreneurship derives from the fact that actors may have difficulty in realizing what exactly is in their interest: Faced with a given issue ... actors may have trouble identifying their interests clearly.... Thus, on many of the most important economic questions of the day, it cannot always be said that the interests of a group or individual are "given" by their socio-economic position. On the contrary, those interests have to be derived via a process of interpretation. (Hall1997: 197)

One perhaps especially salient form of political articulation is the invention of "collective memories" through which events of today are filtered

Class and Social Attitudes

25

(Rothstein 2000). Collective memories supply actors with a sense of who they are and suggest what their interests should be and where their friend and foe are to be found. Such collective memories may be deliberately manipulated by strategic actors, in processes where "the past is reinterpreted for political purposes of the present" (Bates et al. 1998: 250). This is not to say that collective memories are infinitely manipulable, for to gain acceptance they have to concur with known historical facts and be of relevance to life experiences. In class terms, collective memories suggest to present-day actors where they belong and whether to think of themselves and others as members of a class at all. They suggest which classes share common interests and which have conflicting ones. Just as with political institutions, it is still important not to overstate the likely impact of political articulation. It seems unlikely that political articulation can break or reverse common patterns in the class-attitude nexus; rather, political articulation may act as a modifier of this relationship. Political institutions and political articulation are significant not only for their modifying effect on attitudinal patterns. Their role in the translation of social attitudes into social action is perhaps of even greater significance. While this is in itself an immense topic, which is beyond the scope of this book, a few words are in order. By modifying the opportunity structure, political institutions are paramount in making or breaking the link between attitudes and action. Simply put, the available choices determine what may be chosen. For example, a preference for private welfare cannot be translated into choosing private welfare if no private welfare supplier exists. Even short of actually limiting the range of available choices, the institutional framework may render some choices harder to make than others and thus affect the link between social attitudes and social action. Political articulation also plays an important role in modifying the link between attitudes and action. One of the most simple and yet widespread forms of social action is voting. As has been repeatedly stressed, voting is determined by not only values, attitudes, and preferences but also the options available to voters. If parties fail to present themselves in their manifestos and policies as class parties, there is no reason to expect the association between class and voting behavior to be strong, regardless of how strongly class impinges on social attitudes (Evans 1993b: 451-52). The proposed correlation between class, social attitudes, social action, political institutions, and political articulation can be summed up in a simple

26

Class and Social Attitudes

Figure 2.2. Political Institutions and Political Articulation

diagram (Figure 2.2). We can see from this that institutions and articulation are thought to affect the relationship between class and attitude as well as that between attitudes and action. However, it is only the first of these two relationships that is the concern of this book. Moreover, institutions and articulation can influence the general structure and level of attitudes regardless of the strength of their impact on the class-attitude nexus. We will be returning to a modified and elaborated version of this diagram in the concluding chapter. Once having stressed the significance of political institutions and political articulation, it would seem constructive to adopt a comparative, intercountry perspective in the study of class and social attitudes. By comparing the relationship between class and social attitudes in countries with different institutional setups and different traditions of political articulation, we will be able to determine, or at least illuminate, the extent to which these factors matter for class patterns in social attitudes. In so doing, we open up possibilities of demonstrating the impact of the intracountry degree of dearticulation on the field of class politics, which we touched upon in Chapter 1.

METHOD AND DATASETS

The terms, assumptions, and issues in the preceding discussion will serve as the basis of the empirical analyses of the following chapters. It must, however, be remembered that these issues will be addressed to different degrees by the specific analytical discussions, some appearing as points of departure and discursive backdrops. We will be returning to such issues in the final chapter; first, though, an empirical foundation must be laid.

Class and Social Attitudes

27

We now have to ask ourselves what kind of data we should use to make our analyses using the framework sketched. The first thing to stress is that well-founded conclusions must be based on survey data and representative population samples. There are two reasons for this. First, it is postulated in the analytical framework that the relation between class and attitudes is probabilistic in kind (cf. Goldthorpe 2000d). This means that the argument on the nature of the class-attitude relationship cannot be substantiated or dismissed by relying on case studies or other kinds of smaller and less intensively studied cohorts. Second, it is important that data do not derive from any particularly distinctive or highly specific group but represent the entire (or in certain cases the working) adult population. The need to use relatively large representative samples necessarily limits the type of information that can be gathered to fairly standardized indicators. The sort of data that must be used are thus of a rather thin kind, whereby only a few indicators are used to illuminate more broad-based attitude patterns. The fact that we must restrict ourselves largely to survey data, as is the case here, does not mean that other information or data sources are considered redundant. Two kinds of complementary data also seem particularly interesting. One is detailed ethnographical data on strategically selected groups of workers, useful for revealing mechanisms and providing explanatory backup (Goldthorpe 2000c). The "wide-but-thin" data obtained from surveys can thus be supplemented by "narrow-but-thick" data from groups judged of particular interest to a specific line of reasoning (see further Chapter 7, pp. 142).1t is, however, crucial that such "intensive" group-specific data be selected from the broader perspective provided by the results of surveybased research if we are to take into sufficient consideration variations between places and groups. 7 Another important kind of data comprises contextual and institutional information on the countries and contexts being compared. If we are to interpret results from comparative research (in this case explain and interpret why there are similarities and dissimilarities in the class-attitude relationship in different countries), we need to have a solid grasp of how institutions and political discourses differ from one country to the next. In this respect, several kinds of data are worth considering, from historical accounts to analyses of contemporary political discourses and data on income distributions

28

Class and Social Attitudes

and welfare state rights. The empirical chapters of this book will make frequent use of such information from secondary sources. The primary source of data for our analysis will be survey material nonetheless. What information should we then search for in such studies if we are to be able to use our analytical framework? The first, and most obvious, is a reliable measure of class, such as occupation and form of employment; but we also need detailed information on the degree of freedom and power of decision making offered by the job in question, as well as the level and form of remuneration. This will allow us to determine with precision the aspects of class that are the most salient in terms of structuring attitudes. As much information as possible relating to occupation and employment status should be available for not only the individual's current job but also any previous ones (this is understandably of particular importance for those temporarily out of work), as well as for his or her spouse or partner and parents. Such information is needed to study the effect of working life and intergenerational mobility and to compare the influence of an individual's class position with that of his or her "family class." The second obvious target of our search is a broad set of proxies of relevant attitudinal domains, which should cover the activities and arenas previously described in Table 2.1. They should also be available as time series, or at least from more than one measurement occasion, and ideally as longitudinal data that allows individuals to be traced through a succession of social processes. Third, such data should also provide comprehensive details of positions in other critical dimensions of social division and of resources, level of living, and life trajectories. This makes it possible to analyze the extent to which such factors operate as the link between class and attitudes; it should also allow us to assess the relative significance of class differences compared with other stratification dimensions. Finally, such survey data should derive from strictly comparative studies, which will enable an analysis of the influence of political institutions and political articulation on the class-attitude nexus. Ideally, then, data should come from countries categorized as democracies with capitalist market economies but that differ nonetheless in terms of institutional frameworks, the relative power resources of political actors, and the nature of the political discourse.

Class and Social Attitudes

29

Reading this desiderata, it becomes clear that no existing-and perhaps no conceivable-dataset could satisfy all wishes, and the researcher will have to make do with ones that fulfill some of the criteria and attempt to use them in some innovative combination. The datasets that form the basic material of this book satisfy to varying extents and in varying ways the possible requirements (see Appendix A for a fuller discussion on the datasets). The comparative surveys carried out within the International Social Survey Program (ISSP) are conducted annually in different areas and among an expanding group of countries (currently 40). These studies provide a wealth of relevant attitude data of considerable content breadth. The time series of greatest interest to this book are found in the following four modules: The Role of Government (1985, 1990, and 1996), Social Inequality (1987, 1992, and 1999), Family and Gender (1988, 1994, and 2002), and Work Orientations (1989 and 1997). The downside of the ISSP studies is that background data are sometimes not sufficiently detailed or standardized (this is particularly true of the earlier surveys). Occupational data are sufficiently detailed in the later studies to construct class categories for the respondents themselves but not always for the spouse or partner or parents. 8 Data on resources, employment conditions, etc., are rather sparse, rendering the link between class and attitudes illuminable to only a limited degree. Data from some ISSP countries (although not those used here) also are encumbered with nonrepresentative samples and extremely low response rates. In order to obtain longer time series and more frequent replications, I have used data from the Swedish Welfare State Surveys (Svallfors 1989, 1996b, 1999). The absence of corresponding surveys in the other countries precludes any intercountry comparison. A third type of survey of value is (for Sweden's part) more general levelof-living surveys, such as those carried out by Statistics Sweden (the ULF surveys) and by the Swedish Institute for Social Research (the LNU surveys). These surveys provide a wealth of data on levels of living, resources, employment conditions, etc., and are of a high quality as regards sample populations and interview material. However, the attitude data they contain is sparse, leaving only a small sample of attitudes open to analysis. This book uses data from these ULF surveys; normally the ULF surveys contain no attitude data at all, but the one conducted in 1998, which I will be using in

30

Class and Social Attitudes

Chapters 5 and 8, contains supplementary questions relating to attitudes to the welfare state and distribution. To sum up, none of the data used in my analyses is problem-free or perfect (see Appendix A). What we have to do is put together the pieces that we hold to build up a picture of the complex relationship between class and attitudes. However, such piecemeal construction, from which conclusions are drawn from data that are rarely ideal, is the stuff of everyday life to the social scientist and is neither unique nor particularly problematic for this book.

CHAPTER

THREE

Work: In the Sweat of Their Brows?

There is good reason to have attitudes toward work serve as the introduction to our empirical analyses. As we saw in the previous chapter, work is the basis upon which we define and classify class: attitudes to what lies at the very heart of the concept of class are therefore a natural point of departure. It is not unreasonable to assume that classes would differ in their attitudes to work. Since power and influence at work, such as workload and reward, differ from class to class (indeed it is almost integral to the definition of class), we can also expect people from different classes to possess divergent attitudes to work. We can also expect that those, by virtue of their position, who have more power over their own and other people's work and receive greater material and immaterial rewards through the work they do will also develop a more positive outlook qn both tpeir own work and work in general. However, this is just a very general assumption that can be supplemented with a number of questions to which the answers are less obvious. First, we must ask ourselves what it is about work that people have attitudes about. Studies on attitudes to work differentiate between a range of dimensions (Mowday et al. 1982; Warr 1982; Lincoln and Kalleberg 1990). This chapter will deal with attitudes toward the value of having a job in the first place and how it is perceived by the incumbent, as well as job satisfaction, commitment to the company or organization's goals and values, and experiences of workplace relationships. One question that could be asked is whether class patternings will resemble each other in all these areas or whether they are especially salient in certain respects only. Second, we can ask if class differences in attitudes to work depend on institutional context or, more specifically in this case, if they differ among the

31

32

Work

four countries in our study. Comparative research into the actual organization of work, in which authors have pointed out the major differences in typical production environments and work structures in each country, suggests that this could indeed be the case. Dobbin and Boychuk (1999) differentiate between what they refer to as "rule-oriented" and "skill-oriented" production systems. In the former, the management focus is on the close, detailed control of the content and goals of the work being performed; vocational training is rare and employees are expendable, something facilitated by a relative ease of dismissal. As for the latter, employees are free, within limits, to decide how their own work should be performed and vocational training is common both within the normal educational system and at the workplace, something abetted by the relative difficulty of dismissing employees and recruiting new ones. Dobbin and Boychuk contend that rule-oriented production systems are mainly found in the Anglo-Saxon liberal market economies, while the most salient examples of skill-oriented systems are found in Scandinavia. The manner in which work is organized in a particular country is contingent upon the wider institutional environments that characterize it. Fligstein points out that work relations are embedded in institutional "employment systems," which he defines as "the rules governing relations between groups of workers and employers that concern the general logic of how 'careers' are defined and how groups organize to maintain these conceptions" (Fligstein 2001: 101). He identifies three principal employment systems: "Vocationalism," in which the unions have a strong presence and vocational training and career building are common; "Professionalism," of which professional associations and recruitment via higher education are key features; and "Managerialism," in which the company or organization is the dominant arena and internallabor markets and recruitment through tertiary education are common. Fligstein stresses that all existing national employment systems actually comprise a mix of these three principal types, but that the relative weighting of each can differ radically; for instance, the United States is dominated by a combination of professionalism and managerialism, while Germany is dominated by vocationalism. There is also good reason to assume that the work and the labor market climates would differ significantly from one institutional environment to another. In the rule-oriented and managerial production environment, the aim is to achieve low workforce costs and so we could expect the work

Work

33

organization to be more authoritarian and subject to unilateral management control. As regards the skills-oriented, vocational systems, we could expect to see completely different efforts to involve the workforce in production and the relations between management and employees to embody a greater spirit of cooperation and negotiation. There is also in such production environments a more prominent element of job creation and training schemes designed to sustain the reproduction of a skilled, motivated workforce. Classifications of this nature can be nothing but simplifications. It is extremely difficult to identify comparable indicators and to generalize from specific studies of individual industrial sectors. What is clear, however, is that Sweden and Germany have extensive systems for vocational schooling and the reschooling of workers that are designed specifically to maintain and replenish the workforce, while in both Anglo-Saxon countries corresponding systems are remarkably weak (Blossfeld 1992; Weir 1992; King 1995; Shavit and Muller 1998: Chap. 3, 5, 12, 15; Estevez-Abe et al. 2001). On the other hand, we know less about the extent to which there are any systematic differences between countries in the organization of work and the autonomy and codetermination that accompany it. We cannot simply assume that the wider institutional environment has a genuine impact on how work is organized, and comparative research into the area seems somewhat neglected, as Gallie and colleagues lament after reviewing the field (Gallie et al. 2003 ). For instance, conclusions regarding differences in work organization between Germany and other industrial countries seem to be largely based on experiences drawn from case studies of advanced industrial production (Soskice 1999: 115-17), and just how much they can be applied to working life in general is unclear. Nevertheless, the cumulative research would suggest that Sweden is profoundly shaped by a more cooperative, participatory approach to the organization of labor, while Britain and, in particular, the United States are typified by more authoritarian and unilaterally controlled work organizations. Dobbin and Boychuk's comparative study substantiates the notion that autonomy at work is clearly greater in the Scandinavian countries than in the Anglo-Saxon, even when keeping all other conceivable influences constant. The observation that the quality of work and codetermination at work is higher in Scandinavia than elsewhere is also supported by Gallie (2003 ), who compares Sweden and Denmark with other European Union (EU) countries. This can also be taken as corroboration of Dobbin and Boychuk's

34

Work

thesis that the type of production system in place impinges on the content of the day-to-day work performed. However, one would be hard put to find evidence of Germany having a radically different work organization than Britain, and indeed Gallie does not find this to be the case (Gallie 2003). Moreover, one of the few extensive comparative surveys in the field shows that the Germans hardly have systematically greater autonomy and codetermination at work than Britons. Different indicators point in different directions, and both countries hover around the EU average in such matters. Meanwhile, Swedes enjoy a higher degree of autonomy and codetermination than either of these nationals. Common to all countries is that higher-level nonmanuals have greater autonomy at work than workers and have a greater say in the decision-making process. Class differences in Sweden are, however, less salient than they are in Germany and Britain, where variations in the scale of class differences are hard to discern. 1 How can we expect the different ways of organizing production and labar markets to affect attitudes toward work, and, more specifically, on class differences in such attitudes? At first glance, we might expect that the participant"centered, skills-oriented production systems would foster more positive attitudes to work as well as narrower class differences. But even though production systems that encourage greater codetermination and autonomy could well have just this effect, things are not that simple: these types of political economy also tend to have much more powerful and ideologically more articulated unions than the managerial and rule-oriented systems (see the different contributions to Hyman and Ferner 1998). All else remaining constant, we would expect that strong unions encourage more of a partisan mentality (whereby the interests of the management only partly coincide with those of the employees) among the employees than weak unions are able to do. In the more liberal market economies, the management of a company or organization has much more room formaneuver when it comes to trying to shape the employees' opinions. This is clearly seen in the management strategies that prevail in different European countries, where unilateral management-controlled incentive systems are more common in Britain than in continental Europe and Scandinavia, the latter favoring more cooperation-oriented human resource management (Gooderham et al. 1999). In many respects, the attitudes examined in this chapter indicate whether an employee has assimilated the goals and values that the company

Work

35

TABLE3.1

Employment commitment in Sweden, Britain, Germany, and the U.S. (labor force only)

A job is just a way to earn money-no more I would enjoy having a paid job even if I did not need the money

Sweden

Britain

Germanya

u.s.

%agree Scale mean

18 2.65

34 2.16

27 2.43

22 2.53

%agree Scale mean

75 2.75

58 2.30

74 2.61

65 2.49

"Refers to the former West Germany.

represents and whether he or she finds the work and the workplace atmosphere satisfying. We can therefore assume that strong unions can actually contribute to a more negative image of the organization or company and stronger class antagonisms. In Sweden's case there is thus reason to assume that greater codetermination and autonomy generate more positive attitudes to work, while the stronger union presence promotes more skepticism toward the company's or organization's way of operating.

THE VALUE OF WORK

The data used in this chapter are taken from the 1997 ISSP survey (see Appendix A). Here, the interviewees were asked about their attitudes toward a range of different work-related aspects. To begin with, class differences are analyzed as regards the value of having a job in the first place and the perceived benefits it brings. The value of having a job, over and above the need to provide for oneself, is referred to in the literature as "non-financial employment commitment" (Warr 1982; Nordenmark 1999), a term that denotes the individual's appreciation of the intrinsic value of work. Table 3.1 shows the two statements of relevance to this from the ISSP survey, 2 the percentage of respondents who agree ("strongly" or otherwise) to each of them, and the mean values. The mean value has been so calculated that a higher value denotes a stronger belief in the intrinsic value of work. So for the first statement, those who "agree strongly" score 0 (the lowest value), those who "agree" score 1, those who "neither agree nor disagree" score 2, those who "disagree" score 3, and those who "disagree strongly"

36

Work

80

•

75 70

e

65

+

• •

•

D

.6.

0

*

60

0

li1

55 50 45 40 Sweden

+ Service class I .6.

Figure 3.1.

Service class II

Britain

Germany

+ Routine nonmanuals D

Skilled workers

U.S.

* Unskilled workers 0

Self-employed

Employment Commitment Index

score 4 (the highest value). For the other statement, the responses have been coded in the reverse order. 3 As can be seen in the table, the belief in the intrinsic value of work is most widespread in Sweden and least so in Britain. The difference between the countries is, however, relatively minor. To obtain an overall gauge of attitudes toward the intrinsic value of work, both indicators have been condensed into one index, which is then divided by its maximum value and multiplied by 100. 4 The index can subsequently vary between 0 and 100, whereby the higher the value, the greater the appreciation of the intrinsic value of work. Figure 3.1 shows the class differences in the four countries. As we can see from Figure 3.1, we find similar class patterns in all four countries. 5 The greatest emphasis on the intrinsic value of work is found consistently in service class I and least among the manual workers. Our initial assumption that those with more privileged positions at work are more likely to develop a favorable view of work per se is clearly substantiated by the results as shown in the figure. Moreover, class differences in all countries, with the

Work

37

TABLE 3.2 job characteristics in Sweden, Britain, Germany, and the U.S.

My job is secure My income is high My opportunities for advancement are high My job is interesting I can work independently In my job I can help other people My job is useful to society a

%agree Scale mean %agree Scale mean %agree Scale mean %agree Scale mean %agree Scale mean %agree Scale mean %agree Scale mean

Sweden

Britain

Germanya

u.s.

57 2.47 17 1.52 20 1.66 75 2.85 83 3.02 65 2.67 67 2.81

52 2.30 17 1.48 20 1.58 71 2.72 79 2.89 69 2.74 57 2.57

70 2.81 27 1.82 19 1.51 83 3.05 86 3.14 50 2.29 64 2.67

70 2.76 25 1.75 31 1.93 74 2.83 81 2.98 80 3.02 71 2.88

Refers to the former West Germany.

exception of the United States, are considerable-almost 20 points on an index that can vary between 0 and 100 (see note 4 about the reliability of the index in the United States). So if classes differ widely in terms of attitudes toward the intrinsic value of work, is the same true for how they look upon their own jobs? The ISSP survey required the respondents to judge the extent to which a group of statements applied to their own jobs (Table 3.2). The mean value for each statement was calculated by giving the "agree strongly" response a value of 4, "agree" a value of 3, "neither agree nor disagree" a value of 2, "disagree" a value of 1, and "disagree strongly" a value of 0. It should be stressed here that what we are interested in is the individual's own judgment of whether he or she has a high income, an interesting job, etc. This self-appraisal is of course not just contingent on the actual nature of the job in question but also on the individual's own levels of expectation. One person's fat pay check is another person's chicken feed. As an indication of how people view their own work, it can still hold some interest for us. Table 3.2 shows the differences between countries and between indicators. Many people claim to have interesting jobs that allow them a degree of independence, while relatively few say that they have a high salary and good career opportunities. The Germans seem to be more satisfied than the others

38

Work

on several of the indicators, but all in all the national differences are small and asymmetric. Is there any link between the different aspects, so that those who consider themselves to be well paid also claim to have interesting jobs, and so on? Correlation and dimensional analyses reveal similarities and dissimilarities between the countries in terms of how the different indicators are interrelated. In Sweden and Germany, the last two indicators have only weak links with the others and form their own separate variable when subject to factor analysis. 6 In Britain and the United States, on the other hand, the last four indicators tend to correlate while the first three form their own separate grouping. These dimensional analyses suggest that the root of the national differences in attitude structures lies in the last two statements; and as they do not lend themselves easily to interpretation, in that the respondents are invited to judge the value of their work to other people, there is good cause to eliminate them from our comparison. A factor analysis that excludes these last two indicators produces a one-factor solution in all countries except Britain, where statements four and five form their own factor. In order to create an overall measure of what people judge they have at work, the values for the first five statements in Table 3.2 have been aggregated. The index was then divided by its maximum value and multiplied by 100 to produce a value that varies between 0 and 100. As this gives an index of acceptable reliability, even in Britain, there is reason to assume that the different indicators measure similar things. 7 The design of the index thus means that those who "agree strongly" with all five statements in Table 3.2 score 100, while those who "disagree strongly" with all five statements score 0. Another interesting point of inquiry is the correspondence between what people feel they derive from their work and what things they consider important in a job. A low income might not matter for those who do not consider it important to earn a high income. A lack of something becomes a problem only if you want it in the first place. To quantify this, I used the fact that for each of the indicators in Table 3.2, respondents were asked (much earlier in the questionnaire) to state how important these things were at work. The responses were recorded on a five-point scale, from "not at all important" to "very important," and then recoded as values of 0 (not at all important) to 4 (very important). We can now construct a kind of deficit coefficient for each of the statements and questions from the responses to both batteries by subtracting

39

Work

80

0

75

-5

~

0

•> 37.7

0.62''*'' 2.4

27.9

IV

• Refers to the former West Germany. *** = p < 0.001; '"' = p < 0.01; * = p < 0.05.

model V also incorporates the "job-satisfaction" index, from which we can see what happens to the class and national differences if we not only keep the "have" and "relations" indexes constant but also assume that people are equally satisfied with their work regardless of job and country. In the first column in Table 3.6, we can see the very modest class and national differences as regards job satisfaction. The second column tells us what happens when we take into account what people believe they have in their jobs. Interestingly, what emerges here is that the nonmanuals, after control for what people believe they derive from their work, evince a lower degree of job satisfaction than the workers. The third column shows us, however, that the class differences are largely insensitive to how one perceives workplace relations. The last column shows that, controlled for both the "have" index and the "workplace relations" index, there are no significant class differences at all. The fact that the workers are more satisfied with their jobs than service class I when keeping constant their perception of what they derive from work suggests that there is a certain measure here of adaptive preferences. The upper nonmanuals have, quite simply, higher expectations than the

48

Work TABLE3.7

Employment commitment index (0-100). Regression analysis (ordinary least squares). Unstandardized regression coefficients MODEL

Constant Service class I Service class II Routine nonmanuals Skilled workers Self-employed (unskilled workers reference group) Sweden Britain Germanya (U.S. reference group) Have index (0-100) Relations index (0-100) Rz (%) a

II

Ill

IV

55.2'"'* 15.3""'"'' 10.6'"""* 6.3'"'"* 0.3 8.5""'"*

42.4***

45.4***

38.5***

11.8*** 8.2*''* 5.7*** -0.5 6.1 ***

15.5''*'' 10.7**'' 6.3*''* 0.9 7.3 * **

12.2**'' 8.5*''* 5.8'"''* -0.1 4.8**

5.6"'** -5.9''"** 1.2

6.1 *** -5.0**'' 0.9

7.3*** -6.7'"** 0.8

6.1 **'' -5.6**'' 0.7

0.24*"* 0.13*'"" 9.4

12.2

10.9

0.22*** 0.07** 12.9

Refers to the former West Germany.

workers, and more is needed for them to be able to claim job satisfaction. The very small differences in job satisfaction that we saw in Table 3.3 can thus be interpreted as the cumulative effect of the way members of the service class have the better jobs and higher expectations than the workers. Table 3.7 shows the very large class differences found in opinions toward the intrinsic value of work and the substantial national differences. The class differences narrow somewhat when we take into account how people perceive their own work, but only slightly. National differences do not shrink at all when we include this factor and workplace relations; on the contrary, they actually grow slightly. Swedes have by far the highest values on the index, while Britons have the lowest. Model I in Table 3.8 shows the considerable class differences in organizational commitment that we could also discern in Figure 3.3. It also shows that the United States scores higher than the three European countries. The table's second model shows that the class differences virtually disappear when one takes into account what people claim to derive from their work. The reason why the service class shows more commitment to their organizations has thus very much to do with how they consider themselves to derive much more from their jobs than do the workers. On the other hand, the

49

Work TABLE 3.8 Organizational commitment index (0-100). Regression analysis (ordinary least squares). Unstandardized regression coefficients MODEL

Constant Service class I Service class 11 Routine nonmanuals Skilled workers Self-employed (unskilled workers reference group) Sweden Britain Germanya (U.S. reference group) Have index (0-100) Relations index (0-00) Job satisfaction (0-100) R2 (%) a

11

Ill

IV

V

56.4***

24.0***

28.5***

12.1 ***

5.3**''

10.5*** 6.1 **'' 0.3 1.7 14.9***

2.9** 1.6 0.2 -0.2 9.3**''

10.5*''* 6.6*** 0.5 2.3 13.4* * ,,

4.2*** 2.7** 0.4 0.5 9.0*''*

5.0*'"'' 3.0''*'' 0.7 0.7 7.9**''

-5.2*** -3.6**'' -3.3*'"

-3.3*** -0.7 -4.0**''"

-4.7**'' -4.4''** -3.7*'"*

0.58*'"* 0.37''"*'' 9.5

34.7

23.9

- 3.4* ,. * -3.6*'"" -1.6 -0.9 -4.1 *** -2.6'"** 0.49* * * 0.22*H 39.6

0.29'"** 0.09'"** 0.40''** 51.3

Refers to the former West Germany.

*** = p < 0.001; ** = p < 0.01; * = p < 0.05.

self-employed still score much higher than the workers, even controlled for the perceived benefits of their work. Table 3.8's third model shows that class and national differences are hardly affected by the inclusion of appraised workplace relations despite the fact that the latter has a powerful effect in itself on commitment to the organization. Finally, we can see from the table's last two models that certain class and national differences in terms of commitment to organization remain, even controlled for what one has at work. Even if we were to assume that the classes do not differ in terms of what people believe they derive from their jobs or their own job satisfaction, certain class differences still remain. It is not unreasonable to interpret these remaining class differences as an expression of a greater distance from or stronger antagonism toward management among workers than among the upper nonmanuals. It should, however, be remembered that such differences are only minor since most of the class differences in commitment stem from what people feel they derive from their jobs (cf. Hult 2005). There is also cause to comment on the strong relationship between what people believe they derive from work and perceived work relations on the one

50

Work

hand, and job satisfaction (Table 3.6) and commitment to the organization (Table 3.8) on the other. Each unit change on the "have" index is matched, for example, by a 0.62 unit change in job satisfaction (Table 3.6, model 11), and the other attitudinal connections are almost equally as strong. These relationships are, of course, not to be compared with those we find between class and attitudes; rather, they are an expression of a correlation between different attitudinal components that raise questions about the direction of the causal relationship. It is, however, interesting to note just how strongly related the different aspects of how people appraise their own work situations actually are, since this would suggest that the attitudes to work analyzed in this chapter are clearly structured and do not arise by chance.

CONCLUSION

With all these results in our possession, what can we say about class differences in attitudes toward work in a comparative perspective? First, we find corroboration for our original assumption that the classes would genuinely diverge in attitudes toward work. Whether we are looking at attitudes toward the intrinsic value of work, people's appraisals of their own jobs, or commitment to the company or organization, class differences are considerable in all the four countries studied. On the other hand, this is not the case for differences in job satisfaction, something that can be interpreted partly as the result of each class having its own levels of expectations. And as regards attitudes toward work place relations, there are hardly any class differences at all. We can also find differences between the countries as to the nature of the class differences. They are particularly large in Sweden when we consider commitment to the company or organization and how people appraise their own work. They are similar between the three European countries for attitudes toward the intrinsic value of work. Moreover, when we juxtapose the class differences and the aggregated levels, we find that the large class differences as regards organizational commitment in Sweden coexist with a relatively low degree of commitment to the company or organization compared with the United States. As regards the intrinsic value of work, however, we find that Swedes exhibit the very strongest commitment. Both these results can be seen as a measure of support for the differences in production systems and work organization discussed at the beginning of the chapter. The stronger commitment to the intrinsic value of work in

Work

51

Sweden could arguably be the result of a production and work organization that places a stronger emphasis on participation and imposes a less authoritarian style of leadership than the Anglo-Saxon countries. The weaker commitment to the company or organization and the relatively large class differences can be interpreted instead as an effect of the stronger union base, which we can assume fosters more distinct partisan attitudes among the workforce. Furthermore, employees in Sweden are only lightly dependent on their employers for their insurance protection and access to welfare services. Overstating things a little, we could say that many Americans "have to" be committed to their employer, as it is the only organization that cares about their welfare (cf. Hacker 2002). However, we should not put too much weight on these national differences. All things considered, they are fairly small, both on an aggregated level and in terms of the mutual order of and differences in magnitude between the classes. The relationships between the institutional and organizational environments in which people work and the attitudes to the jobs they do have are no more than weak tendencies. Class differences in attitudes toward work are found in the more skill-oriented, participatory production environments and in the more rule-oriented, unilateral managementdominated work organizations. The fact that class differences toward perceptions of work appear in the first place is very possibly rooted in how people from the different classes look upon the qualities of their own work. As we have seen, this basic assumption is strongly corroborated by the analyses discussed in this chapter, in which class differences as regards commitment to company or organization can largely be attributed to class differences toward what people consider they derive from their jobs. Some of the class differences that have emerged regarding attitudes toward the intrinsic value of work can be explained by people's different appraisals of their own jobs-we can talk here about class-structured extrapolations from their own work situations. Those who derive only limited material or immaterial rewards from their own work find it harder to imagine that work in general has any value over and above that of simply earning a living. However, considerable class differences remain here, even after taking into account what people perceive they derive from their own jobs. We can speculate that such class differences originate from class-specific norms and values concerning the value and role of work, imparted to the incumbent through socialization in their family and social networks.

CHAPTER

FOUR

The Market: Whosoever Hath Shall to Him Be Given?

Any attempt to understand class differences in attitudes must include the relationship of class to the market. Market transactions seldom take place between players whose resources and power are matched. Instead, the labor market is characterized by power imbalances between those who buy and those who selllabor or between those whose traits or skills are in great demand and those whose qualities are in excess supply. Just as consumer markets behave differently for those with plentiful resources and those with scant resources, so capital markets differ for those with knowledge and capital and for those bereft of such assets. Consequently, our working assumption could be that people who by virtue of their greater assets are the market winners will look upon the market's transactions as more legitimate and be less inclined to redress market distributions than those who wield less power on the market. This means that we can expect groups with higher class positions to be more positive toward the market than manual workers. This is, however, a sweeping assumption that ignores several complicating factors. First, attitudes are not created out of self-interest alone, as was maintained in the first chapter. They are the product of a more complex and composite "moral economy" that centers upon notions of reasonableness and justice in different social settings. Attitudes to the market and the way It operates are naturally no exception to this; the disprivileged can also accept a certain measure of distribution if it is seen as a reasonable outcome of fair procedures. Second, there is the argument that the operational nature and distributive properties of the market breed much less discontent and attitudinal differences

52

The Market

53

than politics. Lane (1986) maintains that we often see market distribution as something that "just happens" without holding anyone responsible. We tend, contends Lane, to take the market's distribution as a given, unchallengeable phenomenon; the market apportions its resources in an arbitrary and automatic manner, and, if we are unhappy with our lot, there is no authority to which we can turn to complain. In politics, however, things work differently: we choose representatives whom we can hold accountable and with whom we can express dissatisfaction if they fail to stick to their promises or conduct themselves inappropriately. 1 Lane goes on to argue that we have a tendency to see market distribution as an expression of individual input and personal credit. The social forces behind it are quite simply not as visible as they are in politics. If Lane is correct, we might not expect to see any major differences between the classes in how they view the market. Market fatalism would eradicate any differences in class interests, leaving the distribution of the market to enjoy abiding legitimacy. Class differences might then arise only on matters of state redistribution (the subject of the next chapter), and perhaps not even then if the original market distribution gains sufficiently widespread acceptance. A third complication of relevance to the relationship between class and attitudes toward the market is that the institutional framing of national markets differs considerably from one country to another. Self-regulating markets have never been anything but a liberal utopia, as all real markets need a vast institutional framework to avoid being subverted by their own internal dynamics (Polanyi 1944; Granovetter 1985; Fligstein 2001). For instance, if market transactions are to operate properly, there needs to be a legal framework (laws, courts, police, etc.) through which ownership rights are regulated as well as contracts sanctioned. However, over and above the legal framework that each market needs in order to function, there are other ways markets become institutionally embedded, and it is here where different kinds of political economy clearly diverge. They include the extent to which work conditions and pay levels are legally regulated; whether wage bargaining is institutionalized through negotiations between organized parties or determined in private between the individual buyers and sellers of labor; and the nature of relations between unions and employers. Such institutional conditions are often loosely integrated in configurations or regimes (Hall and Soskice 2001), depending on the internal connections

54

The Market

between these different aspects. It is, for instance, hard to imagine a system for centralized negotiations on pay and other working conditions without a high degree of organization on the part of both the employees and the employers. Any agreements reached centrally would then impinge only lightly on the actual market transactions as they would concern only some companies and particular workers. One attempt to define these institutional dissimilarities between advanced market economies is to place them on a "coordinated" vs. "liberal" axis (Soskice 1999; Hall and Soskice 2001). In coordinated market economies we find relatively regulated labor markets, financial systems that allow farsighted planning, and advanced cooperation and negotiation systems between businesses as well as between unions and employers (Soskice 1999: 106 -7). Liberal market economies on the other hand are typified by a lower degree of labor market regulation, financial systems preoccupied with immediate gains and stock market prices, competition-based relations between companies, and weak systems for collective bargaining (Soskice 1999: 110-11). The four countries of concern here can be arranged along this axis, with Sweden and Germany being textbook examples of coordinated market economies and Britain and, much more so, the United States being instances of liberal ones. Such a dichotomy is particularly salient when we consider the wage bargaining systems that are in place in these countries (Rueda and Pontusson 2000; Pontusson et al. 2002). In Germany and Sweden we find welldeveloped, centralized employer-employee negotiation systems for pay and other work conditions; in the United States and Britain, such systems are almost totally absent. There is also a marked difference in the status of the unions between the liberal and coordinated market economies. In Sweden and Germany (albeit less so in the latter) the unions participate much more in the decisionmaking process at both a corporate and societallevel than in Britain and the United States (see the relevant chapters in Hyman and Ferner 1998). In the case of Sweden, this is combined with a high degree of labor organization; the United States, on the other hand, has the lowest union membership of all wealthy Western economies (Western 1997; Kjellberg 1998). The presence of wage bargaining institutions and the organization and strength of the workforce are critical factors in determining current pay discrepancies and their development over time (Rueda and Pontusson 2000;

The Market

55

Pontusson et al. 2002). A comparison of the pay levels of full-time employees illustrates this. In the mid-1990s, people in the 90th percentile of the income distribution curve in each country earned approximately 4.6 times (United States), 3.4 times (Britain), 2.8 times (Germany), and only 2.2 times (Sweden) as much those in the 10th percentile. Of all the developed countries, the United States has the greatest pay spread, while Sweden and the other Nordic countries have the smallest. Wage and salary differentials have widened in all four countries since the early to mid-1980s, but much more so in the United States and Britain than in Sweden and Germany (Rueda and Pontusson 2000: 354-56). If we broaden our scope to cover differences in disposable household income, we also have the effects of household composition, labor force participation, self-employment, capital income, and taxation and transfer systems to take into account. In the mid-1990s, the disposable household income of the 90th percentile in each country was 5.6 times (United States), 4.6 times (Britain), 3.2 times (Germany), and 2.6 times (Sweden) higher than the 10th percentile. In this respect, the United States and Sweden are polar opposites in the industrialized world (we have to look to Latin America to find greater income gaps than in the United States; Smeeding 2002: 195). By the same measure, inequalities have also increased in all four countries since the early to mid-1980s, and in this respect too they have grown more sharply in Britain and the United States than in Sweden and Germany (Smeeding 2002: 195). One question we can therefore ask is how attitudes to income distribution are affected by these institutional and distributory differences between countries. Or more specifically, how will class differences in attitudes to income distribution manifest themselves in each of the four countries? Two possible and diametrically opposed hypotheses can be applied here. Since income distribution is very much the object of discussion and negotiation in the coordinated market economies, it is here that we will find the most salient class differences. We can hardly expect to find Lane's "market fatalism" in countries where income distribution is as politicized as it is in the coordinated market economies. On the other hand, it might possibly be stronger in countries such as Britain and the United States, where wage levels are the result of (ostensibly) impersonally operative market forces. The opposing hypothesis is that since actual income gaps are much wider and have grown much more quickly in Britain and the United States

56

The Market

than in Sweden and Germany, they are much more likely to trigger class antagonisms in these liberal market economies. Put rather crassly, the workers have more to gain and the upper nonmanuals more to lose from redistribution in countries where inequality is greater. If this is one of the instruments behind the formation of attitudes toward income distribution, we would also expect to find greater class differences in the United States and Britain than in Sweden and Germany. Another issue of interest for this chapter is how class differences in attitudes toward the market change over time. As we have noted, pay and income differentials have gradually increased in all four countries, particularly so in the United States and Britain; does this also mean that class polarizations in attitudes toward income differentials have become more entrenched as time has passed, and mostly so in these two countries? Or do we find instead that the countries gradually start to resemble one another? In closing we can also question the general assumption that a lower position on the class hierarchy engenders a more skeptical view of market distribution by noting that attitudes in this area are multidimensional. Believing something about pay differentials between different occupations is not the same as having ideas about income distribution or the causes of inequality. Class patternings need not be identical in all areas, not even within one and the same attitudinal domain. So which aspects of attitudes toward the market are we to compare? The first issue we will be examining is how wide a pay or income disparity is considered reasonable in all four countries. To do this, we will be constructing indexes of legitimate income differentials on the basis of responses to questions concerning what specific occupations should earn, which we will then compare by country and class and over time. The second area of inquiry concerns attitudes toward income distribution. To what extent do people consider income differentials excessive and redistribution necessary? Are there any national and class differences in this regard, and do these opinions change over time? The third issue concerns the perceived causes of inequality. Is it primarily the outcome of distributive conflicts or of the need to create incentives for personal development and hard work? And to what extent are these explanations mutually exclusive: can whoever sees inequality as the result of distributional conflict also consider it motivating?

The Market

57

Fourth, we will be examining one or two points relating to people's views on the ideal compass of the market. What do the different classes in each country believe that money should buy? Is it fair that people on high incomes use their greater wealth to purchase better health care or better schooling for their children? There are of course other opinions about the market that could be examined, such as those about its efficiency and the level and targets of consumption, but comparable data on these areas are lacking. Instead, the concerns of this chapter are confined to aspects that focus on the distribution of income and other resources, which is certainly one, but by no means the only, important feature of market operation. The data analyzed in this chapter are taken from the 1992 and 1999 ISSP surveys on Social Inequality (see Appendix A for samples and response frequencies). 2 These data allow the first three aspects of the preceding four to be examined as a function of time, while it is only the 1999 survey that provides data for the fourth.

WHO SHOULD EARN WHAT: LEGITIMATE INCOME DIFFERENTIALS

What do the different classes consider legitimate income differentials between groups? One way to examine this, as was applied in the ISSP surveys, is simply to ask how much people with a particular occupation should earn (after asking how much they think they actually earn). An almost infinite number of indexes can be constructed from this information, each highlighting a particular aspect of what is considered a reasonable income span. Table 4.1 shows the occupations that were included in the 1992 and 1999 surveys and that form the basis of the composite indexes that I have constructed and then analyzed. 3 In Table 4.1, all occupations have been placed in relation to what people in each of the four countries consider that an unskilled factory worker should earn. At the bottom of the table are four composite indexes. The first of these, the "pay-spread" index, gives the average relation between the six highest paying occupations and the three lowest in the table. The "top-pay" index gives the relation between chairperson and director on the one hand and both worker categories on the other. The "high-pay" index shows how

TABLE 4.1 Legitimate income differences in Sweden, Britain, Germany, and the U.S., 1992 and 1999 SWEDEN

u.s.

GERMANY•

BRITAIN

What do you think people in these jobs ought to be paid?

1992

1999

1992

1999

An unskilled worker in a factory? A shop assistant? A skilled worker in a factory? A doctor in general practice? A lawyer? A cabinet minister in the national government? A judge in [country's highest court]? A chairman of a large national company? The owner-manager of a large factory? Pay-spread index Top-pay index High-pay index Low-pay index (nmin)

100 105h 121 195 240h 226

100 105 120 199 236 244

100 98 155 366 401 466

100 116 152 370 386 470

100 112 146 384 382 446

115 147 386 413 455

100 105 186 614 566 500

100 114 167 576 520 448

276h 239

273 326

584 843

604 886

418 711

479 875

627 1,114

580 1,052

301h

328 245 291 296 123 (947)

585 454 554 663

686 477 613 724 162 (603)

977 465 686 656 148 (1,858)

940 488 734 684 142 (700)

934 551 718 881 173 (1,066)

799 516 684 856 172 (924)

" Refers to the former West Germany. b Created value.

224 243 275 126 (752)

150

(885)

1992

1999

lOO

1992

1999

The Market

59

much people think the six high-paying occupations should earn in relation to what they believe that an unskilled worker earns, while the "low-pay" index gives the corresponding relationships for the three lowest paying occupations in the table. As we can see, the differences between the four countries are considerable, with Sweden occupying a particularly unique position. The average Swede believes that top business executives should earn just under 3 times as much as the manual groups; in the United States, 7 times. In Sweden people think that the high-paying occupations should award an average remuneration almost 2.5 times that of the low-paying occupations; in the United States, over 5 times. In both cases Britain and Germany end up nearer the U.S. end of the scale than the Swedish. Given that the legitimate pay spread is growing in Europe and shrinking slightly in the United States, there was a certain tendency toward convergence in the 1990s; despite this, Sweden shows no signs of approaching British or German attitudes in this respect. The mutual differences are much smaller as regards what people believe lowpaid workers should earn, and it is around the subject of high-paying occupations that the major disparities arise (cf. Kelley and Evans 1993). This is also true if we compare the attitudes of the different classes in each country. The class differences for the "low-pay" index are small across the board, even if the workers want to give this group a somewhat higher income than the service class do. For the three other pay-spread indexes, we find statistically significant class differences in each country, both in 1992 and 1999. Since the results for these three indexes are largely consistent, I will give the results of the "pay-spread" index only. Figure 4.1 shows how these class differences differ by country and year. Notwithstanding the expected tendencies (e.g., that the service class looks more favorably upon large pay differences than the workers do), what is interesting are the many exceptions and small class differences. While it may be true that the upper nonmanual group advocates large pay differentials to a much greater extent than any other group, differences in attitude among the other classes are small and not fully consistent. The unskilled workers usually place themselves at the other end of the scale, but apart from that any deviation from the attitudes of the other classes is minimal. Compared with the attitudinal dimensions we have been looking at, and have still to look at in this book, opinions regarding desired income differences seem to vary only slightly between the classes. 4

60

The Market

700

•

600

•

a

0

500

+

~

400 300

•

200

i

*

• ~

• 0

+

•~

~

0

• ~

100~----.------r-----r-----r-----r----~-----.-----.------

Sweden Sweden 92 99

Britain 92

+ Service class I 6.

Figure 4.1.

Service class II

Britain Germany Germany 99 92 99

+ Routine nonmanuals D

Skilled workers

U.S. 92

U.S. 99

* Unskilled workers 0

Self-employed

Pay Spread Index

Does this mean that differences in class attitudes toward the (re)distribution of income would also be negligible? Table 4.2 shows two statements from the surveys relating to income distribution in each country. In both cases there is a five-point response scale, from "agree strongly" to "disagree strongly" (with "neither agree nor disagree" being the midpoint). In order to calculate the mean, the responses have been recoded so that "disagree strongly" scores 0 and "agree strongly" scores 4 and the other response options 1 to 3 accordingly. The "class difference" is calculated by subtracting the mean value of the upper nonmanuals (the class that with only a few isolated exceptions returned the lowest value) from that for the unskilled workers (the class that with only a few isolated exceptions returned the highest value) for each of the statements. As can be seen from Table 4.2, people in all four countries believe that income differentials are excessive. Moreover, we find that the three European countries converged in this respect during the 1990s, since in 1999 there were more Swedes who thought so than at the start of the decade, and fewer Germans. In the United States, however, the number of people believing that

61

The Market TABLE 4.2 Class differences in attitudes to income redistribution in Sweden, Britain, Germany, and the U.S., 1992 and 1999 SWEDEN

Differences in income %Agree in [ COUNTRY] Scale mean are too large Class difference It is the responsibility %Agree of the government Scale mean to reduce the Class difference differences in income between people with high incomes and those with low incomes

BRITAIN

GERMANYa

u.s.

1992 1999 1992 1999

1992 1999 1992 1999

61 2.63 1.24 54 2.31 1.32

84 3.07 0.58 66 2.60 0.83

71 2.87 0.80 60 2.59 0.78

81 3.08 0.58 65 2.67 1.00

81 3.05 0.31ns 68 2.69 0.71

72 2.85 0.66 47 2.29 1.17

77 2.92 0.26 38 1.93 0.80

66 2.76 0.34 35 1.92 0.52

ns = Class differences not significant at the 0.05 level. Refers to the former West Germany.

NOTE: a

pay differentials are too large declined and in 1999 it was lower than in the three other countries, even though, as we discussed earlier, actual income differences are greater here than elsewhere. The United States also stands apart on two other points in the table: the number of people who believe that political measures should be taken to redress income distribution is much lower, and class opinions are more homogeneous than in the three European countries. Sweden, closely followed by Germany, exhibits the greatest class differences, while Britain leans toward the United States, in that class differences on this issue seem to have waned during the decade. This is also true of Sweden, while in Germany the classes were attitudinally entrenched, if not even more divergent. In order to create a composite measure of attitudes toward income differentials, both indicators have been aggregated into one index, which has then been divided by its maximum value and multiplied by 100. The index then ranges between 0 and 100, the higher values denoting a more critical view of income disparity. Figure 4.2 shows the class differences by country and the trends during the 1990s. One interesting observation here is that the two market liberal countries (the United States and Britain) display the smallest class differences in attitudes toward income (re)distribution. Moreover, in both cases they narrowed

62

The Market

85 80 75

*

70

D

*D

* +

+ 0

e

+ D

•

.6.

65

.6.

60

0

55

•

0

.6.

•

*

lil

D

+

+

g

•

50

* 11!

*

~

6

•

45

*

~

0 .6.

•

•

0

•

40

35~----,-----,-----,-----,-----,-----,-----,-----,-----

Sweden Sweden 92 99

Britain 92

+ Service class I .6.

Figure 4.2.

Service class II

Britain Germany Germany 99 92 99

+ Routine nonmanuals D

Skilled workers

U.S. 92

U.S. 99

* Unskilled workers 0

Self-employed

Income Redistribution Index

during the 1990s, something that we also notice in Sweden. It is also interesting to note that the only country in which the class differences remained stable, Germany, is also the only country in which wage bargaining institutions have not inclined toward market liberalism over the past few decades. There will be reason to return to this observation later.

WHAT CAUSES INEQUALITY?

Let us now take a step back in the process of income distribution and ask how the different classes tend to account for the causes of inequality. Here we can expect two diametrically opposed lines of reasoning. One sees inequality as the result of intergroup conflict and power imbalances; the fact that some people have more than others, the argument goes, is because certain groups have managed to appropriate more of what society has to offer than others. The second line of reasoning adopts a functionalist take on inequality, which it considers a necessary result of the higher remuneration

The Market

63

warranted by certain levels of pos1t10n and responsibility that either command particular authority or require specialist in-demand skills or competencies. Naturally our job here is not to examine which of these two models has the greatest empirical backing or to judge how much they conflict with or complement each other. It is to investigate the extent to which the different classes in each country subscribe to them. It would be tempting to assume in advance that groups that are less favored by the market's distribution are more likely to see inequality as the result of power struggles than of any functional demands. Such an expectation assumes that attitudes toward the causes of inequality are one-dimensional, in that whoever advocates the one will reject the other. This cannot, however, by any means be taken for granted. First, the two hypotheses are not mutually exclusive. Second, we cannot simply expect 100 percent consistency in the way people think; indeed, people can often hold contradictory and inconsistent views. Consequently, the extent to which explanations are one-dimensional or composite has to be an empirical issue. The 1992 and 1999 ISSP surveys contain a collection of statements on the causes of inequality, which are reproduced in Table 4.3 along with the mean values and the class differences. As can be seen, some of the statements used in the 1992 survey were not repeated in the 1999 survey. The mean value for each item has been calculated by recoding the responses on a 0-4 scale, on which the lower values, throughout, denote the more inequality friendly responses. Thus whoever agrees with statements 1 and 3 in the table scores high, while whoever agrees with the other three statements scores low. The class differential has been calculated by subtracting the value for a member of service class I from that for an unskilled manual worker. We can make a number of interesting observations from Table 4.3. First, we can see that the countries vary only slightly in their opinions and that any differences that are discernible are not at all what we might have expected. For instance, the Americans are less inclined than the Europeans to agree with the statements dealing with the necessity of inequality for raising the incentive for self-improvement and responsibility (2, 4, and 5). The Swedes are remarkably prone to agree that inequality is needed to create incentives and there are no signs that the long-standing dominance of social democracy has established any critical stance on the roots of inequality in this particular

TABLE 4.3 Class differences in explanations for inequality in Sweden, Britain, Germany, and the U.S., 1992 and 1999 SWEDEN

Inequality continues to exist because it benefits the rich and powerful No one would study for years to become a lawyer or doctor unless they expected to earn a lot more than ordinary workers Inequality continues to exist because ordinary people don't join together to get rid of it People would not want to take extra responsibility at work unless they were paid for it Workers would not bother to get skills and qualifications unless they were paid extra for having them

u.s.

GERMANYa

1992

1999

1992

1999

1992

1999

1992

1999

%Agree Scale mean Class difference %Agree Scale mean Class difference

53 2.46 0.85 72 1.20 -0.49

61 2.58 0.71 71 1.24 -0.22

64 2.68 0.66 71 1.25 -0.39

62 2.59 0.29 69 1.39 -0.41

75 2.88 0.63 88 0.78 -0.36

69 2.77 0.45 87 0.74 -0.30

58 2.49 0.35 70 1.27 -0.25

50 2.36 0.32 62 1.43 -0.32

%Agree Scale mean Class difference

36 2.05 0.82

39 2.15 0.62

41 2.09 0.69

37 2.03 0.69

47 2.20 0.79

43 2.26 0.79

49 2.25 0.36

47 2.26 0.36

%Agree Scale mean Class difference

76 1.20 -0.06ns

-

78 1.17 -0.49

-

72 1.29 -0.23

-

66 1.44 -0.37

%Agree Scale mean Class difference

70 1.32 -0.24ns

-

-

64 1.43 -0.43

-

77 1.17 -0.27

-

55 1.71 -0.39

ns = Class differences not significant at the 0.05 level. Refers to the former West Germany.

NOTE: a

BRITAIN

-

The Market

65

respect. It should be noted, however, that the items in question do not give any clear indication about the magnitude of inequality respondents feel is necessary to create incentives. It could well be that respondents in Sweden have in mind a much smaller amount of inequality than respondents in the United States. Another quite unexpected result is that class patternings are inverted on items touching on the incentive-generating effects of inequality. It is more common in all four countries for the workers to agree with these statements than the upper nonmanuals. Indeed, the class differences may not be as great here as they were for the conflict statements, but they are nonetheless statistically significant (except on two occasions-statements 4 and 5 in Sweden). In other words, the adoption of a power and conflict perspective on inequality, as reflected in statements 1 and 3, does not necessarily imply the rejection of explanations that stress the incentive aspects of inequality. Manual workers are simultaneously more inclined to see inequality as the outcome of power and conflict and yet more inclined to believe in its powers as an incentive. The finding that the conflict and incentive perspectives are actually two separate dimensions is also substantiated by correlational and dimensional analyses of the statements. The first and third form one dimension, the others a second. 5 We can then, with good theoretical and empirical backing, consider these two dimensions as separate and sensitive in their own ways to class differences. Against this backdrop, two composite indexes were constructed by adding the values for statements 1 and 3, dividing the result by the maximum value for the index, and multiplying by 100 to create a "conflict" index. Similarly, an "incentive" index was formed using the same procedures for the other statements. 6 The indexes are thus designed so that the higher values denote more "leftist" responses (i.e., those who score high on the "conflict" index and who therefore attribute inequality more to power differences, and those who score high on the "incentive" index and who thus believe less in the necessity of inequality for generating incentive). Given that the relevant statements for the "incentive" index were used only in 1992, we can only study the class and country differences for that year. Figure 4.3 shows the results for the "conflict" index. The national class differences are as we would have expected, with the highest index values among the unskilled workers (in Sweden we find largely the same index

66

The Market

75

* 65

D

*D

~

*D

0

+

+

+ 55 0

+ ;::,.

45

~

•

•

;::,.

+

.a

*

i

~

0

0

;::,.

lil

•

•

•

*

Ill

;::,.

~

•

35~----.-----.-----.-----.-----.-----.-----.-----.-----

Sweden 92

Sweden 99

Britain 92

+ Service class I D.

Figure 4.3.

Service class II

Britain Germany Germany 99 92 99

+ Routine nonmanuals D

Skilled workers

U.S. 92

U.S. 99

* Unskilled workers 0

Self-employed

Conflict Index

values among the skilled workers) and the lowest among the upper nonmanuals. The lower nonmanuals and the self-employed are placed somewhere between the workers and service class. The class differences are slightly larger in Sweden than in the other European countries and the lowest in the United States. 7 Moreover, the gap closes slightly in all countries between 1992 and 1999, something particularly noticeable in Britain. Any intercountry differences are insignificant. Figure 4.4 shows the pattern of class differences for the 1992 "incentive" index. As could already be seen in Table 4.3, it was inverted, in that we find the most "rightist" opinions among the working class, with the upper nonmanuals believing less in the necessity of inequality for creating incentive. The self-employed in Sweden and Germany score similarly to the workers, while in Britain and the United States they fall somewhere between the workers and service class I. We cannot overemphasize the importance of the results given in this section. The complexity of the class-attitude nexus is manifested in the way that

The Market

67

50

!

40 ~

•

+

~

D

•

0

~

30

t

*

1:;.

~

• 0

20

10~---------.----------,---------,----------.----------

Sweden

+ Service class I L;.

Figure 4.4.

Service class II

Britain

Germany

+ Routine nonmanuals D

Skilled workers

* 0

u.s. Unskilled workers Self-employed

Incentives Index

the demarcators between classes and their relation to politics depend on which aspect of the causes of inequality we are considering. If it is inequality as the product of some need for incentive, the workers take a rightist position; this is possibly because university educated people believe that there are incentives other than a higher income for entering higher education, something that may be less obvious to those without similar experiences. A related possibility is that the "incentive environment" in which the upper nonmanuals work differs from the environment of the workers, the latter offering fewer inherent rewards and incentive that is more purely pecuniary than the former. 8 However, if our focus of inquiry is inequality as the result of power and conflict, the workers tend toward a "leftist" position. We can note here that the statements used to highlight this aspect of inequality have a rather "radically populist" wording (see Table 4.3 ). A more "reformist" approach would have been desirable, with the focus instead on the opportunities of redressing inequality through piecemeal political reform, but unfortunately

68

The Market

no such data is available. It is not inconceivable that national differences on this point would have been greater if such questions had been studied instead.

WHAT SHOULD MONEY BUY?

Instead of taking a step back, as we did in the previous section, to examine which explanations of inequality are favored by the different classes, we can take a step forward and ask what it is thatthey believe that money should be able to buy. What do people see as the acceptable scope of inequality? How far can resource inequality affect different spheres of life without it becoming unjust? These issues are anchored in Walzer's (1983) theory of justice. Walzer dismisses what he refers to as a "simple equality" principle, which advocates equality within all spheres of life. Instead, he advances a "complex equality" principle, according to which different spheres of society can and should be characterized by different forms of equal treatment, whereby a high degree of inequality within any one sphere may be considered fair provided that it does not become unduly dominant by spreading to other spheres. An unequal distribution of income need not be unjust if it has no impact, say, on people's chances of survival. A cardinal point in Walzer's thesis is therefore the importance of having "blocked exchanges" to prevent the inequality of one sphere spilling over into others. For example, all democracies embody rules that prevent the buying and selling of votes, and all welfare states have regulations preventing the trade in human organs. My purpose here is not to discuss the practical possibilities of maintaining these boundaries between spheres, even less to debate Walzer's theory of justice and the criticism it has provoked (see, e.g., Rothstein 1998: 39-44). My intention is to compare classes in the four countries and their opinions on the extent to which income differentials should impinge on societal spheres other than the economy. More precisely, I am interested in studying whether people believe it fair for such differentials to show through in the ability to purchase better health care and education. Health and education can be considered staple resources, and so it would be interesting to see to what extent people feel that income differentials should be allowed to affect the quality of these services. The 1999 ISSP survey contained two questions about this very issue; these are reproduced in Table 4.4, which also gives the percentage of people who

The Market TABLE

69

4.4

Class differences in attitudes to what money should buy in Sweden, Britain, Germany, and the U.S., 1999 Is it just or unjust-right or wrong-that people with higher incomes can ... Buy better health care than people with lower incomes? Buy better education for their children than people with lower incomes?

o/o "Just" Scale mean Class difference %"Just"

Scale mean Class difference

Sweden

Britain

Germany"

u.s.

10 3.14 0.39 11 3.11 0.49

41 1.94 -0.24ns 44 1.88 -0.14ns

12 2.87 0.40ns 12 2.86 0.34ns

28 2.36 -0.08ns 32 2.30 -0.04ns

ns = Class differences not significant at the 0.05 level. Refers to the former West Germany.

NOTE: a

considered it "just" (either "very" or "somewhat") that high incomes can be used to gain access to better health care or education, along with the mean value and the class difference. The mean value is recoded so that the response "very just" scores 0, "somewhat just" scores 1, "neither just nor unjust" scores 2, "somewhat unjust" scores 3, and "very unjust" scores 4. As in the earlier section, the class difference expresses the difference between the value for an unskilled manual worker and that for a member of the upper nonmanuals. As we can see, the intercountry differences are considerable, while the interclass differences are not. Only about 10 percent of Swedes and Germans consider it fair that high-earners should be able to buy better health care or education. In Britain it is 40 percent. While Britain differs from Sweden and Germany by more than one scale point in mean value for both questions, the corresponding differences between unskilled workers and service class I are less than half a scale point in Sweden and Germany. Class differences are even less marked in Britain and the United States and are even inverted, in that the upper nonmanuals are less inclined than the unskilled workers to believe in the fairness of this income-welfare purchase relationship. Both questions have been combined into a "welfare purchase" index by adding them together, dividing by the maximum value for the index, and multiplying by 100. This gives an index that, like all the indexes published here, ranges between 0 and 100,9 with the values being directly proportional to the belief in the unfairness of this relationship. Figure 4.5 shows the class and national differences.

70

The Market

85 liE

lli

80 75

•

70 65

•

60 55

i 8

0

•

50 45 40

0

35~---------,----------.---------,----------,----------

Sweden • Service class I Service class II

A

Figure 4.5.

Britain

+ Routine nonmanuals 0

u.s.

Germany

Skilled workers

liE 0

Unskilled workers Self-employed

Welfare Purchase Index

Here, the marked intercountry differences and the modest interclass differences show up clearly. If we study the class patternings more closely, we find that it is only Sweden that displays the expected class patterns, with the upper nonmanuals and self-employed being the least skeptical toward the fairness of what money can buy and the workers and lower nonmanuals most so. In Germany it is only the upper nonmanuals that differ, and in the United States the class differences are nonexistent (i.e., not even statistically significant at the 0.05level). A puzzling pattern emerges in Britain, where the only distinct difference in attitude is found among the self-employed, who are the strongest advocates of the fairness of this income-welfare purchase relationship. The results presented in this section are in many ways fascinating from a class perspective. The two liberal economies (the United States and Britain) have a much greater acceptance of the ability of money to buy health care and education and display nonexistent or unexpected class differences in this respect; in the coordinated market economies of Sweden and Germany,

The Market

71

however, the patternings are completely different. This suggests that the national institutional configurations profoundly shape people's views of what money should be able to buy or, to put it another way, of the legitimate compass of the market. It also suggests that there is an inverse correlation in this regard between the accepted legitimacy of market principles and class differences.

WHAT IS THE RELATIONSHIP BETWEEN DIFFERENT ASPECTS OF ATTITUDES TOWARD INEQUALITY?

So far, our analyses have taken one thing at a time and examined class differences for each aspect or dimension separately. There is more to it, however, than that. It is just as interesting to see how views on various aspects of the market and inequality combine to form "mental packages." Are those who entertain certain ideas about the causes of inequality also inclined to believe certain things about ideal income spreads or what money should be able to buy? Do these "mental packages" vary between countries and classes? It is not difficult to imagine that the different attitudes toward inequality would perhaps be related to these other ideas; so that those who advocate small pay differences would also tend (1) to be critical of current income differences and in favor of redistribution, (2) to argue that inequality is created by power and conflict and is not a necessary incentive, and (3) to want to limit the ability of money to improve other aspects of life. Conversely, those who favor large pay differences should (1) have a more skeptical view of redistribution, (2) see inequality as a necessary incentive rather than the product of power struggle, and (3) consider it self-evident that money should affect other domains of life. However, is this what we actually observe among the different classes and countries? One simple way to illustrate how different aspects of market and inequality relate is to correlate all the composite indexes that have been constructed in this chapter and compare by class and country the nature of the patterns that thus emerge. Table 4.5 provides such a correlation matrix, in part by country and in part for unskilled workers and the upper nonmanuals separately. To keep the numbers manageable, we have reproduced only the results for Sweden and the United States; in most respects discussed in the following, the other two countries adopt a midway position along this

The Market

72

TABLE 4.5 Correlations between indexes among unskilled workers and service class I in Sweden and the U.S. Pears on's R X 100

Welfare Pay spread Redistribution Conflict Incentive purchase

Whole Sample Sweden

u.s.

Pay spread Redistribution Conflict Incentive Welfare purchase Pay spread Redistribution Conflict Incentive Welfare purchase

-35** -18** - 05 - 16**

56** - 01 35**

-09* 25**

-16** -15** -01 -13**

46** -19** 24**

-30* 13**

Unskilled workers

Service class I Sweden

u.s.

**

Pay spread Redistribution Conflict Incentive Welfare purchase Pay spread Redistribution Conflict Incentive Welfare purchase

-39•• 57** 29* * 35 **

-12

- 08 - 32** -32** 03 -25**

- 1s•

47••

-46** - 35** -16

- 07 - 22••

- 11 2s•• 23 ..

05 30*''

- os

37••

53 ** - 15* 33* *

- 13

07 - 2t ••

-30 ..

- 12 2o•• 00

- 24** 17*

= significant at 0.01 level; • = significant at 0.05 level.

axis. When correlation data for both 1992 and 1999 have been available, I have chosen those relating to the latter year. This has no impact on the results. Table 4.5 yields some interesting information. If we start by considering the upper part of the table, it would seem that attitude patterns in Sweden are more ideally related than they are in the United States. Correlations are generally higher in Sweden than in the United States, indicating a more integrated attitude spectrum. 10 At first glance, it might appear that Sweden's more integrated attitudinal pattern would not apply to the correlation between the "incentive" index and the other aspects. However, it is important to remember that the manner of its construction (i.e., the higher the value, the less the adherence to the "incentive" rationale) means that higher values represent more "leftist"

The Market

73

responses. Bearing this in mind, the moderately strong negative correlations that we see between this index and the "redistribution" and "conflict" indexes in the United States should be seen rather as somewhat inconsistent, or at least as an indication of a low degree of attitudinal integration. Comparing the indexes' correlations with each other, we find a very close match between the "conflict" index and the "redistribution" index. This means that those who consider inequality the product of power and conflict are more likely to consider income differentials excessive and in need of redistribution.H There is also a close relationship between the "redistribution" index and the "welfare purchase" index, suggesting that those who believe in income redistribution are also less inclined to accept that money should buy better health care and education. Using Walzer's terminology, we could say that proponents of a "simple equality" principle in the domain of income distribution also tend to support a "complex equality" principle as regards the scope of monetary power. In the lower part of the table, we can compare attitudinal integration among the upper nonmanuals (the lower left-hand trilateral for each country) with that among the unskilled workers (the upper right-hand trilateral). Doing this, we find that the relationship between indexes, with a few exceptions, is stronger among the upper nonmanuals than the workers. 12 Notions of inequality are somewhat more integrated in the former group than in the latter. 13 For instance, it is more common for the workers to both agree that inequality is the product of power and conflict and that it is necessary to generate incentive, and again that incomes should be more equally distributed. It is only among the Swedish upper nonmanuals that we find the expected correlation: i.e., that adherents to the power or conflict rationale of inequality also tend to dismiss its incentive claims. This tendency is, however, weak (Pearson's R = 0.05, not significant). The class differences in attitudinal integration are more salient in the United States than in Sweden. When we compare the American and Swedish upper nonmanuals, we find that these differences are not so great; however, if we compare the American and Swedish working classes, they become highly conspicuous. In other words, we find that the class differences in correlational patterns in Sweden are not as strong as they are in the United States.

74

The Market

CONCLUSION

Do class differences on the market also create class differences in attitudes toward the market and its distribution? Or is Lane (1986) correct in proclaiming the dominance of market fatalism and in his contention that market distribution generates only a small measure of dissatisfaction? If so, is this manifested by a tendency for all classes, regardless of their relative strength on the market, to have the same attitudes toward inequality and distribution? The answers to these questions must be both qualified and ambiguous. Lane's argument is substantiated by the findings that class differences are often remarkably small and that groups that can hardly be counted among the market's winners also accept an appreciable spread of income and recognize the legitimacy of market principles. This is particularly the case in the market liberal economies of Britain and the United States, where the working class's acceptance of money's ability to buy health care and education in the former and of large income gaps in the latter is astonishing. Lane is also backed up by the finding that attitudes toward income disparity as a necessary incentive are not only widespread in all four countries but also favored by the working classes more than any other. It could of course be debated whether believing in the need of income differences for propagating responsibility and improvement is tantamount to accepting the principles of market legitimacy. To some extent it is rather a "meritocratic" belief that hard work and effort need to be rewarded, a belief that is in harmony with many different conceptions of justice. It is, however, highly consistent with prevailing legitimizing principles for market distributions, and from that point of view it is interesting to see that workers to a larger extent than higher nonmanuals tend to accept that inequality is necessary to induce effort. It is, however, interesting to note that in most respects the coordinated economies of Sweden and Germany evince more distinct class differences than the liberal market economies. This suggests that the greater the politicization of market distribution, the more sharply the class differences tend to differ. One possible interpretation of this is that when principles of distribution are politically articulated, people are more likely to think of themselves as members of a collective with certain interests and standards rather than as atomized market actors. National variations of this kind never appear in Lane's rationale since it lacks an institutional and comparative focus.

The Market

75

The economy that has undergone the most radical transformation toward "more market" or, in the terms of this book, has been recommodified most rapidly-that of Britain-also seems to be the clearest example of how class differences in attitudes to market distribution narrowed during the 1990s. This is yet another indication of how more clear-cut market distributions, paradoxically enough, tend to be accompanied by smaller class differences in attitudes. The reduced class differences in attitudes toward income distribution that we found in all countries except Germany (see Figure 4.2) are also of interest in this context, as they suggest that the only country where wage bargaining institutions have remained unchanged over the past decade is also the one with the most stable class differences. This is probably because such institutions, like those of a more pronounced welfare state nature, act as focal points for distributive conflicts, with clearly drawn battle lines and conspicuously normative choices. In more out-and-out market situations, class conflicts tend to become atomized and Lane's "market fatalism" comes into play. The significance of bringing institutions and political articulation into consideration also becomes obvious when endeavoring to understand the results in the last part of this chapter. Here we showed that attitudinal integration was more apparent in Sweden than in the United States and more distinct among the upper nonmanuals than among the workers. Differences between classes regarding the integration of attitudinal patternings were thus larger in the United States, where the working class displays a high degree of attitudinal fragmentation. One way of interpreting such differences is to see them as the outcome of political organization and articulation. Granberg and Holmberg (19 8 8) find, as I do, that Swedish attitudes are more integrated than American; they also find that Swedes are more inclined to vote in accordance with these attitudes than are Americans. The principal explanation for this, they argue, is to be found in the differences in party systems between the countries. They also note that education seems to have a more profound effect on the degree of convergence between ideology and voting behavior in the United States than in Sweden, a finding that concurs with the class differences reported here. However, the institutional framework and political articulation extend far beyond the compass of the party system, and it is this broader perspective that we have applied here. In a country such as the United States, with its feeble labor market institutions, weak union organization, lack of left-wing

76

The Market

party, and virtually market-controlled media exposure, it is perhaps hardly surprising that the attitudinal integration among the working class is anything but tight. In a way, it is more interesting that class patternings emerge at all in spite of this lack of political articulation. In all these respects, the Swedish situation is different and, in terms of the ideological integration of the working class, is much more favorable. In many respects, it is much easier for the upper nonmanuals to achieve ideological integration. Their higher education, positions of power and responsibility, and generally more resource-rich networks make it easier for them to interlink specific attitudes and to register any contradictions that may arise. The upper nonmanual who entertains fundamentally right-wing political beliefs also has such values confirmed and embedded in a political context through the many editorials and commentaries published in the privately owned mass media. This is of course particularly marked in the United States, where alternative publicity channels are few and far between, but is also a salient feature of the otherwise socially democratic Sweden. We will be returning to this in the closing chapter; at this present juncture we will simply conclude by reminding the reader that a key argument in this chapter has been that the integration of opinions into coherent ideologies, or even programs of action, is just as important as the standpoints adopted within specific attitudinal dimensions. A thought that remains fragmented and incoherent is hardly a solid platform for successful action.

CHAPTER

FIVE

The Welfare State: My Brother's Keeper?

One of the most important arenas for contemporary class politics is the welfare state. The reason why the welfare state has come to have such a pivotal role is that it does much to redistribute resources and life chances and regulates risks stemming from market dependency. In this way, the welfare state intervenes variously in the processes that connect class position with social outcomes, and by providing social insurance (which safeguards against market dependency) and financing and delivering public welfare services, it shapes the relationship between class, resources, and risks. We should therefore expect those classes that are exposed to greater risks and that are less endowed with resources on the market to be more in favor of a comprehensive, collectively financed, and publicly organized welfare state. That should lead us, for instance, to expect workers to display stronger support for welfare policies than upper nonmanuals and the self-employed. However, this argument needs elaboration in several ways. First, analyses of attitudes to welfare policies show that they are multidimensional (Taylor-Gooby 1982; Svallfors 1996b: 64-70; 1999: 27-30). The financing and organization of welfare policies constitute a complex institutional system, and we should not expect class patterns in attitudes to be necessarily the same across different dimensions. Second, the characteristics of welfare policies vary between countries in ways that may affect the relation between class and attitudes. A particularly influential work in this respect is Esping-Andersen's (1990) regime typology. Without going into details as to the contents of this "modern classic" it can be briefly said that Esping-Andersen argues that the ways welfare policies 77

78

The Welfare State

have become institutionalized will go a long way toward determining which conflict dimensions arising around the welfare state will be the most prominent. In the social democratic welfare regimes, of which Sweden is the prime example, Esping-Andersen predicts that the class dimension will become less important, by comparison with a gender- and sector-based conflict: public sector women vs. private sector men, to put it crudely. In the "conservative" welfare model, of which Germany is the closest real-world case, we will find an "insider-outsider" split between those who depend on the welfare state for survival, such as pensioners and the unemployed, and the working population. In the "liberal" welfare states, exemplified by the United States, the class dimension will become dominant because the market is the prime shaper of the conflicts. However, the results of comparative attitude research suggest that these conflict scenarios have so far failed to materialize. The Scandinavian countries, in particular Sweden, seem still to be dominated by the class dimension; and there are no signs that the "insider-outsider" split has become particularly prominent in the German welfare state (Svallfors 1997, 2003a). Third, it could be asked how the relationship between class and attitudes toward welfare policies changes over time. Some argue that the welfare state will dig its own grave, so to speak, by virtue of its very success in redistributing resources and alleviating poverty. Offe (1987) maintains that when large groups in society no longer suffer from acute problems of sustenance, pure self-interest eventually will be insufficient to support astrong base for welfare policies. What is needed instead is the maintenance of normative orientations, which support public intervention in order to make distributive processes more egalitarian. Offe argues that the class-homogeneous environments that were conducive to such orientations have been largely dissolved by urbanization and increased social mobility, processes that are, to a certain degree, the effect of welfare policies themselves. The result is ideological and normative fragmentation, which weakens the link between class position and normative orientation and by extension threatens the survival of the welfare state. On this analysis, democracy no longer underpins the welfare state; it undermines it. However, if the perspective presented by Offe suggests a dissolution of the relation between class and attitudes to welfare policy, other indications suggest that this relationship should be strengthened. As mentioned in Chapter 1, others claim that the class relations have undergone a process of

The Welfare State

79

recommodification in recent decades (Breen 1997). Should this prove to be the case, we should expect not a dissolution of the link between class and welfare state attitudes but a gradual strengthening of these relationships. Those who lose out from recommodification (i.e., groups with a weak market position) could be expected to increase their support for comprehensive, collectively financed, and publicly organized welfare policies, while we could expect the opposite among groups who gain from it. The exact form of the relation between class and attitudes to welfare policies is therefore an issue that should be addressed using both comparative and time series data. Unfortunately, available data do not permit relevant and reliable comparisons over time. In the data we will be comparing, sourced from ISSP, I am unable to construct in any comparative manner the class variable that would make it possible to compare attitudes toward welfare policies among classes over time. Instead, we must make do with the last available ISSP module on the Role of Government from 1996. In the second part of the chapter we therefore turn to Swedish data, which allow us to track changes over time from the early or mid-1980s to the present. These data have no equivalents in the other countries at hand and therefore no comparisons can be drawn. In the third part of the chapter, we will embark on a hunt for the mechanisms that might link class with attitudes toward welfare policy.

WELFARE STATES IN COMPARATIVE PERSPECTIVE

The four countries under comparison in this book exhibit large differences in the design of their welfare regimes. As pointed out previously, the United States, Germany, and Sweden have each been typecast for representing its own particular "world of welfare" (Esping-Andersen 1990; see also Castles and Mitchell1992; Korpi and Palme 1998). In the United States, state-subsidized corporate welfare plays the main part in allocating resources and supplying insurance protection for the employed (Hacker 2002). The publicly provided welfare policy is meager and, with the important exception of an income-related public pension system, largely targeted to those unable to obtain insurance protection through their employment. Benefits are largely means tested and as of 1996 there is no legal right to social assistance. The country has no universal health insurance, although the combination of Medicare for the elderly and Medicaid for the destitute provides coverage for large sections of the population. Public child

80

The Welfare State

care and elderly care services are mainly targeted to the poorest citizens, but private child care is subsidized through tax credits (O'Connor et al. 1999: 114-21; Dobbin 2002; Olsen 2002: Chap. 2). In short, when only the publicly provided welfare policies are in the focus, the United States stands out as a particularly meager not to say rudimentary welfare state, but when the focus is widened to include also privately provided but heavily subsidized policies, the picture changes. Then the social spending in the United States becomes comparable to many European countries. However, this "private welfare state" tends to distribute coverage and benefits according to corporate hierarchy, producing a very "top-heavy" distribution (Hacker 2002). In Germany 1 the public social insurance system is comprehensive but benefits are distributed more or less strictly according to contributions. Here we find an emphasis on the role of the family; women are supposed to stay at home while the children are small, something that is underlined both by the shortage of public day care and by the fact that family supplements and tax deductions are used to support the income of the family provider (the man) (Stjerno 1995: Chap. 6-10; Sundstrom 2002: 30-36). The goal of Swedish welfare policy is for most of the population to be included as citizens. Social insurance is income-related as in Germany, although we also find a form of citizenship-based compensation that provides a "floor" in the system. In this case, it is the individual rather than the family at the center, as benefits and taxes are both largely individually based. Another important difference from Germany is that public child and elderly care is extensive (Olsen 2002: Chap. 2; Sundstrom 2002: 70-75). The British welfare state is a blend of different system principles. Just as in the United States, there is no national income-related social insurance in Britain, as benefits are a mix of universal, low, flat-rate benefits and occupational income-related insurance, which is private, although largely regulated by the state. In contrast to the United States, however, we find a universal child benefit, a free national health service, and a "proto social democratic" postwar history, during which successive Labour governments would embark on building up an income-related social insurance only to have it dismantled by their Conservative successors (O'Connor et al. 1999: 121-28; Rhodes 2000). Already from these selective points of comparison it is obvious that we are dealing with four very different welfare states. In the perspective of this

The Welfare State

81

book it is important to note that they intervene to very different degrees in class-related distributive processes. Where the U.S. welfare state provides a support of only last resort to those who cannot provide for themselves and distributes corporate welfare benefits in a regressive way, the British welfare state lays a floor under which no one is allowed to fall, the German welfare state provides a comprehensive insurance coverage that to a large extent preserves market distributions, while the Swedish welfare state more actively redistributes living conditions between groups. The question here is what effects such redistributory systems will have on citizens' attitudes toward welfare policies and the responsibilities of the state, particularly on class orientations toward these issues. Do class differences tend to be especially large in polities with small welfare states that leave most of the distributive processes to the market? Or, on the contrary, is it the case that more comprehensive welfare states, in which distributive processes have become more extensively politicized, tend to foster greater class differences? Or do class differences look more or less the same regardless of welfare model? In the search for items to use as indicators of support for government intervention, the ISSP survey on the Role of Government offers a range of possibilities. One battery of questions concerns "what the government might do for the economy," in which respondents were invited to respond to a list of government actions by stating how much they were for or against each one. However, the questions seem far too dependent on the prevailing circumstances in each of the countries (for example, "cuts in government spending" or "less government regulation of business" have quite different connotations in Sweden and the United States), and, besides, several of the measures listed have little to do with welfare policies. Instead it was decided to use a question battery that asked respondents how much they considered that it "should or should not be the government's responsibility to" intervene in certain ways. It was felt that such a battery would give a more accurate estimation of levels of support for government redistribution in different countries; moreover, it provided an index with attractive technical properties in all four countries (see note 4 ). Another question battery listed different areas of government spending and invited respondents to state whether they thought more or less tax money should be spent on them. By extracting the answers for the three welfare policy areas (health care, pensions, unemployment benefits) and forming them into an index, we obtain an indication of how willing the different classes are

82

The Welfare State

TABLE 5.1 Class differences in attitudes to state intervention in Sweden, Britain, Germany, and the U.S.

Should it be the government's responsibility to ... Provide a job for everyone who wants one? Keep prices under control?

Provide health care for the sick? Provide a decent standard of living for the old? Provide a decent standard of living for the unemployed? Reduce income differences between the rich and the poor? Give financial help to university students from low-income families? Provide decent housing for those who can't afford it?

% Definitely Scale mean Class difference % Definitely Scale mean Class difference % Definitely Scale mean Class difference % Definitely Scale mean Class difference % Definitely Scale mean Class difference % Definitely Scale mean Class difference % Definitely Scale mean Class difference % Definitely Scale mean Class difference

Sweden

Britain

Germany a

u.s.

35 1.89 0.98 45 2.26 0.71 71 2.66 0.23 69 2.66 0.36 39 2.26 0.35 43 2.03 1.17 36 2.11 0.40 27 2.05 0.50

29 1.87 0.89 44 2.26 0.70 82 2.80 0.12 73 2.71 0.26 29 2.01 0.41 36 1.93 0.90 38 2.26 0.14 37 2.23 0.47

28 1.97 0.69 23 1.87 0.57 51 2.46 0.24 47 2.43 0.30 17 1.93 0.29 25 1.76 0.69 27 2.11 0.07 20 1.93 0.33

14 1.25 0.56 25 1.82 0.74 39 2.19 0.37 38 2.22 0.39 13

1.41 0.48 17 1.40 0.64 35 2.16 0.12 20 1.78 0.34

' Refers to the former West Germany.

to increase welfare spending (see Appendix C for an exact description of the index design). This measure should not be compared between the countries, since views about "more or less" are naturally affected by actual domestic spending levels. Having said this, it does give an indication of class differences in attitudes to welfare spending within the four countries. Table 5.1 shows the percentage of respondents in the ISSP 1996 survey who thought it "definitely should be" the government's responsibility to take a certain measure, along with the scale means and class difference for each one. 2 The scale mean is computed by assigning a response of "definitely should be" a score of 3, "probably should be" a score of 2, "probably should not be" a score of 1, and "definitely should not be" a score of 0. Consequently, the higher the score, the stronger the support for state intervention. The class difference is calculated by subtracting the value for service class I.

The Welfare State

83

(who without exception score lowest) from that for the unskilled workers (who without exception score highest). Table 5.1 shows that class differences vary substantially between the different indicators. They are particularly large when it comes to views about government responsibility for full employment and for reducing income differences. These are issues where class interests regarding correction of market distributions could be assumed to be highly divergent, so it is not surprising that class differences are large here. Class differences are particularly small with regard to health care, the living standards of the elderly, and economic support for university students. These public measures have a much less clear-cut class profile, so it is hardly remarkable that classes differ little on these points (cf. Korpi and Palme 2003: 431). There is a tendency for large class differences and large country differences to coincide. This indicates that policies that generate clear class differences also are the ones where citizens from different welfare regimes differ most substantially from one another. Since these welfare regimes are very much the outcome of past political power struggles, reflecting differences in power resources between classes and other groups (Huber and Stephens 2001a; Swank 2002; Korpi and Palme 2003 ), we should not be surprised to find that large class and country differences tend to coincide. In order to create a composite index of views about state intervention, the indicators were collapsed into one measure. 3 This index was then divided by its maximum value and multiplied by 100 so as to have a range of 0 to 100. Class differences in the four countries are displayed in Figure 5.1. As regards class, there are both substantial similarities and interesting differences between the countries. From the United States to Europe the class rankings are almost identical, the manual groups being the most favorably disposed toward state intervention and the upper nonmanuals being the most skeptical. Differences between workers and lower-level nonmanuals are hardly noticeable, while service class 11 and the self-employed position themselves somewhere between the lower and upper nonmanuals. At the same time we find that the magnitude of class differences varies quite substantially between countries. The largest class differences, in this sense, are to be found in Sweden; they are somewhat smaller in Britain and the United States and the smallest in Germany. The low index values in the Swedish service class I are particularly noteworthy. They have, for instance, substantially lower values than their counterparts in Britain, while other

84

The Welfare State

90 85

*

80

Ill

ili

75

Ill

/::,.

B

70

0

+

•

65

•

60

fi

*

•

Ill

R

55

•

50 45

40~---------,----------.---------.----------,----------

Sweden

+ Service class I l::.

Figure 5.1.

Service class II

Britain

+ Routine nonmanuals D

u.s.

Germany

Skilled workers

liE 0

Unskilled workers Self-employed

Government Intervention Index

class categories differ little between these two countries. The distance between service class I and unskilled workers in Sweden is more than double that between the Swedish and the U.S. upper nonmanuals. All in all there is little to suggest that this group of employees is part of any social democratic consensus regarding welfare policies. Table 5.2 shows the results from a regression analysis, including indicators of Esping-Andersen's two other conflict dimensions alongside the six class categories and the four countries. In addition to comparing men to women, we also compare three groups representing different degrees of labor market status: the unemployed, pensioners, and the rest. 4 The first four columns of the table show the results of separate regressions performed on each of the four countries, which are integrated in the last column into a composite analysis. 5 The table gives additional support to an important finding from Figure 5 .1: class differences are largest inS weden and smallest in Germany. This finding does not fit well with Esping-Andersen's thesis about an emerging

85

The Welfare State TABLE5.2

Government intervention index. Regression analysis (ordinary least squares). Unstandardized regression coefficients

Constant Service class I Service class II Routine nonmanuals Skilled workers Self-employed (unskilled workers reference group) Sweden Britain Germany" (U.S. reference group) Woman (man = 0) Unemployed Retired (Others reference group) R2(%)

Sweden

Britain

Germanya

u.s.

All four

79.9'" ''* -19.3*"* -9.7'"''* -4.7''''" -2.1 -10.2"''''

79.5'"''' -15.3*** -9.9'''"'' -8.6*'''' -6.9'""" -11.1 *''''

70.7*'''' -12.2**'' -6.2*''''" -5.1 ,,,,

63.7*'''' -15.9'c'f* -14.7*''* -8.7'""'' -3.7 -12.2*''*

63.4* ,, * -15.5*''* -10.1''''"'' -6.8*'"'' -3.0'"" -10.1''*''

0.3 -7.1 *"

14.1 **" 15.1"*" 8.8*'''' 2.4" 6.9'''' 9.1''*'' 18.7

4.2*'' 12.7* ''"* 4.0*''"

4.2''*'' 4.8 0.6

7.0''*'' 10.2'' 3.3

15.2

7.7

10.6

4.4'''"' 8.0'''''' 4.4'''''' 22.3

Refers to the former West Germany. *** = p < 0.001; *'' = p < 0.01; ,, = p < 0.05.

a

gender- and sector-based conflict in Sweden. Nor does his thesis receive any support from the results in other respects. There is a particularly strong difference between pensioners and other groups in Sweden, while gender differences are smallest in Sweden but largest in the United States. The unemployed are especially supportive of the welfare state in Britain and the United States, while the split between "insiders" and "outsiders" is smallest in Germany. The last column of Table 5.2 shows that national differences are roughly of the same magnitude as class differences. Between the United States on one hand and Sweden and Britain on the other we find a difference of about 1415 units on the index, which is about the same difference as the one found between unskilled workers and service class I. Differences between men and women and between groups with different labor market status are considerably smaller. An illustrative example from the last column of the table shows the substantial differences that appear once we take into account the impact of different factors. An American man in service class I, who is neither unemployed nor retired, has, according to these estimates, an index value of 47.9

86

The Welfare State

TABLE 5.3 Welfare spending index. Regression analysis (ordinary least squares). Unstandardized regression coefficients

Constant Service class I Service class II Routine nonmanuals Skilled workers Self-employed (unskilled workers reference group) Woman (Man = 0) Unemployed Retired (Others reference group)

R2 (%) a

Sweden

Britain

Germany"

u.s.

68.6*''* -14.9*''* -8.3'"'"" -2.0 0.1 -10.0'" '""

73.0'' ,, ,,

63.5'"''''" -12.0*** -8.5*''* -5.7*'"" -0.5 -8.9*''*

64.0''*'' -9.5'""* -9.5''*'' -5.6''*'' -0.5 -5.9''*

5.4*** 5.4'' ,. 3.5''* 18.5

-9.1 '"''' -4.7'"" -5.4)~'~-),~

-3.1 -7.3'""'' 1.6 8.9'""'' 5.3'"'*

3.3*'' 8.5'''' 3.2'"

9.6

9.8

4.4'"' * 7.4'' -0.3 5.8

Refers to the former West Germany. = p < 0.001; *'' = p < 0.01; * = p < 0.05.

*"*

(63.4 - 15.5). The corresponding figure for an unemployed British woman in the unskilled working class is 90.9 (63.4 + 4.4 + 15.1 + 8.0). 6 Are these class differences in government's responsibility reflected in views about spending on the welfare state? And do the relative class distances between countries stay the same when spending is in the focus? Table 5.3 shows a corresponding regression analysis for the welfare spending index. Just as for the previous index we find that class differences are particularly large in Sweden. However, we also find that class differences are larger in Germany than in Britain and the United States; moreover, these results fail to substantiate Esping-Andersen's conflict scenarios. Differences between "insiders" and "outsiders" are, for example, not the largest in Germany, and even though gender differences are somewhat larger in Sweden than elsewhere, Sweden is clearly dominated by class differences.

SWEDEN 19 8 I-2002

The questions that have been analyzed so far are of a fairly general character. Furthermore, they have not been posed since 1996 and allow for no comparison over time. 7 In order to assay developments in class differences in views about different aspects of welfare policies since the 1980s, we now

87

The Welfare State

OJ

::l

75

50

70

45

65

40

60

35

55

""@

> >0. 7), and the Cronbach's alpha reliability coefficient is therefore greater than 0.9 in all four countries. This suggests that people do not differentiate between health and education in their responses but tend to give similar answers to both. 10. The differences between Sweden and the United States are statistically significant (0.05 level) for six out of nine of the coefficients. 11. In a previous paper (Svallfors 1993) I even treated all these subissues as indicators of redistribution, which would now appear to have been a dubious procedure to adopt. 12. Tests of statistical significance of the differences in coefficients between workers and higher nonmanuals produce mixed results, something that also is the case for the Sweden-United States comparison within particular classes. Only a handful of these are statistically significant, since the small sample sizes tend to result in large standard errors, rendering even large nominal differences statistically

Notes

197

nonsignificant. The results in these respects must therefore be seen as indicative rather than conclusive. 13. It might seem perplexing that the correlation among the upper nonmanuals and the workers is lower or higher on certain matters than the correlation for the entire sample. All this means, however, is that it is in other classes that we find higher or lower correlations.

Chapter Five 1. East Germany had a completely different social policy than West Germany, but since reunification the institutions have more or less been implanted in the former East Germany. 2. The battery included two items on "providing industry with the help it needs to grow" and "imposing strict laws to make industry do less damage to the environment." Since these items do not indicate welfare state commitment and display low correlations with the other items in the table, they have been eliminated from Table 5.1 and the subsequent composite index. 3. In all countries, factor analysis (principal components, varimax rotation) yields a one-factor solution, indicating unidimensionality of the items. Index reliability measures (Cronbach's alpha) vary between 0.79 (Germany) and 0.85 (Sweden), showing that the index is reliable in all four countries. 4. The question on private and public sector employment has not been asked in the United States and is therefore not included. Differences between private and public sector employees are rather small in the other three countries, particularly in Germany. There is therefore no reason to believe that results would have changed significantly if private vs. public sector employment had been included in the analysis. 5. The German dataset contains almost twice as many individuals as the other datasets. In order to correct for this, the German data have been weighted down to the average sample size of the other three countries. 6. However, these figures should not be taken too literally, since they assume that no interaction effects are present, which we can clearly see is not the case just by looking at the first four columns of the table. By way of example, we can estimate that if we take the country-separate regressions as the starting point, the American service class man is unchanged at 47.8 (63.7 - 15.9), while the index value for the unemployed British unskilled worker woman increases to 96.4 (79.5 + 4.2 + 12.7). The discrepancy vis-a-vis the joint model is mainly explained by the difference between the unemployed and the rest being particularly large in Britain. 7. In Britain, Germany, and the United States, most of the questions used here were also surveyed in 1985 and 1990, but the occupational classifications do not allow the construction of comparable class categories over time. At an aggregated level the index is very stable between surveys (Svallfors 2003a: Table 2; 2003b). The 1996 survey was the first on this particular topic to be conducted in Sweden.

198

Notes

8. The results are not shown here but are available from the author on request. 9. It should be noted that the questions relating to who is considered best suited to manage different welfare services are worded slightly differently in ULF than in the Swedish Welfare State surveys. The respondents in ULF sometimes gave more than one option as the "most suited." In the construction of the index, therefore, those who answered "state and local government" as the most suited have been assigned a value of 1 and all others 0 (see Appendix C). Assigning a value of 1 to all those who replied "state and local government" as one possible option for the best suited would have given the index too narrow a spread. 10. The first adult in the household is given a neutral weighting, the second adult a weighting of 0.65. Children are weighted at 0.48 (0-3 years old), 0.57 (4-10 years old), and 0.65 (11-17 years old). Other people in the household (over 18 years old) are assigned a weighting of 0.83. 11. Both these dimensions can be differentiated using factor analysis (principal components). A third dimension is how much control a person has over various organizational aspects of his or her work. Since this "control" index has a nonexistent correlation with the attitudes under study here and the questions were asked only of the employees, it has not been included in the analyses. 12. Example for the "finance" index: the upper nonmanual with the upper nonmanual background who earns 200,000 kronor more per year than the worker raised in a working class home has a -7.3 - 12.7- (200 X 0.05) = 30-unit lower value on this index.

Chapter Six 1. Baxter and Kane's class model is based on Erik Olin Wright's class schema, and they compare attitudes in Australia, Canada, Norway, Sweden, and the United States. 2. For the purposes of this chapter we will be referring to families as if they were bas~d on heterosexual couplings. The extent to which the theories and results discussed can be applied to ho111osexual couplings is hard to assess. 3. The figures provided by Daly are to some extent misleading for the United States, since this country heavily subsidizes private child care through tax credits and to a large extent provides child care through private employers who receive public subsidies for this provision (Dobbin 2002: 64-70). 4. In addition to the statements included in the table, the two question batteries from 2002 contained the statements "Having a job is the best way for a woman to be an independent person" and "Both the man and woman should contribute to the household income." These statements are only slightly related to the others on the view of women in paid work and have therefore been excluded. 5. The expressions "equality oriented" and "gender conservative" are used here to distinguish those who advocate equal work duties and conditions for men

Notes

199

and women from those who prefer clear-cut roles of male breadwinner and female carer. The latter does not necessarily imply an unequal distribution of work. 6. In Sweden and Germany, a factor analysis (principal components, varimax rotation) of the statements in the table provides a two-factor solution. In Britain and the United States, it gives a three-factor solution, the result of the way that the fifth statement ("Being a housewife ... ")is only slightly related to the other statements. The exclusion of this statement from the index, however, has little effect on the results given in this chapter and renders the index less reliable for Sweden and Germany. 7. Another indication that this could be the case is that not even gender differences are especially salient on this index (approximately a 4- to 6-unit difference), even if they are statistically significant. 8. These data analyses are not reported here but are available from the author. 9. We would have expected the class ordering to have had the opposite effect for women, i.e., that we would find more egalitarian opinions the higher relative class position she holds in the family. However, analyses show that this is not the case, which could be because the number of women with a higher class position than their husbands or partners is relatively small. 10. Information on the profession of the spouse or partner is lacking for much of the British sample, and so only the first two models are included for that country. 11. The list also included "makes small repairs around the house." However, this indicator is unrelated to the others in the list and has therefore been eliminated. Since it is a "typical male" activity, gender differences in performing household tasks would therefore be rendered smaller if it were included. 12. The difference between Swedish and German women appears only once the actual distribution of work is controlled for; there are no significant differences at a bivariate level. 13. Interestingly, albeit in a different theoretical and empirical context, Bourdieu makes the same observation, namely, how role expectations on men and women tend to be less rigid and categorical in "the dominant classes" (Bourdieu 1984: 382). He fails, however, to provide us with any explanation (or any data either).

Chapter Seven 1. Yet other authors have more specifically criticized the measures used by Lipset to indicate "authoritarianism" (such as the so-called F scale; e.g., Miller and Riessman 1961; Ray 1983). This literature is not addressed here. 2. For example, many white American workers were disturbed by what they perceived as the low work ethic and lax morals of blacks, while black workers were worried about the lack of companionship and caring among whites. Immigrants,

200

Notes

on the other hand, were not considered morally suspicious. French white workers were less obsessed with skin color, but were worried about immigrants from North Africa, to whom they attributed many of the same negative traits that white Americans ascribed to black people. 3. The concept "authority" should not be understood as denoting only superior persons or others in a position of power. The concept can equally well apply to the consensually dominant values, lifestyles, and living patterns. 4. The Role of Government survey was also conducted in 1990, but at that time Sweden was not yet a member of ISSP, and the class variable cannot be constructed in an adequate way. Unlike preceding chapters, this chapter analyzes data from a post-reunification Germany, although this has no discernible effect on the results. 5. The remaining response categories are "almost always wrong" and "wrong only sometimes." 6. Separate analyses of the 1994 and 1998 data are available from the author. None of the conclusions drawn in this chapter are affected by any differences between the two years. 7. The low scalability of the indicators in this section and the following is a cause of concern, since it implies that the indicators do not measure exactly the same thing, in spite of their face validity. A couple of observations need to be made in order to defend the procedures applied here: (a) As will be shown, class patterns in attitudes are essentially the same across different indicators for each dimension. This also means that the use of summarized indexes-although it means violating the assumptions behind the construction of such indexes (see Appendix D for a discussion)-yields essentially similar results to the ones presented here (SPSS runs available from the author). (b) Changing analytical techniques from the ones applied here to techniques using latent variables (such as LISREL) would not be appropriate, since the low reliability of the manifest indexes would lead to doubts about the content of latent variables. Instead, I decided to stay with manifest indicators and display results item by item. 8. Cronbach's alpha runs from 0.30 (Sweden 1998) to 0.50 (Germany 1994). 9. Education is coded into three categories: (a) those with the socially expected minimum in each country, (b) those with secondary or tertiary education without university degree, and (c) those who have a university degree. Age is coded into six age groups: 18-25,26-35,36-45,46-55,56-65, and 66 and older. This grouping is preferred to applying age as a continuous variable, in order to take into account possible curvilinear relationships and to make computations of probabilities easier and goodness-of-fit measures more reliable. Men, low education, and 66 and older are used as reference categories in the models. 10. It is mainly age that accounts for different attitudes to infidelity, although education also matters. As regards attitudes to homosexuality there are also

Notes

201

distinct gender differences, with women being more accepting than men (SPSS runs available from the author). 11. The net class difference is achieved by the following steps: 1. Estimate the expected probabilities for conformist and nonconformist attitudes in three classes at three different education levels (see note 9 for definitions). 2. Calculate the expected percentage balance, for each class at each education level, by subtracting the expected probability for having a "conformist" attitude from the expected probability for having a "nonconformist" attitude. 3. Calculate the difference between the expected percentage balances for the service class and workers for each of the three education levels. 4. Sum the three differences achieved in step 3 and divide the sum by three to obtain the net class difference. The net education difference is achieved by replacing step 3 with "Calculate the difference between the expected percentage balances for university graduates and those with low education, for each of the three classes." 12. Pearson's R varies from 0.06 in Germany to 0.12 in Sweden and the United States, which denotes very weak correlations. Cronbach's alpha varies from 0.11 in Germany to 0.21 in Sweden and the United States. 13. Attitudes to obeying the law without exception are heavily structured by age, with education also playing a certain role. Attitudes toward miscarriages of justice differ less by age category, being mostly affected by education (except in Germany where absolutely no differences are found between any of the categories). (SPSS runs available from the author.) 14. An example of an important institutional difference in these matters is that homosexual relations are still formally illegal in a large number of U.S. states. 15. The debate about the effects of education on issues such as the ones analyzed here, and the interpretation of findings, is summarized in Weil (1985: 458-60); Bobo and Licari (1989: 286-90); and Coenders and Scheepers (2003: 313-19). 16. A parallel might here be drawn to Paul Willis's (1977) classic ethnographical study of working-class boys at school, which illustrates how the racist and sexist attitudes of "the lads" tend to hasten their exit from the education system and their voluntary assimilation into a male-dominated and underprivileged working-class milieu.

Chapter Eight 1. There are remarkably few cross-references between the U.S. debate, from C. Wright Mills (1951) to Brint (1994), and the European (e.g., Savage et al. 1992; Muller 1999). The debates in this area thus seem to have been conducted in parallel rather than along any comparative lines.

202

Notes

2. "Assistant nonmanual employees I" comprises occupations that are normally members of TCO (the Confederation of White-Collar Employees) and that require 2-year secondary education. A review of the occupations included in this category reveals that they are such whose content is typical for worker occupations and that the reason why they are positioned among the nonmanuals in the SEI system is that (by historical accident) they are organized by the white-collar TCO rather than•by the blue-collar LO (the Swedish Trade Union Confederation). 3. The eta and F values both express the relationship between (a) the variation between the categories of the independent variable and (b) the variation within its categories. As pointed out previously, eta is the coefficient that expresses how much of the total variation in the dependent variable is attributable to differences between categories in the independent variable. If all variation existed between the categories rather than within them, eta would equal 1; if all variation existed within the categories rather than between them, eta would equal 0. The F value expresses the relationship between the average variation between the categories and the average variation within the categories (i.e., taking into account the number of degrees of freedom; see also Tabachnick and Fidell1996: 39-40, 53). A classification that manages to capture the same variation with a smaller number of categories would thus have the same eta value but a higher F value. 4. Expressed in SEI codes, the ranking is 60, 57, 56, 89, 76, 79, 46, 36, 21/22, 33, 11/12. 5. Normal linear regression, or ordinary least squares (OLS) was used for the three attitude indexes. For the two separate indicators, I applied ordinal regression. 6. Significant (but rather small) differences also emerge for the question of an egalitarian society; for the other indexes and indicators, there were no statistically significant differences between the marginal group and the rest.

Chapter Nine 1. Thanks to Staffan Kumlin for providing the idea behind the figure. 2. The line of reasoning followed here is inspired by Casten von Otter's seminar at Umea University's Department of Sociology, 12 December 2003. 3. The manifesto dataset is based on a classification of the smallest significant units ("quasi-sentences") of each election manifesto on a number of themes (56 categories distributed among 7 policy areas). The proportion of the manifestos (i.e., the relative total number of quasi-sentences) dedicated to each of the themes was then calculated. The different dimensions in this analysis have been constructed by merging multiple categories as follows:

Socio-economic dimension "left-wing themes": per403 ("Market Regulation")+ per404 ("Economic Planning")+ per405 ("Corporatism")+ per412 ("Controlled Economy")+ per413 ("Nationalisation")+ per503 ("Social Justice") + per504 ("Welfare State Expansion") + per701 ("Labour Groups: Positive")

Notes

203

Socio-economic dimension "right-wing themes": per401 ("Free Enterprise")+ per402 ("Incentives") + per414 ("Economic Orthodoxy")+ per505 ("Welfare State Limitation") + per702 ("Labour Groups: Negative") + per704 ("Middle Class and Professional Groups") Socio-cultural dimension "left-wing themes": per 201 ("Freedom and Human Rights") + per602 ("National Way of Life: Negative") + per604 ("Traditional Morality: Negative") + per607 ("Multiculturalism: Positive") + per705 ("Underprivileged Minority Groups") + per706 ("Non-Economic Demographic Groups") Socio-cultural dimension "right-wing themes": per601 ("National Way of Life: Positive") + per603 ("Traditional Morality: Positive") +per 605 ("Law and Order") + per608 ("Multiculturalism: Negative"). For a detailed account of the contents of each category, see Budge et al. (2001: Appendix Ill). 4. SPSS runs are available from the author. 5. A perhaps embarrassingly personal example concerns my studies into welfare opinion in Sweden. In both the books that have resulted from this work, one of the analysis objects is the welfare policy articulation of the organized interests (i.e., interest organizations and political parties; Svallfors 1989, 1996b). When reviewing the writings of these bodies, I have occasionally encountered detailed references to my own attitude research as support for one or the other particular view.

Appendixes 1. The list also includes "the environment," "police and law enforcement," "education," "the military and defence," and "culture and the arts." 2. For example, the item on spending on state and local government administration has a slightly stronger connection with the other "spending items" on these occasions, rendering them includable in the "spending" index. This also applies to the item on financing education and its relation to the other financing items (Svallfors 1995, 1996b). 3. An introduction to linear regression analysis can be found in Tabachnick and Fidell (1996: Chap. 5). Chapter 12 of the same book contains an introduction to logistic regression. Brief, nontechnical descriptions of linear regression and logistic regression and the interpretation of different measures can also be found in Schroeder et al. (1986) and Pampel (2000). A somewhat more technical but nevertheless accessible introduction to all forms of regression method for category data can be found in Long (1997).

REFERENCES

Acker,]. (1973 ). Women and social stratification: A case of intellectual sexism; American journal of Sociology, 78,936-45. Ahrne, G., and Roman, C. (1997). Hemmet, barnen och makten: forhandlingar om arbete och pengar i familjen, Fritzes, Stockholm. Bates, R. H., de Figueiredo, R.]. P., and Weingast, B. R. (1998). The politics of interpretation: Rationality, culture, and transition; Politics and Society, 26, 221-56. Baxter,]. (1994 ). Is husband's class enough: Class location and class identity in the United States, Sweden, Norway, and Australia; American Sociological Review, 59,220-35. Baxter,]., and Kane, E. W. (199 5). Dependence and independence: A cross-national analysis of gender inequality and gender attitudes; Gender and Society, 9, 193-215. Beck, U. (1992). Risk society: Towards a new modernity, Sage, London. Becker, J. W., Davies,]. A., Ester, P., and Mohler, P. P. (Eds.) (1990). Attitudes to inequality and the role of government, Sociaal en Cultureel Planbureau, Rijswiik. Bertilsson, M., and Eyerman, R. (1979). Interest as a problematic concept in Marxist social science; Acta Sociologica, 22, 361-75. Bihagen, E. (2000). The significance of class: Studies of class inequalities, consumption and social circulation in contemporary Sweden, Umea Universitet, Umea. Blossfeld, H. P. (1992). Is the German dual system a model for a modern vocational-training system: A cross-national comparison of how different systems of vocational-training deal with the changing occupational structure; International Journal of Comparative Sociology, 33, 168-81. Bobo, L., and Licari, F. C. (1989). Education and political tolerance: Testing the effects of cognitive sophistication and target group affect; Public Opinion Quarterly, 53, 285-308. Bourdieu, P. (1984). Distinction: A social critique of the judgement of taste, Routledge & Kegan Paul, London.

205

206

References

Bradley, D., Huber, E., Moller, S., Nielsen, F., and Stephens, J. D. (2003). Distribution and redistribution in postindustrial democracies; World Politics, 55, 193-228. Breen, R. (1997). Risk, recommodification and stratification; Sociology, 31, 4 73-89. - - - . (1999). Beliefs, rational choice and Bayesian learning; Rationality and Society, 11,463-79. Breen, R., and Rottman, D. (1995). Class analysis and class theory; Sociology, 29, 453-73. Brint, S. G. (1985). The political attitudes of professionals; Annual Review of Sociology, 11, 389-414. - - - . (1994). In an age of experts: The changing role of professionals in politics and public life, Princeton University Press, Princeton, NJ. Budge, I., Klingemann, H.-D., Volkens, A., Bara,]., and Tanenbaum, E. (2001). Mapping policy preferences: Estimates for parties, electors, and governments 1945-1998, Oxford University Press, Oxford. Castles, F. G., and Mitchell, D. (1992). Identifying welfare state regimes: The links between politics, instruments and outcomes; Governance, 5, 1-26. Coenders, M., and Scheepers, P. (2003). The effect of education on nationalism and ethnic exclusionism: An international comparison; Political Psychology, 24, 313-41. Dalton, R. (1999). Political support in advanced industrial democracies; In Critical citizens: Global support for democratic government (Ed., Norris, P.). Oxford University Press, Oxford, pp. 57-77. Daly, M. (2000). A fine balance: Women's labor market participation in international comparison; In Welfare and work in the open economy. Vol II: Diverse responses to common challenges (Eds., Scharpf, F. W., and Schmidt, V. A.). Oxford University Press, Oxford, pp. 467-510. Daly, M., and Rake, K. (2003). Gender and the welfare state: Care, work and welfare in Europe and the USA, Polity Press, Cambridge. Davis, J. A., and Jowell, R. (1989). Measuring national differences. An introduction to the International Social Survey Programme (ISSP); In British social attitudes. Special international report (Eds., Jowell, R., Witherspoon, S., and Brook, L.). Gower, Aldershot, pp. 1-13. Dekker, P., and Ester, P. (1987). Working-class authoritarianism: A re-examination of the Lipset thesis; European Journal of Political Research, 15, 395-415. Devine, F. (1998). Class analysis and the stability of class relations; Sociology, 32, 23-42. Dobbin, F. (2002). Is America becoming more exceptional? How public policy corporatized social citizenship; In Restructuring the welfare state (Eds., Rothstein, B., and Steinmo, S.). Palgrave MacMillan, New York, pp. 51-77. Dobbin, F., and Boychuk, T. (1999). National employment systems and job autonomy: Why job autonomy is high in the Nordic countries and low in the United States, Canada, and Australia; Organization Studies, 20, 257-91.

References

207

Edelman, M. (1964). The symbolic uses of politics, University of Illinois Press, Chicago. - - - . (1971). Politics as symbolic action: mass arousal and quiescence, Markham, Chicago. - - - . (1977). Political language: Words that succeed and policies that fail, Academic Press, New York. - - - . (1988). Constructing the political spectacle, University of Chicago Press, Chicago. Edlund,]. (1999). Citizens and taxation: Sweden in comparative perspective, Department of Sociology, Umea University, Umea. - - - . (2003). The influence of the class situations of husbands and wives on class identity, party preference, and attitudes towards redistribution: Sweden, Germany and the United States; Acta Sociologica, 46, 195-214. Elster,]. (1983). Sour grapes: Studies in the subversion of rationality, Cambridge University Press, Cambridge. - - - . (1989). The cement of society: A study of social order, Cambridge University Press, Cambridge. Erikson, R. (1984). Social class of men, women and families; Sociology, 18,500-14. - - - . (2004). Has the individual become the unit in the class structure, or may it still be the family? Working Paper, Institute for Social Research, Stockholm University. Erikson, R., and Goldthorpe, J. H. (1992a). The constant flux: A study of class mobility in industrial societies, Clarendon Press, Oxford. - - - . (1992b). Individual or family: Results from 2 approaches to class assignment; Acta Sociologica, 35, 95-105. Esping-Andersen, G. (1990). The three worlds of welfare capitalism, Polity Press, Cambridge. Esping-Andersen, G., Assimakopoulou, Z., and van Kersbergen, K. (1993). Trends in contemporary class structuration: A six-nation comparison; In Changing classes. stratification and mobility in post-industrial societies (Ed., EspingAndersen, G.). Sage, London. Estevez-Abe, M., Iversen, T., and Soskice, D. (2001). Social protection and the formation of skills: A reinterpretation of the welfare state; In Varieties of capitalism: The institutional foundations of comparative advantage (Eds., Hall, P.A., and Soskice, D.). Oxford University Press, Oxford, pp. 145-83. Evans, G. (1992). Testing the validity of the Goldthorpe class schema; European Sociological Review, 8, 211-32. - - - . (1993a). Class, prospects and the life-cycle: Explaining the association between class position and political preferences; Acta Sociologica, 36, 263-76. - - - . (1993b). The decline of class divisions in Britain: Class and ideological preferences in the 1960s and the 1980s; British Journal of Sociology, 44, 449-71.

208

References

- - - . (1996). Putting men and women into classes: An assessment of the cross-sex validity of the Goldthorpe class schema; Sociology, 30, 209-34. - - - . (1998). On tests of validity and social class: Why Prandy and Blackburn are wrong; Sociology, 32, 189-202. - - - . (Ed.) (1999). The end of class politics? Class voting in comparative context, Oxford University Press, Oxford. Evans, G., Heath, A., and Lalljee, M. (1996). Measuring left-right and libertarian authoritarian values in the British electorate; British Journal of Sociology, 47, 93-112. Evans, G., and Mills, C. (1998a). Assessing the cross-sex validity of the Goldthorpe class schema using log-linear models with latent variables; Quality and Quantity, 32,275-96. - - - . (1998b). Identifying class structure: A latent class analysis of the criterionrelated and construct validity of the Goldthorpe class schema; European Sociological Review, 14, 87-106. - - - . (2000). In search of the wage-labour/service contract: New evidence on the validity of the Goldthorpe class schema; British Journal of Sociology, 51,641-61. Fligstein, N. (2001). The architecture of markets: An economic sociology of twenty-first century capitalist societies, Princeton University Press, Princeton, NJ. Forsman, P. (1995). Det kluvna samhallet: Perspektiv pa samhallets omvandling, Carlsson, Stockholm. Fraser, N. (2000). Adding insult to injury: Social justice and the politics of recognition, Verso, London. Gallie, D. (2003). The quality of working life: Is Scandinavia different? European Sociological Review, 19, 61-79. Gallie, D., O'Connell, P., Paugam, S., Scherer, S., and Tahlin, M. (2003). Skills and the quality of work, Prepared for workshop in Mannheim Aprill0-12 2003, Draft report from the Changequal Work Theme Group. Ganzeboom, H. B. G., and Treiman, D.]. (1996). Internationally comparable measures of occupational status for the 1988 International Standard Classification of Occupations; Social Science Research, 25, 201-39. - - - . (2003). Three internationally standardised measures for comparative research on occupational status; In Advances in cross-national comparison (Eds., Hoffmeyer-Zlotnik, J. H. P., and Wolf, C.). Kluwer, New York, pp. 159-93. Gilens, M. (2002). An anatomy of survey-based experiments; In Navigating public opinion: Polls, policy, and the future of American democracy (Eds., Manza, J., Cook, F. L., and Page, B. I.). Oxford University Press, Oxford, pp. 232-50. Glotz, P. (1984). Die Arbeit der Zuspitzung: Vber die Organisation einer regierungsfahigen Linken, Siedler, Berlin.

References

209

Goldthorpe, J. H. (1983 ). Women and class analysis: In defence of the conventional view; Sociology, 17,465-88. - - - . (1984). Women and class analysis: A reply to the replies; Sociology, 18, 491-99. - - - . (1995). The service class revisited; In Social change and the middle classes (Eds., Butler, T., and Savage, M.). UCL Press, London, pp. 313-29. - - - . (2000a). Rent, class conflict, and class structure: A commentary on Sorensen; American Journal of Sociology, 105, 1572-82. - - - . (2000b). Social class and the differentiation of employment contracts; In On sociology: Numbers, narratives and the integration of research and theory (Ed., Goldthorpe, ]. H.). Clarendon Press, Oxford, pp. 206-29. - - - . (2000c). Social ethnography today: Problems and possibilities; In On sociology: Numbers, narratives and the integration of research and theory (Ed., Goldthorpe, ]. H.). Clarendon Press, Oxford, pp. 65-93. - - . (2000d). Sociology and the probabilistic revolution 1930-1930: Explaining an absent synthesis; In On sociology: numbers, narratives and the integration of research and theory Clarendon Press, Oxford, pp. 259-94. Goldthorpe,]. H., and Heath, A. (1992). Revised class schema 1992, Joint Unit for the Study of Social Trends, Nuffield College, Oxford. Goldthorpe, ]. H., and Marshall, G. (1992). The promising future of class analysis: A response to recent critiques; Sociology, 26, 381-400. Gooderham, P. N., Nordhaug, 0., and Ringdal, K. (1999). Institutional and rational determinants of organizational practices: Human resource management in European firms; Administrative Science Quarterly, 44,507-31. Grabb, E. G. (1979). Working-class authoritarianism and tolerance of outgroups: A reassessment; Public Opinion Quarterly, 43, 36-47. - - - . (1980). Marxist categories and theories of class: The case of working-class authoritarianism; Pacific Sociological Review, 23, 359-76. Granberg, D., and Holmberg, S. (1988). The political system matters: Social psychology and voting behavior in Sweden and the United States, Cambridge University Press, Cambridge. Granovetter, M. (1985). Economic action and social structures: The problem of embeddedness; American Journal of Sociology, 91, 481-510. Grusky, D. B., and S0rensen,]. B. (1998). Can class analysis be salvaged? American Journal of Sociology, 103, 1187-1234. Grusky, D. B., and Weeden, K. A. (2001). Decomposition without death: A research agenda for a new class analysis; Acta Sociologica, 44, 203-18. Hacker,]. S. (2002). The divided welfare state: The battle over public and private social benefits in the United States, Cambridge University Press, Cambridge. Hadenius, A. (1986). A crisis of the welfare state? Opinions about taxes and public expenditure in Sweden, Almqvist & Wiksell International, Stockholm. Hall, P.A. (1992). The Movement from Keynesianism to Monetarism: Institutional Analysis and British Economic Policy in the 1970s; In Structuring Politics.

210

References

Historical Institutionalism in Comparative Analyses (Eds., Steinmo, S., Thelen, K., and Longstreth, F.). Cambridge University Press, Cambridge, pp. 90-113. - - - . (1997). The role of interests, institutions, and ideas in the comparative political economy of the industrialized nations; In Comparative politics: rationality, culture and structure (Eds., Lichbach, M. I., and Zuckerman, A. S.). Cambridge University Press, Cambridge, pp. 174-207. Hall, P.A., and Soskice, D. W. (2001). Varieties of capitalism: The institutional foundations of comparative advantage, Oxford University Press, Oxford. Hall, P.A., and Taylor, R. C. R. (1996). Political science and the three new institutionalisms; Political Studies, 44, 936-57. Hamilton, R. F. (1972). Class and politics in the United States, Wiley, New York. Heath, A., and Britten, N. (1984). Women's jobs do make a difference: A reply to Goldthorpe; Sociology, 18,475-90. Hedstrom, P., and Swedberg, R. (1998). Social mechanisms: An analytical approach to social theory, Cambridge University Press, Cambridge. Herek, G. M. (2000). The psychology of sexual prejudice; Current Directions in Psychological Science, 9, 19-22. Hochschild, A. (1989). The second shift: Working parents and the revolution at home, Viking, New York. Holton, R. ]., and Turner, B. S. (1989). Has class analysis a future? Max Weber and the challenge of liberalism to Gemeinschaftlich accounts of class; In Max Weber on economy and society (Eds., Holton, R. ]., and Turner, B. S.). Routledge, London, pp. 160-96. Hout, M., Brooks, C., and Manza,]. (1995). The democratic class struggle in the United States, 1948-1992; American Sociological Review, 60, 805-28. Houtman, D. (1997). Welfare state, unemployment, and social justice: Judgments on the rights and obligations of the unemployed; Social Justice Research, 10, 267-88. - - - . (2000). The working class and the welfare state: Support for economic redistribution, tolerance for nonconformity, and the conditionality of solidarity with the unemployed; Netherlands Journal of Social Sciences, 36, 3 7-5 5. Huber, E., and Stephens, J. D. (2001a). Development and crisis of the welfare state: Parties and policies in global markets, University of Chicago Press, Chicago. - - - . (2001b). Welfare state and production regimes in the era of retrenchment; In The new politics of the welfare state (Ed., Pierson, P.). Oxford University Press, Oxford. Huckfeldt, R. R., and Sprague,]. (1995). Citizens, politics and social communication: Information and influence in an election campaign, Cambridge University Press, Cambridge. Hult, C. (2005). Organizational commitment and person-environment fit in six western countries; Organization Studies, 26, 249-70.

References

211

Hult, C., and Svallfors, S. (2002). Production regimes and work orientations: A comparison of six western countries; European Sociological Review, 18, 315-31. Hyman, R., and Ferner, A. (1998). Changing industrial relations in Europe, Blackwell, Oxford. Iversen, T., and Soskice, D. (2001). An asset theory of social policy preferences; American Political Science Review, 95, 875-93. Jowell, R. (1998). How comparative is comparative research? American Behavioral Scientist, 42, 168-77. Kelley, ]., and Evans, M. D. (1993). The legitimation of inequality: Occupational earnings in 9 nations; American journal of Sociology, 99,75-125. King, D. S. (1995). Actively seeking work? The politics of unemployment and welfare policy in the United States and Great Britain, University of Chicago Press, Chicago. Kingston, P. W. (2000). The classless society, Stanford University Press, Stanford. Kitschelt, H. (1994). The transformation of European social democracy, Cambridge University Press, Cambridge. - - - . (1999). Continuity and change in contemporary capitalism, Cambridge University Press, Cambridge. Kitschelt, H., and Rehm, P. (2004a). New social risk and political preferences, Paper for the 14th International Conference of Europeanists, Chicago 11-13 March 2004. - - - . (2004b). Socio-econorrzic group preferences and partisan alignments, Paper for the 14th International Conference of Europeanists, Chicago 11-13 March 2004. Kjellberg, A. (1998). Sweden: Restoring the model? In Changing industrial relations in Europe, 2nd ed. (Eds., Hyman, R., and Ferner, A.). Blackwell, Oxford, pp. 74-117. Klingemann, H.-D. (1999). Mapping support in the 1990s: A global analysis; In Critical citizens: Global support for democratic government (Ed., Norris, P.). Oxford University Press, Oxford, pp. 31-56. Knudsen, K., and Wa:rness, K. (1999). Reactions to global processes of change: Attitudes towards gender roles and marriage in modern nations; Comparative Social Research, 18, 161-96. - - - . (2001). National context, individual characteristics and attitudes on mothers' employment: A comparative analysis of Great Britain, Sweden and Norway; Acta Sociologica, 44, 67-79. Kohn, M. L. (1977). Class and conformity: A study in values, University of Chicago Press, Chicago. Kohn, M. L., Miller,]., and Schooler, C. (1983). Work and personality: An inquiry into the impact of social stratification, Ablex, Norwood, NJ. Kohn, M. L., Naoi, A., Schoenbach, C., Schooler, C., and Slomczynski, K. M. (1990). Position in the class-structure and psychological functioning in the

212

References

United States, Japan, and Poland; American Journal of Sociology, 95, 964-1008. Kohn, M. L., Slomczynski, K. M., Janicka, K., Khmelko, V., Mach, B. W., Paniotto, V., Zaborowski, W., Gutierrez, R., and Heyman, C. (1997). Social structure and personality under conditions of radical social change: A comparative analysis of Poland and Ukraine; American Sociological Review, 62,614-38. Kohn, M. L., Slomczynski, K. M., and Schoenbach, C. (1990). Social structure and self-direction: A comparative analysis of the United States and Poland, Basil Blackwell, Oxford. Korpi, W. (1978). The working class in welfare capitalism: Work, unions and politics in Sweden, Routledge & Kegan Paul, London. Korpi, W., and Palme, J. (1998). The paradox of redistribution and strategies of equality: Welfare state institutions, inequality, and poverty in the western countries; American Sociological Review, 63, 661-87. - - - . (2003). New politics and class politics in the context of austerity and globalization: Welfare state regress in 18 countries 1975-1995; American Political Science Review, 97, 425-46. Kuechler, M. (1987). The utility of surveys for cross-national research; Social Science Research, 16, 229-44. - - - . (1998). The survey method: An indispensable tool for social science research everywhere? American Behavioral Scientist, 42, 178-200. Kumlin, S. (2003). Snedvridet ansvarsutkravande? In Demokratitrender (Ed., Oscarsson, H.). SOM-institutet Goteborgs universitet, Goteborg, pp. 31-53. Lamont, M. (1992). Money, morals, and manners: The culture of the French and American upper-middle class, University of Chicago Press, Chicago. - - - . (2000). The dignity of working men: Morality and the boundaries of race, class, and immigration, Harvard University Press, Cambridge, MA. Lane, R. E. (1986). Market justice, political justice; American Political Science Review, 80, 383-402. Laurin, U. (1986). Pa heder och samvete: Skattefuskets orsaker och utbredning, Norstedts, Stockholm. Levi, M. (1990). A logic of institutional change; In The limits of rationality (Eds., Cook, K. S., and Levi, M.). University of Chicago Press, Chicago, pp. 402-18. Lewis, J. (1992). Gender and the development of gender regimes; Journal of European Social Policy, 2, 159-73. Lewis, J., and Ostner, I. (1995). Gender and the evolution of European social policies; In European social policy. Between fragmentation and integration (Eds., Leibfried, S., and Pierson, P.). Brookings Institution, Washington, DC, pp. 159-93. Lincoln, J. R., and Kalleberg, A. L. (1990). Culture, control, and commitment: A study of work organization and work attitudes in the United States and Japan, Cambridge University Press, Cambridge.

References

213

Lindgren, G. (1992). Doktorer, systrar och flickor: Om informell makt, Carlsson, Stockholm. Lipset, S. M. (1959). Democracy and working-class authoritarianism; American Sociological Review, 24, 482-502. - - - . (1981). Second thoughts and recent findings; In Political man. The social bases of politics (Ed., Lipset, S. M.). Johns Hopkins University Press, Baltimore, pp. 459-523. Lipsitz, L. (1965). Working-class authoritarianism: A re-evaluation; American Sociological Review, 30, 103-9. Long, J. Scott. (1997). Regression models for categorical and limited dependent variables, Sage. Manza, J., and Cook, F. L. (2002). The impact of public opinion on public policy: The state of the debate; In Navigating public opinion. Polls, policy, and the future of American democracy (Eds., Manza, J., Cook, F. L., and Page, B. I.). Oxford University Press, Oxford, pp. 17-32. Marshal!, G. (1988). Social class in modern Britain, Hutchinson, London. Marshal!, G. (1997). Repositioning class: Social inequality in industrial societies, Sage, London. Marshal!, G., Swift, A., and Roberts, S. (1997). Against the odds? Social class and social justice in industrial societies, Oxford University Press, Oxford. Marx, K., and Engels, F. ([1848] 1967) The Communist manifesto, Penguin, Harmondsworth. Mau, S. (2003). The moral economy of welfare states. Britain and Germany compared, Routledge, London. Mettler, S., and Soss, J. (2004). The consequences of public policy for democratic citizenship: Bridging policy studies and mass politics; Perspectives on Politics, 2, 55-73. Meyersson, E. M., and Petersen, T. (2003). The glass ceiling in the US and Sweden: Lessons from the family-friendly corner of the world, Working paper, 21 Nov 2003. Miller, S. M., and Riessman, Frank. (1961). "Working-Class Authoritarianism": A critique of Lipset; British Journal of Sociology, 12, 263-76. Mills, C. (1994). Who dominates whom: social-class, conjugal households and political identification; Sociological Review, 42, 639-63. Mills, C. W. (1951). White collar: The American middle classes, Oxford University Press, New York. Morris, L., and Scott, J. (1996). The attenuation of class analysis; British Journal of Sociology, 47,45-55. Mowday, R. T., Steers, R. M., and Porter, L. W. (1982). Employee-organization linkages. The psychology of commitment, absenteeism and turnover, Academic Press, New York. Mueller, C., Wallace, J., and Price, J. (1992). Employee commitment: Resolving some issues; Work and Occupations, 19, 211-36.

214

References

Muller, W. (1999). Class cleavages in party preferences in Germany: Old and new; In The end of class politics? Class voting in comparative context (Ed., Evans, G.). Oxford University Press, Oxford, pp. 137-80. Nisbet, R. (1959). The decline and fall of social class; Pacific Sociological Review, 2, 11-17. Nordenmark, M. (1999). Unemployment, employment commitment and wellbeing: The psychosocial meaning of (un)employment among women and men, Department of Sociology, Umea University, Umea. North, D. C. (1990). Institutions, institutional change, and economic performance, Cambridge University Press, Cambridge. O'Connor, J. S., Shaver, S., and Orloff, A. S. (1999). States, markets, families: Gender, liberalism and social policy in Australia, Canada, Great Britain and the United States, Cambridge University Press, Cambridge. OECD (2001). Education at a glance. OECD indicators, OECD, Paris. Offe, C. (1985). Work: The key sociological category? In Disorganized capitalism (Ed., Offe, C.). Polity Press, Cambridge, pp. 129-50. - - - . (1987). Democracy against the welfare state? Structural foundations of neo-conservative political opportunities; Political Theory, 15, 501-37. Olsen, G. M. (2002). The politics of the welfare state: Canada, Sweden and the United States, Oxford University Press, Don Mills, Ont. Orloff, A. S. (1993). Gender and the social rights of citizenship: The comparative analysis of gender relations and welfare states; American Sociological Review, 58, 303-28. Osterman, T. (2000). Svenska befolkningens attityder till homosexuella och homosexualitet, Forskningsgruppen for samhalls- och informationssstudier, Stockholm. Pakulski, J. (2005). Foundations of a post-class analysis; In Approaches to class analysis (Ed., Wright, E. 0.). Cambridge University Press, Cambridge, pp. 152-79. Pampel, F. C. (2000). Logistic regression: A primer, Sage, Thousand Oaks, CA. Pierson, P. (2001). The new politics of the welfare state, Oxford University Press, Oxford. Plutzer, E., and Zipp, J. F. (2001). Class, gender, and the family unit: A dynamic model of stratification and class politics; Social Science Research, 30, 426-48. Polanyi, K. (1944). The great transformation, Farrar & Rinehart, New York. Pontusson, J. (1995). From comparative public policy to political economy: Putting political institutions in their place and taking interests seriously; Comparative Political Studies, 28, 117-47. Pontusson, J., Rueda, D., and Way, C. R. (2002). Comparative political economy of wage distribution: The role of partisanship and labour market institutions; British journal of Political Science, 32, 281-308. Prandy, K. (1990). The revised scale of occupations; Sociology, 24, 629-55.

References

215

Ray,].]. (1983 ). The workers are not authoritarian: Attitude and personality data from six countries; Sociology and Social Research, 67, 166-189. Rhodes, M. (2000). Restructuring the British welfare state: Between domestic constraints and global imperatives; In Welfare and work in the open economy. Vol II: Diverse responses to common challenges (Eds., Scharpf, F. W., and Schmidt, V. A.). Oxford University Press, Oxford, pp. 19-68. Rief, K.-H. and Marlier, E. (2001). Eurobarometer 44.2. Working conditions in the European Union, Interuniversity Consortium for Political and Social Research & Zentralarchive, Ann Arbor, Ml. Rothstein, B. (1996). Political institutions: An overview; In A new handbook of political science (Eds., Goodin, R. E., and Klingemann, H.-D.). Cambridge University Press, Cambridge, pp. 133-66. - - - . (1998). Just institutions matter: The moral and political logic of the universal welfare state, Cambridge University Press, Cambridge. - - - . (2000). Trust, social dilemmas and collective memories; Journal of Theoretical Politics, 12, 4 77-501. Rueda, D., and Pontusson, J. (2000). Wage inequality and varieties of capitalism; World Politics, 52, 350-83. Rydgren, J. (2003a). Mesolevel reasons for racism and xenophobia: Some converging and diverging effects of radical right populism in France and Sweden; European Journal of Social Theory, 6, 45-68. - - - . (2003b). The populist challenge: Political protest and ethno-nationalist mobilization in France, Berghahn Books, New York. - - - . (2004). Explaining the emergence of radical right-wing populism: The case of Denmark; West European Politics, 27,474-502. - - - . (2005). Is extreme right-wing populism contagious? Explaining the emergence of a new party family; European Journal of Political Research, 44, 413-37. Sainsbury, D. (1996). Gender, equality, and welfare states, Cambridge University Press, Cambridge. - - - . (1999). Gender and welfare state regimes, Oxford University Press, Oxford. Savage, M., Barlow, ]., Dickens, P., and Fielding, T. (1992). Property, bureaucracy, and culture: Middle class formation in contemporary Britain, Routledge, London. SCB (2003 ). Sysselsattning och arbets!Oshet 1975-2002, Statistiska Centralbyran, Stockholm. Scharpf, F. W., and Schmidt, V. A. (2000). Welfare and work in the open economy, Oxford University Press, Oxford. Schegloff, E. A., and Kruytbosch, C. (1961). Some comments on 'working class authoritarianism'; Berkeley Journal of Sociology, 6, 99-105. Schroeder, L. D., Sjoquist, D. L., and Stephan, P. E. (1986). Understanding regression analysis: An introductory guide, Sage, Beverly Hills, CA.

216

References

Scott,]. (1996). Stratification and power: Structures of class, status and command, Polity Press, Cambridge. Shavit, Y., and Muller, W. (1998). From school to work: A comparative study of educational qualifications and occupational destinations, Clarendon Press, Oxford. Slomczynski, K. M., Miller, J., and Kohn, M. L. (1981). Stratification, work, and values: A Polish-United States comparison; American Sociological Review, 46, 720-44. Smeeding, T. M. (2002). Globalization, inequality and the rich countries of the G-20: Evidence from the Luxembourg Income Study (LIS); In Globalisation, living standards, and inequality: Recent progresses and continuing challenges (Eds., Gruen, D., O'Brien, T., and Lawson, J.). MacMillan, Sydney, pp. 179-206. Smith, T. W. (2005). The International Social Survey Program Research, National Opinion Research Center, Chicago. S0rensen, A. (1994 ). Women, family and class; Annual Review of Sociology, 20, 27-47. Sorensen, A. B. (2000). Toward a sounder basis for class analysis; American journal of Sociology, 105, 1523-58. Soskice, D. W. (1999). Divergent production regimes: Coordinated and uncoordinated market economies in the 1980s and 1990s; In Continuity and change in contemporary capitalism (Eds., Kitschelt, H., Lange, P., Marks, G., and Stephens,]. D.). Cambridge University Press, Cambridge, pp. 101-34. Soss,]. (2005). Making clients and citizens: Welfare policy as a source of status, belief and action; In Deserving and entitled: Social constructions and public policy (Eds., Schneider, A., and Ingram, H.). State University of New York Press, Stony Brook, pp. 291-328. Stanworth, M. (1984). Women and class analysis: A reply to Goldthorpe; Sociology, 18, 159-70. Stjerno, S. (1995). Mellom kirke og kapital: Tysk velferdspolitikk: med sideblikk til britisk, svensk og norsk, Universitetsforlaget, Oslo. Streeck, W. (1997). German capitalism: Does it exist? Can it survive? In Political economy of modern capitalism (Eds., Crouch, C., and Streeck, W.). Sage, London, pp. 55-70. Sundstrom, E. (1999). Should mothers work? Age and attitudes in Germany, Italy and Sweden; International Journal of Social Welfare, 8, 193-205. - - - . (2000). Gender and attitudes towards female employment in Germany, Italy and Sweden; In Families and family policies in Europe (Eds., Pfenning, A., and Bahle, T.). Peter Lang, Frankfurt am Main, pp. 200-220. - - - . (2002). National policies, local policies and women's right to work, Umea University, Umea. Svallfors, S. (1989). Vem alskar valfardsstaten? Attityder, organiserade intressen och svensk valfardspolitik, Arkiv, Lund.

References

21 7

- - - . (1993). Dimensions of inequality: A comparison of attitudes in Sweden and Britain; European Sociological Review, 9, 267-287. - - - . (1995). The end of class politics? Structural cleavages and attitudes to Swedish welfare policies; Acta Sociologica, 38, 53-74. - - - . (1996a). National differences in national identities? An introduction to the International Social Survey Program; New Community, 22, 127-34. - - - . (1996b). Valfardsstatens moraliska ekonomi: Valfardsopinionen i 90-talets Sverige, Bon~a, Umea. - - - . (1997). Worlds of welfare and attitudes to redistribution: A comparison of eight western nations; European Sociological Review, 13, 283-304. - - - . (1999). Mellan risk och tilltro: Opinionsstodet for en kollektiv valfardspolitik, Umea University, Umea. - - - . (2003a). Welfare regimes and welfare opinions: A comparison of eight western countries; Social Indicators Research, 64, 495-520. - - - . (2003b). Valfardsstatens legitimitet: Asikter om svensk valfardspolitik I komparativ belysning; In Sjukskrivning: Forsakring eller forsorjning Forsakringskasseforbundet, Stockholm, pp. 85-104. Svallfors, S., Halvorsen, K., and Andersen, J. G. (2001). Work orientations in Scandinavia: Employment commitment and organizational commitment in Denmark, Norway and Sweden; Acta Sociologica, 44, 139-56. Swank, D. (2002). Global capital, political institutions, and policy change in developed welfare states, Cambridge University Press, Cambridge. Swift, A. (2000). Class analysis from a normative perspective; British Journal of Sociology, 51, 663-79. Tabachnick, B. G., and Fidell, L. S. (1996). Using multivariate statistics, HarperCollins College Publishers, New York. Taylor-Gooby, P. (1982). Two cheers for the welfare state: Public opinion and private welfare; Journal of Public Policy, 2, 319-46. Thelen, K. (1999). Historical institutionalism in comparative politics; Annual Review of Political Science, 2, 369-404. Thompson, E. P. (1963). The making of the English working class, Gollancz, London. - - - . (1971). The moral economy of the English crowd in the eighteenth century; Past and Present, SO, 71-136. Treas, J., and Widmer, E. D. (2000). Married women's employment over the life course: Attitudes in cross-national perspective; Social Forces, 78, 1409-36. Turner, F. C. (1992). Social mobility and political attitudes: Comparative perspectives, Transaction Publishers, New Brunswick, NJ. van Parijs, P. (1989). A revolution in class theory; In The debate on classes (Ed., Wright, E. 0.). Verso, London, pp. 213-41. Walzer, M. (1983 ). Spheres of justice: A defense of pluralism and equality, Basic Books, New York.

218

References

Warr, P. (1982). A national study of non-financial commitment; Journal of Occupational Psychology, 55, 297-312. Weil, Frederick D. (1985). The variable effects of education on liberal attitudes: A comparative-historical analysis of anti-Semitism using public opinion survey data; American Sociological Review, 50, 458-74. Weir, M. (1992). Politics and jobs: The boundaries of employment policy in the United States, Princeton University Press, Princeton, NJ. Western, B. (1997). Between class and market: Postwar unionization in the capitalist democracies, Princeton University Press, Princeton, N]. Willis, Paul. (1977). Learning to labour. How working class kids get working class jobs, Saxon House, Farnborough. Wright, E. 0. (1979). Class, crisis and the state, Verso, London. - - . (1985). Classes, Verso, London. - - - . (1997). Class counts: Comparative studies in class analysis, Cambridge University Press, Cambridge Young, I. M. (1990). Justice and the politics of difference, Princeton University Press, Princeton, NJ. Zipp, ]. F., and Plutzer, E. (1996). Wives and husbands: Social class, gender, and class identification in the US; Sociology, 30, 235-52.

INDEX

Italic page numbers indicate tables or figures.

age groups, 200n9 aggregation, 15, 160, 169 analytic framework, 26-28 analytical methods, 189-92 attitudes: toward authority, 124, 125; and childhood class, 96; and class relations, 19; context-dependent nature of, 178; and dearticulation, 5; defined, 18; toward education, 201n13; and family, 19-20; toward income distribution, 56; indexes of, constructed, 19, 18688, 190; toward law, obeying, 201n13; and national differences, 105; and political articulation, 24, 26; politics, as shaper of, 175; vs. positive beliefs, 18; and recommodification, 5; toward redistribution, 19; research into, 182; scope of, 19; vs. values, 18; toward welfare state (see welfare states); toward work, 19 attitudinal domains, 19, 20, 20, 28 attitudinal fragmentation, 75 attitudinal integration, 72, 73, 76 attitudinal surveys, 180-81 authoritarianism, 124, 125, 199n1 authority, attitudes toward, 124, 125 authority, defined, 200n3 autonomy, at work, 33, 34 Beck, Ulrich, 193n3 blocked exchanges, 68 bread riots, 1 breadwinner, man as, 105, 106, 150, 151, 198n5

British Standard Occupation Classification (SOC), 183 buy-sell systems, 3 Cambridge scale, 11-12 "canned surveys," 182 career commitment, 42-44 child care, 80, 106, 186, 187, 198n3 childhood class, 91, 92, 95, 96, 101, 164 childhood socialization, 92, 95, 101-2 Christian Democrats, 146 civil liberties and rights, attitudes toward: 122; and authoritarianism, 124; class differences, 128,136-39,140, 141; content of, 129; and education, 139, 140

class: and attitudes, 8, 26, 26, 27, 28, 102, 171 (see also class differences, attitudinal); background (see childhood class); and class identity, 154; classification schemes for, 153-54; and collective memories, 25; conflicts, 75; and conformism, 13 9, 141; and dearticulation, 3, 4; definition of, 8, 9, 11, 12, 15, 162; demarcation of, 13; dissolution of, 2; and dynamism, change, 2-3, 16; economic vs. social, 12; and employment relations, 12; ethnographic studies of, 176-77; and exploitation, 193n1 (chap. 2); factors influenced by, 12; fragmentation of, 169, 172; gender and age composition of, 113; and history, 3; identity, 14, 23, 24; individual vs. household models of, 150-55, 164; and 219

220

Index

institutions, 26, 173; interests, 20-21, 23, 92; and market, 52; minimalist conception of, 12; mobility, 16; models, 15, 154, 160; nominalistic stance on, 13; and occupation, 15, 150; operationalization of, 162; and party sympathy, 154; and political articulation, 26, 162; and political discourse, 4; and political institutions, 26, 162; politics, and welfare state, 77; position, 16, 28, 91, 155; position vs. mobility, 16; and recommodification, 3, 4; relations, 17, 19; research into, 176; and social attitudes, 3; and social conditions and processes, 6; as social demarcator, 171; and social outcomes, 7; static vs. dynamic aspect of, 16; and status scale, 10; structuration, mechanisms of, 91-97; structure, and recommodification, 17; and voting behavior, 25 class analysis: aggregated, 15, 16; communitarian approaches to, 11; disaggregated, 16; "dominance" method, 14; and institutions, 23; strengths and limitations of, 6-7; units for, 13-14, 15; and work, attitudes toward, 14 class context, 102 class differences, attitudinal: and attitude change, 56; and attitude integration, 73; and authoritarianism, 122; civil liberties and rights, 128, 129, 136-39, 140, 140, 141, 164; and classlessness, 16869; in commitment, career, 42, 44-45, 46; in commitment, organizational, 4245, 43, 46, 50, 162; in conformism, 122, 126, 164, 165; in coordinated vs. liberal market economies, 74; toward extramarital sex, 130-35; and financing index, 88; and gender ideology, 105,109,110,111,112,164, 199n13; toward homosexuality, 130-35; toward household chore distribution, 104, 106, 110,112,116,117,118,119, 164;and incentives, 67; and income disparity, 55-56,57,59, 63, 64,54-68,74, 163; and institutions, 166, 176; and job, sense of have/lack in, 39, 39-41; and job satisfaction, 41, 41-42, 46, 47, 162; among jobless vs. employed, 15560; and living conditions, 98; and marginalization, 155-60; toward market,

change in attitudes, 56; and market distribution, 53, 74, 75; and market fatalism, 53; market penetration as reinforcer of, 172; mechanisms underlying, 168, 176; toward military force, use of, 129; toward minorities, 126-27, 131; and morality, 128, 129, 164; and national differences, coincidence of, 83; and occupational self-direction, 126; and political articulation, 166, 173; toward redistribution, 53, 163, 165; and risks, 77, 168; and self-interest, 170; toward sexual behavior, 129-35, 141; size of, 194n1; and skill specificity, 170; stability of, 88, 91, 163; statistical significance testing of, 194n5; and "tribalization," 171; toward welfare state, 77, 78, 81, 82, 82, 83-86, 163; toward work, intrinsic value of, 35, 35-36, 37, 38, 46, 50, 51, 162; and workplace relations, 46 class perspective, vs. marginalization perspective, 160 class struggle, vs. inclusion-exclusion, 155 class-attitude nexus: and class position, 91, 92; and economic resources, 91, 92-95; and institutions, 167, 167; and marginalization, 147, 155-60; mechanisms underlying, 91, 93-94, 95, 100, 102, 171; national differences in, 168; and rational choice, 171; unemployment risk, 92, 9 5; and welfare state, 167; and work, nature of, 91, 92, 93, 95; and work conditions, 163-64 class-based norms, 20, 21 class-conscious outcomes, 7 classlessness, 168-69 class-redistributive institutions, 167 class-related distributive processes, 81 collective memories, 24-25 collective wage bargaining. See wage bargaining; unions communitarian approach, to class analysis, 11 communitarian perspective, and John Scott, 12 company commitment. See organizational commitment comparative attitude research, 178 -79 Comparative Class Analysis Project, 182 comparative social attitudes, 6

Index competition, and institutions, 23 complex equality, 68, 73 conflict index, 65-66, 66 conflict index, reliability of, 196n5 conformism: vs. authoritarianism, 124; and class differences, 122, 126, 164, 165; class structure of, 139; indicators of, 129; and national differences, 127, 139, 141; and occupational self-direction, 126; vs. redistribution, 165; and social issues, 128-29; and work situation, 126 Conservatives, 80 consumption, attitudes toward: research into, 182 contingency model, 1 75 contracts, employment, 9-10, 17,23 conventional classification, 154 coordinated market economies, 54 corporate welfare, 79 Cronbach's alpha, 187, 188 "cultural turn," 173, 174 data bases, 178-83 datasets, 29 dearticulation, 3, 4-5, 26 democracy, 125, 129 Democrats, 146 disaggregation, 16, 169 disposable household income, 55, 93 distribution, 1, 52, 121, 151 division of labor, 9 dominance method, 14, 154 dominance principle, 151 dominant classes, and gender role expectations, 199n13 drop-out rate, for surveys, 179, 180, 181, 182 early retirement, 156, 157 economic classes, vs. social classes, 12 economic crisis of 1990s, 88, 99-100 economic liberalism, 121 education: and authoritarianism, 125; categories coded into, 200n9; and civil rights and liberties, attitudes toward, 139, 141; and class effects, 126; and class position, 133; and conformism, 125, 126, 133, 134; and infidelity, attitudes toward, 200n10; and organizational commitment, 195n9; purchase of,

221

attitudes toward, 68, 69, 187; and survey sample respondents, 180 educational attainment, 6 egalitarianism, and gender roles, 104 elderly care services, 80, 106, 186, 187 elites, 175 employment, private vs. public sector, 197 employment commitment, national differences in, 35 employment contracts, 9-10, 17,23 employment policies, 186 employment relations, 3, 9, 12, 19, 21, 166 employment status, 14-15 Employment Status Variables, 183 employment systems, 32 equality. See civil liberties and rights, attitudes toward; inequality Erikson, Robert, 152, 152-53, 184 eta, defined, 195n8 ethnographic studies, 176-77 exclusion. See marginalization exploitation, 193n1 (chap. 2) external authority, 124 extramarital sex, 130-35, 200n10 extremist parties, 124 factor analysis, 189 families, defined, 198n2 family: class position of, 103; class relations within, 113; formation of, 19-20; and household class, 13-14, 28, 15055; household income proportions, 107; and intrafamily relations, 104; and male breadwinner models, 105; and national differences, 104; opinions about, 103; power balance in, 103, 104, 107; valuations of, 103. See also gender female career role, 105, 198n5. See also gender female labor force participations, 6, 104, 106-7. See also gender finance index, 186 financial problems index, 187-88 financing index, 87, 88 friendships, 12 gender: differences, 86; equality index, 112, 112; female career role, 105, 198n5; and homosexuality, attitudes toward, 200n10; and "housewife sod-

222

Index

ety," 172; ideology, 105; male breadwinner models, 105, 198n5; man's class position, 150, 152; patterns, 6; regimes, 105, 106-9; role expectations, 199; and welfare policies, views on, 89, 91; women, occupations of, 151; women, and party-class nexus, 151; women, role of, 172; women and household class, 150; women in labor market, 6, 104, 106-7, 119-20, 151-52, 154, 164. See also household work, attitudes toward general-attitude indexes, 19 geographical mobility, 172 global capital and financial markets, 3 government intervention. See welfare states Greens, 146 Grusky, David, 167, 169 health care, 68, 69, 186, 187 higher education, 67, 122, 141. See also education homosexuality, attitudes toward, 130-35, 200n10 household. See family household class vs. individual class, 15055 household model, of class, 164 household work, attitudes toward: and actual work done, 116, 118; and class differences, 104,106, 110, 116; and distributive fairness, 106, 113, 116-19; and gender, 103-4; males performing, 164 "housewife society," 172 housing allowances, 186 housing standards, 3 human assets, 9-10 identity, and work, 171-72 immigration, attitudes toward, 126 incentive index, 65, 66, 66 incentives, 23, 65, 67, 74, 163 inclusion-exclusion. See marginalization income disparity, 3, 19; causes of, 56, 6263; and class differences, 55; as incentive, 65, 74; legitimacy of, 56, 57, 58, 59-62; and national differences, 5556,59,61 income redistribution index, 62

income-welfare purchase relationship, 68, 69, 70, 73, 187 individual class vs. household class, 15055, 160-61 individual model, of class, 105, 106, 152, 164 individual vs. family debate, 14 individual-contra-family class problem, 147, 152 "individualised inequality," 193n3 industrial society, and class, 2 inequality: causes of, perceived, 56; and conflict, 65-66, 66, 67; fairness of, 6871; as incentive, 65, 67, 164; and income-welfare purchase relationship, 68, 69, 70, 73, 187; "individualised," 193n3; and market, relationship to, 71; and political reform, 67; and theory of justice, 68. See also income disparity; gender infidelity, 130-35, 200n10 infidelity, and age, 200n10; and education, 200n10 institutional conditions, 53-54 institutionalist literature, 193n5 institutionalist theories, 22, 23, 24 institutions: and class differences, 166, 167, 167; countries, 27; definition of, 22, 23; and inequality, 75; rational choice approach to, 193n6; and welfare purchase, 71 insurance coverage, 3, 51, 81 interests, 20, 21-22 internal markets, 3 International Social Survey Program (ISSP), 29, 178, 179 International Standard Classification of Occupations, 1968 (ISC068), 182183 International Standard Classification of Occupations, 1988 (ISC088), 183 intrinsic value of work, 35, 35-36, 37, 38,50-51,162, 194n2 Iversen, Torben, 170 job creation and training, 33 job satisfaction, and class, 41, 41-42, 163 joblessness, vs. employment, as class divide, 155-60 Jonsson, Janne, 184

Index Kingston, Paul, 168-69 Kahn, Melvin, 126 labor, social distribution of, 3 labor market, and power imbalances, 52 labor market institutions, 75 Labour, 80 lack, sense of, in jobs, 39, 39-41 law, attitudes toward, 201n13 leftism, 165, 166, 173-75 liberal market economies, 54, 56, 61-62, 74 liberalism, 121 libertarian-authoritarian dimension, 124 life chances, 19 linear regression analysis, 191, 203n2 living conditions, 6 logistic regression, 116, 191, 192 male breadwinner models, 105, 198n5 management strategies, 34 managerialism, 32 managers vs. professionals, 145-46, 148 Manifesto Research Group, 173 man's class position, 150, 152 marginalization, 147, 155-60, 161, 165 marginalization perspective, vs. class perspective, 60 market distribution, 7, 53, 75, 81 market fatalism, 53, 55, 75 market penetration, and class differences, 172 market research, 181 market transactions, and power imbalances, 52 market-based systems, 176 markets: and class differences, attitudinal, 52, 74; compass of, ideal, 56; coordinated vs. liberal, 54; distributive properties of, 52-53; institutional framework of, 53; and moral economy, 52; operational nature of, 52-53; selfregulating, 53 means of production, ownership of, 9 media, 5, 76, 166 Medicare/Medicaid, 79 mental packages, 71 military force, attitudes toward use of, 129 Mills, C. Wright, 146, 201n1 minorities, attitudes toward, 124, 126-27

223

miscarriage of justice, attitudes toward, 201n13 mobility. See social mobility Moderates, 173-74 monotonous job index, 188 moral conservatism, 121 moral economy of class, 52, 162, 171, 173 morality, and attitudes, 128, 129, 164 multiple linear regression, 190 national differences, attitudinal: and attitudinal integration, 72, 73; and authoritarianism, 122-23; in child care, 106; in conformism, 122-23, 127, 139, 141; and coordinated vs. liberal markets, 54; in disposable household income, 55; and elderly care, 106; and gender, 104, 106-9, 112, 119-20, 164; and household work, 119-20; and income disparity, 55-56, 58, 59; and inequality, 63, 64, 65-68, 163; and markets, institutional framing of, 53; and redistribution, 165-66; size of, 51; and social integration, 127, 128; stability of, 163; and state intervention, 82, 82; statistical significance testing of, 194n5; and tax systems, 106; and unions, 54; validity of indicators for, 178; and welfare state, 79-86, 163; and work, aggregational levels, 167-68, 169; work organization, 33,34,163 national health service, 80 neoclassical economics, 22 "neo-classical soup," 169, 172 neoliberalism, 4 neo-Marxist class models, 15 net class difference, 201n11 net education difference, 201n11 nominalistic stance, on class, 13 noneconomic liberalism, 121 nonfinancial employment commitment, 35,35 nonmanual workers, 18 Nordic Occupational Classification (NYK), 183 normative feedback mechanism, 2 norms, 20-22, 23 objects of self-gratification, 21 obligation, 1

224

Index

occupational codes, 183-85 occupational labels, 15 occupational self-direction, 126 occupations, 9 odds ratios, 116-17 ordinal regression, 191, 192 ordinary least squares (OLS), 190 organizational commitment: vs. career commitment, 42, 43; class differences in, 44, 50, 51, 162; and education, 195n9; index, 48, 48-49, 49; meaning of term, 42; national differences, 4243, 44, 44-45; and union strength, 51; and workplace relations, 46 outsourcing, 3 parental leave systems, 107 party sympathies, 151-52 pay-spread indexes, 57, 58, 59, 60 pensions, 79, 187 perceptions, and institutions, 23 physically demanding work index, 188 point estimates, 194n5 political articulation: and attitudes, modification of, 24, 25; and class differences, 22, 24-25; and class-attitude nexus, 26, 26, 166, 167; importance of, 24, 25, 162; and inequality, attitudes toward, 75, 76; lack of, 76; of principles of distribution, 74 political discourse, class content in, 4 political entrepreneurship, 24 political institutions, 1-2, 22-23, 24, 25, 26, 162 political left, and class content articulation, 5. See also leftism political reform, and inequality, 67 political right, 4, 76, 98, 173 political sociology, 121, 124 political sympathies, 5, 14, 129 politics, 53,98-99,175, 176, 195n1 populism, 173 positive beliefs, 18 "postmodern" stratification analysis, 17172 poverty, attitudes toward, research into, 182 power imbalances, and labor market, 52 principles of distribution, political articulation of, 74 private vs. public sector employment, 197

private welfare state, U.S. as, 80 production systems, rule-oriented vs. skilloriented, 32 professionalism, 32 professionals vs. managers, 145-46, 148 public care systems, 105 public discourse, class content in, 5 public health services, 104 public policies, 1-2 pure market distribution, 17 race, attitudes toward, 126, 199n2. See also minorities, attitudes toward randomness, of social processes, 169 rational choice, 23, 171, 193n6 reciprocity, 1 recommodification, 3, 4, 5, 17, 75 recruitment, and institutions, 23 redistribution: attitudes toward, 19; class differences, 165; and conflict, 56, 73; and dearticulation of class, 4-5; and income disparity, 56; indicators of, 196n11; and man's class position, 152; and Moderates, 174; national differences, 81; and Social Democrats, 174; and welfare purchase, 73 regime typology, 77-78 regression coefficient, 191 regression methods, 190, 191-92, 203n2 relations, and markets, 53 reliability tests, 190 "rents," 18 reproduction, and the family, 19 research, 176 -77 research environments, growth of, 6 resources, 168 responsibility, 1 rhetoric, 24 right-wing political beliefs. See political right rule-oriented production systems, 32, 33 samples, weighting of, 180, 181 sampling and response rates, 179 Scott, John, 12 SEI (Socio-economic Index) codes, 14647, 149, 185 self-gratification, objects of, 21 self-interest, and class differences, 170 service contract, 10 service index, 87, 89, 187

Index sexual behavior, attitudes toward, 122, 128,129-35,141 sickness insurance, 186 simple equality, 68, 73 skill specificity, 170 skill-oriented production systems, 32, 33 social action, 25, 26 social assistance, 156, 186. See also welfare states social attitudes: and class differences, 7, 171; and class patterning, 23; and class position vs. class mobility, 16; and social action, 25, 26; and social mobility, 16; and values, 6; and work and employment relations, 11 social democracy, 4, 5 Social Democrats, 146, 173-74 social distribution of labor, 3 social exclusion. See marginalization social identity, 5 Social Inequality survey, 195n3 social insurance, 5, 77, 104, 105. See also welfare policies; welfare states social issues, 128 social liberalism, 4 social mobility: actual vs. anticipated, 17; and childhood class, 95; and class analysis, 6; and institutions, 23; intergenerational, 16; and occupation, 150; and social attitudes, 193n2 social outcomes, and class, 7 social processes, randomness, 169 social relationships, in workplace, 42 social services, 3 social work, 187 socialization, 101-2 sociocultural cleavage dimension, 124 Socio-economic Classification (SEI), 183. See also SEI sociological theories, 22 Sorensen, Aage, 17,169,172 Sorensen, Jesper, 167 Soskice, David, 170 spending index, 87, 87 "spot market" contracts, 17 stability, 163 state intervention. See welfare policies; welfare states statistical significance testing, 194n5 Statistics Sweden, 29 status scale, 10-11, 11-12

225

"survey fatigue," 181 survey subjects, representativeness of, 179-80 survey-based research, 176-81 surveys, 27, 28, 182 Swedish Institute for Social Research, 29, 184 Swedish Social Science Data Service, 178 Swedish Trade Union Confederation, 172 Swedish Welfare State surveys, 29, 181 Tahlin, Michael, 184 taxes, 5, 23 theory of justice, 68 "third-way" policies, 4, 5 Thompson, E. P., 1 "Three Worlds of Welfare Capitalism, The" (Esping-Andersen), 104 three-class model, 152 trade unions. See unions "tribalization," of society, 171 "2/3 society," 155, 156 ULF (Statistics Sweden survey), 181-82 unemployment, 3, 156, 157 unemployment insurance, 79, 186 unions: and class, creation of, 15; and national differences, 15, 54, 75, 166, 172; and partisan mentality, 34; and service class, 15; and trade union movement, 4; and vocationalism, 32; and work attitudes, 35; and workplace, relations in, 45-46, 49-50 universal child benefit, 80 universal health insurance, 79 varimax rotation, 189 vocational training, 32, 33 vocationalism, 32 voting behavior, 6, 25, 75, 146 wage bargaining: and class differences, 75; and class-attitude nexus, 167; and coordinated vs. liberal markets, 54; and dearticulation, 5; institutionalization of, 53; and market liberalism, 62; in market-based system, 176; national differences, 166; and recommodification, 3, 17; views on, 19. See also unions weighting, of samples, 180, 181

226

Index

welfare policies: areas studied, 186, 187; attitude-employment nexus, 155-60; and conflict, 78; and gender views on, 89, 91; and market compatibility, 100; and national differences, 77-78, 98, 99; and 1990s economic crisis, 88; from 1997 to 2002, 89; and regime typology, 77-78 welfare purchase, 69, 70, 73. See also income disparity welfare purchase index, 70 welfare services, 5, 51 welfare spending index, 186 welfare states: and attitudes, and class, 97, 167; and attitudes, and gender, 104; attitudes, index on, 83, 84; and attitudes, and marginalization, 157; and attitudes on cohabitation, 104; benefits, 23; and class politics, 77; and democracy, 78; liberal, 97, 98; and life chances, 77; and male breadwinner models, 105; national differences in, 80, 98, 166; programs, 3; and redistribution, 77; and risk regulation, 77; social democratic, 97; stability of, 100; survival of, 78

White Collar (C. Wright Mills), 146 Willis, Paul, 201n16 women, roles of, 150, 151, 152, 154, 172. See also gender work: access to, 104; attitudes toward, 31, 34, 35, 39; conditions, and classattitude nexus, 163-64; data on, 28; and identity, 171-72; intrinsic value of, 35,35-36,37,38,162, 194n2 workers, and welfare state, 77 working-class authoritarianism, 121-22, 123 working-life mobility, 17 workplace, relations in, 42, 45-46, 4950 Wright, Erik Olin, 15, 20-21, 182, 193n1 (chap. 2), 198n1

xenophobia, 131 Zentrum fiir Umfragen, Methoden und Analysen (ZUMA), 184