Ethnic Diversity and Solidarity : A Study of Their Complex Relationship [1 ed.] 9781443891905, 9781443881715

Citation preview

Ethnic Diversity and Solidarity

Ethnic Diversity and Solidarity: A Study of their Complex Relationship Edited by

Paul de Beer and Ferry Koster

Ethnic Diversity and Solidarity: A Study of their Complex Relationship Edited by Paul de Beer and Ferry Koster This book first published 2017 Cambridge Scholars Publishing Lady Stephenson Library, Newcastle upon Tyne, NE6 2PA, UK British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Copyright © 2017 by Paul de Beer, Ferry Koster and contributors All rights for this book reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the copyright owner. ISBN (10): 1-4438-8171-6 ISBN (13): 978-1-4438-8171-5

CONTENTS

Acknowledgements ................................................................................... vii Introduction ................................................................................................. 1 Paul de Beer Chapter One ............................................................................................... 13 Towards an Interdisciplinary Theory of Solidarity Paul de Beer, Maarten Berg, Laurens Buijs, Ferry Koster and Dorota Lepianka Chapter Two .............................................................................................. 43 Why Would Diversity Harm Solidarity? Paul de Beer Chapter Three ............................................................................................ 55 Heterogeneity and Human Sociality Ferry Koster Chapter Four .............................................................................................. 95 Diversity, Similarity, and Solidarity: Results of a Vignette Study Maarten Berg and Ferry Koster Chapter Five ............................................................................................ 127 More or Less Strangers: Diversity in the Dutch Media Depiction of Ethnic Minorities and its Implications for Interethnic Solidarity Dorota Lepianka Chapter Six .............................................................................................. 155 Solidarity in a Multicultural Neighbourhood: Results of a Field Experiment Paul de Beer and Maarten Berg

vi

Contents

Chapter Seven.......................................................................................... 189 Conclusion Paul de Beer Affiliation of the Authors ........................................................................ 199

ACKNOWLEDGEMENTS

This book is the result of a research project that was conducted by the Amsterdam Institute for Advanced Labour Studies (AIAS) (www.uvaaias.net) of the University of Amsterdam and was made possible by a grant of Instituut Gak (www.instituutgak.nl). We thank Aafke Komter, Ruud Koopmans and Wim van Oorschot for their very helpful comments and suggestions. A very special thanks to Merle Zwiers for her invaluable assistance during the research project.

INTRODUCTION PAUL DE BEER

In the past decade, both policy makers and academic scholars have called attention to the possible detrimental effects of ethnic heterogeneity on socially desirable outcomes, such as public goods provision, economic growth, trust, social capital and solidarity (e.g., Alesina and LaFerrara 2005; Leigh 2006; Putnam 2007; Habyarimana et al. 2007; Hooghe et al. 2009). In their view, in ethnically heterogeneous neighbourhoods, cities, regions or countries, citizens find it more difficult to identify with each other, trust each other less and are more reluctant to co-operate. This would be related to the perceived social distance between people who belong to different ethnic groups (cf. van Kempen and Bolt 2009). Some scholars claim that ethnic diversity might also hamper trust and cooperation between people within the same ethnic group (Putnam 2007). The consequences of ethnic heterogeneity might thus be broad and farreaching. Ethnically diverse neighbourhoods may be characterized by less community activities and be plagued by vandalism and crime (Wilson 1987; Ellen and Turner 1997; Morenoff et al. 2001; Haynie et al. 2006), ethnically diverse cities may lack effective governance and may have less public provisions such as parks and sports facilities , and ethnically diverse countries may offer its citizens less protection by social services and income transfers and may turn out to be less prosperous in the long run. As a consequence of immigration and, in addition, fertility rates which are often higher among ethnic minority groups, ethnic diversity is growing in many regions and countries (Zorlu and Hartog 2002). If the worries expressed above are justified, this may have serious detrimental effects. Ultimately, increasing ethnic diversity might even result in a higher incidence of riots, civil war and ethnic cleansing. Both from a scientific and from a societal point of view it is, therefore, of great importance to know more about the societal consequences of ethnic diversity. Up to date, the results of empirical studies are, at best, mixed. Some studies confirm the expectation that ethnic diversity erodes trust, social capital and public goods provisions (cf. Kleinhans et al. 2007), while other studies find no

2

Introduction

relationship at all between ethnic diversity and various outcomes at the community level, and a few studies even find positive effects of ethnic diversity (cf. Portes and Zhou 1993; Portes 1995). This volume adds to the existing knowledge of the effects of ethnic diversity in three important ways. Firstly, of the many societal outcomes of ethnic diversity that have been studied, this study focuses on just one, namely social solidarity, rather than social capital, trust or public goods provision. By solidarity we mean any act that purposefully benefits another person at a cost for the agent, without any guarantee of an equivalent return. While social capital, trust or co-operation does not impose any cost on the actor, solidarity does. Even though this cost may be repaid in the future, there is no guarantee that this will happen. Thus, solidarity is a stronger indicator of pro-social behaviour than most other indicators. Consequently, an empirical study that focuses on the intention to act solidary or on solidary behaviour itself offers a more robust test of the effects of ethnic diversity than most previous studies. Secondly, in most previous studies ethnic diversity or heterogeneity is measured by a single, one-dimensional indicator, such as the fractionalization index, which measures the probability that two persons, meeting each other at random, belong to different ethnic groups. In this volume, we take a broader perspective on diversity by using various ways to measure ethnic diversity. Thus, we do not assume a priori that there is one best way to measure diversity or that there is a simple linear relationship between ethnic diversity and solidarity. To illustrate, we do not assume a priori that there is a dichotomy between natives (autochthones) and non-natives or immigrants (allochthones), but we also examine the possible tensions between various ethnic groups and consider the possibility that religious differences or linguistic differences matter more than ethnic differences. Thirdly, this study applies various research methods, including a field experiment in a multicultural neighbourhood, a content analysis of mass media, a vignette study of a nation-wide survey and a statistical analysis of a large cross-country survey. The use of different research methods allows us to cross-examine the outcomes of the various studies that this volume reports on. Thus, the conclusions we draw do not depend on one particular research method and one particular pool of data, but are corroborated by various methods.

Ethnic Diversity and Solidarity: A Study of their Complex Relationship

3

In this introductory chapter, we will take a closer look at the two central concepts of this volume, viz. solidarity and diversity.

What is solidarity? As mentioned above, we define a solidary act as any act that purposefully benefits another person at a cost for the agent, without any guarantee of an equivalent return. This definition of solidarity includes a broad range of pro-social behaviours, such as alms-giving to a beggar, helping your neighbour, raising children, voluntary community work, donating money to a charity organization, paying an insurance premium and paying taxes. These examples show that solidarity can be both informal and formal, either voluntary or mandatory, and one-sided as well as two-sided or reciprocal (cf. De Beer and Koster 2009, chapter 2). Informal solidarity, which one might also call ‘warm’ solidarity, springs from a direct involvement and sympathy with concrete other persons, for whom one has ‘warm feelings’. The most intense forms of informal solidarity are usually found within the family: between husband and wife, and between parents and children. This solidarity can be so strong that one is prepared to sacrifice one’s life to save another, e.g., a parent that risks her life in trying to save her drowning child. Less intense forms of informal solidarity are often found between good friends, neighbours or colleagues. The compassion expressed by giving alms to a beggar is also an example of informal solidarity. Formal or ‘cold’ solidarity originates from commitment to or sympathy with anonymous others, whom one does not know personally, but to whom one is nevertheless connected through a formal bond. Some examples are the solidarity with poor fellow citizens, which are supported by a social benefit system to which everybody contributes, solidarity with the unlucky person whose house burns down and who is compensated by the insurance company to which one regularly pays an insurance premium, and the solidarity with poor people in developing countries to which your government pays development aid, which is financed from taxes. Whereas informal solidarity is expressed by the direct relation between persons, the formal solidarity between persons who do not know each other is always mediated by an institution such as the state or an insurance company.

4

Introduction

A solidarity act can be voluntary as well as compulsory. Showing solidarity voluntarily is a choice; compulsory solidarity is enforced, usually by the state. Compulsory solidarity does not rule out that one is actually prepared to act out of solidarity, but this is, contrary to voluntary solidarity, not a prerequisite. There is often a close relationship between the (in)voluntariness of solidarity and the organization of solidarity. However, not all informal solidarity is voluntary and not all formal solidarity is compulsory. Contributing a premium to a fire and theft insurance is an example of formal voluntary solidarity, and the solidarity of parents with their children is informal but compulsory, since they are legally obliged to take care of their children. In case of two-sided solidarity the agent expects, on balance, to benefit just as much from others as s/he contributes to others. It should be stressed that this only refers to the expected, ex ante, balance between contributions and receipts. The actual, ex post, contributions and receipts will, in general, differ. Indeed, this is the feature which distinguishes two-sided solidarity from an ordinary market exchange, which one undertakes knowing that the benefits outweigh the costs. In case of two-sided solidarity one does not know beforehand whether one will end up being a net-contributor or a netreceiver. In case of one-sided solidarity a person expects beforehand, ex ante, that his/her contributions and receipts will not balance. This is clearly the case if one expects nothing in return for one’s contribution. In giving alms to a beggar or in donating money for the victims of an earthquake in a faraway country, one cannot reasonably expect to get anything in return. In this sense, one-sided solidarity is always unselfish, although one may of course feel satisfied or get a ‘warm glow’ from showing solidarity. The taxes levied on rich persons for financing social assistance, from which they will probably never benefit themselves, is also a form of one-sided solidarity. Our definition restricts solidarity acts to relations between persons. The persons towards whom one acts solidary may be either concrete persons or anonymous representatives of a group (e.g., the victims of a natural disaster). In the last case, one could also say that solidarity is a relationship between a person and a group. The term solidarity is sometimes also used for the contribution of an individual to a public good or an ideal. For instance, a financial

Ethnic Diversity and Solidarity: A Study of their Complex Relationship

5

contribution (by means of taxes) to the construction of roads or taking part in the activities of a political party or an environmental organization can also be called solidary acts. We will, however, leave these kinds of solidary acts aside and limit ourselves to solidarity between persons. In our study, solidarity is merely a descriptive term for a particular kind of behaviour. We do not address the normative question whether solidarity is desirable from a moral point of view. Although the term solidarity probably has a positive, desirable flavour for most readers, the relevance and valence of our research does not depend on this positive connotation. Since our definition of solidarity refers to a particular kind of behaviour, the underlying values, motives or attitudes do not determine whether a particular act is considered to be a solidary act. Thus, a solidary act need not be motivated by altruism or benignity, but may also arise from selfish or egotistic motives, as long as the act benefits others. This does not mean that we are not interested in the motives for solidarity and the feelings or emotions that trigger solidary behaviour. On the contrary, one of the main purposes of this research is to find out which attitudes, feelings or emotions motivate solidary behaviour. However, explaining solidary attitudes or feelings is not our goal as such, but only a way to understand solidary behaviour.

What is diversity? Diversity or heterogeneity – two terms which we will use interchangeably – can be simply defined as the negation of uniformity or homogeneity. But, while a group can be uniform or homogeneous in only one way, that is, that all group members are identical with respect to a particular characteristic, there are many ways in which a group can be diverse or heterogeneous.1 Most measures for group diversity or heterogeneity that are used in the literature start from a dichotomous relationship between group members: two arbitrary group members are either equal or different with respect to a particular characteristic (ethnicity, mother country, language, etc.). Next, an indicator is constructed to aggregate the dichotomous relationships between all possible pairs of persons. A wellknown indicator is the fractionalization index or, alternatively, one minus 1

Cf. the distinction between income equality and inequality: income equality can only be measured in one way (all incomes are equal), but there are many different ways to measure income inequality, e.g., the Gini coefficient, the Theil coefficient, the coefficient of variation, the D9/D1 decile ratio, etc..

6

Introduction

the Herfindahl index, which measures the probability that two randomly selected individuals belong to a different (ethnic) group. Although this index may be a useful indicator for (ethnic) diversity in many contexts, it has two serious limitations. First, it is based on the assumption that there is no gradation in the extent of difference between two persons. Two persons are either the same or different, but not a little different or very different. As we know from studies of social distance, the extent to which people experience or perceive a distance to people from other ethnic groups, may vary considerably depending on the particular ethnic group. For example, the (perceived) distance between a native Dutch and a fellow citizen with a Surinamese background may be smaller than the distance with a Moroccan Dutch. In studying the effects of ethnic heterogeneity, it is important to take these differences into account. Secondly, the fractionalization index is only one way to aggregate differences between individuals into an overall indicator of the heterogeneity of a group or community. It assumes, for example, that the diversity of a community consisting of many different, relatively small ethnic groups is larger than the diversity of a community that is made up of two large ethnic groups. In a literal sense, this is of course true, but when studying the societal impact of ethnic heterogeneity, two large ethnic groups that confront each other may have worse consequences for social cohesion or solidarity than when a larger number of smaller groups cohabitate. To illustrate, compare the tensions between the Flemings and the Walloons in Belgium to the more harmonious relationship between the various ethnic groups in Switzerland. It is also important to take into account, that the ethnic diversity of a community may vary considerably, depending on the level at which we measure diversity. To illustrate, an ethnically homogeneous neighbourhood may be part of an ethnically diverse city, which is situated in a largely homogeneous region in an ethnically divided country. Depending on the level at which one measures ethnic diversity, one may find quite different societal effects. The concept of ethnicity is, of course, also far from self-evident. It can refer to a common ancestry or tradition, to people with the same culture, religion, language, et cetera. We do not claim that ethnicity has an objective, essentialist meaning, but only that ethnicity is one of the dimensions along which people tend to categorize each other and one of the characteristics that people use to evaluate other people whom they do not know in person (Brubaker 2002). Thus, the fact that another person is

Ethnic Diversity and Solidarity: A Study of their Complex Relationship

7

(perceived as) Turkish, may conjure up a particular image of that person, referring, for example, to his/her competence (‘Turkish people are diligent’) or his/her sympathy (‘Turkish people are callous’). Whether this stereotype is correct – which, as we know, is usually not the case – is not relevant for our research, but only whether people act on it. Do people show less solidarity towards people from another ethnic group because they have an unfavourable image of that group? For this reason, we also study (in chapter five) the images that are depicted of various ethnic groups in the Dutch mass media.

Outline of the book This book is divided into two parts. The first part, consisting of chapters one and two, sketches a theoretical framework that can be used to study the relationship between diversity and solidarity. The second part, consisting of chapters three to six, reports on a number of empirical studies of diversity and its relation to solidarity. The book concludes with a chapter which draws some general conclusions. The first chapter sketches the outlines of an interdisciplinary theoretical framework for studying solidarity. The chapter first briefly discusses the explanation of solidary behaviour offered by five scientific disciplines: sociology, anthropology, social psychology, economics and socio-biology. Next, it combines the common and complementary elements of these disciplines, to construct an interdisciplinary framework of solidary behaviour. This framework distinguishes between other-regarding and self-regarding motives for solidary behaviour on the one hand, and between particularistic and generalized kinds of solidarity on the other hand. These two dimensions give rise to four ideal-typical kinds of solidarity: empathic solidarity, bilateral solidarity, multilateral solidarity and normative solidarity. The chapter concludes that most solidary acts in real life are a mix of these four ideal-typical kinds of solidarity. The second chapter discusses, from a theoretical point of view, which kind of effects ethnic diversity is expected to have on solidarity. Although many recent studies start from the assumption that ethnic diversity is detrimental to social goods, such as trust, social capital and solidarity, this chapter argues that, from a theoretical perspective, the relationship between ethnic diversity and solidarity is ambiguous. The chapter distinguishes between the aggregate effect of an ethnic difference between pairs of individuals, on the one hand, and the direct effect of ethnic diversity as a characteristic of a group or community, on the other hand. It concludes

8

Introduction

that solidarity is possible both between similar and between different persons and both in ethnically homogeneous and in heterogeneous groups. Very large differences or very large heterogeneity are generally detrimental to solidarity. This is simply so, because solidarity presupposes recognition of the other or some common purpose, interest or values. However, this does not imply that the more similar persons are the stronger will be their willingness to act solidary towards each other. This depends on the motives for solidary behaviour and on the possibility (and willingness) to identify with others. Chapter three investigates the relationship between ethnic diversity and a number of solidarity intentions (the willingness to contribute to the welfare state) using international comparative data. Prior studies rely either on archival data at the national level or use census data at the neighbourhood level within a single country. Both approaches have some limitations. The first approach does not allow to investigate variation in diversity within countries and the second approach misses the possibility to investigate cross national differences. This chapter brings these two approaches closer together by constructing diversity measures based on the European Social Survey (ESS). The ESS includes individual level data that allow replicating earlier measures of ethnic, linguistic, and religious diversity for thirty European countries. Furthermore, since respondents are asked to indicate in what region they live, it is possible to construct disaggregated measures at the regional level. Our analyses show that regional diversity is more strongly related to different kinds of sociality than diversity at the national level. Chapter four presents the outcomes of a so-called vignette analysis, based on a representative survey among the Dutch population, to analyse the preferences of the respondents for helping fellow citizens belonging to different ethnic groups. As resources are scarce and solidarity can never be completely unrestricted, we need to think about whom to direct our solidarity at. For this chapter, we conducted two vignette analyses. Respondents were asked to express their preferences for helping fellow citizens with varying characteristics. Because each vignette (a fictional person) is constructed from several characteristics, this methodology is less vulnerable to socially desirable answers. In the public solidarity study, we varied eight characteristics: the sex and age of the beneficiary, his/her ethnic background, working experience, type of welfare state entitlement, willingness to do volunteer work, familiarity in the neighbourhood and the number of under-age children. We identify the variables that foster (public) solidarity, and also study the interaction with the characteristics of the

Ethnic Diversity and Solidarity: A Study of their Complex Relationship

9

respondent. In the private solidarity study (about concrete helping situations directed at neighbours) the same basic design was used. Five attributes varied between the vignettes: sex, age, ethnic background, interactions in the neighbourhood and willingness to reciprocate. We found that people systematically favour certain beneficiaries over others. In particular, natives were more willing to help people with a Surinamese than with a Moroccan background. From these findings it is likely that persons are motivated by motives such as helping the needy, rewarding deservingness and favouring similarity, including ethnic similarity. The same mechanisms seemed to underlie both public and private solidarity, although a few interesting differences emerged as well. Since the attitudes of citizens towards people with a different ethnic background depend strongly on their image of various ethnic groups, chapter five analyses the images of ethnic minority groups as depicted in some of the most widely read and viewed Dutch news media. The investigation rests on the assumption that by voicing specific norms and values, and by presenting a particular image of social life in a society, the media contribute to creating and maintaining symbolic boundaries between groups, thus influencing the nature of inter-group relations. The analyses revolve around the question of diversity in the media presentation of ethnic minorities, focusing on the visibility of different ethnic groups in the Dutch news media (diversity in presence) as well as the content of their portrayal (diversity in presentation). The results show an imbalance in the media presence of various groups and some noteworthy discrepancies in the content of their presentation, with Turks enjoying the most favourable descriptions and Moroccans the most unfavourable ones. The analyses uncover important nuances in the media construction of ethnic minorities, showing how different aspects of negativity (e.g. hostility, illegality and/or incompetence) take precedence in the description of various ethnic groups. The results of the media analysis are juxtaposed with some factual information on the “performance” of various groups in Dutch society. Chapter six studies the effects of ethnic diversity of a group on the solidarity between the group members by conducting a field experiment in the multicultural Dapper neighbourhood in the city of Amsterdam, the Netherlands. In the experiment, the participants played a game in groups with varying ethnic composition. The experiment allows to discriminate between Putnam’s (2007) constrict theory, which states that ethnic diversity hurts in-group solidarity, and the alternative conflict theory, which maintains that an ethnic difference between two persons results in

10

Introduction

less solidarity. We find some evidence for the latter theory, but not for Putnam’s constrict thesis. A difference in ethnic background between two players has a strong negative impact on the gift they bestow each other in the experiment. This result provides support for the conflict theory, which asserts that members of different ethnic groups discriminate against each other. We also find an effect of the ethnic composition of a group on solidarity, but there is no straightforward linear relationship between ethnic diversity and solidary behaviour. We find no evidence that group diversity reduces in-group solidarity. On the contrary, in-group solidarity among natives is the lowest in fully native groups. Moderate ethnic diversity increases ingroup solidarity, although maximum diversity reduces it again. We suggest that these results may be explained by the fact that individuals try to avoid belonging to a minority within their group. In chapter seven we draw some general conclusions from the empirical studies in the previous chapters. We conclude that ethnicity is indeed an important factor in understanding patterns of solidarity. However, there is not a simple linear relationship between ethnic diversity and solidarity. Even though ethnic difference in itself may be a source of discrimination, one cannot conclude from this that increasing ethnic diversity will necessarily result in less solidarity. Under particular circumstances, ethnic diversity may even be beneficial for overall solidarity.

References Alesina, A., and LaFerrara, E. (2002). Who trusts others? Journal of Public Economics, 85, 207-234. Brubaker, R. (2002). Ethnicity without groups. Archives Européenes de Sociologie, 43, 163-189. de Beer, P., and Koster, F. (2009). Sticking together or falling apart? Solidarity in an era of individualization and globalization. Amsterdam: Amsterdam University Press. Ellen, I., and Turner, M. (1997). Does neighbourhood matter? Assessing recent evidence. Housing Policy Debate, 8, 833-866. Habyarimana, J., Humphreys, M., Posner, D.N., and Weinstein, J.M. (2007). Why does ethnic diversity undermine public goods provision? American Political Science Review, 101, 709-725. Haynie, D.L., Silver, E,. and Teasdale, B. (2006). Neighborhood characteristics, peer networks, and adolescent violence. Journal of Quantitative Criminology, 22, 147-169.

Ethnic Diversity and Solidarity: A Study of their Complex Relationship

11

Hooghe, M., Reeskens, T., Stolle, D. and Trappers, A. (2009) Ethnic diversity an generalized trust in Europe. A cross-national multilevel study. Comparative Political Studies, 42: 198-223. Kempen, R. van and Bolt, G. (2009). Social cohesion, social mix and urban policies in the Netherlands. Journal of Housing and the Built Environment, 24, 457-475. Kleinhans, R., Priemus, H., and Engbersen, G. (2007). Understanding social capital in recently restructured urban neighbourhoods: two case studies in Rotterdam. Urban Studies, 44, 1069-1091. Leigh, A. (2006) Trust, inequality and ethnic heterogeneity. The Economic Record, 82, 268-280. Morenoff, J.D., Sampson, R.J., and Raudenbusch, S.W. (2001). Neighborhood inequality, collective efficacy, and the spatial dynamics of homocide. Criminology, 39, 517-560. Portes, A., and Zhou, M. (1993). The new second generation. Segmented assimilation and its variants. Annals of the American Academy of Political and Social Science, 530, 74-96. Portes, A. (1995). The economic sociology of immigration. Essays on networks, ethnicity, and entrepreneurship. New York: Russell Sage Foundation. Putnam, R.D. (2007) E Pluribus Unum. Diversity and community in the twenty-first century. The 2006 Johan Skytte Prize Lecture. Scandinavian Political Studies, 30, 137-174. Wilson, W.J. (1987). The truly disadvantaged. The inner city, the underclass and public policy. Chicago and London: The University of Chicago Press. Zorlu, A., and Hartog, J. (2002). Migration and immigrants. The case of the Netherlands. In Rotte, R. and Stein, P. (eds.). Migration policy and the economy: International experiences. Munchen: Hans Seidel Stiftung.

CHAPTER ONE TOWARDS AN INTERDISCIPLINARY THEORY OF SOLIDARITY PAUL DE BEER, MAARTEN BERG, LAURENS BUIJS, FERRY KOSTER AND DOROTA LEPIANKA

Introduction Pro-social behaviour, or solidarity, for short, has puzzled social scientists for a long time. Acting for the benefit of another at a cost for oneself has often been considered as anomalous or aberrant behaviour, at odds with the self-interest of the acting person. For economists and evolutionary biologists, in particular, pro-social behaviour seemed to be inconsistent with their basic assumptions and was thus largely disregarded as an anomaly, which did not fit into their theories. For sociologists, social psychologists and anthropologists, pro-social behaviour has always been a more familiar phenomenon, but for them, too, it was often difficult to reconcile solidarity with the simultaneous existence of self-interested behaviour. At first sight, both the salience and the explanation of solidarity vary widely between the various scientific disciplines. It may thus seem an impossible task to integrate insights from these various disciplines into one overarching and encompassing interdisciplinary theory. However, on further consideration one notices remarkable similarities between the various disciplinary approaches, which are hidden from sight due to the fact that different disciplines use different terms for similar concepts. In this chapter, we attempt to draw the outline of a comprehensive theory of solidarity which combines insights from sociology, anthropology, social psychology, economics and socio-biology.

14

Chapter One

We first give a brief overview of the interpretation and explanation of solidarity in each of these five disciplines. Next, we enumerate the similarities and differences between the disciplines and draw up an interdisciplinary framework by using the common and complementary elements of the various disciplines. Before starting with the overview of the five disciplines, we first explain what we mean by pro-social behaviour or solidarity.

Solidarity Pro-social behaviour and solidarity will be used interchangeably in this volume. We restrict the definition of solidarity to a particular kind of behaviour, irrespective of the underlying attitudes, motives or preferences. Pro-social behaviour or a solidary act is defined as any act that purposefully benefits another person at a cost for the agent, without any guarantee of an equivalent return. Some examples of pro-social behaviour are alms-giving to a beggar, helping your neighbour, raising children, voluntary community work, donating money to a charity organization, paying an insurance premium1 and paying taxes. The last example shows that solidary acts, according to our definition, need not be voluntary. Moreover, solidarity includes both informal behaviour and formal behaviour, that is, solidarity behaviour that is mediated through a formal organization, such as the state or an insurance company. This definition implies that we exclude normative interpretations of solidarity and focus only on objective interpretations of actual behaviour. Therefore, we refrain from the question whether solidarity is a good thing, worthwhile to be promoted or not. The fact that we restrict our definition of solidarity to behaviour, does not mean that we are not interested in the motives for solidarity and the feelings or emotions triggering solidarity. On the contrary. One of the main purposes of this chapter is to find out which attitudes, feelings or emotions, according to the various disciplines, motivate solidary behaviour. However, explaining solidary attitudes or feelings is not our goal as such, but only a way to understand solidary behaviour.

1

A voluntary insurance may be considered a borderline case, since it can also be interpreted as an ordinary market transaction, in which the guarantee for the insured person of receiving a compensation in case of a damage or an injury is equivalent to the premium.

Towards an Interdisciplinary Theory of Solidarity

15

Sociology From the very beginning, solidarity has been a central concept in sociology, since it relates to social order and collective action, generally considered as the alpha and omega of classical sociological theory (Hechter 1998). Many scholars, both classical and contemporary, placed the concept at the core of their theories and empirical research. This has led to a wide variety of definitions, operationalisations, and applications of the term. In early sociology, the concept of solidarity came to the fore in response to the changes wrought by the development of industry, the rise of market economies and the expansion of cities. New kinds of communities emerged that were bigger, denser and more diverse than the rural communities that dominated before. Most classical theorists regarded the cohesion or solidarity of agrarian communities as unproblematic. In their view, such communities were technologically and demographically stable and were not involved in long-distance trade. Social mobility was minimal and most children were destined to follow in the (professional) footsteps of their parents (cf. Hechter 1998). Since these agrarian communities offered little scope for individual choice, the attainment of solidarity was viewed as relatively straightforward. According to Tönnies (1887), such communities (which he termed Gemeinschaften) were breeding grounds for social relations based on strong emotional, quasi-familial commitments. Émile Durkheim (1893) coined the term ‘mechanical solidarity’ for this kind of society, implying that it had a certain automatic quality. Mechanical solidarity is based on common values and internalized social norms and beliefs in a community with strong moral (often religious) norms, which tie the individuals to the society as a whole. In these communities the individual consciousness is fully integrated into the collective consciousness. This binds the individual to society and makes him or her act in accordance with the shared norms (Durkheim 1893: 60-61). Mechanical solidarity appeared to be threatened, however, by the rise of modern industry and the expansion of market economies in the Western world, starting in the late eighteenth century. Among many other effects, this transformation increased the size and scope of social networks, thereby offering individuals more options in their daily life, ranging from the choice of a marriage partner to the choice of an occupation. This resulted in a growth of individualism and the concomitant decline of the conscience collective (Simmel 1922). It was far from evident how groups, communities, and societies could maintain solidarity in the wake of this

16

Chapter One

increasing individualism. At that time, many theorists were deeply worried: “This transformation means the doom of culture itself if none of its scattered seeds remain alive.” (Tönnies 1887) Durkheim, the foremost early theorist of solidarity, shared Tönnies worries of the decline of ‘mechanical solidarity’, but believed it was replaced by a new form of solidarity. In his book The Division of Labour in Society (1893), he posed that industrial societies were held together by individuals’ mutual functional interdependence. This new form of solidarity was a result of the division of labour – a form of solidarity that he termed ‘organic’. By distinguishing organic from mechanic solidarity, Durkheim emphasized that a coherent social organization based on individualism and large communities was indeed possible. He acknowledged that the nature of solidarity changed, but it remained the foundation of any society. In a modern industrial society, Durkheim (1893) said, social cohesion depends upon the division of labour. Individual specialization leads to mutual interdependence and ensures cooperation for a final end. However, Durkheim warned that an extreme division of labour would result in anomaly and endanger the priority of the whole over the individual. He believed that the intensity of social conflicts was positively correlated with the lack of perception of solidarity. This made social tensions ultimately a political problem. For Durkheim, this showed the importance of the modern democratic state. Solidarity had to be maintained by keeping the right balance between social pressures and individual liberty, constantly enhancing the perception of solidarity (Hechter 1998). With these early attempts to define the influence of societal processes on solidarity, sociology established itself among the social sciences. Durkheim was not very explicit about the mechanism by which the division of labour translates into social solidarity. For a self-interested individual, acknowledging that one is dependent on others does not necessarily result in solidary behaviour, since it might also elicit freeriding on the efforts of others. This would mean that people take advantage of the solidary behaviour of their fellow men without contributing to the common good themselves. However, if free-riding becomes the dominant strategy, the free riders will ultimately crowd out the co-operators and solidarity will disappear in the end. According to Durkheim (1893) civil law, in particular contract law, plays an important role in preventing freeriding behaviour, but contract law can only prevent non-compliance with

Towards an Interdisciplinary Theory of Solidarity

17

private, bilateral contracts. More is needed to explain why individuals contribute to the common good if this does not directly affect the benefits they reap from the common good. In rational choice theory, it is assumed that individuals can only be made to contribute to the common good if the expected cost of free-riding is larger than the cost of contributing. Thus, there has to be a credible threat of a sanction in case of non-compliance with the social norm of helping others. This sanction can take a wide range of forms, from a simple verbal rebuke to a physical beating or ostracism. A self-interested person “will comply only if the probability of detection multiplied by the sanctions imposed given detection equals or exceeds the benefits from noncompliance” (Chai & Hechter 1998: 36, 37). However, the monitoring and punishment of defectors are also costly. These costs will have to be added to the expected contributions to the common good, for example in the form of a tax. If these costs are very high (for example, because detecting non-compliance is difficult), it may be impossible to maintain solidarity, because many group members will leave the group (Hechter 1987, Chai & Hechter 1998). An alternative way to deter free-riders is spontaneous punishment by other group members. However, since punishing is costly (partly because it requires some effort, partly because the defector may retaliate by hurting the punisher), a purely selfish person will not punish a defector (unless a non-punisher is also punished, but this results in an infinite regress). Thus, this explanation needs the additional assumption that people are willing to unselfishly punish a defector (so-called altruistic punishment, cf. Boyd et al. 2003), which does not fit well within the rational choice framework which starts from the assumption of self-interest (Coleman 1990: 31-32).

Anthropology In anthropology the classical example of solidarity is the reciprocal exchange of gifts. The act of gift-giving binds groups and individuals; it creates and reinforces a network of rights and obligations that generates and sustains social cohesion. Exchange of gifts in archaic societies is far more than an economic transaction. As noted by Lévi-Strauss, “[g]oods are not only economic commodities, but vehicles and instruments for realities of another order, such as power, influence, sympathy, status and emotion” (Lévi-Strauss

18

Chapter One

1967 [1949]: 54). The goods that are exchanged embody not only – or not even primarily – material value but also social meanings (Malinowski 1985 [1922]). These goods are by no means limited to material objects; they include borrowing things, getting help, accepting an obligation, or assuming responsibilities for another. Gifts are common and ever present in archaic societies. They do not constitute an exception for special occasions (such as our birthday gifts), but are rather the norm. Lévi-Strauss (1969:54) talks about a “passion for the gift, accompanied by the ritual obligation on the recipient to accept and to give”. A gift is rarely purely voluntary. All anthropologists stress the compulsory nature of reciprocity: gift-giving is generally considered to be a social obligation, just as the duty to repay the gift (Malinowski 1985 [1922]:182). Accepting a gift is obligatory, too: a gift cannot be properly refused. The exchange of gifts is not the result of economic self-interest (Sahlins 1974:160). Most often, the exchange of gifts brings no tangible results or profits, and after the exchange, the partners are no richer than they were before. In fact, from an economic point of view, most transactions do not make any sense, as usually neither partner acquires any real material benefit. Actually, it is the exchange itself that matters and not the object of exchange! Anthropologists distinguish between various types of gifts, primarily depending on the extent of reciprocity, although there is no agreement on the number and kind of categories. Malinowski (1985 [1922]) distinguished seven types of gifts. At one extreme, there is the pure gift, defined as “an act, in which an individual gives an object or renders a service without expecting or getting any return”, which according to him is exceptionally rare. Mauss rejected the idea of ‘free gift’ altogether. The idea of a ‘free gift’ is to him a contradiction, a misunderstanding. Every gift triggers an obligatory counter-gift and perpetuates a (dominant) system of reciprocity. At the other extreme is ordinary trade (e.g. barter), in which mutual advantage to the trading partners is clearly present. In between are various forms of customary gifts that are partially or conditionally returned. The value of the counter gift is (almost) never strictly equivalent to the value of

Towards an Interdisciplinary Theory of Solidarity

19

the original gift and the economic value of the counter gift might be even symbolic. Sahlins (1974) makes an interesting distinction between reciprocity, i.e. “vice-versa movements between two parties”, and pooling or redistribution, i.e. “centralized movements: collection from members of a group … and redivision within this group”. An important difference is that “pooling is socially a within relation, the collective action of a group … [while] reciprocity is a between relation, the action and reaction of two parties” (Sahlins 1974:188). Within the category of reciprocal exchanges, Sahlins distinguishes between generalized reciprocity, balanced reciprocity and negative reciprocity (Sahlins 1974:191-196). Generalized reciprocity is similar to Malinowski’s pure gift, but it seems broader and is not necessarily restricted to the closest kin. It is help given freely, altruistically, out of disinterested concern and without any (open) stipulation of return. It includes ‘sharing’, ‘hospitality’, ‘free gift’, ‘help’, ‘generosity’, but also ‘noblesse oblige’. If there is an obligation to reciprocate, it is vague, meaning that the counter-obligation “is not stipulated by time, quantity or quality: the expression of reciprocity is indefinite”. Generalized reciprocity is a “sustained one-way flow”. As stated by Sahlins, “[f]ailure to reciprocate does not cause the giver of stuff to stop giving: the goods move one way, in favour of have-not, for a very long time” (Sahlins 1974:194). Balanced reciprocity refers to direct mutual exchange. Reciprocation always takes place within a finite and narrow period of time in the form of “the customary equivalent of the thing received” (Sahlins 1974:194). The economic and social interests of the parties are central and clearly stipulated, and the failure to reciprocate within the given time may cause a disruption of the relation between the parties involved (Sahlins 1974:195). Transactions have usually a utilitarian purpose as well, but the ‘moral’ purpose of “renunciation of hostile intent or of indifference in favour of mutuality” remains central (ibid., p. 220). Sahlins also distinguishes negative reciprocity, which is defined as “an attempt to get something for nothing with impunity” (Sahlins 1974:195). However, this form of transaction is exceptionally rare and might be charged with negative social sanctions. Central to anthropological theory is, that gift-exchange is not considered to perform an economic function, but is seen primarily as a social act to

20

Chapter One

acquire and maintain social status, to create and consolidate good relations or to comply with social norms. One of the main motives for gift giving, is according to Malinowski, the “fundamental human impulse to display, to share, to bestow” (Malinowski 1985 [1922]:175). This fundamental impulse may be fed by vanity, but may also serve the purpose of acquiring and maintaining prestige and power (ibid., p. 175). The exchange of gifts also helps in creating, sustaining and strengthening social ties. According to Douglass, this purpose lies at the heart of gift giving: “A gift that does nothing to enhance solidarity is a contradiction” (Douglass 2004: x). To Lévi-Strauss, “reciprocity is … the most immediate from of integrating the opposition between self and others” and the gift, which “makes … individuals into partners”, constitutes the symbolic agent of this integration (Lévi-Strauss 1969 [1949]:84). According to Sahlins, gift-giving “underwrites or initiates social relations” (Sahlins 1974: 186). Malinowski states that the purpose of ceremonial repayments is “to thicken the social ties from which arise the obligations” (ibid., p.182). Malinowski repeatedly stresses the obligatory nature of gifts. Gifts are almost always given to fulfil social obligations “and with a great deal of formal punctillo” (Malinowski 1985 [1922]:174). The obligation to (give, accept and) reciprocate is frequently enforced by social sanctions. However, as noted by Sahlin, “[t]he norms are relative and situational rather than absolute and universal. A given act… is not so much in itself good or bad – it depends on who the “Alter” is” (Sahlins 1974:199). It is important to note, that the obligatory nature of the gift-exchange rules out pure altruism. Mauss notes a peculiar integration of interest and disinterest, generosity and egoism, altruism and selfishness of the gift exchange.

Social psychology The term ‘solidarity’ is not frequently used in the social psychological literature. The related terms in social psychology are ‘pro-social behaviour’ (any act performed with the goal of benefiting another person) and ‘altruistic behaviour’ (pro-social behaviour that involves a cost to the helper). Social psychology is not based on one theoretical perspective but combines insights from various perspectives, which often originate from other disciplines. For example, evolutionary psychology stems from

Towards an Interdisciplinary Theory of Solidarity

21

evolution theory in biology and will here be discussed under the heading of socio-biology. Social exchange theory borrows from economics and rational choice theory the concept of social exchange. It stresses the role of self-interest in pro-social behaviour by assuming that people aim to maximize their own utility or happiness. The (psychological) social exchange theory is broader than its economic equivalent, in the sense that it is less focused on monetary rewards. Relationships between people are also explained in terms of costs and benefits. Helping others can be beneficial if others return the favour (now or in the future) or when a confrontation with misery is stressful for the helper (e.g. Eisenberg and Fabes, 1991). Moreover, by helping others we gain psychological rewards, such as social approval and increased feelings of self-worth. However, if the costs of helping are large (for example when it is dangerous, embarrassing or timeconsuming), people are less inclined to help (e.g. Dovidio et al., 1991). Another motive for solidary behaviour can be empathy, which social psychology borrows from psychology and socio-biology. Batson (1991) argues that people sometimes help other people purely out of the goodness of their hearts, which presupposes the ability to feel empathy. Toi and Batson (1982) manipulated the level of empathy and found that when empathy is low, there is a big difference in willingness to help, depending on whether or not helping is costly. When empathy is high, however, willingness to help is almost independent of the cost of helping. The specific contribution of social psychology to explaining solidary behaviour is its focus on the conditions and circumstances that promote solidarity. For example, the willingness to help others depends on someone’s mood. Isen and Levine (1972) manipulated mood (‘mood induction’) by leaving small amounts of money on a telephone cell. People who had just found the money seemed to be much more willing to help others than people who did not find any money. When people feel good, they do good. When they perform well on a test, receive a gift, have happy thoughts or listen to pleasant music, they contribute more to charity, are more willing to donate blood or to help co-workers on the job (e.g. Carlson, Charlin and Miller, 1988; Salovey, Meyer and Rosenhan, 1991). Feeling good makes us look on the bright side of life. Moreover, helping others is an excellent way of prolonging our good mood (e.g. Williamson and Clark, 1989).

22

Chapter One

But also negative emotional states might foster solidary behaviour. An example is feeling guilty (e.g. Estrada-Hollenbeck and Heatherton, 1998). Churchgoers were more likely to donate money to charity before attending confession than afterwards, presumably because confessing to a priest reduced their guilt. But also feeling sad can boost helping behaviour, as sad people are motivated to engage in activities that make them feel better (Wegener and Petty, 1994). This is sometimes called the ‘negative-state relief hypothesis’ (e.g. Cialdini et al, 1987). According to this theory, the helping behaviour does not have to be related to the source of our sadness. Although not directly related to solidary behaviour, an important strand of social psychology that is relevant in studying solidarity is social identity theory. According to this theory, people try to build a positive self-image through social categorization (thinking in terms of groups), social identification (identifying with certain ‘in-groups’) and social comparison (differentiating their in-group positively on particular dimensions from relevant out-groups). Each individual has a repertoire of (social and personal) identities open to her/him, each identity informing the individual of who he/she is and what this identity entails. Which of these many identities is most salient for an individual at any time varies according to the social context. Where personal identity is salient, the individual will relate to others in an interpersonal manner, depending on their character traits and any personal relationship existing between the individuals. However, under certain conditions a group identity might take precedence. What is a relevant out-group? According to the concept of distinctiveness, a relevant out-group is not too different from the corresponding in-group. The most rivalry exists between quite similar football-clubs, such as Ajax and Feyenoord, or between two Ivy-league universities, or between Protestantism and Catholicism (rather than Buddhism). Important for social identity theory is that no interaction is required to speak of ‘a group’. When a single woman is in a meeting with ten other men, her social identity (female) is salient, despite the fact that there is no (physical) contact with other women. Self-esteem is an important motive for identification. People strive for positive differentiation of their in-group. Self-categorization theory is an elaboration of social identity theory (e.g. Turner, Oakes, Haslam and McGarty, 1994). According to this theory, identity is a flexible and multi-dimensional concept. What part of your identity is active is dependent on the social context.

Towards an Interdisciplinary Theory of Solidarity

23

Economics The mainstream economic perspective on human behaviour, neoclassical economics, starts from the assumption of the homo economicus, i.e. an individual who rationally seeks her/his self-interest. That is, he/she has only self-regarding preferences and is not interested in the well-being or ‘utility’ of other persons. Altruism, in the sense of wilfully acting for the benefit of others, is, thus, impossible. This basic assumption seems to be at odds with solidary or pro-social behaviour. Nevertheless, even purely selfinterested individuals may act solidary if this is in their (long-term) selfinterest. There may be three self-interested motives for solidarity in economics. A first motive is that a risk-averse individual2 wants to insure her/himself against the risk of a harmful future event. This self-interested individual prefers paying a premium for an insurance policy which guarantees compensation in case a harmful event occurs, even if it costs (slightly) more than the expected (probabilistic) loss. The premium paid by the insured person is used to compensate other policyholders who incur a loss. If the harmful event never occurs, the insured person will have contributed to the benefit of others without having received anything in return. In this sense, paying an insurance premium is a solidary act, since there is no guarantee that it will be refunded. However, since the insurance policy stipulates the legal right to compensation, it is usually considered to be equivalent to other market transactions in which no solidarity is involved. A second motive for pro-social behaviour is that one expects to benefit from a future return of the beneficiary (direct reciprocity). If the beneficiary has no legal obligation to pay back the favour bestowed on her/him, the benefactor apparently has no incentive to favour the beneficiary. However, if both persons interact with each other repeatedly, pro-social behaviour may be optimal for both parties. An optimal strategy may be to help the other person as long as he/she helps you, but to retaliate in the next meeting as soon as the other party defects (the so-called tit-fortat-strategy, cf. Axelrod 1984). Both parties have then an interest in continuing their co-operative behaviour as long as it is uncertain when the sequence of interactions will end. 2

A risk-averse person prefers a fixed amount of money to a variable amount with the same expected monetary value. Thus a sum of € 1,000 is preferred to a fiftyfifty chance of € 500 and € 1,500, respectively.

24

Chapter One

A third self-interested motive for solidary behaviour is that it elicits admiration and praise from others, which the benefactor values positively. Thus, a feeling of pride may compensate for the costs of the solidary act. However, this explanation of self-interested solidarity raises the question why people who are purely self-interested would praise the solidary behaviour of a benefactor from which they do not benefit themselves. These three motives explain why a self-interested individual may act solidary voluntarily. However, economic theory can also explain why people may support mandatory solidarity. This can be the case if a voluntary, private insurance scheme is not feasible or is much more costly than a mandatory scheme. Various market failures may render compulsory insurance optimal from the point of view of self-interested individuals (e.g. Barr, 2004). These market failures include interdependent risks, such as inflation and unemployment, adverse selection, which may occur if there is asymmetric information about risks between the individual persons and the insurer, and economies of scale, which make a single public scheme more efficient and thus cheaper than a multitude of private schemes. It is important to realize that even if an individual prefers a compulsory public insurance scheme, he/she still has a strong incentive to free ride and evade paying the mandatory contribution if that does not nullify her/his entitlements to a benefit. Thus, a reasonable probability of detection and a credible sanction are needed to sustain such a compulsory scheme. The three motives that have been discussed so far exhaust the possible motives of a purely self-interested person for acting solidary. However, if one stretches the basic assumption of neoclassical theory somewhat, one may also include other-regarding preferences in the utility function of an individual. Actually, this means that insights from other disciplines, such as psychology and sociology, are incorporated into the economic framework. Three kinds of other-regarding preferences may straightforwardly result in solidary behaviour. First, someone may enjoy gift-giving. More formally, this means that, ceteris paribus, giving something to someone else (not necessarily an arbitrary person) raises your utility. Second, one may care about the well-being of others. This is usually called sympathy or empathy (Smith 1759). As a consequence, your utility

Towards an Interdisciplinary Theory of Solidarity

25

(or well-being, happiness, etc.) increases if someone else’s utility increases. If you care about the well-being of someone else, you may be willing to act in favour of that person if it raises her/his well-being. A rational person will weigh the increase of her/his own well-being as a consequence of the higher well-being of the beneficiary against the reduction of her/his well-being due to the cost of acting for the benefit of the other. In this case, contrary to the joy of giving, it does not matter to your wellbeing whether the well-being of the beneficiary is promoted by your own act or by someone else’s solidary behaviour. This implies that sympathy may give rise to the problem of free-riding. Even though you care about the well-being of another person, you would rather have someone else help that person instead of you. To sustain solidarity based on sympathy, especially if it concerns the rather weak sympathy with ‘distant’ persons, it might, once again, be necessary to oblige individuals to contribute to it by way of compulsory solidarity. A third other-regarding motive for solidarity is commitment (Sen 1985). This means that a person feels obliged to act in the interest of someone else, even if it does not raise her/his utility or well-being. This feeling of obligation may result from the internalization of a social norm, which may have been the result of the socialization process during upbringing. Contrary to the two preceding other-regarding motives, commitment cannot be traded off against the utility of self-interested behaviour, since commitment is not directly related to your utility. The internalized social norm is actually a second (higher) order preference, which is superior to first order preferences (compare your preference for a fair treatment of your child to your preference for marmalade). Thus, the analysis of solidary behaviour because of commitment cannot be easily incorporated into the usual neoclassical economic framework.

Socio-biology Studying solidarity using insights from socio-biology implies trying to understand the social behaviour of different species from the perspective of evolutionary theory. An application of this perspective is based on the following points of departure. First, it means that the main mechanisms explaining social behaviour involve a reference to evolutionary processes such as natural selection, reproduction, fitness, and adaptation. One of the basic premises of

26

Chapter One

evolutionary approaches to social behaviour is that focusing on these processes helps to uncover the ultimate causes of behaviour rather than more superfluous explanations based on proximate causes (Diamond 1997; Henrich and Henrich 2007). Secondly, these social behaviours can be found in different species. As such, to explain a certain kind of behaviour it does not matter whether it is performed by a germ, an ant or a human being (Kropotkin 1902 [2009]; Axelrod and Hamilton 1981; Ridley 1997). In that sense, this perspective places stronger emphasis on the similarity between species than on their differences. As a result, little attention is paid to the question whether certain kinds of behaviour are expressed consciously or not. Therefore, the question whether the consequences of a particular behaviour are rationally considered or consciously reflected on, has a less prominent place in this field of inquiry compared to investigations in psychology, sociology, and economics. In socio-biology, the term solidarity is rarely used. Most research focuses on altruistic behaviour, while far less attention is paid to cooperation between individuals. Within socio-biology, altruism has a specific meaning, namely behaviour decreasing the fitness of the individual who behaves altruistically, while increasing the fitness of another individual (e.g. Hamilton 1964; Trivers 1971; Wilson 1975; Bell 2008). This definition emphasizes costs of altruism for the giver and the benefits for the receiver. One of the central questions that socio-biology tries to answer is how such costly behaviour can have evolved through natural selection (e.g. Komter 2010). Understanding altruistic solidarity is much more difficult than cooperative solidarity, since the latter also yields gains to the individual who acts solidary. As in other scientific disciplines, this puzzle is translated into game theoretic notions, such as the prisoner’s dilemma, in which those acting selfish can take advantage of the benevolence of altruists. In the socio-biological literature several mechanisms are proposed to explain how altruism may evolve (Nowak 2006). A central explanation of altruism is based on the idea that natural selection can favour cooperation between genetic relatives. If the giver and the receiver are genetically close to each other, organisms carrying an altruistic gene that increases the fitness of their relatives, will also foster the reproduction of that gene, since these relatives carry a copy of that gene. This mechanism is referred to as kin selection (Fisher 1930; Haldane

Towards an Interdisciplinary Theory of Solidarity

27

1955; Hamilton 1964; Smith 1964). It explains why there is more help between relatives than between non-relatives, thus maximizing the inclusive fitness of the organism (Hamilton 1964). However, kin selection is not able to explain altruism within groups consisting of non-relatives. The occurrence of altruism in groups of unrelated organisms can be explained with a reference to reciprocal altruism (Trivers 1971). If the organisms interact more than once, there is the possibility that an altruistic act from one is reciprocated with an altruistic act later on by the other. Clearly, for reciprocal altruism to work, organisms should recognize the other to be able to return the favour. By following a tit-for-tat strategy, adjusting their behaviour to the behaviour of others, organisms can yield higher outcomes than following an egoistic strategy (Axelrod and Hamilton 1981). Direct reciprocity is based on symmetric exchange relationships between individual species. In contrast, indirect reciprocity involves asymmetric relationships. This occurs when an organism helps another, without the possibility (or perhaps the need) for reciprocation. Reputation is the main mechanism explaining altruism in such situations. If someone having a good reputation is likely to receive more help from others, helping others is beneficial to oneself (Nowak and Sigmund 1998). Reputational effects involve a complex system of calculation to keep track of who helps others. This is supported by the development of language to enable information sharing and the evolution of social norms (Nowak 2006). Another possible explanation of altruism focuses on the mechanism of group selection (Wynne-Edwards 1962; 1986). The basic idea of group selection is that groups that are better able to adapt to their environment have a greater chance of survival and that nature therefore benefits groups with certain traits, rather than individuals. If altruism strengthens the chance of survival of a group, a group with a large share of altruists among its members is expected to have a higher chance of survival than a group with a smaller share or no altruists. The idea of group selection was accepted for quite some time until it was heavily criticized by biologists like Smith (1964) and Williams (1966), who argued that group selection is not a strong enough evolutionary force, because selection only takes place at the level of individuals or genes (Dawkins 1976). Clearly, a weak point of the theory of group selection concerns the possibility that individuals free ride on the altruism of others within the group. Whereas a group with a large share of altruists may have

28

Chapter One

a higher chance of survival, it will also benefit egoistic individuals who do not contribute to the survival of the group but take advantage of the altruists. Ultimately, if the egoists have a higher survival probability than the altruists, the proportion of egoists within the group will increase and the probability of survival of the group drops. Only recently has the possibility of group selection regained interest. The most recent development in this area is the development of multilevel selection theory, which argues that selection takes place at different levels – both the genetic level and the group level – at the same time (Wilson and Sober 1994; Wilson and Wilson 2008). This latter theoretical development also fuelled ideas about the co-evolution of genes and cultures with an emphasis on social learning and norms (Gintis 2003; Boyd and Richerdson 2009). Socio-biological perspectives on altruistic and cooperative solidarity are inherently functional, as they explain the actions of individuals by referring to the impact on their own fitness and the fitness of others without reference to the underlying motivations. In an attempt to explore the motivational basis of altruism and to go beyond models that try to understand altruism by focusing on the benefits that it may have for individuals, de Waal (2008) suggests that empathy is among the most important impulses explaining altruistic acts. These impulses result from emotional identification and social bonding. If others experience pain, sadness or distress, the empathy mechanism leads to an emotional response and comforting attempts.

Overview: two dimensions of solidarity Despite the widely varying premises and assumptions of the various theoretical perspectives on solidarity, there is a common tendency within most disciplines to distinguish different kinds of solidarity along two dimensions. The first dimension refers to the prime motive for solidary behaviour, in particular whether a solidary act serves the actor’s selfinterest (we will call this self-regarding solidarity) or whether it is meant to further the interests of someone else (other-regarding solidarity). The second dimension refers to the relationship between the donor and the beneficiary, in particular whether there is a personal tie between the two (personal or particularistic solidarity) or the benefactor and the beneficiary do not know each other personally (communal or generalized solidarity). We will discuss these four kinds of solidarity in relationship to the various disciplines below.

Towards an Interdisciplinary Theory of Solidarity

29

Self-regarding solidarity The first common approach to solidarity in the various disciplines assumes that it is in the interest of the individual her/himself to act solidary, since this contributes, mostly indirectly, to her/his well-being, her/his survival or the transfer of her/his endowments to others. This approach does not need other-regarding preferences or motives, such as altruism, empathy, affection or charity, to explain solidary behaviour. It suffices that individuals are aware of their mutual dependence on others or that they are genetically programmed to show solidarity. This perspective is most prominent in economics and in socio-biology. In the standard version of neoclassical economic theory the fully rational homo economicus will always pursue her/his self-interest. Consequently, solidarity is only possible if it is in the self-interest of the agent, since s/he expects at least an equivalent return, i.e. the value of the expected return must be sufficient to compensate the agent for the cost of the solidary act. This equivalence can be achieved in case of an insurance (if the agent is risk averse), in case one expects the beneficiary to return the favour equivalently in the future (direct reciprocity) and in case a solidary act increases one’s reputation which fosters the likelihood of support by others in the future (indirect reciprocity). In the socio-biological approach solidary behaviour can only be explained if it yields some evolutionary advantage. This means that the genetic traits that enhance solidary behaviour can be successfully transferred to the next generation by increasing their likelihood of survival and their procreation. From this reasoning, pro-social behaviour towards genetically related individuals (kin selection), such as solidarity of parents towards their children and solidarity between siblings, can be easily explained. Solidarity towards non-kin is only possible if the beneficiary is likely to do something in return (direct reciprocity) or if showing solidarity enhances one’s reputation, which will trigger support by others (indirect reciprocity). These latter two motives are identical to those in neoclassical theory. Within social psychology, the branch of evolutionary psychology is closely related to socio-biology. It, too, focuses on kin selection and reciprocity as the main motives for solidary behaviour. The social exchange theory is quite similar to the economic approach in focusing on the weighing of costs and benefits. Solidary behaviour presupposes that the (long-term) benefits outweigh the costs. These benefits can either be an equivalent return by the beneficiary (direct reciprocity), or a psychological benefit of

30

Chapter One

feeling proud for helping another or being esteemed by others, which may contribute to pleasant feelings of self-esteem and self-respect. Although in anthropology gift-giving is never purely self-interested, reciprocity is generally required, implying that the costs of the solidary act are partly or fully repaid at a later moment (balanced reciprocity, in Sahlins’ terms). Consequently, the net cost for the benefactor is generally modest. Moreover, gift-giving can benefit the benefactor in indirect ways, for example by increasing her/his status. Within sociology the rational choice approach comes closest to the preceding perspectives. It, too, assumes that individuals make a rational trade-off between the costs and the benefits of various options. Costly prosocial behaviour, thus, assumes that there are positive revenues. These may be due to a reciprocal service by the beneficiary (direct reciprocity) or to praise by the community for acting in conformity with a social norm. Violating a social norm of pro-social behaviour – by refusing to help someone who one should help according to the dominant norm – may be punished by various sanctions, ranging from a mere rebuke to ostracism. In the functionalist branch of sociology, which builds on Durkheim, (organic) solidarity is explained by the mutual interdependence of citizens in a modern society. This interdependence is caused by functional (occupational) specialization resulting in a strong division of labour, rendering individuals dependent on each other.

Other-regarding solidarity Opposite to the self-regarding kinds of pro-social behaviour, discussed above, most disciplines also distinguish a kind of solidarity based on motives which put the interests of others – or of the group as a whole – forefront. This perspective is most prominent in sociology, which almost by definition emphasizes group mechanisms instead of individual considerations. Durkheim’s classical explanation of mechanical solidarity in traditional rural, homogeneous communities, traces solidary behaviour back to shared values and social norms which prescribe pro-social behaviour that enhances the survival of the group. If these norms of solidarity are internalized during the upbringing of children (primary socialization) and are reinforced by peers in later phases of life (secondary socialization),

Towards an Interdisciplinary Theory of Solidarity

31

people will generally comply with these norms, even if there is little social control and no threat of sanctioning in case of violation of norms. In anthropology, gift-giving and repaying the gift are generally considered to be a social obligation. Solidary behaviour is thus tantamount to complying with a social norm, which may or may not be internalized. A purely altruistic gift is considered to be rare, but the obligation to repay a gift is often vague. According to Sahlins, generalized reciprocity often boils down to a “one-way flow”. In social psychology, the innate capacity of humans for empathy may explain why they act in the interest of others and, in particular, help others who are in need. Witnessing the suffering of someone else evokes similar feelings of distress with the observer, who is thus stimulated to relieve the suffering of the other person. However, social psychology cannot explain why humans have a tendency to feel empathy with others. Social identity theory states that people are willing to help members of their in-group, simply because they identify with this group. Even though this help may not be reciprocated, it may enhance his or her self-image and self-esteem. Also in socio-biological theory empathy may explain pro-social behaviour towards non-related individuals who are not expected to reciprocate solidarity. An abundance of empirical evidence supports the claim that humans as well as higher animal species (such as primates) have the capacity to feel empathy – sometimes even towards animals from other species. However, from an evolutionary perspective, it is still a puzzle how to explain the survival of empathic genes. Individual selection processes seem to weed out empathic genes in the long run (although de Waal claims that they may be a by-product of empathy for an animal’s offspring). A process of group selection may better explain the origin and survival of empathy, but it is still a matter of controversy among evolutionary theorists whether group selection processes really exist, because they seem to be at odds with the dominant mechanism of individual (or gene) selection. The assumption of homo economicus in mainstream neoclassical economic theory effectively excludes the possibility of other-regarding behaviour. One has to stretch the assumptions underlying the homo economicus or, actually, abandon the concept of this rational, selfish individual altogether to create room for an other-regarding economic subject. One option, which

32

Chapter One

does not depart too much from the standard assumptions, is to include the well-being of others in an individual’s utility function. This may be justified by the fact that people enjoy giving to others (‘warm glow’) or are directly affected by the well-being of others. Actually, this latter assumption is identical to introducing empathy in the economic model. However, the existence of this tendency cannot be explained within the economic discipline.

Particularistic versus generalized solidarity The second dimension along which we can distinguish different mechanisms that foster solidarity refers to the relationship between the benefactor and the beneficiary. At one extreme, the target of a solidary act (the beneficiary) is a particular person (or particular persons) with whom one has a personal relationship, at the other extreme it is an anonymous person, with whom one has no personal relationship and whose identity is not relevant as such. We will call the former kind of solidarity particularistic solidarity and the latter generalized solidarity. Particularistic solidarity is central to the explanation of solidary behaviour in social psychology and economics. In social psychology, it refers to mechanisms such as empathy or affection that can motivate a person to act in the interest of another person. In this case, it matters whom one helps or supports, since one has to experience a personal tie with the beneficiary. In economics, the personal relationship with the other person matters, too. Since the homo economicus will only support another person if he expects an equivalent return in the future, it is essential that the two persons interact with each other repeatedly and trust each other. In case of generalized solidarity the benefactor does not necessarily know the beneficiary of a solidary act personally. It may be an anonymous person, of whom one does not know more than that s/he is in need and is a member of your own community or group, who therefore deserves support. This kind of solidary behaviour is usually strongly influenced by values, social norms, customs or social pressures. Both sociology and anthropology emphasize these mechanisms as a cause of solidary behaviour. Even if the benefactor expects some return for his or her solidary behaviour, it need not be repaid by the same person that benefited from the solidary act. Thus, one may be willing to help an anonymous person, expecting to be helped oneself by another anonymous person in the future in case one should be in need.

Towards an Interdisciplinary Theory of Solidarity

33

Four ideal-types of solidary behaviour Combining the two dimensions of solidarity that we distinguished in the previous section, other-regarding versus self-regarding solidarity and particularistic versus generalized solidarity, we can define four idealtypical kinds of solidarity (see Figure 1).3 We have chosen names for these ideal-types which are not directly derived from one particular discipline, to emphasize that the distinction between the ideal-types does not coincide with disciplinary borders. Figure 1. Two dimensions and four ideal-typical kinds of solidarity particularistic solidarity

empathic solidarity otherregarding solidarity

bilateral solidarity

hybrid solidarity normative solidarity

selfregarding solidarity

multilateral solidarity

generalized solidarity

Although, up till now, we treated the different kinds of solidarity in sociology, anthropology, social psychology, socio-biology and economics as mutually exclusive, polar cases, it is actually more accurate to consider them as the ends of a continuum ranging from pure altruism to pure selfinterest on one dimension and from purely particularistic to purely generalized solidarity on the other dimension.

3

A similar fourfold distinction is made by van Oorschot & Komter (1998).

34

Chapter One

Empathic solidarity is a form of other-regarding solidarity between two persons that is driven by a strong sense of connectedness (affection and empathy), which may cause feelings of pain or pleasure. Altruism might be the clearest example of this. Empathic solidarity will occur if the benefactor is emotionally affected by the experience of someone else who is suffering or who needs support. Bilateral solidarity is also a form of solidarity between two persons, which is, however, based on self-interest. Whereas empathic solidarity is usually one-sided (unilateral), bilateral solidarity is, by definition, two-sided and manifests itself as a bilateral exchange (quid pro quo), in which the (initial) benefactor expects a return from the beneficiary which is at least of equal value to the benefactor. Thus, reciprocity is essential for bilateral solidarity. Purely bilateral solidarity will only occur if the value of the expected future return is larger than the costs of the solidary act. Both empathic solidarity and bilateral solidarity originate from the personal relationship between two individuals and are not (strongly) influenced by societal forces, although the way in which the two persons interact may of course be shaped by a particular culture or tradition. Normative solidarity, on the contrary, is strongly embedded in the culture (norms and values) of a particular group or community. Acting solidary is dictated by specific social norms, which are usually internalized during upbringing. As a consequence, most people really believe in the norms and comply with them more or less voluntarily. The only reward they get from complying with the norm of solidarity is a feeling of pride or self-esteem (and shame or guilt in case they violate the norm). Normative solidarity may also be the result of more personal norms or values, which may not coincide with the prevailing norms and values in a person’s community. For example, one may endorse particular values regarding a fair or just society or the desired level of income inequality, which urge one to show solidarity in accordance with these values. Thus, a potential benefactor will act out of normative solidarity if he/she has a strong belief or conviction that acting solidary is the (morally) right thing to do. Consequently, normative solidarity predominantly goes one way: from the benefactor to a largely anonymous beneficiary (or beneficiaries). Multilateral solidarity differs from normative solidarity in that the dominant social norm is not internalized. Individuals realize that they are

Towards an Interdisciplinary Theory of Solidarity

35

mutually dependent on other group members, but they are also aware of the advantage of free-riding on the solidary behaviour of others. Therefore, monitoring and sanctioning are indispensable to maintain the norm of solidarity. Fear of loss of reputation, which may result in less support by others in the future, may also be an effective way to foster norm compliance. Multilateral solidary behaviour will occur if the benefactor values the costs of non-compliance with a social norm, such as a verbal rebuke or punishment, higher than the costs of compliance, i.e. acting solidary. Figure 1 may suggest that there exist different, clearly distinguishable kinds of solidarity, which are prevalent under different circumstances or among different individuals. For example, in market relationships bilateral solidarity would be expected to dominate, while in family relations we would primarily expect empathic solidarity. However, this is not what we claim. On the contrary, we postulate that actual solidary behaviour is mostly a mixture of these four ideal-types. Thus, a particular solidary act may be partly bilateral, partly empathic, partly normative and partly multilateral at the same time. As an example, consider the case that you witness your neighbour, carrying a heavy bag of groceries, stumble and fall on the ground. If you assist her to get up and pick up the groceries, this act of solidarity may be motivated by the fact that you feel sorry for her (empathic solidarity), by the expectation that she will help you in the future in a similar case (bilateral reciprocity), by the conviction that you should help someone who is in need (normative solidarity), and by the social pressure that you feel from passers-by (multilateral solidarity), all at the same time. To illustrate, in Figure 2 we have ordered some examples of solidary behaviour along the axes of other-regarding versus self-regarding and particularistic versus generalized solidarity. It should be stressed that the exact position of the various solidary acts on the two dimensions is somewhat arbitrary and may vary between persons and circumstances. For example, lending money to a friend is placed on the self-regarding side of the horizontal dimension, because it is supposed that the donor will expect her/his friend to repay the loan on a later moment when s/he is able to. However, it is also conceivable that you would not expect your friend to ever pay back the loan. In that case, the solidary act should be placed more on the other-regarding side. In both cases, it is probably essential that it is a particular friend who requests a loan, and therefore it is placed at the particularistic side of the vertical dimension in Figure 2.

Chapter One

36

generalized solidarity

particularistic solidarity

Figure 2. Examples of solidary acts ordered along two dimensions empathic solidarity

raising children

donating a kidney

babysitting for grandchildren

lending money to a friend

bilateral solidarity

neighbourly help

alms-giving to beggar gift-giving voluntary community work

helping a driver with car trouble

donation to charity org. development aid

unemployment insurance social assistance

normative solidarity other-regarding solidarity

multilateral solidarity self-regarding solidarity

Voluntary community work is placed on the other-regarding side of Figure 2, because it is assumed that the volunteer does not expect any concrete return from her/his volunteering. Nevertheless, some people may volunteer because they actually do expect some reward, such as increasing employment opportunities. In that case, voluntary work should be shifted to the self-regarding side of the horizontal dimension. The extent of particularism may also vary in case of voluntary work. If you help elderly in a local nursing home, this is probably motivated by involvement with the people in your neighbourhood. If you join a non-governmental development organization, this is more likely motivated by general feelings of human solidarity. In real life, the relative weight and the mixture of the four ideal-typical kinds of solidarity may vary between persons as well as between circumstances. If there are no bystanders, external social pressure cannot play a role. If you do not know the woman who stumbles, you will probably not expect any future return benefit from her. In case of solidarity with anonymous persons, apparently only the normative motives

Towards an Interdisciplinary Theory of Solidarity

37

play a role. But even then, particularistic motives can be important, for example if you donate money to help the victims of a natural disaster after seeing a TV report showing some individual victims. We believe that an act of solidarity which can be characterized by only one of the ideal-types in Figure 1 is exceptional. Moreover, there are probably no persons who cannot be motivated by one of the four idealtypical motivations in Figure 1, except in case of rare psychological disorders, which cause people to have no capacity for empathy or never experience feelings of compassion, such as extreme cases of autism or other kinds of serious psychopathology. Thus, our synthetic proposition regarding solidary behaviour reads, that any solidary act is motivated by a combination of empathy, bilateral reciprocity, internalized norms and values, and social pressure.

General conditions for solidarity In the previous sections we focussed on the motives that people may have to act solidary towards particular people. In doing so, we took for granted that there was some reason to act solidary in the first place. As a conclusion to this chapter, we address three more general conditions that have to be met for solidarity to occur. The first refers to the need for solidary behaviour, the second to the capacity for solidary behaviour and the third to the efficacy of solidary behaviour. According to our definition, solidary behaviour refers to a (costly) act of a benefactor that benefits another person (the beneficiary). In most cases, the very reason to act solidary is that the beneficiary is in need.4 This need may refer to a broad range of causes, such as illness, frailty, disability, poverty, et cetera. Apart from the fact that the beneficiary is in need, the potential benefactor must also be aware of this. Often, this is unproblematic, because it can be easily seen (e.g., a blind person who has trouble crossing the street). But there are also cases in which it is not selfevident that a person is in need, for example because s/he is ashamed to show it or because the potential benefactor does not know the person(s) in need (for instance, in case of a famine in a third world country). 4

There may be some exceptions to this rule. The most important one is gift-giving. If you give a present at a birthday or a wedding, this is not because the recipient is in need, but because it contributes to continuation and strengthening of the relationship (even though the giver may not be aware of this motive).

38

Chapter One

In those cases in which the benefactor acts out of self-interest, i.e. in the case of bilateral and multilateral solidarity, s/he must face the risk of being in need at some future date her/himself, too. Thus, uncertainty and risk can also be important conditions for solidary behaviour. Again, an additional condition is that one must be aware of this risk. The second general condition is, that the benefactor must be able to act solidary. In general, this means that s/he should have the command over resources that can be employed to support the beneficiary. Usually, the person who gives money to a poor person, has a higher income her/himself, or at least can spare some of her/his income. Apart from money, these resources can also be spare time or even a kidney. The third general condition for solidarity is, that a solidary act is efficacious. This means, that the beneficiary is expected to really benefit from the support offered by the benefactor. Often, efficacy is enhanced by a personal relationship between the benefactor and the beneficiary, because the benefactor can quite easily determine what kind of support the beneficiary needs. Moreover, it is relatively easy to check whether the support achieves its goal. But even then, it is not always obvious what kind of support is in the interest of the beneficiary. Even though an homeless person may beg for alms, a passer-by may doubt whether this will actually help him (“probably he will by a beer from it …”) and may decide it is better not to help than to give the wrong incentive. The efficacy of a solidary act is also affected by the provision of help by others. If other persons are willing to help someone in need, the additional contribution of your help will probably be smaller than if you are the only helper. The condition of efficacy is often a more serious hindrance in case of generalized solidarity. Since the benefactor does not know the beneficiary personally, it is more difficult to decide what the best support is and if it is really effective. To illustrate, the many reports about the alleged ineffectiveness of developmental aid, have probably eroded the public support for transferring money to developing countries. Another example is the recurrent criticism of the alleged ineffectiveness of the welfare state, which, according to some critics, enhances rather than alleviates poverty.

Towards an Interdisciplinary Theory of Solidarity

39

Conclusion In this chapter we have combined insights from various disciplinary perspectives in order to construct an interdisciplinary framework for analysing solidarity. At first sight, different disciplines, such as sociology, anthropology and economics, approach the phenomenon of solidarity from completely different angles. However, closer examination reveals that there are many parallels, similarities and overlaps, which are hidden from sight due to the use of different terms and concepts by the various disciplines. Two common dimensions, which enable to us to distinguish between different motives and mechanisms that explain solidary behaviour, have emerged from this overview. Along the first dimension, self-regarding and other-regarding motives are distinguished. At one extreme, a solidary act is motivated by the fact that it benefits the benefactor her/himself, because s/he expects to get something in return for her/his solidary act, either from the beneficiary (in the case of bilateral solidarity) or from the community at large (in the case of multilateral solidarity). At the other extreme, a solidary act has no tangible benefit for the benefactor, apart from the fact that it satisfies an internal urge to help. Along the second dimension, we distinguish between motives which are directed at concrete persons, with whom the benefactor has a personal relationship, and motives which are directed at anonymous members of one’s community. The combination of these two dimensions results in four ideal-typical kinds of solidarity: empathic solidarity, bilateral solidarity, normative solidarity and multilateral solidarity. However, we do not claim that the wide variety of solidary acts in ‘real life’ should be classified into four categories, but rather that (almost) all solidary behaviour is a mixture of these four ideal-types. Although the weight of the four motives may differ, depending on the person and on the circumstances, in most cases several motives underlie a solidary act.

References Axelrod. R. (1984). The evolution of cooperation. New York: Basic Books. Axelrod, R., and Hamilton, W. D. (1981). The evolution of cooperation. Science, 211, 1390-96. Barr, N. (2004) Economics of the welfare state (4th ed.). Oxford: Oxford University Press.

40

Chapter One

Batson, C.D. (1991). The altruism question: Toward a socialpsychological answer. Hillsdale, NJ, England: Lawrence Erlbaum Associates. Bell, G. (2008). Selection: the mechanism of evolution. Oxford: Oxford University Press. Boyd, R., and Richerson, P.J. (2005) The origin and evolution of cultures. Oxford: Oxford University Press. Boyd, R., Gintis, H., Bowles, S., and Richersen, P.J. (2003). The evolution of altruistic punishment. Proceedings of the National Academy of Science (PNAS), 100, 3531-3535. Carlson, M., Charlin, V., and Miller, N. (1988). Positive mood and helping behaviour: A test of six hypotheses. Journal of Personality and Social Psychology, 55, 211-229. Chai, S.-K. & Hechter, M. (1998). A theory of the state and of social order. In P. Doreian & T. Fararo (eds.). The problem of solidarity. Theories and models (33-60). Amsterdam: Gordon and Breach Publishers. Cialdini, R.B., Schaller, M., Houlihan, D., Arps, K., Fultz, J,, and Beaman, A.L (1987). Empathy-based helping: Is it selflessly or selfishly motivated? Journal of Personality and Social Psychology, 52, 749758. Coleman, J.S. (1990). Foundations of social theory. Cambridge, MA: Belknap Press of Harvard. Dawkins, R. (1976). The selfish gene. Oxford: Oxford University Press. de Waal, F.B.M. (1996). Good natured: The origins of right and wrong in humans and other animals. Cambridge, MA: Harvard University Press. de Waal, F.B.M. (2008). Putting the altruism back into altruism: The evolution of empathy. Annual Review of Psychology, 59, 279-300. Diamond, J. (1997). Guns, germs, and steel. W.W. Norton. Durkheim, E. (1893). The division of labour in society (translated by George Simpson). New York: The Free Press. Eisenberg, N., and Fabes, R.A. (1991). In M.S. Clark (Ed.), Review of Personality and Social Psychology: Vol. 12. Prosocial Behavior. (3461) Newbury Park, CA: Sage. Estrada-Hollenbeck, M., and T. Heatherton. (1998). Avoiding and alleviating guilt through prosocial behavior. In J. Bybee (Ed.), Guilt and children (215-231). San Diego: Academic Press. Fisher, R.A. (1930). The genetical theory of natural selection. Clarendon Press.

Towards an Interdisciplinary Theory of Solidarity

41

Gintis, H. (2003). The hitchhiker’s guide to altruism: Genes, culture, and the internalization of norms. Journal of Theoretical Biology, 220, 407– 18. Haldane, J.B.S. (1955). Population genetics. New Biology, 18, 34–51. Hechter, M. (1987). Principles of group solidarity. Berkeley: University of California Press. Henrich, N., andHenrich, J. (2007). Why humans cooperate: A cultural and evolutionary explanation. Oxford: Oxford University Press. Isen, A. M., and Levine, P. F. (1972). Effect of feeling good on helping: Cookies and kindness. Journal of Personality and Social Psychology, 21, 384-388. Komter, A. (2010). The evolutionary origins of human generosity. International Sociology, 25, 443-464. Kropotkin, p. (2009 [1902]). Mutual aid. London: Heinemann. Lévi-Strauss, C. (1969 [1949]). The elementary structures of kinship. London: Eyre & Spottiswoode. Malinowski, B. (1985 [1922]). Argonauts of the western Pacific: An account of native enterprise and adventure in the archipelagoes of Melanesian New Guinea. London etc.: Routledge and Kegan Paul. Maynard Smith, J. (1964). Group selection and kin selection. Nature, 201, 1145-1147. Nowak, M.A. (2006). Five rules for the evolution of cooperation. Science, 314, 1560–1563. Nowak, M.A., and Sigmund, K. (1998). The dynamics of indirect reciprocity. Journal of Theoretical Biology, 194, 561-574. Ridley, M. (1997). The origins of virtue. Viking. Sahlins, M. (1974). Stone age economics, London: Tavistock Publications. Salovey, P., Mayer, J.D., and Rosenhan, D.L. (1991). Mood and helping: Mood as a motivator of helping and helping as a regulator of mood. Review of Personality and Social Psychology, 12, 215, 237. Sen, A. (1985). The standard of living. The Tanner Lectures. Cambridge: Cambridge University Press. Simmel, G. (1955 [1922]). Conflict and the web of group affiliations (translated and edited by Kurt Wolff). Glencoe: Free Press. Smith, A. (1975 [1759]). An inquiry into the nature and causes of the wealth of nations. London: J.M. Dent & Sons. Toi, M., and Batson, C.D. (1982). More evidence that empathy is a source of altruistic motivation. Journal of Personality and Social Psychology, 43, 281-292. Tönnies, F. (1988 [1887]) Community and society (Gemeinschaft und Gesellschaft). Chicago: Michigan State University Press.

42

Chapter One

Trivers, R. L. (1971). The evolution of reciprocal altruism. The Quarterly Review of Biology, 46, 35–57. Turner, J. C., Oakes, P.J., Haslam, S.A., and McGarty, C. (1994). Self and collective: Cognition and social context. Personality and Social Psychology Bulletin, 20, 454-63. van Oorschot, W., and A. Komter (1998). What is it that ties …? Sociale Wetenschappen 41, 5-25. Wegener, D. T., and Petty, R. E. (1994). Mood management across affective states: The hedonic contingency hypothesis. Journal of Personality and Social Psychology, 66, 1034–1048. Williams, G.C. (1966). Adaptation and natural selection. Princeton, N.J.: Princeton University Press. Williamson, G. M., and Clark, M. S. (1989). The communal/exchange distinction and some implications for understanding justice in families. Social Justice Research, 3, 77-103. Wilson, D.S., and Sober, E. (1994). Reintroducing group selection to the human behavioral sciences. Behavioral and Brain Sciences, 17, 585608. Wilson, D.S., and Wilson, E.O. (2008). Evolution “For the Good of the Group”. American Scientist, 96, 380-389. Wilson, E.O. (1975). Sociobiology: The new synthesis. Cambridge MA: Harvard University Press. Wynne-Edwards, V.C. (1962). Animal dispersion, in relation to social behaviour. Edinburgh: Oliver and Boyd. Wynne-Edwards,V.C. (1986). Evolution through group selection. Oxford: Blackwell Scientific.

CHAPTER TWO WHY WOULD DIVERSITY HARM SOLIDARITY? PAUL DE BEER

Introduction In the past decade, a rapidly growing body of academic literature has addressed the question whether ethnic diversity is detrimental to social capital, trust, public good provision or other measures of social cohesion (e.g., Alesina and LaFerrara 2005; Leigh 2006; Putnam 2007; Habyarimana et al. 2007; Hooghe et al. 2009; Gijsberts et al. 2012). Although the empirical evidence for a negative relationship between ethnic diversity and various socially desirable outcomes is still mixed, the starting point of all these studies is that such a negative relationship may be expected from a theoretical point of view. The common idea underlying these theoretical expectations is that (ethnic) similarity between people fosters recognition, communication, cooperation and trust. Consequently, the more ethnically heterogeneous a community is, the less similar its citizens are and the less they will recognize each other as equals, communicate and cooperate with each other, and trust each other. Although the main subject of this book, social solidarity, is clearly related to the concepts of social capital, trust and social cohesion, these terms do not refer to identical phenomena and, thus, need to be studied separately. The most important difference is that the latter concepts refer primarily to attitudes (recognition, trust) or to more or less enduring relationships (social capital, social cohesion), rather than to the actual behaviour of citizens, which is a crucial element of our definition of solidarity. This difference is important both from a theoretical and from an empirical perspective. Theoretically, attitudes and behaviour may not coincide. As we have argued in the preceding chapter, solidary behaviour need not be motivated by other-regarding motives, such as empathy, but may also be based on self-regarding motives, such as the expected return from helping

44

Chapter Two

or from a gift. Even if ethnic diversity would erode other-regarding attitudes, this would not necessarily result in less solidary behaviour if self-regarding solidarity would be enhanced. Empirically, testing the effect of ethnic diversity by focussing on solidary behaviour may yield more robust outcomes than focussing on attitudes, since the opinions and preferences that people express in surveys are generally more susceptible to social desirability. Moreover, their answers may strongly depend on the exact wording of the questionnaire. To illustrate, if the share of the respondents in different countries who confirm that they think most people can be trusted, differs, this may simply reflect a difference in social desirability of this answer in the various countries rather than point to a real difference in the way that people treat strangers. Another point to note about most empirical studies of the effect of ethnic diversity is that they do not (explicitly) distinguish between two ways in which ethnic diversity may affect individual behaviour. The most obvious effect runs through the interaction between people who belong to different ethnic groups. The more diverse a community is, the more often will people encounter people from a different ethnic group (unless, of course, they live in segregated ethnic enclaves). If these interactions are in some sense less satisfying or positive than interactions with people from the same ethnic group, than an ethnically diverse community will, in the aggregate, have less satisfying relations than an ethnically homogeneous community. However, there may also be a direct effect of the ethnic composition of the community, independent of the dyadic relationships between individual persons. To give an example, an ethnically diverse community may be characterised by a plurality of social norms, which gives individual citizens more latitude to act in the way they prefer instead of complying with the dominant social norm. Moreover, the informal enforcement of social norms may be more difficult in ethnically diverse communities, if people are only sensitive to social pressure which originates from their own group. Hence, ethnic diversity of a community may affect the relationship between people belonging to the same ethnic group as well as the relationship between people from different ethnic groups. In this chapter we will discuss, from a theoretical perspective, what the impact of ethnic diversity on solidary behaviour may be. We take the four ideal-typical kinds of solidarity that were introduced in the previous chapter as our starting point. We identify which (societal) conditions are conducive and which are detrimental for the various kinds of solidarity.

Why Would Diversity Harm Solidarity?

45

We start with conditions that affect the dyadic relation between two persons and analyse what role ethnic similarity or difference plays. Next, we discuss the conditions that are related to the community and examine the effect of ethnic diversity, independently of the individual characteristics of the persons involved.

Solidarity between ethnically different people According to a folk wisdom, likeness breeds solidarity and difference hampers solidarity. Thus, the more similar people are, the more likely they are to act solidary towards each other, and the more they differ, the less solidary behaviour we will observe. However, as we will show in this section, solidarity between two persons is often based on some differences as well as similarities between the two persons, at the same time. The verbal similarity of ‘being like someone’ and ‘to like someone’ suggests that affection or empathy is strongly promoted by the similarity of the (potential) benefactor and beneficiary. In case of solidarity between blood relatives this is indeed so, since, in the socio-biological theory of kin-selection, the genetic relatedness determines the costs that individuals are willing to make to support one another. However, as we know from field studies of empathy, the capacity to experience similar feelings and emotions as another individual is certainly not restricted to blood relatives or to subjects who are very close to each other. The Internet abounds with movie-clips of helping behaviour among animals that are not familiar with each other and sometimes even belong to different species. Everyone is familiar with the feeling of sympathy or empathy with a victim of a famine or a natural disaster which is shown on television, and which can encourage people to make a substantial donation to a charity. Although closeness and likeness may make it easier to trigger empathy, they are certainly not a precondition for empathy and, thus, for altruistic behaviour. In many circumstances, the difference between the benefactor and the beneficiary is even the main reason to act solidary. Because of this difference the former person supports or helps the latter, for example when a parent helps her child, a young woman supports a feeble old man or a native citizen assists a foreigner who does not speak the language. Actually, empathic solidarity often occurs if two persons differ in an important respect but nevertheless are able to identify with each other. Thus, identification is much more important than similarity in understanding empathic solidarity.

46

Chapter Two

Consequently, a difference in ethnicity will only be detrimental to empathic solidarity if ethnicity is a barrier to identification with the other person. Whether this is the case probably depends strongly on the dominant values in a society and the extent to which these have been transferred during upbringing (socialization). Purely bilateral solidarity requires that the cost to the benefactor of a solidary act is lower than the value of the expected return from the beneficiary. Consequently, two conditions have to be met for bilateral solidary behaviour to occur. First, the benefactor must know – or at least expect – that the beneficiary will be able to repay the gift in the future. This presupposes that the beneficiary is only temporarily in need, e.g. due to illness, or that the benefactor and the beneficiary have different skills or abilities, so that they both benefit from an exchange between them. Thus, in general, the benefactor and the beneficiary should consider each other as equivalent. A priori, it is an open question whether ethnic differences go together with differences in skill or ability. If particular skills are related to cultural background, such as craftsmanship (e.g., cookery or music), ethnic difference may even enhance the conditions for bilateral solidarity if the skills that are characteristic for different ethnic groups complement each other. However, if people from a specific ethnic group simply lack skills or education, this can hinder bilateral solidarity, since there is not much room for a mutually beneficial exchange between the groups. Secondly, the benefactor must be confident that the beneficiary will repay the gift. Since a contractual obligation to repay a gift is excluded from our definition of solidarity (because this is tantamount to a market exchange), this requires mutual trust between the benefactor and the beneficiary. Trust between two persons can be enhanced if they interact regularly. Both have then an interest not to betray the other in order to continue their cooperation in the future. As is know from game theory and laboratory experiments, a strategy of (initially) co-operating with another until the other defects, in which case one retaliates by defecting too (tit-for-tat), is generally quite successful. Thus, the more frequently people from different ethnic groups interact, for example because they live in a mixed neighbourhood, the more likely mutual trust will develop. Whether two persons will interact repeatedly depends primarily on the value of these interactions to both of them. Put differently, two individuals

Why Would Diversity Harm Solidarity?

47

may interact repeatedly either because they like each other or because they benefit from each other. The chance that they like each other may be enhanced by some similarity between them, but mutual beneficence may depend, as mentioned above, on some differentiation, such as specialization. When people interact for the first time, the willingness to help each other may depend on the level of generalized trust that is prevalent in a community. We come back to this condition in the next section. The more two people recognize each other as equals, the more likely it is that they are willing to support each other, expecting some return benefit in the future. So, generally speaking, a different ethnic background need not hamper bilateral solidarity as long as both persons consider each other equal (or, at least, equivalent) and trust each other. Whether this is the case, may partly depend on the prevailing norms and values in their community. Although the motives for empathic solidarity and bilateral solidarity are clearly different, they are not mutually exclusive. On the contrary, it is quite likely that repeated interactions between two persons, who first act purely out of self-interest, breed some sympathy or affection for each other. This enhances the willingness to help each other, even if the two persons would gradually grow apart with the passing of time (for example, if one of them would become chronically ill; cf. Putnam’s (2000) famous example of two bowling partners).

Ethnic diversity of the community and solidarity As we stated above, apart from the difference in ethnicity between two persons, the ethnical composition of the community itself may affect solidary behaviour. The reason for this is that the extent of ethnic heterogeneity influences the social context in which individuals act. An ethnically diverse community may differ from an ethnically homogeneous community with respect to the prevailing norms and values and the way that people behave in public. Both normative and multilateral solidarity are strongly dependent on the prevailing social norms regarding solidary behaviour in a community or society. These norms may vary considerably between groups within a community or between communities. The next three conditions are crucial.

48

Chapter Two

First, it is important which persons are considered to be a member of the in-group and are, thus, eligible for assistance if in need. Secondly, a social norm may determine under what conditions a group member deserves help. Is being in need the only condition, or is it also important whether the beneficiary is her/himself responsible for her/his distress and whether the beneficiary has assisted other group members in the past? In other words, which needy persons are considered to be deserving and which are non-deserving? These eligibility conditions may be more or less stringent. Thirdly, since normative solidarity is directed towards anonymous persons, it is important in which situations a person is expected to offer help. This may be guided purely by considerations of efficacy – the person who is able to help another person effectively at the lowest cost – but also by other criteria, for example based on some ranking of the population (e.g., younger people are expected to assist older people, or men should assist women). The first condition, the definition of the in-group, is of particular importance and may show considerable differences between groups. One important element is how large the group is with whom one identifies and whose members one is supposed to help. The broader the definition of the in-group, the larger the number of persons that people are willing to act solidary towards. It makes a lot of difference whether the in-group is the extended family, the neighbourhood or a village, or the nation-state. Of course, these groups need not exclude each other. It is conceivable that the solidarity within the extended family is the strongest, but that people are also expected, albeit to a lesser extent, to help neighbours, villagers or fellow countrymen. The more similar people are, the more likely it is that they are included in the in-group. But, what characteristic is crucial for people to be regarded as similar can vary widely. Thus, an in-group can be very homogeneous with respect to one characteristic (e.g. religion), but strongly heterogeneous with respect to another characteristic (e.g. ethnicity, cf. the Jews in Israel; see also chapter 3). Consequently, although ethnic or cultural diversity may result in the partitioning of the society in more homogeneous groups, this need not be the case if there is some other common characteristic which binds these groups together. However, if ethnicity is closely related to other characteristics, such as religion, language and culture, such a partitioning becomes more likely.

Why Would Diversity Harm Solidarity?

49

Another important element is whether ‘strangers’ who live amidst the members of a group are recognized as group members or are treated as outsiders who do not deserve the solidarity of the group members. The treatment of newcomers or ‘strangers’ varies widely between groups and communities. Nevertheless, each in-group has some restrictions on the acceptance of newcomers, otherwise there would be no distinction between group members and outsiders. Acceptance of a newcomer will, in general, be easier, the more similar the newcomer is to the group members. Once again, this similarity may refer to one specific trait of the newcomer, for example, religion, language or skin colour. An additional characteristic that distinguishes group members from outsiders is simply that the former have been living in this particular community for a longer time. Thus, all else remaining the same, the longer one lives in a particular place, the more one is recognized as an insider, irrespective of any other characteristic. This means that the group members can differ in many respects, except that they share the same history. But even if individual characteristics do not play an important role, it may be pretty hard for ‘immigrants’ to be recognized as members of the group (cf. the new inhabitants of a small closely-knit village community). The second condition refers to the prevailing norms and values in a community, that prescribe under what circumstances a person in need is considered to be deserving to receive help. If deservingness depends on the responsibility of the potential beneficiary for being in need or if it depends on the (solidary) behaviour of the potential beneficiary in the past, fewer people will be entitled to help than if these criteria do not apply. These deservingness criteria are not directly related to the diversity of the group, unless the characteristics of the persons which determine the heterogeneity of the group, are an indicator for their deservingness. For example, if the members of a specific ethnic group are thought to be responsible for their neediness themselves and if self-responsibility is part of the deservingness criteria, then group diversity may hamper solidarity. Put differently, the broader the definition of deservingness, the less likely it is that group diversity hampers solidarity. There might be a relationship with generalized trust, too. If generalized trust is strong, most people may be considered to be deserving, even if deservingness criteria are quite strict, since people tend to believe that others will only ask for help if they meet these criteria. In this case, diversity may not hamper solidarity.

50

Chapter Two

Some communities are characterized by a higher level of generalized trust than other communities. One might expect the level of generalized trust to be higher in communities with a homogeneous population than in diverse communities. Nevertheless, it is a matter of (historical) contingency to which personal characteristics (e.g. ethnicity, religion, age, gender) the homogeneity or heterogeneity of a community relates. Thus, ethnic diversity will hamper generalized trust only if ethnicity is considered to be an important individual characteristic that warrants mistrust of persons belonging to a different (ethnic) group. The existence of particular norms regarding solidarity is not sufficient to elicit solidary behaviour. These norms must also be internalized or externally enforced. As explained in the previous chapter, the main difference between normative and multilateral solidarity is that, in the case of the former, the solidarity norm is internalized, while in the latter case it has to be externally enforced. Internalization of a social norm presupposes a socializing process in which the norms are learned and gradually become the subject’s own. If the norm is fully internalized, i.e. if it has become part of a subject’s belief system, s/he complies with the norm out of conviction, without the need of an external pressure. According to developmental psychologists, the early phases of upbringing are crucial for the internalization of social norms (primary socialization). However, in later phases of life, these norms may be reinforced or weakened by other external influences from, for example, peers and mass media (secondary socialization). It should be noted that internalization of a norm is usually more effectively enhanced by the conduct of others which confirms the norm, than by verbally transmitting the norm (‘practice what you preach’). According to Durkheim, in a homogeneous society, the socialization of social norms may show itself in a collective consciousness (Durkheim 1893: 61). “The solidarity that derives from similarities is at its maximum when the collective consciousness completely envelops our total consciousness, coinciding with it at every point. At that moment our individuality is zero.” (Durkheim 1893: 84) Ethnic diversity of a community may imply a differentiation between ethnic groups with respect to the norms that are internalized. Of course, this presupposes that different ethnic groups endorse different social norms. This can result in a variety of solidary behaviours that may affect the secondary socialization process by peers, mass media, et cetera. For

Why Would Diversity Harm Solidarity?

51

example, if your peers’ interpretation of deservingness is stricter than yours, this may persuade you to adapt your own deservingness criteria to your peers’ standards. Thus, prevailing solidarity norms may evolve over time. If these changing norms translate into other solidary behaviour, this process may reinforce itself. However, this process can move in both directions, either increasing or decreasing solidarity. Thus, there is no general presumption that increasing ethnic heterogeneity will result in less socialization of norms of solidarity. In case of multilateral solidarity, the norms of solidary behaviour need to be externally enforced, since people have not fully internalized the norm. The external pressure that forces people to act according to the social norm, can take various forms. It can be formal or informal, material or psychological, negative or positive. Informal social pressure usually presupposes that others have internalized the social norm and are therefore willing to reprimand a person who violates the norm or to praise a person who complies with the norm. Thus, the larger the proportion of group-members who have internalized the norm, the larger the probability that one will experience external pressure to comply with the norm (cf. de Beer & Mosch 2007). The form of the social pressure exercised by group members can vary from a verbal rebuke to a physical beating or ostracism. The social pressure can also be positive, in case a person who helps an anonymous group member is praised by other group members. Apart from the pleasure of being praised, the status that one derives from this may also enhance the probability of being helped by other group members in the future (indirect reciprocity). Formal social pressure is exercised by a formally appointed supervisor, who is legally entitled to inflict a punishment. In the case of formal, compulsory solidarity, this can be, for example, a tax inspector who is authorized to fine a tax evader. Formal pressure may also be positive, for example if people who act solidary (e.g., volunteers) are awarded a decoration. In a case of multilateral solidarity in which the subject does not believe in the social norm at all, the subject will only comply with the norm if the (expected) costs of non-compliance are higher than the costs of acting solidary. However, a more common situation is probably that a person has partly internalized the norm, but is nevertheless tempted to violate the norm because of the costs of compliance. In that case, the expected

52

Chapter Two

punishment need not equal the full costs, since the subject will be willing to bear at least some of the costs of acting solidary. The effectiveness of deterrence of norm violation depends on the extent to which individuals care about the rebukes or punishment by others. Informal deterrence usually does not take the form of physical coercion, but of a verbal rebuke or a reprimand. It is then crucial that the potential benefactor recognizes the punisher as someone about whose rebuke s/he cares because the punisher belongs to her/his own group. This requires that they are not too different and have something in common. Thus, a reproach from a by-stander from your own ethnic group, inflicted because you fail to help another in-group member who is in need, may be a much stronger incentive to act solidary, than a reproach from a by-stander from another ethnic group. Consequently, increasing ethnic diversity may reduce the effectiveness of reprimands. This need not be the case if there is a formal supervisor who is legally entitled to apply a sanction. However, the effectiveness of a formal supervisor usually depends not only on her/his formal competence but also on her/his personal authority, which may be less if the trespasser belongs to a different ethnic group.

The importance of perceptions It is important to recognize that many of the possible effects of ethnic diversity that have been addressed in the preceding sections depend strongly on the perceptions of people who belong to other ethnic groups. Consequently, it is not so much the actual differences between (typical members of) ethnic groups that matter for solidarity, but the perceived differences. This is self-evident with respect to generalized trust, but it also applies to seemingly objective characteristics, such as the skills and needs of people from other ethnic groups. Of course, also the perception of the members of one’s own ethnic group matters. As a consequence, if perceptions change, without any real change taking place, this may affect solidarity between groups. There is abundant empirical evidence that the perception of the members of another (ethnic) group is often based on stereotypes which do not coincide with their real characteristics. These stereotypes may, for example, be based on a few experiences with atypical members of the group. But they are probably also strongly influenced by the mass media. In particular, regarding people from (ethnic) groups that one rarely meets

Why Would Diversity Harm Solidarity?

53

in person, the image that the media depict of these groups may have a strong influence on the perception of any member of these groups. Although it is notoriously difficult to establish whether the mass media affect the dominant views among the population or that they simply reflect these dominant views, it is hard to deny that the media play an important role in disseminating the prevalent views of various ethnic groups. Therefore, in chapter 5, we will pay attention to the way that the mass media portray various ethnic groups in the Netherlands.

Conclusion From the preceding discussion we can conclude that solidarity is possible both between similar and between different persons and both in ethnically homogeneous and in heterogeneous groups. Nevertheless, very large differences or very strong heterogeneity are generally detrimental to solidarity. This is simply so, because solidarity presupposes recognition of the other or some common purpose, interest or values. If two persons or groups have nothing in common, there is simply no sense in acting solidary. However, this does not imply that the more similar persons are, the stronger will be their willingness to act solidary towards each other. It depends on the motives for solidary behaviour and on the possibility (and willingness) to identify with others. In case of other-regarding solidarity (both empathic and normative solidarity), the identification with another person or with a group is a crucial condition for solidarity. Identification is often fostered by similarity, but this need not always be the case. Identification can also be (purposefully) stimulated or manipulated, for example by the mass media. In case of self-regarding solidarity (bilateral solidarity or multilateral solidarity), the awareness of mutual interdependence is the crucial condition for solidarity. In general, interdependence is created by differentiation and specialization. Thus, some heterogeneity may actually breed self-regarding solidarity. However, if individuals or groups become too different, this may weaken interdependence and reduce trust and, consequently, erode self-regarding solidarity. A crucial condition is then the extent to which the differences between ethnic groups are perceived as complementary, in the sense that they can easily benefit from each other, or as an unbridgeable divide. In the remaining part of this volume, we study the relationship between ethnic diversity and solidarity empirically in various ways. We start with

54

Chapter Two

taking a cross-country perspective by analysing the relationship between various indicators of ethnic diversity at the national and at the regional level and some measures of solidarity in a large scale survey (Chapter 3). Next, we present the outcomes of a so-called vignette analysis, based on a representative survey among the Dutch population, in which the respondents were asked to express their preferences for helping fellow citizens with varying characteristics, including their ethnicity (Chapter 4). In the following step, we focus on the role of the mass media, by analysing the images that are depicted of various ethnic minority groups in some of the most widely read and watched Dutch news media (Chapter 5). Next, we examine the solidary behaviour – or its absence – between citizens with a similar or different ethnic background at the local level, by conducting an experiment with the visitors to a renowned market in a multicultural neighbourhood, called the Dapperbuurt, in the city of Amsterdam (Chapter 6). In the concluding chapter we sum up the main results and draw some general conclusions.

References Alesina, A., and LaFerrara, E. (2002). Who trusts others? Journal of Public Economics 85, 207-234. de Beer, P.T., and Mosch, R.H.J. (2009). The waning and restoration of social norms. DNB Working Paper No. 131, Amsterdam: De Nederlandsche Bank. Gijsberts, M., van der Meer, T., and Dagevos, J. (2012). ‘Hunkering down’ in multi-ethnic neighbourhoods? The effects of ethnic diversity on dimensions of social cohesion. European Sociological Review 28, 527-537. Habyarimana, J., Humphreys, M., Posner, D.N., and Weinstein, J.M. (2007). Why does ethnic diversity undermine public goods provision? American Political Science Review 101, 709-725. Hooghe, M., Reeskens, T., Stolle, D., and Trappers, A. (2009). Ethnic diversity an generalized trust in Europe: A cross-national multilevel study. Comparative Political Studies 42, 198-223. Leigh, A. (2006). Trust, inequality and ethnic heterogeneity. The Economic Record 82, 268-280. Putnam, R.D. (2000). Bowling alone. New York/London/Toronto/Sydney: Simon & Schuster. —. (2007). E Pluribus Unum. Diversity and community in the twenty-first century. The 2006 Johan Skytte Prize Lecture. Scandinavian Political Studies 30, 137-174.

CHAPTER THREE HETEROGENEITY AND HUMAN SOCIALITY FERRY KOSTER

Introduction The term social distance refers to the classification of individuals or groups based on how far they are apart from each other relationally, rather than geographically (Park 1924; Bogardus 1925; Akerlof 1997). Social distance is not a unilateral concept and may refer to any socially relevant difference between actors, such as from group membership, social class, ethnic background, and so forth. Measures of social distance may be based on liking, in-group and out-group distinctions, and the extent to which there is contact between people having different backgrounds. The present chapter focuses on the latter of these three approaches to social distance by examining how social structural diversity, indicated as the chance that two actors with a different background meet each other within a given space (Smith 1960; Blau 1977; Koster 2013), relates to solidarity. For that purpose, international comparative survey data are analysed. Given that such surveys do not include in-depth information on actual behaviour, this chapter investigates the attitudes, motivations, and preferences underlying and enabling solidarity. Therefore, the central topic closely relates to what is called human sociality governing the social relationships between individuals (e.g. Henrich, Boyd, Bowles, Camerer, Fehr and Gintis 2004). In what follows, the focus is on two kinds of sociality that differ with regard to the way in which they are organized, namely informally, based on interpersonal relationships, and formally, involving impersonal means of organization. As an indicator of informal sociality, the level of interpersonal trust is examined and formal sociality includes trust in a country’s institutions and support for the welfare state. This chapter investigates to what extent variation in people’s level of interpersonal trust, institutional trust, and welfare state support can be

56

Chapter Three

explained by the social structure in which these individuals reside. There are several dimensions along which the heterogeneity of social structures can be classified. Three of these dimensions received considerable attention in previous research, namely those based on ethnical, linguistic, and religious distinctions between individuals (Alesina, Devleeschauwer, Easterly, Kurlat and Wacziarg 2003). Since these social structural divisions may have an impact on human sociality through distinguishable mechanisms, their effects are considered separately.

Ethnic heterogeneity There are contrasting theoretical expectations regarding the effects of ethnic heterogeneity. In his review of the “contact hypothesis” (which states that contact between people from different ethnic groups increases mutual understanding and social trust) and “conflict theory” (which in contrast states that in-group solidarity increases at the cost of less outgroup solidarity as ethnic groups compete for scarce resources), Putnam (2007) argues that both of these theoretical approaches miss an important aspect of social reality, namely that ethnic heterogeneity has a negative impact on solidarity altogether as in-group as well as out-group relationships are negatively affected by ethnic heterogeneity. The idea that ethnic heterogeneity decreases in-group as well as out-group solidarity is the core of what is termed the “constrict hypothesis” (Putnam 2007). To a certain extent, this criticism holds when focusing on interpersonal trust, as the majority of studies do in line with Putnam’s study. Nevertheless, in the case of public goods provision, meaning that everyone benefits once the public good is provided, the distinction between the conflict and the constrict hypothesis may be less relevant as they both lead to the prediction that ethnic heterogeneity decreases welfare state support. According to conflict theories, support for certain types of welfare state arrangements may be supported more strongly by one ethnic group than the other if the social risks are unevenly spread across different ethnic groups. Even though this may mean that some groups are more in favour of some of the welfare state arrangements, it also follows that other groups will be less supportive of these arrangements. The end result is that in ethnically heterogeneous social structures there is less support for the welfare state compared to ethnically homogeneous societies, which is the same prediction that follows from the constrict hypothesis. In both cases, the main mechanism underlying this expectation is that heterogeneity decreases the possibilities for providing positive and negative sanctions which are required to sustain a norm of cooperation that in turn generates

Heterogeneity and Human Sociality

57

institutional trust and support for government redistribution (Habyarimana, Humphreys, Posner and Weinstein 2007; Miguel and Gugerty 2005).

Linguistic heterogeneity Linguistic heterogeneity is often viewed as a kind of social structural diversity that is strongly related or even similar to ethnic heterogeneity. Empirically, it is not uncommon to use measures based on the ethnolinguistic fractionalization (ELF) of countries (e.g. Easterly and Levine 1997). Although the ethnic and linguistic heterogeneity of a society may be strongly related, they do not always overlap completely. As Alesina et al. (2003) show, combining ethnic and linguistic heterogeneity into a single measure may lead to a misspecification of the actual level of diversity of a country. Their data show, for example, that most countries in Latin America and the Caribbean are much more divided in terms of groups sharing the same language than along the lines of ethnic groups, while the situation is completely the opposite in European countries. As a result, a measure consisting of ethnic and linguistic heterogeneity blends different mechanisms explaining human sociality and using distinct measures allows theorizing about their unique effects instead. In discussing the different mechanisms explaining why ethnic diversity undermines cooperation, Habyarimana et al. (2007) – drawing on Deutsch (1966) and Hardin (1995) – argue that a shared language is part of the reservoir of common cultural material. Speaking the same language accommodates interpersonal communication and cooperation that gives ethnic homogeneous groups an advantage in producing public goods. In other words, linguistic homogeneity may lead to more efficient social interactions (Anderson and Paskeviciute 2006). However, instead of interpreting a shared language as a part of ethnic heterogeneity, linguistic heterogeneity is investigated here as a specific kind of heterogeneity affecting both informal and formal human sociality through a different mechanism, namely fewer possibilities for collective action due to communication problems, instead of lack of social sanctions.

Religious heterogeneity Religious heterogeneity refers to the presence of people from different religious denominations within a certain social setting. Although religious affiliation may be one of the dividing lines between ethnic groups, it turns out that religious heterogeneity is only weakly related to ethnic and linguistic heterogeneity (Alesina et al. 2003). Thus, there is an empirical

58

Chapter Three

reason for distinguishing their effects. In addition to that, it is also theoretically sound to separate the effects of religious heterogeneity on welfare state attitudes. Religious denominations provide the possibility for members to organize themselves, developing feelings of belonging to a specific group, and exclude those who are not a member of the group (Koster, Goudriaan and van der Schans, 2009). This means that membership of a religious group may create a strong dividing line between insiders and outsiders as it facilitates a clear difference between members and non-members. As such, religious affiliations provide a clear point of reference to distinguish between “them” and “us” which may in turn have consequences for the social behaviour of group members (Yamagishi and Mifune 2008). Furthermore, since the members interact on a regular basis, such divisions may strengthen over time, increasing interpersonal trust within religious groups while decreasing the level of interpersonal trust in outsiders (Stolle 1998). As a result, sharing the same religious background may provide the opportunity structure for interacting with similar others while decreasing social contacts with outsiders and making these differences more pronounced. This, in turn, may have negative consequences for both informal and formal human sociality. In short, these different theoretical considerations lead to the general expectation that social structural diversity has a negative effect on human sociality through different mechanisms. Ethnic diversity is believed to hinder social sanctioning, linguistic diversity hinders efficient communication, and religious diversity can lead to a stronger division between members and non-members. Based on these mechanisms it is hypothesized that interpersonal trust, institutional trust, and support for government redistribution are negatively related to heterogeneity. These individual level attitudes may in turn have macro level consequences. Although interpersonal trust does not necessarily have a direct influence on people’s political preferences, it may have an indirect effect on citizenship behaviours, for example. (It should however be noted that the opposite can also be hypothesized: if interpersonal trust is higher, people may be less trusting towards the government.) The macro level consequences are more clearly related to both institutional trust and support for government redistribution through welfare state legitimacy and support, based on the following considerations. The production of collective goods depends on social structural conditions (e.g. Oliver 1993) and structural heterogeneity may be among the conditions restricting the mechanisms bringing about public support for such goods. In terms of theories emphasizing explanations based on social mechanisms, as for example proposed by authors like Coleman (1990), Hedström and Swedberg (1998) Gross

Heterogeneity and Human Sociality

59

(2009) and Hedström and Ylikoski (2010), the focus is on “situational mechanisms” that connect the macro level with the micro level as social structures and environments shape actions, desires, and believes. In the analyses presented here, the macro level conditions are the level of heterogeneity at the regional level and the micro level concerns individual welfare state attitudes. Each of the three kinds of heterogeneity – ethnic, linguistic, and religious – relates to these individual outcomes through a different mechanism. Ultimately, these individual actions, desires, and believes generate intended and unintended social outcomes through socalled “transformational mechanisms” (Hedström and Swedberg 1998), which translate support into actual social policies (Brooks and Manza 2006).

Prior studies To date, two strands of literature can be identified in which societal heterogeneity and human sociality are investigated empirically. The first branch, which may be called the Putnam strand – as one of the main references is his article “E Pluribus Unum” (2007) – focuses on the relationship between heterogeneity and interpersonal trust. Most of the studies that belong to this part of the literature consist of single country analyses of social structural heterogeneity at relatively low levels of aggregation (e.g. the neighbourhood) using census data (Leigh 2006; Putnam 2007; Tolsma, van der Meer and Gesthuizen 2008). These investigations provide mixed evidence for a negative relationship between heterogeneity and interpersonal trust. A strong point of these studies is that they rely on fine-grained indicators of heterogeneity. They do, however, not allow for a comparison of the effects of diversity across countries. Therefore, they do not provide information about how the impact of diversity differs depending on cross-national variation in institutions and economic circumstances and the consequences they have for the welfare state. This latter issue is the topic of the second branch of the heterogeneity literature, which can be referred to as the Alesina strand with “Fractionalization” by Alesina et al. (2003) as a main source of reference, consisting of international comparative studies that measure diversity at the national level using archival data and examine how it relates to social policies at the national level (Fearon 2003; Alesina and Glaeser 2004) and individual level outcomes such as social capital, citizenship behavior, and welfare state attitudes (Anderson and Paskeviciute 2006; Gesthuizen, van der Meer and Scheepers 2009; Mau and Burkhardt 2009; Burgoon, Koster and van Egmond, 2012). These

60

Chapter Three

studies show that, at the national level, heterogeneity and the welfare state are negatively related. However, when relating national level heterogeneity with individual outcomes, these studies have only found a moderate or weak association. Each of the two approaches to heterogeneity and human sociality lead to important insights, but they also overlook some aspects that need further investigation. The Putnam studies investigate data from one country, usually focus exclusively on ethnic heterogeneity, and do not investigate attitudes towards formal institutions. The Alesina studies, in contrast, do include a wider range of countries, more dimensions of heterogeneity, and different kinds of attitudes, but do not generate very strong conclusions about the relationship between heterogeneity and human sociality. A potential problem with the latter approach is that measures of heterogeneity at the national level may simply be too abstract and general to have an impact on individual attitudes. It may be questioned whether national level indices measure the relevant parts of people’s social context and their preferences, attitudes and behaviour. The present chapter tries to bridge the gap between these two empirical approaches by investigating to what extent individual attitudes relate to regional diversity across European countries (see also Koster, 2013). Two different steps are central for achieving this goal. First, new measures of heterogeneity are constructed at lower levels of aggregation than the ones that are investigated in the Alesina studies. Clearly, these new measures are not at the same level of aggregation as the Putnam studies, but they do allow for cross-national comparisons, which is largely understudied in this part of the literature. The newly constructed measures are compared with external sources to investigate their external validity. Having new measures is not a goal in itself, but it is a stepping stone to answer a more relevant question, namely whether these disaggregated indicators of heterogeneity do explain variation in interpersonal trust, institutional trust, and support for government redistribution. The second step, therefore, involves an empirical comparison of the effect of regional and national level heterogeneity.

Data and method Survey data This study uses the European Social Survey (ESS) as its main source to investigate the relationship between heterogeneity and human sociality. The ESS has been held every two years since 2002 and at the time of this research the data of four rounds were available (ESS 2002; 2004; 2006;

Heterogeneity and Human Sociality

61

2008). To include as much information as possible, these four rounds are combined and analysed together. All of the variables investigated here are asked in all rounds of the ESS as they belong to its core module. Table 1 shows which countries participated per ESS round. In total 33 countries participated in these four rounds. However, the number of times that countries participated varies since not every country participated in every round. There are 15 countries for which data are available for all four rounds, 7 countries participated three times, 7 other countries participated two times, and in 4 countries data have been gathered only once. Round 1 includes 22 countries, in Round 2 26 countries participated, Round 3 includes 23 countries, and Round 4 consists of 28 countries. To get data from a broad range of countries, the average scores are computed for each country across the four rounds of the ESS. Table 1 provides an overview of the ESS Rounds, the countries, and the number of respondents. Table 1. Number of respondents per country and per ESS round

Austria Belgium Bulgaria Switzerland Cyprus Czech Republic Germany Denmark Estonia Spain Finland France United Kingdom Greece Croatia Hungary Ireland Iceland Italy Luxembourg Latvia

Round 1 2,257 1,899 --2,040 ---

Round 2 2,256 1,778 --2,141 ---

Round 3 2,405 1,798 1,400 1,804 995

Round 4 --1,760 2,230 1,819 1,215

6,918 7,235 3,630 7,804 2,210

1,360

3,026

---

2,018

6,404

2,919 1,506 --1,729 2,000 1,503

2,870 1,487 1,989 1,663 2,022 1,806

2,916 1,505 1,517 1,876 1,896 1,986

2,751 1,610 1,661 2,576 2,195 2,073

11,456 6,108 5,167 7,844 8,113 7,368

2,052

1,897

2,394

2,352

8,695

2,566 --1,685 2,046 --1,207 1,552 ---

2,406 --1,498 2,286 579 1,529 1,635 ---

----1,518 1,800 ---------

2,072 1,484 1,544 --------1,980

7,044 1,484 6,245 6,132 579 2,736 3,187 1,980

Total

Chapter Three

62

Netherlands Norway Poland Portugal Romania Sweden Slovenia Slovakia Turkey Total

2,364 2,036 2,110 1,511 --1,999 1,519 ----39,860

1,881 1,760 1,716 2,052 --1,948 1,442 1,512 1,856 47,035

1,889 1,750 1,721 2,222 --1,927 1,476 1,766 --38,561

1,778 1,549 1,619 2,367 2,146 1,830 1,286 1,810 2,416 48,141

7,912 7,095 7,166 8,152 2,146 7,704 5,723 5,088 4,272 173,597

Source: European Social Survey

Comparing heterogeneity and fractionalization International comparative surveys offer a unique opportunity for constructing indicators of social structures. For the present study, data from the ESS are examined for the following reasons. First, this dataset allows construct heterogeneity scores and to compare them empirically across the majority of European countries with the existing fractionalization scores. Secondly, the ESS includes a variable indicating the region where people live, thus allowing for the construction of indicators at disaggregated levels. The variables of interest are the country of origin, which is measured with the question “In which country were you born?”, language, which is asked with the question “What language or languages do you speak most often at home?” and religion, measured with the item “Do you consider yourself as belonging to any particular religion or denomination” and if the respondent answers yes, the follow up question is asked: “Which one?”. These individual answers are aggregated to the country level to construct the national level indicators and to the NUTS level 1 level to construct the regional indicators. This procedure results in a mean value for the categories within a country or a region, which is used to construct heterogeneity scores for ethnic, linguistic and religious diversity per country and per region. These heterogeneity scores are in line with the most common measures found in the literature. Both the Putnam studies and the Alesina studies use the same logic in constructing these scores. The basic formula underlies these measures (which is referred to as Simpson’s diversity index, Blau’s index or the Herfindahl-Hirschman index).

Heterogeneity and Human Sociality

63

where p = proportion of individuals per category N = number of categories. A score of 0 indicates a perfectly homogenous population and 1 a perfectly heterogeneous one (every person belongs to a different group). The score of the index indicates the probability that two randomly selected individuals from the population belong to different groups (e.g. Alesina et al. 2003).

Comparing the number of foreigners in the ESS Even though the ESS aims at including a representative sample of citizens of each country, the question can be asked to what extent the data reflect the national population in terms of nationality or country of birth. To investigate this, data from the Eurostat database on the proportion of foreigners in a country (defined as non-nationals, meaning persons who are not citizens of their country of residence) (Eurostat 2007) are compared to the ESS data. These data are available for 25 of the 33 countries for the period between 2000 and 2008. Table 2 summarizes the outcomes. On average, the foreign population constitutes 7 per cent of the total population of the 25 countries according to the Eurostat dataset. The Eurostat data are compared to the ESS indicator asking about people’s country of origin. Overall, the number of foreign-born people in the ESS is slightly higher than the percentage foreigners reported by Eurostat (0.08 versus 0.07). As table 2 shows, in most countries these numbers are somewhat higher in the ESS than in the Eurostat database. On the other hand, the differences between this ESS indicator and the Eurostat measures are considerable for countries like Luxembourg and Latvia (both 9 per cent difference). In addition to the differences between the Eurostat and the ESS indicators, they are related to each other to investigate their similarity. Figure 1 represents the relationship between the proportion of foreigners according to Eurostat and those interviewed in the ESS. Figure 1 shows that these indicators are positively related. This comparison shows that the ESS includes a sizable number of foreigners and that the proportion of foreigners in the ESS is strongly related to that of an external source (Eurostat), but also that it should be realized that the numbers slightly underestimate the number of actual foreigners (non-nationals) and

64

Chapter Three

that this particularly holds for some of the outliers in the dataset. Table 2. Shares of foreigners, Eurostat and ESS % Foreigners Austria Belgium Bulgaria Switzerland Cyprus Czech Republic Germany Denmark Estonia Spain Finland France United Kingdom Greece Croatia Hungary Ireland Israel Iceland Italy Luxembourg Latvia Netherlands Norway Poland Portugal Romania Russia Sweden Slovenia Slovakia Turkey Ukraine Total

0.09 0.08 0.00 --0.11 0.02 0.09 0.05 0.19 0.05 0.02 0.06 0.05 0.08 --0.01 0.06 ----0.03 0.39 0.23 0.04 --0.00 0.02 0.00 --0.05 0.02 0.00 ----0.07

% Foreign born 0.08 0.09 0.01 (0.19) 0.07 0.03 0.08 0.06 0.20 0.07 0.03 0.09 0.10 0.09 (0.09) 0.02 0.09 (0.35) (0.03) 0.02 0.30 0.14 0.08 (0.07) 0.01 0.06 0.01 (0.06) 0.11 0.08 0.03 (0.01) (0.12) 0.08

Sources: Eurostat 2000-2008, European Social Survey 1-4

Difference 0.01 -0.01 -0.01 --0.04 -0.01 0.01 -0.01 -0.01 -0.02 -0.01 -0.03 -0.05 -0.01 ---0.01 -0.03 ----0.01 0.09 0.09 -0.04 ---0.01 -0.04 -0.01 ---0.06 -0.06 -0.03 -----0.01

Heterogeneity and Human Sociality

65

Figure 1. Comparison of percentage foreigners

Comparing heterogeneity with fractionalization Table 3 shows the fractionalization scores of Alesina et al. (2003) together with the heterogeneity measures based on the ESS data. The average scores of the indicators that are constructed using the ESS are consistently lower than the fractionalization scores. While the ethnic fractionalization of the European countries is 0.25, their average ethnic heterogeneity is 0.15 based on the country of origin indicator of the ESS. According to the fractionalization dataset the mean level of linguistic fractionalization is also 0.25, whereas the ESS reports a mean of 0.18. The level of religious fractionalization is 0.37 compared to 0.26 in the ESS data. Again there are some noteworthy country differences, as can be read in Table 4. For a number of countries it holds that the level of fractionalization and the ESS based heterogeneity is quite similar. For some countries, the measures are exactly the same: Norway, Portugal, and Sweden have a difference of zero. There are 15 countries in the dataset for which the difference

66

Chapter Three

between the ethnic fractionalization score and the ESS heterogeneity measure based on the nationality of respondents falls within a range of a 0.10 and this number increases to 0.18 if ethnic fractionalization is compared to the country of origin measure of the ESS. On the other hand, there are countries with a difference that is higher than 0.30 (Bulgaria and Croatia, for example) and there are two countries with a difference of 0.46 (Belgium and Ukraine). Focusing on the country of origin indicator of the ESS, the differences are somewhat smaller, but it also turns out that for some countries (namely Sweden and Israel) the level of ethnic heterogeneity according to the ESS is higher than the ethnic fractionalization scores. Comparing the two datasets with regard to linguistic diversity shows that the fractionalization scores and the ESS based indicator are closer to each other than the ethnic diversity measures. The indicators fall within a range of 0.10 for 25 of 33 the countries. The difference between the level of linguistic fractionalization and the level measured with the ESS dataset is particularly large in the Czech Republic (a difference of 0.30), Cyprus (a difference of 0.36), and the Netherlands (a difference of 0.42). The linguistic fractionalization score of Italy is lower than the ESS based measure of heterogeneity (a difference of 0.13). With regard to the level of religious diversity, the two types of indicators show the following similarities and differences. There are 18 countries with a difference within the 0.10 range and five countries have a difference that is larger than 0.30, notably the Czech Republic (0.41), Cyprus (0.37), Spain (0.35), Ukraine (0.32), and Luxembourg (-0.34). The relationships between the different indicators of diversity are shown in Figures 2 through 4. There is a positive relationship between the fractionalization scores and the corresponding indicators measured with the ESS. The correlation coefficients presented in Table 5 confirm this relationship. Ethnic fractionalization is correlated with the country of origin measures of the ESS (r = 0.455), the two indicators measuring linguistic diversity are positively related (r = 0.818) and the same holds for religious diversity (r = 0.695). Table 5 also shows the interrelations between the different indicators. In accordance with the correlation coefficients calculated for the whole sample of the fractionalization data, ethnic and linguistic fractionalization are positively related to each other (r = 0.770) and are less strongly related to religious fractionalization (r = 0.177 for ethnic fractionalization and r = 0.346 for linguistic fractionalization). The patterns are somewhat similar when the ESS measures

Austria Belgium Bulgaria Switzerland Cyprus Czech Republic Germany Denmark Estonia Spain Finland France United Kingdom Greece Croatia Hungary Ireland Israel Iceland Italy Luxembourg

Ethnic 0.11 0.56 0.40 0.53 0.09 0.32 0.17 0.08 0.51 0.42 0.13 0.10 0.12 0.16 0.37 0.15 0.12 0.34 0.08 0.11 0.53

Fractionalization Linguistic Religious 0.15 0.41 0.54 0.21 0.30 0.60 0.54 0.61 0.40 0.40 0.32 0.66 0.16 0.66 0.10 0.23 0.49 0.50 0.41 0.45 0.14 0.25 0.12 0.40 0.05 0.69 0.03 0.15 0.08 0.04 0.03 0.52 0.03 0.16 0.55 0.35 0.08 0.19 0.11 0.30 0.64 0.09

Table 3. Fractionalization and ESS heterogeneity

Origin 0.15 0.16 0.02 0.35 0.12 0.06 0.15 0.11 0.33 0.14 0.04 0.17 0.18 0.16 0.16 0.04 0.17 0.57 0.06 0.04 0.50

Heterogeneity and Human Sociality

Heterogeneity Linguistic 0.06 0.51 0.25 0.53 0.04 0.02 0.07 0.04 0.43 0.25 0.12 0.10 0.07 0.07 0.01 0.01 0.05 0.51 0.01 0.24 0.50

Religious 0.21 0.22 0.36 0.58 0.03 0.25 0.57 0.13 0.54 0.10 0.07 0.24 0.58 0.07 0.14 0.43 0.10 0.37 0.16 0.03 0.43

67

0.59 0.11 0.06 0.12 0.05 0.31 0.25 0.06 0.22 0.25 0.32 0.47 0.25

0.58 0.51 0.07 0.05 0.02 0.17 0.25 0.20 0.22 0.26 0.22 0.47 0.25

0.56 0.72 0.20 0.17 0.14 0.24 0.44 0.23 0.29 0.57 0.00 0.62 0.37

Chapter Three

0.25 0.16 0.12 0.03 0.11 0.01 0.11 0.20 0.15 0.05 0.02 0.21 0.15

0.42 0.09 0.09 0.00 0.02 0.14 0.10 0.10 0.04 0.20 0.23 0.50 0.18

Sources: Alesina, Devleeschauwer, Easterly, Kurlat and Wacziarg (2003) and European Social Survey 1-4

Latvia Netherlands Norway Poland Portugal Romania Russia Sweden Slovenia Slovakia Turkey Ukraine Total

68

0.70 0.62 0.20 0.03 0.07 0.22 0.22 0.28 0.13 0.32 0.03 0.30 0.26

Heterogeneity and Human Sociality

69

Table 4. Differences between fractionalization and ESS heterogeneity

Austria Belgium Bulgaria Switzerland Cyprus Czech Republic Germany Denmark Estonia Spain Finland France United Kingdom Greece Croatia Hungary Ireland Israel Iceland Italy Luxembourg Latvia Netherlands Norway Poland Portugal Romania Russia Sweden Slovenia Slovakia Turkey Ukraine Total

Ethnic/origin -0.04 0.40 0.38 0.18 -0.03 0.26 0.02 -0.03 0.18 0.28 0.09 -0.07 -0.06 0.00 0.21 0.11 -0.05 -0.23 0.02 0.07 0.03 0.34 -0.05 -0.06 0.09 -0.06 0.30 0.14 -0.14 0.07 0.20 0.30 0.26 0.09

Source: European Social Survey 1-4

Linguistic 0.09 0.03 0.05 0.01 0.36 0.30 0.09 0.06 0.06 0.16 0.02 0.02 -0.02 -0.04 0.07 0.02 -0.02 0.04 0.07 -0.13 0.14 0.16 0.42 -0.02 0.05 0.00 0.03 0.15 0.10 0.18 0.06 -0.01 -0.03 0.07

Religious 0.20 -0.01 0.24 0.03 0.37 0.41 0.09 0.10 -0.04 0.35 0.18 0.16 0.11 0.08 -0.10 0.09 0.06 -0.02 0.03 0.27 -0.34 -0.14 0.10 0.00 0.14 0.07 0.02 0.22 -0.05 0.16 0.25 -0.03 0.32 0.10

70

Chapter Three

are investigated. The ethnic diversity measure based on country of birth is positively related to linguistic diversity (r = 0.567 and r = 0.652). The correlations between religious diversity on the one hand and ethnic and linguistic diversity on the other hand are different than those reported for the fractionalization measures: ethnic and religious heterogeneity measured with the ESS are positively related (r = 0.462 and r = 0.472), while linguistic and religious heterogeneity have the lowest correlation (r = 0.419). Figure 2. Ethnic diversity

Heterogeneity and Human Sociality Figure 3. Linguistic diversity

71

72

Chapter Three

Figure 4. Religious diversity

Furthermore, Table 5 provides insight how the diversity scores from the different sources are related to each other. With regard to ethnic fractionalization, it is worth mentioning that this indicator of diversity is most strongly related to the linguistic heterogeneity measured in the ESS (r = 0.821). Moreover, linguistic fractionalization is also strongly related to the ESS based indicator of linguistic heterogeneity (r = 0.818). A possible interpretation of this result, in combination with the strong relation between ethnic and linguistic fractionalization and the fact that the ethnic fractionalization and ethnic heterogeneity measured with the ESS are less strongly correlated, is that despite the effort to distinguish ethnic and linguistic fractionalization, the ethnic fractionalization indicator mainly reflects the linguistic diversity within a country. However, before this conclusion can be stated succinctly additional data explorations are required, which can be the subject of future studies. The results regarding religious fractionalization are clear-cut. This variable is not related to any of the heterogeneity scores based on the ESS except for religious

Heterogeneity and Human Sociality

73

heterogeneity (r = 0.695), suggesting that these indicators essentially measure similar aspects of the countries investigated in this sample. Table 5. Correlation coefficients fractionalization and heterogeneity

Ethnic

Fractionalization Linguistic Religious

Fractionalization Ethnic 0.770** Linguistic 0.177 0.346* Religious Heterogeneity 0.455** 0.628** Origin 0.821** 0.818** Linguistic 0.414* 0.506** Religious N = 33 countries ** p < 0.01; * p < 0.05 Sources: Alesina, Devleeschauwer, Easterly, European Social Survey 1-4

0.053 0.160 0.695**

Heterogeneity Origin Linguistic

0.652** 0.472**

0.419*

Kurlat and Wacziarg (2003) and

The comparison of the ESS data with two external sources supports the following conclusions. First, on average, the ESS includes shares of foreigners that are quite similar to the figures provided by Eurostat. However, since for some countries the ESS shares of foreigners deviate considerably from the Eurostat numbers, it is advisable to explore whether this affects the resulting indices. In empirical analyses, it is possible to investigate how much inclusion or exclusion of these countries affects the outcomes by conducting a sensitivity analysis. Secondly, comparing different indicators measuring societal diversity, leads to the conclusion that the ESS based scores are consistently lower than the fractionalization scores, which were taken as a point of reference. With regard to this outcome, it cannot simply be stated that the ESS underestimates the level of diversity, as it is not clear what the true level of diversity is in the countries included in the analysis. It may as well be concluded that the fractionalization scores overestimate the level of diversity. There may be a number of reasons for these differences. First, the ESS may include too few respondents from different backgrounds, thus lacking the ability to capture the diversity to the fullest. Secondly, as noted above, related but different dimensions of diversity may be measured in both datasets. In particular, this remark relates to the close relationship between ethnic and linguistic diversity. And, thirdly, part of the difference can result from the difference between using a more objective source (archival data measuring fractionalization) and self-reports involving a certain level of subjectivity.

74

Chapter Three

Given the differences between the ESS indicators and the external sources, it is acknowledged that ESS data are not completely the same as the Eurostat and fractionalization data. However, a focus on the similarities between the measures, by investigating country plots and correlations, also shows that the indicators are interrelated and thus provide evidence that the measures at least reflect a relatively equal ordering of the countries in the dataset in terms of diversity.

Regional heterogeneity measures So far, the focus was on cross national variation in diversity, a topic that has received quite some attention in prior research, whereas other studies have focused on diversity at lower levels of aggregation. A combination of these two strategies, investigating diversity at regional levels across a sample of countries, is not available yet. With the ESS data this can be investigated as respondents are asked about their region of residence. Appendix 1 shows the number of respondents and the ESS heterogeneity scores for each country disaggregated to the NUTS level 1 region. For 11 countries the country level is the same as the NUTS level. This means that the diversity measures cannot be disaggregated to a lower level based on this classification scheme. The level of intra-country diversity of the remaining 19 countries can be further examined as they fall into a number of different NUTS level 1 regions. The number of regions ranges from 2 (Bulgaria and Finland) to 16 (Germany). The data in Appendix 1 provide some evidence that NUTS level 1 regions within countries differ with regard to their level of heterogeneity measured with the ESS indicators. Table 6 investigates these differences more closely. Table 6 reports the country means, standard deviations (based on the NUTS level 1 regions) and the range of heterogeneity (the difference between the minimum and the maximum value of the NUTS level 1 region within a country). With regard to ethnic heterogeneity, Table 6 shows the following. Bulgaria and Finland have the lowest value on this measure with a standard deviation of 0.00 and a range of 0.01. In contrast, Belgium and the United Kingdom have a standard deviation of 0.08 and 0.10 and a range of 0.30 and 0.39, respectively. In Hungary and Poland, the level of linguistic heterogeneity varies the least across the NUTS level 1 regions of the countries in this sample (the standard deviations for both countries are 0.01 and the range is also 0.01). Quite different is the situation for Turkey and Spain where the standard deviations are 0.22 and 0.18 and the scores

Austria Belgium Bulgaria Croatia Cyprus Czech Republic Denmark Estonia Finland France Germany Greece Iceland Italy Latvia Luxembourg Netherlands Norway Poland Portugal Romania

Ethnic heterogeneity Mean SD 0.140 0.026 0.240 0.154 0.025 0.007 0.160 0.044 0.130 --0.060 --0.110 --0.320 --0.055 0.007 0.148 0.083 0.139 0.069 0.104 0.062 0.060 --0.046 0.024 0.240 --0.430 --0.135 0.049 0.124 0.046 0.033 0.022 0.110 --0.010 0.008

Table 6. Regional heterogeneity Linguistic heterogeneity Mean SD 0.063 0.025 0.187 0.177 0.250 0.057 0.007 0.006 0.040 --0.020 --0.040 --0.430 --0.130 0.028 0.101 0.060 0.072 0.039 0.041 0.037 0.010 --0.238 0.093 0.420 --0.500 --0.108 0.116 0.084 0.032 0.005 0.008 0.020 --0.130 0.177

Heterogeneity and Human Sociality

Religious heterogeneity Mean SD 0.207 0.050 0.307 0.220 0.360 0.000 0.137 0.038 0.030 --0.250 --0.130 --0.540 --0.075 0.007 0.214 0.111 0.434 0.108 0.216 0.196 0.160 --0.036 0.013 0.700 --0.430 --0.508 0.181 0.184 0.093 0.032 0.015 0.070 --0.210 0.211

75

0.060 0.150 0.133 0.180 0.303 0.045 0.152

Source: European Social Survey 1-4

Slovakia Slovenia Spain Sweden Switzerland Turkey United Kingdom

76

----0.053 0.066 0.049 0.109 0.105

0.200 0.040 0.170 0.093 0.310 0.179 0.060

Chapter Three

----0.169 0.040 0.108 0.234 0.059

0.320 0.130 0.094 0.260 0.512 0.028 0.537

----0.044 0.100 0.076 0.049 0.077

Heterogeneity and Human Sociality

77

on linguistic heterogeneity have a range 0.62 and 0.41, respectively. Finally, religious heterogeneity also turns out to vary across the NUTS level 1 regions within countries. With respect to this kind of diversity, Bulgaria and Finland stand out with a standard deviation of 0.00 and a range of 0.00 and 0.01, respectively. On the other side there are countries like Romania and the Netherlands with a standard deviation of 0.18 and 0.16 and a range of 0.47 and 0.41, respectively. This quick overview of how the measures of heterogeneity constructed with the ESS data vary across regions within countries provides evidence for the conclusion that in some countries the different regions are far more similar to each other than in other countries. In practice, this means that in some countries people from different regions experience the same level of diversity, whereas in other countries this depends much more on where the person lives. With respect to the impact that diversity may have on people’s behaviour, attitudes and opinions, these differences may indicate that studies relating national level diversity to individual level outcomes miss an important share of the variation in diversity across regions within countries.

Effects of national level heterogeneity versus regional level heterogeneity In this section, three kinds of sociality are investigated that are close to the ones examined in previous studies, namely interpersonal trust, institutional trust, and support for government redistribution. The main aim of this part of the analysis is to examine whether regional heterogeneity affects human sociality using international comparative data rather than single level data. To examine whether regional level indicators of heterogeneity yield different individual outcomes than the ones at the national level the results for these indicators are compared.

Multilevel analyses To take the nested structure of the data into account a three-level multilevel model is constructed. The three dependent variables are investigated following the same steps. First a control model (Model 1) is estimated that functions as a point of reference to compare subsequent models. Then, three models are constructed investigating the effects of ethnic, linguistic, and religious heterogeneity at the regional level (Models 2a-c). And finally, three models are estimated investigating ethnic,

78

Chapter Three

linguistic and religious heterogeneity at the national level (Models 3a-c). The deviance (the difference in -2 log likelihood) is computed using full information maximum likelihood to assess the change in the fit of the models (Snijders and Bosker 1999). The parameters in these models are estimated by the maximum likelihood method and the regression coefficients are tested by Wald tests (Goldstein 2003; Snijders 2003). The results of the multilevel analyses are reported in Table 7, 8, and 9. Concentrating on the models with the control variables, the analyses show that the outcomes for the three kinds of sociality differ markedly. While people report higher levels of interpersonal trust and support for government redistribution in later rounds of the ESS compared to the first, their trust in institutions is lower in the later rounds of the ESS. Women report higher levels of interpersonal trust and they are more in favour of government redistribution, but gender does not explain variation in institutional trust. Interpersonal trust and support for government redistribution is negatively related to age, while institutional trust is not related to the age of respondents. Educational level is positively related to interpersonal trust and institutional trust and negatively related to support for government redistribution. Finally, while interpersonal trust and institutional trust are higher in more wealthy countries, GDP per capita is negatively related to support for government redistribution. Turning to the models relating regional heterogeneity and national level fractionalization to interpersonal trust, Table 7 shows that none of these indicators explains variation in this kind of sociality. All forms of regional and national level diversity, except religious heterogeneity of the region, have a negative sign, but none of these effects is significantly different form zero. Investigating institutional trust, Table 8 shows that it is positively related to ethnic heterogeneity and linguistic heterogeneity at the regional level and negatively to ethnic fractionalization and religious fractionalization of the country. Nevertheless, focusing on how much the variables contribute to the fit of the model, it is evident that including both effects at the regional level leads to a significant improvement (Deviance = 3.900; p < 0.05 for ethnic heterogeneity and Deviance = 15.800; p < 0.01 for linguistic heterogeneity), while including both heterogeneity indicators at the national level does not result in significant improvements of the model (Deviance = 3.800; n.s. for ethnic fractionalization and Deviance = 3.300; n.s. for religious fractionalization).

ESS Round 3

ESS Round 2

Control variables

Religious fractionalization

Linguistic fractionalization

Ethnic fractionalization

Country level

Religious heterogeneity

Linguistic heterogeneity

Ethnic heterogeneity

Regional level

0.038** (0.013) 0.111** (0.013)

(1)

0.038** (0.013) 0.111** (0.013)

-0.295 (0.377)

(2a)

Table 7. Multilevel analysis of interpersonal trust

0.038** (0.013) 0.111** (0.013)

-0.169 (0.227)

(2b)

0.038** (0.013) 0.112** (0.013)

0.316 (0.233)

(2c)

Heterogeneity and Human Sociality

0.038** (0.013) 0.111** (0.013)

-1.197 (0.726)

(3a)

0.038** (0.013) 0.111** (0.013)

-0.515 (0.692)

(3b)

0.038** (0.013) 0.111** (0.013)

0.662 (0.586)

(3c)

79

4.827 (0.129) 0.462 (0.126) 0.052 (0.009) 3.136 (0.011)

4.826 (0.130) 0.465 (0.127) 0.052 (0.009) 3.136 (0.011) 0.616

0.129** (0.013) 0.071** (0.009) -0.003** (0.000) 0.054** (0.001) 0.625** (0.125)

0.129** (0.013) 0.071** (0.009) -0.003** (0.000) 0.054** (0.001) 0.609** (0.125)

0.557

4.829 (0.129) 0.459 (0.125) 0.052 (0.009) 3.136 (0.011)

0.129** (0.013) 0.071** (0.009) -0.003** (0.000) 0.054** (0.001) 0.613** (0.123)

Chapter Three

1.844

4.830 (0.129) 0.456 (0.125) 0.051 (0.008) 3.136 (0.011)

0.129** (0.013) 0.071** (0.009) -0.003** (0.000) 0.054** (0.001) 0.604** (0.123)

2.613

4.847 (0.125) 0.425 (0.116) 0.052 (0.009) 3.136 (0.011)

0.129** (0.013) 0.071** (0.009) -0.003** (0.000) 0.054** (0.001) 0.591** (0.119) 4.844 (0.128) 0.443 (0.121) 0.052 (0.009) 3.136 (0.011) 1.248

0.546

0.129** (0.013) 0.071** (0.009) -0.003** (0.000) 0.054** (0.001) 0.628** (0.122) 4.832 (0.129) 0.457 (0.125) 0.052 (0.009) 3.136 (0.011)

0.129** (0.013) 0.071** (0.009) -0.003** (0.000) 0.054** (0.001) 0.630** (0.126)

Individuals = 166,458; Regions = 123; Countries = 30 ** p < 0.01; * p < 0.05 Sources: European Social Survey 1-4, Alesina, Devleeschauwer, Easterly, Kurlat and Wacziarg (2003), and Eurostat

Deviance

Variance individual level

Variance regional level

Variance country level

Intercept

GDP per capita

Years of education

Age

Gender (1 = female)

ESS Round 4

80

ESS Round 4

ESS Round 3

ESS Round 2

Control variables

Religious fractionalization

Linguistic fractionalization

Ethnic fractionalization

Country level

Religious heterogeneity

Linguistic heterogeneity

Ethnic heterogeneity

Regional level

-0.178** (0.014) -0.092** (0.014) -0.140**

(1)

-0.178** (0.014) -0.092** (0.014) -0.140**

0.811* (0.388)

(2a)

Table 8. Multilevel analysis of institutional trust

-0.178** (0.014) -0.092** (0.014) -0.140**

0.933** (0.224)

(2b)

-0.178** (0.014) -0.092** (0.014) -0.140**

0.314 (0.241)

(2c)

Heterogeneity and Human Sociality

-0.178** (0.014) -0.092** (0.014) -0.140**

-1.202* (0.592)

(3a)

-0.178** (0.014) -0.092** (0.014) -0.140**

-0.531 (0.570)

(3b)

-0.178** (0.014) -0.092** (0.014) -0.140**

-0.873* (0.462)

(3c)

81

4.840 (0.131) 0.334 (0.093) 0.052 (0.009) 3.113 (0.012)

4.998 (0.107) 0.302 (0.086) 0.062 (0.010) 3.113 (0.012) 3.900*

(0.014) -0.006 (0.009) 0.000 (0.000) 0.042** (0.001) 0.476** (0.109)

(0.014) -0.006 (0.009) 0.000 (0.000) 0.042** (0.001) 0.520** (0.103)

15.800**

4.843 (0.130) 0.347 (0.096) 0.052 (0.009) 3.113 (0.012)

(0.014) -0.006 (0.009) 0.000 (0.000) 0.042** (0.001) 0.496** (0.108)

Chapter Three

1.500

4.843 (0.131) 0.333 (0.093) 0.052 (0.009) 3.113 (0.012)

(0.014) -0.006 (0.009) 0.000 (0.000) 0.042** (0.001) 0.514** (0.107)

3.800

4.861 (0.126) 0.267 (0.077) 0.052 (0.009) 3.113 (0.012)

(0.014) -0.006 (0.009) 0.000 (0.000) 0.042** (0.001) 0.500** (0.098)

0.800

4.846 (0.130) 0.295 (0.084) 0.052 (0.009) 3.113 (0.012)

(0.014) -0.006 (0.009) 0.000 (0.000) 0.042** (0.001) 0.542** (0.104)

Individuals = 143,470; Regions = 123; Countries = 30 ** p < 0.01; * p < 0.05 Sources: European Social Survey 1-4, Alesina, Devleeschauwer, Easterly, Kurlat and Wacziarg (2003), and Eurostat

Deviance

Variance individual level

Variance regional level

Variance country level

Intercept

GDP per capita

Years of education

Age

Gender (1 = female)

82

3.300

4.858 (0.129) 0.263 (0.076) 0.052 (0.009) 3.113 (0.012)

(0.014) -0.006 (0.009) 0.000 (0.000) 0.042** (0.001) 0.500** (0.097)

ESS Round 3

ESS Round 2

Control variables

Religious fractionalization

Linguistic fractionalization

Ethnic fractionalization

Country level

Religious heterogeneity

Linguistic heterogeneity

Ethnic heterogeneity

Regional level

0.044** (0.007) 0.082**

(1)

0.044** (0.007) 0.082**

-0.825** (0.226)

(2a)

0.044** (0.007) 0.082**

-0.219 (0.142)

(2b)

Table 9. Multilevel analysis of government redistribution

0.044** (0.007) 0.082**

-0.461** (0.128)

(2c)

Heterogeneity and Human Sociality

0.044** (0.007) 0.082**

0.002 (0.237)

(3a)

0.044** (0.007) 0.082**

-0.163 (0.215)

(3b)

0.044** (0.007) 0.082**

-0.372* (0.163)

(3c)

83

3.634 (0.042) 0.040 (0.013) 0.022 (0.003) 0.948 (0.003)

3.633 (0.040) 0.034 (0.012) 0.026 (0.004) 0.948 (0.003) 11.993**

(0.007) 0.041** (0.007) 0.130** (0.005) -0.002** (0.000) -0.028** (0.001) -0.103** (0.039)

(0.007) 0.040** (0.007) 0.130** (0.005) -0.002** (0.000) -0.028** (0.001) -0.149** (0.039)

2.297

3.627 (0.041) 0.037 (0.012) 0.025 (0.004) 0.948 (0.003)

(0.007) 0.041** (0.007) 0.130** (0.005) -0.002** (0.000) -0.028** (0.001) -0.144** (0.040)

Chapter Three

12.187**

3.628 (0.038) 0.029 (0.010) 0.024 (0.004) 0.948 (0.003)

(0.007) 0.041** (0.007) 0.130** (0.005) -0.002** (0.000) -0.028** (0.001) -0.138** (0.037)

0.000

3.633 (0.041) 0.034 (0.012) 0.026 (0.004) 0.948 (0.003)

(0.007) 0.040** (0.007) 0.130** (0.005) -0.002** (0.000) -0.028** (0.001) -0.149** (0.040)

0.568

3.634 (0.040) 0.033 (0.011) 0.026 (0.004) 0.948 (0.003)

(0.007) 0.041** (0.007) 0.130** (0.005) -0.002** (0.000) -0.028** (0.001) -0.143** (0.040)

Individuals = 167,367; Regions = 123; Countries = 30 ** p < 0.01; * p < 0.05 Sources: European Social Survey 1-4, Alesina, Devleeschauwer, Easterly, Kurlat and Wacziarg (2003), and Eurostat

Deviance

Variance individual level

Variance regional level

Variance country level

Intercept

GDP per capita

Years of education

Age

Gender (1 = female)

ESS Round 4

84

4.964*

3.626 (0.037) 0.027 (0.010) 0.026 (0.004) 0.948 (0.003)

(0.007) 0.040** (0.007) 0.130** (0.005) -0.002** (0.000) -0.028** (0.001) -0.153** (0.036)

Heterogeneity and Human Sociality

85

Table 9 shows the results for support for government redistribution. At the regional level, two of the three heterogeneity measures are significantly related to this kind of sociality, namely ethnic heterogeneity (b = -0.825; p < 0.01) and religious heterogeneity (b = -0.461; p < 0.01). In both these cases, the model fit improves significantly (Deviance = 11.993; p < 0.01 for ethnic heterogeneity and Deviance = 12.187; p < 0.01 for religious heterogeneity). Besides, support for government redistribution is significantly (at the 5 per cent significance level) negatively related to religious fractionalization (b = -0.372; p < 0.05; Deviance = 4.964; p < 0.05). These results refute both the Putnam studies and the Alesina studies. In contrast with the expectation found across the Putnam studies, no relationship is found between heterogeneity and interpersonal trust in this international comparative study. Clearly, the outcome with regard to interpersonal trust may be attributed to the data that were used, as they can be criticized for being too crude to take the social context of individuals into account. Nevertheless, a different interpretation of this outcome is that cross-national differences in interpersonal trust should be explained by other factors than heterogeneity alone. Although a large share of the Alesina studies do not address individual values and voting behaviour explicitly, it is a central part of the explanation for why national heterogeneity negatively affects the welfare state. Again, for this prediction, the analyses presented here offer very little support (with the relationship between religious heterogeneity and support for government redistribution being the only exception). This seems to suggest that decreasing lack of institutional trust and support for the welfare state are not the prime explanation for Alesina’s finding that national level heterogeneity undermines the welfare state. Nevertheless, at the regional level, heterogeneity is related to institutional trust and support for government redistribution. Somewhat surprisingly, institutional trust is positively related to ethnic and linguistic heterogeneity at that level. This finding is in contrast with the mechanisms most often discussed in the literature. Thus, rather than undermining trust in formal institutions, ethnic and linguistic heterogeneity may strengthen it. With regard to support for government redistribution, the analyses do support the expectation that heterogeneity – the ethnic and religious dimension – decreases this attitude of individuals. Combining these outcomes leads to the conclusion that the link between heterogeneity and human sociality is not as straightforward as is suggested by a large share of the literature. While regional heterogeneity is positively related to institutional trust, it is negatively related to support

86

Chapter Three

for government redistribution. This suggests that people are less willing to support the welfare state, but not because their trust in government institutions is in decline.

Conclusion Both social structure and solidarity belong to the core topics of the social sciences. Societal changes such as migration and ageing of the population have spurred scientific and societal interest as they provide prime examples of how these two abstract concepts relate to each other empirically. This chapter argues that to date research efforts have focused on two distinct but related approaches, namely the more micro oriented research relating heterogeneity at the neighbourhood level (in particular the ethnical dimension) to interpersonal trust (which was referred to as the Putnam strand of research) and the macro oriented studies that relate national level heterogeneity to welfare state arrangements (termed the Alesina strand). The analyses discussed in this chapter aim at bringing these two strands closer together as their theoretical underpinnings and empirical findings are relevant for both of them. Nevertheless, given the observed gaps between the two kinds of studies it is impossible at the moment to integrate them completely. To name a few important differences: while the Putnam strand includes single country data, the Alesina studies aim at understanding cross national variation, and whereas the Putnam studies investigate interpersonal trust, the Alesina studies examine welfare state arrangements. As a result, prior studies do not allow relating the macro level to the micro level as they have not been studied in unison. This chapter is an effort in offering such an integrated investigation by using international comparative data. These data allow studying the effects of heterogeneity at multiple levels of analysis on a series of individual attitudes. A number of conclusions can be drawn from this effort. First, as the empirical comparison with which this chapter starts shows, survey data may be used to generate useful information about the social structure of countries and regions. Such a research strategy is attractive in a number of ways. For one thing, as international surveys aim at generating international comparative data, researchers gathering the data put a lot of effort in defining and measuring the data unequivocally. Therefore, measurement errors due to differences in meaning of the variables should be reduced as much as possible. Existing datasets, for example the one presented in Alesina et al. (2003) do not have this advantage as they are compiled from different sources, which may increases

Heterogeneity and Human Sociality

87

such errors. Furthermore, it offers a way to update the existing datasets and to examine changes in heterogeneity. While some of the data on which a number of international comparative studies are based go back to the 1960s, the data presented here are quite recent. As such, they offer a means to use more recent data. Finally, it opens the way to investigate longitudinal patterns of heterogeneity, whereas the existing studies are static. Certainly, there is a downside to the use of these survey data. For example, its potential relies on the quality of the data. If certain groups are underrepresented, this will lead to unreliable estimates of the level of heterogeneity. Furthermore, the number of countries included in the dataset investigated here is smaller than in the one used by Alesina et al. (2003). Nevertheless, there is no reason not to extend the present data with other datasets, such as the World Values Survey (WVS) and the International Social Survey Programme (ISSP), to get a wider range of countries. With regard to the effects of heterogeneity, the outcomes are mixed. The expectation found in the Putnam studies, stating that heterogeneity decreases interpersonal trust, was neither confirmed at the national nor at the regional level. The implications of these findings need to be considered. Clearly, one may argue that the data were simply not accurate enough to find the proposed association. Even though that may be true, if we assume that there is some merit in the analyses provided here, it does shed a light on the theoretical expectations and the empirical outcomes of previous studies. To begin with, it underlines the importance of the level at which heterogeneity is assumed to affect interpersonal trust. If the Putnam studies are correct, the present analysis shows that heterogeneity affects interpersonal trust at low levels of aggregation and that the effect disappears as the analyses move to higher levels of aggregation. This may confirm the idea that the neighbourhood is indeed an important point of reference when it comes to trusting others in society. Another interpretation should also be considered. Whereas the Putnam study is based on data from the US while the current analyses are based on European data, the possibility that heterogeneity plays out differently in these two parts of the world cannot be excluded (see for example the studies by Hooghe, Reeskens, Stolle and Trappers (2009) and Gesthuizen, van der Meer and Scheepers (2009) that do investigate the Putnam thesis across European countries and arrive at different conclusions than the original study). Finally, a point worth mentioning is that interpersonal trust is positively related to GDP per capita. This finding shows that country characteristics do explain the level of trust that people have in others, but heterogeneity turns out not to be among them. In addition to that, the

88

Chapter Three

expectations based on the Alesina branch of heterogeneity studies receive some support. However, it cannot be concluded that the empirical support is conclusive. Most and for all, the original hypothesis in this part of the literature concentrates on national level heterogeneity, for which no support was found in the current analyses. When looking at the effects of regional level heterogeneity, opposite outcomes emerge: institutional trust is higher and support for redistribution is lower. Somehow, this seems to suggest that the public views the government as a reliable and legitimate actor as heterogeneity is higher, but they do not seem to demand a more extensive welfare state. This latter point means that we do not find evidence for a compensation hypothesis type of logic that was found in a study where the stock of migrants within an economic sector was taken into account (Burgoon, Koster and van Egmond 2012). The first point means that the public does expect something from the government, but for now it remains an open question what that is. All in all, this chapter shows that heterogeneity may affect human sociality. Nevertheless, it also shows that the present research is far from conclusive and needs additional work in all kinds of directions. For starters, more theoretical progress is required to understand the effects of heterogeneity at different levels of analysis. To date, the focus has been too much on the direction of the relationship rather than its theoretical justification. In addition to that, the theoretical mechanisms need to be explicated and investigated. For a full understanding, controlled experiments will be extremely valuable. The same holds for qualitative accounts, which may range from participant observation to an in-depth content analysis to understand processes such as stereotyping in heterogeneous societies. In sum, for the moment, we have patterns and regularities, but the whole range of social scientific methods should be used to understand them.

References Akerlof, G.A. (1997). Social distance and social decisions. Econometrica, 65, 1005-1027. Alesina, A., Devleeschauwer, A., Easterly, W., Kurlat, S., and Wacziarg, R. (2003). Fractionalization. Journal of Economic Growth, 8, 155-194. Alesina, A., and Glaeser, E.L. (2004). Fighting poverty in the US and Europe. A world of difference. Oxford: Oxford University Press.

Heterogeneity and Human Sociality

89

Anderson, C.J., and Paskeviciute, A. (2006). How linguistic and ethnic heterogeneity influence the prospects for civil society: A comparative study of citizenship behavior. Journal of Politics, 68, 783-802. Blau, P.M. (1977). Inequality and heterogeneity: A primitive theory of social structure. New York: Free Press. Bogardus, E.S. (1925). Measuring social distances. Journal of Applied Sociology, 9, 299-308. Burgoon, B., Koster, F., and van Egmond, M. (2012). Support for redistribution and the paradox of immigration. Journal of European Social Policy, 22, 288-304. Brooks, C., and Manza, J. (2006). Why do welfare states persist? Journal of Politics, 64, 816-827. Coleman, J.S. (1990). Foundations of social theory. Cambridge: The Belknap Press. Deutsch, K.W. (1966). Nationalism and social communication: An inquiry into the foundations of nationality. Cambridge: MIT Press. Easterly, W., and Levine, R. (1997). Africa's growth tragedy: Policies and ethnic divisions. Quarterly Journal of Economics, 112, 1203-1250. European Social Survey (2000-2010). European Social Survey. Bergen: European Social Survey Data Archive. Eurostat (2007). Regions in the European Union. Nomenclature of territorial units for statistics. Luxembourg: Office for Official Publications of the European Communities. Fearon, J.D. (2003). Ethnic and cultural diversity by country. Journal of Economic Growth, 8, 195-222. Gesthuizen, M., van der Meer, T., and Scheepers, P. (2009). Ethnic diversity and social capital in Europe: tests of Putnam's thesis in European countries. Scandinavian Political Studies, 32, 121-142. Goldstein, H. (2003). Multilevel statistical models. London: Edward Arnold. Gross, N. (2009). A pragmatist theory of social mechanisms. American Sociological Review, 74, 358-379. Habyarimana, J., Humphreys, M., Posner, D.N., and Weinstein, J. (2007). Why does ethnic diversity undermine public goods provision? American Political Science Review, 101, 709-25. Hardin, R. (1995). Self-interest, group identity. Nationalism and Rationality, 14-42. Hedström, P., and Swedberg, R. (1998). Social mechanisms. An analytical approach to social theory. Cambridge: Cambridge University Press. Hedström, P., and Ylikoski, P. (2010). Causal mechanisms in the social sciences. Annual Review of Sociology, 36, 49-67.

90

Chapter Three

Henrich, J., Boyd, R., Bowles, S., Camerer, E., Fehr, E., and Gintis, H. (2004). Foundations of human sociality. Oxford: Oxford University Press. Hooghe, M., Reeskens, T., Stolle, D., and Trappers, A. (2009). Ethnic diversity and generalized trust in Europe A cross-national multilevel study. Comparative Political Studies, 42, 198-223. Koster, F. (2013). Sociality in diverse societies. A regional analysis across European countries. Social Indicators Research, 111, 579-601. Koster, F., Goudriaan, H., and van der Schans, C. (2009). Shame and punishment. An international comparative study on the effects of religious affiliation and religiosity on attitudes to offending. European Journal of Criminology, 6, 481-495. Leigh, A. (2006). Trust, inequality and ethnic heterogeneity. Economic Record, 82, 268-280. Mau, S., and Burkhardt., C.(2009). Migration and welfare state solidarity in Western Europe. Journal of European Social Policy, 19, 213-229 Miguel, E., and Gugerty, M. K. (2005). Ethnic diversity, social sanctions, and public goods in Kenya. Journal of Public Economics, 89, 23252368. Park, R.E. (1924). The concept of social distance. Journal of Applied Sociology, 8, 339-344. Putnam, R.D. (2007). E Pluribus Unum: diversity and community in the twenty-first century. Scandinavian Political Studies, 30, 137-174. Smith, M.G. (1960). Social and cultural pluralism. Annals of the New York Academy of Sciences, 83, 763-785. Snijders, T.A.B. (2003). Multilevel analysis. In M.S. Lewis-Beck, A.E. Bryman, and T.F. Liao (eds.), The SAGE encyclopedia of social science research methods (673-77). London: Sage. Snijders, T.A.B., and Bosker, R. (1999). Multilevel analysis: An introduction to basic and advanced multilevel modeling. London: Sage. Stolle, D. (1998). Bowling together, bowling alone: The development of generalized trust in voluntary associations. Political psychology, 497525. Tolsma, J., van der Meer, T., and Gesthuizen, M. (2009). The impact of neighborhood and municipality characteristics on social cohesion in the Netherlands. Acta Politica, 44, 286- 313. Yamagishi, T., and Mifune, N. (2008). Does shared group membership promote altruism? Fear, greed, and reputation. Rationality and Society, 20, 5-30.

Heterogeneity and Human Sociality

91

Appendix 1. Heterogeneity per region, NUTS level 1 Country Austria

Belgium

Bulgaria

Croatia

Cyprus Czech Republic Denmark Estonia Finland France

Nuts 1 region Ostösterreich Südösterreich Westösterreich Vlaams Gewest Région de BruxellesCapitale/Brussels Hoofdstedelijk Gewest Région Wallonne Severna i iztochna Bulgaria Yugozapadna i yuzhna centralna Bulgaria Sjeverozapadna Hrvatska Sredisnja i Istocna (Panonska) Hrvatska Jadranska Hrvatska Cyprus

N 2926 1513 2479 4451

Origin 0.17 0.12 0.13 0.11

Language 0.09 0.06 0.04 0.10

Religion 0.26 0.20 0.16 0.17

466

0.41

0.39

0.56

2318

0.20

0.07

0.19

1927

0.03

0.29

0.36

1703

0.02

0.21

0.36

413

0.14

0.00

0.11

420

0.21

0.01

0.18

560 2210

0.13 0.13

0.01 0.04

0.12 0.03

Czech Republic

6400

0.06

0.02

0.25

Denmark Estonia Manner-Suomi Åland Île de France Bassin Parisien Est Bassin Parisien Ouest Nord - Pas-de-Calais Est Ouest Sud-Ouest Sud-Est Méditerranée

6108 5167 6259 1854 1109 583 714 546 734 1014 925 952 791

0.11 0.32 0.05 0.06 0.31 0.10 0.10 0.09 0.13 0.07 0.11 0.16 0.26

0.04 0.43 0.11 0.15 0.20 0.08 0.12 0.07 0.20 0.04 0.05 0.07 0.08

0.13 0.54 0.07 0.08 0.46 0.09 0.12 0.20 0.27 0.13 0.18 0.25 0.23

92 Germany

Greece

Hungary

Iceland Ireland Italy

Latvia Luxembourg Netherlands

Norway

Chapter Three Schleswig-Holstein Hamburg Niedersachsen Bremen Nordrhein-Westfalen Hessen Rheinland-Pfalz Baden-Württemberg Bayern Saarland Berlin Brandenburg Mecklenburg-Vorpommern Sachsen Sachsen-Anhalt Thüringen Voreia Ellada Kentriki Ellada Attiki Nisia Aigaiou, Kriti Közép-Magyarország Dunántúl Alföld és Észak Iceland Ireland Nord Ovest Nord Est Centro (IT) Sud (IT) Isole (IT) Latvia Luxembourg (Grand-Duché) Noord-Nederland Oost-Nederland West-Nederland Zuid-Nederland Oslo og Akershus Hedmark og Oppland Sør-Østlandet Agder og Rogaland Vestlandet Trøndelag Nord-Norge

295 207 854 82 2028 614 440 1112 1356 107 698 647 535 1123 693 665 2400 1496 2457 691 1511 2122 2612 579 6132 591 468 521 799 357 1980 3187 878 1718 3635 1681 1530 572 1291 1000 1292 672 738

0.13 0.22 0.15 0.13 0.20 0.25 0.15 0.22 0.19 0.18 0.15 0.05 0.06 0.06 0.04 0.05 0.13 0.11 0.20 0.15 0.07 0.04 0.03 0.06 0.16 0.04 0.08 0.06 0.03 0.02 0.24 0.43 0.08 0.13 0.20 0.13 0.21 0.08 0.15 0.14 0.10 0.11 0.08

0.07 0.13 0.06 0.07 0.11 0.13 0.07 0.11 0.10 0.07 0.10 0.05 0.03 0.03 0.01 0.01 0.08 0.04 0.09 0.06 0.00 0.02 0.00 0.01 0.05 0.16 0.28 0.12 0.29 0.34 0.42 0.50 0.28 0.04 0.07 0.04 0.14 0.04 0.07 0.08 0.08 0.11 0.07

0.29 0.45 0.43 0.28 0.58 0.54 0.53 0.59 0.46 0.43 0.58 0.35 0.38 0.31 0.30 0.44 0.07 0.05 0.07 0.07 0.44 0.31 0.50 0.16 0.10 0.05 0.03 0.03 0.02 0.05 0.70 0.43 0.49 0.61 0.67 0.26 0.30 0.09 0.32 0.20 0.13 0.14 0.11

Heterogeneity and Human Sociality Poland

Portugal Romania

Slovakia Slovenia Spain

Sweden

Switzerland

Turkey

Centralny Poludniowy Wschodni Pólnocno-Zachodni Poludniowo-Zachodni Pólnocny Continente (PT) Macroregiunea unu Macroregiunea doi Macroregiunea trei Macroregiunea patru Slovakia Slovenia Noroeste Noreste Comunidad de Madrid Centro (ES) Este Sur Canarias (ES) Östra Sverige Södra Sverige Norra Sverige Région lémanique Espace Mittelland Nordwestschweiz Zürich Ostschweiz Zentralschweiz Ticino Istanbul Bati Marmara Ege Dogu Marmara Bati Anadolu Akdeniz Orta Anadolu Bati Karadeniz Dogu Karadeniz Kuzeydogu Anadolu Ortadogu Anadolu Güneydogu Anadolu

1580 1428 1393 874 652 1239 8152 502 647 561 436 5061 5723 1143 735 971 998 2140 1574 283 2728 3412 1564 1465 1642 1317 1187 1437 544 212 533 88 277 280 214 314 67 80 31 68 129 335

93 0.01 0.02 0.02 0.04 0.07 0.04 0.11 0.00 0.02 0.01 0.01 0.06 0.15 0.09 0.11 0.22 0.07 0.17 0.11 0.16 0.24 0.19 0.11 0.39 0.24 0.30 0.30 0.31 0.28 0.40 0.01 0.00 0.05 0.08 0.01 0.01 0.00 0.02 0.00 0.00 0.00 0.01

0.00 0.01 0.02 0.00 0.00 0.00 0.02 0.39 0.04 0.00 0.09 0.20 0.04 0.38 0.13 0.06 0.03 0.44 0.04 0.11 0.13 0.10 0.05 0.39 0.49 0.27 0.26 0.24 0.21 0.35 0.08 0.02 0.01 0.04 0.04 0.06 0.06 0.00 0.00 0.55 0.50 0.62

0.02 0.02 0.06 0.03 0.03 0.03 0.07 0.51 0.09 0.04 0.20 0.32 0.13 0.04 0.10 0.11 0.05 0.17 0.08 0.11 0.36 0.26 0.16 0.54 0.50 0.58 0.51 0.57 0.37 0.16 0.05 0.02 0.00 0.10 0.01 0.00 0.00 0.02 0.00 0.00 0.00 0.00

94 United Kingdom

Chapter Three North East (ENGLAND) North West (ENGLAND) Yorkshire and The Humber East Midlands (ENGLAND) West Midlands (ENGLAND) South West (ENGLAND) Eastern London South East Wales Scotland Northern Ireland

Total Sources: European Social Survey 1-4

411 1035 749 666 727 783 843 760 1180 510 752 279 171,619

0.06 0.13 0.14 0.10 0.18 0.11 0.21 0.45 0.18 0.06 0.11 0.09 0.12

0.02 0.06 0.06 0.02 0.09 0.02 0.05 0.23 0.04 0.08 0.04 0.01 0.12

0.54 0.59 0.57 0.48 0.56 0.43 0.51 0.74 0.51 0.51 0.52 0.49 0.26

CHAPTER FOUR DIVERSITY, SIMILARITY, AND SOLIDARITY: RESULTS OF A VIGNETTE STUDY MAARTEN BERG AND FERRY KOSTER

Introduction Solidarity and sharing behaviour are necessary elements of viable, wellfunctioning societies. Solidarity can take different forms, and a relevant distinction is between ‘public’ (formal) solidarity and ‘private’ (informal) solidarity. The prototypical example of public solidarity is the financial support of certain societal groups through the mechanism of the welfare state. Private solidarity refers to voluntary helping behaviour directed at concrete individuals, e.g. taking out the garbage for an elderly neighbour (e.g. de Beer and Koster, 2009). Solidarity, whether measured in terms of time, money, consumer goods, practical assistance or emotional support, has an economic aspect in the sense that it is always scarce. As all resources are limited, solidarity can never be unrestricted. This raises questions about the degree of our potential for solidary behaviour, but also about the allocation of our solidarity. In this chapter we address the latter question. We study what types of potential beneficiaries have the strongest appeal on our sense of solidarity. When forced to choose, towards whom do people feel most solidary? To study this question, we use a vignette approach (Rossi and Nock, 1982). Respondents are asked to choose between two fictive beneficiaries who are both described on several dimensions (‘attributes’). Although reading a vignette is different from experiencing a stimulus or action in real life, the vignette approach is thought to reveal the values and perceptions of respondents. Moreover, this is done in a way that is less vulnerable to socially desirable responses than the standard survey

96

Chapter Four

approach. We will elaborate on the vignette approach, its pros and cons, and its implementation below. The theoretical starting point of this paper is the assumption that ‘diversity’ in a society or a community affects the potential for and the direction of solidarity. Although it is not always clear and it cannot be fully determined what the relevant aspects of diversity are, common ‘distinctions’ between people that are mentioned both in the literature and in policy debates relate to the old and the young and different ethnic groups (de Beer and Berg, 2012). These attributes coincide with some of the ones used in the current chapter. In this chapter we conceptualize diversity not as a characteristic of a group, but as a dyadic relationship between two individuals. Although solidary behaviour necessarily takes place within a broader social context, in this chapter we are interested in the attributes of potential beneficiaries that trigger solidarity and how these attributes relate to attributes of the contributor.

Hypotheses When we think of solidarity as primarily ‘other-regarding’, we should place great importance on the attributes of the beneficiary. His/her neediness is the main reason that solidarity exists in the first place. We hypothesize that needier beneficiaries will generally appeal more strongly to our sense of solidarity (van Oorschot, 2006). This might have to do with internalized social norms to help the needy or with feelings of empathy. In study 1 of this chapter (on public solidarity), we compare different types of entitlements (retirement, unemployment, disability and welfare benefits). We assume that disabled persons are generally perceived to be the most needy and, therefore, will evoke most solidarity. We assume that people on pension benefits are perceived to be the least needy and will evoke the least solidarity1. Study 1 also compares people with varying numbers of under-aged children. We assume that people with many underaged children are perceived to be needier than people with few under-aged children and will evoke more solidarity. In the context of study 2 (on private solidarity), we assume that elderly people are perceived as relatively needy (e.g. because of physical limitations) and will receive the most solidarity. 1

Note that in the Netherlands the term ‘pension’ refers to a supplementary occupational (second pillar) pension and not to state pension (first pillar).

Diversity, Similarity, and Solidarity: Results of a Vignette Study

97

Apart from classifying beneficiaries in terms of their neediness, we must also look at their ‘deservingness’ (e.g. van Oorschot, 2000). For example, it seems likely that people are generally more willing to help others who are not to blame for their needy situation than those who brought it onto themselves. Deservingness also depends on past contributions by the beneficiary. In study 1, we compare people with varying numbers of working years. We assume that people with a lot of working experience are perceived to be more deserving than people with lesser working experience and predict that the number of years of working experience contributes to the willingness of others to show solidarity. Helping behaviour is arguably not only other-regarding, but also (partly) self-regarding in some (or even most) situations (see chapter 1). Helping might be functional in the sense that it encourages reciprocation, thereby strengthening the social fabric of communities. We expect people to be more willing to help people who are prepared to make some future contribution (if possible) in return (Fong, Bowles and Gintis, 2003). In study 1, we compare people who are willing to do volunteer work with people who are not. We predict most solidarity with people who are very willing to volunteer. In study 2, we compare people who are offering to reciprocate favours in the future with people who do not explicitly offer help in the future. In this chapter we are especially interested in the role of ethnic diversity. There are several theoretical reasons to expect people from different ethnic backgrounds to be less inclined to help one another than people from the same ethnic background (see chapter 2). For example, they might have greater difficulty in identifying with one another. Traditionally, left-wing political parties (and their supporters/voters) in the Netherlands have more favourable attitudes towards multiculturalism and (the role of) migrants than right-wing parties (Koopmans, 2013). Therefore, we expect right-wing voters to react more strongly to the ethnicity of the potential beneficiaries than left-wing voters. That is, they will be more unfavourable when the beneficiary has a different ethnic background. The fact that some groups of voters are expected to respond differently to different ethnic groups illustrates a more general notion. Solidarity needs not be just a unilateral relationship directed at and affected by (characteristics of) the beneficiary. It might be truly relational in the sense that both the attributes of the contributor and the beneficiary matter, as

98

Chapter Four

well as the ‘distance’ between them. We hypothesize that the differences between contributor and beneficiary have an independent effect up and above the characteristics of the beneficiary per se. In this chapter we look at both public solidarity (provided through the welfare state) and private solidarity (shown voluntarily between individuals). This distinction might coincide with different psychological mechanisms (Koster, 2009). Social norms and social pressure might be of greater importance in the context of the welfare state. On the other hand, personal contact and visibility might be more important for private solidarity than for public solidarity and may trigger empathy and direct reciprocity. Therefore, we hypothesize that some different results will emerge for the public and the private solidarity studies. The following hypotheses are tested. Hypothesis 1: All other things being equal, people express more solidarity towards people who are in need. More specifically, we predict most solidarity towards disabled persons and the least solidarity towards people on pension benefits. Also, we expect more solidarity towards people with relatively many under-aged children. In the context of private solidarity, we expect more solidarity towards the elderly. Hypothesis 2: All other things being equal, people express more solidarity towards people who are deserving. More specifically, we predict a positive relationship between the number of working years of a beneficiary and the solidarity that s/he encounters. Hypothesis 3: All other things being equal, people express more solidarity towards people who are willing to reciprocate. More specifically, we predict that people who are willing to do volunteer work will encounter more solidarity. In the context of private solidarity, we predict the same for people who offer to reciprocate a favour. Hypothesis 4: All other things being equal, people express more solidarity towards people with the same ethnic background than towards people with a different ethnic background. Hypothesis 5: Persons with a ‘right-wing’ political orientation will be more sensitive to ethnic differences and discriminate more strongly towards people with a different ethnic background than left-wing persons.

Diversity, Similarity, and Solidarity: Results of a Vignette Study

99

Besides these hypotheses, we explore (post hoc) the relative importance of the several attributes and discuss differences between public and private solidarity.

Methods The vignette approach A vignette is a short, carefully constructed description of a person, object or situation, representing a systematic combination of characteristics. McFadden et al (2005) describe vignettes as “hypothetical situations that can be used to elicit preferences, judgments or anticipated behaviour”. In this chapter a vignette refers to a person. An example of a vignette from our study on public solidarity (as shown to our respondents) is: x x x x x x x x

Male 47 years old Moroccan background 22 years working experience On welfare 4 under-aged children Living in your neighbourhood Willing to do unpaid work (volunteering)

There are many different variations of the vignette approach (for a review: Atzmüller and Steiner, 2010). Sometimes respondents have to evaluate only one vignette, whereas in other cases multiple vignettes are used. In the current chapter, respondents deal with several pairs of vignettes. From each pair, they have to choose one. When studying solidarity from a standard survey approach, we could ask respondents to what extent they are willing to support certain groups (e.g. elderly people, people from different ethnic backgrounds, etc.) and to formulate their answer on a verbal or numerical scale. This approach has several limitations. First, the question may seem artificial, because of a lack of contextual information. People might not know if they want to express solidarity towards ‘an elderly person’, because they do not know who this person is. Is the person a male or a female? Is s/he a nice person? Et cetera. The vignette approach is more realistic in this respect, as it presents hypothetical persons in a multidimensional manner.

100

Chapter Four

Secondly, the multidimensional approach is ideal for studying interaction effects (Wason et al., 2002). In this study, we include such interactions by examining how solidarity depends on the combination of characteristics of the hypothetical person and those of the respondents. Thirdly, the standard survey method can be sensitive to socially desirable (or strategic) statements (e.g. de Wolf and van der Velden, 2001). For example, persons with negative stereotypes about migrants might actually express positive attitudes when they want to appear (to the experimenter or even to themselves) open-minded and tolerant. This kind of ‘impression management’ is much harder with the vignette approach, as the respondents have to deal with complex stimuli and several dimensions at the same time. They are forced to answer mainly intuitively (rather than cognitively and deliberately). This advantage of the vignette method is especially valuable when sensitive topics are addressed (such as attitudes towards immigrants). Although some impression management might still be possible, one would have to restrict it to one or two dimensions. We expected that the risk of impression management would be greatest with respect to the ethnic dimension. For that reason, we decided to decrease the saliency of this dimension (as explained below). A last advantage of the vignette approach compared to survey studies is that the vignette method is ideally suited for assessing the relative weight of different aspects that are relevant for the decision or evaluation at hand. A big advantage of the vignette approach over laboratory experiments is that the former method is much less time (and money) consuming. This makes it possible to (drastically) increase the number of subjects and to study several dimensions simultaneously. In other words: when we want to study multiple dimensions simultaneously, the experimental method is often simply infeasible.

Participants The data were obtained from the LISS-panel (Longitudinal Internet Studies for the Social Sciences), an Internet panel survey that aims to be representative for the Dutch population2. The LISS panel consists of 5,000 households, comprising 8,000 individuals (although in our specific study 5,605 people participated). Some distributional characteristics are 2

See Knoef and De Vos (2009) for an elaborate discussion of the representativeness of the panel.

Diversity, Similarity, and Solidarity: Results of a Vignette Study

101

presented in Table 1. The panel is based on a true probability sample of households drawn from the population register by Statistics Netherlands. Every month, panel members complete online questionnaires of about 15 to 30 minutes in total. They are paid for each completed questionnaire. One member in the household provides the household data and updates this information at regular time intervals Table 1: Distribution of some characteristics of 5,605 respondents and of the total Dutch population3 Variable

Value

Sex

Male Female 18-29 years 30-39 years 40-49 years 50-59 years 60-69 years 70 and older Dutch

Age

Ethnic background

N

Other Missing

2,594 3,011 889 747 965 1,102 1,183 719 5,081

% in the sample 46.3 53.7 15.9 13.3 17.2 19.7 21.1 12.8 91.1

% in the population 49.5 50.5 18.7 15.3 19.3 17.7 15.0 14.1 79.1

4994 25

8.9 --

20.9 --

Researchers from different universities can incorporate their studies or questions in the LISS-panel. Moreover, they can use the data on the background characteristics of the respondents. In our case, this enabled us to study not only how vignette conditions relate to respondents’ solidarity, but also how these effects interact with the characteristics of the contributor (respondent).

Study 1: public solidarity All participants were given ten pairs of vignettes, the first five pairs relating to public solidarity and the last five pairs to private solidarity.

3

Data are from Statistics Netherlands (CBS), 2012. Of the 499 non-Dutch respondents, 60 had a Surinamese background and 33 a Moroccan background.

4

Chapter Four

102

Regarding public solidarity, the respondents were asked to imagine the Dutch government having to balance the budget and cut spending on social welfare. Given that context, the question was: What person should be spared most should the government decide to cut its expenditures?5 The ‘public’ vignettes consisted of eight variables: -

-

-

Sex (two categories: ‘male’ and ‘female’). Age (all values between 18 and 90). Ethnicity (five categories: ‘autochthone Dutch’, ‘born in Morocco’, ‘Moroccan background’, ‘born in Surinam’, ‘Surinamese background’6). Number of years of working experience (all values between 0 and ‘age minus 18’, with a maximum of 50). Type of entitlement (four categories: ‘welfare’, ‘unemployment benefit’, ‘disability benefit’, ‘pension’). People above 65 were automatically placed in the ‘pension’ category, people under 65 in one of the other three categories.7 Number of under-aged children (all values between 0 and 6). Distance (two categories: ‘lives in your neighbourhood’, ‘lives in another part of the Netherlands’). Willingness to do volunteer work (three categories: ‘is eager to do volunteer work’, ‘might be interested to do volunteer work’, ‘prefers not to do volunteer work’).

The theoretical number of vignettes (combining all values of all factors) is enormous (even compared to the large number of respondents). Therefore, we constructed 500 vignettes independently from each other, using a random generator. In principle, the different values of the independent variable had an equal chance to be included in a specific vignette.8 There were, however, some exceptions: 5

It was a forced choice question. Maybe respondents felt like both persons (or neither person) should be spared. Still, they had to make a choice. 6 For the analyses, we recoded the five categories into three categories: Dutch, Moroccan, Surinamese. 7 This means that the different independent variables were not perfectly independent from one another. This creates some multicollinearity in the analyses, but we considered this problem to be smaller than using highly unrealistic vignettes. 8 This did not mean that the numbers had to match perfectly. We did not strive for exactly 250 male and 250 female vignettes, for example.

Diversity, Similarity, and Solidarity: Results of a Vignette Study

-

-

103

For ethnicity, we decided to use a 70% chance for the vignette to be ‘autochthone Dutch’ and to use equal chances of 7.5% for the four other categories of this variable. We wanted respondents to face sufficient pairs of two autochthone Dutch, in order to decrease the saliency of the ethnic dimension in this study. With only 1/3 (or even 1/5) of the vignettes being Dutch, it would have been very clear to the respondents that we were most interested in the role of ethnic diversity. Still, compared to the actual ethnic distribution within the Netherlands, we actually included a lot of people from Morocco and Surinam in our vignette population. For the number of under-aged children, we used different distributions for different age groups (as shown in Table 2).

Table 2: Distribution of numbers of under-aged children by age group (%)

Age group 18-19 20-29 30-39 40-49 50-59 60-69 70-79 80-90

0 95 50 40 30 40 80 90 100

Number of under-aged children 1 2 3 4 5 20 15 15 15 10 5 0

0 20 15 15 15 5 5 0

0 5 10 10 10 5 0 0

0 5 10 10 10 0 0 0

5

6

0 0 5 10 5 0 0 0

0 0 5 10 5 0 0 0

Although we took measures (using certain distributions), to guarantee realistic vignettes (we wanted to avoid 87 year old women with 6 underaged children), we loosely checked the vignettes after they were generated. No vignettes had to be excluded.9 For each respondent, ten different vignettes (five pairs) were randomly drawn from the population of 500. The eight attributes were placed vertically and two vignettes (a pair) were placed next to one another. For every respondent, the order of the eight attributes was identical for each 9

When asked to comment on the study, very few (still some) respondents mentioned unrealistic vignettes. However, we have to realize that atypical people exist in the real world as well.

Chapter Four

104

pair (in order to avoid confusion). The order varied between subjects. It was possible that a respondent had to choose between two pairs that were identical on one or more dimensions, although it was impossible that they had to choose between the exact same vignettes.

Study 2: private solidarity The study on private solidarity dealt with helping behaviour in the neighbourhood. Again, there were five pairs of vignettes. The recurring question was what neighbour would be most likely to get the help from the respondent. In contrast to the study on public solidarity, in the private solidarity study five different helping situations were used for the five pairs of vignettes. Question 1: Your neighbour rings at your bell at night and asks if you are willing to look after his/her (grand)child for about an hour. Which neighbour are you most inclined to help? Question 2: Your neighbour rings at your bell at night and asks if you can help buying groceries for a party (as there are too many groceries for one person to carry). Which neighbour are you most inclined to help? Question 3: Your neighbour rings at your bell at night and asks if you can help with some paper work that gives him/her some trouble. Which neighbour are you most inclined to help? Question 4: Your neighbour rings at your bell at night and asks if you can help take out the household refuse. Which neighbour are you most inclined to help? Question 5: Your neighbour rings at your bell at night and asks if you can help him/her to construct a small cabinet (since it is much easier for two persons to do this). Which neighbour are you most inclined to help? The vignettes consisted of five variables (the first three were identical to the ones used in the study on public solidarity): -

Sex (two categories: ‘male’ and ‘female’). Age (all values between 18 and 90).

Diversity, Similarity, and Solidarity: Results of a Vignette Study

-

-

105

Ethnicity (five categories: ‘autochthone Dutch’, ‘born in Morocco’, ‘Moroccan background’, ‘born in Surinam’, ‘Surinamese background’).10 Encounters (two categories: ‘run into him/her in the neighbourhood frequently’, ‘seldom run into him/her in the neighbourhood’). Reciprocity (two categories: ‘offers to return the favour in the future’, ‘says nothing about possibly returning the favour in the future’).

Although the situations differ, and hence it is possible that the effects of the vignette conditions differ across these situations, separate analyses and analyses in which the situations are combined lead to similar results (not shown here). Therefore, the results of the combined analyses are shown here.

Procedure Data were gathered in the beginning of 2012. Participants filled out the LISS panel questionnaire, of which our study was only a part. When coming to our part of the questionnaire, all participants (after reading the instructions) first dealt with the vignette pairs on public solidarity and then turned to the vignette pairs on private solidarity. For all participants, the same order of questions (helping situations) was used for the private solidarity part. After completing the main part of their task, respondents were asked whether they had based their choices mainly on intuition or on conscious deliberation (evaluating and weighing the different attributes in a systematic way). Subsequently, they answered the question about ethnic identity (already mentioned above). Next, there were several questions about their attitudes regarding the Dutch welfare state (scoring from less to more favourable) and regarding solidarity towards the young, the old, and persons with different cultural backgrounds. Also some questions about media usage were included. Respondents had to indicate how often they watched the NOS journal and RTL nieuws (TV news), or read De Volkskrant and De Telegraaf (newspapers) or nu.nl (a news website).

10

The same distribution was used as in the study on public solidarity.

106

Chapter Four

Finally, they answered some questions about the study itself. Was it comprehensible, interesting, and so on? They were also able to comment on the study.

Method of analysis In both studies, the dependent variable is dichotomous since respondents choose either vignette 1 or vignette 2. In our analyses, each choice between two vignettes was taken as a separate observation. Thus, an ‘act of solidarity’ is a dyadic relation between the respondent and a (fictive) beneficiary. Furthermore, we have observations nested within individuals, as respondents answered 5 vignettes for both kinds of solidarity. Therefore, we applied logistic multilevel regression analysis. The total number of observations is 56,050 (distributed among 5,605 persons and two studies). 16 missing observations from the public solidarity study and 91 observations from the private solidarity study could not be analysed. Since respondents compared several pairs of vignettes, their choice may have been affected by fatigue. To make sure that this does not affect the outcomes, we tested, using interaction effects for each vignette condition, whether there were such fatigue effects (not shown here). This analysis shows that the results are unaffected and remain the same. Our independent variables were either continuous (e.g. age) or nominal (e.g. ethnic background). The continuous variables (using the difference between vignette 1 and vignette 2) were easy to include in our multivariate regression models and to interpret in combination with our dependent variable. The dataset includes information about the fictive person displayed on the vignette and information about the respondents. By combining these two sources of information, it is possible to calculate the similarity between the person on the vignette and the respondent. We approached this as follows:  With regard to the effect of ethnic identity, we are interested in investigating whether similarity in ethnic background matters. To assess this, it is necessary to combine characteristics of the vignettes with the ethnic background of the respondents. For the ethnic identity dimension, we created a dummy variable that scored 0 when the persons in the vignettes both had a different ethnic background than the respondent, and scored 1 when the person on the vignette had the same ethnicity as the respondent.

Diversity, Similarity, and Solidarity: Results of a Vignette Study

107

 Age similarity was approached in the same way as ethnic identity. Nevertheless, since age was measured as a continuous variable instead of as categorical variable, we did not calculate a dummy variable. To assess age similarity (or distance in terms of years), the absolute difference between the age of the respondent and the age of the person on the vignette was calculated. To account for the fact that the effect of age may be curvilinear, also the squared root of absolute age was included in the analyses. Finally, to investigate whether the individual characteristics matter for solidarity preferences, interaction effects were added to the analyses. For example, to see whether these preferences are affected by the political orientation of respondents, interactions were included between the answers on this variable and each vignette condition, and to investigate whether there is a gender bias in solidarity, we added an interaction term between the gender of the respondent and the gender of the person on the vignette. Respondents had to indicate on a five-point scale whether or not they thought it was difficult to choose between the vignettes. Recoding the answers to three categories gave the following distribution: easy (60.6%), neither easy nor difficult (18.0%), and difficult (21.4%). Moreover, 76.7% of the respondents thought the questions were clear, 16.3% were ambivalent and 7.0% thought that they were not clear. Considering these findings, it is fair to say that the results below do in fact reflect the (conscious or unconscious) attitudes and preferences of the panel.

Study 1: Public solidarity The descriptive statistics of the study on public solidarity are shown in Table 3 and based on 28,009 observations/choices. Respondents chose more female (fictive) persons than male persons (compared to their prevalence in the vignette population): 52.1% versus 47.9%. The difference is significant: Ȥ2(1)=98.630, p