When marriage ends: economic and social consequences of partnership dissolution

When Marriage Ends

When Marriage Ends Economic and Social Consequences of Partnership Dissolution

Edited by

Hans-Jürgen Andreß Faculty of Management, Economics and Social Sciences, University of Cologne, Germany

Dina Hummelsheim Max Planck Institute for Foreign and International Criminal Law, Freiburg, Germany

Edward Elgar Cheltenham, UK • Northampton, MA, USA

© Hans-Jürgen Andreß and Dina Hummelsheim 2009 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical or photocopying, recording, or otherwise without the prior permission of the publisher. Published by Edward Elgar Publishing Limited The Lypiatts 15 Lansdown Road Cheltenham Glos GL50 2JA UK Edward Elgar Publishing, Inc. William Pratt House 9 Dewey Court Northampton Massachusetts 01060 USA

A catalogue record for this book is available from the British Library Library of Congress Control Number: 2009921832

ISBN 978 1 84844 193 4 (cased) Printed and bound in Great Britain by MPG Books Group, UK

Contents List of contributors Acknowledgements

vii xii

Introduction: economic and social consequences of partnership dissolution – what do we know and what are we looking for? Hans-Jürgen Andreß and Dina Hummelsheim PART I

1

THEORETICAL, METHODOLOGICAL AND MEASUREMENT ISSUES

Conceptualization and measurements of institutional contexts: a review Haya Stier

PART II

1

29

SINGLE COUNTRY STUDIES

2

Germany: will the male breadwinner model survive? Dina Hummelsheim

3

Belgium: economic hardship despite elaborate childcare and leave time programmes Dimitri Mortelmans, Laurent Snoeckx and Peter Raeymaeckers

78

New holes in the safety net? Economic and social consequences of divorce in Denmark M. Azhar Hussain and Olli Kangas

98

4

5

Divergences in the Nordic model: economic consequences of partnership dissolution in Sweden and Finland Heikki Hiilamo

6

Great Britain: ‘things can only get better . . .’ Wendy Sigle-Rushton

7

Marital disruption in Spain: class selectivity and deterioration of economic conditions Lluís Flaquer and Anna Garriga v

51

132 155

178

Contents

vi

8

Low divorce incidence in Greece: facts and figures Haris Symeonidou

PART III 9

10

11

12

COMPARATIVE STUDIES

Female-supportive policies and women’s employment after divorce Maike van Damme and Wilfred Uunk

233

Divorce and housing: a European comparison of the housing consequences of divorce for men and women Caroline Dewilde

263

The economic consequences of partnership dissolution: a comparative analysis of panel studies from Belgium, Germany, Great Britain, Italy and Sweden Hans-Jürgen Andreß, Barbara Borgloh, Miriam Bröckel, Marco Gießelmann and Dina Hummelsheim

286

Marital disruption and economic well-being: poverty, income and lifestyle deprivation Arnstein Aassve, Gianni Betti, Stefano Mazzuco and Letizia Mencarini

PART IV 13

211

CONCLUSION

When marriage ends: results and conclusions Hans-Jürgen Andreß and Dina Hummelsheim

PART V

330

361

DATA APPENDIX

Data appendix

387

Index

407

Contributors Arnstein Aassve obtained his PhD in Economics at Bristol University in 2000 and is now an Associate Professor in Demography at Bocconi University of Milan. He is the deputy director of the Carlo F. Dondena Centre for Research in Social Dynamics. He is a research associate at the Institute of Social and Economic Research (Essex), Statistics Norway, and the Centre for Analysis of Social Exclusion (London School of Economics). His research interests are wide and include demography, applied statistics, micro-econometrics, social policy, simulations and life-course analysis. Hans-Jürgen Andreß studied Social Sciences and Research Methods at the Universities of Frankfurt/Main and Michigan (Ann Arbor). He gained his PhD at the University of Frankfurt/Main in 1983 and is at present Professor for Empirical Social and Economic Research at the University of Cologne. He has published in the field of labour markets, poverty research, social policy, and multivariate statistical methods. Gianni Betti obtained his PhD in Applied Statistics at the University of Florence in 1998 and is currently Associate Professor in Economic Statistics and member of the ‘Centro Interdipartimentale di Ricerca sulla Distribuzione del Reddito C. Dagum’ (CRIDIRE) at the University of Siena, Italy. He has worked for several projects for the World Bank and the European Commission, and has been closely involved with the development of the EU Statistics on Income and Living Conditions (EU-SILC). Barbara Borgloh studied Sociology at the Universities of Hamburg and Bielefeld and holds a Diploma in Sociology from the University of Bielefeld. She has been a Researcher at the University of Bielefeld and published on the economic consequences of divorce and maintenance law. Miriam Bröckel holds a Diploma in Sociology from the University of Bielefeld. She has been a researcher at the University of Bielefeld and published on economic consequences of divorce and subjective well-being after marital disruption. Caroline Dewilde is an Assistant Professor in Sociology at the University of Amsterdam. Prior to January 2009 she was a Postdoctoral Research Fellow with the Research Foundation – Flanders (FWO-Vlaanderen) vii

viii

Contributors

and associated with the Research Centre on Inequality, Poverty, Social Exclusion and the City (OASeS) of the University of Antwerp. Her main research interest is the study of social inequality from a life-course perspective, with special attention to cross-national variations related to the impact of cultural patterns, family structures and institutional arrangements. Her current research focuses on the stratification over the life course in different institutional settings. Lluís Flaquer is currently Professor of Sociology at the Universitat Autònoma de Barcelona. He took a degree in Political Science at the University of Paris Institut d’Études Politiques and then he received his PhD in Law at the Universitat Autònoma de Barcelona. His current research topics include the changing structure of the family, new family forms, family policy and individualization. He was the director of WELLCHI network on ‘The well-being of children: the impact of changing family forms, parents’ working conditions, social policy and legislative measures’, financed under the 6th Framework Programme of the European Union (2004–07). Anna Garriga is currently a PhD candidate at the Department of Political and Social Science in the Universitat Pompeu Fabra in Barcelona. She holds a degree in Political Science from Universitat Autònoma de Barcelona. Her current research topics include the causes and consequences of divorce and cohabitation, and their effects on adults and children. Currently, she is the coordinator of the research team on ‘The effects of marital instability on children’s well-being’ of the project ‘Economic Change, Quality of Life and Social Cohesion’ (EQUALSOC) financed under the European Union’s Sixth Framework Programme (2005–10). Marco Gießelmann studied Sociology and Research Methods at the Universities of Bielefeld and Essex and holds a Diploma in Sociology from the University of Bielefeld. Currently he is a Researcher at the Chair for Empirical Social and Economic Research at the University of Cologne and is preparing a PhD thesis on the working poor phenomenon in Germany and Great Britain. Heikki Hiilamo has received a PhD from the University of Turku in Social Policy and from the University of Tampere in Public Health. He specializes in family policy and poverty issues, and works as a director at the Diaconia University of Applied Sciences in Finland. His recent publications include studies on Nordic family policy and measurement of poverty. Dina Hummelsheim studied Sociology as well as German and Dutch Studies at the Universities of Cologne and Ghent. She recently finished her PhD in Sociology at the University of Cologne. Her dissertation focuses on the

Contributors

ix

impact of family policies and culture on maternal employment in Belgium and West and East Germany. At present she works as a Researcher at the Max Planck Institute for Foreign and International Criminal Law in Freiburg. M. Azhar Hussain obtained his PhD in Economics from the Aarhus School of Business in 2002. Currently he is an Assistant Research Professor at the Department of Economics of the University of Copenhagen. His main research interests include income distribution analysis (inequality, poverty, polarization and deprivation), comparative welfare state studies, and econometric methods. He teaches quantitative methods. Olli Kangas is currently Research Director at the Social Insurance Institution of Finland. He defended his PhD at the University of Helsinki 1991, was nominated as a Visiting Research Fellow at the Swedish Institute for Social Research in Stockholm and appointed as Professor of Social Policy at the University of Turku 1994. From 2004 to 2007 he was Research Professor at the Danish National Institute for Social Research in Copenhagen. His main interest is in causes and consequences of social institutions and comparative methods. Stefano Mazzuco is Assistant Professor of Social Statistics at the Faculty of Statistical Sciences of the University of Padova. He graduated from the University of Padova in 2003, where he obtained his PhD in Statistics Applied to Economics and Social Sciences. His main research interests cover family formation and dissolution, poverty and social exclusion, and longitudinal analysis. He is currently involved in a research project on health inequalities in European Community countries. Letizia Mencarini, PhD, is Associate Professor of Demography at the Political Science Faculty of the University of Turin and she is an associate of the DONDENA Centre for Research on Social Dynamics at Bocconi University in Milan. Her research interests include poverty and demographic dynamics (fertility and migration) both in developed and developing countries, life cycle and family formation, time use and gender issues. She has been involved in a range of national and international research projects and has been published widely in demography, social policy and gender studies. Dimitri Mortelmans, PhD, is Associate Professor in Sociology at the Faculty of Political and Social Sciences of the University of Antwerp. He teaches Qualitative Research Methods and Applied Multivariate Statistics. He is head of the Research Centre for Longitudinal and Life Course Studies (CELLO). His principal research interest lies in the sociology

x

Contributors

of family and labour. His main research topics cover divorce, work–life balance and career patterns. Peter Raeymaeckers has a Master’s degree in Political and Social Sciences at the University of Antwerp (2005). Since October 2006 he has been working at OASeS (Research Centre on Inequality, Poverty, Social Exclusion and the City). His main topics of research concern the measurement of multidimensional poverty in the European Union. Wendy Sigle-Rushton obtained a PhD in Economics from Brown University in 1998 and is currently a Senior Lecturer in Social Policy at the London School of Economics. She is also affiliated with the Economic and Social Research Council’s (ESRC’s) Centre for the Analysis of Social Exclusion. Her research interests include a broad range of issues related to family, gender, social change, and complex inequalities. Recent publications have examined the associations between childhood circumstances and subsequent disadvantage. Laurent Snoeckx studied Political and Social Sciences at the University of Antwerp. Since September 2005 he has been working on a PhD at the Department of Sociology’s Research Centre for Longitudinal and Life Course Studies (CELLO) at the University of Antwerp. His main topics of research cover determinants and consequences of partnership dissolution. Haya Stier received her PhD in Sociology from the University of Chicago in 1990, and her BA and MA degrees in Sociology from Tel Aviv University. She is currently an Associate Professor at the Department of Labour Studies and the Department of Sociology at Tel Aviv University. Haya Stier’s research interests include issues of work, gender and the family, poverty and inequality. In her current research she is focusing on the effect of institutional arrangements on women’s work patterns, the organization of households, and their economic consequences, in a comparative framework. Haris Symeonidou received her PhD from the University of London and is currently Research Director in the Greek National Centre of Social Research. As a demographer with a background in economics, she is a member of the board of directors of the Greek Society of Demographic Studies. She has published several books and articles in national and international journals in the areas of fertility patterns, women’s employment, family policy and social welfare. She was a visiting Professor in the Department of European Studies of the Université Sorbonne-Nouvelle (Paris III) and a former member of the European Union Network on

Contributors

xi

Women’s Employment and of the European Observatory on National Family Policies. Wilfred Uunk is Assistant Professor in Sociology at Tilburg University. He received his doctoral degree in 1996 from the Department of Sociology at the University of Nijmegen with a dissertation on ‘Who marries whom? The role of social origin, education and high culture in mate selection of industrial countries during the twentieth century’. His research interests are social inequality, marriage and the family, ethnic minorities, crosscomparative research, and longitudinal analyses. Maike van Damme is a PhD student in Sociology at Tilburg University. Her dissertation concerns the influence of institutions on the economic consequences of divorce for women. She obtained her MSc in Sociology at Tilburg University in 2004 and subsequently worked at the Social and Cultural Planning Office (SCP) in the Hague. Her research interests include divorce, labour market, welfare state, poverty, and gender inequality.

Acknowledgements We acknowledge funding of our research and for the establishment of our research network from the German Research Foundation (Grant no. AN 210/5-1: The economic risks of separation and divorce in international comparison) and from the Centre for Empirical Research in the Economic and Social Sciences (CERESS) at the University of Cologne. Chapter 11 is a slightly modified version of a 2006 publication in volume 22 of the European Sociological Review (pp. 533–60). Chapter 12 is a slightly modified version of a 2007 publication in volume 170 of the Journal of the Royal Statistical Society (Series A) (pp. 781–99).

xii

Introduction: economic and social consequences of partnership dissolution – what do we know and what are we looking for? Hans-Jürgen Andreß and Dina Hummelsheim WHAT THIS BOOK IS ABOUT Over the last several decades, the probability of divorce and separation among married or cohabiting couples has increased significantly in most European countries, although the level and pace at which this has occurred differs across countries. Much research has been devoted to the causes of this change – for example, to the effects of increased secularization and increased women’s employment on marriage stability. Nevertheless, an equally important area of research is the analysis of the consequences of partnership dissolution. Separation and divorce do not only end unhappy partnerships and open up new perspectives for life. They are also connected with psychological distress, financial strain and many changes in individual living conditions. As we all know, couples are not only bound by love and affection. They live together in one common household, share economic resources, support each other emotionally and practically, try to be good parents to their children and enjoy social relationships with relatives, neighbours and friends. As a consequence, separation and divorce also imply a new organization of personal finances, taking up or increasing employment, applying for financial support from the state, arguing about childcare and custody, and searching for social support and new social relationships. Each of these changes can be problematic in itself, and will thus increase the emotional stress caused by partnership disruption. As we have already mentioned, the causes of divorce are an area of intensive research (see the reviews in Wagner and Weiß, 2006). In contrast, not so much attention is directed to the consequences of divorce, at least in the European context. As the examples show, the consequences of separation and divorce are manifold, and they are noticeable both at the individual and the societal levels. This book deals with consequences of partnership 1

2

When marriage ends

dissolution at the individual level and focuses on their economic and social aspects, leaving aside all kinds of psychological and somatic changes. In an early meta-analysis of 37 studies dealing with the long-term consequences of parental divorce for adult well-being, Amato and Keith (1991a) distinguished 15 different areas in which people are affected by divorce, ranging from psychological well-being to physical health. We focus here on what they have called occupational quality and material quality of life. More specifically, we are interested in men’s and women’s occupation, income, housing and living conditions, unemployment and welfare dependency. Admittedly, this list excludes important topics such as social networks and family relationships. But it still includes a lot of important topics that deserve thorough empirical investigation, especially in the European context.

WHAT DO WE KNOW SO FAR? Most research on the consequences of partnership dissolution has been conducted using US-American data and focuses on economic well-being. A classical example is the study by Weitzman (1985), who argued that women’s average post-divorce economic well-being deteriorates by no less than 73 per cent, while men experience an improvement of 42 per cent. The accuracy of such high averages has been strongly questioned (Peterson, 1996), but the overall finding is that women experience a drop in economic status after a divorce, while the economic impact is not so great for men (some studies even find that men’s economic position improves after divorce). A nice summary of the corresponding US studies can be found in McKeever and Wolfinger (2001). Depending on the methodology of each study (time period, data source, income operationalization, measure of change), women lose between 9 and 51 per cent of their former income during marriage. The few studies focusing on men’s incomes (for example, Andreß and Bröckel, 2007a; Burkhauser et al., 1990, 1991; Finnie, 1993; Jarvis and Jenkins, 1999; Poortman, 2000) report slightly negative or slightly positive changes (ranging between 27 and 16 per cent). This body of research was recently contested by McManus and DiPrete (2001), who argued that many of the aforementioned studies do not adequately deal with men’s support payments and that recent changes in gender-role specialization and law enforcement have levelled off the gender gap in post-divorce incomes. Contrary to conventional thinking, they found that the majority of partnered men in the USA lose economic status when their unions dissolve. A primary source of economic decline after union dissolution is their inability to compensate for the loss of their

Introduction

3

partner’s income, which constitutes an increasing part of the household income as women become more and more active on the labour market (for a related finding in the British context, see Jenkins, 2008). A secondary source of economic decline is an increase in compulsory and voluntary support payments. Another strand of research focuses on changes in labour supply. Given the aforementioned economic decline after partnership dissolution, (increased) employment is one of the ways to cushion the income loss. Correspondingly, many studies find that the majority of women increase their employment after partnership dissolution (Andreß and Bröckel, 2007a; Bradbury and Katz, 2002; Duncan and Hoffman, 1985; Finnie, 1993; Haurin, 1989; Johnson and Skinner, 1986; Morgan, 1991; Peterson, 1989). However, there are also some research findings that are not fully consistent with this view. For example, Mueller (2005) finds no change in women’s labour supply and Jarvis and Jenkins (1999) even detect a decrease. These conflicting results may be due to anticipation (women increase employment in advance because they expect a divorce) or to institutional effects (women decrease employment because welfare payments guarantee their standard of living). Using more recent survey data, Jenkins (2008), for example, shows that employment rates for British mothers rose steadily throughout the 1990s and were given a particular stimulus by the introduction of the Working Families Tax Credit in 1998. Both changes had positive effects on the labour supply of separated women with children and reduced the number of solo mothers living on welfare. Contrary to research on women’s employment, little is known about the divorce effect on men’s labour supply – which is quite surprising given the extensive literature on the positive effects of marriage on men’s careers and wages (for an overview, see Waite and Gallagher, 2000). One plausible hypothesis to be drawn from this body of research is the assumption that if there is a ‘marriage premium’, there should also be a ‘divorce penalty’. In line with this assumption, Kalmijn (2005) finds that after divorce, Dutch men experience an increased chance of becoming downwardly mobile, an increased chance of becoming unemployed and an increased likelihood of becoming disabled. On the other hand, most men living in partnerships work full-time even before separation, so that further increases in employment are hardly feasible for them. Hence, it is only a minority for which we can expect positive changes, while for the majority, any change in employment status implies either a lateral or a downward move. But in a recent German study (Andreß and Bröckel, 2007a), the latter event is observed as often for women as it is for men. Hence, there is evidence that men’s labour supply is negatively affected by partnership dissolution, but it is not quite clear whether this effect is gender specific.

4

When marriage ends

In addition to changes in income and employment, separation and divorce are almost always connected with a household split. Sooner or later, at least one of the two partners leaves the common home (Feijten and van Ham, 2007). Surprisingly little is known about this residential mobility, although it is a necessary by-product of partnership dissolution. According to Dieleman and Schouw (1989), the high number of divorces has a strong impact on the housing market. Their analysis shows that in any given year, the increased demand for housing in a particular region is equal to about one-third of the number of divorces which occurred six to seven years previously. Quite generally, residential mobility can either be a forced choice or a voluntary decision. In the latter case, researchers usually assume that people will only change their residence if the benefits of the new home exceed the costs of moving. Hence, voluntary residential mobility is usually associated with improved housing conditions and upward mobility on the housing market. Accidental circumstances such as unemployment and divorce often result in involuntary residential mobility, in which the individuals have to accept inferior housing conditions compared with their former home. Separation also causes a decrease in household size and disposable income, both of which lead people to look for smaller and less expensive homes. Furthermore, researchers have found that part of the residential mobility is preliminary (Andreß et al., 2003; Holmans, 1990; Wasoff and Dobash, 1990). Hence, living with family and friends is one option; temporarily accepting an inadequate home is another. Finally, if a couple separates, there is the question of how to divide savings and assets. If they own a house jointly, this is usually their greatest asset. Splitting up this asset requires one partner to buy the other partner out, which may be difficult given the limited financial capacities of both partners (apart from the fact that the house may be much too large and expensive for one partner alone). Overall, it is a plausible assumption that separation and divorce are associated with a great deal of downward mobility on the housing market, both in terms of residential status (from owner-occupied housing to rented apartments or temporary living arrangements) and in terms of size and quality of the dwelling. Correspondingly, many studies have observed a negative effect of partnership dissolution on people’s housing careers (see, for example, Andreß and Bröckel, 2007a; Andreß et al., 2003; Feijten, 2005; Holmans, 1990; Sullivan, 1986). But it is not quite clear whether men and women are affected differently, or what happens if the couple has children. It seems to be common practice in most countries for the children to stay with their mother in the majority of cases. If there is a dispute about child custody, family courts will usually decide in favour of the children’s well-being, which often means assigning the former parental home to the mother

Introduction

5

and children. However, women experience financial difficulties much more often than do men (see the previous discussion), and hence may have problems retaining the former family home. Tentative data from an analysis of home ownership points in this direction. According to Feijten’s (2005) analysis of Dutch event-history data, women have a consistently higher probability of dropping out of owner-occupation than do men (see also Symon, 1990). Feijten argues that women’s lower degree of economic independence is the most likely cause of this finding. However, an analysis of German data shows that both genders have to cope with residential mobility to the same extent (Andreß and Bröckel, 2007a). Focusing on individuals with minor children, the overall mobility rate is slightly lower, but contrary to the expectations, mobility is higher for mothers than it is for fathers. On the one hand, this result is in accordance with the common assumption in divorce proceedings that children should stay in their familiar environment and not change homes. On the other hand, it is surprising that this is particularly true for men and less so for women. Yet a more refined analysis of the German data shows that residential mobility is a forced choice, especially for women and most particularly for mothers. Staying in their former family home would have meant having to cope with disproportionately high housing costs. Both the Dutch and the German studies seem to tell the same story that is already familiar to us from the analysis of household incomes: separation and divorce have negative economic and social consequences, but it is women who bear the brunt of the burden – and among them, mothers are an especially vulnerable group. Yet we should not forget the aforementioned results, which indicate that this picture may be gradually changing due to increased women’s employment, enforced support payments and possible negative effects of partnership dissolution on men’s work careers. A careful reading of the literature on residential mobility provides additional clues. For example, low-income men seem to have very limited options on the housing market. Sullivan (1986) shows that the parental home is an extremely important source of housing for young, working-class men. Those men who cannot afford to stay in owner-occupation and who do not find suitable accommodation in the private rental sector have to rely on temporary arrangements such as bed and breakfasts, hostels, caravans or even homelessness. Furthermore, the social and psychological costs of divorce for men are very well documented in the literature (for an overview, see Amato, 2000; for a recent German study see, for example, Andreß and Bröckel, 2007b). Taking all this into account, it is no surprise that studies looking at a wider range of indicators find that divorce results in an increase of lifestyle deprivation both for men and for women. For instance, in an earlier study comparing ‘multidimensional’ poverty in Belgium and

6

When marriage ends

Britain, Dewilde (2004) found that divorce significantly elevates the risk of entering situations of housing deprivation, financial stress and limited financial resources for both genders. In sum, all of these findings call for a much more comprehensive analysis of the consequences of partnership dissolution than those that focus exclusively on changes in income. Another drawback is certainly the predominance of US-American data in divorce research. Research on the economic and social consequences of partnership dissolution in European countries is still scarce and often restricted to specific countries where the necessary statistical information is available through large-scale representative household surveys. The situation is somewhat better with respect to household income, which has been researched both in single countries (Andreß et al., 2003; Andreß and Bröckel, 2007a; Fritzell, 1990; Gähler, 1998; Jarvis and Jenkins, 1999; Jenkins, 2008; Manting and Bouman, 2006; Poortman, 2000) and comparatively across the EU member states (Dewilde, 2002; Uunk, 2004). But analyses of changes in employment, housing, and living conditions are hard to find (see, however, Andreß and Bröckel, 2007; Feijten, 2005; Kalmijn, 2005). On the other hand, the great variation in institutional arrangements between European countries and the availability of cross-nationally equivalent survey data like the European Community Household Panel (ECHP) or the European Union Statistics of Income and Living Conditions (EUSILC) also give us the opportunity to study the effects of different welfare regimes and family policy models on the consequences of partnership dissolution. The analyses by Dewilde (2002) and Uunk (2004) are examples of this approach. Both use welfare state typologies to describe their results. Additionally, Uunk (2004) uses quantitative indicators to study the effects of income-related and employment-related welfare provisions on women’s post-divorce incomes. Both types of welfare provisions improve women’s economic well-being. Uunk (2004) finds that higher social welfare payments increase women’s post-divorce income position more than higher public childcare provision does. Overall, both studies demonstrate that not only individual factors such as gender, children and labour force experience, but also institutional characteristics of the societal context determine the consequences of partnership dissolution.

HOW CAN WE EXPLAIN THE ECONOMIC AND SOCIAL CONSEQUENCES OF PARTNERSHIP DISSOLUTION? A necessary step towards explanation is to present a good description of what actually happens at the individual level. Thus, this book has two basic

Introduction

7

aims. First, we try to provide a comprehensive picture of the social and economic consequences of partnership dissolution. Who moves out of the family home? Who takes care of the children? How is the resident parent remunerated for childcare? How do the separated individuals earn a living on their own? Do they increase or decrease their labour supply? Do they get financial help from their former partners? Do they have to rely on state support? Do they suffer from material hardships, and so on? Second, based on this account, we try to explain differences between the two genders, between our target group and the continuously married, and between their present living conditions and those during their former partnership. For example, if women experience larger income declines than men, as the evidence from many studies suggests, what are the causes of this gender inequality? Is it a result of less economic activity on the part of women? If so, what are the causes of low female labour supply? Are women restricted in their employment possibilities because of childcare? If not, do women earn lower incomes than men, and why is that the case? Do they have lower qualifications? Have they lost part of their human capital during marriage due to temporary retreats from the labour market? What about the rights that family law provides for claiming maintenance payments from former partners? How are they used and enforced? And, finally, does the state help out in cases of financial need, insufficient maintenance payments or a lack of childcare? Understanding the social and economic consequences of partnership dissolution necessitates economic and sociological theories, which can provide answers to at least three broad questions: (1) What are the gains involved in consensual or marital unions, and which of these are lost in cases of separation and divorce? (2) Why are the losses unequally distributed between the genders? (3) What incentives and restrictions do both partners have to improve their situation after their partnership has broken down? Economies of Scale Beginning with the first question, a natural starting point is the economic theory of private households. It describes such an association of individuals as a small firm, which allows for the production of private and public goods, as well as protection against common life risks, at a much lower price than those individuals would have had to pay were they to produce or buy these goods on their own at market prices (for a textbook summary see, for example, Bryant, 2006; Cigno, 1991). In cases of separation or divorce, most of these advantages are lost: separate homes are necessary, someone is needed to care for the children or to earn money for living,

When marriage ends

8

life becomes more expensive because economies of scale are lost, and so on. Sørensen (1994: 178) has estimated these losses of economies of scale at between 27 and 36 per cent. It should also be noted that economies of scale are reinforced by state policies in some countries, which redistribute money to married families (for example, by joint taxation). These benefits are lost when couples divorce. Gender Differences Gender differences in post-separation incomes – the subject of our second question – are usually attributed to the division of labour within the family, for which strong economic incentives exist when the two partners have different levels of productivity in home and market activities. Economic theory of the household asserts that it is efficient if the partner with the lower market income (usually the wife) specializes in home work and childcare, while the other partner (usually the husband) specializes in market activities (Becker, 1981).1 Additional incentives may result from the institutional context, which in some countries benefits the male breadwinner family (for example, by granting tax privileges), while other countries strive for more gender-neutral policies. This unequal division of labour remains efficient as long as the relationship between the two partners continues. But as rising divorce rates demonstrate (data on consensual unions are not available), marriage is no longer a lifelong commitment. In case of an early breakup, the chances of an unequal distribution of marital gains and losses between both partners are high (Ott, 1999). In this case, we can usually identify four risk factors (Holden and Smock, 1991: 68; Sørensen, 1994: 173): ●

●

●

Risk 1: If there are children, economic needs after separation are higher for the resident parent, which in most cases is the mother. Because she has to care for the more economically dependent household members, she is often restricted in her earning capacities, especially if childcare facilities are scarce. Risk 2: Moreover, the risk that the resident parent will not receive the necessary child maintenance costs from the non-resident liable parent is high. Risk 3: Income sharing during marriage allows women to benefit from men’s market incomes, which are higher on average than women’s market incomes. It is questionable whether the amount of maintenance payments women receive from their former spouses after separation compensates for the loss of the economic support they received during marriage.

Introduction ●

9

Risk 4: Changes in material wealth, pension entitlements and human capital that both partners experience during marriage are only imperfectly measurable and therefore hard to equalize after separation. Even if they are measurable, the economic loss may be high when they are sold to pay off the partner (for example, in the case of a common house).

Risk 4 is especially true for gains and losses in human capital. Given the frequent division of labour in many countries – with a husband employed full-time and a (maximum) part-time working wife who takes care of the household and children – it is probable that the gains in human capital achieved by continued employment on the part of the husband (Kenny, 1983) are not set off against the losses of human capital caused by interrupted or delayed working careers on the part of the wife. At least three objections have been made against this gendered view of divorce outcomes. One is that married women’s employment has increased significantly within the past decades such that their economic dependency on the male partner’s income may be less pronounced than expected. This introduction is not the place to discuss the mechanisms underlying this secular trend (but for further exploration, see Stier, in this volume). However, it should be mentioned (1) that welfare states increasingly support women’s employment (Gornick and Meyers, 2003) and (2) that women may increase their labour supply simply in order to be secured against the pitfalls of marriage instability (Beck and Hartmann, 1999; Diekmann, 1994; Johnson and Skinner, 1986; Poortman, 2005; Rogers, 1999). A second objection is that complaints about the unreliability of support payments may be a little bit outdated, since many countries have introduced advanced maintenance systems and have tried to improve the legal regulations and the practice of support payments (see Part II of this volume). Thirdly, it has been noted that an increasing number of partnerships consist of consensual unions which are not regulated by a (marriage) contract, and thus may be less binding relationships in which each partner is more or less economically independent from the other. Coping with Separation and Divorce Finally, looking for theories to address the third question is a difficult task, simply because there are so many options open to both partners after separation. They can economize on their expenses, they can improve their financial situation by increased employment, they can search for additional economic and social support, both from the state and from their family and friends, and so on. To keep this theoretical overview short, we focus on two

10

When marriage ends

aspects, housing and employment, which have already been discussed in our literature review. Given the financial strain that occurs during the process of separation, both partners have high incentives to economize on their expenses. Housing costs usually make up a large portion of their monthly budget, and due to the household split, both partners have less demand for housing. Thus, residential mobility is one of the first options, and since household size decreases and financial resources are scarce, the chances are high that this mobility will be downward (less living space, less quality, but also lower costs). This economic motive is a function of each partner’s post-separation income, and since the economic strain seems to be much greater for women than it is for men, downward residential mobility is much more probable for women than for men. A second motive is the wish of both partners to separate physically from each other, which is an incentive for both men and women to decrease the emotional stress. It is quite probable that questions of guilt and redemption play a role in this process (for example, that the partner who initiated the breakup should also leave the couple’s home). Thirdly, there is the child protection motive, which assumes that children of separated parents should stay in their parental home in order to minimize their emotional stress. This implies that the custodial parent – which is often the mother – is less affected by residential mobility. Although the child protection motive is enforced in some legal systems, it is unclear how this motive fits in with the rather strong economic incentives for women – especially for women with dependent children – to economize on their expenses. Finally, there is the question of property division. Although this pertains mostly to homeowners, it should be noted that the percentage of owner-occupation is quite large in some countries, especially in Southern Europe, where outright ownership is the common tenure status (Dewilde, in this volume). Thus, the question of how to divide the common home is a relevant problem in some countries. Most legal systems expect an equal division of the property acquired during marriage, which implies all the problems described in our literature review. Unless one partner is able to buy out the other, property division is usually a financial disaster for both partners. However, in those cases in which the economically dependent partner does not receive any kind of support payments from the breadwinner, the takeover of the common house may be a form of compensation and serve as an economic safeguard for the (formerly) dependent partner.2 Instead of economizing on their expenses, both partners could also increase their employment. The mechanics of labour supply are very well described by microeconomic theories (see, for example, Ashenfelter and Layard, 1987) which model the choice between different uses of time (for

Introduction

11

example, working time and leisure time). According to these theories, the amount of (paid) labour an individual supplies to the market is a function of his or her market wage and reservation wage (that is, the lowest wage the individual would accept). Market and reservation wages, in turn, are functions of the individual’s human capital and alternative income sources. The better educated the individual, the higher the market wage and the incentive to work. The more alternative income sources he or she has, the higher the reservation wage and the lower the incentive to work. Given their financial needs after separation, both men and women have a high incentive to increase their employment. This will be especially true for better educated individuals. They have more employment opportunities and are better remunerated. Income transfers from the former partner or the state (both an alternative source of income) may not only ease the financial strain, but may also be a disincentive to work, especially for those with low market wages and correspondingly low prospects on the labour market. However, decisions about labour supply are not free of constraints. For example, men often work full-time before separation, and increasing their employment is simply impossible. Even searching for a better remunerated job is often not a feasible alternative. The incentive to increase labour supply can also be low if the man feels that his additional incomes will be illegitimately taxed away by support payments. Finally, one should not forget the social and psychological costs of separation and divorce for men. As Kalmijn (2005) notes, there should be a male ‘divorce penalty’ corresponding to the male ‘marriage premium’. For example, lost female support may result in less career stability and more downward mobility for men. Women, on the other hand, who are often the caregivers of the common children, face certain restrictions when combining care and work. Public childcare facilities and family-friendly jobs (for example, part-time jobs) may be scarce, making the (re-)entry into paid employment a difficult task. In sum, although both genders have strong incentives to increase their labour supply, there are certain subgroups for which this assumption is not true (or simply incomplete, because it does not take into account all the relevant costs and benefits).

THE IMPORTANCE OF THE INSTITUTIONAL CONTEXT Thus far, we have discussed socio-economic theories of individual behaviour. They explain how individuals use the resources at their disposal – for example, time and human capital – to create outputs which contribute to their own well-being. Based on their individual preferences and given the

12

When marriage ends

incentives and restrictions of the specific situation, they choose between different options for individual behaviour. This scheme of resources, options and individual action (within given individual preferences and restrictions) is a popular explanatory tool in research on families and partnerships. But it places a strong focus on individual characteristics and properties of the partnership in explaining individual behaviour, too often neglecting the societal context in which individual action takes place. As we have learned, for example, from the results of Dewilde (2002) and Uunk (2004), the economic well-being of women after divorce depends not only on individual-level variables (for example, age, number of children, education, and so on), but also on the general characteristics of the welfare state – for example, whether the state provides cash transfers in cases of economic need or whether it provides childcare services to support the employment of mothers. In other words, when coping with the consequences of separation and divorce, individuals also utilize more general systems of support besides family and friends. Theories of welfare production usually distinguish between three main providers of welfare (Esping-Anderson, 1990; Rainwater et al., 1986; Zapf, 1984). These are (1) the market providing goods, services and income opportunities, (2) the state providing rights, cash transfers and social services, and (3) the extended family and the social network providing social and material support.3 This institutional context – in addition to many individual factors – provides specific incentives and may alter the constraints faced by individuals in cases of separation and divorce. It makes a difference, for example, whether the institutional setting encourages women’s labour force participation, whether the labour market provides enough opportunities for the gainful employment of both genders, whether the state and the legal system offer rights and financial support for dependent family members and how the kinship network operates in case of family failure. In sum, the national configuration of family, market and state institutions moderates the consequences of partnership dissolution. In order to understand the institutional effects, one has to vary the national context systematically. Thus, an analysis of the economic and social consequences of partnership dissolution in different European countries provides us with the opportunity to test these institutional effects more thoroughly. That is why this book is organized as a comparative study. Overall, we try to explain economic and social consequences of partnership dissolution by means of a combination of individual risk factors (for example, low earnings capacity, parenting, lost network contacts) and context effects due to the institutional setting in each country. The importance of the institutional explanation is easily illustrated with some examples from the literature. One could hypothesize, for example,

Introduction

13

that employment-friendly public policies for women (Gornick et al., 1996, 1997) are also advantageous for separated and divorced women in terms of labour market (re-)entry and financial well-being. But it is also possible that employment-friendly policies result in a two-income norm, and this may make single women even poorer relative to married families. Generous cash transfers, on the other hand, are usually assumed to act as strong incentives not to work, because cash transfers like child benefits, home care allowances or social assistance payments may already provide a sufficient income. In sum, cash benefits may buffer the decrease in economic well-being after separation, although at a rather low level. In the long run, however, only gainful employment guarantees participation in the economic benefits of society. It is also well known that female educational attainment and labour market participation has increased dramatically in recent years. This should be a safeguard in case of separation and divorce, but it also depends on the question of whether national labour markets provide sufficient and remunerative jobs for (better educated) women. Current research on welfare states usually assumes that there are more or less coherent clusters of institutional arrangements that embody distinct values and rationales. This assumption has been operationalized in the concept of welfare regimes (Arts and Gelissen, 2002; Esping-Andersen, 1990, 1999), and it is tempting to use one of the existing welfare state typologies to characterize the institutional context of partnership dissolution. But many scholars argue that such a typology should be domainspecific (for example, Gallie and Paugam, 2000), and especially in the field of family relations, several typologies have been proposed. These include the classical distinction between liberal, conservative, social democratic and Mediterranean (rudimentary or familialistic) welfare regimes (EspingAndersen, 1990; Ferrera, 1996; Leibfried, 1993), but also different kinds of ‘gender-sensitive’ typologies (for example, Ostner and Lewis, 1995; Leitner, 2003; Siaroff, 1994), family policy models (for example, Andreß, 2003; Gauthier, 1996) and classifications of legal traditions (Kaufmann, 2002). Historically, two ways of ensuring the welfare of families have emerged: some welfare states primarily provide monetary transfers and fiscal relief, while others focus on public infrastructure enabling family members to participate in the labour market (Gauthier, 1996). With respect to partnership dissolution, another task of the state is to enact laws which protect children and former partners from negative economic consequences. The market, in turn, can secure individual welfare by providing opportunities to earn one’s keep. Finally, as we already mentioned, family and social networks can offer social and financial support, especially in case of a separation or divorce. Andreß (2003) suggested a typology of family support in industrialized

14

When marriage ends

countries which shows how the particular national configuration of state, market and family influences the economic consequences of separation and divorce. One dimension of this typology is public policy orientation, and it distinguishes between extensive and rudimentary family policies.4 As we already mentioned, extensive family policies can have two different orientations. If family policies are only rudimentarily developed, either the market or the family has to step in. Including these subdivisions, four different models of family support result: 1.

2.

Welfare states with an extensive family policy can be found in the Continental European and Northern European countries. There are two different types of governmental support: either through financial compensation for children and other dependent household members in the form of cash transfers or tax relief, or through support for sufficient care facilities and gainful employment for all adult family members. Accordingly, Andreß distinguishes between what he calls: (a) the male breadwinner model and (b) the dual-earner model. In the male breadwinner model, policies support a traditional, genderspecific division of labour. This model assumes that men are the main breadwinners, while women at most contribute a supplemental income, especially if they have children. Given additional financial support from the state for the dependent family members, the male breadwinner’s income should suffice to support the family. In contrast, women’s employment is more supported in the dual-earner model, and childcare responsibilities should be more evenly distributed between men and women. Family policy is oriented towards employment; thus, sufficient care facilities for children and/or the elderly are offered. Accordingly, double-income households should prevail among married couples and consensual unions. Welfare states with a rudimentary family policy support families to a minor degree, if at all. This policy orientation is dominant in the anglophone and Southern European countries. In countries with rudimentary family policies, the few existing benefits are established only to prevent poverty. Here, the safeguard occurs either via the market or through the family, which Andreß calls (a) the market model and (b) the model of extended family solidarity. The market model emphasizes the individual freedom of the members of a society, and therefore prefers private solutions based on the market to state intervention in family matters. The model of extended family solidarity (the family model, for short) also prefers private solu-

Introduction

15

tions to state intervention; but in this case, the family network and its support capabilities should play a greater role than the market. Andreß assumes that these four types of family support create specific dependencies within the family, which in case of separation or divorce may have negative economic consequences for the economically weaker partner. More specifically, one can ask how much ‘autonomy’ (Orloff, 1993) is granted to the weaker family members within each of the four models.5 Obviously, gainful employment and own income are the basis of financial autonomy. On the basis of Andreß’s typology, it can be supposed that there is little gender-specific division of labour in the dual-earner model. Both partners have more or less equal chances to combine work and family life. In the two rudimentary family policy models, a definite form of labour division is not explicitly encouraged by the state. Rather, it is left to the partners to coordinate employment and childcare. The male breadwinner model ranks somewhere in between: On the one hand, a traditional division of labour is supported by granting tax privileges for marriages with one principal earner and by not offering sufficient full-time childcare facilities. Therefore, full-time employment for married mothers is inefficient or utterly impossible. On the other hand, this model affords a compatibility of childcare and work based on the part-time employment of the mother. Like every typology, these four models of family support are Idealtypen in a Weberian sense. They represent hypothetical configurations of state, market and family, while the real world of nation states consists of mixtures and combinations of these models. Although Andreß (2003: 635) – based on a thorough analysis of various national statistics – identifies Germany, Sweden, the USA and Spain as typical examples of the four models of family support, the typology is still a hypothesis in a twofold sense: first, in the sense of whether it is a useful classification of real-world family policies and, second, as to whether it helps to explain the variance of economic and social consequences of partnership dissolution across different countries. As Stier (in this volume) argues, the typological approach has certain methodological limitations. Alternatively, one could utilize direct measurements of the relevant characteristics of the institutional context (see the discussion in Stier and the empirical analyses in van Damme and Uunk, both in this volume).

METHODOLOGICAL APPROACH Studying the consequences of separation and divorce is a difficult matter. Ironically, it is the small number of divorces that is problematic. Why is that

16

When marriage ends

the case? The answer is that although divorce rates are high and increasing, recently divorced individuals are still a small fraction of the population at any given point in time.6 And we need data from recently divorced individuals in order to measure what has changed since the time of marriage. Respondents who were divorced five years ago may have forgotten important details or may provide a biased report of what has actually happened. Because retrospective information may be unreliable as a result of memory loss and memory bias, many scholars prefer prospective research designs. A typical prospective design samples a representative cross-section of the population and then collects information on how these individuals change over time, both in terms of partnerships and living conditions. But again, since divorce is a rare event, one usually needs a long observation period in order to find sufficient numbers of divorcees in such panel data. It is therefore no surprise that many studies of divorce consequences are based on small numbers of cases. Only in a few studies have researchers had access to large-scale surveys that included sufficient sample sizes or that used some kind of oversampling for the divorced. Long-running panel studies are another possibility, because they allow researchers to pool information from all divorces within a long observation period. But all of these alternatives are costly, and hence may not be available for the time point or country in question. This is also true for some of the following analyses. In some places, we have to restrict our analyses to the currently divorced, although the event of divorce may have happened years before and we have no information on the former marriage that allows us to assess the amount of change caused by the divorce. We also sometimes use information on single parents to describe the implications of having a child, although we are not sure about the timing and causes of single parenthood. It may be a result of divorce, but equally the other parent could have well passed away, or the child could have been born outside a partnership. Target Population Having talked about the problems of studying divorce consequences, we are now ready to define our target population more concretely. Naturally, married couples are a central focus of our study. But what is the main event of interest? Up to now, we have used the terms ‘separation’ and ‘divorce’ more or less interchangeably, because in many countries, a period of separation is a prerequisite for legal divorce. However, most of the aforementioned consequences are in effect long before the divorce decree is passed. Therefore, a natural starting point for most of the following analyses is the period of separation, which is often connected with a household split and residential mobility. Separation, on the other hand, is an event which

Introduction

17

is observed not only for married couples. It equally applies to cohabiting couples, although – as we already mentioned – the average economic and social consequences may not be as severe, since cohabitation for many is less binding than marriage, and mutual dependence is less pronounced. We have made no a priori decision as to whether cohabiting couples should be included or excluded from the following analyses. Married couples are the prime target population of all the studies in this book, but some may also include cohabiting couples – for instance, because sample sizes would otherwise be too low. In other words, this book investigates the consequences of different kinds of partnership dissolution, but with a strong focus on marital disruption. Whenever possible, we use longitudinal information and compare living conditions before and after separation for the target population. Unfortunately, such detailed information is not always available, and we have to use retrospective surveys of separated individuals or cross-sectional data that compare presently divorced or separated individuals with continuously married couples. Mainly for pragmatic reasons, we have decided to focus on the experiences of the adult members (husbands and wives, men and women) of these married and cohabiting couples. Obviously, this decision excludes one important subgroup: the children of the separating couples, who are no less affected by partnership dissolution than are the adults. But it is much more difficult to investigate divorce consequences for children than for adults, at least in the case of the socio-economic topics studied here. Representative survey data, on which most of the following analyses are based, usually sample the adult population and thus provide only indirect information about children. Basic socio-demographic data are provided by their (interviewed) parents, but their personal views are missing. Thus, analyses from the vantage point of children are absent from this book (but see the review articles by Amato and Keith, 1991b; Kelly and Emery, 2003; Rodgers and Pryor, 1998).7 But of course we discuss the economic and social implications of having children for the adults. We also give a detailed account of legal regulations concerning child custody and maintenance payments. Comparative Research Design How did we organize the comparative study? Lacking a common database of partnership dissolutions, we had to combine different comparative research strategies. A detailed account of possible methodological approaches can be found in Stier (in this volume), who discusses the conceptualization and measurement of institutional contexts. Basically, our research combines case-study methods with micro–macro analyses. Correspondingly, several chapters consist of in-depth case studies focusing

18

When marriage ends

on single countries and national statistics, while other chapters employ comparative analyses of several European countries using cross-nationally equivalent statistical data. The term ‘micro–macro analyses’ refers to the fact that the latter chapters combine micro data on individuals with macro data describing the institutional context. As Stier’s literature review (in this volume) shows, such macro data may include typologies or quantitative indicators – for example, typical family policy models like those discussed previously, or the number of children in public childcare like in the study by Uunk (2004). But how did we select the countries for our case studies? First, this was a question of availability. We needed scholars who were experts on both the corresponding national context and on research in family relations. Starting from our own research contacts and from our knowledge of the literature, we were able to build a network of interested sociologists, economists and demographers from ten different European countries. The group met at several workshops, and this book represents the output of our collaborative work. Furthermore, when building up the research network, we tried to represent different national institutional settings in order to test the validity of our institutional hypothesis. As a guiding tool for the case selection, we used the aforementioned typology of family policy models (see the previous section). Our case studies include Denmark, Finland and Sweden as representatives of the dual-earner model, Great Britain as a representative of the market model, Belgium and Germany as representatives of the male breadwinner model and, finally, Greece and Spain as representatives of the family model. Countries were assigned to the four models of family support based on the statistical analyses in Andreß (2003: 635). But as we have already stated, this typology is itself a hypothesis, and one of the aims of the corresponding case studies is to test the validity of this assignment.8 Each country chapter provides a descriptive overview of the incidence of divorce and the socio-demographic profile of the divorced together with an account of national family law, labour market institutions and welfare state provisions.9 In doing so, each chapter tries to present a comprehensive picture of partnership dissolution and its consequences in each country. At the end of any given chapter, the reader should know what is typical for the country in question, and what is a common experience that can also be observed in other European countries. The single case studies are complemented by four comparative analyses focusing on specific consequences of partnership dissolution and using cross-nationally equivalent data sets of Western European countries (including the countries from the former case studies). Three analyses are based on the European Community Household Panel (ECHP). One

Introduction

19

analysis uses a pooled data file compiled from five national household panel studies which were post-harmonized by the authors. These comparative studies analyse the consequences of partnership dissolution with respect to labour market participation, residential mobility, household income and, finally, non-monetary indicators of well-being. In explaining these consequences, they refer both to individual determinants and to characteristics of the institutional context. Thus, all comparative analyses are based on a combination of micro and macro explanations.

STRUCTURE OF THE BOOK As can be seen from Figure I.1, this book is organized into four parts and one appendix. Part I discusses theoretical, methodological and measurement issues when analysing the economic and social consequences of partnership dissolution. Haya Stier’s chapter on conceptualization and measurements of institutional contexts has two major goals: first, to review the various approaches to exploring and understanding how macro-level mechanisms affect individuals’ behaviour and outcomes, and, second, to discuss and propose measurements commonly employed in defining the ‘institutional context’. In the first part of the chapter, she addresses two common strategies for comparative research. The first is the ‘case-study approach’, which often centres on comparisons of just a few countries. She elaborates on this strategy to include multiple-country comparisons and the more common approach of using country typologies in order to draw implicit conclusions regarding the way institutions and cultural characteristics affect individuals’ behaviours and destinies. The second approach is termed ‘macro–micro analyses’. It is based on a comparison of a large number of countries, and on a quantitative examination of the (direct) effect of specific institutional characteristics on social and economic behaviours and outcomes. In the second part of the chapter, she discusses specific institutional arrangements that affect – directly or indirectly – the economic position of women, and she reviews the major approaches to measuring them comparatively. Following Haya Stier’s distinction between case studies and macro– micro analyses, we have organized the empirical chapters into two parts: single-country studies on the one hand (Part II) and comparative analyses based on several countries on the other (Part III). As Figure I.1 shows, Parts II and III are structured ‘vertically’ and ‘horizontally’ according to our typology of family policy models and according to four dimensions of divorce consequences. Part II consists of eight10 country reports (Germany, Belgium, Great Britain, Spain, Greece, Denmark, Finland and Sweden),

20

Figure I.1

Part II: Single country studies

Introduction

Part V: Data appendix

Part IV: Conclusion

Poverty and lifestyle deprivation

Household income

Residential mobility and housing

Labour force participation

Family model

Market model

Dual-earner model

Denmark Sweden Finland

Rudimentary family policy

Great Britain

Explicit family policy

Part I: Theoretical, methodological and measurement issues

Male breadwinner model

Germany Belgium

Structure of the book

Part III: Comparative studies (many countries)

Spain Greece

Introduction

21

which have tentatively been classified as members of the male breadwinner (Germany, Belgium), the dual-earner (Sweden, Denmark, Finland), the market (Great Britain) and the family (Spain, Greece) models. All the country reports have a similar internal structure including (1) a sketch of recent demographic and labour market trends, (2) a summary of how the family and family relations are treated in law and social policy, and (3) a description of the social and economic consequences of partnership dissolution using national statistical data. Part III consists of four comparative chapters analysing these and other (Western) European countries with respect to four dimensions of divorce consequences: (1) labour force participation, (2) residential mobility and housing, (3) household income, and (4) poverty and lifestyle deprivation. In explaining these divorce consequences, the comparative chapters will make use of the proposed typology and of quantitative indicators characterizing the relevant institutional determinants. Finally, the Conclusion in Part IV summarizes the results of the empirical chapters and discusses the validity of the proposed typology and alternative explanations of divorce consequences. All quantitative information that is used in the empirical chapters is summarized in the Data Appendix. This includes a listing of socio-demographic indicators and social policy measures for the years 1985, 1990, 1995, 2000 and 2005 as well as descriptions of the cross-national survey data (ECHP, GSOEP, BHPS, and so on) that are used in the comparative analyses.

NOTES 1.

2.

3. 4.

More specifically, economic theory argues that if women’s market skills relative to their domestic skills are lower than men’s relative skills, then women will remain in the home and men will specialize in paid work. It should be noted, however, that this model assumes that a marriage will last forever. If we introduce the possibility of divorce, it may be more efficient in the long run not to specialize. Functionally, it is equivalent to the dower (also called morning gift), which in earlier times was a provision accorded by law to a wife for her support in the event that she should survive her husband (that is, become a widow). A dower (‘mahr’) is still an essential part of traditional Muslim marriages today. In cases of divorce, the bride does not have to give up her dower unless she is the one who requested the divorce. Some authors (for example, Evers, 2004) also mention a fourth provider of welfare: the so-called ‘third sector’ or the ‘civil society’ (foundations, churches, cooperatives, service clubs, and so on). Kaufmann (2002) distinguishes between explicit and implicit family policies. In the first case, a political discourse about family matters can be observed, and family policy is institutionally autonomous in the sense that either a specific ministry (for example, a Ministry of Family Affairs such as that in Germany) or a similar governmental institution exists. However, if certain political measures can be interpreted as family policies by an external scientific observer without this notion being specifically stated in official politics, Kaufmann terms them implicit family policies. Our distinction between extensive

When marriage ends

22

5. 6.

7.

8. 9. 10.

and rudimentary family policies only looks at the number and scope of political measures that support the welfare of families, irrespective of their interpretation in the political discourse. For example, Sweden could be termed a country with an extensive family policy, although in the Swedish political discourse, most measures have been motivated as instruments for gender equality. Orloff (1993) defines autonomy as the freedom not to continue a potentially repressive relationship – for example, because of economic dependency. This problem is easily illustrated using divorce data from Germany. That country experienced a total of 1 031 071 divorces between the years 2001 and 2005. Assuming that every divorcee founded a household of his or her own and relating twice the number of divorces to the total number of households in 2005 (n 5 39 178 000) results in a percentage of about 5.3 per cent. Thus, a representative survey of the population – be it face to face or by telephone – will have a less than 6 per cent chance of sampling a person who has been divorced within the last five years. For example, a sample of 3000 respondents would include about 158 divorcees, provided non-response is not disproportionately high among divorced persons (which is doubtful, as the experience from specialized surveys of the divorced shows). Specialized surveys interviewing children have mostly been used to investigate the psycho-social consequences of divorce, such as changes in subjective well-being, educational aspirations or anti-social behaviour (see, for example, Furstenberg and Cherlin, 1991; Wallerstein et al., 2000). Another strand of research uses comparative micro data on children and teenagers available from international surveys of school achievement (Third International Mathematics and Science Study – TIMSS, Programme for International Student Assessment – PISA, and so on). Since family history (except for the actual marital status of the parents), is not measured in these studies, divorce consequences have to be addressed indirectly by comparing children of single parents and children from complete families without having longitudinal information on the experiences of these children with their parents’ prior partnership dissolution. Hence, analyses like Pong et al. (2003) have to assume that single parents’ children have experienced a partnership dissolution (and not, for example, the death of one parent) and that the dates of these events are recent and more or less the same for all children in the target group – assumptions which may not be true in every case. Data from population registers (Steele et al., 2007) or from cohort studies of children (Sigle-Rushton et al., 2005) provide much more detailed information. Belgium, for example, is such a borderline case, which some researchers would classify as a member of the dual-earner model (see also Andreß, 2003: 635). Some basic statistical information about each country is summarized in the Data Appendix at the end of the book. They are organized into seven chapters, because the situation in Sweden and Finland is discussed in one chapter.

REFERENCES Amato, P.R. (2000), ‘The consequences of divorce for adults and children’, Journal of Marriage and the Family, 62, 1269–87. Amato, P.R. and B. Keith (1991a), ‘Parental divorce and adult well-being: a metaanalysis’, Journal of Marriage and the Family, 53, 43–58. Amato, P.R. and B. Keith (1991b), ‘Parental divorce and the well-being of children: a meta-analysis’, Psychological Bulletin, 110, 26–46. Andreß, H.-J. (2003), ‘Die ökonomischen Risiken von Trennung und Scheidung im Ländervergleich: ein Forschungsprogramm’, Zeitschrift für Sozialreform, 49, 620–51.

Introduction

23

Andreß, H.-J. and M. Bröckel (2007a), ‘Marital disruption in Germany: does the conservative welfare state care? Changes in material well-being and the effects of private and public transfers’, Schmollers Jahrbuch: Journal of Applied Social Science Studies, 127 (2), 193–226. Andreß, H.-J. and M. Bröckel (2007b), ‘Income and life satisfaction after marital disruption. What do measures of subjective well-being tell us about the economic consequences of divorce in Germany?’, Journal of Marriage and Family, 69 (May), 500–512. Andreß, H.-J., B. Borgloh, M. Güllner and K. Wilking (2003), Wenn aus Liebe rote Zahlen werden. Über die wirtschaftlichen Folgen von Trennung und Scheidung, Wiesbaden: Westdeutscher Verlag. Arts, W. and J. Gelissen (2002), ‘Three worlds of welfare capitalism or more? A state-of-the-art report’, Journal of European Social Policy, 12 (2), 137– 58. Ashenfelter, O. and R. Layard (eds) (1987), Handbook of Labor Economics, NorthHolland: Elsevier. Beck, N. and J. Hartmann (1999), ‘Die Wechselwirkung zwischen Erwerbstätigkeit der Ehefrau und Ehestabilität unter der Berücksichtigung des Sozialen Wandels’, Kölner Zeitschrift für Soziologie und Sozialpsychologie, 51, 655–80. Becker, G.S. (1981), A Treatise on the Family, Cambridge, MA: Harvard University Press. Bradbury, K. and J. Katz (2002), ‘Women’s labour market involvement and family income mobility when marriages end’, New England Economic Review, Q4, 41–74. Bryant, W.K. (2006), The Economic Organization of the Household, 2nd edn, Cambridge: Cambridge University Press. Burkhauser, R.V., G.J. Duncan, R. Hauser and R. Berntsen (1990), ‘Economic burdens of marital disruption: a comparison of the United States and the Federal Republic of Germany’, Review of Income and Wealth, 36, 319–33. Burkhauser, R.V., G.J. Duncan, R. Hauser and R. Berntsen (1991), ‘Wife or frau, women do worse: a comparison of men and women in the United States and Germany after marital dissolution’, Demography, 28, 353–60. Cigno, A. (1991), Economics of the Family, Oxford: Clarendon Press. Dewilde, C. (2002), ‘The financial consequences of relationship dissolution for women in Western Europe’, in E. Ruspini and A. Dale (eds), The Gender Dimension of Social Change. The Contribution of Dynamic Research to the Study of Women’s Life Courses, Bristol: Policy Press, pp. 81–110. Dewilde, C. (2004), ‘Vormen en trajecten van armoede in het Belgische en Britse welvaartsregime. Multidimensionele armoededynamieken bestudeerd vanuit de sociologie van de levensloop’, Proefschrift voorgelegd tot het behalen van de graad van Doctor in de Politieke en Sociale Wetenschappen, Antwerp: Universiteit Antwerpen. Diekmann, A. (1994), ‘Hat das steigende Ehescheidungsrisiko das berufliche Engagement von Frauen befördert?’, Soziale Welt, 45, 83–97. Dieleman, F.M. and R.J. Schouw (1989), ‘Divorce, mobility and housing demand’, European Journal of Population, 5, 235–52. Duncan, G. and S. Hoffman (1985), ‘A reconsideration of the economic consequences of marital dissolution’, Demography, 22, 485–97. Esping-Andersen, G. (1990), The Three Worlds of Welfare Capitalism, Cambridge: Polity Press.

24

When marriage ends

Esping-Andersen, G. (1999), Social Foundations of Postindustrial Economies, Oxford: Oxford University Press. Evers, A. (ed.) (2004), The Third Sector in Europe, Cheltenham UK and Northampton, MA, USA: Edward Elgar. Feijten, P. (2005), ‘Union dissolution, unemployment and moving out of homeownership’, European Sociological Review, 21 (1), 57–71. Feijten, P. and M. van Ham (2007), ‘Residential mobility and migration of the divorced and separated’, Demographic Research, 17, 623–54. Ferrera, M. (1996), ‘The southern model of welfare in social Europe’, Journal of European Social Policy, 6 (1), 17–37. Finnie, R. (1993), ‘Women, men and the economic consequences of divorce: evidence from Canadian longitudinal data’, Canadian Review of Sociology and Anthropology, 30, 205–41. Fritzell, J. (1990), ‘The dynamics of income distribution: economic mobility in Sweden in comparison with the United States’, Social Science Research, 19, 17–46. Furstenberg, F. and A. Cherlin (1991), Divided Families, Cambridge, MA: Harvard University Press. Gähler, M. (1998), Life after Divorce. Economic, Social and Psychological Well-being among Swedish Adults and Children Following Family Dissolution, Dissertation Series No. 32, Stockholm: Swedish Institute for Social Research. Gallie, D. and S. Paugam (eds) (2000), Welfare Regimes and the Experience of Unemployment in Europe, Oxford: Oxford University Press. Gauthier, A.H. (1996), The State and the Family: A Comparative Analysis of Family Policies in Industrialized Countries, Oxford: Clarendon Press. Gornick, J.C. and M.K. Meyers (2003), Families That Work: Policies for Reconciling Parenthood and Employment, New York: Russell Sage Foundation. Gornick, J.C., M.K. Meyers and K.E. Ross (1996), Supporting the Employment of Mothers: Policy Variations across Fourteen Welfare States, Working Paper no. 139, Luxembourg: Luxembourg Income Study (LIS). Gornick, J.C., M.K. Meyers and K.E. Ross (1997), ‘Supporting the employment of mothers: policy variations across fourteen welfare states’, Journal of European Social Policy, 7, 45–70. Haurin, D.R. (1989), ‘Women’s labor market reactions to family disruptions’, The Review of Economics and Statistics, 71 (1), 54–61. Holden, K.C. and P.J. Smock (1991), ‘The economic costs of marital dissolution. Why do women bear a disproportionate cost?’, Annual Review of Sociology, 17, 51–78. Holmans, A.E. (1990), ‘Housing demand and need generated by divorce’, in P. Symon (ed.), Housing and Divorce, Glasgow: University of Glasgow, Centre for Housing Research. Jarvis, S. and S.P. Jenkins (1999), ‘Marital splits and income changes: evidence from the British Household Panel Survey’, Population Studies, 53, 237–54. Jenkins, S.P. (2008), Marital Splits and Income Changes over the Longer Term, Working Paper No. 2008-07, University of Essex: Institute for Social and Economic Research. Johnson, W.R. and J. Skinner (1986), ‘Labor supply and marital separation’, The American Economic Review, 76, 455–69. Kalmijn, M. (2005), ‘The effects of divorce on men’s employment and social security histories’, European Journal of Population, 21, 347–66.

Introduction

25

Kaufmann, F.X. (2002), ‘Politics and policies towards the family in Europe. A framework and an inquiry into their differences and convergences’, in F.X. Kaufmann, A. Kuijsten, H.J. Schulze and K.P. Strohmeier (eds), Family Life and Family Policies in Europe, Oxford: Oxford University Press, pp. 427–500. Kelly, J.B. and R.E. Emery (2003), ‘Children’s adjustment following divorce: risk and resilience perspectives’, Family Relations, 53, 352–62. Kenny, L.W. (1983), ‘The accumulation of human capital during marriage by males’, Economic Inquiry, 21, 223–31. Leibfried, S. (1993), ‘Towards a European welfare state?’, in C. Jones (ed.), New Perspectives on the Welfare State, London: Routledge, pp. 133–56. Leitner, S. (2003), ‘Varieties of familialism. The caring function of the family in comparative perspective’, European Societies, 5 (4), 353–75. Manting, D. and A.M. Bouman (2006), ‘Short- and long-term economic consequences of the dissolution of marital and consensual unions. The example of the Netherlands’, European Sociological Review, 22, 413–29. McKeever, M. and N.H. Wolfinger (2001), ‘Reexamining the economic costs of marital disruption for women’, Social Science Quarterly, 82, 202–17. McManus, P. and T. DiPrete (2001), ‘Losers and winners: the financial consequences of separation and divorce for men’, American Sociological Review, 66, 246–68. Morgan, L.A. (1991), After Marriage Ends. Economic Consequences for Midlife Women, Newbury Park, CA, London and New Delhi: Sage. Mueller, R.E. (2005), ‘The effect of marital dissolution on the labour supply of males and females: evidence from Canada’, The Journal of Socio-Economics, 34 (6), 787–809. Orloff, A.S. (1993), ‘Gender and the social rights of citizenship: the comparative analysis of gender relations and welfare states’, American Sociological Review, 58, 303–28. Ostner, I. and J. Lewis (1995), ‘Gender and the evolution of European social policy’, in P. Pierson and S. Leibfried (eds), European Social Policy. Between Fragmentation and Integration, Washington, DC: Brookings Institution, pp. 159–93. Ott, N. (1999), ‘The economics of gender – Der neoklassische Erklärungsansatz zum Geschlechterverhältnis’, in B.V. Dausien, M. Herrmann, M. Oechsle, C. Schmerl and M. Stein-Hilbers (Hrsg.), Erkenntnisprojekt Geschlecht. Feministische Perspektiven verwandeln Wissenschaft, Opladen: Leske 1 Budrich, pp. 167–96. Peterson, R. (1996), ‘A re-evaluation of the economic consequences of divorce’, American Sociological Review, 61, 528–36. Peterson, R.R. (1989), Women, Work and Divorce, Albany, NY: State University of New York Press. Pong, S.L., J. Dronkers and G. Hampden-Thompson (2003), ‘Family policies and children’s school achievement in single- versus two-parent families’, Journal of Marriage and the Family, 65, 681–99. Poortman, A.R. (2000), ‘Sex differences in the economic consequences of separation: a panel study of The Netherlands’, European Sociological Review, 16, 367–83. Poortman, A.R. (2005), ‘Women’s work and divorce: a matter of anticipation? A research note’, European Sociological Review, 21, 301–09. Rainwater, L., M. Rein and M. Schwartz (1986), Income in the Welfare State. A Comparative Study of Family Income, Oxford: Clarendon Press.

26

When marriage ends

Rodgers, B. and J. Pryor (1998), Divorce and Separation: The Outcomes for Children, York: Joseph Rowntree Foundation. Rogers, S.J. (1999), ‘Wives’ income and marital quality: are there reciprocal effects?’, Journal of Marriage and the Family, 61, 123–32. Siaroff, A. (1994), ‘Work, welfare and gender equality: a new typology’, in D. Sainsbury (ed.), Gendering Welfare States, London: Sage, pp. 82–100. Sigle-Rushton, W., J. Hobcraft and K. Kiernan (2005), ‘Parental disruption and adult well-being: a cross cohort comparison’, Demography, 43, 427–46 Sørensen, A. (1994), ‘Women’s risk and the economic position of single mothers’, European Sociological Review, 10, 173–88. Steele, F., W. Sigle-Rushton and K. Oystein (2007), ‘Consequences of family disruption on children’s educational outcomes in Norway’, paper presented at the Fifth Meeting of the European Network for the Sociological and Demographic Study of Divorce, London, 17–18 September. Sullivan, O. (1986), ‘Housing movements of the divorced and separated’, Housing Studies, 1, 35–48. Symon, P. (1990), ‘Marital breakdown, gender and home ownership: the owneroccupied home in separation and divorce’, in P. Symon (ed.), Housing and Divorce, Glasgow: University of Glasgow, Centre for Housing Research. Uunk, W. (2004), ‘The economic consequences of divorce for women in the European Union: the impact of welfare state arrangements’, European Journal of Population, 20, 251–85. Wagner, M. and B. Weiß (2006), ‘On the variation of divorce risks in Europe: findings from a meta-analysis of European longitudinal studies’, European Sociological Review, 22, 483–500. Waite, L. and M. Gallagher (2000), The Case for Marriage: Why Married People Are Happier, Healthier, and Better Off Financially, New York: Broadway Books. Wallerstein, J.S., J.M. Lewis and S. Blakeslee (2000), The Unexpected Legacy of Divorce: The 25 Year Landmark Study, New York: Hyperion. Wasoff, F. and R.E. Dobash (1990), ‘Moving the family: changing housing circumstances after divorce’, in P. Symon (ed.), Housing and Divorce, Glasgow: University of Glasgow, Centre for Housing Research. Weitzman, L. (1985), The Divorce Revolution: The Unexpected Social and Economic Consequences for Women and Children in America, New York: Free Press. Zapf, W. (1984), ‘Welfare production: public versus private’, Social Indicators Research, 14, 263–74.

PART I

Theoretical, Methodological and Measurement Issues

1. Conceptualization and measurements of institutional contexts: a review Haya Stier INTRODUCTION Students of family and employment processes, as well as those interested in gender inequality, have drawn attention in recent years to the institutional contexts within which individuals make their work and family decisions. Country-specific variation in women’s labour force participation, differences in their working patterns along the life course and in market achievement, as well as diversity in family behaviour and its consequences, have led scholars to search for macro-level mechanisms that determine women’s economic behaviour and their economic interdependence (Gornick et al., 1998; Reskin, 2003; Stier et al., 2001; Van der Lippe and Van Dijk, 2002). Women’s involvement in paid employment, which largely determines their economic independence, is affected by the constraints imposed by family obligations on their ability to allocate more of their time and energy to market work. These constraints explain why, even in the absence of discrimination, women accumulate fewer economic resources than do men, and why women with family responsibilities (especially mothers) earn less than other women (Budig and England, 2001; Ginn and Arber, 1998). This line of explanation emphasizes women’s employment interruptions, which take place in relation to childbearing and child-rearing and to caring for adult family members (Moen and Han, 2001; Rosenfeld, 1996; Stier, 1996; Uunk et al., 2005); their tendency to work part-time in order to accommodate their family work (Blossfeld and Hakim, 1997; O’Reilly and Fagan, 1998; Warren and Walters, 1998); and their concentration in relatively few occupations which offer poor salaries and limited promotion prospects (Charles, 1992; Reskin and Roos, 1990). The focus on institutional arrangements that affect the interplay between work and family, then, is imperative, because women’s involvement in paid employment is an important determinant of their economic well-being, within or outside of marriage. 29

30

Theoretical, methodological and measurement issues

One consequence of women’s inferior status in the labour market and their economic dependency within the family is a deterioration of their economic status following divorce, a phenomenon which is evident in most industrialized countries. The state, through its welfare and legal systems, has developed mechanisms of support to alleviate the economic difficulties of single mothers (through direct or indirect transfers). Thus, in order to understand women’s economic standing following divorce, it is important to take into account both the arrangements that support women’s incorporation into paid employment and their economic gains in the labour market, as well as states’ transfer systems.1 The institutional arrangements within which women make their work and family decisions, and the way they influence women’s economic wellbeing, have been studied in different ways. This chapter has two major goals: first, to review the various approaches to exploring and understanding how macro-level mechanisms affect individuals’ behaviours and outcomes; and second, to discuss and propose measurements which are commonly employed in defining the ‘institutional context’. Studies which focus on institutional arrangements are comparative in nature. In the following section, I address two general strategies for comparative research: the case study approach, which often centres on comparisons of a few countries but could be elaborated to include multiple countries or country typologies, and a combination of micro- and macro-level analysis (see Van der Lippe and Van Dijk, 2002). In the second part of the chapter, I discuss specific institutional arrangements that directly or indirectly affect the economic position of women, and I review the major approaches to measuring them comparatively.

STRATEGIES FOR COMPARATIVE RESEARCH There are several ways to construct studies which measure the effect of institutional arrangements on work and family behaviours. In this section, I discuss two major approaches: country comparisons (the ‘case study’ strategy) and a micro–macro approach which offers a direct measure of the institutional context. The Case Study Approach Analysis of single countries The first strategy concentrates on analyses based on a small number of countries, studied either separately or comparatively. In a ‘single-country’ analysis, studies often focus on a specific research question (say, the economic

Conceptualization and measurements of institutional contexts

31

consequences of divorce) and study in depth the determinants affecting it within a specific context. Part II of this volume – which offers an in-depth account of countries’ characteristics and arrangements relevant to understanding the economic consequences of divorce – constitutes an example of this strategy. This approach does not provide a direct comparison between countries, but it is possible to draw implicit conclusions regarding the way institutions and cultural characteristics affect individuals’ behaviours and destinies. This is done either by focusing on a single country and comparing its unique characteristics against a conventional background (Del Boca, 1998; Papadopoulos, 1998), or, as researchers have done in this volume, by gathering several ‘single cases’ in which the institutional arrangements relevant to the research question are studied in depth, after which the differences in consequences and behaviours are attributed to context divergence (see, for example, Blossfeld, 1995; Blossfeld and Hakim, 1997; Drew et al., 1998; O’Reilley and Fagan, 1998; Shavit and Müller, 1998). In many cases, one country’s unique characteristics are compared to a group of countries such as the European Union, Organisation for Economic Co-operation and Development (OECD) countries, and so on (see, for example, Del Boca, 1998; Papadopoulos, 1998). In other studies, several single-country analyses are presented separately, and then similarities and differences are drawn (cf. Blossfeld, 1997) or even modelled (Shavit and Müller, 1998; Chapter 1). The main advantage of this research strategy is that it allows for an in-depth examination of national contexts in order to tap cross-country variation – an impossible task when analyses are based on many countries (Van der Lippe and Van Dijk, 2002). In this way, it is possible to gain detailed information on each country’s idiosyncrasies and to provide a careful analysis of the institutional background and countryspecific factors (see, for example, the detailed description of family law in each of the participating countries in the following chapters). It is also possible to conduct a detailed individual-level analysis for each country in order to understand causes and effects, using national datasets and appropriate detailed measurements of variables and indicators. Nonetheless, the in-depth case study approach has serious limitations. First, this strategy does not allow for a direct examination of structural effects on individual behaviours. It is not possible, in most cases, to control for the specific influence of one institutional arrangement or another. In this case, for example, differences in marriage patterns, employment behaviours or education cannot be attributed to a single contextual characteristic, but rather to a variety of institutions – some of which are complex entities that involve the interdependence of specific history, culture and organizations. Secondly, because analyses are based on nation-specific datasets (at the macro or micro levels), it is not easy to compare the outcomes. Finally, the choice of

32

Theoretical, methodological and measurement issues

countries (or the ‘standard comparison’) may be arbitrary, and hence not always instructive in understanding the ways in which social institutions and specific policies affect individual consequences. Multiple country comparison One way to more directly address the effect of institutional contexts on individual behaviours and outcomes is to design a comparison between a small number of countries. Hiilamo’s (in this volume) comparison of Finland and Sweden represents such a study design. The main assumption underlying such a strategy is that each of the countries represents a unique context, so that differences in individual behaviours or outcomes could be attributed to the particular country’s characteristics (see, for example, Brinton et al., 2001; Cooke, 2003; Diprete and McManus, 2000; Harkness and Waldfogel, 2003; Natti, 1995; Rosenfeld and Kalleberg, 1990). One advantage of comparing just a few countries is the possibility of obtaining and generating identical indicators and similar variables – a difficult task when more countries are involved or when we conduct unique singlecountry analyses. The countries should differ in some major features in order to make a meaningful comparison, but they should also have enough in common to isolate the effects of specific macro-level characteristics on individual behaviours. This situation is not always easy to find. Because countries differ on a large number of dimensions, it becomes almost impossible to determine which specific social institutions affect women’s market prospects and behaviours. Take, for example, a comparison between Sweden and the USA: Sweden has generous maternity and paternity leave policies, fully subsidized day-care facilities for children, a commitment to increased gender equality, a highly regulated labour market, and a high level of decommodification and defamilialization (Esping-Andersen, 1999). The USA is very different in all of these dimensions: The level of decommodification is low, family policy barely exists and the market is not regulated. Which of these institutional characteristics best explains the differences in women’s labour force participation and their work patterns along the life course? Which explains gender inequality? It is almost impossible to decipher the effect of each of these dimensions, so it is not easy to test the outcomes of specific institutional arrangements. One approach scholars use to overcome this problem is to choose similar countries which differ on a single dimension (see, for example, Natti’s (1995) comparison of women’s part-time employment in the Scandinavian countries; Brinton et al.’s (2001) comparison of women’s employment in South Korea and Taiwan; and Cooke (2003) on Eastern and Western Germany). However, this type of research framing limits even more the number of countries that can be included in the analysis.

Conceptualization and measurements of institutional contexts

33

Macro–Micro Analysis Country typologies As a result of the limited information that can be obtained from a comparison of a few countries and the lack of systematic rules for comparison, several typologies have been attempted which identify similarities and differences in institutional contexts among countries. The best known typology is that proposed by Esping-Andersen (1990; 1999) which, using different institutional and ideological characteristics, separates countries into three specific welfare regimes. Esping-Andersen’s basic distinction between the three regime types focuses on the ways in which social welfare is produced and allocated between the state, the market and the family. The contextual differences – best characterized by the level of decommodification and, in the more recent work, by the level of defamilialization – are expected to affect women’s work behaviour and their pathways to achieving economic independence in different ways. Other categorizations, such as the one proposed by Lewis (1992), focus on the familial ideology underlying different welfare arrangements as the basis for distinguishing groups of countries. Accordingly, countries can be ordered on a range of dimensions from a male breadwinner model to a dual-earner model. The male breadwinner ideology assumes a gendered division of labour whereby men are responsible for market work to provide for their families, while women do the caregiving work. In the dual-earner model, men and women are both perceived as workers, so women’s work for pay is encouraged. Similarly, Pfau-Effinger (1998) and Crompton (1999) suggested distinguishing countries based on their ‘gender order’, or ‘gender regime’. The underlying assumption is that gender relations are constructed in particular social contexts in a complex way, where ‘the major dimension structuring the gender division of labour has been the gender coding of caring and market work’ (Crompton, 1999: 204). Hence, societies differ in their dominant form of gender organization, ranging from the historic ‘male breadwinner-female caregiver’ model to the still unattained ‘dual-earner/dual-caregiver’ model (Gornick and Meyers, 2003). A handful of studies have utilized these (and other) typologies, to compare women’s employment behaviour and market positions across countries which typically fall into specific categories of the various typologies (for example, Andreß et al., in this volume; Pfau-Effinger, 1998; Stier et al., 2001; Uunk, 2004; Van Damme and Uunk, in this volume). The advantage of this strategy is that it allows researchers to pool data from a large number of countries without necessarily providing an in-depth and separate analysis for each one.2 However, using a one-dimensional categorization (as most typologies do) is problematic in many respects. Uunk

34

Theoretical, methodological and measurement issues

(2004) argues that not all countries fall clearly into one category; and once classified, within-category variation in countries’ policies and behaviours is considerable (Gornick et al., 1997; Stier et al., 2001). Gornick et al. (1997) demonstrated variation among countries in employment-related policies which do not entirely correspond to existing typologies of welfare regimes or gender relations. They categorized countries by the level of support they provide for working women and found higher support for women in Sweden than in Norway – two social-democratic countries. In their study of women’s employment patterns and their economic consequences, Stier et al. (2001) combined Esping-Andersen’s three-category welfare regimes with Gornick et al.’s (1997) division of countries according to the level of support they provide for working mothers, and suggested a more elaborated way to capture country variation between welfare regimes as well as within them. This strategy is useful for studies which focus on the interplay between markets and other institutions, such as families, as it allows for the combination and control of multiple dimensions. However, the major drawback is the need to find a sizeable number of countries to represent each of the categories. This is not always possible: For example, in Stier et al.’s study, there was no country which fitted the category of a liberal welfare regime with high support for women’s employment. Similarly, only one country (Canada) fell into the liberal-intermediate policy category. In addition, different institutions and arrangements that fall into one typology may affect the various aspects of family–employment relations in different ways. For example, family-supportive policies are expected to encourage women’s employment in general (and they do, as numerous studies suggest); however, generous maternity leaves may in fact be detrimental to women’s ability to accumulate resources through market work, since they provide incentives to remain separated from market work for a long time (Gornick and Meyers, 2003; Stier and Mandel, 2003). Typologies, then, are crude measures that may conceal conflicting influences on individual behaviours which cannot easily be discerned. Finally, this strategy – which has become popular in the study of women’s employment behaviour and their market prospects – is also problematic since a number of typologies exist, each of which relies on different theoretical bases and utilizes different institutional characteristics. It is not easy to choose the ‘right’ typology, which may at times depend on the specific outcome in question. A direct account of macro-level effects A third approach tries to overcome the above mentioned drawbacks in conducting comparative research. It is based on a comparison of a large number of countries and on a quantitative examination of the (direct)

Conceptualization and measurements of institutional contexts

35

effects of specific institutional characteristics on social and economic behaviours and outcomes. This approach, then, offers a test for the direct effects of various institutional arrangements – such as specific policies or country characteristics – on country variation in market outcomes. One example of a study that used macro–micro linkages in its analysis is the work of Gornick et al., who used a variety of indicators to construct an index of family policy which then was applied to explain the effect of young children on their mothers’ employment (Gornick et al., 1998). Likewise, using more developed multi-level statistical models, Mandel and Semyonov (2005) tested the consequences of family-supportive policies on women’s labour force participation, their concentration in female-type occupations, and gender inequality in pay. Similarly, Uunk et al. (2005) tested for the net effect of childcare arrangements on women’s labour supply, controlling for alternative explanations (that is, economic affluence and egalitarian gender ideology). A combination of the effects of welfare state typology and state-specific characteristics on women’s economic standing following divorce can also be found in Uunk (2004). His findings suggest an effect of welfare state arrangements (depicted by welfare regimes) as well as a direct influence from specific measures of welfare and childcare provisions. Several studies have examined the effect of policies, country characteristics, and other macro-level variables on a household’s division of labour (Fuwa, 2004; Hook, 2006; Stier and Lewin-Epstein, 2007). This approach allows researchers to explain country variation in specific behaviours (for exmaple, women’s labour force participation, the division of household labour and women’s economic standing following a divorce) as a product of country-specific institutional arrangements (such as policies, family structure, labour market characteristics, and more). Moreover, using data on countries as well as on individuals within those countries makes it possible to test for macro–micro interactions. That is, this approach allows researchers to examine how institutional contexts mediate the relationship between two individual-level characteristics – for example, the effect of gender on earnings (Mandel and Semyonov, 2005). The advantage of this method is clear, since it facilitates disentangling the influence of individuallevel and country-level characteristics on behaviours and outcomes (Van der Lippe and Van Dijk, 2002) and directly testing the extent to which institutional arrangements (at the country level) account for differences in the outcome variable after taking into account the effect of individual-level characteristics. However, there are limitations to this approach. First, the measurement of institutional arrangements, while direct and clearly defined, is rough because it is hard to find comparable indicators for a large number of countries. This limits researchers’ ability to study contextual effects in

36

Theoretical, methodological and measurement issues

depth (Van der Lippe and Van Dijk, 2002). Moreover, it is also difficult to measure individual-level characteristics comparatively because the level of details (for example, in the measure of education, labour force involvement and familial arrangements) may vary across countries. Consequently, some information which could otherwise be used in in-depth individuallevel analysis is lost. A second limitation results from the method used. In a multi-level analysis, which utilizes fixed and random effects, a small number of countries may cause unstable estimations, and only a limited number of country indicators may be included in the empirical model. Increasing the number of countries is not always a viable solution, since it is not easy to combine macro- and micro-level data for a large enough number of countries. The use of indices – such as that proposed by Gornick et al. (1997), Gornick and Meyers (2003) and Mandel and Semyonov (2005), can solve part of the problem. However, as I will argue below, indices may also create problems, especially when different components of an index affect the outcome variable in different ways. A problem common to all methods of analysis is our inability to determine causal relationships. To what extent do institutions and policies affect individual behaviours? Since the measures are crude and often cannot take the time dimension into account, it is hard to determine whether the specific institutional arrangement affects individual behaviours, or vice versa. Childcare arrangements, for example, have been found to affect women’s labour force participation. But it is plausible to argue that an ample supply of female workers creates pressures and demands for the provision of more childcare facilities (Gornick and Meyers 2003). The problem of causality is almost impossible to overcome with the data available: its resolution demands longitudinal data on institutional arrangements as well as individual behaviours. The use of panel data at specific periods of time may provide a way to sustain the directionality of effects from macro-level characteristics to micro-level behaviours.

HOW CAN WE MEASURE THE INSTITUTIONAL CONTEXT? With the massive entry of women to the labour force, greater awareness has been given to social policies and institutional arrangements that could contribute to alleviating gender inequality and reducing work–family incompatibilities. In this section, I consider three types of policies that directly or indirectly affect the economic status of women and have important implications for women’s economic standing following divorce. While these policies have different aims and may reflect different assumptions regarding

Conceptualization and measurements of institutional contexts

37

the relationships between gender, family and the state, they also coincide at different levels. Policies of the first type are aimed directly at eliminating gender inequality in market opportunities and in pay. They are directed to labour market processes and pertain to women who already participate in paid employment. Direct measures such as anti-discrimination acts have been implemented in most industrial countries. These include equal opportunity in hiring and promotion and affirmative action, as well as legislation that deals directly with discrimination (Blakemore and Drake, 1996). This policy type presupposes gender similarity in market involvement: in other words, it pertains mainly to women who are willing to adopt a ‘male’ work pattern, based on continuous attachment to the labour market, long working hours and high motivation for career building (Gornick and Meyers, 2003; Hakim, 1999). Policies of the second type are designed to increase the employability of women, especially those constrained by family demands, by providing the necessary conditions for combining work and family. These policies – usually termed ‘employment supportive’ or ‘women-friendly’ – are found frequently in most industrialized countries at present. Nonetheless, the underlying assumption regarding gender roles and women’s economic independence and the specific policies that have been implemented differ considerably among countries (Brewster and Rindfuss, 2000; EspingAndersen, 1999; Gornick and Meyers, 2003; Hantrais, 1997; Morgan and Zippel, 2003; Orloff, 1993, 2001; Stier et al., 2001). Some of these policies are indeed constructed to encourage women’s – and especially mothers’ – employment, while others are intended, directly or indirectly, to preserve the gendered division of labour and the role of families in society (Morgan and Zippel, 2003). These policies include various caretaking leave arrangements, such as maternity and paternity leave, tax schemes which include a deduction for parents, subsidized day-care facilities, long school days, and so on (Gornick et al., 1997). In the Scandinavian countries, for example, these policies are aimed at alleviating women’s reliance on the family and increasing their economic independence (Esping-Andersen, 1999; Hantrais, 1997). In other countries, such as Germany or Italy, where the family is still seen as the major social institution responsible for individuals’ well-being (Hantrais, 1997), these policies are implemented so as to preserve the gendered division of labour, and hence contribute only partly to women’s economic independence (Orloff, 2001). Employment-supportive policies are perceived as providing the necessary conditions for women to participate in paid employment, and hence are expected to reduce gender inequalities. Numerous studies have shown that such policies do facilitate the employment of mothers and make possible its continuation at times when the demand for family time is high (Gornick et

38

Theoretical, methodological and measurement issues

al., 1998; Stier et al., 2001; Uunk et al., 2005). Access to money of their own contributes to women’s economic independence within marriage, and still more so after a family breakdown. However, the long-term consequences of these policies are not as straightforward as they might seem, because while they allow women to maintain a continuous attachment to the labour force, they also contribute indirectly to the preservation of a gendered division of labour and limit women’s access to greater economic resources (Mandel and Semyonov, 2005; 2006; Stier and Mandel, 2003). Generous maternity leave (taken almost entirely by women) detaches mothers from their work place (Gornick and Meyers, 2003); part-time jobs are marginal positions with low economic rewards and few prospects for future advancement (Warren, 2004); and a high degree of gender segregation – as exemplified in countries with a highly developed and relatively large public sector – while offering more convenient work conditions for mothers, also limits their access to lucrative jobs (Mandel and Semyonov, 2006). In the long run, to what degree these women-friendly policies actually promote women’s economic independence is unclear. A third route to improving women’s economic position and protecting their economic standing is state support for economically vulnerable citizens, either directly, or indirectly by providing universal transfers (for example, child allowances) to all members of society. In many countries – especially those characterized by a high degree of decommodification – allowances are provided for single mothers whether or not they participate in paid employment. Scholars have argued that countries which adopt a universal transfer system with a relatively high level of grants reduce poverty in general, and gender inequality in poverty rates in particular (Casper et al., 1994; Korpi and Palme, 1998). To understand women’s degree of economic independence – and specifically, their economic standing following divorce – it is imperative to take into account the state transfer system, its nature and arrangements. This type of policy reduces the costs of divorce but does not necessarily contribute to women’s ability to accumulate resources along their life course or to improve their economic independence in the long run, since they then become dependent on the state.

WHICH INSTITUTIONAL ARRANGEMENTS ARE RELEVANT TO THE STUDY OF DIVORCE OUTCOMES? The economic well-being of women, whether in or out of marriage, depends mainly on three major sources of income: the first is derived from

Conceptualization and measurements of institutional contexts

39

the market, the second from state transfers, and the third from family support. In this section, I discuss specific measures of institutional arrangements that can be expected to contribute to the economic independence of women, focusing mainly on market and state arrangements. To put it differently, women’s economic position after they divorce is a product of the resources they managed to accumulate while they were married, through market work and/or other sources (mainly independent income derived from assets, savings and family transfers), and the extent to which the state, through its transfer system, reduces their reliance on market income. The measures I propose in the following section are therefore direct and indirect indicators of specific policies and arrangements that allow women to gain access to independent resources. Earlier in this chapter, I referred to two types of policies: those aimed at encouraging women’s employment and facilitating their involvement in the labour market, and those dealing directly with discrimination and constraints in the labour market. Along the same lines, I now distinguish between measures of employment-supportive policies and those that serve directly to enhance gender equality in the labour market. Several measures of policies that support mothers’ employment were suggested by Gornick and Meyers (2003). Among these are various family leave policies, childcare arrangements and working-time regulations. Family leave policies mainly include maternity and paternity leaves and a more general caregiving leave for family reasons (Gornick and Meyers, 2003: 114; Morgan and Zippel, 2003). Various aspects of leave policies can be measured: whether they are implemented in a particular country; who is eligible and under which conditions (for example, prior market work); the rate of coverage; and the benefit level. Information on family leave schemes is available for most industrialized countries, and in some cases (for example, fully paid maternity leave) for non-industrialized countries as well (see, for example, Kamerman, 2000). Working-time regulations are important in determining men’s and women’s time allocation between paid and unpaid work and the extent to which women are forced to work in a segregated labour market because of the time constraints imposed on them by family demands. Most countries have working-time policies (for example, normal and maximum working hours), and although the variation among them is relatively small, countries differ considerably in their ways and means of enforcing such regulations. While there is ample evidence of country variation in actual working time and time arrangements (Plantenga, 2004), relatively few studies have examined the effect of time regulations on men’s and women’s working lives (see, for example, Manfred, 2002). More often, it has been argued that the availability and demand for reduced-time employment and

40

Theoretical, methodological and measurement issues

flexible-time arrangements should be seen as a way to incorporate women into paid employment, thereby giving them access to independent income (see Gornick and Meyers, 2003). While working-time regulations and provisions can be measured directly, most studies that look at women’s employment and family-time organization make use of indirect measures – mostly the availability of part-time work. Reduced hours of employment is not a clear and direct policy measure. Nonetheless, it is an important employment feature which, on the one hand, facilitates women’s work but, on the other hand, limits their access to well-paid jobs. Countries differ in their rates of part-time employment and in the legal arrangements and benefits associated with this mode of work (Gornick and Meyers, 2003). Similarly, non-standard working hours (for example, flexible hours, shift work) can serve as a measure of incorporating women – especially those with high family demands – into the labour market. The last component of ‘employment-supportive policies’, childcare arrangements, can be measured directly (through school-day schedules, tax benefits for care or financial support for childcare) or indirectly (for example, the percentage of children in fully subsidized day-care facilities). See, for example, Gornick et al. (1997), Stier and Mandel (2003) and Stier and Lewin-Epstein (2007). Measures of gender equality in the labour market are also present in all countries. For example, most industrialized countries have equalopportunity regulations, and some have more direct measures of affirmative action. While studies could use direct indicators of such policies and arrangements, these measures are complicated and can hardly provide information on law enforcement. It is imperative to use indirect measures of inequality in general – and gender inequality in pay in particular – to indicate country variation on this dimension. It can be argued that the level of gender inequality in pay is a result of different types of policies, not necessarily those directed toward the elimination of discrimination. For example, because unions and government regulate labour markets and reduce overall inequality, they also affect the gender wage gaps, as suggested and demonstrated by Blau and Kahn (1995). Nonetheless, these indirect measures could serve as acceptable indicators of labour marketbased policies and arrangements. To summarize, measures of institutional arrangements should include both direct (that is, specific policies and regulations) and indirect indicators (for example, behavioural outcomes at the macro level). State transfers can be indicated by different welfare policies, such as support for single mothers, child allowances, income maintenance regulations and the extent to which the state is involved in enforcing alimony arrangements. These different arrangements are relatively comparable

Conceptualization and measurements of institutional contexts

41

across countries. Uunk (2004), for example, in his comparative study of the economic consequences of divorce, uses the guaranteed net allowance for single-parent families as a measure for social welfare. Other indicators of child support, state liability and law enforcement are also available (see, for example, Corden and Meyer, 2000). Money transfers can also take place through private family arrangements (with or without state involvement). Child support transfers from non-resident fathers are a most frequent example. Although countries differ in the regulations and control of such payments – as is evident from the various chapters in this book – it is still possible to indicate them comparatively (Corden and Meyer, 2000). However, family support could also be less formal and come from different sources (for example, from parents, siblings, former spouses and other relatives) and is not always easy to measure (for example, in-kind support or informal care for children). Thus, measuring the nature and network of family support systems is still a challenging goal for comparative studies. Since, as I argued earlier, it is often too complicated to incorporate many indicators into a comparative analysis, several studies have suggested using an assemblage of indices in order to capture a more general construct of policies. Gornick et al.’s (1997) index of family-supportive policies was based on information from 14 European countries and included several components of public parental leave schemes (for example, legislated job protection, weeks of paid maternity leave, wage replacement rate, eligibility terms, availability of extended leave and paternity benefits) and childcare policies (including tax relief for childcare, guaranteed childcare for the 0–2 and 3–5-year-old age groups; childcare expenditures, percentage of children aged 0–2 in publicly funded childcare, percentage of children aged 3–5 in publicly funded childcare, percentage of children aged 5 in pre-primary school, and percentage of children in public after-school programmes). Gornick and Meyers (2003) later proposed two types of indices, one pertaining to families with pre-school children and another that applies to families with school-age children. A similar index was constructed by OECD researchers (OECD, 2001), which includes maternity pay entitlements, childcare coverage for children under 3, voluntary family leave, flex-time working hours and voluntary part-time employment. In their studies based on 20 industrialized countries, Mandel and Semyonov (2005, 2006) simplified this approach by compiling an index with only three indicators: the number of fully paid weeks of maternity leave, the percentage of children in public day-care centres and the size of the public sector. Fuwa (2004) used the United Nations (UN) measure of women’s empowerment as a general measure to denote women’s status in society. It is also possible to construct indices that will take into account elements of policies

42

Theoretical, methodological and measurement issues

that support women’s employment and elements of policies and arrangements that reduce gender inequality in the labour market. However, it is important to bear in mind the disadvantage of this strategy for those interested in understanding which specific policies or institutions affect women’s labour market behaviour and their economic prospects, and which are only remotely related to their economic well-being. Moreover, while the different components of an index may be highly correlated, their individual effects on different outcomes may differ. Consequently, an index may conceal the effect of specific indicators, each of which may operate in a contradictory way, thus yielding non-significant effects for the index as a whole (see Mandel and Semyonov, 2005; Stier and Mandel, 2003). For example, countries that provide generous and fully paid maternity leave tend also to provide and subsidize high-quality childcare facilities. However, while good and affordable day-care facilities may encourage women to enter demanding full-time occupations that yield high salaries and thus contribute to women’s economic independence, generous maternity leave may have the opposite effect. These policies encourage women to enter the labour force but not necessarily to take on full-time, demanding jobs (Gornick and Meyers, 2003; Stier and Mandel, 2003). Or, as Hansen (1995) and others have argued, they may encourage women to find employment in female-type occupations that tolerate long separations from market work. Consequently, they may not contribute as much to women’s economic independence. In terms of theoretical and policy considerations then, measuring specific policies and arrangements seems more useful than using a combined measure of policy that makes it impossible to distinguish the effect of one type of policy from another.

CONCLUSIONS The study of individual behaviours and their related outcomes should take into account the context within which they take place. Comparative studies are thus important for understanding how institutional and cultural contexts shape individual lives and determine their consequences. Yet identifying the relevant contexts and assessing their impact on specific behaviours and outcomes is not easy. As I outlined in this chapter, there are several approaches available to researchers, but it is difficult to answer the question of what is the ‘best’ strategy for comparative research. The choice of strategy depends first and foremost on the specific research question. In studies that focus on an in-depth comparison of contexts, researchers should choose the case study comparison. This way of conducting comparative research allows us to understand how different institutions and

Conceptualization and measurements of institutional contexts

43

policies are related to specific outcomes and the general conditions under which different consequences and behaviours appear. Nonetheless, one has to bear in mind that this strategy helps to outline a general story but not to identify specific mechanisms that may affect individual behaviours. In this volume, for example, one major goal was to present an in-depth account of country characteristics and institutional arrangements related to divorce and divorce outcomes. Thus, it is possible to compare different contexts that affect the consequences of divorce without necessarily testing the effect of specific policies or arrangements on the outcome in question. When the research goal is to understand a specific effect of one or more institutions on specific outcomes, a different strategy is more suitable. For example, a study interested in the effect of childcare provisions on women’s labour force participation and work patterns should adopt a multivariate micro–macro approach (see, for example, Mandel and Semyonov 2005). Similarly, in her chapter on residential mobility following divorce, Dewilde (in this volume) tested the effect of specific policies and arrangements (support for families and the structure of housing market) on women’s housing arrangements following their divorce. In between these two examples there is a wide variety of strategies, each of which has its advantages and disadvantages; each answers different questions and provides different insights. Regarding specific measures of ‘institutional arrangements’, most of the choices researchers make depend on the data available at both the macro and the micro levels and on the extent to which the information can be used in a comparative framework. Studies have utilized a variety of institutional indicators and macro-level characteristics, and it is difficult to conclude which of these best describe policies, arrangements and structures. Moreover, the indicators that are commonly available for all countries are often only a rough measure of the ‘context’, because they pertain to specific periods, and are not sensitive enough to the dynamic nature of policies and arrangements. Finally, most comparative studies (whether they use individual cases, multi-country comparisons or a macro–micro analysis) are limited in their ability to determine causality between institutional arrangements and individual-level behaviours and outcomes. In many cases, the time frame is not well specified, so it is not clear whether a specific policy affected behaviour or whether pressure from below resulted in new policy programmes. It is also not always clear which specific policy is affecting people’s behaviour, since many programmes are composed of a variety of arrangements and provisions, where some effects are more difficult to disentangle than others. That said, it is still extremely interesting to assess these relationships in order to better understand the role of macro-level processes and arrangements as they shape the destinies of individuals.

44

Theoretical, methodological and measurement issues

NOTES 1. It is important to note that the family also serves as a source of economic support for divorced women, either through direct payments from former husbands, parents or other relatives, or through services to accommodate women’s employment. However, the role of family as a source of economic support is beyond the scope of this chapter. 2. Many studies that used Esping-Andersen’s (or other) categorizations actually used only one country to represent each category (for example, Pfau-Effinger, 1998). This is problematic because it does not provide any support for the theoretical assertion underlying the specific categorization. A stronger test can be provided by pooling together several countries which fall under a specific category.

REFERENCES Blakemore, K. and R. Drake (1996), Understanding Equal Opportunity Policies, London: Prentice Hall. Blau, F.D. and L.M. Kahn (1995), ‘The gender earning gap, some international evidence’, in R.B. Freeman and L.F. Katz (eds), Differences and Changes in Wage Structures, Chicago, IL: University of Chicago Press. Blossfeld, H.-P. (1995), The New Role of Women: Family Formation in Modern Societies, Boulder, CO: Westview Press. Blossfeld, H.-P. (1997), ‘Women’s part-time employment and the family cycle: a cross national comparison’, in H.-P. Blossfeld and C. Hakim (eds), Between Equalization and Marginalization: Women Working Part-time in Europe and the United States, Oxford: Oxford University Press, pp. 315–24. Blossfeld, H.-P. and C. Hakim (1997), Between Equalization and Marginalization: Women Working Part-time in Europe and the United States, Oxford: Oxford University Press. Brewster, K.L. and R.R. Rindfuss (2000), ‘Fertility and women’s employment in industrialized nations’, Annual Review of Sociology, 26, 271–96. Brinton, M.C., Y.-J. Lee and W.L. Parish (2001), ‘Married women’s employment in rapidly industrializing societies: South Korea and Taiwan’, in M.C. Brinton (ed.), Women’s Working Lives in East Asia, Stanford, CA: Stanford University Press, pp. 38–69. Budig, M.J. and P. England (2001), ‘The wage penalty for motherhood’, American Sociological Review, 66 (2), 204–25. Casper, L.M, S.S. McLanahan and I. Garfinkel (1994), ‘The gender-poverty gap: what we learn from other countries’, American Sociological Review, 59, 594–605. Charles, M. (1992), ‘Cross national variation in occupational sex segregation’, American Sociological Review, 57, 483–502. Cooke, P.L. (2003), ‘The gendered division of domestic labour and family outcomes in East and West Germany’, paper presented for ‘European Societies or European Society?’ Euro Conference on Institutions and Inequality, Helsinki, 20–24 September. Corden, A. and D.N. Meyer (2000), ‘Child support policy regimes in the United States, United Kingdom, and other European countries: similar issues, different approaches’, Focus, 21 (1), 72–9.

Conceptualization and measurements of institutional contexts

45

Crompton, R. (ed.) (1999), Restructuring Gender Relations and Employment: The Decline of the Male Breadwinner, Oxford: Oxford University Press. Del Boca, D. (1998), ‘Labour policies, economic flexibility and women’s work: the Italian experience’, in E. Drew, R. Emerek and E. Mahon (eds), Women, Work and the Family in Europe, London: Routledge, pp. 124–30. Diprete, T.A. and P.A. McManus (2000), ‘Family change, employment transitions and the welfare state: household income dynamics in the U.S. and Germany’, American Sociological Review, 65, 343–70. Drew, E., R. Emerek and E. Mahon (1998), Women, Work and the Family in Europe, London: Routledge. Esping-Andersen, G. (1990), The Three Worlds of Welfare Capitalism, Princeton, NJ: Princeton University Press. Esping-Andersen, G. (1999), Social Foundations of Postindustrial Economies, New York: Oxford. Fuwa, M. (2004), ‘Macro-level gender inequality and the division of household labor in 22 countries’, American Sociological Review, 69 (6), 751–67. Ginn, J. and S. Arber (1998), ‘How does part-time work lead to low pension income?’, in J. O’Reilly and C. Fagan (eds), Part-time Prospects, London: Routledge, pp. 156–73. Gornick, J., M.K. Meyers and K.E. Ross (1997), ‘Supporting the employment of mothers: policy variation across fourteen welfare states’, Journal of European Social Policy, 7 (1), 45–70. Gornick, J., M.K. Meyers and K.E. Ross (1998), ‘Public policies and the employment of mothers: a cross national study’, Social Science Quarterly, 79, 35–54. Gornick, J.C. and M.K. Meyers (2003), Families that Work: Policies for Reconciling Parenthood and Employment, New York: Russell Sage Foundation. Hakim, C. (1999), ‘Models of the family, women’s role and social policy: a new perspective from preference theory’, European Societies, 1 (1), 33–58. Hansen, M.N. (1995), ‘The vicious circle of the welfare state? Women’s labour market situation in Norway and Great Britain’, Comparative Social Research, 15, 1–34. Hantrais, L. (1997), ‘Exploring relationships between social policy and changing family forms within the European Union’, European Journal of Population, 13, 339–79. Harkness, S. and J. Waldfogel (2003), ‘The family gap in pay: evidence from seven industrialized countries’, Research in Labour Economics, 22, 369–413. Hook, J. (2006), ‘Care in context: men’s unpaid work in 20 countries, 1965–2003’, American Sociological Review, 71 (4), 639–60. Kamerman, S.B. (2000), ‘Parental leave policies: an essential ingredient in early childhood education and care policies’, Social Policy Report. A Publication of the Society for Research in Child Development, 14 (2), 3–16. Korpi, W. and J. Palme (1998), ‘The paradox of redistribution and strategies of equality: welfare state institutions, inequality, and poverty in the western countries’, American Sociological Review, 63 (5), 661–87. Lewis, J. (1992), ‘Gender and the development of welfare regimes’, Journal of European Social Policy, 2 (3), 159–73. Mandel, H. and M. Semyonov (2005), ‘Family policies, wage structures and gender gaps: sources of earnings inequality in 20 countries’, American Sociological Review, 70 (6), 949–68. Mandel, H. and M. Semyonov (2006), ‘A welfare state paradox: state interventions

46

Theoretical, methodological and measurement issues

and women’s employment opportunities in 22 countries’, American Journal of Sociology, 111 (6), 1910–49. Manfred, A. (2002), ‘The relationship between paid work and parenthood – a comparison of structures, concepts and developments in the UK and Austria’, Community, Work, and Family, 5 (2), 203–18. Moen, P. and S.-K. Han (2001), ‘Gendered careers: A life course perspective’, in R. Hertz and N.L. Marshall (eds), Working Families: The Transformation of the American Home, Berkeley, CA: University of California Press, pp. 42–57. Morgan K.J. and K. Zippel (2003), ‘Paid to care: the origins and effects of care leave policies in Western Europe’, Social Politics, 10 (1), 49–85. Natti, J. (1995) ‘Are women trapped in part-time jobs?’, Labour, 9 (2), 343–58. O’Reilly, J. and C. Fagan (eds) (1998), Part-time Prospects, London: Routledge. Organisation for Economic Co-operation and Development (OECD) (2001), ‘Balancing work and family life: helping parents into paid employment’, Employment Outlook, 129–66. Orloff, A.S. (1993), ‘Gender and the social rights of citizenship: the comparative analysis of gender relations and welfare states’, American Sociological Review, 58 (3), 303–28. Orloff, A.S. (2001), ‘Gender equality, women’s employment: cross-national patterns of policy and politics’, paper prepared for the meeting of International Sociological Association Research Committee 19, Poverty, Social Welfare and Social Policy ‘Old and New Social Inequalities: What Challenges for Welfare States?’, University of Oviedo, Spain, 6–9 September. Papadopoulos, T.N. (1998), ‘Greek Family Policy from a Comparative Perspective’, in E. Drew, R. Emerek and E. Mahon (eds), Women, Work and the Family in Europe, London: Routledge, pp. 47–57. Pfau-Effinger, B. (1998), ‘Culture or Structure as explanations for Differences in Part-time Employment in Germany, Finland and the Netherlands?’, in J. O’Reilly & C. Fagan (eds), Part-time Prospects, London: Routledge, pp. 177–98. Plantenga, J. (2004), ‘Changing work and life patterns: examples of new working arrangements in the European member states’, Advances in Life Course Research, 8, 119–35. Reskin, B. and P. Roos (1990), Job Queues, Gender Queues, Philadelphia, PA: Temple University Press. Reskin, B.F. (2003), ‘Including mechanisms in our models of ascriptive inequality’, American Sociological Review, 68 (1), 1–21. Rosenfeld, R. (1996), ‘Women’s work histories’, Population and Development Review, 22 (suppl.), 199–222. Rosenfeld, R. and A.L. Kalleberg (1990), ‘A cross-national comparison of the gender gap in income’, American Journal of Sociology, 96 (1), 69–106. Shavit, Y. and W. Müller (1998), From School to Work: A Comparative Study of Educational Qualifications and Occupational Destinations, Oxford: Clarendon Press. Stier, H. (1996), ‘Continuity and change in women’s occupations following first childbirth’, Social Science Quarterly, 77, 60–75. Stier, H. and N. Lewin-Epstein (2007), ‘Policy effects on the division of housework’, Journal of Comparative Policy Analysis, 9 (3), 235–59. Stier, H. and H. Mandel (2003), ‘Inequality in the family: the institutional aspects of wives’ earning dependency’, paper presented at the ISA Research Committee on Stratification (RC28), Tokyo, 1–3 March.

Conceptualization and measurements of institutional contexts

47

Stier, H., N. Lewin-Epstein and M. Braun (2001), ‘Welfare regime, family-supportive policy, and women’s employment along the life course’, American Journal of Sociology, 106 (6), 1731–60. Uunk, W.J.G. (2004), ‘The economic consequences of divorce for women in the European Union: the impact of welfare state arrangements’, European Journal of Population, 20, 251–85. Uunk, W.J.G, M. Kalmijn and R. Muffels (2005), ‘The impact of young children on women’s labour supply: a reassessment of institutional effects in Europe’, Acta Sociologica, 48 (1), 41–62. Van der Lippe, T. and L. van Dijk. (2002), ‘Comparative research on women’s employment’, Annual Review of Sociology, 28, 221–41. Warren, T. (2004), ‘Working part-time: achieving successful work–life balance?’, British Journal of Sociology, 55 (1), 99–122. Warren, T. and P. Walters (1998), ‘Appraising a dichotomy: a review of the use of “part-time/full-time” in the study of women’s employment in Britain’, Gender, Work and Organization, 5 (2), 102–18.

PART II

Single Country Studies

2.

Germany: will the male breadwinner model survive? Dina Hummelsheim

INTRODUCTION Even though partnership stability in the past few decades has become doubtful, it still provides for a high degree of social security – an aspect that is often underestimated. Although partnerships are gainful and advantageous, the consequences in case of a break-up are often severe, emotionally and financially. In view of the general rising tendency towards divorce, more and more people are affected by this phenomenon and accordingly by its consequences. Germany like many other Western industrialized countries experiences increases in divorce rates. This development is closely linked with an increasing employment orientation of women. But at the same time German women are, to an extremely high degree, confronted with difficulties to reconcile work and family. This means that some of them abandon or postpone the wish to have children, while others retreat – temporarily or even long-term – from the labour market. The latter is maybe the most important reason for the comparatively bad financial situation of separated and divorced German women with children. Most of the difficulties to reconcile work and family have to do with the traditional focus of the German welfare state on the male breadwinner model (cf. Lewis and Ostner, 1994; Sainsbury, 1999). Therefore, the aim of this chapter is to explain the roots of this policy orientation in German family law and welfare policies, its influence on the constitution of marriage and cohabitation, and the resulting consequences in case of a breakup. The chapter is structured as follows: after giving some general information on the country and its recent demographic and labour market trends, the subsequent section deals with the conception of the family in law and social security with particular regard to the special situation of separated or divorced persons. The following section starts with some hypotheses concerning the economic consequences of partnership dissolution based on the information given in the previous sections and finally summarizes 51

52

Single country studies

the main results from empirical research on the consequences of partnership dissolution in Germany. The chapter closes with a classification of the German welfare state and an evaluation of the consequences of separation and divorce in Germany.

RECENT DEMOGRAPHIC AND LABOUR MARKET TRENDS Changes in the Family Structure According to Figure 2.1, there has been a substantial increase in the number of divorces since the mid-1960s (barring the decrease in 1977/78 due to the reform of divorce law (see the next section)). While in the 1970s about 15 of 100 marriages in West Germany failed, in the 1990s twice as many couples got divorced. Today, more than every third marriage ends in divorce. Different factors have contributed to this increase in marital instability, for example, a change of faith, gainful employment of both spouses, higher education of the wife, childlessness, divorced parents, and so on (cf. Klein and Kopp, 1999; Wagner and Weiß, 2003). The increasing divorce rate comes along with a devaluation of the traditional marriage- and family-model and an increasing cultural acceptance of divorce as a legitimate form of conflict resolution (cf. Diekmann 1994). The more marriages end in divorce, the more divorce is socially accepted. In this way, different divorce risks can reinforce each other and the divorce rate increases further. In East Germany, the divorce risk rose earlier than in West Germany. In this regard, it is important to consider that a divorce in the former German Democratic Republic (GDR) had only marginal economic and social consequences due to the legal circumstances, the predominant full employment of men and women, and the minor importance of the Christian churches (cf. Wagner, 1997). According to this, divorce was linked with minor social and economic barriers in contrast to West Germany. But in the beginning of the 1990s – immediately after reunification in 1990 – the East German divorce rate decreased noticeably. This is often traced back to the adoption of the West German divorce law that extended the divorce procedure in East Germany (Böttcher, 2006: 7) but also to the great uncertainty concerning the social and economic development in the course of reunification (Engstler, 1997: 89). In recent years, even though the divorce rate in East Germany has come closer to that of West Germany, it is still clearly lower. Not only the stability of marriages but also the disposition to get married

Germany: will the male breadwinner model survive?

53

140 West East

Divorces per 10000 marriages

120

100

80

60

40

20

0 1960 1965

Source:

1970

1975

1980 1985 Year

1990

1995

2000

Emmerling (2005).

Figure 2.1

Divorce rate for West and East Germany (1960–2003)

has changed in the last decades. Since the birth cohorts of the early 1950s, more young women remain unmarried for a longer time. Between 1994 and 2003, the average age of the first marriage increased for men from 29.4 to 32 years and for women from 27.1 to 29 years (Emmerling, 2005: 101). The increase of marriage age is associated with the educational expansion because the extended duration of training postpones the timing of the first marriage. Furthermore, because of improved employment possibilities for women, marriage becomes less attractive for them – a reason why, simultaneously, the rate of people who remain forever unmarried rises. In 2000, 18 per cent of all men and 11 per cent of all women aged between 40 and 44 were still unmarried. The proportion of unmarried individuals in this age has risen by 6 percentage points for men and 7 percentage points for women since 1991 (Engstler and Menning, 2004: 18). This development is also reflected in the declining crude marriage rate in the past 20 years: while in 1985 still 6.39 of 1000 inhabitants got married, in 2004 there were only 4.8 (see Data Appendix). As a result of the declining disposition to marry, more and more couples are living together without being married. Between 1994 and 2003, the

54

Single country studies

percentage of unmarried couples in the population rose from 5.1 per cent to about 6.9 per cent (cf. Statistisches Bundesamt Deutschland, 2006). This living arrangement is especially widespread among young, mainly childless, couples. Today, the vast majority of couples have a phase of cohabitation. But there is an important difference between East and West Germany. East German couples are less likely to start a family after getting married. They have longer periods of cohabitation which are not necessarily ended by the birth of one or more children. West German couples, on the other hand, are more likely to marry latest, after the birth of a child. While in 2000 in West Germany 23 per cent of all non-married couples lived with children, in East Germany the rate was twice as high. So, non-married partnerships in West Germany are basically living arrangements for couples without children, while in East Germany they are more frequently with common children as well (cf. Gysi and Meyer, 1993; Konietzka and Kreyenfeld, 2005). But non-married partnerships are not only a typical living arrangement before or in place of marriage; they also become a more widespread living arrangement after the breakup of a marriage. Almost one-quarter (24 per cent) of non-married partnerships without children included at least one divorced partner. The same applied to almost one-half (47 per cent) of non-married partnerships with children (Engstler and Menning, 2004: 10). In view of the increase of partnership dissolution, it is necessary to take into account that single parents and their children constitute 6 per cent of the population (see Data Appendix). About 87 per cent of all single parents are women and only 13 per cent are men (Engstler and Menning, 2004). Therefore, living alone with children is predominantly a female pattern of life. Most single parents are separated or divorced (63 per cent). Only a small number are widowed. In the West, almost two-thirds (65 per cent) of all single parents are divorced, while in the East – where single motherhood is widespread – only 56 per cent are divorced. 36 per cent of single mothers in Eastern Germany are unmarried, as opposed to 22 per cent in West Germany (Engstler and Menning, 2004: 9). The explanation for this is not only the greater reluctance to marry in the Eastern states, but also a less pronounced normative effect of marriage as an institution. The comparative reluctance to marry in spite of parenthood in East Germany is also favoured by an easier access to childcare institutions and the more widespread double-earner model (cf. Chapter 3). Aside from the declining stability of the family, family size has been shrinking in the past decades. Household and living arrangements without children have become more frequent in the past years. The rate of childless people living alone or with a partner has continued to increase. The birth rate has declined since the beginning of the 1970s, with Germany currently having one of the lowest fertility rates in the European Union (EU).

Germany: will the male breadwinner model survive?

55

During the past 25 years, the total fertility rate ranges between 1.3 and 1.4 children per woman (see Data Appendix). The rate of women that remain without their own children is still on the rise. The increase of childlessness in East Germany in comparison to West Germany is a younger phenomenon, maybe due to the better reconcilement of familial and professional life in the former GDR. However, after reunification the East German birth rate decreased below the level of West Germany (Engstler and Menning, 2004: 20). In summary, there are three different trends in the German family structure. First, the percentage of unmarried couples with children and the rate of single parents has increased, while the percentage of married couples with children has decreased. Second, the rate of couples living without children has risen. Third, the importance of living alone has increased due to the rising percentage of older widowed people who often live alone. Despite the aforementioned increase of alternative living arrangements, living together as a married couple is still the predominant living arrangement (60 per cent of the population). Cohabiting couples are still a minority (7 per cent), while more than 20 per cent of the population lives in single households (cf. Statistisches Bundesamt Deutschland, 2006). Labour Market Trends For about 30 years, the German labour market has been characterized by an imbalance of supply and demand. The gap between persons seeking work and the number of job vacancies began to grow in the course of time. Unemployment is now one of the greatest social problems in Germany as well as in other European countries. Particularly in East Germany, previously unknown labour market problems appeared after reunification and the transition from a planned to a market economy. Unemployment rates are subject to cyclical variations, but overall, they have continued to increase in East and West Germany. In 2006, about 4.5 million people were affected by unemployment. However, not all population groups are affected by unemployment to the same degree. The unemployment rate for men is higher than for women (see Data Appendix), foreigners are more affected than Germans and persons older than 50 years belong to the so called ‘problem groups’ of the labour market. Last but not least, an unequal distribution between the regions is observable: the North is more affected by unemployment than the South of Germany, and in East Germany unemployment rates are more than twice as high as in West Germany. In comparison to other EU countries, the unemployment rate for Germany in 2002 was above average (Statistisches Bundesamt Deutschland, 2004a: 114).

56

Single country studies

Owing to the dominating male breadwinner model and the associated gender-specific division of labour (cf. Lewis and Ostner, 1994; PfauEffinger, 2004; Sainsbury, 1999), the labour market participation of men is much higher than that of women. Referring to 15–65-year-old men, the employment rate in 2004 added up to 79.3 per cent, and to 65.2 per cent for women of the same age (cf. Cornelißen et al., 2005). Especially in West Germany, the discrepancy in labour participation between men and women is greater than in East Germany. The proportion of employed women, and in particular the labour force participation of married women, is about 10 percentage points higher in the East than in the West due to the longer tradition in women’s employment in the former GDR and the better childcare infrastructure. Since 1991, the overall employment rate of women slowly but continuously increased. Thus, the discrepancy between the sexes becomes also smaller in West Germany, where the employment rate of mothers with children in school-age has risen (cf. Cornelißen et al., 2005). Starting a family influences particularly the employment behaviour of women. In view of the predominant male breadwinner model and the insufficient public childcare facilities, at least in West Germany it is not surprising that female labour market participation is heavily influenced by the fact of having children and the age of the children. Women living with dependent children confine their professional career noticeably (cf. Klammer and Klenner, 2003). During the time before reunification, nine out of ten East German mothers went out to work. At that time, the majority of women returned to work one year after the birth of a child at the latest (cf. Klenner, 2002). Even though the employment rate for East German women with children in the household has declined since 1991, it is still at 72 per cent about 10 percentage points higher than in West Germany (62 per cent) (cf. Klenner, 2002). Since the introduction of parental leave and child-raising allowance in West Germany in 1986 and in East Germany since the reunification,1 the disposition to take parental leave is very high in both regions. The duration of parental leave has gradually increased to three years since 1992. Empirical research shows that about 85 per cent of the parents in Germany – and nearly exclusively women – make use of parental leave after the birth of the first child (Vaskovics, 2000: 234). For the remaining 15 per cent, the women are not entitled to benefit because they were not employed before the birth of the child, and due to financial reasons the man does not make use of the parental leave. At the birth of a second child, the eligibility for parental leave of the woman often does not exist anymore because she does not return to employment after the expiration of the parental leave

Germany: will the male breadwinner model survive?

57

for the first child. The proportion of fathers that apply for parental leave indeed tripled since initiation, but it is still very low, about 2.5 per cent (Vaskovics, 2000: 236). Financial reasons, normative beliefs and expectations on gender roles play in this context an important role. In West Germany the proportion of working mothers rose steadily, particularly in the case of women whose children had reached school age. However, this increase is largely due to a general rise in part-time work. Part-time work has become increasingly more relevant for women as well as for men. While in 1991, 5.4 million employees worked part-time, in 2001 there were 9.4 million part-time employees. Part-time work is primarily a female domain. Almost one-half of all employed women and only less than 10 per cent of the employed men worked part-time in 2001 (cf. Engstler and Menning, 2004). This increasing female labour market participation can be traced back to the expansion of part-time work. Due to the predominant male breadwinner model in West Germany, women and especially mothers have to combine house and family work with their occupational career. In addition, in Germany, women earn about 25 per cent less than men. Even women employed full-time earn about 12 per cent less than their male colleagues with similar human resources in the same occupation and firm (cf. Hinz and Gartner, 2005). This is also a factor that supports the male breadwinner model and encourages women to (at least temporarily) interrupt their occupational career when starting a family. The families are not able to pass on the significantly higher male wage.

CONCEPTION OF THE FAMILY IN LAW AND SOCIAL SECURITY Family Law Since 1958, the legal marital property regime in Germany is the community of goods (§1363 BGB2). In case of divorce, the gain in property accumulated during the marriage is divided between the former partners. The underlying idea is the assumption that everything acquired jointly or based on division of labour during the marriage has to be fairly split between the partners. Also pension rights acquired during marriage have to be divided in the case of divorce. The purpose of the splitting of pension rights is to give women, especially housewives, a share in their husband’s retirement pension (cf. Münder, 2005; Scheiwe, 1999). In 1977, the divorce law was reformed and fault-based divorce was abolished and replaced by the principle of matrimonial breakdown (cf.

58

Single country studies

Münder, 2005; Scheiwe, 1999). Thus, marital breakdown is now the only reason to get a divorce. The former partners have to live apart for one year. Then the divorce may be pronounced on the mutual consent of both partners. If only one spouse demands the divorce and the other spouse disagrees, the former partners have to be separated for at least three years before the divorce may be granted (§1566 BGB). Separation in this context means separation from bed and board. So, for a divorce to take place, the question of fault is no longer relevant. During the time of separation the same standard of living as in the former marriage has to be guaranteed for the economically dependent members of the family. Therefore, a claim for support payments exists from the first day of separation. In general, divorced spouses are expected to be self-sufficient. But this principle is suspended for numerous exceptions. A partner has a right to get maintenance payments in case of indigence, that is, if he or she is not able to earn his or her keep (§1569 BGB). When and to what extent gainful employment of a divorced mother or father can be expected, is differently construed by the courts. Number and age of the children play a major role. Owing to insufficient childcare, it is usually expected that women with children under 6 years of age do not have to work (§1570 BGB) (cf. Münder, 2005; Scheiwe, 1999). However, if current own income does not allow a standard of living comparable to the former standard during marriage, a special maintenance claim exists (Aufstockungsunterhalt) (cf. Münder, 2005). There is much discussion in the public about the significance of this maintenance claim, but the empirical knowledge about its application is scarce. Therefore, its practical relevance for the ‘average divorce’ is hard to determine. Given the expectation in maintenance law, which is shared by the public, that adult persons primarily care for their own living, it can be assumed that spousal support, although applicable in many situations, is a rather residual claim, which is demanded only under certain circumstances and after child support payments have been made. Whether it is used to offset the marriage-specific changes in human capital remains an open question. For cohabiting couples there are no legal regulations in Germany. But both partners are free to make an official legal partnership agreement that, for instance, defines maintenance regulations during partnership and in case of a breakup. Generally there are no maintenance obligations for unmarried partners, except in the case of maintenance claims due to a common child. When the parents are married, as a rule both parents maintain joint custody following divorce if there are no cogent reasons that argue against it. But another issue is the factual execution of joint parenting that presumes mutual agreement between the former partners. In cases

Germany: will the male breadwinner model survive?

59

of disagreement the court has to decide. Normally the parent the child lives with (in most cases the mother) is free to arrange autonomously the child’s affairs concerning everyday life, while matters with substantial relevance (like decisions concerning education, financial assets of the child, change of residence and the like) have to be decided by both parents. When the parents are unmarried, generally the mother receives the right of custody. The father may also get custody rights, when both parents make together a declaration at the youth welfare office (Jugendamt) (cf. Münder, 2005). In terms of child support payments, separated and divorced parents have to support children who are under 18 years of age. This obligation continues even longer if the children have not yet completed their education or training. From the very beginning of a separation, the spouse who does not live with the child – normally the father – has to comply with the maintenance obligations. Child support takes priority over other support claims, for example, spousal maintenance. For unmarried children under 18 years of age an increased maintenance obligation exists. This means that parents are bound to share everything with their child even if they have little. Furthermore, the person who is liable for maintenance is obliged to maintain his or her ability to pay. Thus, he or she may not resign from a job without a plausible and important reason. In addition, in the case of unemployment, the liable person has to make every effort to find work (cf. Münder, 2005). The amount of child maintenance depends on the income of the liable person, the age of the child entitled to it, and the overall number of persons entitled to maintenance, for example, further children. A minimum amount of maintenance is required by law. As a guide to assess the amount of support payments, there is the so-called ‘Düsseldorfer Tabelle’ (2005) – a means to appraise the ‘adequate child maintenance’ based on the income of the liable person and the age of the child. For example, for a liable father with a 3-year-old child and a monthly net income of about €2400 the maintenance payment amounts to €290 and increases with the increasing age of the child. The income of the person who is responsible for the upbringing of the child is not important for the calculation of the maintenance payments for the liable person (cf. Münder, 2005; Scheiwe, 1999). However, depending on the total number of entitled persons and the available income, neither the child’s economic need nor the actual child costs are guaranteed (cf. Scheiwe 1999). Besides that, the liable parent is allowed to retain an amount for his or her basic needs while the resident parent’s basic needs are not considered. Thus, already the legal norms suggest that in some cases child costs are not met and disproportionately shared by both parents (cf. Andreß and Bröckel, 2007).

60

Single country studies

Family Policy Support for married couples From a historical point of view it was not necessary to distinguish between policies for married couples and for families because in the majority of cases policy for both was combined: marriage was basically supported by the state because it was seen as the basis for the birth and upbringing of children. Therefore, formal marriage is still very important for the acquisition of social benefits and the privileges for married couples are independent from the existence of children. In particular, the male breadwinner model is supported by granting tax privileges for married partners who divide labour – the so-called Ehegattensplitting. Married couples may choose between joint or individual taxation. In the case of joint taxation, which has become the standard form, husband and wife each pay income tax on half the total of their combined incomes. In this way, the tax system principally benefits one-earner couples and two-earner couples where the differences between partners’ earnings are high. Thereby, the traditional gender-specific division of labour with one principal earner – in most cases the husband – is facilitated. In addition to fiscal relief, spouses are co-insured free of charge in compulsory health insurance of their partner if their monthly income does not exceed €3503 (Bundesministerium für Familie, Senioren, Frauen und Jugend, 2006). In view of the economic consequences of separation and divorce, it is important to keep in mind that in case of separation, married couples lose these privileges (Scheiwe, 1999: 246). Support for parents: family benefits Aside from the support for married couples, parents are also provided for by the state, independently of their marital status. While children are an enrichment for a couple, they cost a lot of money. Food, clothes and education all have to be paid for. Therefore, there is special financial compensation in the form of cash transfers and tax relief aimed at compensating for familial burdens (Familienlastenausgleich). One component of this is child allowance (Kindergeld). Germany has a child allowance system with universal and standardized benefits for all children up to the age of 18. Child allowance is not income related and the amount increases according to the number of children. Under certain conditions, like education or training of the child, the allowance is prolonged up to the age of 27 (given the child’s income does not exceed the maximum of €7680). For unemployed children, child allowance can be claimed up to the age of 21. There is no age limit for handicapped children who are unable to earn their living. Although the amounts of child

Germany: will the male breadwinner model survive?

61

allowance have increased several times in the past years, in comparison to other European countries they are generous but not the highest. For the first, second and third child it currently amounts to €154 per month and for the fourth and subsequent children to €179 per month. Only one person can receive child allowance for each child. Parents can choose who claims child allowance for the children living in their household. If the parents are separated or divorced, child allowance is paid to the parent the child lives with (Bundesministerium für Familie, Senioren, Frauen und Jugend, 2006). Tax-free allowance for children and tax-free allowance for childcare, child-raising and vocational training (Kinderfreibetrag) are, besides child allowance, another important political measure to compensate for familial burdens. But in those cases where the amount of the child allowance does not exceed the burden-easing effect of the tax allowance for children, the tax allowance schemes are taken into consideration in the following year, as income tax is assessed by the local tax office. This concerns mostly families with higher incomes. Thus, in most cases, the tax abatement is fully covered by the child allowance already paid out.4 As already mentioned, in order to get tax advantages, being married is far more important than having dependent children. Raising children is a great responsibility. Mothers especially, but also fathers, want to leave work for a while or wish at least to reduce their working hours to concentrate on childcare. To enable working parents to assume the care of their children by themselves, Germany has some measures (like maternity and parental leave and corresponding benefits) to release parents for a certain time from work and to compensate for the income loss during this time. First, expectant mothers gainfully employed have a period of maternity protection (Mutterschutz) of 14 weeks – normally six weeks before and eight weeks after the child’s birth – to protect them and their child from workplace hazards. During this period, the mother receives maternity benefits. Mothers who are insured with the statutory health insurance receive up to €13 per day. The difference between the €13 and the women’s average daily net income is paid by the employer, regardless of whether the woman is a member of the statutory health insurance scheme or not. Furthermore, working women are protected from dismissal from the beginning of their pregnancy to four months after the birth of their child. Subsequent to maternity leave, gainfully employed mothers – and fathers – are entitled to take parental leave (Elternzeit) up to the child’s third birthday. This is, in comparison to other European countries, a very long parenting break in the period of employment. Parents can take parental leave separately or simultaneously for the whole or for a limited

62

Single country studies

period of time. During the parental leave, a parent may work for up to 30 hours a week. Moreover, parents are entitled to do part-time work during the parental-leave period. Mothers (or fathers) who make use of the parental leave receive a childraising allowance (Erziehungsgeld; Elterngeld). For children (born before 1 January 2007) the parents have the choice between a standard amount of €300 per month for two years or a budget amount of €450 per month for one year. While in receipt of this benefit, the person in parental leave may work up to 30 hours a week. The child-raising allowance is income dependent. During the first six months, after the child’s birth, the caring person receives the full rate of €300, or in case of the budget amount €450, per month, if the net annual income of a married couple (with one child) does not exceed €30 000 or €23 000 for a single parent. Parents who exceed this income limit do not receive child-raising allowance. Once the child is seven months old, the child-raising allowance is gradually reduced for parents who have an annual income exceeding €16 500 (or €13 500 in case of single parents). The income limit rises by €3140 in each case with every additional child in the family. For births following the 1 January 2007 the flat-rate child-raising benefit was replaced by an income-related benefit.5 A very important measure for separated or divorced parents is the advanced child maintenance payment (Unterhaltsvorschuss). With this cash transfer, the state supports single mothers or fathers if the other parent fails to make his or her maintenance payments (cf. Münder, 2005: 129; Scheiwe, 1999: 133). But this amount is often far from enough to cover the need (the amounts for children under 6 years of age were, in 2006, €127 per month in West Germany and €111 in East Germany; for children from 6 to 12 years of age €170 per month in West Germany and €151 in East Germany). Furthermore, the advanced child maintenance payments are paid for a maximum period of 72 months and there exists a maximum age limit of 12 years for the child. If possible the state tries to recover the amount from the parent who has defaulted on the payments. The caring parent will be barred from claiming the advanced maintenance if he or she remarries. Overall, the provision of a broad parental leave as well as financial incentives for parents to take over the childcare by themselves, in the form of a child-raising allowance and the support of the male breadwinner model (with one principal earner), shows that childcare is, predominantly, accounted for as a private matter. In particular, mothers are put in charge for the care of their children. For children under 3 years of age especially the possibilities for using public childcare are very restricted. But because of different traditions in childcare and women’s employment, there is a considerable East–West divide in the available number of childcare

Germany: will the male breadwinner model survive?

63

Table 2.1 Childcare facilities in Germany (number of places per 100 children, 2002) Age of the children

West Germany East Germany Total Source:

0–3 years

3–6 years

> 6 years

3 37 9

88 105 90

5 41 9

Statistisches Bundesamt Deutschland (2004b).

facilities as well as in the employment rates and employment wishes of mothers (see Table 2.1). There are only few childcare facilities for children under 3 years of age available in West Germany. In a European comparison, the former Federal Republic of Germany (FRG) is one of the countries with a relatively low level of childcare facilities for this age group. In 2002, only 3 per cent of those aged under 3 visited such a childcare facility in West Germany. The children of working single parents had an above average attendance. In contrast, East Germany had a high number of day-care facilities in comparison with other European countries. In 2002, 37 per cent of all children in East Germany under 3 years used public day care (Statistisches Bundesamt Deutschland, 2004b). Since 1996, there is a legal claim for 3-year-olds to attend nursery school (Kindergarten). For this reason, there are sufficient kindergarten places for nearly 90 per cent of all children between 3 and 6 years of age. But there is also an East–West divide: in East Germany there are more places than children, while in West Germany only 88 per cent of the children can visit a kindergarten. In addition, most of the East German kindergarten places are full-time, while in West Germany most of the kindergartens are only open for half a day. The situation for schoolchildren is similar. Compulsory education starts at the age of 6 and schools in Germany are in most cases half a day. But only 9 per cent of all schoolchildren between 6 and 11 years have facilities for after-school care available (for 5 per cent in the West and for 41 per cent of the schoolchildren in the East) (cf. Statistisches Bundesamt Deutschland, 2004b). In summary, the state does not provide an institutional framework which makes it easy for both parents to work simultaneously. For West German parents in particular it is still difficult to reconcile work and family. If both parents want to work, they have to decide whether one

64

Single country studies

parent works part-time or they use informal childcare, if the latter is available. Social assistance Aside from a set of different social benefits for families, couples or individuals, Germany has a basic financial security system. Social assistance is also important for separated or divorced people without children, in case of unemployment, disease or disabilities, because it provides a safety net to protect people from poverty, social exclusion and hardship (cf. Bundesministerium für Arbeit und Soziales, 2007). Regarding separation and divorce, social assistance is particularly relevant for many single parents because assistance for living expenses makes up for insufficient maintenance payments. The aim of the social benefit system in Germany is to ensure the necessary standard of living for people and households who are unable to meet their own needs and who do not get help from other people or institutions. Where income and savings fall short, social assistance covers the minimum needed to maintain a socially acceptable living standard. Social assistance consists of a standard rate, accommodation rate, heating costs, supplementary assistance for certain groups (for example, single parents) and health insurance. Furthermore, a nonrecurring assistance is provided, for example, for setting up a household. In 2005 social assistance has been integrated with unemployment assistance into one system of basic income security (Arbeitslosengeld II).

ECONOMIC CONSEQUENCES OF PARTNERSHIP DISSOLUTION The described institutional framework suggests that the German system of family law and social security provides measures to safeguard the living of spouses and children in case of separation and divorce, for example, claims to maintenance, advanced child maintenance and social assistance. But how effective are these measures in practice? Are the existing legal regulations indeed able to protect the individual from negative financial consequences in case of separation and divorce? Theoretical Assumptions An analysis of the economic consequences of partnership dissolution usually starts with reflecting on the financial advantages linked with marriage and partnership that are lost in case of separation. According to Gary S. Becker (1981), joint housekeeping – which is in most cases associated

Germany: will the male breadwinner model survive?

65

with partnership – reduces the living expenses for the individual. First, gains arise from using household goods conjointly, that is, housing space, electricity, domestic appliances, cars, and so on. Second, advantages result from a material safeguard in risk situations like unemployment or illness. Two incomes, or merely the possibility of compensating for a loss in income of one partner, contributes to a reduction of poverty risks. Third, economic gains emerge from the division of labour, that is, the specialization of both partners on gainful employment and housework respectively. These financial advantages are lost – independently from marital status – after separation. Additionally, married persons lose the specific benefits associated with marriage, for example tax relief for married couples or joint health insurance (see previous section). Since most separated and divorced persons live – at least at the beginning – without a new partner,6 the risk for them to degrade financially is high: separate homes are necessary, someone is needed to care for the children or to earn money for living, life becomes more expensive because economies of scale are lost, and so on. Hence, it is assumed that both involved spouses suffer financial losses. But, as discussed in preceding sections, the German socio-political system is designed to strongly support the male breadwinner family and thereby the gender division of labour within the family. While this unequal division of care tasks and market activities remains efficient as long as the partnership continues, it also produces an unequal distribution of marital gains and losses between both partners in case of a breakup. As already mentioned in the introduction to this book, four main risk factors can be identified (cf. Holden and Smock, 1991; Sørensen, 1994): Risk 1: The economic needs are substantially higher for the parent who lives with the children – in most cases the mother. At the same time the caring person is often not able to participate in the labour market because public childcare places are scarce. Risk 2: The risk of the resident parent not to receive the child maintenance from the non-resident parent is high. Risk 3: Income sharing during marriage allows women to participate in men’s market income, which is on average higher than women’s market income. Presumably, the amount of spousal maintenance from the former partner after separation does not compare with the economic support women received during marriage. Risk 4: Changes in material wealth, pension entitlements and human capital that both partners experience during marriage are only imperfectly measurable and therefore hard to equalize after separation. Gains and losses in human capital especially are hard to measure and therefore hard to counterbalance.

66

Single country studies

Based on these considerations one can assume that the financial risks in consequence of partnership dissolution are not equally distributed between men and women. Women are likely to suffer higher losses than their former partners. Furthermore, the above mentioned risks demonstrate the main problem of the German welfare state from the point of view of separated women: care work is essentially secured through private transfers and mediated entitlements between the spouses (for instance, maintenance claims) rather than individualized entitlements to social security. The outcome is that women are financially dependent on others – during marriage as well as after marriage. Therefore one can expect that women are, on the one hand, more dependent on maintenance payments than men and, on the other hand, more dependent on social assistance benefits than men. Empirical Evidence Economic and social consequences of partnership dissolution in Germany have been researched with cross-sectional data on single-parents (cf. Sørensen, 1994), retrospective surveys of divorced and married individuals (cf. Andreß et al., 2003) and panel data of the overall population (cf. Andreß et al., 2003 and Chapter 11 in this volume; Burkhauser et al., 1990, 1991; Dewilde, 2002; DiPrete and McManus, 2000; Kampmann et al., 1996; Schwarze and Härpfer, 2000; Uunk, 2004). Even though these studies differ with respect to data and methodology, the overall picture of the empirical findings is consistent: economic gains and losses due to partnership dissolution are unequally distributed between men and women. By comparing the economic situation of the involved spouses before and after separation7 it becomes obvious that the event of separation is linked with economic changes for both partners (cf. Table 2.2). The change of equivalized household income8 two years before (t 2 2) and one year after separation (t 1 1) for separated German men and women shows that partnership dissolution is in most cases linked with income losses, however with great gender differences. While half of all men suffer moderate income losses (median: 24 per cent) and most of the other half actually benefits from separation (third quartile: 127 per cent), virtually all separated women have to bear income losses which are also more severe (median: 227 per cent). Besides the overall income changes, Table 2.2 differentiates with respect to re-partnering, employment, parenting and number of income earners in the former partnership. Re-partnering following separation seems predominantly to affect the female post-divorce income, while men’s economic situation rather keeps untouched by a new partner (cf. Table 2.2). Women living in a new partnership dispose of higher incomes, and experience – if at all – lower financial

67

t11 Euro

Euro 15 334

18 007 14 974 17 465 10 353 14 941 15 375

15 949

16 481 15 949

18 267 9 890

16 655 15 949

t22

a

a

–20 –26

–24 –33

–26 –26

226

Q1

b

–10 0

–5 14

–4 –4

–4

Me

% change

Men

117 128

127 125

129 127

127

Q3

75 191

217 49

82 184

266

N

13 051 21 387

14 321 12 945

16 487 14 321

14 582

Euro

t–2

a

9 232 13 459

11 314 8 146

13 977 9 706

10 184

Euro

t11

a

–42 –43

–41 –52

–49 –41

–42

Q1

–26 –29

–28 –27

–1 –29

–27

Me

% changeb,c

Women

–4 18

15 –10

133 –10

–1

Q3

207 93

208 92

82 218

300

N

Changes in equivalized household income for separated men and women by household type and employment status (median and quartiles)

Equivalized household income New Partner t11 Yes No Employed t11 Yes No Children t11 Yes No

Table 2.2

68

18 007 15 616 12 716

18 356

19 920 12 647

–27 –39

–30

–8

Q1

–17 16

–1

118

Me

% changeb

Men

–7 131

127

150

Q3

55 33

103

59

N

20 155 13 748

17 346

10 624

Euro

t – 2a

13 459 12 903

9 706

10 315

Euro

t 1 1a

–42 –22

–53

–39

Q1

–32 0

–35

–27

Me

% changeb,c

Women

–22 134

–15

–5

Q3

59 40

105

70

N

Source: Andreß et al. (2003: 53, 56); DM converted to euros.

a

Notes: Median equivalized household income. b Distribution of change scores obtained by comparing equivalized household income between t – 2 and t 1 1 for each individual (Q1: 1st quartile, Me: median, Q3: 3rd quartile).

13 893

Euro

Euro

12 744

t 1 1a

t 2 2a

(continued)

Household type t–2 Husband single earner Husband main earner Dual earner Wife main/single earner

Table 2.2

Germany: will the male breadwinner model survive?

69

losses after separation (median: 21 per cent) than women without a new relationship (median: 229 per cent). Obviously a new partner is able to enhance the economic situation of women presumably by offering an income source that compensates for the income losses and substitutes for the income of the former partner. In contrast, living with dependent children has a strong negative impact. In most cases the parent who lives with the children after separation disposes of a significantly lower income and undergoes higher income losses (cf. Table 2.2). But this again applies particularly to women: while women with children lose on average 28 per cent of their pre-divorce income, men living with dependent children lose only 10 per cent. Taking up employment or expanding working hours is often seen as a way to improve the economic well-being after partnership dissolution – especially for women. According to Table 2.2, employed women indeed dispose of a higher household income than unemployed women but gainful employment seems not to prevent economic losses: both employed and unemployed women lose about 27 per cent. One important reason might be great differences in working hours and wages between men and women (cf. previously in this chapter). Women’s wage is on average considerably lower, and for this reason they are not able to reach the income level of their former partner. Therefore the loss of the male income can only be partially compensated by own employment. When looking at the impact of the employment behaviour during partnership (cf. Table 2.2, household type t 2 2) it appears that the female losses in almost all household types average out at 30 per cent – regardless of whether they were full-time, part-time or not employed when with an employed husband. Just in the few households with a female main or single earner the gains and losses for women tend to balance each other out. In contrast the income losses of men after separation increase, the more the woman had contributed to the household income during marriage (cf. Table 2.2) (again with the exception of the few female main or single earner households, which are probably a very selective group). The observation that also the financial losses of men are linked with the female employment status during marriage, probably indicates that the financial contribution of women to household income should not be underestimated. Evidently separation and divorce are linked with negative income changes. But another question is, whether these income losses result in income poverty. According to a cross-country study with longitudinal data from Great Britain and Germany (cf. Ruspini, 1998), divorce is one of the greatest poverty risks in Germany. Andreß et al. (2003: 75) show that poverty rates significantly increase after separation, in particular for women. Also cross-sectional data show that separated and divorced

Single country studies

70

Table 2.3

Payment of child and spousal maintenance (only respondents with minor children)

Spousal maintenance during separation (all respondents) N Thereof: entitled (%) Thereof: regular and complete receipt of spousal maintenance (%) insufficienta receipt of spousal maintenance (%) without receipt of spousal maintenance (%) Child maintenance during separation (only respondents with minor children) N Thereof: entitled (%) Thereof: regular and complete receipt of child maintenance (%) insufficienta receipt of child maintenance (%) Without receipt of child maintenance (%) Note: Source:

a

Women

Men

768 76 28

741 13 9

9 65

4 87

436 95 54

387 23 14

20 26

2 84

Irregular payments and/or amount does not meet the maintenance agreement. Andreß et al. (2003: 167, 189).

persons are twice as often hit by poverty than are married persons (cf. Statistisches Bundesamt Deutschland, 2004a: 632). Against the background of these results, the question arises as to what extent the German system of family law and security is able to safeguard the economically weaker family members – women and children – in case of a breakup. When looking at the experiences of separated persons entitled to spousal maintenance (cf. Table 2.3),9 it appears that the majority – mostly women – either do not receive any kind of support or receive irregular and inadequate payments. Only 28 per cent of all entitled women and 9 per cent of the few entitled men can count on regular and complete payments. The situation is better for child support, but still one-half of all women report insufficient or no support payments for their children (cf. Table 2.3): 26 per cent receive no financial child support and 20 per cent insufficient payments. For the few men entitled to child maintenance the private support turns out worse: more than two-thirds do not receive any child maintenance. One possible reason for the comparatively poor paying habits of the few women liable for support is that they are often part-time employed and achieve significantly lower wages than men.

Germany: will the male breadwinner model survive?

71

All facts considered, there are a lot of problems resulting from inadequate legal procedures, insufficient amounts of maintenance and incomplete enforcement procedures that are increased by poor paying habits and insufficient solvency due to low incomes or unemployment (cf. Martiny, 1987). Results from panel data suggest that public transfers often have to fill the gap and are of higher importance within the income package of women than private support payments (cf. Table 2.4) (Andreß and Bröckel, 2007; Andreß et al., 2003). While private transfers constitute about 11 per cent of the household income of separated women, public transfers amount to 26 per cent, for women with children even to 33 per cent of the household income (cf. Table 2.3). While the male income package changes only marginally after separation, the female income package alters significantly: the percentage of private and public support increases while the percentage of labour income decreases. Because of the loss of the male labour income, women become more dependent on private as well as on public transfers following separation. Besides private and public transfers and own gainful employment, financial and social assistance can be rendered by friends and the extended family network. Even if today some of the family’s previous social and economic functions are shifted to other institutions such as companies, schools and other public agencies (Coleman, 1994: 585), the relationships within the kinship are still very important. Especially in private crises like separation and divorce when other institutions react slowly or do not intervene, the family of origin and particularly an individual’s own parents again become more important (cf. Herlyn et al., 1998; Wagner, 2002). In many cases, they grant relief and can offer, to a greater or lesser extent, financial support. In fact, separated or divorced people more often receive financial help than do married people. Eighteen per cent of the separated people living without a new partner are financially supported by their parents – twice as many as married people (9 per cent).10 Even if the financial help provided by parents, other relatives or friends is of little importance for a general improvement of the economic situation after separation, their social help should not be underestimated. In view of the lack of public childcare facilities in Germany, grandparents are the most important informal caregivers (Esch and StöbeBlossey, 2005: 8). Provided that they live nearby, they can enable the mother to participate in the labour market by taking care of the grandchildren. Overall, empirical research on the economic consequences of partnership dissolution in Germany shows that women have to bear the main part of income losses. This is especially true if there are dependent children, which boost the economic needs and at the same time restrict the employment chances of the caring person. According to our data, the existing public and private transfers are not able to eliminate this gender gap.

72 (N 5 369) 2 12 79 11

t–2 (N 5 367) 11 26 56 24

t11

(N 5 63) 0 19 73 13

t11

(N 5 276) 2 14 78 12

t–2

(N 5 257) 12 33 49 31

t–2

Women with children

(N 5 216) 0 13 82 6

t–2

Men with children

(N 5 263) 1 11 77 11

t11

(N 5 93) 0 8 82 9

t11

(N 5 110) 9 10 70 10

t–2

Women without children

(N 5 110) 0 8 74 5

t–2

Men without children

Source:

Andreß and Bröckel (2007).

Notes: a Mean percentage of disposable household income (after deducting taxes and support payments, but including private and public transfers). b Gross earnings as percentage of pre-government income (before taxation and not including private and public transfers), mean value.

Structure of household income private transfersa (% of income) public transfersa (% of income) labour incomeb (% of income) public transfers more than 50% of household income (%)

(N 5 326) 1 12 76 11

t11

All women

(N 5 326) 0 11 79 5

t–2

All men

Labour income, private and public transfers before and after separation (in percentage of total household income)

Structure of household income private transfersa (% of income) public transfersa (% of income) labour incomeb (% of income) public transfers more than 50% of household income (%)

Table 2.4

Germany: will the male breadwinner model survive?

73

CONCLUSION Germany represents the conservative, corporatist model of welfare states (cf. Esping-Andersen, 1990). This conservative model encourages gender differences by supporting the male breadwinner model. Even though there are many services in support of families, the main orientation of policy is to ease rather than substitute the mother’s traditional role as the main household and early child education manager (cf. Daly, 2000; Sainsbury, 1999). Aside from joint taxation of the spouses, poor local childcare provision as well as financial incentives for parents to take over childcare by themselves are reasons for the low labour force participation of women. At the same time, there is not only a close relationship of social services to people’s position in the labour market, but social benefits are often earnings related. By this means, the German social security arrangements reproduce income inequalities and determine great gender-specific economic and social inequalities (Daly, 2000: 79). Therefore, great differences in the social and economic situation of men and women after separation and divorce arise. While gender-specific division of labour is gainful and advantageous for stable partnerships, it produces diastrous consequences in case of separation and divorce for the non-earning and economic dependent women. Empirical research shows that the greatest financial risk for women in partnerships in general, but married women in particular, occurs when having to live without access to a male income (cf. Andreß et al., 2003 and Chapter 11 in this volume; Daly, 2000). Maintenance payments are often of prime importance for divorced wives and especially for mothers, but – as previously mentioned – are often insufficient to cover need. However, maintenance is only one kind of financial safeguard after the breakup of a partnership. For women without dependent children employment is one option to assure subsistence. Mothers, on the other hand, are often not able to improve their financial situation after a matrimonial breakup by being employed. First, the possibility of finding adequate, alternative childcare is restricted – in particular for children under 3 years of age (cf. Table 2.1). In this context, informal childcare through relatives and friends can be an important factor enabling mothers to participate in the labour market. Second, long career breaks for mothers due to the broad parental leave system entail a depreciation of human capital and thereby lower the chances of finding adequate work (cf. Mincer and Ofek, 1982). The consequences of these extensive interruptions in a woman’s working life are severe and long-lasting for her occupational career and her achievable wages (cf. Ziefle, 2004). It can be concluded that female dependency on external financial resources (from the partner, social

Single country studies

74

services or family networks) in case of separation is a continuation of the economic dependencies culturally and governmentally generated within partnership and marriage. In consideration of the increase in divorces and women’s employment, the question arises of whether the German male breadwinner model is able to survive in the future. Some recent family-political reforms suggest that the German welfare state is actually changing. Recent measures like the efforts to improve the day-care supply for children under 3 years or the new child-raising allowance from 2007 are aimed at distributing gainful employment and care work more equally between the genders: mothers should be encouraged to participate in the labour market and men should be encouraged to take part in childcare. A more equal distribution of labour and family work – of paid and unpaid work respectively – between men and women should also contribute to more equally distributed and less severe financial consequences in case of partnership dissolution.

NOTES 1. 2. 3. 4. 5.

6.

7. 8. 9. 10.

In the former GDR a parental leave of one year (the so-called ‘baby year’) was provided. BGB stands for ‘Bürgerliches Gesetzbuch’ (German Civil Code). For persons with marginal jobs (with special social insurance conditions) the income limit amounts to €400. The tax-free allowance in 2006 amounts to €3648 for married couples and €1824 for single parents, and the tax-free allowance for childcare, child-raising and vocational training amounts to €2160 for couples and €1080 for single parents. For a period of 12 months, one parent that interrupts his or her career in order to raise children will receive 67 per cent of the last net income per month, but at most €1800. In the case of unemployment, the caring parent receives at least €300 per month. For low-income earners, families with many children and families with twins or triplets the amount is increased. If the other parent also makes use of the parental leave the incomerelated amount can be received two month longer, so altogether up to 14 months. The new parental benefit favours upper-income earners, who will receive substantially higher benefits, and double earners who exceeded the income limits of the previous child-raising allowance system (before 2007) will now receive support. Furthermore this measure should encourage fathers – in most cases the parent with the potentially higher income – to participate in childcare. It is necessary to consider that many of divorced people sooner or later have a new partnership and more than half of them marry again (Bundesministerium für Familie, Senioren, Frauen und Jugend and Cornelißen, 2005). It is assumed that the earlier the new partnership starts, the sooner people benefit again from the gains of joint housekeeping and can compensate for possible losses due to separation. Data refer to persons with the event of separation between 1984 and 1999 observable in the German Socio-Economic Panel Study (GSOEP). Household incomes have been adjusted with respect to household size and inflation. Data based on a representative survey from 2001 of about 1500 men and women who got divorced following 1995. Own calculations with data of the Family Survey 2000 of the German Youth Institute (DJI).

Germany: will the male breadwinner model survive?

75

REFERENCES Andreß, H.-J. and M. Bröckel (2007), ‘Marital disruption in Germany: does the conservative welfare state care? Changes in material well-being and the effects of private and public transfers’, Schmollers Jahrbuch: Journal of Applied Social Science Studies, 127 (2), 193–226. Andreß, H.-J., B. Borgloh, M. Güllner and K. Wilking (2003), Wenn aus Liebe rote Zahlen werden. Über die wirtschaftlichen Folgen von Trennung und Scheidung, Opladen: Westdeutscher Verlag. Becker, G.S. (1981), A Treatise on the Family, Cambridge, MA: Harvard University Press. Böttcher, K. (2006), ‘Scheidung in Ost- und Westdeutschland. Der Einfluss der Frauenerwerbstätigkeit auf die Ehestabilität’, MPIDR Working Paper 2006016. Bundesministerium für Arbeit und Soziales (2007), website, updated 2 April 2007, http://www.bmas.bund.de/Coremedia/generator/16194/filter5Thema:Sociale1 Sicherung/ergebnisse.html, accessed May 2008. Bundesministerium für Familie, Senioren, Frauen und Jugend and W. Cornelißen (eds) (2005), 1. Datenreport zur Gleichstellung von Frauen und Männern in der Bundersrepublik Deutschland, Munich: Bundesministerium für Familie, Senioren, Frauen und Jugend. Bundesministerium für Familie, Senioren, Frauen und Jugend (2006), ‘Der Unterhaltsvorschuss, Bundesministerium für Familie, Senioren, Frauen und Jugend’, updated 26 September 2006, www.bmfsfj.de/Politikbereiche/ familie,did534088.html, accessed May 2008. Burkhauser, R.V., G.J. Duncan, R. Hauser and R. Berntsen (1991), ‘Wife or frau, women do worse: a comparison of men and women in the United States and Germany after marital dissolution’, Demography, 28, 353–60. Burkhauser, R.V., G.J. Duncan, R. Hauser and R. Berntsen (1990), ‘Economic burdens of marital disruptions: a comparison of the United States and the Federal Republic of Germany’, Review of Income and Wealth, 36, 319–333. Coleman, J.S. (1994), Foundations of Social Theory, Cambridge, MA: Belknap. Daly, M. (2000), The Gender Division of Welfare. The Impact of the British and German Welfare State, Cambridge: Cambridge University Press. Dewilde, C. (2002), ‘The financial consequences of relationship dissolution for women in Western Europe’, in E. Ruspini and A. Dale (eds), The Gender Dimension of Social Change. The Contribution of Dynamic Research to the Study of Women’s Life Courses, Bristol: Policy Press. Diekmann, A. (1994), ‘Hat das steigende Ehescheidungsrisiko das berufliche Engagement von Frauen gefördert?’, Soziale Welt, 45, 83–97. DiPrete, T.A. and P.A. McManus (2000), ‘Family change, employment transitions, and the welfare state: houshold income dynamics in the United States and Germany’, American Sociological Review, 65, 343–70. Düsseldorfer Tabelle (2005), website, updated 1 Juli 2005, www.olg-duesseldorf. nrw.de/service/ddorftab/ddorftab5/2005-07-01 percent20duess-tab.pdf, accessed March 2007. Emmerling, D. (2005), ‘Ehescheidungen 2003’, Wirtschaft und Statistik, 12/2005, 97–108. Engstler, H. (1997), Die Familie im Spiegel der amtlichen Statistik, Bonn: Bundesministerium für Familie, Senioren, Frauen und Jugend.

76

Single country studies

Engstler, H. and S. Menning (2004), Families in Germany – Facts and Figures, Berlin: Bundesministerium für Familie, Senioren, Frauen und Jugend. Esch, K. and S. Stöbe-Blossey (2005), ‘Arbeitsmarkt und Kinderbetreuung – Anforderungen an die Neustrukturierung eines Dienstleistungsangebots’, in Institut Arbeit und Technik (ed.), Jahrbuch 2005, Gelsenkirchen: Institut Arbeit und Technik. Esping-Andersen, G. (1990), The Three Worlds of Welfare Capitalism, Cambridge: Polity Press. Gysi, J. and D. Meyer (1993), ‘Leitbild: Berufstätige Mutter - DDR-Frauen in Familie, Partnerschaft und Ehe’, in G. Helwig and H.M. Nickel (eds), Frauen in Deutschland 1945–1992, Bonn: BpB, pp. 166–90. Herlyn, I., A. Kistner, H. Langer-Schulz, B. Lehmann and J. Wächter (1998), Großmutterschaft im weiblichen Lebenszusammenhang. Eine Untersuchung zu familialen Generationenbeziehungen aus der Perspektive von Großmüttern, Pfaffenweiler: Centaurus. Hinz, T. and H. Gartner (2005), ‘Geschlechtsspezifische Lohnunterschiede in Branchen, Berufen und Betrieben’, Zeitschrift für Soziologie, 34 (1), 22–39. Holden, K.C. and P.J. Smock (1991), ‘The economic costs of marital dissolution: why do women bear a disproportionate cost?’, Annual Review of Sociology, 17, 51–78. Kampmann, C., J. Schupp and G. Wagner (1996), Erwerbs- und Einkommensentwicklung nach wichtigen Lebensereignissen in Westdeutschland, Bonn: Bundesministeriums für Arbeit und Sozialordnung. Klammer, U. and C. Klenner (2003), ‘Geteilte Erwerbstätigkeit – gemeinsame Fürsorge. Strategien und Perspektiven der Kombination von Erwerbs- und Familienleben in Deutschland’, Jahrbuch für Europa- und Nordamerika-Studien, 2, 177–207. Klein, T. and J. Kopp (eds) (1999), Scheidungsursachen aus soziologischer Sicht, Würzburg: Ergon Verlag. Klenner, C. (2002), Geschlechtergleichheit in Deutschland, Aus Politik und Zeitgeschichte, Beilage zu das Parlament, B 33–34, pp. 17–28. Konietzka, D. and M. Kreyenfeld (2005), ‘Nichteheliche Mutterschaft und soziale Ungleichheit. Zur sozioökonomischen Differenzierung der Familienformen in Ost- und Westdeutschland’, MPIDR Working Paper 2005-001. Lewis, J. and I. Ostner (1994), ‘Gender and the Evolution of European Social Policies’, Zentrum für Sozialpolitik der Universität Bremen Arbeitspapier No. 4. Martiny, D. (1987), ‘Des widerspenstigen Schuldners Zähmung – Zur Soziologie des Unterhaltsrechts’, Zeitschrift für Rechtssoziologie, 8, 24–56. Mincer, J. and H. Ofek (1982), ‘Interrupted work careers: depreciation and restoration of human capital’, Journal of Human Resources, 17 (1), 3–24. Münder, J. (2005), Familienrecht. Eine sozialwissenschaftlich orientierte Darstellung, Munich: Luchterhand. Pfau-Effinger, B. (2004), Development of Culture, Welfare States and Women’s Employment in Europe, Aldershot: Ashgate. Ruspini, E. (1998), ‘Women and poverty dynamics: the case of Germany and Britain’, Journal of European Social Policy, 8, 291–316. Sainsbury, D. (1999), Gender and Welfare State Regimes, Oxford: Oxford University Press. Scheiwe, K. (1999), Kinderkosten und Sorgearbeit, Frankfurt a. M.: Klostermann.

Germany: will the male breadwinner model survive?

77

Schwarze, J. and J. Härpfer, (2000), ‘Die wirtschaftlichen Folgen von Trennung und Scheidung für Familien: Erwerbstätigkeit, Einkommen und Lebenszufriedenheit’, ifb-Materialien 4/2000. Sørensen, A. (1994), ‘Women’s risk and the economic position of single mothers’, European Sociological Review, 10, 173–88. Statistisches Bundesamt Deutschland (ed.) (2004a), Datenreport 2004, Bonn: Bundeszentrale für Politische Bildung. Statistisches Bundesamt Deutschland (ed.) (2004b), Kindertagesbetreuung in Deutschland. Eintrichtungen, Plätze, Personal und Kosten 1990 2002, Wiesbaden: Statistisches Bundesamt. Statistisches Bundesamt Deutschland (2006), website, www.destatis.de, accessed 8 February 2008. Uunk, W. (2004), ‘The economic consequences of divorce for women in the European Union: the impact of welfare state arrangements’, European Journal of Population, 20, 251–85. Vaskovics, L.A. (2000), ‘Erziehungsurlaub – Akzeptanz und Inanspruchnahme’, in B. Jans, A. Habisch and E. Stutzer (eds), Familienwissenschaftliche und familienpolitische Signale, Grafschaft: Vektor-Verlag. Wagner, M. (1997), Scheidung in Ost- und Westdeutschland, Frankfurt a. M.: Campus Verlag. Wagner, M. (2002), ‘Familie und soziales Netzwerk’, in R. Nave-Herz (ed.), Kontinuität und Wandel der Familie in Deutschland: eine zeitgeschichtliche Analyse, Stuttgart: Lucius und Lucius. Wagner, M. and B. Weiß (2003), ‘Bilanz der deutschen Scheidungsforschung. Versuch einer Meta-Analyse’, Zeitschrift für Soziologie, 32, 1–21. Ziefle, A. (2004), ‘Die individuellen Kosten des Erziehungsurlaubs: Eine empirische Analyse der kurz- und längerfristigen Folgen für den Karriereverlauf von Frauen’, WZB Discussion Paper SP I 2004-102.

3.

Belgium: economic hardship despite elaborate childcare and leave time programmes Dimitri Mortelmans, Laurent Snoeckx and Peter Raeymaeckers

INTRODUCTION From a European standpoint, Belgium is a country of averages in many respects. The country lags behind in terms of the demographic evolution that has taken place in the Northern European countries, but it is ahead of developments in Southern Europe. Neither Belgium’s employment rates nor its poverty figures have ever been exceptional from a European perspective. One of the domains in which Belgium takes a leading position, however, is the field of marriage dissolution. Longitudinal divorce calculations show that 60 per cent of the marriages contracted in 1995 will eventually end in divorce (Vanhove and Matthijs, 2002). Divorce rates have quadrupled over the past 30 years and are still increasing steadily. In the European and North American context, many explanations have been put forward for these trends. From a demographic perspective, this evolution is integrated in detail into the theory of the Second Demographic Transition as formulated by Lesthaeghe (1995a, 1995b; Lesthaeghe and Neels, 2002; Lesthaeghe and Van de Kaa, 1986) and Van De Kaa (1994). According to this theory, the dramatic changes in demographic behaviour are due to the cultural shift during the 1950s and 1960s. In fact, a causal link between micro changes in belief systems and changes in demographic behaviour (for example, births out of wedlock, cohabitation and divorces) has resulted in shifting macro trends. As indicated above, Belgium is undergoing a delayed evolution compared with Northern Europe, but it is catching up, and currently shows one of the highest divorce rates in Europe. Explanations for the Belgian demographic evolution are always complicated by the country’s idiosyncratic state structure. Even though Belgium is one of the smallest countries in Europe, its political system is very complex and incomprehensible to outsiders. After World War II, the 78

Belgium: economic hardship

79

Belgian state structure evolved from a unitary state to a federal state, in which the central federal authority weakened and the regional authorities were granted legislative and executive powers in the fields of employment, education, and welfare. Unlike the German Federal State, in which the Länder operate independently from the federal authority, Belgium has two types of regional authorities. First, the Communities (three in total: the Flemish Community, the French Community and the German-speaking Community) have jurisdiction over person-related matters. At this level, policy-making is carried out in the fields of cultural issues (theatre, libraries, and so on), education, the use of languages and matters relating to the individual – which concern health policy (curative and preventive medicine) on the one hand and assistance to individuals (protection of youth, social welfare, aid to families, immigrant assistance services, and so on) on the other. Next, the Regions (also three, but geographically somewhat different from the Communities: the Flemish Region, the Brussels-Capital Region and the Walloon Region) execute powers on territorial issues. Their powers extend to economic matters, employment, agriculture, energy, transport, local authorities and other domains. Divorce legislation, which is rooted in the French Code Napoléon of 1804 (Senaeve, 2004a), has remained under the jurisdiction of the federal authority. This is also true for tax policy, national defence, foreign affairs and social security legislation. Flemish as well as Walloon individuals seeking a divorce have to go through identical divorce procedures and will be confronted with the same tax policy and social security consequences.

DEMOGRAPHIC PROFILE AND LABOUR MARKET TRENDS Before looking at divorce legislation and family policy in Belgium, we first sketch a general demographic profile and give an economic outline of the country. In the demographic section, we pay special attention to the evolution of marriage formation and dissolution, to new living arrangements such as cohabitation and to fertility trends. The sketch of the economic profile will mainly offer an overview of poverty rates and the composition of the Belgian labour market, and highlight trends in (un)employment figures and male and female part-time employment. Demographic Changes Just like other European countries, Belgium has witnessed quite dramatic demographic changes since World War II. The family has changed

80

Single country studies

tremendously, and new trends in partnership formation have arisen. This so-called ‘Second Demographic Transition’ (Lesthaeghe and Neels, 2002) – in which individual autonomy and female emancipation play a major role – is characterized by an accelerating upward trend in divorces, advancement in the age of marriage, and a rise in new family forms such as premarital and non-marital cohabitation (which lead to a rise in extra-marital births), single-parent families (most of them headed by women), remarriages and post-marital cohabitations (Lesthaeghe, 1995b). When we look at Belgian nuptiality figures and compare them with European figures, it becomes clear that, on the one hand, the Belgian demographic evolution started later than that in Northern European countries. On the other hand, the latest figures (2000) show that the divorce rate, at 2.6 per thousand population (see Data Appendix), is among the highest in Europe, while the marriage rate is among the lowest, namely 4.4 (see Data Appendix). Data on family structure in Belgium (see Data Appendix) show that the share of one-person households has been rising since 1985. In 2000, it became the most common household type (31.2 per cent). The number of single female households is higher than that of single males because of the longer life expectancy of women (FPS Economy – Directorate-General Statistics Belgium, 2004b; not shown in Data Appendix). Another large share of Belgian households consists of couples with and without children (30.2 per cent and 22.2 per cent, respectively). As the percentages show (see Data Appendix), the proportion of couples in the population of households has been declining since 1985. However, the relative number of single parents had risen to 11.1 per cent in 2000. Within this group, single-mother households are in the majority – a situation partly due to custody arrangements, which almost automatically assigned children to the mother until 1995 (when Belgian divorce legislation was amended substantially). Information on the number and distribution of family units (FPS Economy – Directorate-General Statistics Belgium, 2004b) indicates that single-father households account for 161 000 units, whereas the number of single-mother households adds up to 431 000 (out of a total of about 2 760 000 households). High divorce rates, declining marriage rates, and the rise in new living arrangements such as pre-marital, non-marital and post-marital cohabitation have led to a higher proportion of children born out of wedlock, as can be clearly seen in the Data Appendix. The crude divorce rates show an accelerating trend from 1985 until 2004 (with a temporary peak in 1995 due to the shortened and simplified divorce procedure thereafter), together with a declining crude marriage rate (5.8 per cent in 1985 to 4.1 per cent in 2003) and a rise in births outside of wedlock (7.07 per cent in 1985 to 31.0 per cent in 2003). The Total Fertility Rate (TFR), which had been

Belgium: economic hardship

81

declining since 1960, shows a rather modest rise from 1985 onward (see Data Appendix) due to a recuperation of first child postponements. Still, part of the TFR increase is due to the Belgian population with migrant roots. Within this migrant population, the fertility rate is higher than the national average. If fertility rates among these groups decrease, Belgium will have a negative population growth. Since 1970, a growing share of non-marital cohabitations and remarriages has been observed. In 1970, 11 per cent of all marriages were remarriages. In 2000, this proportion had already risen to 31 per cent (FPS Economy – Directorate-General Statistics Belgium, 2004b). Unfortunately, Belgium does not publish official statistics on cohabitation. Recently, Corijn (2004) published estimates of the share of cohabiting households based on the National Register. In 1970, about 4.3 per cent of all households were cohabiting partners. The number of cohabiting households increased to 4.7 per cent in 1991 and 7.6 per cent in 2004. About half of these couples have never been married, one-fifth consist of two divorced partners and 18 per cent are households containing at least one divorcee (Corijn, 2005). Parallel to a rise in second marriages, we can observe that the increase in cohabitation is partly due to a rise in divorces. A large number of divorcees start new relationships and either remarry or cohabitate. Belgium has recognized same-sex marriages (see below). At the time of the policy’s official instalment in 2003, a total of 1631 samesex marriages were concluded (58.8 per cent were male couples). In 2004, this number rose to 1904 (54.8 per cent male couples; FPS Economy – Directorate-General Statistics Belgium, 2004a, 2005). The Belgian Labour Market The structure of the labour market is a crucial factor in the post-divorce coping process (Raeymaeckers et al., 2008b; Uunk et al., 2005). Access to the labour market, and possibilities for increasing one’s labour market participation – or, inversely, of easily reducing one’s working hours – are important to former spouses who are re-adapting their lives to a new situation. As in many European countries, the Belgian labour market is characterized by extensive regulation. Although laws on working times and employment protection became more flexible during the 1980s and 1990s, a strict regulation of labour contracts still prevails (Dewilde, 2006). Whereas unemployment rates for men have fluctuated around 6 per cent since 1985, the rates for women have decreased, from 16.7 per cent in 1985 to 9.5 per cent in 2005 (see Data Appendix). These general figures, however, conceal the structural heterogeneity of labour market risks – for example,

82

Single country studies

migrants and people who leave school early are quite vulnerable on the labour market. In addition, there are significant inequalities between the regions in terms of unemployment: 3.8 per cent in Flanders, 6.8 per cent in Wallonia and 9.9 per cent in Brussels (Steunpunt WAV, 2005). Long-term unemployment is rather high in Belgium, as repeatedly indicated by several European and international studies (see De Lathouwer et al., 1999, for figures). One cause of this is the country’s rather high unemployment benefits, which are (almost) unlimited in duration (De Lathouwer et al., 1999). Employees under the age of 50 who are cohabiting are the only people who can lose their unemployment benefits. Singles and heads of households can never be suspended. Belgian men have an employment rate of 68.3 per cent, which is rather low compared with other European countries. While employment figures for women have increased since 1990, and reached 53.8 per cent in 2002 (see Data Appendix), the women’s employment rate has nevertheless remained rather low from a comparative perspective. Especially in comparison with Northern European countries like Denmark, Finland and Sweden (but also Great Britain), women are not employed as often as one would expect when taking Belgium’s elaborate childcare facilities into account (see below). The percentage of working mothers with children under the age of six (see Data Appendix), however, amounted to 68.8 per cent in 2005. Only Denmark and Sweden do better and have higher participation rates. The influence of generous family benefits and extensive childcare services that facilitate the work–caregiving combination clearly comes into play. In addition – and unlike policies in countries such as Germany, Great Britain and Spain – this social policy, which is situated on the boundary between a traditional male breadwinner and a dual-earner model (Andreß et al., in this volume), protects single mothers from poverty (see Data Appendix). It should be noted, however, that single-parent families in Belgium (single mothers making up 80 per cent of the cases) have a high poverty rate (36 per cent) compared with couple households (13 per cent on average; see Data Appendix). Looking at part-time employment, the proportion of Belgian men working part-time is low and hardly changes over time (from 6.3 per cent in 1985 to 7.6 per cent in 2005). The share of women in part-time jobs, on the other hand, has risen from 16.4 per cent in 1985 to 40.5 per cent in 2005. Moreover, male part-time employment is found almost exclusively in the groups of men under 19 years or above 65 years of age. Part-time jobs are relatively common; small part-time jobs are virtually non-existent (Steunpunt WAV, 2005). The average number of hours worked weekly in part-time jobs is 21.5, placing Belgium between France (22.6) on the one hand and Germany (19.5) and the Netherlands (18.4) on the other (De Lathouwer et al., 1999).

Belgium: economic hardship

83

Belgium’s demographic as well as its labour market characteristics provide some indication of what divorce entails for Belgian men and women. Although divorce is a prevalent life-course risk in Belgium, there are no specific measures to assist individuals who divorce and/or become single parents. However, Belgium’s well-regulated labour market and low poverty rates lead us to expect that the negative effects of divorce will be less severe in comparison with other European countries.

FAMILY LAW AND DIVORCE IN BELGIUM In this section, we present an overview of the different options Belgian law provides as far as marriage and separation are concerned. We discuss the history of Belgian divorce legislation and its recent evolutions, and present the different divorce options and their consequences concerning property, children and alimony. Because cohabitation is rapidly gaining importance as a household type, we also pay attention to the protective measures available in case a cohabiting relationship breaks up. Introduction: Getting Married Before turning to Belgian legislation on divorce and separation, it is important to describe the ways in which couples can legally unite. The traditional union is, of course, marriage. As an alternative to marriage, people can cohabit. This factual union has neither legal consequences, nor does it offer any form of protection when things turn out badly. In 2000, Belgium introduced a ‘Declaration of Cohabitation’,1 according to which two cohabiting people can have their cohabitation registered and enter a state of ‘legal cohabitation’. The declaration introduces several protective measures for cohabiting couples, which resemble those for married ones. The statement can be given by any couple, regardless of the gender of both partners. Therefore, the Declaration of Cohabitation was seen as the first official registration for homosexual couples. A next step in the development of family rights for same-sex couples was taken in 2003, when Belgium allowed people of the same sex to marry. There is one restriction: adoption is still reserved for heterosexual couples. Belgian Divorce in Historical Perspective Belgian divorce law originates from the French legal system. In the Ancien régime de la France, most family law matters fell within the jurisdiction of the Roman Catholic Church (Senaeve, 2004a). Belgium became

84

Single country studies

independent in 1830 and ratified the French Code Napoléon in 1884. One of the regulations which were left out at that time was divorce by mutual consent. The 1884 law remained in force until 1974, when a major breakthrough was realized in terms of no-fault divorce. Spouses of mentally ill persons were allowed to apply for divorce, and married couples could get a divorce after they had lived separately for ten years. In 1994, Belgian divorce legislation again went through some major changes which made the dissolution of marriage much easier. The so-called ‘double introduction’, which consisted of a divorce claim and an application, was replaced by a mere application preceding the trial (Senaeve and Pintens, 1997; Senaeve et al., 1994). The divorce procedure was thereby simplified and shortened by six months. Another new element in the altered divorce procedure was a modification of the obligation that both partners try to save their marriage. This obligation was dropped and it is the partners’ free decision whether to do so or not. The announcement of the new law in 1993 kept many couples from applying for a divorce, which is clearly visible in the divorce rates halfway through the 1990s. Belgian Divorce Legislation Current divorce legislation includes three standard procedures: first, there is divorce by mutual consent, which is by far the most frequently used procedure; secondly, couples can terminate their marriage based on factual separation. The third procedure is marriage dissolution based on ‘facts’: one of the spouses is confronted with misconduct on the part of his or her partner. In addition to these three standard procedures, there is ‘separation from bed and board’, which is often mentioned as a fourth way of divorcing (Heyvaert, 2002). Technically, this is not a divorce in legal terms, but only a temporary measure for married people who decide to live separately. Therefore, we limit our discussion to the three standard procedures. In the case of divorce by mutual consent, both spouses agree to dissolve their marriage. In contrast to the other two procedures, the question of guilt is not raised (Senaeve and Pintens, 1997). Two conditions are required in order to qualify for a divorce by mutual consent: first, both spouses need to be older than 20; secondly, the marriage must have lasted for at least two years (Heyvaert, 2002; Senaeve, 2004a). In case of a divorce by mutual consent, the court does not need to intervene in matters relating to property, family and personal rights. It is up to the partners to agree on the partition of joint property (Senaeve, 2004a), their new addresses, the parental authority and the financial arrangement concerning children – if necessary, an alimony agreement. The mutual consent procedure takes about five months.

Belgium: economic hardship

85

The legal state of factual separation is determined by two conditions. First, it requires a factual separation of a certain period of time. In 2002, the duration of the separation was two years (Heyvaert, 2002). Before that, the time span had already been gradually reduced, from ten years in 1979 to five years in 1982. Secondly, the separation must reveal an irretrievable breakdown of the marriage. This means that the marriage is beyond saving. The married couple does not live together, and they have cut off all contact with each other. Until the moment of the actual divorce, the court takes temporary measures based on a consensus. Whereas before 1994, the authority of the civil court was limited to temporary measures concerning children, the Act of 1994 expanded its jurisdiction to include temporary measures concerning the persons involved, their goods and the maintenance of the spouses (Senaeve and Pintens, 1997). The third procedure, divorce based on facts, can be initiated when just one member of the couple is confronted with misconduct on the part of his wife or her husband for example, the guilty spouse seriously neglects his or her marital duties, or some factual grounds are present, such as adultery, acts of violence, ill treatment or gross insults (De Corte, 2001). If the offended spouse wins the case, he or she will retain all the advantages of the marriage (De Corte, 2001). Consequences of Divorce in Belgian Family Law During several of the above mentioned procedures, temporary and complementary measures are taken by the couple or the court. Family law regulations govern the process of the divorce and stipulate rules concerning the living situation of the ex-spouses and their children. By the end of the procedure, measures need to be taken with respect to the division of property and the custody of the children, if there are any. If the divorce procedure is based on facts, the question of alimony for the ex-spouse will also arise. The division of property after marriage dissolution depends upon the couple’s marriage settlement. Belgian family law mentions three different contracts (Gerlo, 2001): joint property, division of goods and participation of property. Most Belgian couples prefer joint ownership of property (this is also the standard contract when a couple does not opt for any system), which means that in case of a divorce, all property will be divided equally between the husband and the wife. When opting for the division of goods, spouses keep all property rights at the beginning of their marriage and maintain them in case of a divorce. Marriage or divorce actually has no effect on the spouses’ properties. The third system concerns a participation of property. During the marriage, partners divide the ownership of

86

Single country studies

their goods. In case of a divorce, all wages, interests and property acquired during the marriage are divided according to the division that was made during the marriage. Children are the most important subject of debate in almost every divorce trial. In general, separated couples fight more often over children than they do over property. In 1995, one year after the adaptation of the divorce procedure, a major amendment to custodial law was implemented which stated that both parents have an equal right to raise their children (Uyttenhove et al., 2004). Before 1995, one of the spouses was granted custody rights and the other had visiting rights. Joint parenting has now become the standard. Nevertheless, when partners do not agree on the custody arrangements, or if one of the partners refuses to raise the children, the court takes compulsory measures. Belgian family law clearly stipulates that the judge has to take one criterion into consideration, and that is the interest of the child (Senaeve, 2004b). Children can be placed under guardianship or parental authority can be granted to only one parent, while the other is granted visiting rights. Although property regulation is a decisive element of the economic consequences of divorce, the financial situation of the ex-partners often depends on the amount of alimony received (or paid) by the ex-partner or children. Maintenance obligations can only be adjudicated in the case of a divorce based on facts, with the result being that the guilty spouse pays alimony to the innocent party. Belgium has no fixed alimony rates. The basic idea is that the alimony has to guarantee that the innocent spouse can maintain the standard of living that he or she had before the divorce. The court also considers the financial capacity of the claimant in terms of income and job opportunities. The financial situation of the debtor is also taken into account: the amount of alimony may never exceed one-third of his or her income (Gerlo, 2003; Senaeve, 2004b). However, as Cantillon (1999) states, 50 per cent of divorced women report that they do not receive any maintenance payments. Many partners are reluctant to pay or, in many cases, are not capable of financially supporting two families. Moreover, the mean amount of these payments only comes to about €125 (1999 figures), which implies that they represent only 8 per cent of the family income (Cantillon, 1999). The obligation to pay alimony terminates as soon as the spouse receiving it starts cohabiting with someone, or when he or she remarries. In the case of children’s alimony, other factors are taken into consideration. Both parents pay equally for the upbringing of the children. When one parent is given full custody, the other parent pays alimony for the children. The determination of the alimony rate is based on the same principle as the spouse’s alimony: the children’s standard of living has to be identical

Belgium: economic hardship

87

to the pre-divorce situation. Alimony is obligatory when the children are under age, or as long as their education is not finished. In the case of nonpayment, there is no official child maintenance regulation that provides additional benefits for single mothers.2 Cohabitation in Belgian Family Law Two people are officially cohabiting when they hand in a written application to the municipal registry (Gerlo, 2003). In terms of the couple’s legal protection, the state of cohabitation resembles the marital state. Even though protection is not as extensive as it is for married couples, the law provides a minimum of protection to officially cohabiting partners. Because of the importance of the couples’ material situation, cohabitation legislation provides insurance against any unacceptable behaviour on the part of one of the partners (Senaeve, 2004b). Three main types of protection can be discerned. First, the couple’s joint property is protected. This aspect refers primarily to a joint rent law, which implies that both partners become holders of the tenancy agreement. Thereby, both partners have to approve any cancellation, notification and judicial demands (under penalty of nullity). Secondly, a joint debt liability has been installed. When one of the cohabiting partners contracts a debt, the creditor can recover the outstanding debt from both partners. Third – and comparable to the marital system – legal cohabitation comes with particular mutual obligations. The cohabiting couple is obliged to live in the same residence; they have to be faithful and give moral support and material assistance to each other (Senaeve, 2004b). A registered partnership ends when one of the partners dies, gets married, or when a concerted or unilateral declaration is presented at the municipality. When a cohabiting relationship ends, none of the partners or children has a right to alimony.

WELFARE STATE ARRANGEMENTS AND DIVORCE IN BELGIUM Having discussed the consequences of divorce as they are described in Belgian family law, we will now take a look at welfare state regulations. In Belgium, divorce is not directly related to the social security system, because social security rights are primarily based on labour market relations and not on household composition. In that respect, the Belgian social welfare system is based on the logic of the classic male breadwinner. Divorce is not officially recognized as a social risk (it only plays a minor part in the

88

Single country studies

regulation of pensions). In practice, however, it is clear that the loss of a spouse can have huge implications on the social and economic situation of both ex-partners. In this brief overview, we touch upon family allowances, childcare services, leave systems and unemployment insurance. With respect to Esping-Andersen’s welfare state typology, Belgium is usually assigned to the group of Continental European welfare regimes (Andries, 1997; Esping-Andersen, 1991). It has a low score in the cluster of liberal regimes and an average score in the cluster of social democratic regimes. Together with a high degree of decommodification, Belgium scores considerably high on ‘corporatism’ and ‘etatism’ indices. The system makes a strict division between regulations for employees, civil servants and self-employed persons. A second trait is the degree to which health and unemployment benefits and (partially) pensions are related to earnings. This is a carry-over from the classic male breadwinner model which is deeply rooted in the system. This principle is also represented in generous family benefits and social transfers which support a traditional division of labour. Nevertheless, an elaborate childcare infrastructure, enabling women to combine work and family, provides evidence that Belgium also has characteristics of the dual-earner model (Andreß et al., in this volume). Belgium scores high in the area of public expenditures for family services. Only Denmark, Finland and Sweden do better (see Data Appendix). Concerning the share of its gross domestic product (GDP) spent on family cash benefits, Belgium is the absolute front-runner, with 1.8 per cent. The allowance system for families with children is one of the most universalistic in the world (Cantillon, 1999). However, there are some forms of selectivity which are very efficient – for example, for unemployed persons with children, more than 60 per cent of the enlarged benefits are granted to lowincome families, which leads to large poverty reductions. Child benefits are paid until a child reaches the age of 18 (or even 25 if the child is a student or is disabled). Child benefits are socially and age adjusted. The basic amount (from €77 per month for the first child to €213 per month from the third child onwards) can be raised for long-term unemployed persons (from an additional amount of €39 per month for the first child to an additional amount of €4 per month from the third child onwards) and as the child gets older (for the first child from an additional €13 per month [6–12 years old] to an additional €26 per month [>18 years old] and from an additional €26 per month to an additional €51 per month for other children). The Belgian childcare system is quite comprehensive, providing care for 30 per cent of all children up to the age of 3 and for 97 per cent of children between 3 and 6 years old (see Data Appendix). Moreover, there are tax deductions for the costs of childcare for children up to the age of 3 (Cantillon, 1999).

Belgium: economic hardship

89

In addition, Belgium provides maternity, paternity and parental leave. Working women are entitled to 15 weeks of maternity leave, while working men are given 10 days off. A three-month parental leave is also available. The partners can either split the period between the two of them or decide that one of them will take the full leave. In practice, it is most often the mother who takes the three-month leave. The Belgian unemployment insurance system is a rather elaborate and complex legislative structure. The calculation of unemployment benefits depends on the earnings of the unemployed person before he or she lost his or her job, the period of unemployment and the family situation. However, there is a tendency to loosen the link between employment earnings and social rights (Dewilde, 2006). Coverage rates are very high and the duration of benefits is next to unlimited, resulting in low poverty rates.

DIVORCE RISKS AND ECONOMIC CONSEQUENCES OF DIVORCE IN BELGIUM Research on the consequences of partnership dissolution consistently shows a larger breakup penalty for women in comparison with men. Overall, women experience a net income loss of 30 to 40 per cent during the first few years after the dissolution (Bianchi et al., 1999; Kalmijn and Broese van Groenou, 2005; Poortman, 2000; Smock, 1994). The extent of the financial consequences, however, is highly dependent on the national setting. In their review of the economic consequences of divorce, Andreß et al. (in this volume) find differences ranging from an income loss of 80 per cent (a German study from 1984) to an income gain of 7 per cent (a Belgian study from 2003). In the literature, these vast differences are usually explained by the income-related and employment-related welfare arrangements typical for each country. As Uunk (2004) and Dewilde (2002) show, welfare regime typologies explain, to a great extent, the way in which economic consequences are reduced after a divorce. Either welfare transfers are available to compensate for the income loss, or institutional arrangements enable the individual’s (re-)entry into the labour market. For those who are already working, these arrangements facilitate the combination of work and caregiving. Family-friendly policies (Raeymaeckers et al., 2008a, 2008b) and the accessibility of large-scale childcare provisions seem to facilitate women’s working opportunities (Uunk et al., 2005). In comparative analyses, Belgian women often manage to reduce the negative economic consequences of divorce better as compared to women in other countries. In most analyses, Belgium is classified within the group of conservative welfare regimes – even though it is acknowledged that on

90

Single country studies

family policy indicators, Belgium scores close to the social-democratic group in the Esping-Andersen typology (Andreß et al., in this volume; Uunk, 2004). Uunk finds the lowest income changes (minus 20 per cent) among the conservative regimes, while Andreß et al. find better outcomes in comparison to Germany. A combination of high social transfers and an elaborate childcare infrastructure accounts for this result. Even though childcare facilities do not moderate the negative effects as such, they are responsible for high labour participation rates among mothers. Available childcare facilities help divorced mothers to combine their labour activities with an increased care responsibility in case they are given custody of the children. To conclude this chapter, we focus on the income effects of relationship dissolution for Belgian men and women, based on data from the Belgian Household Panel (PSBH). Before doing so, we illustrate the range of different divorce risks in Belgium. Since divorce – together with inactivity linked to unemployment or disability – is the major life event leading to poverty in Belgium (Dewilde, 2006), it is important to gain an insight into the risk factors for divorce. Controlling for socio-demographic background variables, relationship characteristics are the best predictors of divorce in Belgium. Premarital cohabitation is the most important negative predictor of divorce, while marriage duration appears to be the most protective element (Snoeckx et al., 2005). In addition, significant differences are found between the regions (Flanders and Wallonia). A close look at these regional differences shows that the effects of premarital birth and the presence of children (as a protective factor) are stronger in Flanders. This goes hand in hand with the effect of the number of religious persons in the regions. It appears that cultural determinants – especially the historical Roman Catholic character of Flemish society – still exert an important protective force on the risk of getting a divorce (Snoeckx et al., 2007). Bearing these risk factors in mind, we now look at the financial consequences of divorce in Belgium. For this purpose, we have used the Belgian Household Panel. This panel ran for 11 years between 1992 and 2002. From 1994 onwards, it served as the basis for the Belgian section of the European Community Household Panel (ECHP). The motivation for using the PSBH instead of the ECHP is its more detailed data (especially concerning income) and the higher number of waves available for analysis (for additonal information on the PSHB see the Data Appendix). As we look at all divorces occurring during the panel, the PSBH enables us to increase the statistical power of the analysis. The sample we analyse consists of people who dissolved their marriage3 during the panel. We adopted all legal divorces, but also all married couples no longer living together in a subsequent wave. There may be a delay in the official dissolution of the marriage. Therefore, the marital status of the partners alone does not capture all

Belgium: economic hardship

91

Table 3.1 Equivalized household income (in euros) before and after relationship dissolution N Men

88

Women

108

Source:

Mean Std. Deviation Median Mean Std. Deviation Median

Before (t 2 1)

After (t 1 1)

1148.5 387.2 1165.8 1036.8 396.2 970.4

1418.4 751.8 1223.0 871.1 409.3 787.7

Change

14.9%

218.8%

PSBH (11 waves).

divorcees. Since we focus on the economic consequences of a divorce, the year of the split marks the timing of the targeted event more closely. The descriptive statistics in Table 3.1 show equivalized disposable household incomes4 for men and women one year before and one year after a relationship dissolution. Before relationship dissolution, the incomes of men and women do not differ significantly.5 After a divorce, income differences become highly significant (p , 0.001). For men, we find an increase in equivalized income of nearly 5 per cent. For women, the difference is negative, as expected. The difference in median equivalized income amounts to 18.8 per cent, which is close to the result (20 per cent change) of Uunk’s (2004) analysis with the ECHP. In Table 3.2, we present the results of an ordinary least squares (OLS) regression on the proportional change in income one year after a marital breakup. The aim is to look at pre-divorce and post-divorce indicators explaining the size of the economic change. Lower values on the dependent variable indicate a small or even negative change in income after divorce. Two models were estimated. In model 1, the dependent variable is the ‘raw’ proportional change variable (the change score, for short). Model 2 uses a multiplicative inverse transformation of this variable to correct for the non-normal distribution of the change scores.6 The parameters of model 2 do not allow for a straightforward interpretation. The signs of the parameters are the inverse of those in model 1. A positive parameter in model 2 indicates an increase in the inverse change in equivalized household income or – stated in reverse – a decrease in the proportional change in income (as analysed in model 1). For this reason, we will interpret the parameters in model 1, even though the significance levels and relative size of the standardized regression coefficients (B) in model 2 are more exact.

Single country studies

92

Table 3.2

Proportional change of equivalized household income (in euros) after relationship dissolution (OLS regression coefficients) Model 1 Proportional change in income b

OECD-equivalized income (in euros) before dissolution Gender (woman) Number of children Age Not working at t21 Started working at t11 Education Higher secondary Higher education Has new partner at t11 Region (Flanders) Intercept R² Adj. R² N Notes: Source:

B

0.0005***

Model 2 Inverse of proportional change in income b

B 0.349

20.340 20.00037***

20.407*** 20.043 0.007 20.244 0.266

20.335 20.078 0.097 20.167 0.121

0.121 0.221 0.108

0.290 0.185 20.049 0.201 20.170

20.248*** 20.072* 20.003 20.207 20.262

0.094 20.072 0.178 20.180* 0.077 20.132

20.031 1.504

20.080 20.206 20.132 0.115

20.025 20.098 20.668 0.27 0.21 208

0.32 0.27 208

* p , 0.05; ** p , 0.01; *** p , 0.001. PSBH (11 waves).

First, we find that pre-divorce income level is a highly significant predictor. The parameter of model 1 suggests that persons with higher income levels before the breakup of a marriage experience a larger loss in income. Next, women clearly suffer more from divorce than do men, as already shown in the descriptive analysis. Pre-divorce income level and gender are the two most important predictors of the change in income.7 Model 2 further shows a significant effect of being highly educated and of the number of children in the household. Surprisingly, entering the labour market after a divorce turns out not to be significant for the prediction of income change. About 8 per cent of divorcees start working after a divorce. The effects in model 1 show that entering the labour market does help to reduce the negative effects, but note that the parameter is not statistically significant.

Belgium: economic hardship

93

CONCLUSION From a European perspective, Belgium shows divorce rates which are comparable to those in the Scandinavian countries and the UK. Since 2005, the continuously rising trend has been stabilizing at a high level. Since cohabitation is quickly gaining popularity among younger cohorts, and dissolutions of cohabiting relationships are not officially registered in national statistics, a spurious decline in the divorce rate may even occur in the coming years. The Belgian welfare state can be classified as a conservative welfare regime (Esping-Andersen, 1991) with a strong accent on the family as a provider of welfare. In social security regulations, divorce is not recognized as a social risk. Poverty measures and labour market policies are generic and work related. Therefore, the male breadwinner model can be seen as having a major influence on the effects of ending a marriage. Since many pillars of the social security system are based on income, the non-working partner (usually the woman) comes up against the most severe consequences of divorce. Starting to work or increasing one’s working hours are effective strategies for fighting the income decline after a divorce, but the regression analysis in this study does not show a significant effect of these strategies. This indicates that Belgian minimum income protection succeeds both in keeping poverty rates low and in moderating the income decline after a divorce. In recent years, Belgian labour market policies have focused on two domains: the increase of labour force participation after 55 (not at stake here) and the improvement of work–life balance. With respect to divorce and its consequences, leave time schedules are trump cards in fighting the negative economic outcomes of relationship dissolution. The high labour market participation of mothers – especially younger women – proves that being a solo parent and working full-time is a tough, though feasible, combination. In addition, generous child allowances and an elaborate childcare infrastructure – which almost labels Belgium as a country with a social-democratic welfare regime – ease the economic consequences for women. The crucial question remains as to why these regulations do not succeed in fighting the dramatic economic deterioration shown in our empirical analyses. Even though women may catch up in the long run, the short-term relapse is relatively severe in light of the aforementioned safety net. Since 2000, families and family law have enjoyed renewed political interest. The most visible results of this shift are the aforementioned measures to improve the work–life balance of young households and same-sex marriages. From 2004, the breakup of cohabiting unions also seemed to be included in the political agenda. Many legislative initiatives have been

94

Single country studies

taken to change divorce legislation once more. Even though few changes have already been adopted, it is clear that the relatively stable Belgian family law will undergo several major changes. The rapidly changing demographic nature of the society, with its numerous new household forms, is putting the classic (male breadwinner-based) system under pressure. Although large internationally comparative surveys (for example, the ECHP) have reference periods that do not yet capture the consequences of these changes, it is important to note that the legal context of Belgian households is changing fast and will probably evolve even more in the years to come. Important in this respect is that legislative policies are moving towards lowering the thresholds for getting divorced without compensating for the short-term economic consequences. From a European perspective, Belgian poverty rates are low. However, if we look at divorce as a particular life-course event, income declines remain sharp, and recovery from income declines may take a long time. Due to the breadwinner-based foundations of the system, labour market policies and childcare facilities alone do not suffice. Generous and elaborate leave time schedules are predominantly favourable to dual-earner households. The absence of ‘family composition’ as a criterion for social security system benefits is largely responsible for the unfavourable position of single parents. Recently, political attention to single-parent families has increased, but as yet without any legislative result.

NOTES 1. The regulation of the ‘Declaration of Cohabitation’ was introduced by a law passed in 1998, but the actual declaration has only been in use since 2000 (Gerlo, 2003). Even though the legislation holds for the entire country, there are minor differences in the benefits adjudged to cohabitation depending on the region. 2. There is no official regulation regarding benefits available to single parents in case the other parent does not pay maintenance. However, there is a possibility for social services to pay the single parent the alimony in advance (in 2003 €125 per month). The social service will then try to regain the alimony from the other parent. This regulation only holds for single parents whose yearly income does not exceed €11 544 (situation in 2003). 3. Only marital dissolutions are considered. Cohabiting couples are not included in the sample. 4. Disposable household income has been adjusted (‘equivalized’) to household size using equivalence insights from the Organisation for Economic Co-operation and Development (OECD). Each first adult in a household is assigned a weight of 1. Further household members aged 14 years and older are assigned a weight of 0.5, while children below the age of 14 count for 0.3 (Atkinson et al., 1995; Hagenaars et al., 1994). 5. Because of intra-household attrition, the equivalized disposable household income of men and women is not identical. However, no imputations have been made. 6. We first used a log-transformation of the change score as in Uunk (2004). However, this transformation was insufficient to correct for non-normality. The multiplicative inverse transformation (see Mosteller and Tukey, 1977) turned out to be more appropriate.

Belgium: economic hardship

95

7. Interaction effects of gender with re-partnering and the number of children turn out to be non-significant.

REFERENCES Andries, M. (1997), ‘De Belgische sociale zekerheid in de typologie van de welvaartsstaten’, CSB Berichten, October, pp. 1–35. Atkinson, A., L. Rainwater and T.M. Smeeding (1995), Income Distribution in OECD Countries: Evidence from the Luxembourg Income Study, Paris: OECD. Bianchi, S.M., L. Subaiya and J.R. Kahn (1999), ‘The gender gap in the economic well-being of nonresident fathers and custodial mothers’, Demography, 36 (2), 195–203. Cantillon, B. (ed.) (1999), De welvaartsstaat in de kering, Kapellen: Pelckmans. Corijn, M. (2004), ‘Ongehuwd en gehuwd samenwonen in België. Feiten en opvattingen vanuit een sociaal-demografisch perspectief’, CBGS Werkdocumenten, 2004(8), 77. Corijn, M. (2005), ‘Huwen, uit de echt scheiden en hertrouwen in België en in het Vlaamse gewest’, CBGS Werkdocumenten, 2005(5), 98. De Corte, R. (2001), Overzicht van het burgerlijk recht: een globaal overzicht van het burgerlijk recht, aangevuld met een algemene theorie over het recht, het burgerlijk procesrecht en het strafprocesrecht, Gent: Mys and Breesch. De Lathouwer, L., B. Cantillon, M. Andries and I. Marx (1999), ‘De Belgische route in perspectief: markt, gezin en sociale zekerheid in de OESO-welvaartsstaten’, in B. Cantillon (ed.), De welvaartsstaat in de kering, Kapellen: Pelckmans, pp. 13–58. Dewilde, C. (2002), ‘The financial consequences of relationship dissolution for women in Western Europe’, in E. Ruspini and A. Dale (eds), The Gender Dimension of Social Change. The Contribution of Dynamic Research to the Study of Women’s Life Courses, Bristol: Policy Press, pp. 81–110. Dewilde, C. (2006), ‘Becoming poor in Belgium and Britain: the impact of demographic and labour market events’, Sociological Research Online, 11 (1), 19 pages. Esping-Andersen, G. (1991), The Three Worlds of Welfare Capitalism, Cambridge: Polity. FPS Economy – Directorate-general Statistics Belgium (2004a), Aantal huwelijken tussen personen van hetzelfde geslacht, in 2003, Brussels: Directorate-general Statistics Belgium. FPS Economy – Directorate-general Statistics Belgium (2004b), Bevolking en huishoudens. Huishoudens en familiekernen, Brussels: Directorate-general Statistics Belgium. FPS Economy – Directorate-general Statistics Belgium (2005), Aantal huwelijken tussen personen van hetzelfde geslacht, in 2004, Brussels: Directorate-general Statistics Belgium. Gerlo, J. (2001), Huwelijksvermogensrecht, Brugge: Die Keure. Gerlo, J. (2003), Personen – en familierecht, Brugge: Die Keure. Hagenaars, A.J.M., K. de Vas and A. Zaidi (1994), Poverty Statistics in the late 1980s: Research Based on Micro-data, Luxembourg: Eurostat Heyvaert, A. (2002), Het personen- en gezinsrecht ont(k)leed: theorieën over

96

Single country studies

personen – en gezinsrecht rond een syllabus van de Belgische techniek, Gent: Mys and Breesch. Kalmijn, M. and M.I. Broese van Groenou (2005), ‘Differential effects of divorce on social integration’, Journal of social and personal relationships, 22, 455–76. Lesthaeghe, R. (1995a), ‘The second demographic transition in western countries: an interpretation’, in K.M. Oppenheim and A.-M. Jensen (eds), Gender and Family Change in Industrialized Countries, Oxford: Clarendon Press, pp. 17–62. Lesthaeghe, R. (1995b), ‘The Second Demographic Transition in western countries: an interpretation’, in K.O. Mason and A.-M. Jensen (eds), Gender and Change in Industrialized Countries, Oxford: Clarendon Press, pp. 17–62. Lesthaeghe, R. and K. Neels (2002), ‘From the First to the Second Demographic Transition: an interpretation of the spatial continuity of demographic innovation in France, Belgium and Switzerland’, European Journal of Population, 18, 325–60. Lesthaeghe, R. and D.J. Van de Kaa (1986), ‘Twee demografische transities?’, in D.J. Van de Kaa and R. Lesthaeghe (eds), Bevolking: groei en krimp, Deventer: Van Loghum Slaterus, pp. 9–24. Mosteller, F. and J.W. Tukey (1977), Data Analysis and Regression: A Second Course in Statistics, Reading, MA: Addison-Wesley. Poortman, A.-R. (2000), ‘Sex differences in the economic consequences of separation. A panel study of the Netherlands’, European Sociological Review, 16 (4), 367–83. Raeymaeckers, P., C. Dewilde, L. Snoeckx and D. Mortelmans (2008a), ‘Childcare strategies of divorced mothers in Europe: a comparative analysis’, European Sociological Review, 24 (1), 115–31. Raeymaeckers, P., C. Dewilde, L. Snoeckx and D. Mortelmans (2008b), ‘The influence of formal and informal support systems on the labour supply of divorced mothers’, European Societies, official journal of the European Sociological Association, 10 (3), 453–77. Senaeve, P. (2004a), Compendium van het Personen – en Familierecht, Boekdeel 3 Familierecht, Leuven: Acco. Senaeve, P. (2004b), Compendium van het Personen – en familierecht, Leuven: Acco. Senaeve, P., W. Pintens (1997), De hervorming van de echtscheidingsprocedure en hoorrecht minderjarigen: commentaar op de wetten van 30 juni 1994, 27 december 1994 en 20 mei 1997, Antwerpen: MAKLU. Senaeve, P., W. Pintens and F. Buyssens (1994), De hervorming van de echtscheidingsprocedure en het hoorrecht van minderjarigen: commentaar op de wet van 30 juni 1994, Antwerpen: MAKLU. Smock, P.J. (1994), ‘Gender and the short-run economic consequences of marital disruption’, Social Forces, 73 (1), 243–62. Snoeckx, L., J. Dronkers, D. Mortelmans and P. Raeymaeckers (2007), ‘Crossnational divorce risks within one nation. Europe’s cultural boundary within one country’, paper presented at the Fifth Meeting of the European Network for the Sociological and Demographic Study of Divorce, London, 17–18 September. Snoeckx, L., P. Raeymaeckers and D. Mortelmans (2005), ‘Relationele kenmerken en echtscheiding in België, Een analyse op basis van de Panel Studie van Belgische Huishoudens’, Tijdschrift voor Sociologie, 27 (2), 157–77. Steunpunt WAV (2005), ‘De Belgische Arbeidsmarkt in cijfers’, www.steunpuntwav.be/stat/kerncijfers/kerncijfers_fs.htm, accessed 5 April 2007.

Belgium: economic hardship

97

Uunk, W. (2004), ‘The economic consequences of divorce for women in the European Union: the impact of welfare state arrangements’, European Journal of Population, 20, 251–85. Uunk, W., M. Kalmijn and R. Muffels (2005), ‘The impact of young children on women’s labour supply: a reassessment of institutional effects’, Acta Sociologica, 48, 41–62. Uyttenhove, A., S. Renette and E. Lenaerts (2004), Feitelijke scheiding, echtscheiding en sociale zekerheid, Brugge: die Keure. Van de Kaa, D.J. (1994), ‘The second demographic transition revisited: theories and expectations’, in H. Beets, J.C. van den Brekel, R.L. Cliquet, G. Dooghe and J. de Jong Gierveld (eds), Population and Family in the Low Countries 1993: Late Fertility and Other Current Issues, Amsterdam: Swets and Zeitlinger, pp. 81–126. Vanhove, T. and K. Matthijs (2002), ‘The socio-demographic evolution of divorce and remarriage in Belgium’, paper presented at the First Meeting of the European Network for the Sociological and Demographic Study of Divorce, Florence, 14–15 November.

4.

New holes in the safety net? Economic and social consequences of divorce in Denmark M. Azhar Hussain and Olli Kangas

INTRODUCTION In his classical study on poverty in York, Seebohm Rowntree (1901) showed that the risk of poverty that an individual was exposed to was associated with various phases in the formation of a family. The first poverty risk was linked to childhood. The individual who was affected by poverty had most likely been born into a family with many children, and due to the parents’ low income level, the child suffered from poverty, or from something that nowadays is called ‘childhood poverty’ or ‘child poverty’. When the individual left his childhood home and became employed, he escaped poverty but, as soon as he married and had children, the risk of poverty increased considerably. Again, when the children left home, the ‘empty nest’ phase for the parent was easier: there were no children to feed, and the person was not too old to earn a sufficient income. Nevertheless, when he got older and was unable to work, the individual fell back into poverty due to lack of proper pension protection. International comparisons on poverty rates have demonstrated that the kind of life-cycle poverty identified by Rowntree over a century ago can still be observed among some of the most advanced industrial nations (Kangas and Palme, 2005). Yet it is evident that in every country, the old cycle has been affected by changes that have occurred during the most recent decades – but to quite varying extents. There has been a universal decline in poverty among the elderly, albeit with substantial variation across different regions. In some countries – primarily in the Nordic region – poverty among families with children has also decreased and is now at the level of families without children (Ferrarini, 2006; Ferrarini and Forssen, 2005). One can accuse Rowntree of being gender-blind, since in his analyses the starting point is the so-called male breadwinner, and his famous poverty cycle overlooks the hardship caused by the premature death of 98

Economic and social consequences of divorce in Denmark

99

the breadwinner. Indeed, widowhood was one severe cause of poverty 100 years ago. Much has changed since then, and social science no longer pays much attention to widowhood. Instead, the debate revolves around the economic and social consequences of marital dissolution and single parenthood, which are perceived to be the most important new welfare risks (Taylor-Gooby, 2004). Denmark – the country we concentrate on in this chapter – is a good example of how poverty risks connected to life phases have been more or less completely eliminated and, consequently, the Rowntree poverty cycle has been flattened out. Yet, in spite of the country’s extensive welfare state, divorce can still hit the individual very hard. When it comes to problems caused by the new risks, the existing body of research indicates that in all countries, regardless of their welfare state arrangements, marital dissolution has an impact on the economic well-being of spouses and their offspring (see, for example, Gähler, 1998; Andreß et al., in this volume). The problem is growing in importance in Denmark as well. According to a survey, 54 per cent of Copenhageners think that their marriage could end up in divorce (Houmann, 2006), and in fact almost the same proportion of marriages do terminate in dissolution, and one-fifth of all families with children are headed by single parents (NOSOSCO, 2002: 39). Thus, marital dissolution continues to be a major problem for the welfare state. The prevailing international testimony is that women experience a drop in economic status after a divorce, while the economic impact is not so great for men. Some studies have even found that men’s economic position improves after divorce. The degree of deterioration and improvement varies from study to study. Perhaps the greatest changes are reported by Weitzman (1985), who argues that women’s average post-divorce economic well-being deteriorates no less than 73 per cent, while men experience an improvement of 42 per cent. The accuracy of such high averages has been strongly questioned (Peterson, 1996). But however correct those results may be, the main underlying story remains: women are hit disproportionately hard by family breakups (see, for example, Holden and Smock, 1991; Smock et al., 1999). The lion’s share of studies is based on the American experience, and the question is to what extent those results are valid for countries with more developed welfare safety nets. Indeed, comparative studies display some country-specific variation, but there are still differences between genders (see, for example, Uunk, 2004). In their study on the dynamics of low income in Sweden, Fritzell and Henz (2001) found that even in the archetypical welfare state, women are more at risk of entering into poverty after divorce than men are (see also Gähler, 1998; Andreß et al., in this volume). An old saying states that behind every successful man there is always a

100

Single country studies

woman. Indeed, there seems to be some truth in the argument: on the one hand, studies have shown that married men tend to have higher incomes than unmarried men. It has also been shown that traditional family men still reap the greatest rewards in career success as well (Schneer and Reitman, 2002). However, there are disputes revolving around this ‘marriage bonus’ as to whether it is purely a question of a selection effect, or whether there is indeed a bonus from being married. On the other hand, there are some studies which suggest that we could turn the adage around and argue that behind a miserable man there is nobody – hinting that divorce may have devastating consequences at least for some men; those men who never manage to remarry after a divorce might be in particular danger. This speaks in favour of gendered economic and social consequences of divorce. In her study, Waite (1995: 498) showed that in America, marriage improves women’s economic well-being, while men benefit more socially than economically from their marriages. Waite explained her findings using three factors that may be of interest for the subsequent study of divorce in Denmark. First, family forms a kind of micro insurance pool that evens out various hazards and risks in life. Second, married couples benefit greatly from economies of scale. Third, marriage connects people to various social networks and by doing so, provides various forms of social capital to cope with life. This was precisely what Peter Abrahamsson (1988) found in his study of Danish urban poverty. Although in general, women were losers in monetary terms in the case of a divorce, many men were hit even more seriously, both economically and socially. Usually the latter form of exclusion from society was much stronger and more severe among single men. One problem that makes it a bit difficult to compare results from previous studies is that the concept of income is operationalized differently from study to study. Therefore, we do not know exactly how commensurable the results are from the different studies referred to above. To shed some light on the impact of the operationalization of income we also study to what extent our results are affected by the use of equivalence scales vis-à-vis non-equivalent income. Thus, the information presented above indicates that the issue of the economic consequences of divorce is perhaps not as established as had been indicated in many previous US-based studies. Our hypothesis is that there are important within-gender differences depending on the duration of separation. This is one aspect that we focus on in the empirical part of our study. Secondly, given the generous welfare state provisions in Denmark, we could expect that the harmful effects for separated women would not be as severe as they are in the USA, for example, as described above. One can suppose that in Denmark, welfare state arrangements – perhaps the most

Economic and social consequences of divorce in Denmark

101

generous in the world, including well-developed public childcare facilities, active and flexible labour market programmes and income transfers for families – should provide a buffer so robust that the harmful economic effects of divorce are more or less minimized. The Danish case is a kind of benchmark in shedding light on the ability of the welfare state to mitigate income loss in the case of divorce. Previous studies have convincingly demonstrated the capacity of the Scandinavian welfare state to reduce poverty caused by traditional social risks, whereas its capacity to cope with the socalled new risks is much less certain. Given the effect of economies of scale and the risk-pooling possibilities that the dual breadwinner family model offers, we could still expect there to be some detrimental economic effects even in generous welfare states like Denmark. It can also be discussed to what extent, if any, the welfare state can – and should – counterbalance various forms of social exclusion in the context of divorce. On the basis of a cursory inspection of previous studies, including Danish ones, we can construct a number of research hypotheses which partially contradict one another: ●

●

● ● ●

As a result of the generosity of the Danish welfare state, the poverty impacts of divorce – a ‘new social risk’ – are negligible compared to ‘old social risks’. Due to economies of scale, even the generous Nordic welfare state cannot combat the detrimental economic consequences of marital dissolution. Economic consequences of divorce (if there are any) are most severe for women. Males, especially those who never remarry, are hit the hardest socially, which consequently leads to economic problems as well. There is a gender-based difference between females and males who never remarry: women are better off than men.

The structure of this chapter is as follows: in order to place the discussion of the economic consequences of marital dissolution in a welfare state context, we first give a short overview of historical developments in Danish divorce legislation and institutional aspects of the Danish welfare state that we can assume to be linked to the mitigation of the effects of divorce. Thereafter, we describe our quantitative data and methodological choices. The core of this chapter comprises longitudinal analyses (over the course of a decade) of people who have never divorced, people who have remarried and people who remained single throughout the entire period from 1990 to 1999. We first present the overall income development of various marital groups over a period of ten years, after which we will look at social

102

Single country studies

assistance recipients and poverty rates among those groups. Although social assistance is not directly linked to the economic consequences of divorce, it may have important indirect ramifications – for example, if the divorced individuals lose a part of their social network and social capital, and consequently end up in a situation where their incomes decrease. In order to provide a snapshot of various social aspects of life among the divorced and non-divorced, we will utilize survey data to cross-check the results obtained from our multivariate longitudinal analyses, which do not say very much about subjective feelings. The final section is a short summary and discusses the findings on a more general level.

DANISH DIVORCE LEGISLATION, WELFARE STATE AND FAMILY STRUCTURE Divorce Legislation Up until the Middle Ages, the church insisted that marriage is a holy sacrament and thus indissoluble. After the Reformation, this interpretation began to change, and marriage was no longer considered irreversible. The changing attitude was mirrored in the Danish 1582 marriage law – inspired by the Protestant German practice – stipulating that divorce can be granted on the basis of adultery, impotence or the disappearance of the partner (Jensen, 2003: 132; Melby et al., 2006: 91–4). Gradually marriage came to be seen as a practical contract between two adults and, consequently, divorce practice and legislation were liberalized further. In the 1700s, divorce could take two forms: the couple could ask the king to accept a private agreement of separation and a division of their assets (bodeling), or the couple could ask the local authorities to give them permission to separate, and after a separation of three years a divorce could be granted (Wegener, 1977: 19). The varying practices were codified in the Royal Decree of 1796, which was a step towards acceptance of the nofault principle (Jensen, 2003: 14–142; Melby et al., 2006: 91–3). The 1922 Marriage Act stipulated separate regulations for divorce and separation. Divorce could be granted on the basis of family violence, unfaithfulness, venereal disease, mental illness or imprisonment. The Act also backed up the practice of allowing separation if both of the partners agreed that it was impossible to continue in cohabitation (Wegener, 1977: 19). The 1969 Marriage Act adopted the basic principle of the 1922 Act, but the basis for separation was liberalized, and it was no longer necessary to give any specific reasons for not being able to continue in a marriage. Furthermore, separation was possible if only one of the spouses wanted

Economic and social consequences of divorce in Denmark

103

to separate and the court considered the relationship between the spouses to be ruined. After a separation period of one year, divorce was possible (Koch-Nielsen et al., 1976: 11). According to the Act, a ‘spousal allowance’ (ægtefællebidrag) was payable to the spouse who had a lower or no income (in most cases, the wife). Usually spouses in short-lived marriages (up to two years) were not entitled to the allowance; marriages that had lasted two to 15 years yielded an allowance for two to ten years, and long-lasting marriages (more than 15 years) could lead to an unlimited entitlement providing that the wife was ‘considerably worse-off ’ (Wegener, 1977: 98–103). The size of the ægtefællebidrag could be agreed upon by the partners or, in the case of a dispute, stipulated by the court (Budzt, 1970: 59–62). However, the court could not set the spousal allowance to be more than one-quarter of the difference between the husband’s gross income (minus alimony paid to children) and the wife’s income (Wegener, 1977: 104–5). Alimony for children was, in principle, negotiated between the partners, but again the court could make a decision on the amount each spouse was supposed to contribute. Since 1933, a construction called ‘normal allowance’ (normalbidrag) – set annually by the authorities to cover costs for children – has been the basis for the adjustment of alimony. The father contributes three-fifths and the mother two-fifths of the normalbidrag. The allowance is precisely the same in all parts of the country; however, it can be increased (never by more than 400 per cent) depending on the partners’ income levels and the number of children (Wegener, 1977: 172–3). The most recent adaptation of Danish divorce legislation took place in 1989. According to the 1989 Act, divorce can be filed for after a separation period of half a year if both partners agree to divorce. Divorce without separation is possible in cases of unfaithfulness, family violence or bigamy – the first reason being the most important cause for immediate divorce (Nielsen and Rasmussen, 2001: 245). The new law also changed the role of the ægtefællebidrag so that it corresponded to increased female labour force participation. Although the issue of spousal support is always addressed, it no longer has the same importance as it did previously, and the maximum duration of the spousal allowance is set at ten years; only in very specific circumstances is it extended beyond that limit (Nielsen and Rasmussen, 2001: 267–77). Consequently, the amount was set at one-fifth of the difference between the gross income (minus alimony for children) of the paying partner and that of the partner with the lower income. The normalbidrag practice is still in use, and in normal cases the allowance is payable for children up to 18 years of age. The spousal allowance and alimony for children must not exceed one-third of the claimant’s income unless the partners agree upon higher amounts (for example, Melby et al., 2006: 91–3).

104

Single country studies

Family Structure and Divorce Patterns Needless to say, Denmark is facing the same demographic challenges as other industrialized Western countries, yet in contrast to many other countries, Danish fertility rates have been rather high and even increased during the past few decades. In 1985, the fertility rate was 1.45 children per woman, and it had increased to 1.78 in 2004 (see Data Appendix). When it comes to marriage patterns and divorce rates, Denmark’s trajectories are similar to those of many other countries. Divorce rates – which had been decreasing since World War II – shot up dramatically in the 1970s, and since then there has been a fluctuating upward trend. As can be seen in Figure 4.1, there is a general trend that the younger ‘marriage generations’ are more prone to separation than previous generations. Out of the marriages that were initiated in the 1950s and 1960s, about 10 per cent lasted less than ten years, while about 30 per cent of the marriages that began in the 1980s and 1990s ended in dissolution before their tenth anniversary. When it comes to the labour market position of single parents, the picture has been rather constant for the last few decades. The major changes which have taken place are the disappearance of homemakers and the increase in unemployment among single parents. There is also a slight tendency for those single parents who do work to work on a full-time basis than was previously the case (Nygaard Christoffersen, 2004: 145). On a more general level, the Danish case is a good example of how the ‘male breadwinner’ model of the past has been transformed into a ‘dual-earner’ model (Gornick and Meyers, 2003; Korpi, 2000; Leira, 2002; Lewis, 1992, 1993; OECD, 1998; Rostgaard, 2004). The childcare enrolment ratio for 0 to 3-year-olds was 48 per cent in 1990 and as high as 64 per cent in 2000 (see Data Appendix). As many as 91 per cent of all children in the 3- to 6-year-old age bracket were participating in publicly administered day care in 2000. As a consequence, the female labour force participation rate – which, in the 1960s, was not especially high (about 45 per cent) by Organisation for Economic Co-operation and Development (OECD) standards – increased rapidly during the two consecutive decades: since the late 1980s, it has hovered around 70 per cent, which is one of the highest female activity rates in the OECD region (OECD, 1999: 41; see Data Appendix). However, the flip side of the coin is that the so-called gender employment gap (percentage point difference between employment rates for men and women) among persons aged 25 to 54 still plays a role even in a dual-earner country such as Denmark. The gap between men and women without children is 8 percentage points – 12 percentage points if there are children. In other countries that gap is greater – especially in Greece, Spain and Italy, where it hovers around 30 percentage points

105

0

5

1990

10

15

20

1980

25

30

1970

35

1950

40 45 50 Duration of the marriage, years

1960

Percentage of divorces by marriage duration for marriages initiated in 1950, 1960, 1970, 1980, and 1990

Nygaard Christoffersen (2004).

Figure 4.1

Source:

0

5

10

15

20

25

30

35

x

Percentage of divorced 40

x

106

Single country studies

(Kangas and Palme, 2005: 37). In the other non-Nordic OECD countries, the gap is 15 to 20 percentage points (Kangas and Palme, 2005: 37; for a more detailed analysis, see Datta Gupta et al., 2006). There are also some gender-specific differences in employment patterns (see Data Appendix). As in most other countries, there is a bias towards part-time employment among women, although this gap has diminished during recent decades: in 1985, about 8 per cent of all men and as many as 44 per cent of all women had part-time labour contracts. In 2005, the corresponding figures were 13 per cent and 33 per cent (see Data Appendix). Danish Family Policy, the Danish Welfare State In comparison to the other Nordic countries – not to speak of countries with less generous social protection – family policy transfers are generous in Denmark. In 1985, Denmark spent 2.6 per cent of its gross domestic product (GDP) on families. Twenty years later, the figure was up to 3.9 per cent (see Data Appendix). About 60 per cent of that spending is devoted to social services (mainly public day care) and the rest to social transfers (mainly child allowances). By 2000, income-substituting cash benefits in the case of childbirth were payable for 31 weeks (which, in the Nordic context, was not such a long period); since 2002, the parental leave period has been one year at the rate of 100 per cent of the previous income (Olsen, 2005). However, there is a benefit ceiling which cuts the replacement level in the higher income brackets. Despite the ceiling effect, the claimant’s disposable income while receiving maternity benefits exceeds 100 per cent of her previous earnings up to the middle-income groups (NOSOSCO, 2002: 42–7). This is due in large part to the rather high child allowance, which in 2000 was about €1400 per year for one child, €2800 for two children and €4200 for three children. The corresponding amounts for single parents were €2400, €4300 and €6200, respectively. Thus, benefits for single parents were 1.5 to 1.7 times higher than benefits for couples. At the average income level, the net compensation rate of child allowances paid for two children was 11 per cent for a two-parent family (with the husband receiving the average wage in industry – APW – and the wife receiving 67 per cent of the APW), whereas the compensation rate for a one-parent household (income 65 per cent of the APW) was as high as 43 per cent in 2000 (OECD, 2001: 170–71). In Denmark, as in the other Nordic countries, in cases where the parents are separated, the parent not living with the child pays a maintenance allowance for any child under 18 years of age. The amount of the maintenance allowance is set as part of the legal divorce process (as described above). The allowance is fixed either by a court decision, according to a

Economic and social consequences of divorce in Denmark

107

consensual agreement by the parents, or by local authorities. The public authorities will always set a minimum amount of allowance that must be paid. In cases where the liable partner is unable to pay the allowance (at present, about 14 per cent of children below the age of 18 receive the maintenance allowance), public authorities are responsible for guaranteeing the benefit. In 2000, the maximum municipal amount paid was DKK10 644 (€1440) a year (NOSOSCO, 2002: 55). As these numbers indicate, there is a strong built-in equalization mechanism that attempts to combat the economic hardships caused by divorce and single parenthood. But the question is, how much do these efforts manage to mitigate the harmful economic effects of divorce? As stated in the introduction, families may play a role as a micro-insurance pool by sharing their income and reducing the poverty risks to which singles and single-earner households are more directly exposed. In this equalization process, it has been important that women also participate in the labour market and bring their earnings into the family income pool (Fritzell and Ritakallio, 2004). In the Danish case, this is very much what has happened during the past few decades. Since the 1970s, the expansion of publicly subsidized day care made it possible for an increasing number of fathers and mothers to combine labour market work with caregiving responsibilities for children (Ferrarini, 2006; Palme, 1999).

DATA AND METHODS We evaluate our hypotheses utilizing the possibilities given by the longitudinal data that we have access to. The database is a 1 per cent representative sample of the Danish population aged 30 to 48 years in 1990, derived from Statistics Denmark. We applied these age limits in order to reduce the effect of students, and to avoid the retirement decision affecting our results. The vast majority of Danish students have completed their education by age 30. Although the official Danish retirement age was (and is now again) 67 years, the so-called post-employment wage (efterløn) scheme already starts at age 60 for eligible persons. Therefore, the oldest person in the sample was 48 years of age in 1990. In total, we had 15 074 individuals in our sample in 1990 (see Table 4A.1 in the appendix to this chapter). By following these individuals over a period of ten years (1990–99), we were able to look at changes in marital status and at the consequences of these changes. Unfortunately, information for some individuals is missing for some years (technically speaking, our data-set is an unbalanced panel). But as can be seen from the right-hand side of Table 4A.1, most of the sample

108

Single country studies

members (n 5 14 181) provided information for all ten years of the observation period. The pooled data-set includes 147 790 observations (5 sum of all [yearly] measurements for all sample members). Given these data, we can compare each individual’s marital status between two consecutive years. We analysed a total of 131 907 transitions: 71 per cent of the transitions consisted of individuals living as couples in both years; 20 per cent were living without a partner in both years, while 8 per cent changed marital status during the one-year period. In addition to marital status, we registered information on individuals’ income, income composition, labour market status and various demographic attributes. Our concept of income includes all family income that consists of market income (that is, wages, income from capital and self-employment, and so on), all public transfers (unemployment, sickness, social assistance, housing and child allowances, and so on), and all private transfers (alimony or allowances from the ex-spouse). To cope with the impact of inflation, all incomes have been deflated into 1990 monetary value using the Consumer Price Index. Furthermore, in order to make incomes for families of different sizes and compositions comparable, we calculated equivalent incomes by dividing the family’s disposable income by the square root of the number of family members. For example, if there is a family of two adults and two children, the equivalence scale is 2 (square root of 4).1 All family members were presumed to receive the same equivalent income. Of course, this supposition is more or less unrealistic, but since we do not know how family income is divided up between family members, we must try to cope with this proxy measure and comfort ourselves with the fact that nearly all other studies suffer from the same problem. Equivalent family incomes were calculated at the family level and then appended to each individual’s data record. In addition to ‘equivalent family income’ we also used ‘per capita family income’, which is equal to the disposable family income divided by the number of family members. Comparing the results based on the two income concepts discloses the importance of the equivalence scale. Henceforth, the unit of analysis is the individual. In addition to changes in income levels, we looked at the recipience of social assistance and at poverty rates, and present an example of a ‘typical divorce case’ in order to highlight the impact of various components upon changes in income levels. An additional way to look at the issue is to utilize surveys and try to take a more subjective approach by putting some flesh on the quantitative bones. In the final empirical section we will use data from the first wave of the European Social Survey (ESS, 2007) to see to what extent subjective experiences of economic hardships and social exclusion/inclusion vary between divorced and non-divorced Danes.

Economic and social consequences of divorce in Denmark

109

The applied multivariate analysis (Tables 4.4 and 4.5) models (the natural log of) equivalent income. One important control is marital status (single or not), whose parameter is a measure of the separation effect. The other controls include gender, age, immigrant status (1 5 yes, 0 5 no; without distinguishing between Western and non-Western immigrants or between different immigrant generations), number of children, university education (1 5 at least a master’s degree from university, 0 5 other), labour market experience (in years), full-time employment (1 5 yes, 0 5 no), and percentage of the year spent in unemployment. The models are estimated with both ordinary least squares as well as fixed effects.

RESULTS The immediate economic effects of leaving a relationship and forming a new family without a partner are shown in Table 4.1. Individuals who stay in couples have the highest equivalent family incomes (DKK142 000). When they become single, these individuals’ income levels drop to DKK100 000, which is a reduction of 30 per cent. As we can see, the drop is more severe for women (about 40 per cent) than for men (about 20 per cent). The results are the same regardless of whether we use mean or median equivalent family income. This drop in income takes place mainly because economies of scale cannot be utilized when people live alone. The conclusion becomes clear if we compare results from analyses where we used equivalent family income with the results from analyses utilizing per capita family income. In fact, under the latter assumption, mean per capita family income increases from DKK88 000 when individuals stay in a couple to DKK93 000 if a separation takes place. Again, we find a clear gender-specific pattern. For males, the per capita income level increases by 24–30 per cent while it decreases by about 15 per cent for females. This discrepancy between per capita and equivalent income can be one reason for the huge gender-specific differences in results obtained in previous research on the consequences of divorce, as discussed earlier. Although equivalent income decreases severely in the case of marital dissolution, the income of new singles is not that different from that of other singles. Individuals who remain single from one year to the next have an average income of DKK106 000, which is only 6 per cent higher than for those who are single due to separation (see the transition from couples to singles). Thus, the economic effects of divorce not only depend very much on the economies of scale, but also on whom the new singles are compared with. When compared with couples, their income loss is high, yet it is not that noteworthy when compared to that of other single people. When it

Single country studies

110

Table 4.1

Income by marital transition between t and t 1 1 (DKK1000 1990–1999) Couple 2 Couple 2 Single 2 Single 2 couple single single couple

All

Change from C-C to C-S DKK1000 %

Mean Equivalent income Male Female All Per capita income Male Female All Equivalent income Male Female All Per capita income Male Female All n %

141 144 142

109 89 100

110 100 106

132 136 134

130 131 131

231 –56 –43

–22 –39 –30

86 90 88

107 77 93

109 91 100 Median

82 84 83

90 88 89

21 –13 5

24 –15 5

130 134 132

101 84 93

102 96 99

123 127 125

121 122 122

–29 –50 –39

–22 –37 –30

77 82 79 94 274 71

100 71 88 5371 4

100 89 95 26 939 20

73 76 74 5323 4

82 81 81 131 907 100

23 –11 9

30 –14 11

Note: All individuals aged 30–48 years in 1990 are included. They were followed from 1990 to 1999. 15 883 transitions where no information was available for the initial year have been excluded. Source:

Statistics Denmark, 1% sample of the Danish population aged 30–48 in 1990.

comes to the income levels of singles who (re)marry, we can see that on average, the change of status from single to a couple causes an increase in equivalent income of 26 per cent (from DKK106 000 to DKK134 000). The effects of marital dissolution for men and women are qualitatively the same, but with very different magnitudes, as indicated. Furthermore, divorced women have an 11 per cent lower equivalent income than other single women, whereas divorced men have about the same income level as other single men. An explanation for this may be that women, more often than men, live with their children. This question is assessed in Table 4.2. In the case of couples without children, women’s and men’s equivalent

Economic and social consequences of divorce in Denmark

111

Table 4.2 Equivalized household income by marital transition and children living at home (DKK1000 1990–99) Couple Couple Single Single All – couple – single – single – couple

Without children Male Female With children Male Female Source:

Change from C-C to C-S DKK1000

%

164 161

110 98

111 107

153 154

137 143

–53 –63

–33 –39

125 129

95 77

99 86

117 121

122 119

–30 –52

–24 –40

Statistics Denmark, 1% sample of the Danish population aged 30–48 in 1990.

incomes are reduced almost equally (39 per cent and 33 per cent, respectively). But the gender gap is quite apparent if we look at families with children. Here, the income losses for men and women are 24 per cent and 40 per cent, respectively. This rather significant effect of divorce is again less dramatic if we compare the income levels of recently divorced persons to those of other singles. This is especially the case for men with or without children. Divorced men without children have an equivalent income amounting to DKK110 000, and men with children have DKK95 000, which is quite close to that of other single men (DKK111 000 and DKK99 000 respectively). Divorced women with children have DKK77 000 at their disposal, while other single women with children have DKK86 000. The main conclusion we can draw from these descriptive analyses is that divorce leads to much lower consumption possibilities compared to the previous couple phase of life, but compared to the situation of other singles, the effect is much smaller. Economically, women suffer from divorce more than men do. Tables 4.1 and 4.2 demonstrate that the use of equivalence scales has a major impact on our results when trying to understand the severity of economic losses in the case of couple dissolution. In order to visualize how the welfare state, private transfers, equivalence scales and other factors contribute to the situation outlined above, we have formulated an example that is based on the average Danish male and female income, and on the income transfers that occur in the case of divorce. This typical case sheds some additional light on the role that economies of scale play in our calculations. Table 4.3 makes the hidden factors and assumptions that affect the results more visible. First, it is assumed that incomes are equally

Single country studies

112

Table 4.3

Example based on Danish average production worker family with two children (DKK, 2000) Before divorce

Male, net wage Female, net wage Child allowance 1 Child allowance 2 Private transfers Single parent allowance Total disposable income Equivalence weight Equivalent net income Reduction in DKK Reduction in % Note:

After divorce

Couple

Single male

Single female

180 000 144 000 9 500 9 500 0 0 343 000 2.00 171 500

180 000 0 0 0 –24 000 0 156 000 1.00 156 000 –15 500 –9

0 144 000 9 500 9 500 24 000 12 600 199 600 1.73 115 239 –56 261 –33

Taxing and deduction effects of child allowance are included.

distributed between different members of the household. This assumption has been questioned by many scholars, who consider an equal division to be totally unrealistic (for a more in-depth discussion, see, for example, Gornick and Meyers, 2003; Lewis, 1992, 1993). There are studies that indicate that the husband’s share of the family income is much higher than the wife’s share, and that the shares may vary between socio-economic groups, income brackets, and the degree of female labour market participation. Therefore, the equivalent incomes displayed in the first column are certainly somewhat biased. Consequently, the degree of income reduction would be less severe for females and not as small for males. Keeping these reservations in mind, we can nevertheless take a look at the two other columns which clarify the income dynamics that take place in the context of a divorce. As a result of private transfers (alimony for two children) the ex-husband loses DKK24 000 from his wallet and, therefore, his disposable income becomes smaller than his equivalent disposable income was during marriage. Thus, in absolute terms, he loses out. The reverse situation is true for the ex-wife: in addition to her own income, she receives child allowances for her two children, and an additional single-parent allowance is payable to her. Furthermore, she receives alimony from her ex-husband. Hence, her equivalent family income increases. However, taking into consideration the impact of family composition and the costs of child-rearing, her

Economic and social consequences of divorce in Denmark

113

equivalent income decreases in comparison to her status during marriage. Despite all the mitigating effects of the generous Danish welfare state, divorce will reduce the wife’s income by as much as one-third. The welfare state cannot fully compensate for the micro-pooling effect and economies of scale experienced in the dual-earner household. Determinants of the Consequences of Divorce The bivariate cross-tabulations presented in the previous section do not say much about exactly what causes the gender differences and what will happen if we control for background variables. In order to quantify the effects of various intervening factors, we applied multivariate regression techniques to model impacts from intervening variables upon disposable income. In order to obtain more homoscedastic residuals and to be able to interpret the estimated coefficients as relative percentage changes, the response variable is the natural log of income. To pinpoint the impact of equivalence scales, we used equivalent family income and per capita family income as dependent variables. (Summary statistics are displayed in Table 4A.2 in the appendix to this chapter.) To evaluate the sensitivity of our results, we ran both OLS and fixed effect regressions, but since the results were more or less the same, for space considerations, we have only presented the results from the OLS runs. As a consequence of the large sample sizes, almost all the parameters in Table 4.4 are significant at the 0.01 per cent significance level and furthermore, all the parameters have the signs we expected. The overall picture on the economic consequences of marital dissolution depicted in Table 4.4 very much resembles the results reported in the descriptive section above. Men lose 31 per cent (see the transition from couple in year t21 to single in year t, couple → single for males) of their equivalent income after marital dissolution, whereas the loss for women is 48 per cent (see the transition from couple in year t21 to single in year t, couple → single for females). (In the fixed effect models, men lose 28 per cent and women lose 47 per cent.) Thus, after correcting for observed as well as unobserved fixed characteristics of individuals, a huge loss in disposable income remains – and it is actually higher than in the descriptive analysis as suggested in Table 4.2. When we assume that there are no economies of scale – that is, when we use per capita family income – men gain from marital dissolution while women lose out. But when we control for intervening variables, the gender gap is reduced from 39 percentage points (Table 4.1) to 25 percentage points (Table 4.4). This time, men’s gains are ‘only’ 4 per cent – while Table 4.1 states 24 per cent – and women’s losses are 21 per cent, which is higher than in Table 4.1 (15 per cent). Thus, in the case where there are no

Single country studies

114

Table 4.4

Economic consequences of divorce (OLS regression coefficients) ln(equivalized household income)

Intercept Female Age (years) Age squared/100 Immigrant Couple → single Single → single Single → couple Number of children University education Labour market experience (years) Full-time employed Unemployment (%) 1992 1993 1994 1995 1996 1997 1998 1999 R2 Sample size

ln (per capita family income)

Males

Females

All

Males

Females

All

3.972

3.754

3.711

3.549

0.029 –0.028 –0.096 –0.309 –0.322 0.007 –0.122

0.040 –0.039 –0.032 –0.480 –0.404 0.002 –0.095

3.920 0.036 0.032 –0.032 –0.070 –0.392 –0.362 0.001 –0.114

0.024 –0.022 –0.090 0.036 0.029 0.007 –0.281

0.032 –0.029 –0.021 –0.212 –0.117 0.002 –0.267

3.684 0.021 0.027 –0.025 –0.062 –0.088 –0.046 0.000 –0.283

0.303

0.331

0.313

0.301

0.334

0.313

0.009

0.016

0.012

0.009

0.016

0.012

0.173

0.108

0.132

0.171

0.104

0.127

–0.003

–0.001

–0.002

–0.003

–0.001

–0.002

0.012 0.017 0.063 0.078 0.084 0.085 0.104 0.105

–0.017 –0.018 0.013 0.027 0.031 0.033 0.046 0.042

–0.003 –0.001 0.038 0.052 0.057 0.059 0.075 0.073

0.012 0.018 0.064 0.079 0.087 0.087 0.106 0.107

–0.020 –0.020 0.011 0.024 0.029 0.031 0.043 0.039

–0.005 –0.002 0.037 0.052 0.058 0.059 0.075 0.074

0.29 66 969

0.30 64 938

0.29 131 907

0.43 66 969

0.42 64 938

0.42 131 907

Note: All parameters, except single to couple transition and some year parameters, are significant at least at the 1% level and often also at the 0.01% level. Source:

Statistics Denmark, 1% sample of the Danish population aged 30–48 in 1990.

economies of scale, men’s gain from marital dissolution almost vanishes when we control for background factors. In Table 4.4, we concentrated on the short-term effects of marital dissolution (the first year after the divorce). But what are the effects of staying

Economic and social consequences of divorce in Denmark

115

single, with or without children, for a longer time? In order to study the long-term effects of divorce in more detail, we limited our data-set to only those persons who were initially married in 1990 and excluded those who were initially single. The ‘old’ singles had to be excluded because we did not know for how long they had been living alone. This limitation left us with 109 380 observations – 53 630 for men and 55 750 for women. The number of years that the divorced persons lived alone was recorded up until 1999, which yields a maximum ‘single’ duration of nine years (from 1991 to 1999). Individuals who divorced after 1991 were also included, but their maximum duration was, of course, less than nine years. Again due to space limitations, we analysed equivalent incomes only, and included only the OLS estimates, as the fixed effect estimates essentially tell the same story. Again, almost all the parameters are very significant and have the expected signs. Excluding the ‘old’ singles does not change parameter estimates of divorce very much when we compare Table 4.5 with Table 4.4. Thus, in their first year as divorcees, males lose 30 per cent of their income compared to when they were living with a partner (31 per cent in Table 4.4, where the larger sample included the initially single) and women lose 51 per cent (48 per cent in Table 4.4). For men, the duration parameters are insignificant, except in the case of divorcees in the eighth year, who have significantly lower incomes than couples. The development for divorced women is very different, as they continuously experience significant income growth as divorcees (about 8.5–13.5 per cent; insignificant growth in the eighth year). We also carried out a formal test for the significance of income difference between couples and divorcees from the second to the ninth year. For males, the difference was very significant, with p 5 0.02 per cent, but when applying heteroscedasticity-robust standard errors, the null hypothesis cannot be rejected at the 5 per cent significance level (p 5 7.51 per cent). For females, however, the robust test is clearly rejected (p , 0.01 per cent). Thus, although women seem to be hit harder by divorce initially, they appear to be able to cope with their lives as divorcees better than men do. The argument gains qualitative support if we look at a number of background factors presented in Table 4.6. An obvious coping strategy is to be active in the labour market. The initial level of unemployment was about 10 per cent for both genders in the first year as a divorcee. This number fell to only 3 per cent for women, yet it was still close to 10 per cent for men in the eighth and ninth years. Thus, those men who never remarry seem to suffer more from labour market exclusion than the corresponding group of women. This interpretation gets some qualified support if we look at the experience of unemployment in different groups of divorced persons as

116

0.0358 –0.0353 –0.0947 –0.2992 0.0243 0.0241 0.0112 0.0105 –0.0382 –0.0121 –0.1054 –0.0687 –0.1249 0.2940

0.0027 0.0031 0.0066 0.0091 0.0138* 0.0153* 0.0171* 0.0191* 0.0217* 0.0249* 0.0292 0.0420* 0.0015 0.0051

0.0465 –0.0465 –0.0298 –0.5061 0.1025 0.1216 0.1166 0.1070 0.1053 0.0843 0.0515 0.1355 –0.0968 0.3337

Estimate

Estimate

Std error

Females

Males

0.0028 0.0031 0.0072 0.0096 0.0146 0.0158 0.0173 0.0192 0.0218 0.0248 0.0305* 0.0463 0.0017 0.0074

Std error

0.0448 0.0389 –0.0392 –0.0637 –0.4012 0.0636 0.0714 0.0612 0.0557 0.0326 0.0317 –0.0270 0.0302 –0.1145 0.3103

Estimate

All

0.0020 0.0019 0.0022 0.0049 0.0066 0.0101 0.0111 0.0122 0.0136 0.0155 0.0177* 0.0213* 0.0314* 0.0011 0.0043

Std error

Long-term income effects of marital dissolution (logarithm of equivalized household income, OLS regression coefficients)

Female Age (years) Age square/100 Immigrant Divorcee Divorcee for 2 years Divorcee for 3 years Divorcee for 4 years Divorcee for 5 years Divorcee for 6 years Divorcee for 7 years Divorcee for 8 years Divorcee for 9 years No. of children University education

Table 4.5

117

0.0081 0.1057 –0.0028 0.33 53 630

0.0004 0.0047 0.0001

0.0153 0.0821 –0.0011 0.32 55 750

0.0004 0.0037 0.0001

0.0121 0.0889 –0.0018 0.32 109 380

0.0003 0.0029 0.0001

Source:

Statistics Denmark, 1% sample of the Danish population aged 30–48 in 1990.

Notes: Intercept and year dummies are not shown. All parameters are mostly significant at the 0.01% level, and some at the 5% level. Parameters with an * are not significant at the 5% level.

R2 Sample size

Labour market experience (years) Full-time employed Unemployment (%)

118

43.5 5.6 0.2 5.4 10.9 11.0 78.0 2.3 3.9 4.7 10.6 11.5 5.4 118.8 43.8 5.3 0.8 3.3 11.1

7.5 116.6 43.1 4.9 0.8 3.2 10.0

2

42.7 6.0 0.2 6.0 9.9 10.7 78.1 2.1 5.2 4.0 9.1 9.3

1

44.5 5.4 0.8 3.4 9.6

117.2

2.8

44.2 6.1 0.2 5.0 11.3 11.2 77.3 2.4 2.7 5.2 8.9 9.8

3

45.4 5.9 0.7 3.9 9.2

117.2

4.3

45.1 6.5 0.1 4.4 11.6 11.4 75.5 2.5 3.0 6.1 11.1 11.9

4

45.8 5.9 0.7 2.9 7.8

120.1

3.7

45.6 6.2 0.1 4.9 10.1 11.6 73.5 2.8 3.7 7.2 11.1 11.8

5

46.4 6.3 0.7 2.2 6.9

115.5

0.0

46.3 7.2 0.1 5.1 8.8 11.6 68.2 2.1 5.9 7.2 11.9 12.8

6

7

47.1 5.1 0.6 1.5 5.8

120.3

0.0

46.7 6.4 0.1 5.8 8.9 11.8 66.1 4.7 4.1 7.3 12.9 14.0

Number of years continuously single

12.5

47.7 5.8 0.1 6.7 7.6 11.4 60.0 2.5 7.5 8.0 12.5 12.5

8

47.3 4.8 0.5 2.4 3.4

115.1

Background factors and number of years of being single (percentages and arithmetic means)

Males Age (years) Immigrant (%) No. of children University education (%) Unemployment (%) Labour market experience (years) Full-time employed (%) Part-time employed (%) Poverty risk (%) Social assistance (DKK1000) Recipient (%) of social assistance, all Recipient (%) of social assistance, without children Recipient (%) of social assistance, with children Income (DKK1000) Females Age (years) Immigrant (%) No. of children University education (%) Unemployment (%)

Table 4.6

47.4 3.9 0.4 0.0 2.4

122.5

0.0

48.1 7.3 0.1 7.3 9.5 11.9 65.5 3.6 3.6 6.6 12.7 14.0

9*

119

Source:

Note:

9.3 69.8 7.2 4.0 4.8 10.1 6.7 13.5 95.4

8.8 69.3 7.5 9.7 4.4 10.6 6.6 14.6 86.7

99.4

12.2

9.7 70.5 6.2 4.0 4.8 9.6 7.2

104.0

10.0

10.1 70.8 4.7 3.7 4.9 8.3 6.9

Statistics Denmark, 1% sample of the Danish population aged 30–48 in 1990.

* Less than 100 observations.

Labour market experience (years) Full-time employed (%) Part-time employed (%) Poverty risk (%) Social assistance (DKK1000) Recipient (%) of social assistance, all Recipient (%) of social assistance, without children Recipient (%) of social assistance, with children Income (DKK1000) 102.7

10.6

10.1 69.7 5.6 3.2 4.7 8.0 5.8

101.0

8.5

10.5 68.5 4.1 1.1 5.4 8.1 7.9

101.8

11.7

10.7 70.2 3.0 2.0 5.2 8.1 5.8

103.1

6.5

10.4 66.1 4.0 2.4 5.1 6.5 6.4

102.8

7.1

9.7 56.9 9.8 2.0 8.0 11.8 13.5

120

Single country studies

well (not presented in the tables). Among divorced men without children, unemployment went up in the first year after divorce, whereas the unemployment rates for divorced men with children went down. At the end of the period under inspection, the men who took custody of their children after divorce and remarried clearly did the best of all the men in terms of employment. The interplay between marital dissolution and unemployment was different among women: unemployment among women with children went up in both situations, whereas it went down for those women who had no children and who never remarried. An explanation for these findings might be related to the various gendered impacts of different welfare state programmes. First, the rapid increase in unemployment among single mothers (but not among single fathers) may be related to the incentive structures within the Danish welfare state and differences in labour market experiences between genders. Single fathers are, on average, the most privileged group in terms of their educational attainment, income and labour market position; whereas single mothers seem to have greater problems in immediately adapting themselves to working life – either due to the availability of suitable employment or to the problems of combining family and working life, or due to the combined effect of both of these factors. Furthermore, their income levels might be low enough that the combination of unemployment compensation and other welfare state provisions causes work incentive problems. Whatever the explanation, in 1999, the overall unemployment level among divorced women with children was still higher than among females without children. Secondly, the increase in unemployment among divorced males without children may result from the fact that divorced men without children suffer from a substantial loss in social contact – or social capital if you prefer – and consequently, they are hit by unemployment more severely than divorced men with children. A conventional way of looking at the severity of exclusion from mainstream society is to look at poverty rates in different population categories. In our study, we used the 60 per cent poverty line and defined individuals as poor if their equivalent income was less than 60 per cent of the median equivalent income for all Danes. By international standards, the poverty rate in Denmark is fairly low, reflecting the equal distribution of income in Danish society and the overall efficacy of the welfare state in combating traditional Rowntree-type poverty. However, once again, these averages hide a number of divergent stories. The immediate impact of divorce is most severe for women, whose poverty rate in the first year is about 10 per cent, while it is only about 5 per cent for men. Thereafter, the difference is negligible up until the sixth year as a single; then poverty turns out to be a problem for divorced males. There seems to

Economic and social consequences of divorce in Denmark

121

be a qualitative difference between men and women who never remarry. The former could be classified as ‘les misérables’, whereas the latter do much better. Poor men and rich women tend to have long-lasting divorce periods. We will return to this issue a little later when discussing the social aspects of divorce. A complementary way to evaluate the degree of social exclusion is to look at the recipiency of social assistance. One can argue that recipiency of social assistance is a kind of dual social exclusion measure. On the one hand, individuals themselves regard their income as too low and seek help; on the other hand, public authorities define the help-seekers as poor according to the commonly accepted poverty line. Therefore, recipiency of social assistance is an important indicator of perceived and experienced economic hardships in society, and it is also perhaps one of the strongest indicators of social exclusion – at least in the case of longer-lasting periods of recipiency (Kangas and Ritakallio, 1998). In many countries, welfare state dependency among single parents is a serious problem. As expected, divorce increases people’s propensity to rely on social assistance in Denmark as well. Among those who had never divorced, the recipiency rate was as low as 1.3 per cent. During the first three years there were no major gender differences, and the recipiency rate varied between 9 and 10 per cent. Beginning in the fourth year after divorce, the figures for women began to decrease, whereas there was an upward trend among men; by the eighth year, the recipiency rate for men (12.5 per cent) was almost double that for women (6.5 per cent) – supporting our interpretation of the qualitative gender difference in the long-term economic consequences of divorce. This hypothesis is further strengthened if we concentrate only on childless cases. On average, the likelihood of being on benefits is significantly higher among single men than among single women, and in all male categories, single men (no wife, no children) have the highest recipiency rates, whereas single women tend to have the lowest rates among all female categories. An inspection of the average amounts of social assistance tells the very same story about gender differences. Women received an average of DKK4400 per year in the first year after divorce, whereas men received DKK400 less, but in the eighth year, the amount for men doubled to DKK8000 while it only increased to DKK5100 for women. When it comes to single parents, fathers living with their children have the lowest recipiency rates, while the corresponding numbers are higher – although they decrease over time – for single mothers. This is related to the fact that men who obtain custody of their children after divorce tend to have higher levels of educational attainment and higher incomes than men who do not take their children (see also Nygaard Christoffersen, 1996). Thus, a selection effect obviously plays a role here.

122

Single country studies

Social Consequences of Divorce To complement our register-based analyses, we have utilized surveys to see to what extent subjective experiences vary between divorced and nondivorced Danes. We also present some cursory comparative information in order to place Denmark in a somewhat wider international perspective. We have applied four compound welfare variables often used to measure attachment to and possibilities for participating in society: social trust (the scale varies from 0 5 no trust to 30 5 the highest possible trust), happiness (0 5 extremely unhappy; 20 5 extremely happy), frequency of social contacts (2 5 lowest frequency; 12 5 the highest frequency), perceived sickness (2 5 good health; 10 5 bad health) and the experience of economic hardship (2 5 no economic problems; 8 5 big economic problems). A more detailed description of the composition of the variables is given in this chapter’s appendix. As a rule, Danes display higher mean values for trust (20.4 for Denmark versus 15.2 for the other European countries), happiness (16.8 versus 14.4), and social contacts (8.3 versus 7.7), lower values for sickness (3.4 versus 3.7) and much lower scores for economic hardship (2.8 vs 4.2) than respondents in the other countries. In principle, in all countries, divorce decreases the level of well-being and there are no substantial differences between Danes and other nationalities in this respect. In order to more closely evaluate the social impact of divorce, we ran multivariate analyses to control for background variables. We used the general linear modelling (GLM) procedure in the Statistical Package of the Social Sciences (SPSS). GLM also allows us to estimate and visualize interaction effects. In this particular case, we were interested in the interaction between gender and divorce, and in order to make a proper comparison, we divided the respondents into three groups: couples, singles who were not divorced and divorced individuals who were living single. Age and age squared as well as educational attainment were used as covariates. When analysing determinants of trust and happiness, we also used economic hardship and health status as explanatory variables; and in analysing economic hardship, health was included as a covariate. Correspondingly, in health analyses, we added the perceived economic status to the covariates. Due to space limitations, we have only plotted estimated mean values for each welfare component when all relevant background variables are controlled for (Figure 4.2). For all dimensions, there is a clear interaction effect, and the results support the analyses we presented above. Interestingly enough, when it comes to trust (social capital), single women seem to be better off than their married sisters, while divorced single males perform worst. Couples

Economic and social consequences of divorce in Denmark

22.00

Economic hardship 3.60 Estimated marginal means

Estimated marginal means

Trust Male Female

21.00

20.00

Male Female

3.40 3.20 3.00 2.80 2.60

19.00 Couple

Non- Divorced divorced single single

Couple

Happiness

Sickness Estimated marginal means

Estimated marginal means

17.10 16.80 16.50 16.20 15.90 15.60

Male Female Couple

Non- Divorced divorced single single

4.00

Male Female

3.80 3.60 3.40 3.20 Couple

Non- Divorced divorced single single

Non- Divorced divorced single single

Estimated marginal means

Social contacts 8.80

Male Female

8.60

8.40

8.20

Couple

Figure 4.2

Non- Divorced divorced single single

Estimated marginal means for couples, singles and divorced singles (2002)

123

124

Single country studies

enjoy the happiest life, and divorced men are again at the bottom – but the difference between them and other single males is not significant. As our panel study showed, divorced women – in particular single mothers – are most exposed to monetary poverty, which is also clearly visible in GLM plots. Finally, women in general are sicker than men, but the most sicknessstricken group seems to be divorced males. Needless to say, it is impossible to argue strongly for causality on the basis of cross-sectional data and, furthermore, cross-sectional data do not control for the time since divorce – that is, the divorced individuals may be a select group of people who have not yet found a new partner. However, the subjective survey results support our reasoning pertaining to the economic consequences for women and social consequences for men.

DISCUSSION Classical poverty studies concentrated on the so-called old social risks caused by childbirth, unemployment, sickness and old age. In this study, we used a longitudinal Danish database and followed the very same individuals from 1990 to 1999. Instead of looking at the old risks – which have been more or less effectively eliminated in Denmark thanks to its generous welfare state – we scrutinized various economic and social consequences of divorce. We had two competing hypotheses: in the first, it was supposed that on this issue as well, the welfare safety net would mitigate any harmful effects and effectively combat the detrimental economic consequences of a family breakup. The alternative hypothesis was that as a result of economies of scale and risk pooling, the problems would persist regardless of the social policy model that is applied in one country or another. Our results showed that even under a generous welfare model such Denmark’s, divorce has a major impact on individuals’ livelihood and is gender dependent. The average reduction in income as a result of divorce is about 30 per cent for women and 21–24 per cent for men. These figures are considerably lower than those arrived at in some influential American studies (for example, Weitzman, 1985), but not so different from those presented in more conservative American studies (for example, Peterson, 1996) and European comparisons. One important explanation for this decline in income in the Danish case is linked to economies of scale and our way of calculating and using equivalent incomes. In a few cases, we ended up with a seemingly contradictory situation in which income in households with a female head increased in absolute terms, yet there was a considerable reduction in the equivalent disposable income. This result could reflect the change in economic constraints that single breadwinner families face.

Economic and social consequences of divorce in Denmark

125

The declines in income levels are more or less directly mirrored in the poverty rates for single parents. Remarrying brings the high single-parent poverty rates down significantly; however, this effect is not as strong as the negative impact of divorce. The same phenomenon – although in a much more accentuated form – is visible in the recipience of social assistance. Assistance rates go up gradually and reach their peak four to five years after a divorce, and they do not go down as rapidly as poverty rates after remarriage. When it comes to the effect of the celebrated Scandinavian welfare state model, these findings raise at least two issues for consideration. First, the social policy network is not able to fully protect individuals against the hazards caused by a family breakup. Secondly, in some cases, the steps the welfare state takes to improve the lot of separated people may lead to incentive problems and create welfare dependency. Our results also indicate that there are some important qualitative differences between genders. The men who did not remarry during the decade we observed them were more miserable than their brothers who managed to find new partners. As a rule, even initial income levels were lower in the former group, and their relative situation turned out to be even worse towards the end of the 1990s. Hourly income lagged behind, and unemployment, poverty and social assistance recipiency rates were consistently higher among this group than among the remarried males. The reverse story applied to women who stayed single throughout that decade. They tended to do better than other groups of single women, and by 1999, their income levels were the highest among all women. The other relatively successful group of divorcees were single fathers – that is, men who took custody of their children when divorced. The unemployment rates for single fathers went down immediately after divorce, while precisely the opposite happened for single mothers. This result draws our attention to the fact that single fathers perhaps had a better foothold in the labour market to start with, whereas single mothers presumably faced greater problems in combining single parenthood and work, and were more likely to be receiving unemployment compensation. The most ‘miserable’ group consisted of men who had lost their children and spouse and did not find anyone new with whom to share their lives. Thus, our longitudinal study supports some of the results found in cross-sectional and qualitative studies on the gendered consequences of divorce: women are hit most severely economically, while men’s sufferings are found to be more in the social sphere. In conclusion, on the basis of the Danish evidence, the best way of improving one’s income level after divorce is to remarry – but the very best way to keep one’s income level up is not to divorce. Those Danes who did not divorce at all displayed the highest initial income levels, and their income curve increased steadily. Money and happy marriage seem to go

126

Single country studies

together well even in Denmark – in spite of the generous welfare state, or thanks to it.

NOTE 1. The equivalence scale we use here is almost the same as the new OECD scale, which gives a weight of 1 to the first adult family member, a weight of 0.5 to the second adult and a value of 0.3 to each child. Thus, the OECD scale for two adults and two children would be 1 1 0.5 1 2 3 0.3 5 2.1.

REFERENCES Abrahamsson, P. (1988), ‘Urban poverty and social dynamics of the welfare state. A comparative review of the narrative and the practices in Europe and in the United States’, Department of Sociology, University of Copenhagen, mimeo. Budzt, H. (1970), Ægteskab, separation, skilsmisse, Rinkøbing: Forlaget IDAG. Datta Gupta, N. R. Oaxaca and N. Smith (2006), ‘Swimming upstream, floating downstream: comparing women’s relative wage progress in the United States and Denmark’, Industrial and Labour Relations Review, 59 (2), 243–66. European Social Survey (ESS) (2007), ‘European Social Survey’, www.europeansocialsurvey.org/, accessed 25 May. Ferrarini, T. (2006), Families, States and Labour Markets: Institutions, causes and Consequences of Family Policy in Post-War Welfare States, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Ferrarini, T. and K. Forssen (2005), ‘Family policy and cross-national patterns of poverty’, in O. Kangas and J. Palme (eds), Social Policy and Economic Development in the Nordic Countries, Houndmills: Palgrave/Macmillan, pp. 118–46. Fritzell, J. and U. Henz (2001), ‘Household income dynamics: mobility out of and into low income over the life course’, in J.O. Jonsson and C. Mills (eds), From Cradle to Grave, York: Sociology Press, pp. 185–211. Fritzell, J. and V.-M. Ritakallio (2004), ‘Societal shifts and changed patterns of poverty’, LIS Working Paper Series, no. 393, Syracuse University, Syracuse, NY. Gähler, M. (1998), Life after Divorce: Economic, Social and Psychological Wellbeing among Swedish Adult and Children Following Family Dissolution, Edsbruk: Akademitryck. Gornick, J. and M. Meyers (2003), Families that Work: Policies for Reconciling Parenthood and Employment, New York: Russell Sage Foundation. Holden, K. and P. Smock (1991), ‘The economic costs of marital dissolution: why do women bear a disproportionate cost?’, Annual Review of Sociology, 17, 51–78. Houmann, A.L. (2006), ‘Hurtig skilsmisse kan redde kvinders økonomi’ (‘Quick divorce can save the economy of women’), Berlingske Tidende, first section, p. 1. Jensen, A. (2003), ‘Fra selvfølge till symbol – ægteskap I krydsfeltet mellem stat, kirke og befolkning I 1800-tallet’, unpublished PhD thesis, Syddansk Universitet.

Economic and social consequences of divorce in Denmark

127

Kangas, O. and J. Palme (2005), Social Policy and Economic Development in the Nordic countries, Houndmills: Palgrave/Macmillan. Kangas, O. and V.-M. Ritakallio (1998), ‘Different methods – different results: approaches to multidimensional poverty’, in H.-J. Andreß (ed.), Empirical Poverty Studies, Aldershot: Ashgate, pp. 167–203. Koch-Nielsen, I., J. Ussing and G. Schmidt (1976), ‘Familie 1975 – lov og tal’, Socialforskningsinstituttet Meddelelse 19, København: Socialforskningsinstituttet. Korpi, W. (2000), ‘Faces of inequality: gender, class, and patterns of inequalities in different types of welfare states’, Social Politics, 7 (2), 127–91. Leira, A. (2002), Working Parents and the Welfare State: Family Change and Policy Reform in Scandinavia, Cambridge: Cambridge University Press. Lewis, J. (1992), ‘Gender and the development of welfare regimes’, Journal of European Social Policy, 2 (2), 159–73. Lewis, J. (1993), Women and Social Policies in Europe: Work, Family and the State, Aldershot, UK and Brookfield, US: Edgar Elgar. Melby, K., A. Pylkkänen, B. Rosenbeck and C. Carlsson Wetterberg (2006), Äktenskap och politik i Norden ca 1850–1930, Lund: Makadam. Nielsen, L. and J. Vostrup Rasmussen (2001), Familjeretten, København: Thomson. Nordic Social Statistical Committee (NOSOSCO) (2002), ‘Social protection in the Nordic countries 2000’, Nordic Social Statistical Committee, Copenhagen. Nygaard Christoffersen, M. (1996), Opvæeks hos fædre: En sammeligning af 3-5 årige børns opvæeks hos fædre og mødre, København: Socialforskningsinsituttet. Nygaard Christoffersen, M. (2004), Familjen udvikling I det 20 århundrede, København: Socialforskningsinsituttet. Olsen, B.M. (2005), Mæend, orlov og arbejspladskultur. Fire danske virksomheder, København: Socialforskningsinsituttet. Organisation for Economic Co-operation and Development (OECD) (1998), Family, Market and Community. Equity and Efficiency in Social Policy, Paris: OECD. Organisation for Economic Co-operation and Development (OECD) (1999), OECD Historical Statistics 1960–1997, Paris: OECD. Organisation for Economic Co-operation and Development (OECD) (2001), Taxing Wages 2000–2001, Paris: OECD. Palme, J. (1999), The Nordic Model of Social Protection and the Modernization of Social Protection in Europe, Copenhagen: Nordic Council of Ministers. Peterson, R. (1996), ‘A re-evaluation of the economic consequences of divorce’, American Sociological Review, 61, 528–36. Putnam, R.D. (ed.) (2002), Democracies in Flux: The Evolution of Social Capital in Contemporary Society, Oxford: Oxford University Press. Rostgaard, T. (2004), ‘With due care: social care for the young and the old across Europe’, PhD thesis, Danish National Institute of Social Research/Southern Danish University. Rowntree, S. (1901), Poverty: The Study of Town Life, London: Macmillan. Schneer, J.A. and F. Reitman (2002), ‘Managerial life without a wife: family structure and managerial career success’, Journal of Business Ethics, 37, 25–38. Smock, P., W. Manning and S. Gupta (1999), ‘The effect of marriage and divorce on women’s economic well-being’, American Sociological Review, 64, 794– 812.

128

Single country studies

Taylor-Gooby, P. (2004), New Risks, New Welfare The Transformation of the European Welfare State, Oxford: Oxford University Press. Uunk, W. (2004), ‘The economic consequences of divorce for women in the European Union: the impact of welfare state arrangements’, European Journal of Population, 20, 251–85. Waite, L. (1995), ‘Does Marriage Matter?’, Demography, 32, 483–507. Wegener, M. (1977), Familieret, Odense: Odense Universitetsforlag. Weitzman, L. (1985), The Divorce Revolution: The Unexpected Social and Economic Consequences for Women and Children in America, New York: Free Press.

Economic and social consequences of divorce in Denmark

129

APPENDIX The Construction of Additive Well-being Indices Social trust is an additive index of three questions that pertain to the degree of social capital (see, for example, Putnam, 2002): ‘Most people can be trusted, or you can’t be too careful in dealing with people’; ‘Most people would try to take advantage of you if they got the chance’, and ‘Most of the time people try to be helpful’. In all three questions, the scale varied between 0 and 10; consequently, the additive index varies between 0 (no trust) and 30 (maximal trust). In various previous studies, this social capital indicator has been shown to be a good predictor of a number of positive social outcomes (health, social relations, and so on). ‘Happiness’ is a collapsed index of two questions measuring self-rated life satisfaction: ‘Taking all things together, how happy would you say you are?’ and ‘All things considered, how satisfied are you with your life as a whole nowadays?’ The index varies between 0 (extremely dissatisfied) and 20 (extremely satisfied). ‘Social contacts’ is based on responses to two questions: ‘How often do you meet socially with friends, relatives or work colleagues?’ (1 5 never . . . 7 5 every day) and ‘Compared to other people in your age, how often would you say you take part in social activities?’ (1 5 much less . . . 5 much more) that were added together. The minimum value of 2 indicates low frequency of contacts, while the maximum value 12 pertains to frequent contacts. ‘Health status’ is based on self-rated health status: ‘How is your health in general?’ (1 5 very good . . . 5 5 very bad) and the question ‘Are you hampered in daily activities by any longstanding illness or disability?’ (1 5 yes, a lot, 2 5 yes, to some extent, 3 5 no). The latter categories were recoded so that the original answer ‘yes, a lot’ was assigned a value of 5, the answer ‘yes, to some extent’ was assigned a value of 3, and ‘no’ was assigned a value of 1. This recoded variable was then added to self-evaluated health. The compound variable ‘Sickness’ ranges from 2 to 10, with the higher values representing worse health status. Finally, to get an evaluation of the respondent’s economic situation, we used the questions ‘If for some reason you were in serious financial difficulties and had to borrow money to make ends meet, how difficult or easy would that be?’ (1 5 very difficult . . . 5 5 very easy) and the question ‘How you feel your household income?’ (1 5 Living comfortably . . . 4 5 very difficult). The values of the former question were reversed, and the two response alternatives (‘very easy’ and ‘easy’) were merged into the value of 1. Thus, the economic hardship indicator varies from 2 (easy life) to 8 (economic problems).

130

Source:

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 All %

Year

15 074 14 986 14 908 14 857 14 795 14 722 14 699 14 626 14 569 14 504 147 740 100

All

10.2 10.1 10.1 10.1 10.0 10.0 9.9 9.9 9.9 9.8 100

%

Sample size

3 615 3 610 3 631 3 624 3 583 3 564 3 596 3 618 3 640 3 660 36 141 24.5

Single 11 459 11 376 11 277 11 233 11 212 11 158 11 103 11 008 10 929 10 844 111 599 75.5

Couple

Individuals

7 694 7 651 7 589 7 552 7 519 7 478 7 469 7 419 7 384 7 338 75 093 50.8

Male 7 380 7 335 7 319 7 305 7 276 7 244 7 230 7 207 7 185 7 166 72 647 49.2

Female

Statistics Denmark, 1% sample of the Danish population aged 30–48 in 1990.

Table 4A.1

10 9 8 7 6 5 4 3 2 1

Observations

14 181 277 141 121 64 82 75 58 59 76 15 134

Individuals

141 810 2 493 1 128 847 384 410 300 174 118 76 147 740

Obs.*Ind.

Panel observations per individual

Economic and social consequences of divorce in Denmark

131

Table 4A.2 Summary statistics for OLS regression (Table 4.4) Mean ln(equivalized household income) Female Age (years) Age squared/100 Immigrant Couple → single Single → single Single → couple Number of children University education Labour market experience (years) Full-time employed Unemployment (%) 1992 1993 1994 1995 1996 1997 1998 1999 Source:

4.803 0.492 44.429 20.130 0.045 0.041 0.204 0.040 0.810 0.055 10.275 0.721 7.310 0.113 0.113 0.112 0.111 0.111 0.111 0.110 0.110

Std.dev. 0.460 0.500 6.253 5.562 0.207 0.198 0.403 0.197 1.014 0.228 5.591 0.449 20.561 0.316 0.316 0.315 0.315 0.314 0.314 0.313 0.313

Minimum Maximum 27.091 0 31 9.610 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

10.022 1 58 33.640 1 1 1 1 8 1 20 1 100 1 1 1 1 1 1 1 1

Statistics Denmark, 1% sample of the Danish population aged 30–48 in 1990.

5.

Divergences in the Nordic model: eE conomic consequences of partnership dissolution in Sweden and Finland Heikki Hiilamo

INTRODUCTION The Nordic countries are known as pioneers in reconciling work and family as well as in recognizing new family forms (see, for example, Bradshaw and Hatland, 2006; Duncan, 1995; Fraser, 1997; Gauthier, 1996; Gornick et al., 1998; Kamerman and Kahn, 1978; Kaufmann, 1993; Korpi, 2000; Lewis and Hobson, 1997; Montanari, 2000; Pfenning and Bahle, 2000). This means, for example, that women are not financially dependent upon their husbands to the same extent that they are in conservative welfare states. Consequently, the economic consequences of partnership dissolution are supposedly not as grave as elsewhere. In comparative studies, Finland and Sweden are often associated together; both countries represent the Nordic welfare model (Kautto, 2000). The more removed the perspective from which the countries are observed, the more similar they look. From a European perspective, both countries belong to the same category of welfare states. Roughly speaking, the development of a welfare state has been identical in both countries (see, for example, Flora, 1986). Sweden and Finland are small, relatively open economies which are very dependent on international economic cycles. Their political backgrounds and social structures closely resemble each other. Other unifying factors include their northern geographical position, Lutheranism as a state religion, social homogeneity and national, linguistic and historical traditions. In both countries, national social policy has been developed on the basis of democratic corporatism (cf. Katzenstein, 1985). When we study the two countries in more detail, however, significant differences begin to emerge behind the overall commonalities with regard to demographic characteristics, family law and family policies (see, for 132

Economic consequences of partnership dissolution in Sweden and Finland 133

example, Castles, 1978; Kosonen, 1998). Some researchers suggest that Finland is moving towards neo-familialism – that is, towards the familial model – while Sweden is committed to strengthening the egalitarian model (Mahon, 2002). This perception challenges the image of both Sweden and Finland as equal representatives of the egalitarian countries. To facilitate an analysis of possible differences between the two countries with regard to the economic consequences of divorce, I refer to two different policy orientations. The egalitarian model stresses the government’s role in family welfare and gender equality (Gauthier, 1996; Nyberg, 2002). The tradition is known for extensive programmes to support female labour force participation – most notably, highly developed parental insurance and public provisions for day care. The aim of the model is to guarantee individual freedom of choice. That implies that the model does not impose restrictions or penalties, for example, in the case of divorce. On the contrary, the egalitarian model tries to mitigate the negative effects of divorce. The familial model, on the other hand, emphasizes the role of the family in guaranteeing the well-being of individuals (Kaufmann, 1993). This tradition stresses the role of the breadwinner – for example, through joint taxation and employment-based social insurance benefits. The stability of the family is the main aim of the model. Therefore, the model includes barriers to divorce because it assumes that policies which support divorced parents and their children work as incentives to divorce. At this point, it is necessary to emphasize that I am not expecting to find dramatic differences between Sweden and Finland – for example, visà-vis other European countries. However, my aim is to add more details to the analysis of the economic consequences of divorce in two Nordic countries. To that end, I provide a sketch of recent demographic trends in Sweden and Finland. Furthermore, I investigate whether there are differences between Sweden and Finland in their institutional orientations and underlying ideologies with regard to partnership dissolution, and if so, whether they result in different economic consequences of partnership dissolution. This chapter is organized as follows: the second section deals with demographic trends, while the third section explores institutional orientations in family law and family policies. I use this background information to formulate hypotheses about the economic consequences of divorce in the fourth section. The fifth section measures the outcomes of partnership dissolution in terms of mothers’ employment and earnings. I utilize and reanalyse employment and earnings data published by Haataja and Nyberg (2006). The data is based on labour force surveys conducted in Sweden and Finland and on household income surveys (Luxembourg Income Study).

134

Single country studies

SOCIO-DEMOGRAPHIC TRENDS OF DIVORCE IN SWEDEN AND FINLAND: 1970–2000 Before discussing the institutional context in detail, it is helpful to have a sound knowledge of the socio-demographic trends of divorce in the two countries. The prevalent theories of divorce assume that a couple divorces when the real or perceived advantages of the outside option exceed those of a partnership (Teachman, 2002: 331–2). The rise in divorce risk has been attributed to the flow of women into the labour market (Finch, 2006). Scholars have explained the association between female labour force participation and a higher risk of divorce in terms of the so-called independence effect: when a woman earns enough to support herself financially, it becomes easier for her to leave a bad marriage (Liu and Vikat, 2004). Empirical research from Sweden and Finland – and from other countries as well – supports this theory: there is a positive relationship to the share of wives’ income. A wife’s income increases the risk of divorce at all levels of the other spouse’s income, but it does so especially when the wife’s income exceeds that of the husband’s, indicating that divorce is a matter of power rather than of absolute earnings (Jalovaara, 2003; Liu and Vikat, 2004). The general picture of marriage in Sweden and in Finland is that of declining popularity. Indeed, divorce rates have increased together with female labour force participation in Sweden and Finland throughout the twentieth century (Härkönen, 2005). Both countries are among those with the highest divorce rates in Europe (see Data Appendix). Statistics indicate that almost half of all recent marriages will end in divorce unless the present trend changes. Marriages contracted in the mid-1970s have reached the highest frequency of divorce; one-third of them have ended in dissolution. Marriages contracted later seem to be ending in divorce even more frequently. However, the growth in divorce rates seems to be stabilizing in both countries. Since 1995, the probability of first marriage has increased in Finland (Finch, 2006). The first marriage rate is clearly higher in Finland than it is in Sweden, while the share of never-married persons is much higher in Sweden than in Finland. Consensual unions are very popular in both countries, especially among the young. It is rare for a young couple to marry without first having lived together. Cohabitation is a less stable relationship than marriage, as cohabitation is more likely to end in dissolution than marriage is (Björnberg, 2001). A major difference between Finland and Sweden is that cohabitation is becoming a more permanent family form in Sweden, while in Finland (as in most Western countries) cohabitation is a temporary union and a precursor to marriage. In 2000, 48 per cent of all Swedish couples between 30 and 39 years of age were cohabiting (Finch, 2006: 19).

Economic consequences of partnership dissolution in Sweden and Finland 135

The corresponding figure for Finland was 36 per cent. These figures are reflected in the fact that more than half of all children in Sweden are born outside marriage, while in Finland the figure is around 40 per cent. The share of single-parent families among all families with children has continued to increase during the early 2000s. In 2003, some 20 per cent of all Finnish households with children were single-parent families, while in Sweden the corresponding figure was 24 per cent. Most single parents are women (NOSOSCO, 2005). The share of women among single parents in Finland was 87 per cent; in Sweden, 79 per cent. In 2000, 17.8 per cent of all children in Sweden lived in single-mother households (LIS, 2006). The figure for Finland was 12.8 per cent. The fertility pattern is very similar in both countries. The age-specific fertility of younger women (under 30) has declined, while that of older women has increased. In a comparative perspective, both countries enjoy a considerably high fertility rate of around 1.8 (see Data Appendix). In both countries, women’s average age of first marriage is above the average age of having the first child (Finch, 2006). In Sweden there are fewer marriages, generally more divorces and more children born outside marriage. The share of single parents is also higher. Moreover, cohabitation is becoming a permanent form of partnership in Sweden, while in Finland it remains a precursor to marriage (Finch, 2006). The institution of the family is without doubt less stable in Sweden than it is in Finland.

INSTITUTIONAL CONTEXT In the following section, I look at family law and family policies with regard to the economic consequences of partnership dissolution. I analyse this institutional context according to the extent to which it gives support to and/or places restrictions on the partners before and after dissolution. As already mentioned in the introduction, I distinguish between two different policy orientations. The egalitarian model stresses the rights of individuals and does not give preference to marriage (Nyberg, 2002). The familial model provides support only to families and gives preference to marriage over other forms of partnership, thereby creating barriers to partnership dissolution (Kaufmann, 1993). Divorce Law Divorce law has major economic consequences – for example, in terms of the division of assets as well as in determining custody, maintenance

136

Single country studies

and alimony issues. Divorce laws in Finland and Sweden are very similar. While married, both spouses must contribute financially to the family’s needs according to their ability, and often the finances are more or less shared. When a marriage is dissolved, not only does cohabitation come to an end, but financial commonality as well. In principle, all the spouses’ assets – after the deduction of debts – should be divided equally. However, spouses may have a mutual agreement which stipulates the division of assets in the case of an eventual divorce. When a marriage is dissolved, both spouses are required to support themselves. However, it is possible that alimony may be paid during a transition period (and in exceptional cases, for an extended period) – for example, to a wife who has been a homemaker. In both countries, the courts decide divorces. If both parties have agreed to get divorced, they submit a joint petition for divorce to the District Court (Social Ministry, 2000). The major difference in divorce law concerns couples without children. It always takes at least six months to get a divorce in Finland, while in Sweden couples without children may get a divorce almost instantly if they are in agreement. If Swedish spouses have a common child under the age of 16 – or if one of them is the guardian of a child under age 16 who lives at home – the divorce can only be decreed after a period of reconsideration lasting at least six months. In Finland, the period of reconsideration applies for all married couples. It is possible, however, to get an instant divorce in Finland as well, if the spouses have lived separately for two continuous years before the petition is filed. In both countries, the period of reconsideration does not automatically end with a divorce petition. After six months have passed, one of the spouses must again inform the District Court as to whether he or she wishes to go through with the divorce. If the District Court does not hear from either spouse within a year from the date of the petition, the court will dismiss the case and the marriage will continue to exist. Consensual unions do not have official status either in Finland or in Sweden. Consequently, the courts do not intervene if a consensual union breaks apart. If a consensual union ends in dissolution, the division of the property includes any residence and furnishings acquired for the partners’ common use. In case of a dispute, the partner claiming that the property has been acquired for common use must produce evidence – for example, a contract or receipt – for his or her claim. No other property can be taken into account in the division. Consequently, any residence or furnishings not acquired for the partners’ common use is excluded from the division of property, as are all other assets. The party who owns the property retains it.

Economic consequences of partnership dissolution in Sweden and Finland 137

Child Maintenance Both in Sweden and in Finland, parents who have separated are entitled to agree on the amount of child maintenance payments. If the partners cannot agree, the court may determine the amount of the maintenance. The authorities have to make sure that the sum agreed upon is not unfair to either party. The average amount of maintenance decreed by the court in Finland in 2005 was around 125 euros per month (Valkama and Litmala, 2006). No uniform practice to determine maintenance is in force in either country.1 The level of maintenance is determined by the needs of the child and the income level of the parents. If the child spends an equal amount of time with both parents, no maintenance is paid. Swedish legislation is somewhat more flexible with respect to divorced parents. In Sweden, joint custody continues if the parents get divorced, while in Finland the parents must agree on custody. Almost all Swedish parents have joint legal custody after separation or divorce or if they are unmarried, but in most cases the children live predominantly or exclusively with the mother (Bernhardt, 2003). Joint custody as such does not have a major influence on the economic consequences of partnership dissolution. Joint parenting is, however, clearly more popular in Sweden than it is in Finland. In Sweden, around a quarter of all children of divorced or separated parents live alternately with each of their parents. Both in Sweden and in Finland, family policy includes a system through which, under certain conditions, the government guarantees payment of maintenance to single parents. The purpose of maintenance allowances is to compensate for or supplement the parental maintenance to which the child is otherwise entitled and to guarantee a minimum level of allowance to the custodial parent. Maintenance allowances are paid until the child turns 18 (in Sweden, if the child is still attending school, an extended allowance is paid until the end of June of the year in which the child turns 20). In Finland, these allowances are paid by the municipal social welfare boards and in Sweden by the social insurance offices. If the parental maintenance paid is less than the maintenance allowance level, the difference is paid as an allowance. The current level (2008) of maintenance allowance in Sweden and in Finland is around 130 euros per month. Non-resident parents are legally obliged to pay back the maintenance allowance. The difference between the two countries lies in the process of claiming maintenance. In Finland, the municipal social offices make claims concerning not only the maintenance allowance (the amount the liable parent owes to the municipality) but also the maintenance payments (the amount above the maintenance allowance that the liable parent owes to the custodial

138

Single country studies

parent). This means that the municipality, if successful in collecting the full amount, also pays the amount of maintenance payments which exceed the maintenance allowance. In Sweden, the non-resident parent’s obligation to the social insurance office is limited to a certain percentage of that parent’s income (assets are also taken into account above a certain threshold), while in Finland the non-resident parent’s ability is assessed when the maintenance is first determined. The Swedish social insurance office is not responsible for collecting unpaid maintenance payments. It is only entitled to claim money spent on maintenance allowances. This means that the Swedish system is more liberal with regard to non-resident parents. A third form of maintenance allowance also exists in Sweden; an allowance for children who live alternately with each of their parents. If the child spends equal time or almost equal time with both parents, both of them may apply for the maintenance allowance, which is 50 per cent of the level of the full maintenance allowance. The benefit is granted without obligation of repayment; however, it is means tested. The existence of a maintenance allowance for children who live alternately with each of their parents and the fact that the collection systems are more flexible towards non-resident parents provides grounds to argue that the Swedish maintenance system is less familial than that in Finland. The Swedish system encourages parents to share care and financial responsibility for their children in the event of a partnership dissolution, thereby mitigating an unequal distribution of the negative effects of such an event. Taxation Joint taxation leads to higher costs than separate taxation in the case of a divorce. The general aim of family taxation is to support the breadwinner. However, in the event of a divorce, the amount of taxes paid will fall, since the incomes of the spouses are no longer combined. In the case of separate taxation, tax deductions may increase the cost of partnership dissolution. This is true for exemptions, which are made from income. Within the partnership, the partners may choose to have the exemptions taken from the income of the higher-earning partner. Thereby the household’s overall tax load is decreased, since the higher-earning partner’s tax rates are relatively higher. Separate tax assessment came into force in Sweden in 1971 (Sundström, 1994). However, some features connected to the number of children in the family remained. All elements of joint taxation were removed in 1991, including a tax deduction for single parents.2 Thereafter, support to single parents was to be provided through housing allowances.

Economic consequences of partnership dissolution in Sweden and Finland 139

Finland adopted completely separate tax assessment for spouses in 1976. However, tax deductions were an important element of family policy until the early 1990s. After that, the situation changed quickly. In the so-called family support reform of 1994 (except for a small deduction for maintenance liability), tax deductions for families with children were abolished and replaced by direct support – that is, an increase in cash child benefits. The decision to keep the deduction for maintenance liability was justified on the grounds that it motivates the payment of alimony (Kaurala, 1993: 26). Finland’s single-parent deduction was replaced by special increased cash child benefits for single parents (Hiilamo, 2002). The prime losers in this revision were single parents cohabiting with new partners. They lost the single parent deductions, and due to their cohabitation, were not eligible for the special cash child benefits which replaced the abolished deduction. Thus, this change increased the financial dependence of cohabiting single parents. Ever since 1994, the taxation systems in Sweden and Finland have been identical with regard to the economic consequences of partnership dissolution. No differences in outcomes are expected as a result of taxation. Cash Child Benefits Tax deductions, directed at spouses according to their choice, are a family-oriented benefit, whereas cash child benefits – especially in the Nordic countries – are typically an individual benefit, usually paid to the child’s mother on a universal basis (Wennemo, 1994). All children who are Swedish citizens and resident in Sweden are entitled to a cash child benefit (103 euros per month for the first child) up to and including the year they turn 16. In Finland, cash child benefits (100 euros per month for the first child) are paid until a child turns 17. The major difference between Sweden and Finland is the special supplement to cash child benefits for single parents (36 euros per month) which is paid in Finland (and in all other Nordic countries except Sweden). In case of partnership dissolution, the cash child benefit is paid to the custodial parent. In Sweden, however, in the case of joint custody, the parents are allowed to split the cash child benefit. In Finland this is not possible. The special supplement to cash child benefits for single parents means that the Finnish system is more individual than the Swedish system in that respect. In sum, the cash child benefit systems do not result in any major difference between Finland and Sweden with regard to the economic consequences of partnership dissolution.

140

Single country studies

Parental Insurance Benefits Parental benefits in Sweden and Finland consist of the same elements: parental leave time and cash benefits. However, measured in terms of the level and coverage of the benefits, parental benefits in Sweden are significantly higher than those in Finland (for single mothers, the leave time is 480 days compared with 263 days in Finland). In 2005, the average cash benefit in Sweden was around 46 euros per day, while in Finland it was around 40 euros per day. Furthermore, Swedish benefits are broader (for example, they include leave time to care for sick children and for contact days) and their payment is more flexible. It is possible to receive the benefits until the child has turned 8 or finished the first grade at school. The days of parental leave are divided equally between both parents and can be transferred between the parents at their own discretion. Single parents have the right to use all the days themselves (Haataja, 2006). In Finland, single mothers cannot use that part of the parental leave which is allotted to fathers (18 working days). Hence, the system in Finland is slightly more familial than that in Sweden. Housing Allowances The purpose of housing allowances is to improve housing levels in terms of quality and spaciousness, to even out differences in costs between different regions and dwellings of different ages, and to improve the economic situation of those on low incomes (SOU, 2001:24: 129–62). In 2005, the average amount of housing allowance for a family with children was 290 euros per month in Finland and 184 euros per month in Sweden. The system in Sweden differs from that in Finland in two important respects: first, a housing allowance may also be payable to a parent in whose home the child is living periodically as a result of custody or visiting rights. Secondly, household income is not combined when a housing allowance is calculated. Sweden introduced the use of individual income as a basis in 1997, replacing the previous reliance on joint gross family income. Both conditions provide support for non-traditional family forms and make partnership dissolution a less costly option. Public Provision of Day Care Public provisions for childcare play a decisive role in single parents’ labour market participation (OECD, 2005). If high-quality child day care is easily available at a low cost, a parent with small children may leave an unhappy partnership without withdrawing from the labour market. Although the

Economic consequences of partnership dissolution in Sweden and Finland 141

comprehensiveness and levels vary, all of the Western European countries provide direct income transfers to families with children, but few other countries provide social services for families with children that are as extensive as those in the Nordic countries (Kvist, 1998: 169; Sipilä, 1997). In both Finland and Sweden, the share of children below school age who are in public day care is very high in comparative perspective. The fact that the share is higher in Sweden is explained by the child home care allowance that is granted to parents with children under 3 years of age in Finland (see below). When it comes to children between 3 and 6 years of age, a clear majority are in public day care in both countries. However, the share of children between 3 and 6 years old who are in public day care is significantly higher in Sweden than it is in Finland (see Data Appendix). In Finland, day-care services were expanded in 1990, when all children under 3 were granted the subjective right to day care. This meant that municipalities were obliged by law to provide day care for every child under 3 years of age. Finland was the first country in the world to implement such a subjective right, but Sweden very soon followed suit (Anttonen and Sipilä, 2000: 128–9). These reforms finally made day-care services universal in the strict sense of the word, and they promoted economic gender equality significantly. In 1996, the subjective right to day care was extended to all children of preschool age in Finland. Almost at the same time, a similar reform was implemented in Sweden. It needs to be emphasized that day care in both countries is affordable. Income-related day-care fees constitute only a fraction of the overall costs of day-care arrangements (Hiilamo, 2002). Single parenthood is taken into account in fee schedules in both countries. However, the maximum fee is higher in Finland (250 euros per month) than in Sweden (137 euros per month), which gives us grounds to expect more negative effects on partnership dissolution in Finland, especially for partners with high incomes. Family Ideologies While studying European childcare more than a decade ago, Kamerman and Kahn (1991) gave telling titles to both countries’ orientations: ‘Finland: supporting parental choice’ and ‘Sweden: supporting work, family and gender equality’. In Finland, parents (in practice, mothers) may choose to care for a child at home with the home care allowance or place the child in day care, while in Sweden no such option exists. The issue of the child home care allowance is the factor that most clearly distinguishes between Swedish and Finnish family policy – and perhaps social policy (Hiilamo and Kangas, 2006). In Sweden, the allowance for families with small children consists solely of parents’ insurance and public day care,

142

Single country studies

whereas in Finland, a child home care allowance paid to parents of children under age three is also in use. In Finland, the child home care allowance began in the form of different municipal and regional experiments in the late 1960s (Sakaranaho, 1992: 33; Takala, 2000: 11). The reason for this was a serious shortage of day care places in most municipalities. The municipalities tried to diminish the demand for day-care places by paying a home care allowance. The payment of the child home care allowance started with a government resolution in 1980. The benefit was a tax-free, flat-rate amount payable to families with at least three children under school age. The coverage was gradually extended to also include families with one child. The national child home care allowance system was established in 1985, when the Act on Child Home Care Allowance came into force. The payment of the allowance was made to all parents or other guardians of children, if the child was not in municipal day care. A means-tested supplementary part was paid to those who cared for their children at home. The child home care allowance has been supplemented by other measures. In 1985, an amendment to the Employment Contracts Act guaranteed employees the right – after the termination of the parental allowance – to take a childcare leave until the child turns 3. Currently, the level of the flat-rate home care allowance is 290 euros per month (with a supplement for other children of 50–84 euros per month), and the income-related supplement is a maximum of 168 euros per month. Most parents are eligible for municipal supplements to the home care allowance ranging from 20 to 300 euros per month. In Sweden, discussion about a child home care allowance started earlier than it did in Finland, but the discussion did not result in legislative measures until the bourgeois coalition government carried the allowance scheme through late in the summer of 1994 (Hinnfors, 1992). The succeeding Social Democratic government immediately abolished the Act. Thus the allowance was in use for only six months. Despite the negative attitude of the governing Social Democrats, some municipalities have proposed – and some have even decided to grant – municipal home care allowances to children who are not in day care arranged by the municipality. As in the Finnish case, the allowance has been justified by referring to the fact that it would save municipal funds. However, courts have cancelled these municipal decisions as illegal. In Finland, equality between genders as an objective of family policy was not as clearly emphasized as it was in Sweden, where the child home care allowance was interpreted as propping up the traditional and oldfashioned family model. The traditional family model was seen as discriminating against women and violating gender equality. The allowance, which compensated for only a small part of the lost income, would not

Economic consequences of partnership dissolution in Sweden and Finland 143

have encouraged men to stay at home to care for children. Thus, women’s dependence on ‘breadwinner men’ would have increased, and the allowance would have created a new class of full-time mothers in Sweden. The aim was to solve women’s double burden with shared parenthood and equal possibilities and the freedom to fulfil oneself both on the labour market and at home. The child home care allowance was branded by the ruling Social Democrats as a ‘trap for women’. A similar line of argumentation has been dominating the debate on tax allowances for domestic help (cleaning, renovation, gardening and so on). Such an allowance has been applied and developed in Finland since 1997. The allowance reduces the price of domestic help in Finland by as much as 40 per cent, up to a maximum amount of 5850 euros per year. The reform has created a substantial number of new jobs in the household service sector (Niilola and Valtakari, 2006). The Conservative Party in Sweden has argued for the introduction of a similar allowance there by referring to positive outcomes in Finland. The Social Democrats have attacked the reform as a step towards a society where lower class women are employed to serve richer families.

SUMMARY OF DEMOGRAPHIC TRENDS AND INSTITUTIONAL CONTEXTS Both Finland and Sweden are often classified as egalitarian countries. The above description does not give us grounds to challenge that understanding. However, several distinctions between the two countries were revealed with regard to both demographics and family policies. Table 5.1 summarizes the distinctions between Finland and Sweden. The institution of the family is clearly less stable in Sweden than it is in Finland. In Sweden there are fewer marriages, more divorces and more children born outside marriage. The proportion of single parents is also higher. Moreover, cohabitation is becoming a permanent form of partnership in Sweden, while in Finland it remains a precursor to marriage (Finch, 2006). These outcomes may be partly attributed to differences in family law. It is easier to get a divorce in Sweden than in Finland. Swedish couples without children may get a divorce almost instantly if they are in agreement. This is one indication of the fact that there is less support for traditional family values in Sweden than there is in Finland. Child maintenance arrangements in Sweden give more support to partners who do not live together. Joint custody is the norm, and parents are allowed to split the maintenance allowance in Sweden – yet another indication of flexibility towards partnership dissolution. Additionally, the

Single country studies

144

Table 5.1

Degree of familialism in divorce law and family policy in 2008 Finland

Sweden

Child maintenance

Period of consideration 6 months Guaranteed maintenance scheme (GM)

Taxation Cash child benefits

Individual taxation A special supplement for single parents

No period of consideration for childless couples GM payable to both parents if child spends equal time with both parents More flexible rules in recollecting maintenance liability Individual taxation Can be divided between parents if child spends equal time with parents More generous both in duration and level of benefits More flexible towards single parents Means testing on individual basis Support for non-resident parent Subjective right Maximum fee lower No such scheme

Divorce law

Parental benefits

Housing allowance

Means testing on household basis

Day care

Subjective right

Cash for childcare

Extensive scheme of home care allowance

recollection system for maintenance allowances is more flexible towards non-resident parents. There are hardly any differences in taxation or in public provisions for child day care between the two countries. However, in Sweden, parents are allowed to split the cash child benefit. Furthermore, parental benefits in Sweden are more flexible towards single parents than they are in Finland. Housing allowances in Sweden are more individualized in two respects: non-resident parents are entitled to a housing allowance, and household income is not combined when the housing allowance is granted. These differences indicate that Sweden is more progressive in supporting nontraditional family forms. Overall, it seems that the Swedish system not only avoids building barriers to separation and divorce, but that it also encourages parents to continue sharing the financial and care responsibilities for their children after a partnership dissolution. If this goal is achieved, the economic consequences of partnership dissolution are shared equally.

Economic consequences of partnership dissolution in Sweden and Finland 145

A major difference between the two countries concerns the ideology of childcare. The egalitarian model stresses the rights of individuals and does not give preference to marriage (Nyberg, 2002). The familial model gives support only to families and gives preference to marriage over other forms of partnership, thereby creating barriers to partnership dissolution (Kaufmann, 1993). Finland emphasizes parental choice between home care and day care, while Sweden supports only public day care. In this respect, the Finnish policy quite clearly reflects familialism or neo-familialism (Mahon, 2002). In practical terms, the option to stay at home with a small child concerns only women in Finland. Single parents with low qualifications in the labour market face strong incentives to stay at home with the cash childcare benefit (including the income-related supplement). If Finnish parents choose the home care allowance instead of employment following a partnership dissolution, they will face more negative effects in terms of income than their counterparts in Sweden. In the following section, I analyse whether the observed differences in institutions are reflected in partnership dissolution outcomes. The previous analysis of institutional contexts and socio-economic trends gives grounds to assume that the economic consequences of partnership dissolution are (1) mild from a European perspective, and (2) due to Finland’s neofamilialism, generally not as harsh in Sweden as they are in Finland.

ECONOMIC CONSEQUENCES: EMPIRICAL ANALYSIS Analysis of the economic consequences of partnership dissolution calls for commensurate longitudinal data-sets, which allow us to take pre- and postdissolution measurements. In the absence of such data-sets for Sweden and Finland, I review the literature to test the first hypothesis and conduct a small preliminary analysis of single parents’ employment status and earnings to explore the second hypothesis. The negative effects of partnership dissolution for women and children are most clearly reflected in single-mother households. When analysing the consequences of partnership dissolution in a longitudinal setting, it is sometimes difficult to capture the dissolutions of consensual unions. This problem does not apply to the analysis of single parents, since this category includes both previously married couples and couples who were in consensual unions. Single parenthood is most often a result of partnership dissolution. However, that is not always the case. The other partner may be deceased or was not present in the first place. In Finland, around 5 per cent of all single

146

Single country studies

parents are widows (Tilastokeskus, 2005). Given the similarities in demographic structures, we can assume that there are no significant differences between Finland and Sweden in the pathways to single parenthood. Results based on data from the European Household Panel Study (ECHP) indicate that women’s median adjusted household income decreases by 13 per cent after a divorce in Finland, which corresponds to the figure for other countries belonging to Esping-Andersen’s (1990) social democratic welfare regime category – for example, 13 per cent for Denmark and 19 per cent for the Netherlands (Uunk, 2004). But it is clearly a lower income loss than that in countries belonging to other welfare regimes in EspingAndersen’s typology. Earlier research has demonstrated that the incomes of Swedish women who divorced between 1981 and 1990 did not increase as much as those of other groups (Gähler, 1997: 244). Equivalized household incomes for those couples who continued their marriages rose by 28 per cent, while incomes for divorced women increased by 11 per cent, and incomes for divorced men increased by 36 per cent. These results have been explained by the fact that women earn lower wages, they lose economies of scale, and they take primary responsibility for any children. The results from Sweden also showed that divorcees, both female and male, reported a higher risk for psychological distress in 1991 than did their married, cohabiting or remarried counterparts (Gähler, 1998). Psychological distress seems to precede divorce among women, whereas among men, it lasts longer following a divorce. However, in a comparative perspective, the Swedish results are rather positive for both females and males. A study conducted in Belgium, Germany, Great Britain, Italy and Sweden showed that income decline after partnership dissolution was the lowest in Sweden (Andreß et al., in this volume). However, although Sweden stands out as the country with the highest gender equalities with respect to postseparation incomes, in the long run, Swedish men and women both have to deal with long-lasting financial consequences after separation. In conclusion, the earlier results seem to confirm my first hypothesis. The economic consequences of partnership dissolution in Sweden and in Finland are rather mild from a European perspective. Secondly, we may compare the situation of single parents in both countries during the 1990s. The recession that took place in the early 1990s provides a backdrop for the analysis. This decade may be described as a test case for single mothers’ work incentives. When a deep economic recession hit Sweden and Finland in the early 1990s, the unemployment rate rose to high levels in both countries. It is plausible that those with the weakest incentives to work – for example, separated mothers with small children

Economic consequences of partnership dissolution in Sweden and Finland 147

Table 5.2 Child poverty rate by family type (percentage) Country

Year

Finland

1991 1995 2000 1992 1995 2000

Sweden

Two-parent family 1.9 1.5 2.1 2.6 1.5 2.3

Single mother 5.8* 4.7* 8.1* 5.7* 6.6 12.9

Difference 3.9 3.2 6.0 3.1 5.1 10.6

Notes: Poverty defined as having an income less than 50% of median equivalized household income. * Estimates based on 15 to 30 observations only. Source:

LIS (2006).

entitled to home care allowances – would have withdrawn from the labour market. The incentive effect could have been aggravated by the fact that Finland continued to improve its home care allowance scheme during the first years of the recession (Hiilamo, 2002). In the following section, I study the living situation of single parents in both countries during the 1990s with employment, unemployment and earnings as outcome variables. Although women’s employment rates have converged over time across the Central and Northern European countries, the Nordic cluster continues to show higher female employment rates (Mayhew, 2006). In 1990, before the onset of the recession, poverty rates for single mothers were comparable to those of other population groups in both countries. The low levels of single-parent poverty in Sweden and Finland were largely explained by the fact that single parents participated in the labour market. Child poverty rates increased in both countries during the 1990s, especially among single-parent families (Table 5.2). Next, I compare employment rates between 1989/90 and 2002 (Table 5.3). The employment rate for single mothers dropped dramatically during the recession and did not recover to earlier levels (Haataja and Nyberg, 2005). In general terms, the situation of single parents degenerated substantially in both countries, as single parents were more seriously hit than couples with children. In 2002, the employment rate for single mothers with children aged 0–6 years was 63.5 per cent in Sweden and 52.9 per cent in Finland. The corresponding figures for 1989/90 were 80 per cent and 84.7 per cent, respectively. In Finland, the employment rate dropped more dramatically than in Sweden. The rate decreased by 31.8 percentage points for single mothers

Single country studies

148

Table 5.3

Employment rates for single and partnered mothers (percentage) Single mothers

Age of the youngest child 0–6 1989/90 2002 Change Age of the youngest child 7–17 1989/90 2002 Change Source:

Finland

Sweden

84.7 52.9 231.8

88.3 73.1 215.2

Partnered mothers

Difference

Finland

Sweden

Finland

Sweden

80 63 217

76.0 67.8 –8.2

88 80 28

28.7 14.9 23.6

8 17 9

90 80 210

90.4 86.3 –4.1

92 90 22

2.1 13.2 11.1

2 10 8

Haataja and Nyberg (2006).

with small children and 15.2 percentage points for mothers with schoolaged children. The corresponding figures for Sweden were 17 and 10 percentage points, respectively. In 2002, single mothers’ employment rates were higher in Sweden. These results give some support to the hypothesis that due to the home care allowance system, more single mothers withdrew from the labour market in Finland than in Sweden. A similar picture emerges if we compare single and partnered mothers in both countries. Single mothers’ employment rates dropped more than partnered mothers’ employment rates, and the development was more pronounced in Finland. However, if we compare differences in employment rates for single and partnered mothers with small children in 2002, single mothers in Finland compared with partnered mothers have higher employment rates than single mothers compared with partnered mothers in Sweden. This could be a result of the home care allowance, which especially decreases employment rates for partnered mothers who can rely on their partners’ income while on home care leave. Unemployment rates for single mothers rose more dramatically in Finland than in Sweden (Table 5.4). The level of unemployment was also higher in Finland in 2002. A similar trend applies to a comparison with partnered mothers. Overall, single mothers suffer from unemployment more than partnered mothers do. However, the difference is even greater in Finland. If we compare single mothers’ share of earnings in relation to the earnings of dual-earner fathers,3 Sweden once again does better than Finland (Table 5.5). The level of single mothers’ earnings was higher in Sweden in

Economic consequences of partnership dissolution in Sweden and Finland 149

Table 5.4 Unemployment rates for single and partnered mothers (percentage) Single mothers

Age of the youngest child 0–6 1989/90 2002 Change Age of the youngest child 7–17 1989/90 2002 Change Source:

Partnered mothers

Finland

Sweden

Finland

Sweden

1.7 14.3 12.6

5 10 5

2.0 7.0 5.0

1.0 3.5 2.5

3.6 9.1 5.5

2 5 3

2.0 5.4 3.4

1.0 2.5 1.5

Difference Finland

Sweden

0.3

27.3 27.6

24.0 26.5 22.5

21.6 23.7 22.1

21.0 22.5 21.5

Haataja and Nyberg (2006).

Table 5.5 Single mothers’ earnings as percentage of fathers’ earnings in dual-earner couples Age of the youngest child 0–6

1987 2000 Change Source:

7–17

Finland

Sweden

59.1 27.5 231.6

36.7 32.4 24.3

Finland

Sweden

68.5 58.7 29.8

53.8 44.0 29.8

Haataja and Nyberg (2006).

2000, and the share had not dropped between 1987 and 2000 as much as it did in Finland (for mothers of school-aged children, the decrease was 9.8 percentage points in both countries). The share of single mothers with earnings was higher in Sweden in 2000, and the decrease between 1987 and 2000 was smaller (Table 5.6). However, this observation applies only to single mothers with small children. If we look at single mothers with school-aged children in 2000, there were more single mothers with earnings in Finland. In summary, the development of the labour market was more negative towards single mothers in Finland than in Sweden during the 1990s. This could partly be a result of the fact that in Finland, more women stayed at

Single country studies

150

Table 5.6

Single mothers gainfully employed (percentage) Age of the youngest child 0–6

1987 2000 Change Source:

7–17

Finland

Sweden

Finland

96.4 61.4 235.0

96.6 73.5 223.1

93.7 88.2 25.5

Sweden 93.0 84.0 29.0

Haataja and Nyberg (2006).

home with the home care allowance in the event of a partnership dissolution. Once the depression hit the economy in the early 1990s, the home care allowance may have provided single mothers with the option to stay at home (Hiilamo, 2002). However, it is possible that the results are affected by third factors – for example, economic developments which could have been different in Sweden than in Finland. Overall, the number of women in employment between 1989 and 2002 decreased in Finland by 3.6 per cent, while the corresponding decrease in Sweden was 1.9 per cent. Changes in the tax benefit system were slightly more favourable for employed single parents in Sweden than in Finland (Hiilamo, 2002: 321–3). These factors may also contribute to the fact that development in the labour market was more negative towards single mothers in Finland than in Sweden during the 1990s. However, as stated above, this analysis is restricted by the available data and does not allow us to draw any firm conclusions. In the absence of longitudinal data, it is very difficult to disentangle the institutional effects from third factors. The present results, however, indicate that there could be a link between differences in the institutional features of family policies and labour market outcomes. The position of single parents in egalitarian Sweden was more favourable than in neofamilial Finland in the early 2000s. Consequently, one could assume that the economic consequences of partnership dissolution are not as grave in Sweden as they are in Finland. However, this conclusion should be explored further using more powerful data than those that were available for the present analysis.

CONCLUSIONS In this chapter, I explored whether there are differences between the institutional orientations and underlying ideologies in Sweden and Finland

Economic consequences of partnership dissolution in Sweden and Finland 151

with regard to partnership dissolution, and if so, whether they will result in different economic consequences of partnership dissolution. Swedish family policy, divorce law and family ideology encourage parents to share care and financial responsibility for their children in the event of a partnership dissolution, thereby mitigating the negative effects of such an event – especially the unequal distribution of the consequences. On the other hand, Finland emphasizes parental choice between home care and day care, thereby creating disincentives to work for mothers with small children. The analysis of labour market and income data indicates – as assumed – that the economic consequences of partnership dissolution are generally not as harsh in Sweden as they are in Finland. This especially seems to be the case for single parents with small children. It is important to bear in mind that while analysing commonalities and divergences in demographic characteristics, family law and family policies, the emphasis in this chapter was placed on divergences. However, this does not challenge the understanding that there are crucial differences between the two countries with regard to the degree of familialism. The results of this study reveal the need for more detailed country studies that go beyond generalized typological approaches (see also Stier, in this volume). However, from a European perspective, the economic consequences of partnership dissolution seem to be mild in both Sweden and Finland.

NOTES 1. However, the Ministry of Justice in Finland has issued a proposal for recommendations to the courts and the social welfare boards. In Sweden, family law also includes recommendations on how to assess the level of maintenance. 2. In 1990, the Swedish tax system still included four tax deductions related to family status: the deduction for a spouse (deduction from tax), the deduction for reduced tax-paying ability on the basis of the number of children in the family (deduction from income), the tax deduction for the person liable to provide maintenance (deduction from income) and the tax deduction for single parents (deduction from tax). As a part of the major tax reform in Sweden, the tax deduction for single parents was abolished at the beginning of 1991 but was compensated for in May 1992 by the introduction of a temporary special allowance for single parents (SOU, 1995:26: 61). These special tax-free allowances were paid retroactively from the beginning of 1991 to the end of June 1993. In principle, the most important tax deduction related to family status was the spouse deduction (hemmamakereduktion), which was not directly related to children, but to the spouse. This deduction, made directly from tax, was granted to couples in which only one of the spouses had an earned income. The deduction was abolished without any compensation in the major tax reform of 1991 (Troedsson, 1999: 62–3). This move abolished the last remnants of joint taxation and made labour force participation by both spouses a more profitable alternative. The other remaining deduction was the deduction for reduced tax-paying ability (where the sum remaining after taxes was not enough to ensure the minimum livelihood of the family). This deduction was also abolished in the 1991 tax reform on the grounds that such families still had the possibility to apply for social assistance to pay taxes

152

Single country studies

(Troedsson, 1999: 63–4). In the same reform, the tax deduction for maintenance payments (alimony, and so on) was also abolished. The government compensated for this by reducing certain maintenance liabilities (SOU, 1995:26: 57). 3. The figure is calculated as a share of single mothers’ average earnings as of fathers’ average earnings in dual-earner couples with children in the same age group.

REFERENCES Anttonen, A. and J. Sipilä (2000), Suomalaista sosiaalipolitiikkaa (About Finnish Social Policy), Tampere: Vastapaino. Bernhardt, E. (2003), ‘Key family issues in Sweden’, in Key Family Issues in the EU Member States: Summary Reports, European Observatory on the Social Situation, Demography and Family. Björnberg, U. (2001), ‘Cohabitation and marriage in Sweden – does family form matter?’, International Journal of Law, Policy and the Family, 15 (3), 350–62. Bradshaw, J. and A. Hatland (eds) (2006), Social Policy, Employment and Family Change in Comparative Perspective, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Castles, F.G. (1978), The Social Democratic Image of Society. A Study of the Achievements and Origins of Scandinavian Social Democracy in Comparative Perspective, London: Routledge and Kegan Paul. Duncan, S. (1995), ‘Theorizing European gender systems’, Journal of European Social Policy, 5 (4), 263–84. Esping-Andersen, G. (1990), The Three Worlds of Welfare Capitalism, Cambridge and Princeton, NJ: Polity Press and Princeton University Press. Finch, N. (2006), ‘Family change’, in J. Bradshaw and A. Hatland (eds), Social Policy, Employment and Family Change in Comparative Perspective, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Flora, P. (ed.) (1986), Growth to Limits: The Western European Welfare States Since World War II, Berlin: De Gryters. Fraser, N. (1997), ‘After the family wage: gender equality and social welfare’, Political Theory, 22 (4), 591–618. Gähler, M. (1997), ‘Ekonomiska konsekvenser av skilsmässor – skillnader mellan kvinnor och män’ (‘Economic consequences of divorce – differences between women and men’), in G. Ahrne and I. Persson (eds), Familj, makt och jämställdhet, SOU 1997:138, Stockholm: Fritzes, pp. 224–61. Gähler, M. (1998), Life after Divorce. Economic, Social and Psychological Well-being among Swedish Adults and Children following Family Dissolution, Dissertation Series No. 32, Stockholm: Swedish Institute for Social Research. Gauthier, A.H. (1996), The State and the Family. A Comparative Analysis of Family Policies in Industrialized Countries, Oxford: Clarendon Press. Gornick, J., M. Meyers and K. Ross (1998), ‘Public policies and the employment of mothers: a cross-national study’, Social Science Quarterly, 79 (March), 35–54. Haataja, A. (2006), Pohjoismainen ansaitsija-hoivaajamalli. Ruotsin ja Suomen perhevapaajärjestelmän vertailu (The Nordic Earner–Carer Model), Helsinki: Sosiaali- ja terveysministeriön selvityksiä 2006:43. Haataja, A. and A. Nyberg (2005), Did the Dual-earner Model Become Stronger or Weaker in Finland and Sweden in the 1990s? Luxembourg: LIS 414. Haataja, A. and A. Nyberg (2006), ‘Diverging paths? The dual-earner/dual-carer

Economic consequences of partnership dissolution in Sweden and Finland 153 model in Finland and Sweden in the 1990s’, in A.L. Ellingsaeter and A. Leira (eds), Politicizing Parenthood in Scandinavia. Gender Relations in Welfare States, Bristol: Policy Press. Härkönen, J. (2005), ‘Divorce risk factors across Finnish marriage cohorts, 1954– 1989’, Yearbook of Population Research in Finland, 41, 151–64. Hiilamo, H. (2002), The Rise and Fall of Nordic Family Policy? Historical Development and Changes during the 1990s in Sweden and Finland, Helsinki: Stakes. Hiilamo, H. and O. Kangas (2006), Trap for Women or Freedom to Choose? Political Frames in the Making of Child Home Care Allowance in Finland and Sweden, Turku: University of Turku A:15/2006. Hinnfors, J. (1992) Familjepolitik. Samhällsförändringar och partistrategier 1960– 1990, Göteborg Studies in Politics 26, Stockholm: Almqvist and Wiksell International. Jalovaara, M. (2003), ‘The joint effects of marriage partners’ socioeconomic positions on the risk of divorce’, Demography, 40 (1), 67–81. Kamerman, S. and A. Kahn (eds) (1978), Family Policy: Government and Families in Fourteen Countries, New York: Columbia University Press. Kamerman, S. and A. Kahn (1991), Child Care, Parental Leave and the Under-3s. Policy Innovation in Europe, New York: Auburn House. Katzenstein, P.J. (1985), Small States in World Markets. Industrial Policy in Europe, Ithaca, NY: Cornell University Press. Kaufmann, F.-X. (1993), ‘Familienpolitik in Europa’ (‘Family policy in Europe’), in 40 Jahre Familienpolitik in der Bundesrepublik Deutschland, Bundesministerium für Familie und Senioren, Neuwied: Luchterhand. Kaurala, M. (1993), Yksinhuoltajat ja elatusvelvolliset – taloudellinen asema ja tukivaihtoehdot (Single Parents and Non-custodial Parents – Their Economic Position and Supports Available to Them), Helsinki: Sosiaali- ja terveysministeriön monisteita 1993:3. Kautto, M. (2000), Two of a Kind? Economic Crisis, Policy Responses and Welfare during the 1990s in Sweden and Finland, Forskningsraport från Kommittén Välfärdsbokslut, Statens Offentliga Utredningar, Stockholm: SOU 2000:83. Korpi, W. (2000), ‘Faces of inequality: gender, class and patterns of inequalities in different types of welfare states’, Social Politics, 7 (2), 127–91. Kosonen, P. (1998), Pohjoismaiset mallit murroksessa (Changing Times for the Nordic Models), Tampere: Vastapaino. Kvist, J. (1998), New Perspectives in Comparative Social Policy, Odense: Odense University and the Danish National Institute of Social Research. Lewis, J. and B. Hobson (1997), ‘Introduction’, in J. Lewis (ed.), Lone Mothers in European Welfare Regimes: Shifting Policy Logics, London: Jessica Kingsley. Liu, G. and A. Vikat (2004), Does Divorce Risk Depend on Spouses’ Relative Income? A Register-Based Study of First Marriages in Sweden in 1981–1998, Rostock: MPIDP Working Paper 20004-010. Luxembourg Income Study (LIS) (2006), ‘Luxembourg Income Study’, www.lisproject.org/keyfigures/povertytable.htm, accessed 27 February 2007. Mahon, R. (2002), ‘Childcare: towards what kind of “Social Europe”’, Social Politics: International Studien in Gender, State and Society, 9 (3), 343–79. Mayhew, E. (2006), ‘The parental employment context’, in J. Bradshaw and A. Hatland (eds), Social Policy, Employment and Family Change in Comparative Perspective, Cheltenham, UK and Northampton, MA, USA: Edward Elgar.

154

Single country studies

Montanari, I. (2000), ‘From family wage to marriage subsidy: controversy and consensus in the development of family support’, Journal of European Social Policy, 10 (4), 207–33. Nordic Social Statistical Committee (NOSOSCO) (2005), Social Protection in the Nordic Countries 2003, Copenhagen: Nordic Social Statistical Committee. Niilola, K. and M. Valtakari (2006), Kotitalousvähennys kotipalvelumarkkinoiden luojana ja työllisyyden lisääjänä (How the Allowance for Domestic Help has Contributed to Markets for Domestic Help and Employment), Helsinki: Työpoliittinen tutkimus 310. Työministeriö. Nyberg, A. (2002), ‘Gender, (de)commodification, (in)dependence, and autonomous households: the case of Sweden’, Critical Social Policy, 22 (1), 72–95. Organisation for Economic Co-operation and Development (OECD) (2005), Babies and Bosses: Reconciling Work and Family Life, Paris: OECD. Pfenning, A. and T. Bahle (eds) (2000), Families and Family Policies in Europe. Comparative Perspectives, Frankfurt am Main: Peter Lang. Sakaranaho, T. (1992) ‘Lasten kotihoidon tuki. Tutkimus tuen kehitysvaiheista ja lakisääteistämisestä suomalaisen perhepolitiikan osana’, unpublished Master’s thesis, Department of Social Policy, University of Tampere. Sipilä, J. (ed.) (1997), Social Care Services: The Key to the Scandinavian Welfare Model, Aldershot: Avebury. Social Ministry (2000), ‘Social ministry’, www.sos.se/FULLTEXT/0000-008/0000008.htm, accessed 28 January 2007. Statens Offentliga Utredningar (SOU) (1995:26), Underhållsbidrag och bidragsförskott. Betänkande av Underhållsbidrags- och bidragsförskottsutredningen –93 (Social Assistance and Maintenance Support. A Government Report by the Research Committee on Social Assistance and Maintenance Support), Stockholm: Statens Offentliga Utredningar. Statens Offentliga Utredningar (SOU) (2001:24), Ur fattigdomsfällan. Slutbetänkande från Kommittén Familjeutredningen (Out of the Poverty Trap. A Final Report from the Government Commission on Family Affairs), Stockholm: Statens Offentliga Utredningar. Sundström, M. (1994), More Children and More Paid Work: Birth-Leave-Work Strategies of Swedish Women in the 1980s, Stockholm University: Stockholm Research Reports on Demography 82. Takala, P. (2000), Lastenhoito ja sen julkinen tuki, Tutkimuksia 110, Helsinki: Stakes ja Kela. Teachman, J.E. (2002), ‘Stability across cohorts in divorce risk factors’, Demography, 39 (2), 331–51. Tilastokeskus (2005), Väestö 2005:80, Helsinki: Tilastokeskus. Troedsson, I. (1999), Den kommenderade familjen. 30 år med Alva Myrdals familjepolitik (The Commanding Family Policy: 30 Years of Alva Myrdal’s Family Policy), Stockholm: Timbro. Uunk, W.J.G. (2004), ‘The economic consequences of divorce for women in the European Union: the impact of welfare state arrangements’, European Journal of Population, 20, 251–85. Valkama, E. and M. Litmala (2006), Lasten huoltoriidat käräjäoikeuksissa (Child Custody Disputes in District Courts), Helsinki: Oikeuspoliittinen tutkimuslaitos. Wennemo, I. (1994), Sharing the Cost of Children. Studies on the Development of Family Support in the OECD Countries, Swedish Institute for Social Research No. 25, Edsbruk: Akademitryck.

6.

Great Britain: ‘things can only get better . . .’ Wendy Sigle-Rushton

INTRODUCTION As is the case in many other industrialized countries, the divorce rate in Great Britain1 has increased markedly over the second half of the twentieth century. There had previously been short-lived increases linked to policy changes (such as the availability of legal aid in 1950), but the divorce rate did not begin to follow a consistent and steep upward trend until the early 1960s (Smith, 1997). The crude divorce rate increased substantially from 1960 to 1980. It subsequently stabilized and decreased (but only slightly) after 1995 (Eurostat, 2004a). At the same time that divorce rates were rising, the evolution of the economy and the expansion of the service sector meant that women had increasing access to jobs that offered flexible hours and did not require physical strength. Among the working-age population, the female participation rate increased from 47 per cent in 1961 to 55 per cent in 1971 (Hakim, 1980: 559). Since then, the rate for women aged 16–59 in Great Britain has climbed to around 70 per cent – well above the Lisbon targets for female employment (DWP, 2006). The coincident rise in divorce and female employment might seem to suggest that as divorce became more common, women were increasingly better prepared for its financial consequences. This is probably more the case for childless women than it is for mothers. Despite their increased presence in the labour market, British mothers continue to reconcile work and family by engaging in part-time work, and often only after their children have entered school. Although the gender pay gap for full-time workers is not extremely wide, part-time workers in Britain face a good deal of occupational segregation and an extremely large part-time pay penalty. For example, in 2003–04, about one-quarter of all female part-time workers in Great Britain were shop assistants, care assistants or cleaners. In contrast, only 4 per cent of part-time (versus 15 per cent of full-time) working women were managers or senior officials (Manning and Petrongolo, 2004). Moreover, research 155

156

Single country studies

shows that a past spell of part-time work is linked to lower pay even after the person returns to full-time work (Francesconi and Gosling, 2005). Hence, the choices that women with children (and other caregiving responsibilities) make early on might lead to considerable economic difficulties years later should they divorce. This chapter considers the British context as it relates to the economic consequences of divorce. We will begin by providing an overview of demographic and labour market trends since the 1970s. Next, we will consider the British approach to family law and family policy. Recent developments in family policy are something of a break from the past and indicate a growing recognition of the need for public support for families. The effects of these changes will not be captured in most empirical studies that rely on data collected during the 1990s, but the overall effect of policy changes is likely a matter of extreme. Conditions have probably got better. But consistent with a historical reluctance to interfere in the functioning of the labour market, the government has tended to rely on exhortation as opposed to regulation (Crompton et al., 2005). Labour market conditions, combined with a fairly weak family policy regime, all point to the persistence of high poverty risks for women following a divorce.

RECENT DEMOGRAPHIC AND LABOUR MARKET TRENDS2 Similar to its European counterparts, Great Britain has experienced a great deal of family and social change, progressing through what some have termed the Second Demographic Transition (Lesthaeghe, 1995). In the first stage of this transition, birth rates begin to decline, age at marriage begins to increase and divorce rates begin to rise. Next, cohabitation becomes more common, as does childbearing within cohabiting unions. In the third stage, divorce rates begin to plateau, post-marital cohabitation increases, and fertility plateaus with an older age profile. In this section, we outline the extent and timing of some of these changes as they have occurred in Great Britain – many of which have taken place coincidentally and are arguably interrelated with one another and with changes in divorce patterns. We begin with a discussion of trends in partnership formation, paying particular attention to changes in the popularity of marriage. We then move on to discuss two trends that are related to these changes: declining fertility and changes in family structure. We conclude with an overview of the British labour market and rates of poverty, with a special focus on the issue of child poverty.

Great Britain: ‘things can only get better . . .’

157

Marriage and Consensual Unions Between 1960 and 1970, marriage was popular, and the crude marriage rate in the United Kingdom was increasing. This trend subsequently reversed, and between 1970 and 2000, as divorce rates were largely climbing, the crude marriage rate fell from 7.7 to 5.1 per thousand midyear population. Part of the decline is due to the fact that marriage has been increasingly postponed. Median age at marriage has progressively increased from below 25 for both men and women in the early 1960s to over 30 in recent years (Bynner et al., 2002; Murphy and Wang, 1999; ONS, 2005). Over this same period, the proportion of first marriages among wives fell steadily, from 90.7 to 70.6 per cent. The latter is due, in part, to an increase in the number of people who are likely to remarry because of rising divorce rates. Ermisch and Francesconi (2000) suggest that changes in marriage patterns should not be interpreted as a rejection of marriage but as a postponement of it, and that the increasing popularity of consensual unions is a contributory factor. Empirical evidence provides some support for their position. Consensual unions have become increasingly common, especially among people who have never been married. Between 1976 and 1998, the proportion of unmarried women who were cohabiting in Great Britain increased from 9 per cent to 29 per cent (Haskey, 2001). In 1968, only 1 in 20 British women had cohabited prior to marriage. By 1990, this number had increased to 12 out of 20 (Bynner et al., 2002). Although at the moment, cohabitation is most common at younger ages and is often a precursor to (first) marriage, projections for England and Wales suggest that among the unmarried, the proportion living in consensual unions will continue to increase across all age groups (Haskey, 2001). These changes in the propensity to cohabit mean that the percentage of women who are partnered has declined, but only slightly, over the past few decades. However, the slight decrease in the percentage of women who have partners is due to the fact that marriage is being foregone to some extent as well. The proportion of females who were never married was only 8 per cent for the cohort of women born in 1955 and increased to 14 per cent for the cohort of women born in 1960 (Eurostat, 2004a). Fertility At the same time that changes in divorce, marriage, and cohabitation have been taking place, the total fertility rate (TFR) in Great Britain has been falling. In 2004 the TFR was 1.74 children per woman in Great Britain. Although higher than the EU-25 average of 1.5, the TFR is still well

158

Single country studies

below the replacement rate, and has been since the early 1970s (Eurostat, 2004a). Of course, the TFR can be misleading when childbearing is being postponed, and this does appear to be the case in Britain (high teenage pregnancy rates notwithstanding). Nonetheless, cohort measures of completed fertility, although less recent, tell a similar story. The cohort of women born in the UK in 1945 had a completed fertility rate that was just above replacement level. Those born five years later had just over two children per woman, on average. The cohort of women born in 1963 – some of whom may still have more children – have only had 1.90 children per woman (Eurostat, 2004a). One possible reason for declining fertility is the increasing popularity of consensual unions. Despite the fact that the percentage of births outside of marriage has been increasing since 1960 and is higher in the United Kingdom (42.3 per cent, see Data Appendix) than the EU-25 average (31.6 per cent) (Eurostat, 2004a, 2005), fertility rates are lower in consensual unions than in marriages. This is due both to lower rates of conception in consensual unions and to the fact that many of those living in consensual unions still choose to marry when they have a child. Single Parents Changes in patterns of family formation and dissolution have led to changes in the structure of families and households in Great Britain. Since the mid-1980s, the proportion of all households headed by a single parent in Great Britain has remained relatively stable at around 9–10 per cent. This figure includes all single-parent families – those with dependent and/ or non-dependent children. Between 1971 and 1991, the percentage of single-parent households with dependent children doubled, from 3 per cent to 6 per cent. The proportion then remained stable up until 2002 (ONS, 2003), and is currently higher than in most other European countries – second only to Sweden in the EU-15 (Eurostat, 2004b). Changing the unit of analysis from families to children, data from 2002 suggest that one-fifth of all dependent children in Great Britain live in single-parent households, a proportion which has doubled since 1981 (ONS, 2003). Only one in seven solo mothers report having never lived with or been married to the father of their children, and very few lone mothers are widows (McKay 2002). This means that partnership breakdown is the predominant pathway into solo parenthood, and that increasing divorce rates are responsible for much of the increase in single-parent families. Taken together, the evidence just presented implies that a considerable proportion of British children experience the divorce of their parents and the economic consequences that follow.

Great Britain: ‘things can only get better . . .’

159

The British Labour Market The British labour market is characterized by high employment rates for both men and women and – relative to the rest of Europe – fairly low levels of regulation. Unemployment rates in Great Britain have halved from 10 per cent in the mid-1980s to around 5 per cent between 2001 and 2005, and long-term unemployment has fallen by more than 80 per cent since 1997. The proportion of children living in workless households has also fallen by 2.6 per cent since 1997 (Sutherland, 2005). Nonetheless, these figures mask a good deal of heterogeneity within the British population. Chronic unemployment remains a problem for some individuals, and figures from 2004 show that two-fifths of long-term unemployed persons who enter work find themselves out of work again within six months. More than a quarter of temporary employees would like a permanent job but cannot find one. The employment rates of women have been rising over time, but employment rates and rates of growth differ substantially by level of education, ethnicity and family structure. Mothers with small children have far lower participation and employment rates than other women. For single mothers, employment rates are even lower, but they have been increasing in recent years. For example, the employment rate of single mothers was just over 55 per cent in 2004, up from 44 per cent in 1996 (Sutherland, 2005). This percentage is also small relative to the general population of women aged 25–49 in the United Kingdom, where 72 per cent report their main activity status as employed or self-employed (Eurostat, 2004b). Only the Netherlands has such a small percentage of single mothers in paid employment. Although the headcount rates for employment are high and suggest some level of gender equality, they obscure some important aspects of the British labour market as they relate to the economic independence of women, and of mothers in particular. Despite high employment rates, many women – and a small but increasing number of men – work part-time, and many women work fairly short part-time hours, often less than 16 hours per week. Although rates of employment among British mothers with children under the age of 6 rose from 43 to 56 per cent between 1990 and 2000, most of that increase was in part-time employment; increases in full-time equivalent employment were much more modest (Harkness, 2005). This has important consequences for gender inequalities in earnings in both the short and long term, and has important consequences for the economic well-being of women and children following a divorce. Relative to other European countries, part-time jobs are more occupationally and vertically segregated (Francesconi and Gosling, 2005; Manning and Petrongolo, 2004). As a consequence, many women (and some men) trade flexible

160

Single country studies

hours for low-paid and low-skilled jobs. Of those working-age adults who earned less than £6.50 per hour in 2004, nearly two-thirds were women and half were part-time workers (Palmer et al., 2004a, 2004b).3 This has led to concern that the segregated full-time/part-time labour market has resulted in an inefficient utilization of skills because a high percentage of part-time workers believe themselves to be working below their potential (Darton and Hurrell, 2005; EOC, 2005). In addition, individuals who experience a spell of part-time work and then later return to full-time employment face wage penalties well into the future, suggesting that part-time work – even when it is a temporary strategy – substantially affects a person’s long-term position in the labour market (Francesconi and Gosling, 2005). This pay penalty affects women more than men because women usually shift to parttime work in the middle of their working lives when they have to reconcile work with childcare and elder care responsibilities (EOC, 2005). This life cycle-related pay penalty is an important reason for women’s inferior economic position following divorce, particularly in the case of women who have had children (and/or taken on caregiving responsibilities) in the past. Compared with the regimes in other Western European countries, the British government is reluctant to intervene in the labour market. There is fairly strong equal opportunity legislation, but the substantive effects for those whose work patterns deviate from the standard male model are somewhat limited. For example, the Part-time Workers’ (Prevention of Less Favourable Treatment) Regulations, which aim to secure parity between part-time and full-time workers in terms of holiday entitlements, sick pay and pensions, were introduced in 2000. Their potentially positive impact is weakened by the degree of occupational and job segregation that exists between full-time and part-time workers, which makes it difficult to find comparable workers in order to make a case for unfair or discriminatory treatment. And as long as the labour market remains highly segregated, women’s access to well-paying jobs will be limited unless they are willing and able to work full-time. The labour market increases the economic risks of divorce for those women who are unable or unwilling to do so. While many part-time workers are poorly paid and unable to exercise their full potential, concerns have also been raised about the fact that many full-time employees – particularly fathers – are stressed as a result of increased work intensification and long hours (Hurrell and Davies, 2005). Despite having some of the longest working hours in Europe, Great Britain is not achieving the same levels of productivity as other European economies like France and Germany (HM Treasury, 2004). About two-thirds of all employees report regularly working overtime (both paid and unpaid), and the most common reason for doing so is work pressure (Stevens et

Great Britain: ‘things can only get better . . .’

161

al., 2004). Even if it is not related to productivity gains, research shows that working overtime is related to men’s wage growth and job retention (Harkness, 2005) – so in the British labour market, it is often an unstated requirement. Unfortunately, this work culture of long hours and high intensity creates an obstacle to decreasing the horizontal and vertical segregation of part-time or other flexible work opportunities. Moreover, it reinforces a gendered division of paid and unpaid work, because men’s time is taken up with long hours of work, which, in turn, shifts caregiving responsibilities onto the female partner. Such incentives for specialization leave neither partner in a position to assume both the economic and caregiving responsibilities for children in the event of a divorce or separation (EOC, 2005). Poverty Using a relative poverty rate and the most common threshold for low income (60 per cent of median income),4 data show that in the late 1990s, the United Kingdom had one of the highest rates of poverty among more economically developed countries. In terms of child poverty, it looked even worse. But this was not always the case. Relative poverty rates, particularly among households containing children, began increasing in the 1980s. In 1979, about 12 per cent of children were living in households with low incomes, and by 1996/97, this figure had risen to 25 per cent. One reason for this was the increase in divorce and single-mother families and, somewhat related, the increased risk for children of living in workless households (Sutherland, 2005). Of course, most studies of the economic consequences of divorce utilize data that were collected during these years of high poverty and high rates of joblessness in single-parent families. Although lone-parent families remain extremely vulnerable, policy changes that have taken place since the late 1990s have reduced poverty and have likely affected the economic position of those divorced families that have dependent children. In 2003/04, an estimated 12 million people in Great Britain were living in households whose income was below the 60 per cent threshold. This figure is 2 million less than it was in 1996/97, but it remains much higher than it was in the early 1980s. Similar to the overall trend, the percentage of children living in low-income households has been falling. In 2003/04, about 3.5 million children were living in low-income households – a drop of 0.75 million since 1996/97. In contrast, the percentage of working-age, childless adults living in low-income households has remained relatively stable and even gone up slightly (Palmer et al., 2004a, 2004b). The different situations of children and childless adults can be explained by changes in the social

162

Single country studies

security system which are targeted at reducing child poverty but do not put much focus on poverty more generally. Despite recent declines in the number of children living in low-income households, children are still 1.5 times more likely to live in low-income homes than adults are. Almost 2 million children live in workless households, and this is an important risk factor for poverty – although high levels of income inequality, particularly important for women facing part-time pay penalties, mean that work is not always sufficient to lift families above the low-income threshold. Children living in single-parent households, many of whom have experienced parental divorce, remain at a greater risk of living in poverty. Estimates for 2004/05 suggest that about 22 per cent of children in lone-parent families are living in low-income households, compared to only 13 per cent of children living in couple households with children (Sutherland, 2005). However, children living with couples are more likely to be poor than children who live with solo parents in many other European countries, suggesting that minimal government support leaves many families economically vulnerable (Rainwater and Smeeding, 2004).

THE FAMILY IN BRITISH FAMILY LAW AND FAMILY POLICY In this section, we describe the way in which British law and policy approach family issues. Historically, families have been expected to make choices and bear the financial (and emotional) costs of those decisions privately. When relationships break down and the government is not required to provide income support, there continues to be a strong emphasis on negotiation and mutual consent. Families are encouraged to sort out custody issues and the division of property among themselves. Moreover, when the government has sought to intervene to a greater extent – as in the case of child support – the more interventionist approach has been far from successful. Family Law Although no-fault divorce was implemented in Scotland in 2005,5 the law for England and Wales continues to be fault-based and, for the most part, based on the Matrimonial Causes Act (1973) which allows for divorce solely on the grounds of ‘irretrievable breakdown’. Irretrievable breakdown is evidenced when one of the following five facts is established as ‘proof’: adultery of the other spouse; unreasonable behaviour of the other

Great Britain: ‘things can only get better . . .’

163

spouse; desertion after two years; separation with mutual consent after two years; separation with no consent after five years. Despite a general reliance on negotiated divorce settlements, when an agreement cannot be reached, the division of marital assets is left to the discretion of the court. During the 1980s and 1990s, discretionary approaches focused predominantly on the needs and reasonable requirements of the spouse with caregiving responsibilities. For families with low levels of assets, this approach generally favoured the parent who cared for the children, most often the female partner. In contrast, for wealthier families – those with more assets than required to meet their needs – a larger share (typically two-thirds) tended to be awarded to the male spouse. Although the court’s decision regarding division of assets was discretionary and therefore uncertain, this rule of thumb likely increased the bargaining power of husbands and affected the division of assets even among those families who reached their own agreements (Dnes, 1998).6 It was not until 2000 that the House of Lords ruled that the starting point should be an equal division of assets. This was the same year that pension-sharing rules were introduced, so for most families analysed in the mid to late 1990s, women in wealthier families were likely to be discriminated against in the division of marital assets following a divorce. The maintenance aspects of the Divorce Reform Act of 1969 and the Matrimonial Causes Act of 1973 were formulated at a time when marriage rates were extremely high, and it was implicitly assumed that remarriage would follow divorce. The assumption of (rapid) remarriage meant that less attention was directed towards support for ex-wives and biological children. Related to this was an emphasis on the need for a ‘clean break’ which would help adults to move from a bad relationship to a good one with fewer remaining financial and emotional ties (Smart, 1999). However, the provision of social security support for first wives, which developed during the 1970s as a way of supporting clean breaks, became increasingly expensive because divorced women either did not remarry, or because they remarried but subsequently divorced. The proportion of single mothers dependent on state benefits rose from 20 per cent in 1961 to 66 per cent in 1987 (Bradshaw and Millar, 1991). In the late 1980s and early 1990s, the government responded to the increasing costs of supporting single-mother families and began to redefine parental obligations with an emphasis on tracing absent fathers and recovering maintenance. This policy development deviated in important ways from the pre-existing approach to divorce and child maintenance agreements. In 1991, a Conservative government passed the Child Support Act, which required all non-resident fathers to support economically their biological children without reference to their current living arrangements

164

Single country studies

and obligations. This was contrary to the previously implicit assumption that men should support the families with whom they currently live. The law was meant to shift responsibility for supporting single-mother families from the state to individual men, increasing both the percentage of fathers providing support and the amount that they provided. The law was implemented retroactively, so that men who believed they had had a ‘clean break’ divorce were suddenly at risk of being required to provide maintenance. The discretion applied by the courts with respect to maintenance decisions was replaced by a complicated formula which took fathers’ living costs into account using social assistance levels. This meant that for many fathers, child support obligations were likely to increase by large amounts. Had they been effective, these policy changes would have had little impact on the economic situation of low-income families following a divorce, because for those mothers who received state benefits, child support payments went not to them but to the Treasury. The net effect of these changes would be to lower the economic well-being of divorced men with children without increasing the economic well-being of low-income divorced women with children to any great extent. But child support policies have been far from effective. The Child Support Agency (CSA) went into operation in 1993 and has faced a series of problems, failures and criticisms from the outset (Blaiklock, 2004; Lewis, 2002). Recently, the Department of Work and Pensions carried out an extensive review of the CSA, hoping to improve collection outcomes and address its £1 billion arrears (Batty, 2005). The result of this review, the 2006 Henshaw Report, called for a less complicated regime that provides opportunities and incentives for parents to negotiate their own maintenance arrangements. The report suggests that the state should only intervene when that is not possible (Henshaw, 2006). In other words, current calls are for a return to the traditional liberal approach characterized by mutual consent with low levels of public regulation and interference, and, if the results are consistent with past outcomes, this policy will probably result in fairly low levels of child maintenance and work to protect the incomes of fathers after they divorce. Family Policy and Anti-poverty Measures In the United Kingdom, family policy has historically been minimal, reflecting the country’s liberal, individualistic and market-oriented ideology. Taxes tended to be individualized (although there was some tax relief for married couples in the past), and for married couples, issues of fertility and childcare were considered private matters (O’Donoghue and Sutherland, 1999; Randall, 1999). On the other hand, the United Kingdom

Great Britain: ‘things can only get better . . .’

165

(along with the Netherlands) expected solo mothers to remain at home and care for their children full-time, and provided low levels of income support so that they could do so (Millar, 1996). But changes in recent years have reflected a marked move towards greater governmental support for families with children, especially when those families are engaged in (low) paid work. Many of the consequent policy developments – which could be characterized as an attempted move to a ‘citizen worker model’ – have affected the circumstances of families with children, and especially of single-parent families (Drenth et al., 1999). Although much of the government’s support for families with children continues to be limited to those on low incomes, increased rights and levels of support have been extended to all families with children. Some of the more important aspects of recent policy reforms and developments include improvements to maternity and parental leave entitlements, including the right of parents with small children to request part-time hours; above-inflation increases in child benefits (counterbalanced for some families by the removal of the lone-parent supplement); work-activating changes to the tax system; and improved government support for childcare. Although these changes took place so recently that their effects will not be reflected in many estimates of the economic consequences of divorce, we can nonetheless speculate about the impact they are likely to have. Maternity leave Statutory maternity leave has existed since the 1970s, but it was originally only available to women who had been continuously employed with the same employer for at least two years (five years for women working short part-time hours). A large percentage of working women failed to meet this requirement. Entitlement was only extended to all working mothers in the early 1990s, but even so, there was a two-tiered system of Ordinary Maternity Leave (OML) and Additional Maternity Leave (AML). Until recently, all women employees were entitled to 26 weeks of OML, and those who had completed 26 weeks continuous employment with their present employer ending with the fifteenth week before the baby was due were eligible for a further 26 weeks of AML. For babies due on or after 1 April 2007, maternity leave was extended to one year for all women, regardless of the length of service. At the same time, entitlement to Statutory Maternity Pay was extended to 39 weeks. Parental leave Parental leave – an unpaid, individual entitlement consisting of 13 weeks per parent per child – was first introduced in December 1999. Leave can be taken in blocks or in multiples of one week up until the child is 5 years old,

166

Single country studies

but parents are not permitted to take more than four weeks of leave per calendar year. All employees with at least one year of continuous employment with their present employer who have – or expect to have – parental responsibility for a child are eligible. Consistent with its reluctance to impose strong regulations on the labour market or to interfere in ‘private matters’, the government’s Summary Guidance states that ‘wherever possible employers and employees should make their own agreements about how parental leave will work in a particular workplace’ (BERR, 2008). Moreover, if employers ‘cannot cope’, they are permitted to postpone a period of leave for up to six months. Right to request part-time hours Legislation was also introduced in 2003 which gives parents of children under the age of 6 (under 18 if the child is disabled) a Right to Request to Work Flexibly in their existing jobs. From April 2007, this right to request was extended to those with caregiving responsibilities for adults as well. This is a potentially positive measure, given the considerable segregation between full-time and part-time work. Research suggests that thus far the majority of requests for flexibility have been granted (Palmer, 2004; see also Holt and Grainger, 2005), but other national evidence suggests the practice is quite limited, and only those who think that their requests will be granted ask for consideration in the first place (Manning and Petrongolo, 2004). Likewise, local labour market studies suggest that flexible working arrangements are not widespread. Some firms have extended this right to all employees (CIPD, 2005), but since the scheme remains discretionary, the decisions that firms make are likely to vary with the state of the labour market. Furthermore, qualitative studies have reported that, in practice, the operation of flexibility on a day-to-day basis rests on decisions made by line managers irrespective of formal company policy (Perrons, 1999; Standing, 2006; Yeandle et al., 2002). Child Benefit The Child Benefit is a universal cash payment made available to all families with children under the age of 16 and, since April 2006, to families with ‘qualifying young persons’ aged 16, 17, 18 and 19. In order to qualify, a young person must meet education or training requirements. The payment is made directly to the main care provider, generally the mother. A higher rate of Child Benefit is paid for the eldest child in a family. From April 2008, the benefit provides £18.80 per week for the eldest child and £12.55 per week for each additional child. Compared with the amounts paid in other EU-15 countries, this level of benefit places Britain in the lower-middle range of the ranking for different amounts of

Great Britain: ‘things can only get better . . .’

167

household incomes, just above the four Southern European countries (de Henau et al., 2004). In addition to the Child Benefit, for all children born since September 2002, there is now a Child Trust Fund into which the government pays £250 at the time of birth (£500 for poor children). Additional funds are added when the child is 7 years old. These savings (along with additional contributions by families or friends up to an annual limit) are tax-free, and the funds can be accessed when the child reaches the age of 18. Taxation changes In the late 1990s, the government introduced a series of targeted tax reforms aimed at providing additional support to families with children. Key policies include an integrated refundable tax credit for out-of-work families and low-paid working families (with a supplement in the first year of a child’s life) and additional benefits for parents who work 16 hours or more per week (the Working Tax Credit). The effect of these policies has been to make the tax system increasingly less individualized, to increase directly the incomes of poor families with children and to increase the number of households eligible for benefits (Sutherland, 2005). More than three times as many working-age adults in the UK received tax credits in 2004/05 as those who received what was then called the Family Credit in the early 1990s (New Policy Institute, n.d.). While changes to the tax system have improved the financial situation of many families, the impact of these policy changes on gender equality among low-income couples with children has been criticized. Because eligibility for tax credits is calculated at the household level, they may work to discourage second earners – usually mothers – from entering the labour market and building up work experience. Hence, tax credits may work to reinforce the male breadwinner model and incentivize decisions which would exacerbate the economic risks faced by mothers should they eventually divorce. Improved access to childcare In May 1998, the government launched a National Childcare Strategy. The strategy aims ‘to ensure good quality, affordable childcare for children aged 0 to 14 in every neighbourhood, including both formal childcare and support for informal arrangements’ (DfEE, 1998: 7). Although the plan has been criticized for providing too few places to children under 3 years old (Rake, 2001), the number of childcare places has increased considerably. In addition, there has been an expansion in free early education for all 4-year-olds and an increasing number of 3-year-olds. All 4-year-olds are guaranteed a free early education place if their parents want one, but the hours provided are part-time, and the minimum provision amounts

168

Single country studies

to just 12.5 hours a week (Rake, 2001). For many dual-earner families, taking advantage of free early childcare means that they must rely on a complex mix of formal, informal and backup support, both for care and for transportation between care sites. Logistical requirements of this sort prevent some mothers from taking up paid work. The situation is particularly difficult for single mothers who are less able to rely on assistance from fathers and for families with more than one child (Skinner, 2003). Anti-poverty measures In order to address high rates of relative poverty and child poverty, the government has, in recent years, implemented a wide range of new measures and reformed some aspects of the welfare state. As Holly Sutherland suggests, taken together, the piecemeal reforms can best be summarized as attempts to reduce poverty ‘without preventing . . . labour market improvements’ (Sutherland, 2005: 12). Some of the more important measures include the introduction of a national minimum wage as well as the introduction of the ‘New Deal’ employment activation programmes. Overall, the government has increased the amount of support that it offers to lowpaid families (with a stronger emphasis on families with children), but it has been more radical in its reforms to the tax system and more cautious in its efforts to impose additional regulations on the labour market. The first legislation for a national minimum wage was passed in 1998. It took effect in April 1999, and has worked to raise the wages of the lowest paid – the majority of whom were women and part-time workers. Hence, this legislation should have an effect on the economic position of singleparent families. Concerns about the costs it would impose on businesses and the effect it might have on unemployment rates meant that the main rate was originally set at a fairly low £3.60 per hour (with a lower youth rate of £3.00 per hour). However, it has been raised in each subsequent year, and as of October 2007 stands at £5.52 per hour (with a lower youth rate of £4.60 per hour for 18–21-year-olds, and an even lower development rate of £3.40 per hour for 16–17-year-olds). Despite periodic increases, the rate remains relatively low, and because it is nationally set, it is well below a ‘living wage’ in more expensive areas like London. At the same time that it introduced a national minimum wage, the government implemented a series of ‘New Deal’ work activation programmes. Different programmes were made available to different segments of the population (younger workers, older workers, the long-term unemployed), including two groups comprised mostly of women: partners of the longterm unemployed and single mothers. Single parents of children under 16 who receive benefits are now required to attend information meetings once a year (more often if the child is older than 14) to discuss the possibility

Great Britain: ‘things can only get better . . .’

169

of finding work and to receive help and assistance should the lone parent choose to do so. However, there is no requirement that the single parent take up work while he or she has a child under the age of 16. These measures were meant to inform those single mothers who were not employed of the new and increased benefits of entering the work force. Relative to New Deal programmes with high levels of male participation, funding for schemes aimed at lone parents and partners of the unemployed is fairly limited, and only about half as much is spent per head (Rake, 2001). Finally, in order to address the persistently high rates of poverty among childless adults, the Working Tax Credit, which had originally been made available only to families with a dependent child, was subsequently made available to couples without children who are low paid and work more than 30 hours per week. Reductions in social security contributions (without a decrease in social security entitlements) were also introduced for all low-paid workers. The Likely Impact of Recent Policy Changes Taken together, recent policy changes should lessen the economic consequences of divorce for women with children. The increased eligibility for maternity leave – if linked to greater labour market attachment and earnings growth – should protect mothers against the economic consequences of divorce, although women may still face part-time pay penalties (and other disadvantages associated with low labour market regulation) if they request and are granted part-time hours. Employment incentives, a minimum wage, and the increased availability of childcare – particularly for lower-income families – should work to increase the number of working mothers after (and to a lesser extent before) a relationship dissolves. Greater economic support for children in general should improve the economic circumstances of single-parent families headed by women, whether or not they work. Although women and children will still bear most of the economic costs of divorce, recent policy changes should work to lessen the burden to some extent. Recent attempts to address the poverty of workingage adults without children should provide a buffer against the economic consequences of divorce for childless adults as well.

CONSEQUENCES OF PARTNERSHIP DISSOLUTION Empirical analyses of longitudinal data suggest that – as in many other countries – women in Great Britain experience a substantial drop in household income following a divorce, particularly when they have

170

Single country studies

children (Jarvis and Jenkins, 1999). Moreover, their short-term average and median changes in income are large relative to those experienced by British men who divorce. Indeed, some studies find that on average, British men experience a modest increase in household income subsequent to divorce (Jarvis and Jenkins, 1999; Andreß et al., in this volume). Estimates from 1991–94 suggest that women’s household income declines by 18 per cent after a divorce, while men’s income increases by about 2 per cent. Although methods and measures differ across studies, making comparability difficult, estimates from the second half of the 1990s suggest that declines in household income for women may have grown even larger as the decade progressed (Dewilde, 2002; Uunk, 2004).7 These studies estimate income declines as high as 27 to 36 per cent for women who divorce. Although similar results have been found for many other industrialized countries, evidence also suggests that compared with other countries where welfare provision is more generous, the short-term economic consequences in Britain are relatively severe (Dewilde, 2002; Uunk, 2004; Andreß et al., in this volume). Given that during the 1990s, women had weak positions in the labour market, weak bargaining power to obtain a fair share of marital assets, little recourse to maintenance from a former husband and received little or no child support when there were children involved (Marsh et al., 2001), large changes in household income relative to other countries are not surprising. Some authors have suggested that men face a good deal of economic disadvantage subsequent to divorce as well, but that a reliance on monetary indicators like earned income tends to mask the short-term deprivation that they experience. Evidence from an analysis with welfare regimes as the unit of analysis suggests that for liberal countries (data for the UK and Ireland are combined), this may be the case. When monetary indicators are used, the changes in well-being for divorced men are small and insignificant. In contrast, when a relative deprivation index is used, although women continue to fare worse, there is evidence that men also face high levels of deprivation following a divorce. The authors point to the fact that women tend to remain in the family home, at least when there are children involved. They suggest that the gender gap in access to good quality housing and household durables – at least in the shorter term – may work to favour women. This may not be reflected well in purely monetary measures but will show up in questions regarding housing quality (Aassve et al., in this volume). Although different measures have the potential to provide a more nuanced picture of changes in men’s economic status after a divorce, it remains true that regardless of the deprivation measure used, women in liberal countries still tend to fare worse both relative to their ex-partners and relative to women who divorce in other countries. The

Great Britain: ‘things can only get better . . .’

171

gender gap in overall well-being narrows because men fare worse on some (less commonly used) measures than others. The situation for women is remarkably stable across different measures; therefore, the conclusion that women bear a disproportionate share of the economic costs of divorce is not overturned. Few authors have examined the economic consequences of divorce in Great Britain or any other country over the longer term, and most existing studies have focused on families headed by single mothers (whether due to divorce or childbearing outside a partnership) rather than on the divorced population. Evidence from an American study of solo parents suggests that one of the most important determinants of economic recovery for single mothers over the longer term is re-partnering (Page and Stevens, 2004). This work suggests that children in families that divorce experience a 50 per cent drop in household income after the divorce. For those who do not remarry, after six years household income is still 45 per cent lower than that of two-parent families. In contrast, among the remarried, income is about the same (insignificantly higher). Similarly, Jenkins and Schluter (2003) show that in Great Britain, most lone-parent families that increase their labour supply (and hence increase their chances of exiting poverty) do so by adding a new partner to the household. Given women’s inferior labour market and economic position, the incentives for re-partnering in Great Britain are probably fairly strong. Analyses of British data show that 70 per cent of previously cohabiting people and 43 per cent of previously married people find another partner within five years of a relationship breakdown. Rates of re-partnering are slower for those who are older or have custody of a child, however, and women with these characteristics are likely to have had the largest declines in household income (Ermisch, 2002). Comparative evidence on rates of re-partnering is scarce, so it is difficult to confirm that these figures are higher or lower relative to other countries. One of the few comparative studies focusing on the longer-term economic consequences of divorce (not single motherhood) also presents some evidence that higher rates of re-partnering may be responsible for what appears to be the relatively fast economic recovery of divorced women in Great Britain (Andreß et al., in this volume). But cross-national differences in re-partnering and its economic consequences is an area that requires further exploration and research (see the literature discussed in Andreß et al., in this volume). Although comparative research provides valuable information about variations in risk and vulnerability, making sense of the long-term economic consequences of divorce – and doing so using a comparative perspective – can be particularly challenging, not least because data are often only available for specific periods of time, and the prevailing contextual

172

Single country studies

factors are likely to differ dramatically across countries. For example, when interpreting the relative speed with which British women recover their predivorce incomes, it is very important to take into account that most studies of divorce in Britain use data collected during the 1990s. Using data from this time period means that recovery of pre-divorce income (measured and deflated for example in year t 2 1, where t is the year of divorce) will tend to overstate the speediness of economic recovery following divorce, at least relative to many other countries. Between 1991 and 1998, Britain went through a period of strong economic growth. Median real adjusted household income rose by 11 per cent over that time period. In contrast, over the same time period, median adjusted incomes in West Germany (and many other European countries) remained stagnant (Jenkins and Schluter, 2003). So pre-divorce income is a more meaningful measure for German women than it is for British women – who, because of the country’s strong economic growth, were chasing a fast-moving target. Put another way, a woman who had a median household income in 1991 may have got back to that level of real income in 1998, but in 1998 that same level of income would now be below the median. This is an important challenge facing researchers who want to examine economic consequences over the longer term and who want to use a comparative perspective to interpret their findings. The economic climate in the countries being compared may make comparisons of economic recovery based on measures of real income difficult to interpret and may require the use of several measures to get a complete picture of differences across countries. Income measured as a percentage of the relative poverty line might yield substantively different results than (inflation-adjusted) real incomes.

CONCLUSION Rates of divorce in England and Wales are among the highest in Europe, and Haskey (1996) estimates that if divorce rates continue at around the mid-1990s level, two in five marriages will eventually end in divorce, and many of these families will include young children. Data from 2004 show that one-fifth of all children affected by divorce were under 5 years old, and just under two-thirds were under 11. These patterns are important because evidence suggests that women and children bear a disproportionate burden of the economic costs of separation. Traditionally, Great Britain has been liberal in its approach to welfare and individualistic in its approach to family matters (both during and subsequent to a relationship breakdown). Relative to other Western European countries, family matters tended to be relegated to the private sphere, while

Great Britain: ‘things can only get better . . .’

173

attempts to deal with their consequences were generally left to the market. Although some important changes have occurred in recent years, most of the empirical evidence on the economic consequences of divorce pre-date those potentially important policy changes, and the earlier policy context is therefore relevant for understanding the findings of existing studies. Although the earlier policy context is important for interpreting the findings of existing studies, we should not forget that important changes have occurred. The national minimum wage, the expansion of childcare, more generous tax credits and increased rights to request work flexibility should all have worked to improve the situation, at least for those women who are able to enter the labour market. Women’s economic circumstances after a divorce have probably improved. Even so, British family and antipoverty policies still do little to address the gender inequalities that result from a relationship breakdown. For women, the economic costs of divorce are likely to be lower, but remain great, especially when there are children involved. Improved but limited family policy means that mothers can expect little compensation for the costs of raising children. Additionally, weak child support policy means limited financial support from the nonresident father. As a consequence, the only feasible route to economic recovery for women subsequent to divorce is via their own or a new partner’s labour market participation. But because the British labour market is highly unregulated and continues to assume a male model of work, it fails to acknowledge or accommodate caregiving responsibilities. This reality makes the reconciliation of work and caregiving difficult, and for some women – despite recent policy changes – re-partnering may still be (and continue to be) the most effective route to economic security.

NOTES 1. When reading this report, it is important to note that Great Britain is not equivalent to, but part of the United Kingdom. Great Britain comprises the contiguous regions of England, Wales and Scotland. Its political system is relatively complex and unique in that, to varying degrees, different regions have different legislative powers. Since 1998, both Scotland and Wales have enjoyed (different) devolved powers. Scotland, in particular, is responsible for a good deal of its own legislation. Moreover, statistical information is sometimes collected and reported for the United Kingdom as a whole (covering Northern Ireland and Great Britain), sometimes just for Great Britain, and – in an increasing number of cases – for England and Wales, Scotland, and Northern Ireland separately. Whenever possible the information presented in this report will be for Great Britain, although sometimes only information for the UK or part(s) of it is available. When this is the case, it will be made explicit in the text. 2. See Sigle-Rushton (2008), where some of the information in this and the following section was first presented and discussed. 3. Low pay is also strongly associated with qualification levels. For example, more than half of those with no qualifications earn less than £6.50 per hour.

174

Single country studies

4. In 2003/04, before deducting housing costs, this equated to £200 per week for a couple with no children, £122 for a single person, £291 for a couple with two children and £214 for a single parent with two children (Palmer et al., 2004a). 5. In 2005, the Scottish Executive passed a divorce reform; thus, its divorce law now differs from that in England and Wales. This bill reduces the mandatory separation period from two years to one where there is mutual consent and from five years to two when the divorce is contested. In addition, desertion is eliminated as a grounds for divorce. It also extends the legal regime to cohabiting couples, granting legal protection in the event of partnership breakdown or death of the partner. 6. For Scotland, the Family Law Act of 1985 states that the starting point should be an equal division of assets. Therefore, the one-third rule of thumb applies to those families divorcing in England and Wales. 7. These latter studies did not consider the economic position of men following a divorce.

REFERENCES Batty, D. (2005), ‘Crackdown on absent dads unfair without CSA reform’, Guardian, 7 November, available at www.guardian.co.uk/society/2005/nov/07/ childrensservices.politics2, last accessed 17 October 2008. Blaiklock, O. (2004), ‘A broken backed agency: the impending collapse of the Child Support Agency’, submission made to the Department of Work and Pensions Select Committee, November. Bradshaw, J. and J. Millar (1991), Lone Parent Families in the UK, DHSS Social Security Research Report No. 6, London: HMSO. Bynner, J., P. Elias, A. McKnight, H. Pan and G. Pierre (2002), Young People’s Changing Routes to Independence, York: Joseph Rowntree Foundation. Chartered Institute of Personnel and Development (CIPD) (2005), Flexible Working: The Implementation Challenge, London: CIPD. Available at www. cipd.co.uk/subjects/hrpract/flexibleworkingpractices/_flxwrkimpchlg.htm, last accessed 17 October 2008. Crompton, R., M. Brockmann and C. Lyonette (2005), ‘Attitudes, women’s employment and the domestic division of labour: a cross-national analysis in two waves’, Work, Employment and Society, 19 (2), 213–33. Darton, D. and K. Hurrell (2005), People Working Part-Time Below Their Potential, Manchester: Equal Opportunities Commission. Department for Business, Enterprise and Regulatory Reform (BERR) (2008), ‘Summary guidance’, available at www.berr.gov.uk/whatwedo/employment/ workandfamilies/parental-leave/guidance-summary/page21300.html, last accessed 17 October 2008. Department of Work and Pensions (DWP) (2006), Opportunity for All, Eighth Annual Report, Norwich: HMSO. Dewilde, C. (2002), ‘The financial consequences of relationship dissolution for women in Western Europe’, in E. Ruspini and A. Dale (eds), The Gender Dimension of Social Change. The Contribution of Dynamic Research to the Study of Women’s Life Courses, Bristol: Policy Press, pp. 81–100. Dnes, A.W. (1998), ‘The division of marital assets following a divorce’, Journal of Law and Society, 25 (3), 336–64. Drenth, A.v., T. Knijn and J. Lewis (1999), ‘Sources of income for lone mother families: policy changes in Britain and the Netherlands and the experiences of divorced women’, Journal of Social Policy, 28 (4), 619–41.

Great Britain: ‘things can only get better . . .’

175

Equal Opportunities Commission (EOC) (2005), Britain’s Hidden Brain Drain – Final Report, Manchester: Equal Opportunities Commission. Ermisch, J. (2002), ‘Trying again: repartnering after dissolution of a union’, ISER Working Paper 2002-19, Institute for Social and Economic Research. Ermish, J. and M. Francesconi (2000), ‘Cohabitation in Great Britain: not for long but here to stay’, Journal of the Royal Statistical Society, Series A, 163, 153–72. Eurostat (2004a), Population Statistics 2004, Luxembourg: European Commission. Eurostat (2004b), ‘Household formation in the EU – lone parents’, Statistics in Focus, Luxembourg: European Commission. Eurostat (2005), ‘EU25 population is up 0.5 per cent in 2004’, Eurostat news release 136/2005, Luxembourg: European Commission. Francesconi, M. and A. Gosling (2005), Career Paths of Part-Time Workers, EOC Working Paper 19, Manchester: Equal Opportunities Commission. Hakim, C. (1980), ‘Census reports as documentary evidence: the census commentaries, 1801–1951’, Sociological Review, 28, 551–80. Harkness, S. (2005), Employment, Work Patterns, and Unpaid Work: An Analysis of Trends Since the 1970s, summary report for the Economic and Social Research Council. Haskey, J. (1996), ‘The proportion of married couples who divorce: past patterns and current prospects’, Population Trends, 83, 25–36. Haskey, J. (2001), ‘Cohabitation in Great Britain: past, present and future trends – and attitudes’, Population Trends, 103, 4–25. Henau, J. de, D. Meulders, S. O’Dorchai and H. Périvier (2004), ‘The relative generosity of the EU-15 member states’ child policies’, final report of the EC Project ‘The Rationale of Motherhood Choices: Influence of Employment Conditions and of Public Policies’, Brussels: European Commission, pp. 43–97, available at http://ec.europa.eu/research/social-sciences/pdf/finalreport/hpse-ct-2001-00096final-report.pdf, last accessed 17 October 2008. Henshaw, D. (2006), Recovering Child Support: Routes to Responsibility, London: The Stationery Office (TSO), available at www.dwp.gov.uk/childmaintenance/ pdfs/Henshaw_complete22_7.pdf, last accessed 17 October 2008. HM Treasury (2004), Opportunity for All: The Strength to Take Long-Term Decisions for Britain (pre-Budget report, December), London: The Stationery Office. Holt, H. and H. Grainger (2005), ‘Results of the second flexible working employee survey’, Employment Relations Research Series No. 39, London: Department of Trade and Industry. Hurrell, K. and K. Davies (2005), Time Use and Childcare Survey, Manchester: EOC. Jarvis, S. and S.P. Jenkins (1999), ‘Marital splits and income changes: evidence from the British Household Panel Study’, Population Studies, 53 (2), 237–54. Jenkins, S.P. and C. Schluter (2003), ‘Why are child poverty rates higher in Britain than in Germany? A longitudinal perspective’, Journal of Human Resources, 38 (2, special issue on Cross National Comparative Research Using Panel Studies), 441–65. Lesthaeghe, R. (1995), ‘The Second Demographic Transition – an interpretation’, in K.O. Mason and J. An-Magrit (eds), Gender and Family Change in Industrial Countries, Oxford: Clarendon Press, pp. 17–62. Lewis, J. (2002), ‘The problem of fathers: policy and behavior in Britain’, in B.

176

Single country studies

Hobson (ed.), Making Men Into Fathers, Cambridge: Cambridge University Press. Manning, A. and B. Petrongolo (2004), The Part-Time Pay Penalty, London: Department of Trade and Industry and Women and Employment Unit. Marsh, A., S. McKay, A. Smith and A. Stephenson (2001), Low Income Families in Britain, DSS Research Report No. 138, London: Corporate Document Services. McKay, S. (2002), ‘The dynamics of lone parenthood, employment and poverty in Great Britain’, www.benefits.org.uk/loneparentdynamics.pdf, last accessed 17 October 2008. Millar, J. (1996), ‘Mothers, workers, wives: comparing policy approaches to supporting lone mothers’, in E.B. Silva (ed.), Good Enough Mothering? London: Routledge. Murphy, M. and D. Wang (1999), ‘Forecasting British families into the 21st century’, in S. McRea (ed.), Changing Britain, Families and Households in the 1990s, Oxford: Oxford University Press, pp. 100–138. New Policy Institute (n.d.), ‘The poverty site: summary key facts’, available at www. poverty.org.uk/summary/key%20facts.shtml, last accessed 17 October 2008. O’Donoghue, C. and H. Sutherland (1999), ‘Accounting for the family in European income tax systems’, Cambridge Journal of Economics, 23, 565–98. Office for National Statistics (ONS) (2003), Social Trends 33, Norwich: The Stationery Office. Office for National Statistics (ONS) (2005), ‘Report: divorces in England and Wales during 2004’, Population Trends, 121, 85–9. Page, M.E and A.H. Stevens (2004), ‘The economic consequences of absent parents’, Journal of Human Resources, 39 (1), 80–107. Palmer, G., J. Carr and P. Kenway (2004a), Monitoring Poverty and Social Exclusion 2004, York: Joseph Rowntree Foundation. Palmer, G., J. Carr and P. Kenway (2004b), Monitoring Poverty and Social Exclusion in Scotland 2004, York: Joseph Rowntree Foundation. Palmer, T. (2004), ‘Results of the first flexible working employee survey’, Employment Relations Occasional Papers, London: Department of Trade and Industry. Perrons, D. (1999), ‘Flexible working patterns and equal opportunities in the European Union: conflict or compatibility?’, European Journal of Women’s Studies, 6 (4), 391–418. Rainwater, L. and T. Smeeding (2004), ‘Single-parent poverty, inequality, and the welfare state’, in D.P. Moynihan, T.M. Smeeding, and L. Rainwater (eds), The Future of the Family, New York: Russell Sage Foundation. Rake, K. (2001), ‘Gender and New Labour’s social policies’, Journal of Social Policy, 30 (2), 209–31. Randall, V. (1999), ‘Childcare policy in the European states: limits to convergence’, Journal of European Public Policy, 7, 346–68. Sigle-Rushton, W. (2008), ‘England and Wales: stable fertility and pronounced social status differences’, Demographic Research, 19 (15), 455–502. Skinner, C. (2003), Running Around in Circles: Coordinating Childcare, Education and Work, Bristol: Policy Press. Smart, C. (1999), ‘Divorce in England 1950–2000: A moral tale’, paper presented at the CAVA workshop on Frameworks for Understanding Policy Change and Culture, University of Leeds, 29 October.

Great Britain: ‘things can only get better . . .’

177

Smith, I. (1997), ‘Explaining the growth of divorce in Great Britain’, Scottish Journal of Political Economy, 44 (5), 519–44. Standing, K. (2006), ‘Swings and roundabouts: work, childcare and women’s career progression’, paper presented at the conference researching women’s experiences of work and family life, Liverpool, 24 May. Stevens, J., J. Brown and C. Lee (2004), The Second Work-Life Balance Study: Results from the Employees’ Survey, Working Paper Series No. 32, Manchester: EOC. Sutherland, H. (2005), ‘Can child poverty be abolished? Promises and policies in the UK’, paper presented at the Australian Social Policy Conference, University of New South Wales, Sydney, 20–22 July. Uunk, W. (2004), ‘The economic consequences of divorce for women in the European Union: the impact of welfare state arrangements’, European Journal of Population, 20, 251–85. Yeandle, S., R. Crompton, A. Wigfield and J. Dennett (2002), ‘Employed carers and family-friendly employment policies’, report for the Joseph Rowntree Foundation, Bristol: Policy Press.

7.

Marital disruption in Spain: class selectivity and deterioration of economic conditions1 Lluís Flaquer and Anna Garriga

INTRODUCTION When the 1981 Divorce Act was passed, the dissolution of marriages had been banned in Spain for 40 years. With the exception of the short-lived 1932 Divorce Act, which was in force for a few years under the II Republic due to the critical circumstances just before the outbreak of the Civil War (1936–39), for all intents and purposes, we can consider that the practice of divorce in modern Spain started in the early 1980s. Given that Spain was one of the last countries in Europe where divorce was enacted and that the number of divorces there had traditionally been quite low, one of the consequences of the 2005 legal reform, which implemented the so-called ‘fast-track’ procedure, was a huge rise in divorce rates. This has happened at a time when there is growing concern about the growth of social inequality and poverty levels in a country where the sheer figures of favourable macroeconomic performance in the last decade have not been matched by the proper redistribution of development profits. One of the most prominent characteristics of the divorce system in Spain is that it is one of the few countries in Europe where a recent rapid and drastic liberalization of access to divorce has not been matched by stronger regulation and effective government control of maintenance payments, nor by increased public provisions to mitigate the economic consequences of marital disruption (Björnberg, 2006). In Spain, there is indeed a clear-cut contradiction between very liberal divorce regulations and a rudimentary family policy. Low social spending – especially in the area of families and children – and limited development in terms of social services are important components that cannot be overlooked when we examine the consequences of partnership dissolution. In this chapter, we focus mainly on the evolution of levels of economic well-being among the Spanish divorced population during the last decade 178

Marital disruption in Spain

179

in a long-term perspective. Our other areas of interest are the impact of changes in the composition of the divorced population as well as the effects of developments resulting from legal and institutional reform. One of the issues we will explore is the extent to which the process of ‘divorce democratization’ which is at work in Spain is connected with increasingly negative consequences of divorce. Our thesis is that a critical ingredient in the consequences of divorce is the social characteristics of the population which experiences it. According to William Goode’s hypothesis (Goode, 1966), in a cross-national perspective, an increase in the rate of divorce is associated with the lessening of its selectivity in terms of class. While in countries with a brief divorce tradition, the upper and the middle strata of the population tend to dissolve their marriages more often than do the working classes, the diffusion of divorce brings about the opposite effect. If this is true (and in Spain there has been a definite trend towards decreasing class selectivity in divorce), the consequences could be more negative, at least in social terms. Another relevant focus of interest when we study the consequences of divorce is gender disparities. In principle, a gradual and late shift towards a dual-earner family model should be linked with more gender-balanced outcomes after divorce. In fact, the picture is more complex. In Spain, at an aggregate level, the rise in female labour market participation has long been a very strong driving force in marital disruption. Women’s employment facilitates female independence and creates a situation in which, in case of a failed and unsatisfactory relationship, separation becomes viable. The growth of marital disruption interacts with the decline in divorce selectivity we discussed previously, and with certain features in the structure of the Spanish labour market. Persistent high unemployment rates – especially for women – up until the end of the 1990s, together with a lack of measures for reconciling work and family are important elements that explain traditionally low female economic activity rates. Other factors, such as the prevalence of undeclared work among women and of fixed-term contracts and atypical and junk jobs, in conjunction with some other kinds of labour discrimination, contribute to considerable gender imbalances after divorce. Finally, while the rate of remarriage (and re-partnering) has been rather low in Spain in the past two decades, especially for women, one of the most popular strategies for alleviating the adverse consequences of partnership dissolution is to move in with relatives, particularly parents, after a divorce. Indeed, household complexity is one of the major family characteristics in Mediterranean countries, and, although it can undoubtedly be traced back to a practice intended to fight poverty in the context of a

180

Single country studies

rudimentary family policy, it is also connected with a long-standing tradition of the force of kith and kin, with the prevalence of strong familialistic values and possibly with low state legitimacy. In the title and in many contexts in this chapter, we use the term ‘marital disruption’ rather than ‘partnership dissolution’ in order to highlight the fact that in Spain, for many years, marital breakups tended to be connected with legal and even informal separation rather than with divorce. On the other hand, we have not been able to find much information about the dissolution of informal partnerships. This chapter comprises seven sections. The second is concerned with the institutional setting in which marital disruptions take place. In particular, we examine the structure and regulation of the labour market, the type of welfare regime, and the system of social protection. The third section deals with the legal regulation of separation and divorce and with the legislative reforms made in the period after Franco’s death in 1975. In the fourth section, we discuss the evolution of marital disruption within the framework of major socio-demographic trends. We pay particular attention to the process of diffusion of divorce, which can possibly be associated with the adoption of the 2005 Divorce Act. In the fifth section, we consider some factors underlying the economic consequences of partnership dissolution, focusing especially on the decline of divorce selectivity in terms of class and on the formation of complex households as an alternative to low remarriage rates and as a buffer against the severe impact of a potentially drastic drop in earnings. Finally, in the sixth section, we will assess the main empirical findings from our descriptive analysis by using 1994 European Community Household Panel (ECHP) and 2004 European Union Statistics on Income and Living Conditions (EU-SILC) survey data (see the Data Appendix for additional information on both data-sets). The seventh section concludes with a summary of the main results.

INSTITUTIONAL CONTEXT In this section, we provide an overview of the institutional setting in which divorce takes place, including the development of the welfare state and the evolution and characteristics of the labour market. One of the most prominent characteristics of the Spanish system of social protection is its low level of provisions for the well-being of families. Despite the fact that in the decades following the transition to democracy, Spain has experienced a fundamental political, economic and social transformation, this scant interest in family policy has remained practically unchanged. While this political stance can certainly be interpreted as a backlash against the social

Marital disruption in Spain

181

policy of the Francoist period, which was heavily centred on upholding the traditional family, the structure of the labour market in the 1980s and 1990s also contributed to maintaining this system. The Development of the Welfare State The Spanish welfare state has often been characterized as an example of a rudimentary welfare regime in the ‘Latin Rim’, where there is a sharp contradiction between constitutional rights to employment and welfare and the poor implementation of these rights in practice (Leibfried, 1992). In the same vein, Ferrera stressed as main features of the Southern model of welfare its fragmented system of income guarantees, its lack of an articulated network of minimum social protection, its deficit in ‘stateness’ and its leaning towards particularism and clientelism (Ferrera, 1996). Other authors have especially emphasized an ambivalent or implicit familialism as a crucial factor underlying the Mediterranean model, in the sense that although the caretaking function of the family is not actively supported, the family is in fact the primary caretaker (Flaquer, 2000; Leitner, 2003; Saraceno, 1994). Finally, Naldini proposed the notion of a ‘family/kinship solidarity’ model to characterize the Southern welfare regime, since in countries such as Spain or Italy, strong intergenerational ties and intense family solidarity along the life course are expected both within the household and between households (Naldini, 2003). The primary outline of the Spanish welfare system was developed under Franco’s regime following Bismarckian lines, and this influence is still visible in the current structure of Spain’s Social Security (Cousins, 1999). At a time when industrial workers were few, since most of the Spanish population was still living in rural areas, the male breadwinner family model was enjoined as the prescribed standard. One of the main hallmarks of social policy under the authoritarian regime was family cash benefits. After the transition to democracy, no debate took place on the subject of updating family policy in the context of the modernization of social policies. Family cash benefits were not discontinued, but they were simply frozen. Some researchers have suggested that the weakness of family policy in democratic Spain can be understood as a backlash in line with the rejection of authoritarian policy legacies in so far as family policy was viewed as one of the hallmarks of the Francoist regime. Beginning in 1975, all of the relevant political and social actors wanted to distance themselves from Franco’s pro-natalist and anti-feminist family policies, which featured heavily in the official discourse of the regime, and they therefore avoided policy-making in this field (Valiente, 1997). The long-term effect of this situation can still be felt today, over 30

182

Single country studies

years later: Spain is one of the lowest spenders on family policy among the developed countries. As we can see in Figure 7.1, the lowest share of social expenditure devoted to families and children is found in Spain, at 0.7 per cent of the GDP in contrast to the 2.1 per cent EU-25 average. This is a very important statistic to bear in mind when we discuss the economic consequences of divorce. As is also the case in other Mediterranean countries, the Spanish welfare state is much more oriented toward retired people than it is toward families and children (Ferrera, 1996; Flaquer, 2007). Clearly, divorce is not an event in the Spanish welfare state. We can expect that separated or divorced people – especially those with children – will be worse off after a divorce, since in contrast to widowhood, there is no provision for that contingency in the Spanish system of social protection (Flaquer et al., 2006). Contributory and non-contributory child benefits are means-tested and only meant for very poor families. However, middle-class families with dependent children receive much better fiscal advantages in the form of income tax credits, involving higher benefit levels and more generous conditions of eligibility. Spain not only has a low level of spending on child benefits but on family services as well. Measures intended to reconcile work and family developed late and are poorly articulated. Maternity leave is paid for 16 weeks at 100 per cent of the mother’s previous earnings; beginning in 2007, a new scheme provides fathers with a fully paid two-week paternity leave which may be used after the maternity leave. There is also an unpaid parental leave available for parents of children under 3, with very limited employment protection. Compulsory schooling in Spain starts at age 6. Attendance at early education and childcare services varies greatly according to age. While there is nearly full coverage available for children aged 3–6 years, enrolment in day care for children under age 3 is only about 21 per cent.2 Finally, Spain has no national system of social assistance. The safety net is very fragmented in nature because social assistance depends on regional government legislation, and the levels and coverage are quite dissimilar from one territory to another. The Structure and Evolution of the Labour Market One of the prominent features of the Spanish labour market was the persistence of high unemployment rates during the 1980s and 1990s.3 This situation has long been associated with a high degree of labour market segmentation according to gender and generation, which has created an unremitting duality of insiders and outsiders.

183

0

0.5

0.7 1

1 0.9 1.5

1.3 1.3 1.3 1.2 1.1 1.1

1.8 1.7 1.7 1.7 1.6

2

2 2 2

2.1

2.5

2.5 2.5 2.5

Expenditure on family and children as percentage of GDP (2004)

Eurostat.

Figure 7.1

Source:

Denmark Luxembourg Iceland Norway Sweden Finland Austria Germany Hungary France Ireland EU-25 Slovenia Cyprus Belgium Slovakia United Kingdom Greece Estonia Czech Republic Switzerland Netherlands Latvia Portugal Lithuania Italy Malta Poland Spain 3

3 3 3 3

3.2 3.1

3.5

4

3.9 3.8

4.5

184

Single country studies

The high unemployment rates mentioned above were caused by structural imbalances in the labour market due to the overlapping of industrial and post-industrial transitions (which resulted in a massive shedding of jobs in both rural and urban areas), coupled with the expansion of (especially female) labour force participation beginning in the 1980s. This shortage of employment was one of the main underlying factors leading to the rise of an insider/outsider labour market structure which set a core of over-protected middle-aged men (usually male breadwinners) in opposition to a large number of women and young people engaged in precarious and atypical employment. While male family heads were getting the major share of the welfare state benefits, young people and middle-aged women were bearing the brunt of unemployment (Cousins, 1999; Esping-Andersen, 1999).4 On the other hand, this employment deficit – which was a strong disincentive for women with low educational qualifications to become economically active – contributes to explaining the traditionally low rates of labour market participation among Spanish women.5 Additionally, it is possible that the low provisions for family services which we discussed in the previous section were associated with a weak demand for them due to an abundant availability of caregiving work – mainly from inactive women in the older (but also younger) generations. This trend was completely reversed in the early 2000s when, in the context of an improved employment situation, an increasing supply of low-paid jobs relative to the demand attracted a significant number of migrants to Spain from Northern Africa and South America, including a large number of female immigrants who provide domestic and caregiving services to families in a formal or informal way. A number of these immigrants work irregularly in the informal economy – often because they are illegal – despite a regularization process which took place in 2005. During the 2000s, the foreign population in Spain has increased rapidly, so that in 2007, it amounts to 4.5 million.6 About one-quarter of all foreign women work in household domestic services. Recent published data from Eurostat show Spanish employment rates to be similar to the EU average.7 Full-time employment has become the norm for women in age cohorts born in the late 1950s and onwards. In 44 per cent of all couples in which at least one partner is working, the man and the woman both work full time. However, there is also a very similar share of couples (43 per cent) in which only the man is working. This divide is connected with a reduced prevalence of part-time jobs for women (only 10 per cent of couples).8 Part-time employment is not very popular in Spain, and it is mainly involuntary because it is associated with precariousness, poor pay and low social protection. Part-time jobs are mainly occupied by women with older children who are trying to return to the labour market after a

Marital disruption in Spain

185

period of inactivity, but not so much by mothers with young children. This characteristic contributes to explaining the expansion of domestic service by female immigrants who provide child and elder care. A final remark should be made concerning the high occurrence of fixedterm contracts in Spain. One of the measures which were intended to combat unemployment beginning in the late 1980s was the spread of fixedterm contracts and casual flexible work. Spain is an extreme case in which fixed-term contracts affect about one-third of the workforce – particularly women and young people, and virtually all new entrants to the job market (Bentolila and Dolado, 1994; Cousins, 1999).9 In sum, this brief overview of the development of the labour market in Spain in relation to the consequences of divorce yields an ambivalent picture. On the one hand, the growth of female employment has undoubtedly reduced gender disparities between partners, and this means that women are in a better position to face a potential dissolution of their partnerships than they were before. However, this progress has been at the cost of a rise in new grounds for labour market discrimination against women and young people. Not only does labour market segmentation by gender and generation still continue (and has even expanded in certain areas), but quite recently, segregation by citizenship has come to the fore as a new form of discrimination in the labour market.

THE LEGAL REGULATION OF DIVORCE Since the onset of the Spanish democratic transition in 1975, the dissolution of marriage has been regulated twice: in 1981 and in 2005. We will discuss both of these Acts below. The first Act has been in force for more than two decades, and most of the data on marriage dissolution presented in this chapter have been generated under its sway. The 2005 Act set a new hallmark and is associated with an important turning point in divorce trends. The 1889 Civil Code recognized both canonical and civil marriage, although civil marriage was limited to the few non-Catholic couples. The marriage bond was exclusive and for life. Even if divorce did not exist, separation without remarriage was possible. Marriages could be annulled, however, if either or both of the partners alleged impotence – or, in the case of arranged marriages, lack of consent – especially if the marriage was not consummated (matrimonio rato y no consumado), in which case remarriage was allowed. Under the Second Republic (1931–39), the Civil Marriage Act (1932) and the Divorce Act (1932) provided for civil marriage that was valid for

186

Single country studies

all citizens including Catholics, allowing divorce by mutual consent, establishing equal rights between the sexes and between spouses, abolishing discrimination between legitimate and illegitimate children, decriminalizing adultery, regulating the right to abortion, and granting women the right to vote (Alberdi, 1979; Flaquer and Iglesias de Ussel, 1996; Iglesias de Ussel, 1998). After Franco’s victory, the old civil legislation was re-established so that marriage was made indissoluble once again. This situation lasted until the end of Franco’s dictatorship in 1975. The new 1978 constitution dealt with the regulation of marriage dissolution, but it did not mention any specific approach to the matter (Iglesias de Ussel, 1998). The 1981 Divorce Act The new civil regulation of marriage involved two different kinds of valid marriages: civil and religious. In order to take legal effect, a marriage, whether civil or religious, must be celebrated in accordance with the formalities laid down in the Civil Code. In order for a religious marriage to be valid, the religion must be registered, its formalities recognized by the Spanish state and the marriage concerned entered in the Registry. In this case, the marriage is blessed by a minister and the ceremony takes place only in church without any further civil formalities (Roca, 2002). The main characteristic of the regulation of marital dissolution in Spain in 1981 was that divorce was introduced – not as a substitute for separation; it was simply added to it as a further legal step. Judicial separations were very often a precursor to divorce. In nearly all cases, the granting of divorce required some kind of previous separation, either legal or simply de facto.10 However, the length of the required separation could vary, and in some cases had to be accompanied by matrimonial misconduct. This meant that in most cases, couples had to go through two separate court proceedings, since divorce required two different court appearances and two filings (Goode, 1993; Martín-Casals et al., 2003; Naldini, 2003; Roca, 2002). Civil separations could be granted by mutual consent ratified by a court after one year of marriage as well as by petition to the court based on certain grounds, but canonical separations were not recognized. A petition for a judicial separation could be brought on any of certain fault grounds11 or on the grounds of a marital breakdown resulting in more than six months of separation (when both spouses consented to it), or more than three years of separation regardless of consent. Furthermore, when it was difficult to identify one of the typified grounds, the courts used to refer quite often to the so-called ‘lack of affectio maritalis’ as a final ground for separation (Martín-Casals et al., 2003; Roca, 2002).

Marital disruption in Spain

187

A divorce could be sought either by mutual consent or in contentious proceedings. Most applications for divorce were brought on the basis of one12 or two years13 of separation and were uncontested. Finally, a divorce could be granted after five years of separation in the case of desertion or lack of matrimonial life for the same period (Roca, 2002). Even if judicial separation was originally conceived as a transitional status allowing for the reconciliation of spouses, in most cases it came to be seen as the waiting room for divorce – that is, as a passage of no return or as a one-way transitional stage (Barbagli and Saraceno, 1998). Nevertheless, while under the 1981 Divorce Act, most divorced people had been judicially separated, not all separated people would finally divorce, so that an important number of legal separations were never transformed into divorces. In fact, ever since the Divorce Act came into force, the number of separations had always been higher than the number of divorces.14 This indicates that a proportion of separated people did not get divorced straight away and did remain at the separation stage for a long period of time. In this way, separation became a status in itself. There are various reasons why people remained at the separation stage for longer than was necessary. First, some people wished to remain separated for religious motives. Considering that the doctrine of the Catholic Church accepts separation but not divorce, this system could allow Catholics to legalize their situation permanently without having recourse to divorce (Martín-Casals et al., 2003). Secondly, some people, especially women, could see certain advantages in remaining at the separation stage, since they retained certain spousal rights such as inheritance or widowhood pensions (Goode, 1993).15 Thirdly, taking into account that, in practice, only separated people who were contemplating a second union filed a petition for divorce, low remarriage prospects could be a considerable disincentive.16 Finally, divorce was a slow, long, expensive and cumbersome process, and the cost of the legal procedure itself could be an important deterrent (Alberdi et al., 1995). The 2005 Divorce Act In mid-2005 the Spanish Parliament passed a number of legislative measures aimed at reforming family law, including the regulation of homosexual marriage and the reform of divorce. The latter was mainly intended to eliminate the double procedure (first separation, then divorce) by giving partners the freedom to choose between these two options, to enhance their possibility of reaching an agreement and to foster their common exercise of parental responsibility through shared custody. The 2005 Divorce Act provides for a no-fault divorce with a very short

188

Single country studies

waiting period (three months after marriage),17 involving the elimination of previous legal separation as a pre-requisite for divorce and the abolition of specific grounds for divorce with the exception of physical, psychological or sexual violence (in which case the waiting period is not applicable). The new Act maintains separation as an option for those who do not wish to resort to divorce for personal reasons. The waiting period of three months is the same irrespective of whether the divorce (or legal separation) is initiated by only one of the spouses, at the request of both spouses or by one of them with the consent of the other. The new Act regulates consultation with teenage minors in divorce proceedings in order to ascertain their wishes and feelings as far as future arrangements are concerned, and it establishes mediation as a voluntary alternative method of conflict resolution, so that a contested procedure can be interrupted at any time in order to seek an agreement.18 One of the novelties of the 2005 Divorce Act is the regulation of joint custody. The custody of the children may either be granted to one spouse alone or shared between the two, but in the latter case (custodia compartida) some restrictions apply. According to the law, a court may grant joint custody only if it is agreed to by both parents. If both parties are opposed to joint custody, or in a case of gender-based violence, the judge must not order them to share custody. However, upon demand by one of the parties, a court may grant joint custody in the best interest of children. No official data exist as yet concerning the percentages of different types of custody that are ordered by the courts. Ancillary Matters The 2005 Divorce Act made no practical changes in the regulation of ancillary matters pertaining to divorce, such as child custody and maintenance. In this section, we discuss the provisions for ancillary matters that exist in the laws we have reviewed so far. In Spain, there is no administrative procedure for obtaining a divorce. It can only be granted through a judicial process. However, anyone whose financial resources are insufficient has the right to obtain legal aid without discrimination on the grounds of nationality.19 The legal consequences of divorce (or separation) depend upon whether the parties have or have not made an agreement regulating ancillary matters. Where there is no such agreement, it is up to the judge to issue orders in this respect (Roca, 2002). If the petition for divorce is filed by both spouses – or by one of them with the consent of the other – they must submit a proposal for an agreement on ancillary matters (convenio regulador). If only one spouse files the petition, he or she must also submit

Marital disruption in Spain

189

a preliminary outline for a proposal. The joint proposal must contain at least the following arrangements and provisions: (1) measures relating to the exercise of parental responsibility, the type of child custody and visiting arrangements; (2) the allocation of the use of the family dwelling; (3) the division of common property; and (4) the determination of child maintenance (pensión de alimentos) and alimony (pensión compensatoria), if applicable. Among the most important ancillary matters are the use of the matrimonial home and the setting of pensions for minor or dependent children and for ex-spouses. Except in cases where the spouses reach another agreement which is endorsed by the judge, use of the common dwelling is allocated to the family unit in which the children live. Spanish law gives judges discretionary powers to determine the amount of pensions, although they usually ratify private agreements made between the parties. There are no tables stipulating minimum amounts per child in connection with family income, and the legal guidelines on this matter are not very specific.20 In the case of unmarried partners, the determination of child maintenance follows similar criteria. Considering that in Spain there is no state law regulating consensual unions, the divorce law also applies in the case of breakdowns between unmarried couples with children.21 According to 2005 data, the father is required to pay for child maintenance in 49.3 per cent of separations (versus 1.3 per cent for mothers and 2.1 per cent for both parents).22 No data exist on parental compliance in making child maintenance payments. Nevertheless, some surveys suggest that a significant share of fathers are not meeting their financial obligations. Beginning in 2008, a scheme of advanced child maintenance payments has been set up in Spain. It is a means-tested measure, intended for lowincome families, with quite restrictive conditions of eligibility and monthly payments up to €100 per child. Finally, the legislation on divorce regulates the payment of alimony (pensión compensatoria) in such cases in which separation or divorce may bring about an economic imbalance between the spouses resulting in a worsening of one party’s previous financial situation. According to 2005 data, alimony to be paid by the husband was granted in 14.2 per cent of separations and in 8.2 per cent of divorces.23

THE EVOLUTION OF MARITAL DISRUPTION IN SPAIN In order to assess the economic consequences of divorce across countries, it is very important to take into account the characteristics and evolution

190

Single country studies

of divorce in each country. The evolution of marital disruption in Spain between 1981 and 2005 roughly corresponds with the onset of the second demographic transition. In the past 30 years, the evolution of nuptiality has been very dramatic. The crude marriage rate24 evolved from 7.6 in 1975 to 5.0 in 2005. In Spain, fertility rates started to fall in the mid-1970s and the drop was very dramatic shortly after the beginning of the democratic transition. Although the fertility rate shifted from 2.8 children per woman in 1975 to 1.6 in 1985, it was still above the European Union (EU) average during this period. The lowest point was reached in 1998 with a rate of 1.1 children per woman.25 Spain is one of the European countries with the highest child gap (gap between desired and real fertility; Esping-Andersen, 2002, 2007).26 Spain has a low extramarital fertility rate by European standards. However, in 2005, nearly one out of every four children was born to an unmarried mother.27 Considering that in Spain the share of unregistered fathers accounts for only 7.3 per cent of births outside of marriage (2002), most of these children are being born to mothers who live in partnerships (Flaquer et al., 2006). These couples may experience separation at even higher rates than couples who are married. Unfortunately, there are no official data on the number of court orders concerning child maintenance connected with the dissolution of informal partnerships. The increase in marital disruption in Spain can be easily seen if we look at the evolution of the absolute numbers of separations and divorces since the 1981 Divorce Act came into effect (Figure 7.2). The chart portrays the number of both separations and divorces, since it is important to recall that under the former legislation, separation was already a valid indicator of marital disruption because few people made petitions for a final divorce. We can observe that during the period from 1983 to 2004, the number of separations was much higher than the number of divorces (see the previous section). The exception in 1982 can be explained because in that year, couples who had been separated de facto for more than five years were allowed to divorce right away (Alberdi, 1986). Starting with modest figures of about 20 000 in the early 1980s, separation reached its peak in 2004 with more than 80 000 cases, amounting to a growth of about 300 per cent in the past 25 years. On the other hand, the increase in divorces during the same period was much more moderate, with an overall growth of about 140 per cent. Since 1985, crude divorce rates in Spain have experienced a gradual progression, and in mid-2000s they were already approaching EU average. An estimation for 2006 yields a rate of 2.8 divorces per 1000 population, and thus Spain is joining the leaders in the divorce league.28 From 2004, Spain seems to have been experiencing a process of intense

191

INE.

1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

Number of separations Number of divorces

Number of separations and divorces (1982–2006)

0

20 000

40 000

60 000

80 000

100 000

120 000

140 000

Figure 7.2

Source:

Absolute numbers

192

Single country studies

divorce diffusion. The extent to which this exponential growth proves to be the result of the adoption of the 2005 Divorce Act by the Spanish Parliament remains to be seen. As we discussed earlier, the new law involves a structural change in the nature of the Spanish legal system by heading towards no-fault and unilateral divorce. This radical turn has involved a shift from a conservative and restrictive legislation to one of the most liberal regulations in the world. It is reasonable to think that the new divorce law is acting as a driving force in divorce diffusion, but it is too early to draw conclusions on the matter. Literature shows that the legal reforms may be having an impact on divorce rates. Friedberg (1998) demonstrates that the no-fault divorce law reforms that began around 1970 in the USA actually contributed to the sharp increase in divorce rates observed over the past 30 years.29 Looking at data, there was an astounding increase of the number of divorces between 2004 and 2006, which is not fully compensated by the decrease in separations. While the number of separations dropped by 77 per cent, the number of divorces increased by 149 per cent. If we add up the number of separations and divorces, we find about 146 000 cases of marital disruption in 2006. On the other hand, it is important to emphasize the dramatic increase in the number of marriages dissolved before completion of the first year (330.6 per cent with respect to 2005; INE, 2007). One explanation for this sharp increase in divorce figures is that a portion of these divorces are the result of the validation of previous separations under the new law. In fact, we know that in 30.4 per cent of the divorces granted in 2006 there was a previous separation (INE, 2007). A second suggestion is that the growth in divorce numbers may be associated with the increase of an existing phenomenon – that is, the informal breakup of marriages to avoid the cost of court procedures. A third suggestion is that, as Friedberg (1998) shows, the increase in divorces may be a direct impact of divorce law reform per se, and therefore the divorce boom will not recede in the next few years. While the first two explanations only consider the short-term effects of the divorce law, the last explanation emphasizes the possible long-term consequences. Nevertheless, it is too early as yet to draw conclusions about the eventual outcomes of the new law.

FACTORS UNDERLYING THE ECONOMIC CONSEQUENCES OF DIVORCE In this section, we will discuss the effects of a number of factors on the economic well-being of individuals after divorce. Under this heading, we

Marital disruption in Spain

193

have grouped together two different kinds of factors: (1) the mediating effect of the population composition in terms of educational attainment and labour market participation, and (2) the types of households in which the divorced and separated individuals live, with a special emphasis on complex households, as well as patterns of remarriage as alternative strategies for alleviating the negative consequences of divorce. Literature on the predictors of divorce shows that they vary across countries and over time. But most literature on the effects of divorce does not combine research on the predictors of divorce and on its consequences. One of the reasons for this gap is that while demographers usually tend to focus on the predictors of divorce, sociologists are more interested in its outcomes. In this section, we suggest that in order to understand the economic consequences of divorce in Spain, we need to pay more attention to the characteristics of divorced people and to their evolution. We argue that these characteristics, whether before or after the event, have changed to a great extent in the past two decades. Goode’s hypothesis suggests that the relationship between class and marital instability depends on the extent to which divorce has become easy and therefore widespread (Goode, 1966). Where divorce is difficult and costly, it is primarily an upper-class privilege. In the early stages of Western modernization, with high legal, social and economic barriers to divorce, it was more common among the middle and upper classes. As these barriers gradually faded away, marital instability became more frequent among the lower classes. Consequently, in countries with low divorce rates, the relationship between class and divorce should be positive – divorce is more common among the upper and middle classes – and in countries with high divorce rates, the relationship should be negative: this means that divorce rates are higher among the working classes. For example, in countries such as the USA, marital dissolution rates are higher among the lower classes, but in Western countries some generations ago, these rates were higher among the upper classes (Goode, 1993; Smock, 1993). Divorce tends to be more selective in countries where the phenomenon is recent than in others where there is a long-standing tradition. Countries such as Spain and Italy, in accordance with their restrictive divorce regulations, had traditionally low divorce rates and hence the relationship between class and divorce had been positive. This relationship may be changing in Spain. Divorce rates have increased significantly during the past few years and, accordingly, the relationship between class and divorce has also been modified. Taking into account the diffusion of divorce in Spain during the past decade, we should expect that the social consequences of marital instability would be more detrimental today than they were before.

194

Single country studies

The Composition of the Divorced Population in Terms of Educational Attainment The economic and social consequences of partnership dissolution can be explained by a combination of individual factors (gender, parenting, loss of network contacts) and contextual effects related to the institutional setting in each country. Other relevant factors are the socio-demographic characteristics of divorced people, such as their educational attainment and labour market participation. It is well known that more educated people have more resources for coping with the disruption experience because they have less unemployment, can get better jobs before and after divorce and can therefore earn higher incomes. Greater human capital investments before and after divorce are likely to reduce its negative economic consequences, since these investments improve employment chances and earnings (Amato, 2000; Härkönen and Dronkers, 2006; Uunk, 2004). In Spain, shortly after the implementation of the 1981 Divorce Act, divorce was a very selective phenomenon. However, a comparison of data on the educational attainment of separated and divorced people in the population for the past two decades suggests that in Spain, the selectivity of marital disruption has dwindled as this has become a more common experience. During the 1980s, high levels of education used to correlate with a high proportion of divorces, but as divorce became more normalized this selective effect tended to disappear (Solsona et al., 2000). As Alberdi (1986) showed, in the early 1980s, few people were divorced and the percentage of people with tertiary education was more than three times higher among divorced people compared with the total population (2 per cent versus 7 per cent). As we can see from Figure 7.3, although a certain overrepresentation of people with tertiary education among the separated and divorced population was still visible during the 1990s and in the early 2000s, this trend was on the wane. Using data from the 1991 Spanish Socio-Demographic Survey, a number of researchers found that the importance of educational level among women decreases in the younger generations. For the cohorts who were married from 1961 until 1980, women with secondary and university degrees were more likely to experience marital disruption. However, for the youngest cohort – women who were married between 1981 and 1990 – the educational level was not significant anymore (Houle et al., 1998; Treviño, 2006). Similarly, in a different study using data from the 1995 Spanish Family and Fertility Survey, researchers found that educational attainment was not a significant variable. Women with higher levels of education were not more likely to experience divorce than women with lower levels of education (Simó and Solsona, 2002).

195

0

0.5

1

1.5

2

2.5

3

3.5

4

1981

1982

1991

2004

Ratio of the share of individuals with tertiary education among the separated and divorced population and the adult population (1981, 1982, 1991, 2004)

Own calculation using data from Alberdi (1986: 105) for 1981–82, Ruiz Becerril (1999: 19) for 1991 and EU-SILC for 2004.

Figure 7.3

Source:

Tertiary education (separated/divorced vs adult population)

196

Single country studies

Activity and Employment Rates of Separated and Divorced Individuals In Spain, the level of female participation in the labour market appears to be correlated with educational attainment. We know that in most countries, labour market participation is very different for men and women and that women’s education is one of the most prominent factors explaining this gap. While the economic activity rates for males and females with higher educational attainment are broadly the same, less-educated women exhibit much lower activity rates. High unemployment rates – which in Spain have been affecting women more severely than men for some time – may also contribute to explaining this gap (see the subsection ‘The structure and evolution of the labour market’ in this chapter). Especially for women, labour market participation can be considered to be both a predictor of divorce and a protection against its most adverse economic consequences. Descriptive analyses of the Spanish population by sex, marital status and participation in the labour market have shown that separated and divorced women have much higher activity and employment rates than married ones (Ruiz Becerril, 1999; Flaquer, 2004; Flaquer et al., 2006). Having some experience on the labour market and having job security are factors positively associated with divorce, but marital breakup is also a factor that induces women to enter the labour market (Solsona et al., 2000). Simó and Solsona (2002) found that for 1966–99 Spanish union cohorts, women’s activity increased their probability of experiencing marital disruption, whereas for men being unemployed had a positive relationship with marital breakup. In countries such as Spain, where women’s labour market participation has traditionally been low, the rise in female activity rates appears correlated with the growth of divorce rates (Flaquer, 2007). At an aggregate level, there are considerable similarities between trends in the rates of increase for marital disruption and for female labour market participation. Since 1982 in particular, the standardized graphs for the rates of separation by mutual agreement and for activity rates of women aged 25–54 follow very similar patterns. However, since the early 2000s, marital disruption appears to be growing at a greater pace than female activity rates. In fact, a correlation analysis of disruption and female activity rate indicators for the Spanish regions shows that the positive relationship between the two factors tends to lose force during this period (Flaquer et al., 2006). Remarriage and Household Complexity Living with relatives or re-partnering can be considered as alternative strategies for alleviating the negative consequences of divorce. One of the

Marital disruption in Spain

197

consequences of the rise in divorce rates is the existence of a number of marriageable partners who are willing to enter second unions. However, as Goode noticed a long time ago (1966), remarriage is an institution that mainly develops in countries with a long-standing divorce tradition. A high rate of marital turnover is only possible when divorce rates are also high. For a long time, this was not the case in Spain, and formal second unions were not very frequent.30 Figure 7.4 plots the number of remarriages among divorced males and females in relation to the number of divorces. We can see from the chart that there is an important and widening gap between divorce and remarriage, and – especially in the early 2000s – there is no correspondence between these two variables. A calculation of remarriage rates by sex31 shows that in Spain these are still quite low in a comparative perspective (8.4 per cent for males and 7.6 per cent for females in 2005).32 An examination of the ratios between remarriage rates for males and females in Spain reveals that historically there has been a gradual progression of sex levelling, although this trend has been more marked in the last few years.33 The fact that under Spain’s 1981 Divorce Act, there was a double filter to getting divorced (first separation, then divorce) may explain the low number of second marriages. Class selectivity – which was much higher for divorced than for separated persons – meant that the larger proportion of remarried people belonged to the middle classes. Another contributory factor was the longer duration of marriage at the time of divorce (about 16 years),34 which was a handicap to contracting new marriages.35 One strategy for facing post-divorce consequences that is frequently adopted in Spain is to live with close relatives, especially with parents. The rise of this residential arrangement is possibly connected with a need for pooling resources, whether they be material or caregiving services. A number of researchers insist on the importance of taking into account household composition in studying the social and economic consequences of single parenthood (Flaquer et al., 2006; Treviño, 2006). Results from previous own research show that in Spain, about one out of four children living with single parents dwell in households where at least one adult is present. In contrast to other extended family forms, between 1991 and 2001, the number of complex single-parent households has experienced a large increase. This shift can be interpreted as a reaction against the intensification of adverse consequences of divorce, but it may also reflect a higher need for the reconciliation between family and work. An analysis of the characteristics of people living in such complex households reveals that their socio-economic profiles are lower than those of other groups. This would suggest that this strategy is most frequently adopted by the lower strata of the population (Flaquer et al., 2006).

198

81

19

Figure 7.4

83

19

85

19

87

19

Remarriages men Remarriages women Number of divorces

89

19

91

19

93 19

95 19

97 19

99 19

01 20

Remarriage of divorced males and females and the number of divorces (1981–2005)

Eurostat.

0

10000

20000

30000

40000

50000

60000

70000

80000

Source:

Absolute numbers

20

03

20

05

Marital disruption in Spain

199

Finally, the high positive correlation that exists between household complexity and levels of child poverty (R 5 0.79) in a comparative perspective indicates that pooling resources in this way is of limited use in fighting severe poverty. However, if these kinds of residential arrangements did not exist in countries such as Spain, levels of child poverty would probably be much higher (Flaquer, 2007).

ANALYSIS OF THE RESULTS FROM SURVEY DATA ON ECONOMIC CONSEQUENCES OF DIVORCE The economic consequences of partnership dissolution are complex. Divorced and separated people are affected differently by the divorce process depending on their characteristics before and after divorce. Most individuals experience an important reduction in income and therefore in their capacity to face unexpected expenses. Some people may even fall under the poverty line. In this section, we analyse survey data on some of the economic consequences of divorce. In particular, we will draw a comparison between data from the ECHP (1994) and the EU-SILC (2004) on several economic dimensions of the population aged 18–59, analysed by sex and marital status. However, we also discuss results from other studies. The main dimensions under descriptive analysis are median equivalized disposable income, poverty levels and the capacity to face unexpected financial burdens. Median Equivalized Disposable Income A comparison of median equivalized disposable income36 for married, separated and divorced people shows that, in general, the married population has higher income levels than the other groups. We find that these differences exist irrespective of the number of children in the household. Of course, in all the categories, the income level depends upon the number of children, so that people with no children are better off than those with children (see Table 7.1). On the other hand, separated people have less disposable income than divorced people. This finding – also reported in previous studies of marital disruption – has to do with the character of the 1981 Spanish Divorce Act (see previously in this chapter) and may be the result of two different effects. Under the 1981 regulation, divorce used to occur a few years later than separation, and the elapsed time between the two events may explain

Single country studies

200

Table 7.1

Number of children 0 1 2 3 4 5 and more Source:

Median equivalized disposable household income by sex, marital status and number of children under 18 years in the household (population aged 18–59, 2004, euros) Male

Female

Married Separated Divorced Married Separated Divorced 12 304 10 873 9 181 7 197 5 214 4 563

11 600 9 227 8 615 5 685 – –

12 206 11 038 8 000 11 596 – 4 518

12 034 10 892 9 276 7 193 5 020 4 730

8 786 8 110 7 856 5 359 1 483 3 715

11 041 8 934 8 905 7 107 4 732 –

EU-SILC.

the improvement in economic conditions for divorced people. Secondly, the fact that the latter exhibited higher income levels than separated people did may suggest a selective effect in divorce. Taking into account that the cost of legal procedure was a deterrent to divorce, it could be expected that people with higher incomes were more likely to get divorced. As expected, a very heavy gender bias appears in the income differences between individuals from various marital statuses. While divorced men have similar (and even higher) disposable incomes than married men, the married women are much better off than the divorced ones. Furthermore, the disparities between separated and married males are smaller than those between separated and married females. Research on the evolution of income levels between 1995 and 2001 shows that, although income growth across the Spanish population was considerable during this period, the increase in the median income of household units with children was less than that of units without children. Single parents are one of the household categories whose increase was differentially less, with a growth of 31 per cent versus 46 per cent for the median of all households as a whole (Meil Landwerlin, 2006).37 The relative deterioration of the position of households with children during the second half of the 1990s in a country such as Spain, where the level of social transfers for families is comparatively very low and where divorce is not a recognized event in the welfare state, must undoubtedly have a strong impact on income levels of separated and divorced people. Figure 7.5 illustrates the previous findings by clearly showing that single parents are lagging behind in relation to other households with or without dependent children.38 The diagram depicts the evolution of

Marital disruption in Spain

201

16000

14000

Purchasing Power Standard (PPS)

12000

10000

8000

6000

4000

Single parent with dependent children (EU-15) Single parent with dependent children (Spain) Households without dependent children (EU-15) Households without dependent children (Spain) Households with dependent children (EU-15) Households with dependent children (Spain)

2000

0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Source:

ECHP, EU-SILC.

Figure 7.5

Median equivalized net household income by selected household types (EU-15 and Spain, 1995–2005)

median adjusted income in purchasing power standards for single parents and other selected household categories over the last decade. It shows how single parents experienced a drastic fall in income in the late 1990s after a substantial improvement in their position shortly after the start of the period covered. At the beginning of the 2000s there was a new recovery, but the last recorded data for 2004 and 2005 show a further worsening in income levels for all households with dependent children.39 Whereas in the EU-15, the evolution of the income gap between single parents and the bulk of households with dependent children reveals an improved position for single parents during the period (from 29.9 per cent less in 1995 to 24.8 per cent less in 2001), the reverse situation exists in Spain, but the changes are minor (from 21 per cent less in 1995 to 22.4 per cent less in 2005).

202

Single country studies

Finally, an analysis of the evolution of income gaps between the EU-15 and Spain in the late 1990s for these two categories of households shows that while the position of households with dependent children tended to improve, that of single parents tended to deteriorate.40 At-risk-of-poverty Rates The analysis of at-risk-of-poverty rates by marital status for the Spanish population aged 18–59 reveals that they are quite severe for separated people (Figure 7.6).41 As expected, gaps between the rates for married and separated individuals are much greater for females than for males. However, while from 1994 to 2004 they tended to decrease for males, the reverse is true for females. In 2004, the at-risk-of-poverty rate for separated women was nearly double that for married women. Again, the gender gap (difference in percentage points between at-risk-of-poverty rates for men and women by family status) is much higher for the separated than for the divorced. However, while in the last decade, gender differences have tended to increase among the separated (they have shifted from about 7 to 14 percentage points), they have tended to diminish among the divorced (from 4 percentage points to less than 1). The relative deterioration of the position of separated women with respect to men can be explained by the worsening situation of families with dependent children during the second half of the 1990s (see the ‘Median equivalized disposable income’ section), which is reflected in the growth of child poverty rates.42 Considering that after a family breakup, most children live with their mothers, and that no scheme of advanced child maintenance existed in Spain at that time, the financial status of separated women was more vulnerable. An explanation of the relatively favourable position of the divorced vis-à-vis the separated has already been suggested in the above section, on account of the greater selectivity of divorce and of the time elapsed after marital disruption. Previous research on poverty levels among single parents in Spain is in keeping with our results. Researchers found that not only do single-parent households face higher poverty risks in relation to other kinds of households with children, but also that between 1995 and 2001, there was an important shift in the poverty levels of Spanish single parents in relation to the EU-15 average. While in the mid-1990s poverty rates for Spanish single parents were much lower than those of their European counterparts, in the middle of the period there was a drastic shift, and in 2001 their position had worsened in comparison to the EU average (Flaquer et al., 2006). These results are in accordance with the findings discussed in the section, ‘Median equivalized disposable income’, on income levels for different kinds of

203

Never married

Male 1994 Male 2004 Female 1994 Female 2004

Figure 7.6

Married

Separated

Widowed

At risk of poverty rates by sex and marital status (population aged 15–59, 1994 and 2004)

ECHP and EU-SILC.

0

5

10

15

20

25

30

35

Source:

At poverty risk rates (poverty threshold set as 60% of median equivalized disposable income)

Divorced

Single country studies

204

Table 7.2

Capacity to face unexpected financial expenses by marital status (population aged 18–59, 2004, row percentages)

Marital status

Yes

No

Never married Married Separated Widowed Divorced

60.1 67.0 43.0 52.6 48.9

39.9 33.0 57.0 47.4 51.1

Source:

EU-SILC.

households, and may suggest that the drop in divorce selectivity could be an explanatory factor for this downturn. Capacity to Face Unexpected Expenses A number of people who do not live below the poverty line may nevertheless have a diminished capacity for facing unexpected financial burdens. For this reason, these data may represent drops in income that do not necessarily reflect situations of poverty. Once again, separated people are the marital status group with a lesser capacity for facing unexpected expenses. However, the share of divorced people in the same situation is also quite high. While the gap between separated and married individuals is 24 percentage points, the gap between divorced and married persons is 18 percentage points (Table 7.2). These results suggest that a reduction in people’s capacity for facing unexpected financial burdens is a more common consequence of marital disruption than the risk of poverty. Even if not all separated and divorced people fall under the poverty line as a result of divorce, a number of them may lose an important part of their economic well-being. As they lose their capacity to face some expenses, they become more vulnerable than married people, and they may be unable to respond to an unexpected event such as unemployment or illness.

CONCLUSION Following our review of the available data and the discussion of our analysis of marital disruption in Spain, a complex picture emerges which results from the interaction of a number of factors. Since the early 1980s, the rise in marital disruption and the diffusion of divorce have been accompanied

Marital disruption in Spain

205

by a steady increase in female employment. In principle, this should have a protective effect for the people involved in divorce to the extent that dual-earning households have better overall income levels, and that we can expect that the reduction of gender disparities in income before divorce is connected with a more balanced drop just after marital disruption. In fact, this does not appear to be the case. According to our findings, in the past decade, the gender gap in relation to poverty has tended to increase among the separated to the detriment of women, and in 2004, income disparities between married and separated females were much greater than between the same categories of males. This may be due to various factors. In most cases, dependent children tend to live with their mothers after a marital disruption, and this places more financial responsibilities on the latter. On the other hand, we have seen that since the enactment of divorce in Spain, there has been no practical tightening of controls for the payment of child maintenance nor has a scheme of advanced child maintenance payments been put into place.43 Moreover, divorce is not a recognized contingency in the Spanish welfare state; spending on family policy is very low, and there is no national safety net for people in need. Indeed, the limited development of the Spanish welfare state goes hand in hand with the strength of family solidarity. One of the ways in which many people try to face the adverse consequences of divorce is by moving in with relatives, particularly parents. In Spain, living with other relatives is clearly a most popular strategy, especially in the lower strata of the population. On the other hand, we have found low rates of remarriage in a cross-national perspective – possibly because second unions typically only occur among the middle classes. For most separated people, there seems to be no second opportunity.44 The growth of gender gaps may be also an indication of the decline of class selectivity in divorce. As we have hypothesized, the diffusion of divorce usually goes together with changing compositional effects in the population. In the case of Spain, the worsening economic position of separated people might be related to a drop in their levels of educational attainment. This is an hypothesis that we have explored in depth. While separated people have lower disposable incomes than divorced people do, relative poverty tends to be concentrated especially among separated females, and their position has tended to deteriorate somewhat in the past decade. Although it comes as no surprise that marital disruption adversely affects women more than it does men, the growth of gender gaps probably reflects the progression of working class divorce, considering that women with low levels of educational attainment tend to get more junk jobs and precarious contracts. Spanish single parents are not only the social category with the lowest

Single country studies

206

income, but their position also deteriorated drastically in the second part of the 1990s. This worsening was also apparent in a significant shift in single parents’ poverty rates in relation to the EU average, such that their previously favourable position was reversed. While we have not been able to prove Goode’s hypothesis in a strong sense by means of our cross-sectional data, our illustrations of the trends are quite suggestive. We think that the reduction of divorce selectivity in terms of class is one plausible explanation for the drop in income among separated people, and correspondingly for the growth of their poverty rates. However, we cannot preclude the possibility that other unknown factors are at work. In any case, more research is needed – particularly a longitudinal analysis, which will hopefully be possible when further waves of EU-SILC data are published. The national family policy and the demography of a given country are likely to be reciprocally related (Garib et al., 2007). In fact, welfare provisions may contribute to an increase in the number of divorces and therefore intensify some of their socio-economic consequences. More research is needed in order to know how to help divorced people without increasing the divorce rates. With this in mind, we would like to express our concern about the consequences of the considerable growth in divorce figures that has occurred in connection with the implementation of the 2005 Divorce Act. It is well known that apart from financial effects, divorce has other serious negative outcomes. If our previous analysis is correct, the new Divorce Act will probably lead to a further drop in class selectivity and to an increase in the diffusion of marital disruption.

NOTES 1.

2. 3. 4. 5. 6.

We would like to express our appreciation to Mr Antoni Vidal, lawyer at the bar of Barcelona, for his useful remarks on the practice of family law, and in particular of divorce court proceedings. We would also like to thank Alexander Schmidt (University of Cologne) for his statistical analyses of the survey data in the penultimate section. Data for 2004 from the Organisation for Economic Co-operation and Development (OECD) Family Database www.oecd.org/els/social/family/database, accessed 13 October 2008. When Spain entered the European Economic Community (EEC) in 1986, its unemployment rate was 21.3 per cent, approximately twice the EEC average. It reached its peak in 1994 at 24.2 per cent (Eurostat data). In fact, one of the permanent features of the Spanish labour market is that unemployment affects women much more severely than it affects men. In 1992, the female activity rate in Spain was only about 40 per cent; in 2006, it was about 60 per cent. Whereas in 1992, the gap between Spain and the EU was 15 percentage points in this respect, in 2006, the gap was only 3 percentage points (Eurostat data). That is, 10 per cent of the Spanish resident population, three times more than in 2001 (data from Consejo Económico y Social).

Marital disruption in Spain 7.

8. 9.

10. 11.

12. 13. 14.

15. 16.

17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

207

Employment rates for the population aged 15–64 were nearly the same: 64.8 per cent in Spain and 64.4 per cent in the EU-27. However, the Spanish female employment rates were four percentage points lower in relation to the EU (53.2 per cent in Spain versus 57.2 per cent in the EU-27; Eurostat data for 2006). Percentages refer to the total number of couples aged 20–49 where at least one partner has a job (Eurostat data from the European Labour Force Survey, 2003; as reported by Aliaga, 2005: 5). Recently published comparative data attest to Spain’s outermost position with regard to temporary employment. In 2005, 35.5 per cent of female and 31.6 per cent of male workers’ contracts were fixed term in Spain, in contrast to only 14.9 per cent and 13.9 per cent, respectively, in the EU-25. Of these, 24.4 per cent and 21.9 per cent were involuntary fixed-term contracts for men and women, respectively; the figures for the EU-25 were 7.5 per cent and 6.7 per cent (Hardarson, 2007: 3). The exception being a case in which the other spouse had been convicted of an attack on the applicant spouse or on his or her progeny or relatives. Such as the unjustified abandonment of the family home, marital infidelity, abusive or offensive conduct and any other serious or reiterated infringement of conjugal obligations; any serious or reiterated infringement of the obligations regarding children or the other spouse, having a criminal conviction for longer than six years, alcoholism, drugaddiction or mental illness resulting in separation. If both parties consented to divorce and agreed to ancillary matters. If the separation was documentarily proven. With the exception of 1982, when there were 19 per cent more divorces than separations, there has been an average of about 50 per cent more separations than divorces, with a peak of 77 per cent in 2002 (own calculations using data from the National Statistics Institute, INE). In cases where the parties have divorced, the widowhood pension is to be divided between the former spouses in proportion to the period of cohabitation with the deceased (Roca, 2002). A characteristic feature of this system was that, while wives were more likely to take the initiative in a separation, husbands were just as likely as wives to take the initiative in a divorce. According to 2004 data from the National Statistics Institute (INE), in 30.3 per cent of separations, the wife filed the petition to the court (versus 52.9 per cent of cases in which the petition was filed by both parties and 11.4 per cent in which it was filed by the husband). On the other hand, in 26.4 per cent of divorces, the petition was filed by the husband, in 26.0 per cent by the wife and in 41.4 per cent by both partners. In Sweden, which has one of the most liberal divorce laws in the world, it is necessary to wait 6 months after marriage. Although no state act regulating family mediation is in force in Spain, some of the autonomous communities have laws that do so. The 2005 Divorce Act announced the drafting of a state law on mediation, but this has not yet been passed. A person is considered to be lacking in resources if his or her income does not exceed twice the national minimum wage (Salario Mínimo Interprofesional). The law only establishes that the amount of a pension should correspond to the family’s economic circumstances and to the needs of the children at a particular time. However, some autonomous communities have passed legislation on the matter and have provided more specific regulations. The figures for divorces are 37.3 per cent, 1.2 per cent and 1.9 per cent, respectively (data from the INE). Data from the INE. The crude marriage rate refers to the number of divorces per 1000 population. Data from the INE. For the evolution of the crude marriage rate and total fertility rate since 1985, see Data Appendix. Women in Spain have about half the number of children that they would like to have. See Data Appendix. In 1975, the proportion of children born outside of marriage was only 2.0 per cent (data from the INE).

Single country studies

208 28. 29.

30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44.

Own estimation using data from the INE. For the evolution of crude divorce rates between 1985 and 2005, see Data Appendix. The analysis by Friedberg (1998) reveals a strong influence of unilateral divorce. Divorce rates would have been about 6 per cent lower if states had not adopted unilateral divorce, which accounts for 17 per cent of the overall increase between 1968 and 1988. We have not been able to find much information about the evolution of informal repartnering, but some evidence on the matter may be found in Flaquer et al., 2006. The percentages of divorced males and females married in one year to the total number of marriages. For example, in Denmark they were 21.9 per cent for males and 21.0 per cent for females in the same year. Danish remarriage rates in 1960 (12.2 per cent and 11.2 per cent, respectively) were much higher than Spanish rates in 2005 (Eurostat data). Shortly after the implementation of the 1981 Divorce Act, men tended to remarry up to twice as often as women did (data from the INE). With the exception of Italy, this is one of the longest durations in the EU. Recent data for 2006 show a shorter duration of 14.7 years (INE, 2007). Disposable income is the resulting income after paying direct taxes and receiving social transfers. Individual income has been equivalized – that is, adjusted to the household structure – using the OECD-modified equivalence scale. Income after direct taxes and before social transfers. Most single parents are separated or divorced. The data gap in 2002 and 2003 is due to a break in the series as a result of the shift from ECHP to EU-SILC. European Union averages were not available for 2004 and 2005. While in 1995, Spanish single parents had 20.5 per cent less income than their EU-15 counterparts, in 2001 they had 25.4 per cent less. For households with dependent children, the shift was from 29.4 per cent to 22.4 per cent during the same period. Poverty threshold fixed at 60 per cent of the median equivalized disposable income. During the 1990s, child poverty rates rose in 17 of the 24 countries for which data are available. In Spain, poverty rates (with a threshold set at 50 per cent of the median) rose by 2.7 percentage points (UNICEF, 2005). Beginning in 2008 a scheme of advanced child maintenance payments has been set up (see previously in this chapter). Little can be said about informal re-partnering because we have practically no information on this matter.

REFERENCES Alberdi, C. et al. (1995), ‘Organización formal: Marco legal de la familia’, in I. Alberdi (ed.), Informe sobre la situación de la familia en España, Madrid: Ministerio de Asuntos Sociales, pp. 377–459. Alberdi, I. (1979), Historia y sociología del divorcio en España, Madrid: Centro de Investigaciones Sociológicas. Alberdi, I. (1986), ‘Divorcio y sociedad en la España actual’, Sistema, Núm. 70, 93–112. Aliaga, C. (2005), ‘Gender gaps in the reconciliation between work and family life’, Eurostat. Statistics in Focus (Population and Social Conditions), Luxembourg: Office for Official Publications of the European Communities. Amato, P.R. (2000), ‘The consequences of divorce for adults and children’, Journal of Marriage and the Family, 62, 1269–87. Barbagli, M. and C. Saraceno (1998), Separarsi in Italia, Bologna: Il Mulino.

Marital disruption in Spain

209

Bentolila, S. and J.J. Dolado (1994), ‘Labour flexibility and wages: lessons from Spain’, Economic Policy, 9 (18), 53–99. Björnberg, U. (2006), ‘Paying for the costs of children in eight North European countries: ambivalent trends’, in J. Lewis (ed.), Children, Changing Families and Welfare States, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 90–109. Cousins, C. (1999), Society, Work and Welfare in Europe, London: Macmillan. Esping-Andersen, G. (1999), Social Foundations of Post-industrial Economies, Oxford: Oxford University Press. Esping-Andersen, G. (2002), ‘A child-centred social investment strategy’, in G. Esping-Andersen et al., Why We Need a New Welfare State, Oxford: Oxford University Press, pp. 26–67. Esping-Andersen, G. (2007), ‘Children in the welfare state’, in G. Esping-Andersen (ed.), Family Formation and Family Dilemmas in Contemporary Europe, Bilbao: Fundación BBVA, pp. 223–63. Ferrera, M. (1996), ‘The “Southern Model” of welfare in social Europe’, Journal of European Social Policy, 6 (1), 17–37. Flaquer, L. (2000), ‘Is there a Southern European model of family policy?’, in A. Pfenning and T. Bahle (eds), Families and Family Policies in Europe, Frankfurt a.m. and New York: Peter Lang, pp. 15–33. Flaquer, L. (2004), ‘Monoparentalidad’, Arbor. Ciencia, pensamiento y cultura, S. del Campo (ed.), La familia en el XXV aniversario de la constitución española, Núm. 178. Madrid: CSIC, pp. 345–76. Flaquer, L. (2007), ‘Family change and child poverty in comparative perspective’, Wellchi Working Paper Series 1/2007, Barcelona: Children’s Well-being International Documentation Centre, available at www.ciimu.org/webs/wellchi/ working_papers/wp1_flaquer.pdf, accessed 13 October 2008. Flaquer, L., E. Almeda and L. Navarro (2006), Monoparentalidad e infancia, Barcelona: Fundación ‘la Caixa’. Flaquer, L. and J. Iglesias de Ussel (1996), ‘The sociology of the family in Spain: an attempt at interpretation’, Marriage and Family Review, 23, 575–98. Friedberg, L. (1998), ‘Did unilateral divorce raise divorce rates? Evidence from panel data’, The American Economic Review, 88 (3), 608–27. Garib, G., T. Martín García and J. Dronkers (2007), ‘Are the effects of different family forms on children’s educational performance related to the demographic characteristics and family policies of modern societies?’, in H. Moerbeek, A. Niehof and J. van Ophem (eds), Changing Families and their Lifestyles, Wageningen: Wageningen Academic Publishers, pp. 27–50. Goode, W.J. (1966), ‘Marital satisfaction and instability: a cross-cultural class analysis of divorce rates’, in R. Bendix and S.M. Lipset (eds), Class, Status and Power. Social Stratification in Comparative Perspective, New York: Free University Press, pp. 377–87. Goode, W.J. (1993), World Changes in Divorce Patterns, New Haven, CT and London: Yale University Press. Hardarson, O. (2007), ‘Men and women employed on fixed-term contracts involuntarily’, Eurostat, Statistics in Focus (Population and Social Conditions), Luxembourg: Office for Official Publications of the European Communities. Härkönen, J. and J. Dronkers (2006), ‘Stability and change in the educational gradient of divorce. A comparison of seventeen countries’, European Sociological Review, 22, 501–17.

210

Single country studies

Houle, R., C. Simó, R. Treviño and M. Solsona (1998), ‘Los determinantes sociodemográficos y familiares de la ruptura de uniones en España’, Working Paper No. 142, Barcelona: Centre d’Estudis Demogràfics. Universitat Autònoma de Barcelona. Iglesias de Ussel, J. (1998), La familia y el cambio politico en España, Madrid: Tecnos. INE (National Statistics Institute) (2007), ‘Estadística de nulidades, separaciones y divorcios. Año 2006’, Notas de prensa, 15 de noviembre de 2007, Madrid: INE. Leibfried, S. (1992), ‘Towards a European welfare state: on integrating poverty regimes into the European Community’, in Z. Ferge and J.E. Kolberg (eds), Social Policy in a Changing Europe, Frankfurt am Main and Boulder, CO: Campus Verlag-Westview Press, pp. 245–79. Leitner, S. (2003), ‘Varieties of familialism: the caring function of the family in comparative perspective’, European Societies, 5 (4), 353–75. Martín-Casals, M., J. Ribot and J. Solé (2003), ‘Spanish report on the grounds for divorce’, in K. Boele-Woelki, B. Braat and I. Sumner (eds), European Family Law in Action, Vol. I (Grounds for Divorce), Antwerpen, Oxford, London and New York: Intersentia. Meil Landwerlin, G. (2006), Padres e hijos en la España actual, Barcelona: Fundación la Caixa. Naldini, M. (2003), The Family in the Mediterranean Welfare States, London and Portland, OR: Frank Cass. Roca, E. (2002), ‘Spain’, in C. Hamilton and A. Perry (eds), Family Law in Europe, London and Edinburgh: Butterworths, pp. 587–618. Ruiz Becerril, D. (1999), Después del divorcio: Los efectos de la ruptura matrimonial en España, Madrid: Siglo XXI. Saraceno, C. (1994), ‘The ambivalent familism of the Italian welfare state’, Social Politics, Spring, 60–82. Simó, C. and M. Solsona (2002), ‘Importancia de la participación laboral como determinante del divorcio en España’, Working Paper No. 129, Barcelona: Centre d’Estudis Demogràfics, Universitat Autònoma de Barcelona. Smock, P.J. (1993), ‘The economic costs of marital disruption for young women over the past two decades’, Demography, 30 (3), 353–71. Solsona, M., R. Houle and C. Simó (2000), ‘Separation and divorce in Spain’, in M.J. González, T. Jurado and M. Naldini (eds), Gender Inequalities in Europe: Women, Work and Welfare in the 1990s, London and Portland, OR: Frank Cass, pp. 195–222. Treviño, R. (2006), ‘Estructura y dinámica de la monoparentalidad en España’, unpublished PhD thesis, Universitat Autònoma de Barcelona. UNICEF (2005), ‘Child poverty in rich countries, 2005’, Innocenti Report Card No. 6, Florence: UNICEF Innocenti Research Centre. Uunk, W. (2004), ‘The economic consequences of divorce for women in the European Union: the impact of welfare state arrangements’, European Journal of Population, 20, 251–85. Valiente, C. (1997), ‘The rejection of authoritarian policy legacies: family policy in Spain (1975–1995)’, Comparing Social Welfare Systems in Southern Europe, Mire Florence Conference, Vol. 3, Paris: MIRE, pp. 363–83.

8.

Low divorce incidence in Greece: facts and figures Haris Symeonidou

INTRODUCTION The theory of utility maximization as applied to marriage (Becker, 1991) suggests that people marry and remain married for the purpose of conducting mutually beneficial exchanges. In an efficient marriage, each of the partners specializes either in home or in non-home production. Men spend more time in the labour market than do women, since they tend to be better paid. On the other hand, women invest more time and effort in household activities and the care of children. The greater the specialization of husband–wife economic roles, the greater the benefits from the exchange and the more stable the marriage (Ross and Sawhill, 1975). Changes in these patterns which occur when women also work outside the home should lead to delayed marriage – or even non-marriage – and to an increased risk of divorce. In other words, women’s economic independence (economic independence hypothesis) reduces the mutual dependency between partners – and consequently women’s gains from being married – and makes it easier for a couple to end an unhappy marriage (Cherlin, 1992; Ross and Sawhill, 1975; Schoen, 2000). Similarly, a number of studies indicate that a better socio-economic position on the part of the wife increases the risk of divorce (for an overview see Poortman and Kalmijn, 2002). Along the same lines, women’s level of education – considered as an investment in the market and used as an indicator of their earnings potential – may also be positively related to the disruption of marital unions. However, women’s improved economic position can have exactly the opposite results (see, for example, Oppenheimer, 1997), since the gains derived from marriage are increased (due to combined incomes in dual-earner families), and consequently the union may strengthen instead of weaken. Children – also considered to constitute a partnership-specific investment (Becker et al., 1977) – and union duration connected with the accumulation process of other shared assets besides children (Becker, 211

212

Single country studies

1991) are often negatively associated with union disruption (for example, Thornton and Rodgers, 1987). Finally, other factors, such as parental divorce (Kiernan and Cherlin, 1999; McLanahan and Bumpass, 1988), religiosity (Klijzing, 1992) and partners’ cultural background (Müller, 2003) seem to be connected to marital dissolution. All these factors have been examined in the case of Greece and have been proven to be significant; Becker’s specialization hypothesis has also been verified (Symeonidou and Mitsopoulos, 2003). With regard to the economic consequences of divorce, research findings show that while men undergo minimal income losses after divorce, women usually experience a significant deterioration in their economic situation (Duncan and Hoffman, 1985; Smock, 1993, 1994; Uunk, 2004). The economic costs of divorce may be decreased through public welfare policies (Burkhauser et al., 1991; Jarvis and Jenkins, 1999). However, only a limited number of studies have been conducted on this subject due to the lack of relevant longitudinal data (for an overview of such studies, see Uunk, 2004). Results from selected countries (Belgium, Germany, Great Britain, Italy and Sweden) show that women may lose as much as 80 per cent of their income (German results) after a divorce, while the maximum loss for men (again, in Germany) amounts to 23 per cent (Andreß et al. in this volume). Studies conducted in the other countries mentioned have shown more moderate income losses (Dewilde, 2002; Uunk, 2004), the highest being in Great Britain (36 per cent of income lost in the first study and 27 per cent in the second). Our aim in this chapter is to present a comprehensive overview of the phenomenon of divorce in Greece. In the next section I present relative demographic data describing fertility trends as well as trends in partnership formation and dissolution and changes in family structure. Some data on women’s labour force participation (employment and unemployment rates) will then be presented; following which, I introduce the main issues of the Family Law of 1983, which is still in force today. I go on to describe the role of the welfare state as it relates to protecting the family. Finally, I briefly describe the financial consequences of divorce for women, followed by my summary and conclusions.

FAMILY STRUCTURE Fertility Trends The fertility rate in Greece does not seem to have been affected by the baby boom of the post-war period. Furthermore, the country was a latecomer

Low divorce incidence in Greece

213

with respect to the Second Demographic Transition. The average number of children per woman remained stable at 2.3 throughout the period from 1956 to 1980. It is only after 1981 that fertility dropped below the replacement level of 2.1. Today, the total fertility rate in Greece (1.29 in 2004) is among the lowest in Europe and worldwide (Symeonidou, 2006; see Data Appendix). These trends can be explained in demographic terms by the decline of marital fertility and nuptiality rates, which have not been counterbalanced by an increase in the rate of births out of wedlock. Although the percentage of non-marital births has shown a noticeable increase during past decades, it still remains low (see Data Appendix); together with Italy, it is actually the lowest among the European Union (EU) countries (Eurostat, www.eu.int). Partnership Formation and Dissolution The crude marriage rate in Greece has shown a slight decrease since 1985 (see Data Appendix), although it remains among the highest in the EU. The same holds true for the total first marriage rate for females under the age of 50, which, despite a decrease, remained high (together with Portugal, the highest among the EU countries – Council of Europe, 2005). It seems that marriage is still considered important as an institution in Greece. Non-marital cohabitation has increased in recent years, although marriage remains the most common type of union. The new Family Law of 1983, which made important changes in the laws governing marriage and divorce, did not provide a legal framework for cohabitation, as was the case in other European countries. Therefore, economic insecurity, further aggravated by the low participation of women in the labour force and increasing levels of unemployment, may partially explain the low nonmarital cohabitation rate in Greece. This is also a major reason for the low number of children born out of wedlock. As far as marital dissolution is concerned, statistics on divorce (Council of Europe, 2005) show interesting differences between Greece and other EU countries. Although we can observe an increase in the crude divorce rates (from 0.4 per thousand population in 1970 to 0.8 in 1985 and 1.1 in 2004) and in the total divorce rates (from 8 divorces per 100 marriages in 1984, to 18 in 2001 and 21 in 2004), these figures are the lowest (together with Italy), among all the EU countries (see Data Appendix). Composition of the Household The composition of Greek households (see Data Appendix) has shown some important changes between 1990 and 2005. While the proportion of

214

Single country studies

single parents has increased (from 6.0 per cent to 7.3 per cent) the proportion of couples with children has decreased, from 49.1 per cent to 37.9 per cent, due to declining fertility. On the other hand, the proportion of couples without children remained stable (24 per cent), while the proportion of one-person households has almost doubled (from 14.6 per cent to 25 per cent). This latter change must be mainly attributed to the increased number of people living alone after the death of their spouses (the Greek population is ageing, mainly because of the aforementioned declining fertility), to the increased number of young people living alone temporarily because of studies (attending a university situated away from the parental home) – and to a far lesser extent, to the increased number of divorces, as well as to young people who leave the parental home permanently to live on their own. In this last case, however, we should consider the fact that unemployment and housing costs are likely to impede this development.

THE LABOUR MARKET The Greek labour force is characterized by a high percentage of selfemployment, a low percentage of salaried work and a relatively high percentage of unpaid family work, especially among women. These trends can be explained by the fact that, while the secondary and tertiary sector are still limited, the primary sector, though declining, continues to play an important role in the Greek economy. On the other hand, Greece shows a very low percentage of part-time work, the lowest among the EU countries (European Commission, 2004). Looking at unemployment, the rates have increased significantly – from 1.5 per cent among men and 3.3 per cent among women in 1977 to 6.1 per cent and 15.3 per cent, respectively, in 2005 (Symeonidou, 2002; see also Data Appendix). Long-term unemployment (longer than 12 months) increased considerably between 1985 and 2002. It must be noted that since 1993, more than 50 per cent of the unemployed are long-term unemployed; among women, the percentages have been this high since 1985 (European Commission, 2004). Women’s economic activity in Greece has remained stable over the past five decades (around 30–37 per cent), with only minor fluctuations being observed during the entire period. This stability (for an analysis of the reasons for it, see Symeonidou, 1997) characterized most European countries from the mid-nineteenth century up until the 1970s or slightly later, after which women’s participation rates started rising (Blossfeld, 1994; Hakim, 1993). The recent increase for Greece – 46.1 per cent in 2002 (see

Low divorce incidence in Greece

215

Data Appendix) – is partially due to changes in the definitions of labour force with regard to the relevant age groups and to unemployment, and to the growth of the service sector, where most women are employed. Nevertheless, even this increased rate is the lowest (together with Spain and Italy) among all the EU countries (European Commission, 2004). Previous research findings (causal analysis) have shown that when we consider various socio-economic variables, fertility has not had any statistically significant effect upon women’s employment, as was the case in other European countries. Inflexible labour market conditions do not permit women to interrupt their employment for any serious length of time after childbirth, if they wish to return to the workforce. What usually happens is that they are either compelled to keep on working, or to withdraw permanently from the workforce upon their marriage or, more commonly, with the birth of their first child. Those women who have never worked have great difficulties in finding a first job (Magdalinos and Symeonidou, 1989; Symeonidou and Magdalinos, 1993). Finally, it has to be pointed out that women’s earnings are still lower than men’s. For example, in 1998, women’s earnings as a share of men’s earnings (among wage earners in manufacturing) were equal to 82.1 per cent (see Data Appendix). All the above facts and figures, combined with the high percentage of unpaid work by women in family businesses (28 per cent in 2005), show that women’s position in the labour market is still much worse than men’s (National Statistical Service of Greece, www.statistics.gr).

FAMILY LAW Until 1983, the legislation governing family relationships in Greece dated back to 1946 and expressed the social values of an earlier age. In 1983, however, a new family law came into force (Law 1329/83). Underpinning the more modern belief that men and women have equal rights and obligations in the family, an acknowledgement of women’s changing role in society is reflected in the 1983 legislation. Articles 1438–1446 of the Civil Code deal with matters relating to divorce. The changes brought about by the Family Law of 1983 could be summed up as follows. Fault divorce is not the only kind available. Divorce can be obtained by mutual consent, after four years of separation and/or by reason of an irretrievable breakdown of the marriage relationship. It is a condition for divorce that the marriage must have lasted at least one year and that two applications are submitted, with a prerequisite of a time lag of at least six months between them.

216

Single country studies

With regard to property belonging to a husband and wife, the 1983 law establishes the right of both spouses to claim participation in whatever property is acquired during the marriage. Thus, if the marriage is dissolved by death or divorce, is annulled, or if the partners are separated (not divorced) for more than three years, each spouse is entitled to participate in the property acquired by the other during the marriage. This participation is proportionate to the amount that each marriage partner can prove that he or she contributed to the increase in the other’s property. Housework and the bringing-up of children also count as contributions. In any event, the spouse with the smaller increase in property is entitled to one-third of the property acquired by the other during the marriage without having to prove his or her contribution to the acquisition of that one-third, even if this contribution is limited to bringing up the children and managing the joint home. The spouse against whom a higher claim is addressed may be exonerated if he or she can prove that the other spouse did not contribute in any manner at all to the increase in property acquired by him or her, or that he or she contributed less than one-third. Anything acquired as a gift or inheritance is not included in the property. Both spouses are obligated to provide maintenance. Regardless of who was at fault for the divorce, either of the former spouses can seek maintenance from the other – but only in the event that he or she cannot support him or herself on his or her income or property, and that at least one of the following conditions applies: 1. 2. 3. 4.

5.

He or she is at an age at which he or she cannot start or continue a job which would support him or her. His or her state of health does not permit the commencement or continuation of a job which would support him or her. He or she has the custody of minor children and is therefore prevented from working. He or she cannot find a stable and suitable job or requires professional training (in these two cases, however, maintenance cannot be sought for longer than three years). In every case where the court judges that maintenance should be awarded for reasons of clemency.

Each of the former spouses is obliged to provide the other with precise details about his or her property and income if these are necessary for determining the amount of maintenance. Furthermore, upon the application of one of the former spouses – which is forwarded through the district attorney – the employer and the tax authorities are obliged to provide all relevant information concerning the financial status of the other spouse, and particularly concerning his or her income.

Low divorce incidence in Greece

217

As far as the custody of children is concerned, in the case of a divorce by mutual consent, the parents are obliged to provide the court with a statement in writing of their agreement as to which of them will have custody of the children before a divorce can be granted. This agreement is then confirmed by the court and has the validity of a provisional decision until a final decision is issued concerning the parental care of the children. This includes care of the child, his or her representation, administration of his or her property, as well as his or her representation in every legal transaction which concerns the child’s personal property. In all cases of divorce (by mutual consent or because of violation of the matrimonial bond), the court awards parental care either to one of the parents, to both of them together (if they agree upon where the child is to live) or even to a third person. The criteria that a judge considers when awarding parental care and custody of children are as follows: 1. 2. 3.

4.

5.

6.

The main consideration is always the best interest of the child. The bond between the child and each of his or her parents and his or her brothers and sisters is taken into consideration. Any agreements that may have been reached between the parents relating to the custody of their children and the administration of their property are taken into account. The court decision must respect the equality of the parents and make no distinction by reason of sex, race, language, religion, political or other beliefs, citizenship, national or social origin, or the fact that one parent is richer or poorer than the other. Depending on the maturity of the child, the judge may call him or her into chambers to hear his or her wishes regarding which parent he or she prefers to live with. The decision establishing the custody of a child may be changed at any time upon the application of the parent not awarded custody, if the real conditions on the basis of which it was issued have changed.

The above regulations appear to establish equality between men and women in the case of divorce, as well as considering the children’s interests. However, the practice of the law is often quite different from the legal regulations. Since the institution of a Family Court – a long-standing demand of feminist organizations – has not been introduced yet, we can assume that an increased risk of an unfair judgement exists in case of divorce, which most often affects the most vulnerable population groups – that is, women and children. And as is usually the case, poorer and less educated women, either ignore or do not have the means to claim for their rights.

218

Single country studies

THE WELFARE STATE Divorce patterns seem to be closely related to welfare regimes. In countries with liberal welfare states, such as Great Britain, divorce rates are relatively high, while lower rates can be observed in countries with social democratic welfare regimes, such as Denmark, Finland, Sweden and Norway. The conservative welfare states of Austria, Germany, the Netherlands, Belgium, Luxembourg and France follow immediately after the Nordic countries. Finally, the lowest divorce rates are found in Southern European ‘familialistic’ welfare states like Greece, Italy and Spain (Blossfeld and Müller 2002–03; Kiernan 2000). Greece is an example par excellence of the Southern welfare state model outlined by Ferrera (1996) – although the term is now being re-discussed – characterized by fragmentation, dualism and ineffectiveness in income maintenance, near universalism in national health care (supplemented, however, by extensive private coverage), a particularistic-clientelistic welfare state and a peculiar mix between public and non-public actors and institutions. In this section, however, I focus mainly on family benefits and their amounts, since their effect on women’s well-being is more direct – especially in the case of divorce. Greece has never had an explicit family policy. Nevertheless, legislation on employment, social insurance and other areas of social security, as well as fiscal policy, also refer to family issues (implicit family policy, cf. Kaufmann, 2002). There are regulations that take the responsibilities and dependencies within marriage into consideration, as well as the duties of parents toward their children and those of adults towards their elderly parents. In some instances, these regulations can also have demographic implications, in the sense that they can mitigate the economic and other costs of having children. Financial aid (subsidies, allowances, tax exemptions, and so on) and services (nurseries, kindergartens, consultative services, and so on) are provided to families (conjugal or mono-parental). Some measures are available for families ‘at risk’ (for example, immigrant families) or for families in crisis (affected by unemployment, disability, dissolution, and so on) as well as for abandoned children (for example, adoption), while there are some specific measures for large families, a concern which first appeared in the legislation of 1926 (Moussourou, 1994; Symeonidou, 1997). However, in spite of the government’s increasing concern for the country’s demographic situation, family benefits in Greece are very limited. Family allowances are the lowest among the European Union (EU) countries (Dumon, 1992; Eurostat, 1994; Symeonidou and Magdalinos, 2007). Moreover, Greece allocates a very low share of its social security budget to

Low divorce incidence in Greece

219

the family as compared to other EU-15 countries (European Commission, 2004). Although its expenditure on social security as a percentage of gross domestic product (GDP) has increased in the last decade (from 21.2 per cent in 1992 to 27.2 per cent in 2001), Greece still falls behind most of the EU-15 countries when expenditure on social security is measured in purchasing power parities (PPP) per head of population. In 2001, this amount was equal to 3971 PPP for Greece, whereas the EU-15 average was equal to 6405 PPP. Social security is directed mainly into old age and survivors’ benefits: 51.3 per cent of the social budget is spent in this sector. Along the same lines, public expenditure for families is extremely low, and although it increased between 1985 and 1995 (from 0.36 per cent of the GDP to 1.9 per cent), in 2000, it stabilized at the same levels as in 1995. Family cash benefits follow the same trend (see Data Appendix). The ageing of the population and the high number of disability pensions explain this situation. The Greek welfare state spends only 6.9 per cent of its social budget on families and children. It is characteristic that while the at-risk-ofpoverty rate in Greece before social transfers was near the EU-25 average in 2001 (23 per cent in Greece versus the EU-25 average of 24 per cent), after social transfers, the gap increased: the average rate for Greece decreased only to 20 per cent, whereas the EU average fell to 15 per cent (European Commission, 2004). Declining fertility and increasing economic activity on the part of women make childcare services and other measures which support compatibility between family and working life especially important. There are already a number of measures in place which favour women, including the protection of working women against dismissal, maternity leave of 17 weeks in the private sector and 20 weeks in the public sector, parental leave of 14 weeks for each parent (see Data Appendix) and an additional maternity leave of two years for a child under 6 years old in the public sector; for each additional child, this maternity leave increases by one year. Nevertheless, parental leave is unpaid, and due to a number of severe restrictions, working parents do not generally make use of it. Mothers in the public sector also tend not to use the extended maternity leave, since it is also unpaid. Thus, these measures fall far short of adequately supporting women’s participation in the labour force. On the other hand, at the end of their maternity leave, women working in the public sector have the right to either a reduced time schedule or to a paid leave of nine months. Women working in the private sector have the right to a reduced time schedule amounting to either two hours per day for one year following the maternity leave, or one hour per day for two years. However, these measures cannot be considered truly helpful in reconciling the demands of family and working life.

220

Single country studies

As far as childcare is concerned, public care facilities are limited – especially for children younger than 3 years of age – and are of poor quality (Symeonidou, 1997). Specifically, only 3 per cent of children younger than 3 years old are accommodated in public nursery schools, while this percentage is much higher (about 70 per cent) for older children – that is, 3 to 6 years of age (see Data Appendix). In this instance, however, grandparents – and especially grandmothers – play an important role. According to the results of a secondary analysis of the Greek ffS,1 in dual earner families, grandparents are the main care providers for children: 46.4 per cent of children younger than 3 years of age, 29 per cent of children 3–5 years old and 24.1 per cent of children 6–15 years old are looked after by grandparents on a regular basis when both partners are working. Public nursery schools are used for a very low percentage of children less than 3 years old (1 per cent). In this case, nannies at the child’s home are preferred (13.8 per cent), while private nursery schools are favoured for older children (24.9 per cent) – a fact which again illustrates the absence of a welfare state (Symeonidou et al., 2001). From everything I have said above, it becomes clear that the institutional setting does not encourage women’s labour force participation in Greece. The supportive role of the family counterbalances the state’s inability to satisfy social needs to an extent, and the role of women is critical for the care of younger and older family members. Family obligations are principally women’s obligations. Men’s participation in household chores and childcare remains low (Symeonidou et al., 2000). Therefore, women have internalized their role as caregivers and are the link that keeps the family together – although their role often turns out to be ‘compulsory altruism’ (Land and Rose, 1985). As a result of this situation, many women – as already mentioned – either do not enter the labour force at all or they stop working in order to provide care for dependent family members. This situation sometimes leads to female social exclusion, especially in the case of separation or divorce, and again as I have already mentioned, may to some extent explain the very low divorce rates in the country.

FINANCIAL CONSEQUENCES OF DIVORCE: THE EMPIRICAL EVIDENCE The following analysis starts by presenting some descriptive information on the economic situation of separated and divorced individuals as compared to people who remain married continuously. It is based on cross-sectional data from the European Union survey of Statistics on Income and Living Conditions (EU-SILC; see Data Appendix) pertaining to the year 2003. It provides recent information on a large sample of individuals. However, as

Low divorce incidence in Greece

221

a single cross-section, it does not control for pre-separation characteristics of the separated and divorced people – which may imply that they are a very selective group (for example, with much lower household incomes than continuously married couples even before the marital breakdown occurred). Thus, in a second step, I discuss the results from panel studies based on the older European Community Household Panel (ECHP; see Data Appendix), which allow us to study the changes in household income after separation and divorce. Unfortunately, these results are based on a very small sample, and hence should be interpreted with care. The hypotheses to be examined in this section are based on the previous description of the Greek socio-economic context and are formulated as follows: 1.

2.

The lack of an effective state policy for the reconciliation of family and working life and the resulting low participation of women in the labour force, the low level of benefits provided for families, and the fact that the practice of the family law often contradicts legal regulations all result in higher poverty risks among divorced and separated women as compared to married women. These discrepancies are even larger when compared with men’s poverty risks. When age is added in the analysis, poverty among divorced and separated women is higher as compared to men’s poverty rates in all age groups.

Recent data on poverty based on marital status and gender are taken from the European SILC survey. The SILC survey was initiated by Eurostat in 2003, in collaboration with the member states, and replaced the European Community Household Panel (ECHP). In Greece, it was carried out in 2003 by the National Statistical Service and the Institute of Social Policy of the National Centre for Social Research. The total sample for Greece was equal to 10 930 households (20 386 adults aged 16 years and over were interviewed), and the results presented here were adjusted to represent the whole population. The main advantage of the SILC data is the fact that it is recent and covers a large sample of individuals. However, it does not control for preseparation characteristics, which may be different among divorced and married women (Symeonidou and Mitsopoulos, 2003). According to the definition used by the EU member states and Eurostat, the at-risk-of-poverty threshold is fixed for each country at 60 per cent of the national median equivalized income. The focus is therefore on the relative rather than on the absolute risk of poverty. The national threshold is computed for the population as a whole and is based on equivalized household incomes in order to take household size and composition into account.2

Single country studies

222

Table 8.1

Family status and risk of poverty of men and women (population aged 161 years, 2003)

Family status

Single Married Separated Widowed Divorced Total Source:

Men

Women

Total

N

%

N

%

N

%

1 347 700 2 615 403 32 345 104 726 56 042 4 156 216

18.7 20.2 14.7 28.2 12.0 19.8

1 000 107 2 612 432 48 502 634 245 90 737 4 386 023

20.4 20.2 34.5 26.4 26.1 21.4

2 347 807 5 227 835 80 847 738 971 146 779 8 542 239

19.3 19.9 26.6 26.6 20.7 20.6

EU-SILC.

The analysis of the risk of poverty for men and women by marital status yields some interesting results. First, Table 8.1 clearly shows that a higher risk of poverty is found among separated women (34.5 per cent) – that is, among women who are not divorced but are living separately from their husbands. The poverty risks for widowed and divorced women are identical (26.0 per cent) and follow immediately after that of separated women, whereas the poverty risk for widowed men is even higher (28.2 per cent). All these figures are much higher than the risk of poverty found among married women and men (20.2 per cent), which is quite near to the overall poverty risk (21 per cent). Separated and divorced men, however, seem to be much better off as compared to men and women in the other marital status categories, since they have the lowest poverty risks (14.7 per cent in the first case and 12 per cent in the second). Finally, the poverty risk for single men equals 19 per cent; for single women, it is 20 per cent. A first conclusion we can draw from the above findings is that, as expected, divorced, separated and widowed women are in a worse financial situation compared with married and single women, whereas the differences for men are even stronger. A possible explanation for this situation is that after the breakup of a family, mothers most often have custody of the children and are solely responsible for bringing them up. If they have never worked – or if they interrupted their jobs upon marriage or with the birth of their children – they find it extremely difficult (as already mentioned) to enter or re-enter the labour force. Moreover, they are not supported by the welfare state, and quite often the financial support provided by the husband is insufficient. Data by marital status and age for women and men (Table 8.2) show that the highest risk of poverty is found among young (20–39 years old), widowed and separated women (the rate being equal to 52.5 per cent in the first case

223

RoP W

29.6 15.0 25.5 36.0 20.4

31.2 14.8 21.5 23.1 18.7

EU-SILC.

,19 20–39 40–60 601 Total

Source:

Single

15.3 17.2 18.4 24.2 20.2

RoP M 27.4 18.4 17.6 26.4 20.2

RoP W

Married

– 16.6 18.3 5.7 14.7

RoP M – 49.4 28.3 27.8 34.4

RoP W

Separated

– 100.0 25.2 28.0 28.2

RoP M – 52.5 14.7 28.0 26.4

RoP W

Widowed

Family status

– 13.9 10.9 12.7 12.0

RoP M

– 27.8 22.4 34.0 26.1

RoP W

Divorced

30.9 15.7 18.5 24.3 19.8

RoP M

29.6 17.6 18.2 27.5 21.4

RoP W

Total

Age, family status and risk of poverty (RoP) of men (M) and women (W) (population aged 161 years, 2003)

RoP M

Age

Table 8.2

224

Single country studies

and to 49.4 per cent in the second).3 It is followed by the risk of poverty for older women (aged 60 years and over), which equals 36 per cent for single and 34 per cent for divorced women. In the case of men, the lowest percentages are found among those who are separated, aged 60 years and over (5.7 per cent) and among those who are divorced, aged 40–60 years (10.9 per cent). Finally, at-risk-of-poverty rates by household type for 2001 (European Commission, 2004, see Data Appendix), show that for Greece, the highest rate (37 per cent) is found among single parents (with at least one child), which is higher than in most EU-25 countries and higher than the EU-25 average (33 per cent). According to SILC 2003 results, 50 per cent of monoparental families have only one child, 42 per cent have two children, while the rest (8 per cent) have three or more children. It is worth mentioning, however, that 90 per cent of all one-parent families are headed by women. Therefore, one can easily infer that the high poverty rates for single parents refer mainly to women with children. Coming next to the question of women’s economic situation after divorce as compared to their economic situation while married, corresponding longitudinal survey data for Greece is rather limited. Uunk (2004), using longitudinal data from the ECHP (1994–2000, in a total of seven waves), investigated the economic consequences of divorce in European countries for 14 member states of the EU. He used data on income, employment and demographic characteristics – such as marital status and the presence of children – for a sample of women aged 18–65 years who experienced a separation from their marriage partners between two consecutive interview years. In total, the analysis refers to 948 divorced women with valid pre- and postdivorce information. However, although this analysis controls for selected pre-separation characteristics, it uses only a small sample size per country (68 cases on average). For Greece, the sample consists of only 34 cases. The results of the analysis for Greece show that women’s median income declines by 11 per cent after a divorce. This decline seems rather weak compared to the decrease observed in many other EU countries. However, this small drop may be influenced by low median pre-divorce income, which amounts to 10 065 PPP in Greece, as compared to at least 12 000 PPP in the other EU countries. Given this low initial income, further income declines are hardly feasible. But this result may also be due to the small number of cases in the Greek sample (Uunk, 2004).

SUMMARY AND CONCLUSIONS In demographic terms, Greece fits into the Mediterranean model: low fertility rates, a low rate of births out of wedlock, declining marriage rates

Low divorce incidence in Greece

225

(although marriage still remains a strong institution), a low incidence of cohabitation and low divorce rates. In terms of welfare, the country is a representative example of the Southern European welfare model outlined by Ferrera (1996): Inadequacy and lack of uniformity in the benefits provided, absence of coordination between individual services, and inequality in its coverage of the population are some of its main characteristics. The family plays a crucial role in substituting for the state’s inability to satisfy social needs, and women are the agents of care for older and younger family members. Therefore, women are quite often impeded from participating in the labour force, or they work in low-status, low-paid jobs – and very often as unpaid family members in their fathers’ or their husbands’ businesses. Overall, women’s participation in the labour market is very low – the lowest, together with Italy, in the EU-15 (European Commission, 2004). The lack of state support for the family and for the reconciliation of family and working life is mainly responsible for this low participation, as well as for the aforementioned low fertility rates. The economic dependence of women – together with the fact that Becker’s specialization hypothesis seems to be valid for Greece (Symeonidou and Mitsopoulos, 2003) – can explain, to an extent, the low divorce rates in the country. The family law of 1983 introduced some important changes with regard to divorce – such as divorce by mutual agreement, the institution of common property and some important regulations governing the custody of children. However, Greece still lacks the institution of a Family Court, which the Council for Equality has been proposing since 1983. Therefore, women (and children) may be seriously affected by an unfair judgment in a case of divorce, since the practice of the law is quite often very different from the legal regulations. As far as the financial situation of divorced and separated women is concerned, cross-sectional survey data (from SILC 2003) show that they have higher at-risk-of-poverty rates than married and single women, and much higher rates than men. Panel data (ECHP 1994–2000) show that the median income for women declines by 11 per cent after a divorce, although the small number of cases in the sample may distort the results. Therefore, the first hypothesis of higher at-risk-of-poverty rates among divorced and separated women as compared to other women and men seems to be verified. The differences in at-risk-of-poverty rates by marital status are more pronounced among young (20–39 years of age) and older women (60 years of age and older). Moreover, in all age groups, the rates for women are higher than the rates for men. Therefore, the second hypothesis – of

226

Single country studies

different at-risk-of-poverty rates by age for women and higher rates for women as compared to men in all age groups – is also verified. In conclusion, the continued predominance of marriage as a social institution in Greece need not be attributed to the society’s more traditional family values relative to other European countries. The weak position of women and their multiple roles in the family make divorce quite a difficult step to take. Cohabitation exists only as a pre-test to marriage, and it is not a viable alternative since it does not yet have a legal framework. Nevertheless, younger generations cohabit more and marry less. If family ties change, serious secondary effects will occur in society as a whole in the long-run. Many economically dependent individuals who are now protected by the family – disproportionately women – may become subject to poverty and social exclusion. It is necessary, therefore, that these social changes and potential consequences be seriously taken into consideration in the development of social policy and the introduction of new legislation concerning union formation, union dissolution and the fate of economically dependent family members. The provision of necessary services for the young and the elderly could relieve women of their multiple burdens and thus enable them to decide more freely whether to remain in or leave a marital union.

NOTES 1. The Greek Fertility and Family Survey (ffS) was carried out in 1999, in the wider context of the United Nations Economic Commission for Europe ffS project. The Greek survey was carried out by the National Centre for Social Research (EKKE), under the direction of the present author. A national random sample of 3048 women and 1026 men, aged 18–49 years, were surveyed in personal interviews on the basis of the United Nations (UN) Population Activity Unit standard questionnaire. The results were published in Symeonidou (2002). 2. According to Eurostat’s definition, total disposable income is computed as the total net household monetary income excluding both taxes and social insurance contributions as well as in-kind income and imputed rents. The equivalence indicator weights total disposable household income for the individual household, which depends on the household size and composition, and the result – that is, equivalized disposable income – is attributed to both the household unit and each member. Total household income is divided by the number of household members weighted by the so-called ‘new’ modified Organisation for Economic Co-operation and Development (OECD) equivalence scale used by Eurostat. This scale gives a weight 1.0 to the first adult, 0.5 to any other household members aged 14 and over and 0.3 to each child below the age of 14. The median is computed on the basis of the equivalized incomes of the total population, and the poverty line is set at 60 per cent of the median. It should be noted that the concept of equivalized income implies that each household member enjoys the same level of welfare. In other words, the equivalized income represents a welfare indicator and not a monetary income indicator. 3. The rate of 100 per cent for widowed men aged 20–39 years may be biased because of the small number of cases in the sample.

Low divorce incidence in Greece

227

REFERENCES Becker, G.S. (1991), A Treatise on the Family, enlarged edn, Cambridge, MA: Harvard University Press. Becker, G.S., E.M. Landes and R.T. Michael (1977), ‘An economic analysis of marital instability’, Journal of Political Economy, 3, 1141–87. Blossfeld, H.P. (1994), ‘Family cycle and growth in women’s part time employment in Western European countries’, Brussels: European Commission, DGV, mimeo. Blossfeld, H.P. and R. Müller (2002–03), ‘Introduction’, in H.P. Blossfeld and R. Müller (guest editors), ‘Assortative partnership selection, division of work in the household, and union separation’, International Journal of Sociology, 32(4), 3–35. Burkhauser, R.V., G.J. Duncan, R. Hauser and R. Berntsen (1991), ‘Wife or frau, women do worse: a comparison of men and women in the United States and Germany after marital disruption’, Demography, 28, 353–60. Cherlin, A.J. (1992), Marriage, Divorce, Remarriage, revd edn, Cambridge, MA and London: Harvard University Press. Council of Europe (2005), Recent Demographic Development in Europe, Brussels: Council of Europe. Dewilde, C. (2002), ‘The financial consequences of relationship dissolution for women in Western Europe’, in E. Ruspini and A. Dale (eds), The Gender Dimension of Social Change. The Contribution of Dynamic Research to the Study of Women’s Life Courses, Bristol: Policy Press, pp. 81–110. Dumon, W. (1992), National Family Policies in EC Countries in 1991, Vol. 1 Brussels: Commission of the European Communities, Directorate General for Employment, Industrial Relations and Social Affairs. Duncan, G.J. and S.D. Hoffman (1985), ‘A reconsideration of the economic consequences of marital dissolution’, Demography, 22, 485–97. European Commission (2004), The Social Situation in the European Union, Luxembourg: Office for Official Publication of the European Communities. Eurostat (1994), ‘Social protection in the European Union’, Statistics in Focus, Population and Social Conditions, 5, 1–5. Eurostat, www.eu.int, accessed January 2006. Ferrera, M. (1996), ‘The southern model of welfare state in social Europe’, Journal of European Social Policy, 6, 17–37. Hakim, C. (1993), ‘The myth of rising female employment’, Work, Employment and Society, 7, 97–120. Jarvis, S. and S.P. Jenkins (1999), ‘Marital splits and income changes: evidence from Britain’, Population Studies, 53, 237–54. Kaufmann, F.X. (2002), ‘Politics and policies towards the family in Europe: a framework and an inquiry into their differences and convergences’, in F.X. Kaufmann, A. Kuijsten, H.J. Schulze and K.P. Strohmeier (eds), Family Life and Family Policies in Europe, Oxford: Oxford University Press, pp. 419–77. Kiernan, K.E. (2000), The State of European Unions: An Analysis of ffS on Partnership Formation and Dissolution, London: Department of Social Policy and Centre for the Analysis of Social Exclusion, London School of Economics and Political Science. Kiernan, K.E. and A.J. Cherlin (1999), ‘Parental divorce and partnership dissolution in adulthood: evidence from a British cohort study’, Population Studies, 53, 39–48.

228

Single country studies

Klijzing, F.K.H. (1992), ‘“Weeding” in the Netherlands: first union disruption among men and women born between 1928 and 1965’, European Sociological Review, 8 (1), 53–70. Land, H. and H. Rose (1985), ‘Compulsory altruism for some or altruistic society for all?’, in P. Bean, J. Ferris and D. Whynes (eds), Defence of Welfare, London: Tavistock, pp. 74–96. Magdalinos, M. and H. Symeonidou (1989), ‘Modeling the fertility–employment relationship: simultaneity and misspecification testing’, European Journal of Population, 5, 119–43. McLanahan, S. and L.L. Bumpass (1988), ‘Intergenerational consequences of family disruption’, American Journal of Sociology, 94, 152. Moussourou, L.M. (1994), ‘Family policy in Greece: traditional and modern patterns’, in W. Dumon (ed.), Changing Family Policy in the Member States of the European Union, Brussels: Commission of the European Communities, DGV, pp. 88–105. Müller, R. (2003), ‘Union disruption in West Germany. Educational homogeneity, children and trajectories in marital and nonmarital unions’, in H.P. Blossfeld and R. Müller (guest editors), ‘Assortative partnership selection, division of work in the household, and union separation’, International Journal of Sociology, 33 (2), 3–35. National Statistical Service of Greece, www.statistics.gr, accessed January 2006. Oppenheimer, V.K. (1997), ‘Comment on Ruggles’, Demography, 34, 467–72. Poortman, A. and M. Kalmijn (2002), ‘Women’s labour market position and divorce in the Netherlands: evaluating economic interpretations of the work effect’, European Journal of Population, 18, 175–202. Ross, H. and I.V. Sawhill (1975), Time of Transition: The Growth of Families Headed by Women, Washington, D.C.: Urban Institute. Schoen, R. (2000), ‘Union disruption in the United States of America’, paper presented at the Cross-National Comparative Workshop on the Effects of Assortative Union Formation and Division of Work within the Household on Union Separation, 10–11 December 1999, University of Bremen. Smock, P.J. (1993), ‘The economic costs of marital disruption for young women over the past two decades’, Demography, 30, 353–71. Smock, P.J. (1994), ‘Gender and the short-run economic consequences of marital disruption’, Social Forces, 73, 243–62. Symeonidou, H. (1997), ‘Social protection in modern Greece’, in M. Rhodes (ed.), Southern European Welfare States. Between Crisis and Reform, London: Frank Cass, pp. 67–86. Symeonidou, H. (2002), Fertility and Family Surveys in Countries of the ECE Region. Standard Country Report – Greece, New York and Geneva: United Nations. Symeonidou, H. (2006), Family and Fertility in Greece. A Cohort Analysis, Athens: Ant. Sakkoulas. (In Greek.) Symeonidou, H. and M. Magdalinos (1993), ‘Fecondité et travail féminin en Grèce: un modèle micro-économique’, Cahiers Québéquois de Démographie, 22, 285–312. Symeonidou, H. and M. Magdalinos (2007), Family Policies in the Countries of the European Union. Reconciliation of Family and Working Life. Cost–Benefit Analysis for Greece, Athens: Ant. Sakkoulas. (In Greek.) Symeonidou, H. and P.G. Mitsopoulos (2003), ‘Union disruption in Greece’, in

Low divorce incidence in Greece

229

H.P. Blossfeld and R. Müller (guest editors), ‘Assortative partnership selection, division of work in the household, and union separation’, International Journal of Sociology, 33 (2), 65–94. Symeonidou, H., M. Cavouriaris, E. Kandilorou, M. Magdalinos, P.G. Mitsopoulos, J. Tsahalides and K. Vezyrgianni (2000), Expected and Actual Family Size in Greece. Life-Cycle Events. A Follow-up Study: 1983, 1997, Athens: The National Centre for Social Research. (In Greek.) Symeonidou, H., P.G. Mitsopoulos and K. Vezyrgianni (2001), The Impact of Policies on the Division of Paid and Unpaid Labour in Families, European Network on Policies of Unpaid and Paid Work, Tilburg: European Science Foundation. Thornton, A. and W. Rodgers (1987), ‘The influence of individual and historical time on marital dissolution’, Demography, 24, 1–22. Uunk, W. (2004), ‘The economic consequences of divorce for women in the European Union: the impact of welfare state arrangements’, European Journal of Population, 20, 251–85.

PART III

Comparative Studies

9.

Female-supportive policies and women’s employment after divorce Maike van Damme and Wilfred Uunk

INTRODUCTION After a divorce, most women experience a substantial income decline. Many divorced women also start to depend on social welfare and experience increased poverty risks (for example, Dewilde, 2002; Duncan and Hoffman, 1985; Finnie, 1993; Hoffman, 1977; Jarvis and Jenkins, 1999; Moore, 1978; Poortman, 2000; Poortman and Fokkema, 2001; Smock et al., 1999; Uunk, 2004). These negative financial consequences can be explained – among other things – by the low labour supply of women before and after divorce (Bouman, 2004; Poortman and Fokkema, 2001), and by (the lack of) institutional support for income and female employment (Uunk, 2004). However, in contrast to income, changes in women’s employment situation after separation have not been well investigated. Although existing studies have described these changes and investigated the effects of individual-level characteristics on post-divorce labour supply (Bouman, 2005; Bradbury and Katz, 2002; Duncan and Hoffman, 1985; Finnie, 1993; Haurin, 1989; Jarvis and Jenkins, 1999; Johnson and Skinner, 1986; Moore, 1978; Mueller, 2005), these studies were limited to single countries and consequently did not address whether a country’s institutional arrangements matter. In this contribution, we want to fill in the gap in the literature and describe and explain changes in labour supply after separation for women from 13 countries of the European Union. In particular, we investigate whether institutional arrangements affect women’s post-divorce labour supply. Scholars have found that female-supportive institutions (in particular, public childcare provision) increase married women’s employment (Gornick et al., 1998; Kalleberg and Rosenfeld, 1990; Stier et al., 2001; Van Dijk, 2001), and they have observed that these institutions mitigate the negative income consequences of divorce for women (Uunk, 2004). But do female-supportive policies influence women’s employment after separation as well? 233

234

Comparative studies

We will answer this question using data from the European Community Household Panel (1994–2001). These cross-national comparative, longitudinal data allow us to investigate short-term changes in employment. In our analyses, we include dissolutions of both marriages and consensual unions, since more and more people in (Northern) Europe now live in cohabiting relationships and never get married. In addition, our focus will be on the de facto end of the relationship and not on the timing of legal divorce, because the legal process of divorce may take a long time (months or even years). During this period, people may already live separately and experience changes in their economic situation.

THEORETICAL BACKGROUND AND HYPOTHESES Previous Findings Studies on the economic consequences of divorce suggest a general increase in labour supply after divorce, yet the findings are not fully consistent. Johnson and Skinner (1986) investigated individual labour supply changes among American women who divorced between 1969 and 1977. They revealed a sharp rise in women’s working hours in the two years after a separation. Most of this increase was due to an increase in levels of employment (from 68 per cent before the separation to 88 per cent afterwards). Others found increases in labour supply as well, but they did not find such large changes (Bouman, 2005; Bradbury and Katz, 2002; Duncan and Hoffman, 1985; Finnie, 1993; Haurin, 1989). Bradbury and Katz (2002), for example, revealed that in the USA during the 1990s, 12 per cent of women who lost their husbands (through death or divorce) started working, yet 10 per cent stopped. Moore (1978) investigated changes in annual working hours among young American women around 1970 and found that whites increased their working hours after a divorce by 60 percentage points, and blacks by 14 percentage points. Moreover, 26 per cent of the women who worked before a separation left the labour market after the split. For the United Kingdom, Jarvis and Jenkins (1999) even found a decrease in women’s employment after separation: 59 per cent of the wives studied worked before the divorce, compared to 51 per cent afterwards. In addition, they found that 15 per cent of the women who were employed before divorce stopped working afterwards. Lastly, a recent study conducted in Canada (Mueller, 2005) found no changes in divorced women’s labour supply at all. The finding that divorced women do not increase their labour supply across the board is initially surprising, since the majority of divorced

Female-supportive policies and women’s employment after divorce

235

women face a substantial reduction in adjusted household income, and employment may be a way to reduce financial distress. One reason why the effect of divorce on female labour supply is not consistently strong and positive is that some women might anticipate a divorce by increasing their labour supply before the separation. For example, Johnson and Skinner (1986) demonstrated a steady increase in working hours from the seventh year before separation to the year before separation. Yet others did not find a substantial pre-divorce increase in labour supply (Finnie, 1993). Poortman (2005) investigated anticipation differently by including the extent to which women expected their divorce to occur. She examined women’s labour force participation, number of working hours, part-time work and full-time work. Only in the case of full-time work did she find some support for anticipatory behaviour, but the effect was marginal. Another reason why divorce does not consistently increase female labour supply is country-specific variations in the divorce effect. In some countries, the majority of women react to divorce with an increase in labour supply, whereas in other countries, women reduce their working hours or stop working. We hypothesize that this cross-national variance in women’s reactions to separation is due to differences in the institutional context. For instance, government policies moderate the income drop women face after union dissolution (Uunk, 2004) and hence influence labour supply decisions. The Effect of Divorce on Women’s Labour Supply Why do some women increase labour supply after separation, others decrease labour supply and still other women do not change? We used a simple microeconomic labour supply model (based on Becker, 1965; Blau et al., 1998) to account for these varying reactions. We assume that women’s labour supply decisions are a trade-off between two commodities: time spent on paid work and time spent outside the labour market. The decision to work or not to work is based on a comparison between the market wage and a reservation wage – that is, the lowest wage rate at which it is worthwhile for a woman to work. If the market wage is higher than the reservation wage, the woman will seek a job; if it is lower, the woman will not seek a job. Moreover, she has to decide how many hours she wants to work depending on her wage and the opportunity costs of non-working hours. A woman makes her labour supply decision in the household context, which means that the employment decision and the decision on how many hours she wants to work is influenced by the total household income (including the income of other household members and her partner in particular). In theory, we assume that at every moment in

236

Comparative studies

time, this work/non-work choice plays a role. Hence, when the household composition and economic situation changes due to separation, women have to reconsider their labour supply decisions (in terms of employment and working hours). Following the logic of this labour supply model, we can expect divorce to have a positive effect on women’s labour supply. The reason for this expectation is the general deterioration in women’s financial situation after divorce. Such deterioration occurs because of a loss of economies of scale (for example, increased housing expenses) and because of the traditional, gender-based task specialization in the prior marriage (for example, Poortman, 2000). Women who lived in traditional male breadwinner-type households can no longer rely on their partners’ income and experience a drop in adjusted household income. These financial changes give divorced women a financial incentive to work. Therefore, non-employed women may enter the labour market after divorce and employed women may increase their working hours. Of course, we assume that not all separated women will increase labour supply to the same extent. We expect variation in labour supply changes depending on labour market, institutional and individual restrictions. On the individual level, human capital and children are important restrictions. Women with low education and little labour market experience are less likely to find a (good) job. Having to take care of (young) children also reduces women’s opportunities to work. Children cost time and money; time spent on caregiving cannot be spent on the labour market, and public childcare can be costly (Uunk et al., 2005; Van Dijk, 2001). The value of non-working time will be highest when women have children under 6 years of age, because these children do not attend school and require more intensive caretaking. Therefore, we expect separated women with young children to face more difficulties in increasing their labour supply than separated women without children or those with school-age children. Mothers with young children will also be more likely to exit the labour market after separation (under the assumption that they have alternative income sources). The costs and availability of (formal) childcare are important in this respect. We discuss the effects of public childcare provisions and other female-friendly policies in the next section. Labour market restrictions concern the availability of (part-time) jobs. The more employment positions there are available on the labour market, the easier it is for women to find jobs and the more likely it is that they will enter the labour market after a separation. Moreover, if part-time work is more common, women may have more opportunities to change (decrease or increase) their working hours: (full-time) working mothers who have difficulties in combining work and caregiving will be less forced to exit the

Female-supportive policies and women’s employment after divorce

237

labour market after separation; and non-employed mothers may more easily find a job in which they are able to combine work and caregiving. To summarize, we expect that women in general will increase their labour supply after divorce. However, some women face restrictions and will decrease their labour supply. In particular, we expect this to happen for mothers with young children. A Typology of Family Support Cross-national differences in family and work policies have an effect on women’s employment. For instance, ‘female-supportive’ policies such as public childcare provisions reduce the negative effect of children on female employment, encourage part-time work among women, and mitigate the costs of employment interruptions (Gornick et al., 1998; Mandel and Semyonov, 2006; Rosenfeld and Birkelund, 1995; Stier et al., 2001; Uunk et al., 2005; Van Dijk, 2001). Do public policies influence women’s employment decisions after a separation as well? We expect that they do. After all, separated women may have to reconsider their employment situation because their living situation (income, household situation) changes. Moreover, several studies have already revealed an impact of institutions on the financial consequences of divorce (Dewilde, 2002; Uunk, 2004). In examining the impact of institutions on women’s labour market participation after divorce, we have used a typology rather than quantifiable macro indicators. We believe that the institutional context of the 13 countries we analysed cannot be captured well by a limited set of macro indicators (such as the percentage of children in public childcare and the level of single-parent allowances), but involves multiple, interrelated institutional factors. These institutional effects can – given the limited degrees of freedom for testing such effects – be better portrayed by distinguishing qualitatively different country clusters. Or stated otherwise: ‘The analysis of country clusters enables us to identify commonalities and differences across regime types, even if we may not always be able to say definitively what it is about these countries that accounts for the variation observed’ (Gornick and Jacobs, 1998: 693). Welfare states differ in the basic principles and organization of their welfare state arrangements. The typology developed by Esping-Andersen (1990, 1999), which takes the role of the state, the market and the family into account, is the most well known. However, a disadvantage of his approach is that he considers women’s entitlement to welfare only as workers, not as wives or mothers. According to Lewis (1992) and others (O’Connor, 1993; Orloff, 1993; Sainsbury, 1996) welfare state typologies need to incorporate family ideology as well. Therefore, we choose to follow

238

Comparative studies

the typology of family support that has been described in the Introduction (Andreß and Hummelsheim, in this volume). The typology is based on the extent of income and employment support provided by the state for women’s economic independence, and the extent to which the market and the family have to step in if family policies are only rudimentarily developed. We distinguish four ideal types of family support: the dual-earner model, the model of traditional division of labour (as a shorthand we use the term ‘male breadwinner model’), the market model and the model of extended family solidarity (here, ‘family model’). In the ideal-typical market model there is low state support (both with respect to income and employment) for women’s economic independence; support for women is left to the market. In contrast, in the dual-earner model, there is high state support for women: countries within this model provide high levels of financial and employment support for women (for example, by providing high child allowances and sufficient childcare facilities). In the male breadwinner model, the state also plays an important role, but it is oriented towards a traditional division of labour in which men are the main breadwinners. Consequently, the state provides high financial support for children but low employment support for women. Finally, in the family model state support for women is low and family support is high. The typology implies that the institutional context may support separated women in three ways: through income support from the state (for example, single-parent allowances, alimony), through employment support from the state (for example, public childcare provisions) and through family support (for example, financial support via the family network). In order to gain insight into how the 13 European Union (EU) countries in our analyses rank in terms of these different types of support and how they should be classified into the typology, Table 9.1 quantifies country scores on multiple institutional indicators. We measured income support according to levels of social assistance, child allowances, singleparent allowances and child maintenance. Employment support is measured by the level of public childcare provision and an index of the quality of the childcare leave system (payment and duration). Family support is measured by a traditional family index (see notes, Table 9.1). Table 9.1 shows that Denmark and Finland score high on income and employment support measures and low on family support, and that they therefore can be grouped into the dual-earner model. The Netherlands, Belgium, France, Germany and Austria are also characterized by high levels of state income support, yet employment support from the state is substantially lower than in Denmark and Finland. We therefore group these countries in the male breadwinner cluster. The United Kingdom is the most obvious candidate for the market model: it provides low state

239

Notes overleaf

Dual earner Denmark Finland Family Italy Portugal Spain Greece Ireland Market United Kingdom Male breadwinner the Netherlands Belgium France Germany Austria

1 20 49 2 21h 283 224 2113 32 0 13 0 30 220 27 224 210

Social assistancea 0 137 163 2 144i 46i 46i 18 76 0 139j 1 171 217 95 231 215

Child allowanceb 1 1010 478 2 474 236 208 208k 492 1 833 0 682 805 535 549l 566l

Single parent allowancec

Income support

0 0 1 2 0n 1 11 11

0 0 0 2 2m 2m 2m 2m

Child maintenanced 11 11 1 2222 0 2222 2222 2222 0 1 1 11 1 11 0 11 11

Total

1 56 27 – 6 12 5 3 20 (2) – 18 (2) – 7 30 26 6 3

Public child care provisionse 1 High High – High Medium Medium Low Low – Low 0 Low Medium Medium Medium High

Childcare leavef 11 11 11 22 0 0 2 22 22 22 22 2 22 1 1 2 0

Total

Employment support

Classification of countries according to four models of family support

Family support model

Table 9.1

– 1.6 2.3 1 11.6 11.6 12.2 10.1 11.1 0 5.6 0 4.5 8.0 6.9 7.0

Traditional family indexg

– – – 1 1 1 1 1 1 0 0 0 0 0 0 0

Total

Family support

240

(continued)

Notes and sources: a Disposable income of social assistance recipients (after housing costs) as a proportion of the mean for all OECD countries: average for nine household types, standardized for GDP per capita, 1992 (in PPP). Source: Gough et al. (1997). b Legislative guaranteed monthly net allowance for families with two children, age 7 and 14 (including supplements), mean over the period 1996–2001 (in PPP). Source: Mutual Information System on Social Protection (MISSOC), European Commission (1997, 2002). c Legislative guaranteed monthly net allowance for single parent families with one child of ten years, mean over the period 1996–2001 (in PPP). This is the sum of basic allowance for welfare dependency, single parent allowance and child allowance. Source: Mutual Information System on Social Protection (MISSOC), European Commission (1997, 2002). d Sum of 2 indicators: (1) formal child maintenance liabilities (amount that will be paid by fathers with 1.5 average income who want to divorce from mothers with average part-time income needing 12 hours of child care a week and having 2 children (5 and 9 years old); (2) relative value of advance maintenance, 1997 (£/month in PPP). Source: Corden (1999: 34, vignette B). e The number of available public child care places per 100 children under age three in publicly funded day care services, mean of 1990–1995 and 1998–2001. Source: Tietze and Cryer (1999), Van Dijk (2001) and OECD (2001). Data from 1998–2001 include both public and private provisions. Due to the high importance of the latter in the UK and Ireland, values are slightly distorted for these countries. Proportion of public child care provisions in 1993 in brackets. f The scores low, medium, and high are created by Koopmans and Schippers (2003). Sum of maternity, parental, paternity and care leave. Payment and duration are given even weight. Source: Koopmans and Schippers (2003: 28). g Index of 5 indicators: (1) the national average household size, (2) the percentage of single adults in the age of 30–64 years, that is after the family formation period, and before mortality affects the outcome (the propensity to stay single), (3) the percentage of adults living in a consensual union, (4) the percentage of young adults under the age of 30 still living with their parents, (5) the percentage women aged 70–84 living with their mid-age children (301), 1994. Source: Vogel (2003: 433). h Unreliable figure. i No child benefit above a certain income level. j For 1998 and 1999, amount for first child refers to single parent families. k No general social security schemes. We assume the allowance level is equal to the benefit level in Spain. l National average of regionally different allowances. m We assume that the child maintenance levels are comparatively low. According to Corden (1999) few lone mothers in the Southern European countries receive significant amounts of child maintenance. n We assume the child maintenance level (for vignette B) is equal to the level in Finland, because the level in Belgium is closest to the level in Finland for vignette A (see Corden, 1999: 33).

Table 9.1

Female-supportive policies and women’s employment after divorce

241

employment support for women, even lower than the male breadwinner countries. Yet, it deviates from the ideal-typical market model in that it provides relatively high income support for separated women. One reason for this rather generous support may be the comparatively high number of single parents in the UK (Rowlingson, 2001). Italy, Portugal, Spain, Greece, and Ireland are classified within the family model. They have relatively low levels of income and employment support for women, but provide much family support. Our classification of countries into different models of family support deviates from the well-known Esping-Andersen typology in the cases of Ireland (Esping-Andersen typology: in the liberal regime type with the UK) and the Netherlands (Esping-Andersen typology: in the social-democratic type with Denmark and Finland). These changes reflect the role of family ideology and female-supportive institutions in these countries. For example, the Netherlands has an extensive welfare state and generous social welfare programmes, and in most respects, it belongs to the social-democratic regime. However, when it comes to supporting female employment, its institutions lag behind. In Catholic Ireland, family ideology is more in line with that of the Southern European countries than with the market regime in the United Kingdom. It values family support and provides comparatively low income and female-employment support. We wish to further note that not all countries fit neatly into one model, but that some countries are more hybrid types, sharing elements of distinct models of family support. Belgium, for example, belongs to the dual-earner model if we look at the level of public childcare provisions, income support for single parents and child maintenance. However, considering the country’s general social assistance level, child allowances, childcare leave and the traditional family index, we classified Belgium into the male breadwinner cluster. In our analyses, we perform sensitivity checks for the two most hybrid types: Belgium and Ireland. Hypotheses Given the institutional differences in income, employment, and family support – as well as our theoretical assumptions about (post-divorce) labour supply – we can expect to see the following differences in women’s labour supply after divorce (see also Table 9.2). First, we expect that separated women in countries belonging to the dual-earner model will have comparatively high odds of increasing their labour supply after a separation. Women’s employment is highly encouraged by female-friendly policies, especially for women with young children. In addition, there is no large family network present that could help women out by giving them financial support; the state is the main safety net. This also encourages

Comparative studies

242

Table 9.2

Hypotheses on post-divorce labour supply and the effect of young children on post-divorce labour supply

Family support model

Dual earner Family Market Male breadwinner

Work incentive Income support

Employment support

Family support

2 1 2 2

1 0 0 0

1 0 1 1

Divorceeffect on labour supply

Effect of young children on post-divorce labour supply

11 1 0 0

0 2 22 222

female employment after divorce. However, a factor which may discourage female employment is state-provided income support. The dual-earner model provides relatively high amounts of income support – for separated women generally (social assistance levels are high) and for single separated parents in particular (cf. Table 9.1). The high level of income support – although it has changed in recent years and is more often means-tested – may pose a disincentive to work. However, we think that this disincentive from income support will not outweigh the incentives provided by employment support and a lack of family support, since state income support only partially mitigates the income loss after divorce (Uunk, 2004): in other words, women can still improve their situation by working. An additional expectation has to do with the effect of young children: given the generous opportunities for combining work and caregiving, we believe that in general, women’s post-divorce labour supply in the dual-earner model will not suffer much from having young children. Secondly, we expect that separated women in the family model will have lower odds of increasing labour supply than separated women in the dualearner model. The family model provides low state income support, which may be a financial incentive for separated women to work. Yet, women’s employment is not encouraged by female-supportive policies: public childcare levels, for example, are comparatively low. In contrast to all the other countries, an extended family system exists in the family model. Separated women – especially mothers – may be more likely than they are in other countries to fall back on their families for financial help or care in case they need it. However, we do not expect a general discouraging effect of strong family support on employment, since family support can take the form of both financial support (negative effect on labour supply) as well as informal caregiving support (positive effect on labour supply). The well-developed

Female-supportive policies and women’s employment after divorce

243

informal caregiving system within extended families may also give rise to a moderate (negative) child effect on post-divorce labour supply: women in countries belonging to the family model may still be able to continue (or start) working, provided the family helps out. Thirdly, we expect that women in the market model will have comparatively low odds of increasing labour supply after a separation. Income support for single parents (single-parent allowances and child maintenance) is high in the United Kingdom. This provides a disincentive to work. Furthermore, the market model provides low female employment support – another financial disincentive. The low levels of childcare provision will also lead to a more negative effect of young children on postdivorce labour supply (cf. Table 9.2). Fourthly, and finally, we expect that women in countries belonging to the male breadwinner model will have comparatively low odds of increasing their employment after separation. Income support in the male breadwinner model is rather high. Furthermore, this model of family support is characterized by a ‘traditional division of labour’ ideology, and consequently by much lower levels of female employment support from the state. The effect of children will be the most negative of all the models investigated, since institutions favouring the combination of work and caregiving are rather poor, and alternative family caregiving support is much less developed than in the family model.

DATA AND METHOD Data We investigated the labour supply changes of separated women as well as country differences using longitudinal data from the European Community Household Panel (ECHP) 1994–2001 (eight consecutive yearly waves; see Data Appendix). This data-set contains comparable cross-national information on employment, income and several demographic characteristics of the 15 original EU member states. The first wave is a sample of approximately 60 500 households and about 130 000 adults aged 16 years and older. The national samples are representative for the respective countries’ populations. Demographic changes in population over time are reflected in the continuous development of the sample (through births, deaths and the creation of new households). Although panel attrition differs across countries, it is plausible to assume that this does not bias our results. A study conducted by Behr et al. (2005) on panel attrition in the ECHP showed that attrition effects were minimal and did not bias estimates of dynamic analyses of income.

244

Comparative studies

We analysed 13 countries: Denmark, the Netherlands, Germany, Belgium, France, the United Kingdom, Ireland, Italy, Greece, Spain, Portugal, Austria and Finland. For Austria, we have data from 1995 onwards; Finland joined the panel in 1996. We excluded Sweden from our analyses because the data for this country are in the form of pooled cross-sections. Luxembourg was excluded because of the small number of cases. Our sample of analysis consisted of women aged 18 to 65 during the panel period who were not enrolled in educational programmes at the time of the interview and who experienced a (marital) separation during the panel period (N 5 2179). We considered all union dissolutions: hence, women who separated twice (130) during the panel period were analysed twice. Widows were excluded. We do not have information on prior marriages and cohabitations, thus we do not know which of the separations we analysed are first separations and which are later separations. We define separation as a transition from marriage or cohabitation in one year to not living as a couple in the subsequent year. We also included women who remarried or began cohabiting quickly (at least one year after a breakup), because living with a new partner may be a strategy which compensates for the income decline caused by separation, and a life event reducing labour supply. Unfortunately, we do not have information about re-partnering within one year (between pairs of consecutive waves). It is plausible that the labour supply of women who found a new partner quickly is less affected by separation. Method In order to examine to what extent country differences in women’s labour supply changes after separation exist, we carried out three analyses. First, we looked at the transition into employment. Then we investigated exits from the labour market. Finally, we examined changes in working hours for those who had worked before separation. In the first type of analysis, we focused on the odds of entering the labour market for women who did not work before a separation. We modelled this using a discrete-time event history analysis and used a time window of seven years (the year of dissolution and six years after the dissolution, see Figure 9.1). First, we observed the year of union dissolution (t 5 0); then we determined whether the event of employment occurred in each calendar year after separation (t 5 1 2 6). However, we were not able to follow all of the women in the panel for seven years. The reason for this is that we used an unbalanced panel; that is, we included all women who separated during the eight-year panel period and who participated in the panel for at least one year before and one year after a separation. Hence, our risk set

Female-supportive policies and women’s employment after divorce

245

Separation Pre-separation Post-separation labour supply labour supply t = –1

Figure 9.1

t=0

t=1

t=2

t=3

t=4

t=5

t=6

Time window for the analysis of changes in labour supply for women after separation

consists of women who experienced a separation between 1994 and 2000 and who did not work in the calendar year before the divorce. We analysed 668 women who experienced a total of 705 spells of non-employment. Of the 705 spells, 378 women entered the labour market between 1995 and 2001; the 327 remaining spells are right-censored. Due to censoring and attrition, our post-divorce labour supply estimates will mainly concern short-term labour supply changes. In the second type of analysis, we examined the odds of exiting the labour market; we did this in the same way that we conducted the entry analysis. Our risk set was made up of women who worked in the calendar year before they divorced. We analysed 1209 women who experienced a total of 1251 spells of employment. Of the 1251 spells, 270 women exited the labour market between 1995 and 2001; the 981 remaining spells are right-censored. Again, for most women, we only know about short-term labour supply changes because of censoring and panel attrition. The third type of analysis was on the sample of women who were working before a separation and remained working afterwards. Here, we examined to what extent women changed their working hours. Again, we used a seven-year time window (Figure 9.1). We observed the year in which the separation took place (t 5 0), then assessed the number of weekly hours a woman worked for pay in the calendar year before separation (t 5 21), and compared this with the number of working hours in the calendar years after separation (t 5 1 2 6). We used an ordinary least squares (OLS) regression analysis and estimated the number of working hours after separation while controlling for the pre-separation number of hours (for a similar pre/post design, see Allison, 1990). We have valid pre-separation and post-separation data for 1086 separations. For all the analyses, we estimated robust standard errors using the cluster option in STATA. In this way we corrected the standard errors of the macro-level factors for within-country correlation. We performed sensitivity checks by including Belgium in the dual-earner model and Ireland in the market model. The results of these checks are mentioned in footnotes if they differ from the results presented here.

246

Comparative studies

Dependent Variables We measured employment status (employed or not-employed) by asking the respondents whether ‘they were working at present in a job or business normally involving at least 15 hours of work a week’. One shortcoming of the ECHP data is that the data for women working less than 15 hours per week are not comparable with the hours information we have for women who work more than 15 hours. Women who worked less than 15 hours a week only reported the total numbers of hours they worked at their main job and any other jobs if they considered working to be their main activity. Women with small jobs who did not consider working to be their main activity did not provide information on their working hours. Therefore, we considered women who worked less than 15 hours a week as not working and women working 15 hours and more as working. Due to this omission in the data, the average number of working hours and changes in these hours will be underestimated, and the entry and exit numbers will be overestimated. Working hours were derived from the following question: ‘How many hours per week do you normally work in your main job or business?’ We truncated working hours greater than 60 per week (1.1 per cent of preseparation hours, 0.8 per cent of post-separation hours). Independent Variables The most important independent variables in our models are the institutional context and the presence of children. As explained, we distinguished four models of family support. The classification of countries into these four models was presented in Table 9.1. The presence of young children was measured by the age of the youngest child in the household after the separation, and was divided into three categories: 0 5 no child; 1 5 the youngest child is under 6 years old; 2 5 the youngest child is older than 6 (time-varying variable). In the ECHP, people under the age of 16 living with the mother are considered to be dependent children. We took bottom and ceiling effects into account by including women’s pre-divorce working hours. Furthermore, we controlled for some individual-level background variables that are known to influence labour supply: human capital (education and labour market experience), re-partnering, and the women’s ages. The population composition on these individuallevel characteristics could be an alternative explanation for country and model-type differences. Education was measured as the highest level of education achieved in the calendar year before separation. Three levels

Female-supportive policies and women’s employment after divorce

247

are distinguished in the ECHP data, using the International Standard Classification of Education (ISCED): 1 5 less than second stage of secondary education (ISCED 0–2) and those who are still at school; 2 5 second stage of secondary education (ISCED 3); 3 5 university degree or comparable level (ISCED 5–7). Labour market experience was measured differently for the three analyses. In the analysis of employment odds, we used the number of years women did not participate on the labour market. In the analyses of exit and the number of working hours, we used the number of years between the most recent working year (before separation) and the year that women started their working life as a proxy for labour market experience before separation. Re-partnering is defined as the transition from not living as a couple in one year to living together with a new partner in the subsequent calendar year. Age is a time-varying variable and refers to current age. In addition, we controlled for each country’s labour market situation. We did this in order to exclude the alternative explanation that differences in separated women’s labour supply between the four models of family support are due to labour market differences. In particular, we controlled for each country’s unemployment rate and its share of part-time employment among women. The standardized unemployment rate is the number of unemployed persons as a percentage of the total labour force. The annual data are obtained from the OECD, are comparable over time, and conform to the guidelines of the International Labour Office (OECD, 2005). The share of female part-time employment is the percentage of employed women who usually work less than 30 hours per week in their main job. Data included only women declaring usual hours (OECD, 2005).

RESULTS In this section, we examine differences in women’s post-separation labour supply. First, we present descriptive country results concerning labour supply changes after separation. Subsequently, we investigate to what extent individual labour supply changes differ between the four models of family support, controlling for the labour market situation and population composition of each country. A Description of Labour Supply Changes To what extent do women change their labour supply after separation? To answer this question, we will present different descriptive tables and figures showing women’s labour supply in the year before and in the year after

Comparative studies

248

Table 9.3

Employment rates of separated and married women in 13 EU countries Before separation

Dual-earner model Denmark Finlanda Family model Italy Portugal Spain Ireland Greece Market model United Kingdom Male breadwinner model The Netherlands France Belgium Germany Austriab Total (13 countries) X2 (df) 4 models X2 (df) 13 countries

After Net change separation

N

Married women

69.9 64.1 77.2 57.1 60.7 58.0 56.1 40.4 65.4 67.4 67.4 61.3

75.7 75.8 75.6 64.6 72.0 67.0 64.4 44.7 63.5 62.5 62.5 69.2

5.8** 11.8*** 21.6 7.5*** 11.2*** 9.0* 8.3* 4.3 21.9 24.9** 24.9* 8.0***

276 153 123 438 107 100 120 47 52 304 304 839

75.8 77.7 74.3 43.1 41.2 59.6 33.8 40.3 42.4 62.1 62.1 52.9

53.1 54.4 55.8 68.0 73.6

65.0 63.3 64.2 75.2 75.5

11.9*** 8.9* 8.3* 7.2** 1.9

177 158 132 278 106

45.8 54.4 60.9 53.4 53.8

62.6

68.0

5.4***

1857

15.7 (3)** 14.7 (3)** 50.5 (12)** 38.1 (12)**

50.5 8105.0 (3)** 12021.8 (12)**

Notes: a The Finnish survey has only six waves (1996–2001). b For Austria only information for 1995–2001 is available. * significant at 5%; ** significant at 1%; *** significant at 0.1%, one-tailed tested. Source:

ECHP 1994–2001, own calculations.

a separation: employment participation rates (Table 9.3), entry into and exit from the labour market (Figure 9.2), and changes in working hours (Table 9.4). In line with earlier single-country studies (see also ‘Theoretical background and hypotheses’ section), we found that in the 13 countries of the European Union, women increased their employment only modestly after separation. The employment rate before separation was 63 per cent, while after divorce it was 68 per cent, a statistically significant increase of 8 per

249

DK

FI

IT PT ES IE

GR

UK

NL

FR

BE

Entry and exit rates into and out of the labour market for women after separation

ECHP 1994–2001, own calculations.

Figure 9.2

Source:

–20.0

–15.0

–10.0

–5.0

0.0

5.0

10.0

15.0

20.0

25.0

DE

EU

Entry Exit Net change

AT

Comparative studies

250

Table 9.4

Average number of weekly working hours of separated and married working women in 13 EU countries Before After separation separation

Net change

N

Married women

Dual-earner model Denmark Finlanda Family model Spain Italy Portugal Ireland Greece Market model United Kingdom Male breadwinner model the Netherlands Germany Austriab France Belgium

38.3 37.1 39.6 38.0 37.9 37.1 40.8 32.0 37.8 36.3 36.3 37.0

38.3 37.4 39.2 37.8 38.4 37.2 40.7 31.3 35.4 36.6 36.6 37.9

20.04 0.33 20.40 20.16 0.54 0.14 20.11 20.73 22.48 0.24 0.24 0.93**

169 85 84 222 65 64 53 15 25 170 170 455

37.2 34.9 38.9 37.2 38.4 35.5 39.4 31.9 38.2 33.6 33.6 33.9

32.3 39.2 36.6 36.7 38.0

34.5 40.2 37.1 37.1 38.2

2.19** 0.98 0.48 0.38 0.17

86 169 71 69 60

27.6 35.4 36.0 35.7 34.5

Total 13 countries

37.3

37.7

0.42

F-value (df) 4 models 1.8 (3) F-value (df) 13 4.5 (12)** countries

1.3 (3) 4.0 (12)**

1016

35.6 929.3 (3)** 869.2 (12)**

Notes: a The Finnish survey has only six waves (1996–2001). b For Austria only information for 1995–2001 is available. * significant at 5%; ** significant at 1%; *** significant at 0.1%, one-tailed tested. Source:

ECHP 1994–2001, own calculations.

cent (or 5 percentage points; see Table 9.3). However, divorced women in Europe differed notably in employment changes. Divorced women in Denmark, Italy, and the Netherlands stood out with a relatively large increase in post-divorce labour supply (about 11 percentage points); in the UK, on the other hand, separated women actually showed a decrease in employment (also see Jarvis and Jenkins, 1999). This country variation cannot be fully explained by the typology we used. In contrast to our hypotheses, separated women in the dual-earner model did not display consistently higher increases in labour supply than women in the male

Female-supportive policies and women’s employment after divorce

251

breadwinner and family model. The figures show that these three models of family support are on a par, showing similar increases in labour supply (between 6 and 8 percentage points). Our multivariate analyses will show whether this unexpected finding still holds when important individual-level (for example, composition effects of pre-divorce labour supply and education) and country-level characteristics (for example, female unemployment and part-time work availability) are controlled for. That the market model (the UK) ranked most negative with respect to changes in employment participation is more in line with our hypotheses: high income support and low employment and family support in the UK may not pose an incentive for women to work after separation. Notwithstanding these differences between our four models of family support, the figures in Table 9.3 also indicate substantial within-cluster variation. In the dual-earner model, for example, Danish women increased their labour supply after separation by 12 percentage points, yet Finnish women decreased their labour supply, although not statistically significant. Again, however, we warn against drawing premature conclusions before more advanced multivariate analyses have been conducted. The modest changes in labour supply observed above do not tell us the whole story, since a null overall change may result from a large proportion of women entering the labour market and a large proportion leaving the labour market. Figure 9.2 therefore displays entry and exit rates by country. It can be seen that in our sample of European countries, 13 per cent of women entered the labour market after separation, whereas almost 8 per cent left. The proposed typology can only partly explain the existing country differences. While the market model (UK) stood out with comparatively low entry rates and high exit rates, the other models of family support did not systematically differ from each other. The rate of entry was, on average, 14 to 15 per cent in the dual-earner, family and male breadwinner models, and the rate of exit 6 to 8 per cent. Yet, we did find substantial within-cluster variation, for example in the male breadwinner model: while 20 per cent of the divorced French women entered the labour market, in Austria only 8 per cent entered. In the European countries we studied, women who continued working after separation did not significantly change working hours: their average working hours before separation were 37.3 and after separation 37.7 (Table 9.4). This holds for all countries, except for the Netherlands: Dutch women worked (significantly) more hours after a separation, yet the increase was small (from 32.3 hours before separation to 34.5 hours after separation). A reason why only Dutch women showed an increase may be the comparatively short working weeks prior to divorce: this bottom effect will lower the possibility to further decrease working hours. We test this

252

Comparative studies

explanation in the multivariate analyses by taking the pre-divorce working hours into account. Our analyses have shown, so far, that divorced women in Europe modestly increase their labour supply, notably by an increase in the share of working women, not by an increase in working hours for employed women. The general increase in employment participation suggests a positive effect of separation, yet women may also anticipate a divorce and increase their labour supply before the actual separation (Johnson and Skinner, 1986). Moreover, union dissolution may be easier for women who can afford it, an argument pointing to reverse causation (employment affects separation). Many scholars have pointed at this selection effect (Poortman, 2005). Indeed, as Tables 9.3 and 9.4 show, separated women were a select group: on average, they were employed more often and worked more hours than married women (compare first column with last column). The difference in employment rates between these categories was especially high in the family and male breadwinner model: separated women in these countries either anticipated their separations more, or they were the ones who could afford a separation (the ‘elite’ according to Goode, 1966). Anticipation and selectivity weaken the divorce effect on labour supply that we observe, and may consequently ‘disturb’ the observed differences between different models of family support. Again, we must wait for our more advanced analyses to find more conclusive evidence. Multivariate Analyses of Labour Supply Changes Our multivariate analyses are presented in Table 9.5. We started with analysing the odds of labour market entry. The first multivariate model (model 1A) estimates country effects, controlling for individual-level determinants (effects of control variables not shown). We have already observed country differences in post-divorce labour market entry in the descriptive section (‘Data and method’). Model 1A shows that after controlling for composition effects (such as education, and the presence of children), these differences still existed; for instance, Denmark (the reference country) had the highest odds of post-divorce labour market entry, and the UK the lowest. In the second model of Table 9.5 (model 1B), we estimated effects of the institutional context on entry odds. The estimates show that the differences between the four models of family support are more in line with our hypotheses than the descriptive figures of Figure 9.2. The ranking is that women in the dual-earner model had the highest chances to enter employment after separation, followed by women in the family model, then women in the male breadwinner model and, finally, women in the market model. This change in ranking is mainly due to the more negative scores in

253

Belgium

The Netherlands

Male breadwinner

United Kingdom

Market

Ireland

Greece

Spain

Portugal

Italy

Family

–0.800** (0.261) –0.998** (0.335)

–1.159*** (0.259)*

–0.455* (0.349) –0.445* (0.362) –0.692** (0.283) –0.356* (0.451) –0.669* (0.435)

— — –0.577* (0.339)

–0.458* (0.217)

–0.966*** (0.227)**

–0.356* (0.193)

—

Model 1A Model 1B

–0.463** (0.183)

–0.987*** (0.178)

–0.246* (0.152)

—

Model 1C

Odds of exit after separationb

Working hours after separationc

–0.145* (0.116)

–0.056* (0.050)

–0.281* (0.141)

—

–0.510* (0.336) –0.531* (0.389)

0.180 (0.250)

–1.056* (0.507) –0.286* (0.372) 0.040 (0.342) 0.842* (0.384) 0.591 (0.413)

— — 0.059 (0.330)

–0.133* (0.231)

0.188*** (0.053)

–0.019* (0.226)

—

0.025 (0.311)

0.419* (0.235)

–0.037* (0.237)

—

–0.176* (0.321)

0.316 (0.252)

–0.208* (0.280)

—

–1.071* (0.593) 1.059 (0.659)

–0.424* (0.514)

0.303 (0.699) 1.118 (0.711) 0.948 (0.659) –2.203** (0.874) –3.257** (1.087)

— — 0.864 (0.623)

0.005 (0.398)

–0.804** (0.258)

–0.302* (0.661)

—

0.751 (0.599)

0.222 (0.549)

–0.367* (0.595)

—

2.029* (0.912)

2.559*** (0.620)*

–0.660* (0.761)

—

Model 1D Model 2A Model 2B Model 2C Model 2D Model 3A Model 3B Model 3C Model 2D

Odds of entry after separationa

Multivariate analyses of the changes in labour supply of separated women (unstandardized coefficients, robust standard errors)

Dual earner (ref) Denmark Finlanda

Table 9.5

254

Dual earner model * child , 6 years (ref) Family model * child , 6 years Market model * child , 6 years Male breadwinner model * child , 6 years Dual earner model * child > 6 years (ref) Family model * child > 6 years

Child > 6 years

Female part time work No children (ref) Child , 6 years

— –0.520*** (0.161)** 0.002 ** (0.173)**

–0.370* (0.278) *–0.635** (0.263) –0.442* (0.366)

— –0.489* (0.252) 0.027 (0.167)

Model 1A Model 1B

— –0.494* (0.249) 0.035 (0.168)

–0.025* (0.015)

Model 1C

Odds of exit after separationb

Working hours after separationc

— –0.199* (0.404)

— 0.138** (0.309)**

–0.011* (0.012) — 0.551* (0.245) 0.068 (0.080) — 0.697* (0.402) 0.380 (0.288) 0.288 (0.430)

— 1.003*** (0.167)* 0.185 * (0.175)*

–0.011 (0.012) — — 0.935*** 0.911*** (0.165) (0.163)* 0.181 0.152 * (0.155) (0.157)*

0.159** (0.520)** –1.347*** (0.339)** –0.861*** (0.429)**

— 0.001 (0.368) 0.159 (0.308) —

–0.020* (0.015)

0.238 (0.303) 0.281 (0.265) –0.804* (0.388)

— –3.047*** (0.373)** –1.160*** (0.311)*v

–0.366* (0.635) **1.246** (0.525) 0.380 (0.639)

— –2.947* (1.320) –0.976* (0.570)

–0.049* (0.024) — –3.070* (1.332) –1.086* (0.582)

0.125 ** (1.241)**

—

1.902*** (0.721)** –6.099*** (0.193)** –3.625* * (1.886)**

–0.060* (0.025) — –0.752*** (0.172)** 0.313 (0.850) —

Model 1D Model 2A Model 2B Model 2C Model 2D Model 3A Model 3B Model 3C Model 2D

Odds of entry after separationa

(continued)

Unemployment rate

Austriab

Germany

France

Table 9.5

255

1634

1644 –9.2 (9)

1642 1.6 (1)

1626 16 (6)**

–1.235*** (0.292) –0.136* (0.319) 1829

1864 –35 (9) ***

1862 1.9 (1)

1856 6.5 (6)

–0.153* (0.152) 0.387** (0.148) 0.31

0.30 –0.01

0.30 0.00

0.31 0.01

–3.583*** (0.850) –1.879* (0.990)

Source:

ECHP 1994–2001, own calculations.

Notes: a Event history analyses of the odds of entry into employment after separation for women who were not working before separation (number of separations: 705; number of events: 378; number of person years: 1650). b Event history analyses of the odds of exit out of employment after separation for women who were working before separation (number of separations: 1209; number of events: 270; number of person years: 3782). c OLS regression analyses of working hours after separation for women who were working before and after separation (number of separations: 1086, number of person years: 3779). * significant at 5%; ** significant at 1%; *** significant at 0.1%, one-tailed tested. Models control for unemployment rate (entry), share of female part time work (exit, hours), pre-divorce working hours (exit, hours), pre-divorce education, pre-divorce labour market experience (exit, hours), duration not working (entry), repartnering, post-divorce age, whether married or cohabited before separation, whether in school before separation (entry).

-2LL / R2 D-2LL Mii-Mi (Ddf) / DR2 Mii-Mi

Market model * child > 6 years Male breadwinner model * child > 6 years

256

Comparative studies

the multivariate analyses for the countries in the male breadwinner model (Belgium, Germany, and the Netherlands; see model 1A). Apparently, initial (descriptive) changes in employment entry for these countries were obscured by compositional differences in the populations of divorced women. Additional support for the clustering of countries into the four models of family support was provided by a comparison of the statistical fit of the models. The deviance test indicated that model 1B (using the typology) did not perform significantly worse than the country model (model 1A). The third model of employment entry (model 1C), which included the national unemployment rates, showed that only a small part of the institutional differences were due to differences in unemployment: Only the position of the family model changed in a way that the entry odds were closer to that of the reference group of the dual-earner model (no statistical difference anymore). This means that initial lower entry odds in countries of the family model (lower than the reference group of countries in the dual earner model) may be explained by higher unemployment rates in these countries. Our predictions about the interactive effects of the family support model and the presence of young children fared quite well. Model 1D of Table 9.5 shows that in the reference category, the dual-earner model, there existed no significant negative effect of young children on the odds of labour market entry (evidenced by the insignificant effects of the presence of young children). The effect of young children for women in the family model was not significantly different from that of women in the dual-earner model. In other words, in the family model, children did not pose a large hindrance for mothers to enter the labour market. This result may be explained by the extensive informal childcare provided by the (extended) family. In the market and male breadwinner models, the effect of young children on women’s post-divorce labour market entry chances was – as predicted – negative, and significantly different from that in the dual-earner model. However, contrary to our expectations, the negative child effect is larger in the market model than in the male breadwinner model. In the market model, even separated women with children over the age of 6 are hampered by the presence of their children. In our second multivariate analysis, we investigated the country and family support model differences on women’s probability to exit the labour market after a separation (controlling for pre-separation working hours; Table 9.5). The first model (model 2A) shows moderate cross-national differences in exit rates. Only Austria, Greece, and Italy differed significantly from the reference country Denmark: divorced women in Austria and Italy were less likely to exit employment, whereas divorced women in Greece were more likely to exit employment – a similar finding to that in

Female-supportive policies and women’s employment after divorce

257

Figure 9.2. Consequently, institutional effects – as measured by our typology – were also weak. Models 2B and 2C showed that although women in the market model had higher odds of exiting the labour market than women in the dual earner model (the odds were 1.5 times higher), women in the family and male breadwinner model did not differ. The interactive effects of family support model and children on the exit odds (model 2D) did not provide much support for our typology either. The effect of having a child under the age of 6 on the odds of exit is positive in all models of family support, meaning that the presence of a young child increases the likelihood that separated women will leave employment. This child effect is stronger in the family model than in the other models – a surprising finding since the entry analyses showed that the presence of a young child in the family model did not hinder employment entry. The third and final analysis estimated the number of post-separation working hours for women who continued to work after a separation (controlling for pre-separation working hours). The first model (model 3A) displayed country variation in working hour changes after divorce. Given the same pre-divorce working hours, divorced women in Germany worked significantly more hours after divorce than the reference group of Danish women. Divorced women in Greece, Ireland and the Netherlands worked significantly fewer hours after divorce than Danish women. This is in line with the post-divorce working hours observed in the descriptive analyses (see Table 9.4). Model 3B of Table 9.5 showed again that only the market model differed from the other models with respect to post-divorce labour supply: the number of working hours after divorce was lower than elsewhere. Interestingly, this difference disappeared after controlling for the share of female part-time work (which had a negative effect in itself; see model 3C). This means that the lower number of working hours for women in the market model (the UK) was attributable to the relatively high rate of female part-time work; or, stated otherwise, divorced women in the UK worked so few hours after divorce because there were many possibilities to do so. Additionally, other explanations may be that divorced UK women often only qualify for part-time jobs, or that they have difficulties in combining full-time work with caring for their children. The interactive effects of family support model and presence of children on post-divorce working hours provided support for our hypotheses: For women in the dual-earner and family model, the effect of young children on post-divorce working hours is significantly less negative than it is for women in the market and male breadwinner model (model 3D). Separated women in the family model even increased their working hours somewhat when they had a child of less than 6 years old. Thus, the presence of young children had a dual effect on the labour supply for divorced women in the

258

Comparative studies

family model: on the one hand young children formed a restriction because they increased the odds of employment exit; on the other hand young children formed an incentive because they increased employed women’s working hours.

CONCLUSIONS AND DISCUSSION In this chapter, we described and explained changes in labour supply of women after separation. Using data collected from 13 EU countries in the 1990s, we quantified labour supply before and after separation and estimated the effect of the institutional context on the changes in labour supply. First, we conclude that on average, European women changed their labour supply only modestly after separation. The employment rate before separation was almost 63 per cent; after separation it was 68 per cent, an increase of 9 per cent. In the 13 countries of the EU, 13 per cent of women entered the labour market after a divorce and almost 8 per cent left. Working women hardly changed their working hours; the overall increase was insignificant. However, this pattern of labour supply changes differed across countries: in the Netherlands and Denmark, for example, women participated more often on the labour market and worked somewhat more hours after separation, whereas in the United Kingdom women participated less often after separation. Secondly, we conclude that female-supportive policies affected women’s post-divorce labour supply. Independent of compositional differences in the populations of divorced women across Europe, and independent of other macro factors such as the national unemployment rate, we observed systematic differences between distinct models of institutional family support. Women in the dual-earner model experienced the highest odds of employment entry after separation, the lowest odds of employment exit, and the highest number of working hours after divorce for those who were already working before the divorce. Divorced women belonging to the market model experienced the worst labour market outcomes: the lowest odds of entry, the highest odds of exit and the lowest number of working hours for those working. Divorced women belonging to the male breadwinner and to the family model ranked in between these extremes with respect to entry chances, and did not differ from the dual earner model with respect to employment exit and post-divorce working hours. In addition, the four models of family support also differed systematically in terms of the effect of young children on women’s post-divorce labour supply. In the dual-earner and family models, the presence of young children posed less

Female-supportive policies and women’s employment after divorce

259

of an obstacle to separated women to increase their labour supply than it did in the market and male breadwinner model. These differences support institutional explanations of women’s changes in labour supply after separation. In specific, they lend credit to the role played by employment support from the state, not to the role played by income support from the state. The countries of the male breadwinner, market and dual-earner models have comparable levels of income support for separated women, yet these countries differ substantially in labour market outcomes for separated women (mothers in particular). The differences in divorced women’s labour market outcomes are more in line with institutional employment support for women: women in the dual-earner model, a model with ample support for female employment, enter employment after separation more often than women in other models of family support. Apparently, female-friendly employment institutions not only affect the labour supply decisions of (married) women, but also affect women’s decisions after important life events like divorce. This is especially true for single mothers, since institutions can facilitate the combination of work and care for children. One non-state institution that seems to be important is family support. Although in countries of the family model state employment support for women is low, women in these countries did not experience large negative changes in labour supply thanks to the role of extended families in providing informal childcare. Therefore, family support seems to be a substitute for employment-supportive policies. However, financial support received from the family (often in the way that divorcees come to live with other family members) may pose a disincentive effect. It would be interesting for further research to investigate these different and possibly counteracting institutional effects of family support. In our view, explaining divorced women’s labour supply changes by a typology of countries based on family support policies has been fruitful, yet we still are ill-informed about the precise mechanisms that affect post-divorce labour supply. We assume that employment and family support are important, but what is the magnitude of these effects, and what precise institutions matter? Country typologies are not successful in answering these questions. In addition, they suffer from classification problems (assigning countries to clusters is difficult and may influence estimates of the institutional effects – as our sensitivity analyses showed), from lack of dynamics (institutions change), and from within-cluster variation. In principle, questions on the precise mechanisms and magnitude of institutional effects can be better answered in analyses with quantifiable institutional indicators. Such a research design was not feasible in our analyses of 13 countries, but it could be in the near future.

260

Comparative studies

We could wait for data from other countries, but it would also be wise to investigate other ways. One approach would be to take institutional measurements for several countries at several points in time, yet this way is still seriously hampered by the (un)availability of information. Another way is to test the effects of institutional changes using longer running panel data.

NOTES 1. A first argument for this expectation is that the income decline caused by separation is of short duration. Separated women may be able to tide over a short period with other income sources. A second argument is that they may not have found a job yet in such a short period. 2. Sensitivity checks – considering Belgium as a member of the dual-earner model and Ireland as a member of the market model – displayed weaker differences between the four models of family support. Only the market model significantly differed from the other three models of family support in terms of women’s post-divorce entry odds. This indicates that the differences are sensitive to country clustering. 3. When assigning Ireland to the market and Belgium to the dual-earner model, the positive child effect in the family model was smaller and did not significantly differ from the child effect in the dual-earner model.

REFERENCES Allison, P.D. (1990), ‘Change scores as dependent variables in regression analysis’, Sociological Methodology, 20, 93–114. Becker, G. (1965), ‘A theory of the allocation of time’, Economic Journal, 75, 493–517. Behr, A., E. Bellgardt and U. Rendtel (2005), ‘Extent and determinants of panel attrition in the European Community Household Panel’, European Sociological Review, 21 (5), 489–512. Blau, F.D., M.A. Ferber and A.E. Winkler (1998), The Economics of Women, Men, and Work, Upper Saddle River, NJ: Prentice Hall. Bouman, A.M. (2004), ‘Financiele Gevolgen Van Echtscheiding Op De Lange Termijn’, Bevolkingstrends, 52 (4), 85–9. Bouman, A.M. (2005), ‘Arbeidsparticipatie Van Vrouwen Rond De Echtscheiding’, Bevolkingstrends, 53 (1), 90–97. Bradbury, K. and J. Katz (2002), ‘Women’s labour market involvement and family income mobility when marriages end’, New England Economic Review, 2002 (4), 41–74. Corden, A. (1999), Making Child Maintenance Regimes Work, London: Family Policy Studies Centre. Dewilde, C. (2002), ‘The financial consequences of relationship dissolution for women in Western Europe’, in E. Ruspini and A. Dale (eds), The Gender Dimension of Social Change. The Contribution of Dynamic Research to the Study of Women’s Life Courses, Bristol: Policy Press, pp. 81–110.

Female-supportive policies and women’s employment after divorce

261

Duncan, G. and S. Hoffman (1985), ‘A reconsideration of the economic consequences of marital dissolution’, Demography, 22, 485–97. Esping-Andersen, G. (1990), The Three Worlds of Welfare Capitalism, Cambridge: Polity Press. Esping-Andersen, G. (1999), Social Foundations of Postindustrial Economies, Oxford: Oxford University Press. European Commission (1997), Missoc: Social Protection in the Member States of the European Union. Situation on 1 July 1996 and Evolution, Brussels: European Commission. European Commission (2002), Missoc: Social Protection in the Member States of the European Union. Situation on 1 January 2002 and Evolution, http://ec.europa. eu/employment_social/social_protection/missoc_tables_en.htm, accessed 30 October 2007. Finnie, R. (1993), ‘Women, men, and the economic consequences of divorce: evidence from Canadian longitudinal data’, Canadian Review of Sociology and Anthropology, 30 (2), 205–41. Goode, W.J. (1966), ‘Marital satisfaction and instability: a cross-cultural class analysis of divorce rates’, in R. Bendix and S. Lipset (eds), Class, Status, and Power, London: Routledge, pp. 377–87. Gornick, J. and J. Jacobs (1998), ‘Gender, the welfare state, and public employment: a comparative study of seven industrialized countries’, American Sociological Review, 63 (5), 688–93. Gornick, J., M. Meyers and K. Ross (1998), ‘Public policies and the employment of mothers; a cross-national study’, Social Science Quarterly, 79 (1), 35–54. Gough, I., J. Bradshaw, J. Ditch, T. Eardley and P. Whiteford (1997), ‘Social assistance in OECD countries’, Journal of European Social Policy, 7 (1), 17–43. Haurin, D. (1989), ‘Women’s labor market reactions to family disruptions’, The Review of Economics and Statistics, 89, 54–61. Hoffman, S. (1977), ‘Marital instability and the economic status of women’, Demography, 14 (1), 67–76. Jarvis, S. and S. Jenkins (1999), ‘Marital splits and income changes: evidence from the British Household Panel Survey’, Population Studies, 53 (2), 237. Johnson, W. and J. Skinner (1986), ‘Labor supply and marital separation’, American Economic Review, 76 (3), 455–69. Kalleberg, A. and R. Rosenfeld (1990), ‘Work in the family and in the labor market: a cross-national, reciprocal analysis’, Journal of Marriage and the Family, 52 (2), 331–46. Koopmans, Y. and J. Schippers (2003), ‘Female employment and family formation – the institutional context’, paper prepared for the joint meeting of MoCho, FENICS, DynSoc and AGIR, Brussels, 18–20 February. Lewis, J. (1992), ‘Gender and the development of welfare regimes’, Journal of European Social Policy, 2 (3), 159–73. Mandel, H. and M. Semyonov (2006), ‘A welfare state paradox: state interventions and women’s employment opportunities in 22 countries’, American Journal of Sociology, 111, 1910–49. Moore, S. (1978), ‘The short-term effects of marital disruption on the labor supply behavior of young women’, thesis, Ohio State University, Ohio. Mueller, R. (2005), ‘The effect of marital dissolution on the labour supply of males and females: evidence from Canada’, Journal of Socio-Economics, 34 (6), 787. O’Connor, J. (1993), ‘Gender, class and citizenship in the comparative analysis of

262

Comparative studies

welfare state regimes: theoretical and methodological issues’, The British Journal of Sociology, 44 (3), 501. Organisation for Economic Co-operation and Development (OECD) (2001), ‘Balancing work and family life: helping parents into paid employment’, in OECD Employment Outlook, Paris: Organisation for Economic Co-operation and Development. Organisation for Economic Co-operation and Development (OECD) (2005), Employment Outlook 2005, Paris: Organisation for Economic Co-operation and Development. Orloff, A. (1993), ‘Gender and the social rights of citizenship: a comparative analysis of gender relations and welfare states’, American Sociological Review, 58 (3), 303–28. Poortman, A. (2000), ‘Sex differences in the economic consequences of separation: a panel study of the Netherlands’, European Sociological Review, 16, 367–83. Poortman, A. (2005), ‘Women’s work and divorce: a matter of anticipation? A research note’, European Sociological Review, 21 (3), 301–09. Poortman, A. and T. Fokkema (2001), ‘Economische Gevolgen Van Echtscheiding Voor Mannen En Vrouwen in Nederland, 1949–1998’, Sociale Wetenschappen, 44, 69–92. Rosenfeld, R.A. and G.E. Birkelund (1995), ‘Women’s part-time employment: a cross-national comparison’, European Sociological Review, 11, 111–34. Rowlingson, K. (2001), ‘The social, economic and demographic profile of lone parents’, in J. Millar and K. Rowlingson (eds), Lone Parents, Employment and Social Policy: Cross-National Comparisons, Bristol: Policy Press, pp. 169–88. Sainsbury, D. (1996), Gender, Equality, and Welfare States, Cambridge: Cambridge University Press. Smock, P., W. Manning and S. Gupta (1999), ‘The effect of marriage and divorce on women’s economic well-being’, American Sociological Review, 64 (6), 794–812. Stier, H., N. Lewin-Epstein and M. Braun (2001), ‘Welfare regimes, family supportive policies, and women’s employment along the life-course’, American Journal of Sociology, 106 (6), 1731–60. Tietze, W. and D. Cryer (1999), ‘Current trends in European early child care and education’, Annals of the American Academy of Political and Social Science, 563, 175. Uunk, W. (2004), ‘The economic consequences of divorce for women in the European Union: the impact of welfare state arrangements’, European Journal of Population, 20, 251–85. Uunk, W., M. Kalmijn and R. Muffels (2005), ‘The impact of young children on women’s labour supply: a reassessment of institutional effects in Europe’, Acta Sociologica, 48 (1), 41–62. Van Dijk, L. (2001), ‘Macro changes in public childcare provision, parental leave, and women’s employment: an international comparison’, in T. Van der Lippe and L. Van Dijk (eds), Women’s Employment in a Comparative Perspective, New York: Aldine de Gruyter, pp. 37–58. Vogel, J. (2003), ‘The family’, Social Indicators Research: Special Issue, 64 (3), 393–435.

10.

Divorce and housing: a European comparison of the housing consequences of divorce for men and women Caroline Dewilde1

SETTING THE SCENE High divorce rates have long been a major concern for both social science researchers and policy-makers. One of the reasons for this is that divorce entails substantial psychological, social and economic costs (see, for example, Amato, 2000; Kitson and Morgan, 1990), the latter costs being the main focus of this volume. Although in recent years, the economic consequences of divorce have been the topic of a fair number of publications, most authors have concentrated on the impact of divorce on household income. The results of this line of research are fairly consistent: while for men, the impact of divorce on (adjusted) household income is modest and in many cases even slightly positive, for most women divorce results in a substantial decline in household income and an increased likelihood of falling into poverty and/or living on social welfare (see, for example, Burkhauser and Duncan, 1989; Duncan and Hoffman, 1985; Holden and Smock, 1991; Jarvis and Jenkins, 1999; Peterson, 1996; Poortman, 2000; Smock, 1993, 1994).2 This outcome is related to women’s weaker attachment to and position in the labour market, and to the fact that, when there are children involved, it is usually the mother who has custody and is thus responsible for their day-to-day care – the costs of which are only partly compensated for by child support payments. The focus on (women’s) income position has perhaps resulted in a somewhat one-sided picture of the economic consequences of divorce. Several factors contribute to this. First, while in many surveys, the received alimony and child support payments are added to the household income measurement, the amounts paid by the ex-partner are often not recorded, resulting in a too rosy picture of men’s income after divorce. Although the available empirical evidence suggests that deducting maintenance 263

264

Comparative studies

payments does not alter the above conclusions (see, for example, Jarvis and Jenkins, 1999), it is quite conceivable that for those men who take their financial obligations seriously, their ‘real’ available income is significantly lower. In a study of the financial consequences of divorce for American men, McManus and DiPrete (2001: 266) concluded that men’s economic outcomes are heterogeneous, ‘with a majority of losers and a sizeable core of winners’, and that the levelling of men’s income is partly explained by judicial transfers to children and former spouses. Secondly, even assuming that men’s living standards directly after the event of divorce are largely unaffected, research suggests that in the long term, men are suffering as well. For instance, some studies suggest that the health consequences of divorce are more severe for men than for women (Williams and Umberson, 2004). This is mainly attributed to the loss of the so-called ‘marriage premium’: marriage proves to be more beneficial for men than for women in terms of a healthy lifestyle, emotional support and physical comfort. Kalmijn (2005) calls the reverse phenomenon the ‘divorce penalty’, and finds it to have a significant impact on men’s employment and social security histories: after divorce, men experience an increased chance of becoming downwardly mobile, unemployed or disabled. Thirdly, next to the ‘pure’ income effect, divorce entails significant additional economic changes. Perhaps the most important other economic impact concerns the housing situation of the ex-partners. By definition, the dissolution of the couple results in two separate households, meaning that at least one partner has to find another place to live. Setting up a new household comes at a substantial cost – particularly for those people at the lower end of the income distribution, finding new housing which is both affordable and meets certain quality standards is a difficult task. Likewise, the partner who remains in the marital home often has to provide for the full mortgage or rent payments, which can cause severe financial strain and/or result in moving away. Taking all this into account, it is no surprise that studies looking at a wider range of indicators find that divorce results in an increase of lifestyle deprivation for both men and women. For instance, in an earlier study comparing ‘multidimensional’ poverty in Belgium and Britain (Dewilde, 2004), I found that divorce significantly elevates the risk of entering situations of housing deprivation (only for British men), financial stress and limited financial resources for both sexes – the effects for women being somewhat stronger. More recently, Aassve et al. (in this volume) conducted a similar exercise for a larger number of countries, and found that while women are more deprived in both monetary and non-monetary terms, men generally suffer from divorce in non-monetary terms.

Divorce and housing

265

However, in those countries belonging to the so-called ‘market model’ of family support (see Andreß and Hummelsheim, in this volume), differences between men and women are small as both sexes are confronted with increased levels of lifestyle deprivation following divorce. This effect is furthermore stronger for men with children, which the authors link to the higher residential mobility of this subgroup, as their partners more often remain in the marital home in order to provide the children with a stable environment. In this chapter, I focus on one aspect of the wider economic impact of divorce – specifically, the impact of divorce on the housing situation of men and women. Although research on housing and divorce is relatively scarce and often focused on the housing tenure changes following divorce, one of the main conclusions is that divorce often implies a fall down the housing ladder. I hope to contribute to the literature in several ways. First, rather than focusing on one dimension – for instance, the exit out of home-ownership – I will sketch an encompassing picture of the impact of divorce on several aspects of the housing situation, such as changes in ownership status, type of accommodation, housing costs, size and quality of the dwelling and neighbourhood characteristics. The main focus of this chapter is on the short-term housing consequences of divorce. Secondly, this study is, to my knowledge, the only one so far that has been conducted from a comparative perspective, analysing post-divorce housing in several countries. My longitudinal analyses are based on data from the European Community Household Panel (ECHP; see Data Appendix), using all available waves (1994–2001) for 12 European Union member states belonging to different models of family support (see Andreß and Hummelsheim, in this volume): Denmark, Finland, Germany, Belgium, France, Austria, the United Kingdom, Ireland, Italy, Spain, Portugal and Greece. In line with the arguments outlined above, my main expectation is that when it comes to the housing situation after divorce, we can generally expect both men and women to experience a deterioration of their housing conditions. I expect this negative effect to occur in all the countries under consideration. The outline of this chapter is as follows. I start with an overview of the literature on divorce and housing. I then go on to further explore the comparative dimension – more specifically, the place of housing in the welfareproviding nexus consisting of the state, the market and the family and whether the housing dimension is related to the typology of family support outlined in this volume. The next section discusses the data, method and operationalizations, followed by the analyses. I end with an overview and discussion of the main results.

266

Comparative studies

WHAT WE KNOW SO FAR As Murphy (1990: 41) states, ‘in the housing game of snakes and ladders, marital breakdown is a snake’. For the majority of the population, the general direction over the life-course of their progress on the housing ladder is upward. Downward mobility is uncommon, usually involuntary and caused by accidental circumstances such as unemployment and divorce (De Decker and Geurts, 2003; Dieleman and Clark, 1995; Feijten, 2005). The negative effect of divorce on people’s housing careers has been documented in several (often British) studies, usually focusing on changes in ownership status. Most authors report a substantial amount of housing mobility following divorce. For instance, Sullivan (1986) found that overall, about half of British women who experience a divorce leave the marital home at the time of the breakdown. However, older women, women with children and women in local authority housing are less likely to move. The figures reported by Holmans (1990) are relatively similar. Among owner-occupiers, about 60 per cent of the dwellings remain occupied by one of the former spouses: in 35 per cent of cases the wife remains; in 25 per cent it is the husband who does so. For local authority tenants, the situation is reversed, with 50 per cent of women and 20 per cent of men remaining in occupation. Furthermore, the destination tenure for homeowners is strongly gender-biased: while women often end up in the social rental sector, men are more likely to stay in or regain access to owneroccupation (Symon, 1990). On the other hand, for low-income men, the options are very limited. For instance, Sullivan (1986) finds that for young, working-class men, the parental home is an extremely important source of housing. Those men who cannot afford to stay in owner-occupation and who do not find suitable accommodation in the private rental sector have to rely on temporary arrangements such as bed and breakfasts, hostels, caravans or even homelessness. Furthermore, most authors have found that people make the decision to end their relationship first, and postpone the search for suitable housing until after their departure from the marital home (for example, Holmans, 1990; Wasoff and Dobash, 1990). The implication of this is that living with family or friends is a relatively common, but usually short-lived phenomenon. Depending on the time frame adopted, survey estimates range from as much as 60 per cent (Britain, percentage of all respondents, immediately following divorce; Sullivan, 1986) to 18 per cent (the Netherlands, percentage of all movers at the time of the survey; Dieleman and Schouw, 1989). The period immediately following marital breakdown is thus one of

Divorce and housing

267

comparatively high mobility. However, divorce tends to have longer-term consequences as well. Although the number of owner-occupiers who stay in the marital home after a breakdown is surprisingly high – earlier, I mentioned one study where 60 per cent of marital homes remained occupied by one of the former spouses – research finds that during the time between separation and the legal settlement of the divorce, there is further attrition out of the marital home (Symon, 1990). Many respondents, especially women, have difficulties coping with housing costs and maintenance, or find that they do not manage to raise the capital necessary to buy their ex-partner’s interest in the house (McCarthy and Simpson, 1991). Furthermore, people might want to make a ‘clean break’, move closer to family and friends or find a house which better suits their needs. This long-term effect is confirmed by Feijten (2005), who – using Dutch data – found that the risk of leaving owner-occupation disappears over time for men, but stays high for women. The author relates this to women’s lower degree of economic independence. Housing movements after divorce are not only relevant for the persons directly involved, but also have larger implications. Since divorce, by definition, results in the creation of an additional household, the need for extra housing has a non-negligible impact on the housing market. Not all divorced persons eventually remarry. Indeed, according to Dieleman and Schouw (1989: 248), ‘divorce has a cumulative effect on the housing market, because each year more persons enter the population of divorced people than leave it’. This additional demand for housing mainly concerns the lower end of the housing market, resulting in imbalances between supply and demand (McCarthy and Simpson, 1991). Especially in countries with a small rental sector, competition at the lower end of the housing ladder might become fierce as the divorce figure rises over time. In addition to tenure changes, divorce also has a significant impact on housing quality. Many divorced people move from single-family homes to flats, from better situated areas to poorer neighbourhoods and from high-quality houses into more deteriorated dwellings. For those who end up sharing with another household, overcrowding and common use of amenities frequently cause stress and tension (McCarthy and Simpson, 1991; Murphy, 1990; Wasoff and Dobash, 1990). As is clear from the above review of the literature, most studies so far have focused on the impact of divorce on changes in housing tenure – specifically, the exit out of ownership. However, the data-set on which this chapter is based allows for a more thorough analysis of the impact of divorce on the housing situation of men and women, looking at a range of dimensions, including changes in the type of accommodation, housing costs, size, housing quality and characteristics of the neighbourhood.

268

Comparative studies

HOUSING IN A COMPARATIVE PERSPECTIVE Many authors consider housing to be an important, but ‘wobbly’, pillar of the welfare state (Kemeny, 2001; Torgersen, 1987). Housing policy is therefore seldom incorporated into comparative welfare regime research. Various reasons can be given to explain this. First, researchers are confronted with incomplete data and with problems of data comparability for countries with different housing systems. Secondly, housing is a clear example of a policy domain characterized by a mix of principles based on different regimes. For instance, Hoekstra and Reitsma (2002) described the Belgian system as a combination of liberal and corporatist policy arrangements. The private rental sector is liberally oriented, while both the social housing sector and the private sector for owner-occupiers are based on corporatist principles. Thirdly, housing differs from the other welfare state pillars in the fact that it is rarely considered as a universal public provision, since the social rental sector is usually only accessible to a restricted part of the population (Kemeny, 2001). Finally, its capital-intensive and ‘permanent’ character makes housing a less flexible policy domain. Despite this complexity, attempts have been made to identify so-called housing regimes, based on an evaluation of ‘domain-specific’ constellations of institutional arrangements. Whereas in the past, for the sake of convenience, the size of the social rental sector was considered to be an adequate reflection of the state component, recently all aspects of housing policy have been taken into account. For instance, Hoekstra and Reitsma (2002) arrived at the following dimensions: subsidizing, price-setting and price regulation, distribution of the housing supply, the position of the renter versus the landlord, the organization of promotion and production of new dwellings and fiscal arrangements related to housing. Each regime type results in a specific housing system (Barlow and Duncan, 1994). State regulation in the liberal Anglo-Saxon countries is aimed at population groups who are unable to manage on the market. Furthermore, in a more covert way, owner-occupiers are systematically supported. All this occurs within the context of a policy regime that promotes the interests of big building firms and credit institutions. Housing policy in the conservative-corporatist countries of ‘Continental’ Europe has a problem-solving and incremental character, aimed at maintaining the existing social differentiations. Personal initiative is strongly encouraged. In the Scandinavian social-democratic regimes, proper and affordable housing is a universal right. State intervention is not limited to the social housing sector, but aimed at different sectors of the housing market. Finally, in the rudimentary welfare states of Southern Europe, the interference of the state is limited. The ability of households to provide for themselves, often in the

Divorce and housing

269

Table 10.1 Tenure status by country (persons aged 16–64, not living in an extended family, 2001) Tenure status

Dual-earner model

Male breadwinner model

Denmark Finland Germany Belgium Owneroccupier, no mortgage Owneroccupier, with mortgage Social rental sector Private rental sector Rent-free Tenure status

Owneroccupier, no mortgage Owneroccupier, with mortgage Social rental sector Private rental sector Rent-free Source:

France

Austria

5.9

33.7

21.0

33.9

28.9

32.6

69.8

38.7

32.6

46.2

37.1

30.9

11.8

12.8

13.1

5.0

13.8

15.2

12.1

14.0

31.5

13.0

16.6

17.2

0.4

0.9

1.9

2.0

3.5

4.1

Market model

Extended family solidarity

United Kingdom

Ireland

Italy

Spain

Portugal

Greece

20.0

41.6

63.6

59.4

43.6

72.8

58.7

43.1

14.4

27.4

25.7

11.4

13.7

11.0

5.3

1.1

5.2

0.5

6.5

3.5

11.4

7.4

15.5

11.8

1.1

0.7

5.4

4.7

10.0

3.6

ECHP 2001 (own calculations).

family context, takes precedence. The absence of the state in the different sectors of the housing market results in a high level of speculation. In order to briefly sketch the housing market in our sample of countries, Table 10.1 shows the tenure distribution for respondents aged 16 to 64 – the age group which we are focusing on in this chapter. Although owner-occupation is the dominant tenure in all the countries under consideration, between-country differences are fairly large. The number of respondents who are home-owners is the lowest in Germany

270

Comparative studies

(53.6 per cent) and Austria (63.5 per cent), while it rises above 80 per cent in Belgium, Ireland, Spain and Greece. The number of outright owner-occupiers is very low in Denmark (5.9 per cent). In the Southern European countries, on the other hand, outright ownership is a common tenure status, ranging from a ‘low’ of 43.6 per cent in Portugal to a high of 72.8 per cent in Greece. As these figures refer to respondents not living in an extended family, crossnational differences in co-residence cannot account for between-country differences in outright ownership. However, the high percentage of outright owners in the Southern European welfare states does reflect the fact that in these countries, it is the extended family which acts as the natural unit for the pooling of resources and needs (Castles and Ferrera, 1996). This is partly linked to specific circumstances in the housing market. Housing policy in these countries mainly encourages owner-occupation, with only a small social housing segment. Furthermore, rigid systems of rent control have resulted in a gradual shrinkage of the private rental sector. Thus, for many households, owner-occupation is the only available tenure. In addition, given the lack of other investment options and the relative inaccessibility of mortgage credit, ‘family resources, savings and self-build play the role performed by mortgagefinanced house purchase in other countries’ (Fahey et al., 2004: 441). In most countries, between 10 and 20 per cent of respondents live in social housing. However, social housing provision is low (5 per cent or less) in Belgium and the Southern European countries. Germany has the largest number of respondents in private rental housing (31.5 per cent). In the United Kingdom, Ireland and Spain, less than 10 per cent of respondents live in this type of tenure. From Table 10.1, we can conclude that – much to the frustration of housing researchers – institutional variation in the domain of housing does not neatly fit the typology of family support outlined in the Introduction to this book. For instance, the cluster dominated by the male breadwinner model is characterized by different housing regimes. Also, in the market model, a comparatively high number of persons are living in social housing. Furthermore, the housing consequences of divorce are not solely influenced by the housing regime in a specific country, but also by the economic consequences of divorce (which determine the affordability of certain housing types and the options men and women have concerning their post-divorce housing situation), by the legal traditions concerning the division of marital property (in case of owner-occupation, mothers might be ‘compensated’ for the economic consequences of divorce by assigning them the marital home) and by the extent of family support, both financial and in terms of co-residence. Consequently, it is difficult to formulate more specific predictions about how the housing consequences of divorce for men and women vary across European countries.

Divorce and housing

271

DATA, METHOD AND OPERATIONALIZATIONS Data As stated in the introduction, this chapter uses data from the European Community Household Panel (ECHP). The ECHP is a cross-nationally comparative longitudinal survey containing data from the 15 original European Union member states. It is a so-called household panel study (HPS); that is, a standardized questioning of an initial sample of households and individuals at regular – in this case annual – points in time. In this way, a prospective database is created (Rose, 2000; for example, Ruspini, 2002). The sample of households and individuals is representative of the population in each of the participating countries. Cross-national comparability is guaranteed through a standardized design and common technical and implementation procedures, with centralized support and coordination of the national surveys carried out by Eurostat (Eurostat, 2003). In the first wave of the ECHP (1994), a sample of more than 60 000 households and approximately 130 000 adult respondents were interviewed. The participating countries were Belgium, Denmark, France, Germany, Greece, Ireland, Italy, Luxembourg, the Netherlands, Portugal, Spain and the United Kingdom. Austria joined the ECHP in 1995 and Finland in 1996. Information for Sweden is available from 1997 onwards. Data for this last country, however, are not longitudinal but pooled cross-sections, and as such do not qualify for our purposes. Furthermore, in the Dutch sample, there seems to be a problem with the variables pertaining to the residence status of households and persons, which makes it impossible to accurately determine housing movements. Due to the small sample size, I also excluded Luxembourg, leaving a sample of 12 countries. Analyses for Germany and the United Kingdom are based on the German Socio-Economic Panel (GSOEP) and the British Household Panel Survey (BHPS), which were adapted to the ECHP format. Although panel attrition differs across countries, it is plausible to assume that this will not bias our results. Studies conducted by Watson (2003) and Behr et al. (2005) show that attrition effects are minimal. Method As mentioned above, ECHP data are collected in a prospective way, allowing us to use a longitudinal research design. This way, problems of selectivity are avoided and the causal effect of divorce on changes in housing conditions can be determined with a higher degree of certainty. Indeed, the group of divorcees in a cross-sectional survey consists of men and women

272

Comparative studies

who experienced relationship dissolution over a longer period of time. Some respondents divorced many years ago, while others only recently split up with their partners. Furthermore, since most people eventually repartner, it is possible that the group of divorced respondents who remain single is quite a selective sub-sample of all people experiencing relationship dissolution. The starting point for my prospective analysis was the construction of a sample of men and women who experienced partnership dissolution between 1994 and 2001. Although the number of divorces on a yearly basis is rather small, the ‘pooling’ of events in a prospective way allowed me to construct a large enough sample to allow for multivariate analysis. Divorce is defined as the transition from marriage or cohabitation in one year (t) to not living as a couple or living with a new partner in the subsequent year (t 1 1), thus resulting in the formation of two separate households. My sample contains both married and cohabiting respondents at time (t). The choice for a de facto measure of partnership dissolution is based on several arguments. First, a growing number of people live in consensual unions, especially in the Scandinavian countries (Kiernan, 2003). Secondly, housing movements following partnership dissolution usually coincide with the de facto ending of the relationship rather than with the formal divorce settlement, which can take several years. Restricting the analysis to respondents aged 16 to 64 who were living as part of a heterosexual couple at time (t), the number of divorce events amounts to 4466 (1958 men and 2508 women). Since I am comparing the housing situation of respondents directly before and after the separation of the couple into two separate households, my main focus will be on the short-term consequences of divorce. Since the ECHP does not contain information on prior marriages or cohabitation relationships, both first and higher-order divorce events are included in the sample. Furthermore, some respondents experienced more than one divorce over the course of the panel. However, this concerns only a very small group (4.2 per cent of the total sample), so I just included them in my sample. Operationalizations As stated in the introduction, in this chapter we are interested in the short-term impact of divorce on the housing situation of men and women, looking at a range of dimensions including housing tenure movements, changes in the type of accommodation, housing costs, size, housing quality and characteristics of the neighbourhood. For most countries, this information is available in the ECHP.

Divorce and housing

273

Regarding changes in housing tenure, I have focused on two transitions which, in the literature on housing and divorce, are considered to be a form of downward mobility: the exit out of ownership and the entry into shared accommodation. The latter category is defined in terms of ‘loss of independence’. Although most people in this category have changed residence – mostly by moving back to the parental home – those respondents who stay at the same address but in some way or another enter a situation of sharing (for example, a parent or another related or non-related adult moving in) are also included. Re-partnering, however, is not considered as sharing. Changes in the type of accommodation following a divorce are explored by mapping different possible transitions built around the – in the literature, assumed – quality differences between living in a single-family house and living in a flat. Housing costs are expressed as a percentage of total household income. However, since the income concept in the ECHP refers to the yearly total household income of all household members during the previous calendar year, in order to track changes in the cost of housing due to relationship dissolution, I have compared housing costs at time (t) with those at time (t 1 2). In order to supplement this ‘objective’ information with the more subjective perception of the respondents, I have also reported the number of respondents who stated that housing costs are a heavy burden, comparing the situation before and after divorce. The impact of divorce on the size of the dwelling is measured by two indicators: the average number of rooms before and after divorce, as well as the average number of rooms per person (in both cases, not counting kitchens, bathrooms and toilets). For reasons of respondent confidentiality, the variable referring to the number of rooms is top-coded with a maximum value of 6. This might result in an underestimation of the number of rooms for the respondents in large households (and thus larger houses). However, top-coding the number of household members to 6 or more does not alter our results (figures available from the author). I measured the housing quality before and after divorce by means of a nine-item summation index of the following items (a situation of deprivation is scored as ‘1’, otherwise ‘0’): no bath or shower; no indoor flushing toilet; no hot running water; shortage of space; accommodation too dark/ not enough light; lack of adequate heating facilities; leaky roof; damp walls, floors, foundations; and rot in window frames or floors. Changes in neighbourhood characteristics were evaluated by means of a three-item summation index: noise from neighbours or outside; pollution, grime or other environmental problems caused by traffic or industry; and crime or vandalism in the area. Owing to reasons of data availability, neither

Comparative studies

274

Table 10.2

Number of divorces, response rates after divorce and mover status by country

Country

Men

Denmark Finland Germany Belgium France Austria United Kingdom Ireland Italy Spain Portugal Greece Total sample

Women

N

Response rate*

% moved

N

Response rate

% moved

208 201 326 136 255 88 283

59.3 68.1 60.6 57.1 58.9 52.7 48.9

62.5 69.7 58.9 45.6 49.0 39.8 67.5

241 206 413 184 300 115 434

68.7 69.8 76.8 77.3 69.3 68.9 75.0

68.1 75.7 55.9 54.9 54.7 43.5 45.2

27 177 95 103 59 1958

37.0 58.6 47.3 59.5 48.0 56.4

48.2 67.2 67.4 49.5 52.5 58.9

60 213 145 117 80 2508

82.2 70.5 72.1 67.6 65.0 72.2

23.3 58.7 40.0 39.3 40.0 53.3

Note: * Number responding at t 1 1 as percentage of the total number of men/women experiencing divorce between t and t 1 1. Source:

ECHP 1994–2001 (own calculations).

indicator of housing quality could be calculated for Germany or the United Kingdom.

ANALYSIS I begin my overview of the results with some descriptive figures concerning the number of divorces in each country, the response rates after divorce and the extent of residential mobility following divorce. Next, using a ‘pre/ post’ research design, I will evaluate the overall impact of divorce on the dimensions of the housing situation identified above. Descriptives Table 10.2 provides an overview of the sample sizes in each country, the response rates after divorce and the overall housing mobility in terms of change of residence after divorce. As is typical for this type of study, the

Divorce and housing

275

sample sizes are rather small but not problematic, except perhaps for Ireland and Greece. The response rate, defined as the percentage of respondents still cooperating after an observed divorce,3 is, in most countries, higher for women than for men. Overall, 72.2 per cent of women remained in the panel, compared with 56.4 per cent of men. This gender difference is also found in other studies (see, for example, Jarvis and Jenkins, 1999). The lowest response rate for men was found in Ireland (37.0 per cent); at the same time, however, Ireland had the highest response rate for women (82.2 per cent). Finland had the highest percentage of cooperating divorced men (68.1 per cent). The lowest number of women who responded following a divorce was found in Greece (65.0 per cent). In most countries, a comparable proportion of men (overall average of 58.9 per cent) and women (overall average of 53.3 per cent) changed their residence immediately after a divorce. Overall, residential mobility was highest in Finland and Denmark, while percentages were low in Austria (both sexes), Ireland (women), Spain (women), Portugal (women) and Greece (women). Gender differences were relatively large in the United Kingdom, Ireland and the Southern European countries, with more men changing residence than women. In Belgium, the situation was reversed, although the gender difference was somewhat smaller. The Impact of Divorce on the Housing Situation Next, we will look at the impact of divorce on several dimensions of the housing situation. I focus on changes in tenure status, accommodation type, housing costs, housing quality and characteristics of the environment. In Table 10.3, I compare the housing tenure mobility of men and women experiencing divorce to respondents in a ‘stable’ couple between wave (t) and wave (t 1 1), focusing on the exit out of ownership and entry into shared accommodation. Compared to owner-occupiers in a ‘stable’ household, men and women experiencing divorce have a much higher chance of leaving the tenure status of homeowner. Furthermore, we see a regional divide. In the dual-earner countries and in the welfare states dominated by the male breadwinner model, more women than men exit out of ownership. In the market model and in two of the countries characterized by a tradition of extended family solidarity – Italy and Spain – the situation is reversed. This is not, however, the case for Portugal and Greece. Nevertheless, in both family support models, women in particular have a comparatively low likelihood of losing their owner-occupied homes. Perhaps the inadequate protection from economic risks following divorce for women in these countries (see, for example, Uunk, 2004) is compensated for by different procedures of property settlement. Also, we saw earlier that in the

Comparative studies

276

Table 10.3

Housing tenure movements following divorce for men and women compared with respondents in stable couples Exit out of ownershipa

Country

Denmark Finland Germany Belgium France Austria United Kingdom Ireland Italy Spain Portugal Greece

Entry into shared accommodationb

‘Stable’ couples

Divorced men

Divorced women

‘Stable’ couples

Divorced men

Divorced women

1.2 1.4 1.7 0.9 1.1 2.3 0.8

27.5 37.5 47.7 28.7 27.9 24.2 24.3

41.3 43.9 52.4 41.1 48.0 63.0 20.5

0.6 0.7 1.0 0.6 0.7 0.7 1.1

2.5 6.4 4.5 4.9 6.0 4.3 9.2

2.5 1.5 1.9 4.6 8.3 5.7 4.6

0.9 1.8 2.8 2.2 1.6

41.4 30.8 47.7 9.5 16.8

12.5 26.7 14.7 22.9 19.4

1.2 1.1 1.5 1.8 0.9

0.6 8.4 14.4 4.0 10.5

5.9 2.6 13.1 4.5 2.9

Notes: a Risk group restricted to owner-occupiers. b Risk group restricted to respondents not in shared accommodation at time t. Source:

ECHP 1994–2001 (own calculations).

Southern European countries more houses are owned outright – albeit with help from the extended family – which, of course, makes it easier for the remaining partner to stay in his or her house. Finally, the relative absence of alternatives on the housing market in both models of family support may explain the comparatively low exit rates. Exit rates for homeowners following divorce are high for both sexes in Finland and Germany. Austrian women have the highest chance (63.0 per cent) of leaving owner-occupation following a divorce. As we saw before, both Austria and Germany have relatively large rental sectors. Compared with respondents who are part of a stable couple, divorced men and women also have higher chances of entering some form of shared accommodation. In 7 out of 12 countries, entry rates are higher for men than for women. Entry rates are highest for men in the United Kingdom, Italy, Spain and Greece. Female entry rates into shared accommodation are rather high in France and Spain. In summary, entry into shared accommodation is more common in the countries characterized by extended family solidarity.

Divorce and housing

277

Next, I will look at changes in the type of accommodation following divorce for the risk group of respondents who changed their residence (see Table 10.4). Bearing in mind that the frequency distribution of possible transitions is dependent on the country-specific distribution of accommodation types and thus not strictly comparable across countries, we find that ‘downward’ changes in the type of accommodation (that is, from a single-family house to a flat and from a flat to ‘other’ accommodation) are relatively common, although more so in some countries than in others. Again, we find a rough divide between Ireland and the Southern European countries on the one hand – where ‘downward mobility’ of this kind seems to be less – and the Northern European countries. Especially in Germany, Belgium and France, many divorced men and women moved from a singlefamily home to a flat. Although not as many respondents moved from a flat to some kind of ‘other’ accommodation, this transition is not negligible either, and seems generally more common for men than for women. Finally, we can note that a not unimportant minority of divorced men and women managed to rise on the housing ladder following divorce and move from a flat to a single-family home. Table 10.5 explores the changes in housing costs following divorce. For both sexes, divorce results in a significant increase in the cost of housing, expressed as a percentage of household income. However, this increase is much greater for women than it is for men (note that any housing allowances have been deducted from the total housing cost). The increase in housing costs is particularly steep in Portugal and in Austria (for both sexes), and for women in Finland, Germany, Belgium, Italy and Spain. According to the more subjective indicator as well, in most countries, housing costs are more often perceived as a heavy burden following a divorce. The impact of divorce is, however, somewhat smaller according to this indicator. Again, divorce has a somewhat larger effect for women compared with men. The impact of divorce on the perceived cost of housing is most evident for women in Denmark, Belgium, France, Austria, the United Kingdom, Spain and Portugal. The next dimension of housing quality concerns the size of the new home for divorced men and women who change their residence (Table 10.6). As between-country differences are rather small, I will focus on the general picture. On average, we find that for men, the number of rooms decreases from 3.6 to 2.9, while for women it goes down from 3.7 to 3.1. The stronger decrease for men is, of course, related to the fact that they generally do not have custody of the children (if there are any). When we look at the changes in the number of rooms per person, we find that men actually experience an increase in the size of their dwellings: in most cases, they are single and even if their new home is small, it does provide enough space to live

278

Source:

Note:

a

Germany

24.1 36.1 29.3 4.0 6.5

Belgium

France

Austria

15.8 50.8 27.4 5.5 0.6

Ireland

18.0 50.2 19.0 8.5 4.3

7.9 47.7 35.2 6.5 2.7 Italy

10.4 37.0 32.7 19.9 0.0

39.4 16.8 35.9 5.1 2.8

Spain

19.9 23.5 45.1 4.2 7.4

20.3 36.2 34.9 5.0 3.6 Portugal

11.2 43.6 31.9 7.8 5.6

3.6 52.8 33.2 1.2 9.2 Greece

15.6 62.7 5.0 16.1 0.7

46.3 11.3 28.0 5.6 8.8

57.2 11.5 16.7 8.4 6.2

37.8 32.3 19.8 0.0 10.1

53.0 17.5 21.7 0.0 7.9

19.8 34.7 16.5 18.8 10.1

14.1 51.2 15.3 15.9 3.5

ECHP 1994–2001 (own calculations).

8.0 59.6 10.3 22.2 0.0

13.5 54.7 16.2 15.6 0.0

45.7 16.8 11.2 18.1 8.2

50.7 13.8 31.6 4.0 0.0

13.2 57.1 29.6 0.0 0.0

13.3 45.1 16.0 22.0 3.7

Men Women Men Women Men Women Men Women Men Women Men Women

United Kingdom

22.8 39.1 22.5 5.3 10.4

Risk group restricted to respondents not in shared accommodation at time t 1 1.

‘Stable’ single-family house ‘Stable’ apartment or flat From single-family house to flat From flat to single-family house From single-family house or flat to ‘other’ accommodation

Type of accommodation

‘Stable’ single-family house ‘Stable’ apartment or flat From single-family house to flat From flat to single-family house From single-family house or flat to ‘other’ accommodation

Finland

Men Women Men Women Men Women Men Women Men Women Men Women

Denmark

Type of accommodation before and after divorce for men and women changing residencea

Type of accommodation

Table 10.4

279

Source:

Note:

a

21.7 22.3 26.4 19.1 18.3 13.5 19.3 10.3 10.2 12.8 8.6 7.5

After (t 1 2) 113.6 113.8 118.9 114.4 113.7 140.6 120.6 222.6 10.34 14.9 195.5 115.4

% change 19.2 19.5 24.5 14.6 17.4 13.7 17.2 11.6 7.9 7.0 3.7 5.5

Before

ECHP 1994–2001 (own calculations).

Any housing allowances are deducted from the total housing cost.

19.1 19.6 22.2 16.7 16.1 9.6 16.0 13.3 7.6 12.2 4.4 6.5

Before

Men

21.5 27.3 33.6 23.0 20.2 19.7 21.4 14.3 11.7 11.1 11.7 5.6

After (t 1 2)

Women

112.0 140.0 137.1 157.5 116.1 143.8 124.4 123.3 148.1 158.6 1216.2 11.8

% change

Average housing costsa (% of household income)

Housing costs before and after divorce for men and women

Denmark Finland Germany Belgium France Austria United Kingdom Ireland Italy Spain Portugal Greece

Country

Table 10.5

6.0 17.4 20.3 32.3 19.6 7.7 13.0 25.5 43.4 35.7 37.4 36.8

Before

Men

11.5 19.4 24.6 31.6 26.4 8.2 13.5 21.8 39.1 39.6 36.6 25.7

After

9.2 25.0 24.1 30.7 22.5 10.6 12.1 50.9 45.0 36.7 23.5 39.1

Before

17.1 22.3 24.1 51.7 37.0 29.1 21.1 49.1 47.6 46.5 47.3 36.1

After

Women

Housing costs are heavy burden (% change)

Comparative studies

280

Table 10.6

Size of the dwelling before and after divorce for men and women changing residence

Country

Average number of rooms Men

Women

Average number of rooms/ person Men

Women

Before After Before After Before After Before After Denmark Finland Germany Belgium France Austria United Kingdom Ireland Italy Spain Portugal Greece Source:

3.6 2.9 3.6 4.0 3.5 4.0 4.4 4.0 3.8 4.1 3.8 3.2

2.9 2.4 2.5 2.7 2.7 3.0 3.4 2.5 2.9 3.6 2.6 2.5

3.8 3.3 3.5 4.3 3.6 4.1 4.3 4.2 3.5 3.8 3.9 3.2

3.0 2.4 2.5 3.3 3.2 2.8 3.7 3.6 3.1 3.7 3.0 2.9

1.3 1.0 1.2 1.3 1.2 1.5 1.5 1.3 1.0 1.4 0.9 1.0

2.3 2.0 1.9 1.8 2.1 2.1 2.8 2.1 2.0 2.6 2.0 1.4

1.4 1.2 1.3 1.4 1.2 1.3 1.5 1.2 1.0 1.3 1.0 0.9

1.7 1.6 1.6 1.9 1.6 1.6 2.0 1.9 1.7 1.6 1.3 1.2

ECHP 1994–2001 (own calculations).

comfortably. For women, we also find an increase in the average number of rooms per person, although it is less impressive. For both sexes, however, part of this ‘positive’ effect is caused by the fact that single persons now have their own bedroom. I also checked the size of the dwelling following divorce for persons entering shared accommodation. As it turns out, women entering shared accommodation are the only ones who experienced a (small) decrease in the average number of rooms per person, from 1.24 to 1.18. This is in line with the ‘forced’ nature of this type of accommodation reported in the literature (McCarthy and Simpson, 1991; Murphy, 1990; Wasoff and Dobash, 1990). We can conclude that for both sexes, the impact of divorce on the size of the dwelling seems rather small. Finally, I will look at the changes in the quality of the dwelling and the characteristics of the neighbourhood following divorce (risk group of men and women changing residence). Again, we learn from Table 10.7 that the differences between the countries in our sample are rather small. Furthermore, for both men and women, divorce does not seem to significantly raise the level of deprivation. On average, men have a score of 0.87 before divorce, which actually goes down to 0.73 after divorce. For women, we find a similar decrease, from 0.92 to 0.78. We find the same trend when

Divorce and housing

281

Table 10.7 Housing quality before and after divorce for men and women changing residence Country

Quality of the dwelling Men

Neighbourhood characteristics

Women

Men

Women

Before After Before After Before After Before After Denmark Finland Germany Belgium France Austria United Kingdom Ireland Italy Spain Portugal Greece Source:

0.7 0.6 — 1.1 1.0 0.3 — 0.4 0.7 0.7 3.4 1.7

0.6 0.4 — 0.9 0.9 0.3 — 0.7 0.9 0.5 2.8 1.1

0.6 0.6 — 1.1 1.0 0.5 — 1.4 0.9 1.0 3.0 1.9

0.8 0.4 — 0.9 0.9 0.6 — 0.7 0.7 0.6 1.6 1.5

0.5 0.9 — 0.7 0.7 0.3 — 0.8 0.6 0.8 0.8 0.5

0.3 0.7 — 0.6 0.5 0.3 — 0.4 0.8 0.4 0.5 0.4

0.5 0.8 — 0.7 0.6 0.4 — 0.2 0.7 0.6 0.5 0.4

0.4 0.9 — 0.6 0.5 0.4 — 0.0 0.7 0.5 0.5 0.4

ECHP 1994–2001 (own calculations).

we look at the characteristics of the neighbourhood. The average score on this indicator decreases from 0.67 to 0.52 for men, while it goes down from 0.63 to 0.58 for women. Although these results seem counter-intuitive, they are in line with research suggesting that the risk of marital breakdown is higher for those who experience the worst housing conditions and live in the least desirable tenure statuses (Murphy, 1990). According to Holmans (1990), part of the explanation for this lies in the association between early marriage and high divorce risks, since couples who marry young are more likely to be tenants and have a higher chance of divorce. Nevertheless, it means that housing conditions also have an ‘independent’ effect on the risk of divorce. Although divorce perhaps results in a lower quality of dwelling and local environment for some respondents, for others the event of divorce actually results in better housing conditions.

CONCLUSION AND DISCUSSION In this chapter, I looked at the impact of divorce on the housing situation of European men and women, using longitudinal data from the ECHP (1994–2001) for 12 countries. Rather than focusing on changes in housing

282

Comparative studies

tenure status, I looked at a range of dimensions such as changes in the type of accommodation, housing costs, size and quality of the dwelling, and neighbourhood characteristics. My main hypothesis was that, although the income consequences of divorce fall mainly on the shoulders of women (and the children they look after), the economic consequences of divorce are more encompassing and concern both men and women. Especially in the case of housing, divorce, by definition, results in the formation of two separate households and thus affects both sexes. Since I used data from a larger number of countries in this chapter, I also sketched a picture of the housing market in order to interpret cross-national variations in the impact of divorce on housing. In line with previous single-country studies, I find that divorce substantially raises the risk of leaving owner-occupation for both men and women. However, this risk seems smaller for women in those countries where the economic consequences of divorce are more severe – that is, in the market model and the countries characterized by extended family solidarity. I related this to the fact that women in these countries might be compensated for the economic consequences of divorce by more advantageous procedures of property settlement. However, other explanations are possible as well. In the Southern European countries, outright ownership is more common, which makes it easier to stay in the marital home. Furthermore, alternative forms of housing are limited, so that divorced men and women might be forced to stay in owner-occupation even if they cannot afford it. Further research will have to clarify this issue. I also found that for both sexes divorce increases the risk of entering shared accommodation. As far as the changes in the type of accommodation following divorce are concerned, I found that for a substantial number of respondents, divorce initiates a process of downward mobility on the housing ladder. In Germany, Belgium and France in particular, many divorced men and women move from a single-family home to a flat. In most countries, a nonnegligible number of respondents move from a flat to some ‘other’ type of accommodation. However, this transition is more common for men. Again, I find that ‘negative’ changes in the type of accommodation are less common in Ireland and the Southern European countries. Here, too, the lack of alternative housing forms or the high percentage of outright ownership in these countries (thanks to the help from the extended family) could explain this regional divide. Although divorce results in a substantial increase in housing costs, this effect is larger for women than it is for men. Contrary to my expectations, the effects of the indicators concerning housing quality (size of the dwelling, quality of the dwelling and neighbourhood characteristics) are small – or even changed in the opposite direction. I interpreted this finding by

Divorce and housing

283

pointing out that housing conditions are also an important determinant of divorce: Housing deprivation significantly increases the risk of divorce. It is thus possible that the aggregate trend is blurred by two opposite underlying tendencies: for some, divorce might result in worse housing conditions, while for others it might signify an escape out of a situation of housing deprivation. I conclude that in line with my main hypothesis, divorce has important, often negative, consequences on the housing situation of both men and women. This process seems to be influenced by macro-level arrangements such as the structure of the housing market, the extent of help provided by the extended family and the legal tradition considering the division of marital property. However, further (multivariate) research will have to identify the most important determinants.

NOTES 1. European Community Household Panel data were made available through the Research Centre for Longitudinal and Life Course Studies (CELLO), University of Antwerp. Eurostat bears no responsibility for the analyses and interpretations presented here. The author would like to thank Pascal De Decker and the participants of the Second Workshop on the Economic and Social Consequences of Partnership Dissolution (3–4 November 2006, Cologne) for useful comments and suggestions. 2. See also the literature discussed in Andreß et al. (in this volume). 3. Divorced couples where both partners dropped out of the survey are not included in this figure. Since neither of the two partners lived in an interviewed household, the divorce remained unobserved.

REFERENCES Amato, P.R. (2000), ‘The consequences of divorce for adults and children’, Journal of Marriage and the Family, 62 (November), 1269–87. Barlow, J. and S. Duncan (1994), Success and Failure in Housing Provision. European System Compared, Oxford: Pergamon. Behr, A., E. Bellgardt and U. Rendtel (2005), ‘Extent and determinants of panel attrition in the European Community Household Panel’, European Sociological Review, 21 (5), 489–512. Burkhauser, R.V. and G.J. Duncan (1989), ‘Economic risks of gender roles: income loss and life events over the life course’, Social Science Quarterly, 70 (1), 3–23. Castles, F.G. and M. Ferrera (1996), ‘Home ownership and the welfare state: is southern Europe different?’, South European Society and Politics, 1 (2), 163–85. De Decker, P. and V. Geurts (2003), ‘Belgium’, in J. Doling and J. Ford (eds), Globalisation and Home Ownership. Experiences in Eight Member States of the European Union, Delft: Delft University Press, pp. 21–48. Dewilde, C. (2004), ‘Vormen en trajecten van armoede in het Belgische en Britse welvaartsregime. Multidimensionele armoededynamieken bestudeerd vanuit

284

Comparative studies

de sociologie van de levensloop’, Proefschrift voorgelegd tot het behalen van de graad van Doctor in de Politieke en Sociale Wetenschappen, Universiteit Antwerpen. Dieleman, F.M. and W.A.V. Clark (1995), ‘Falling out of the home owner market’, Housing Studies, 10 (1), 3–15. Dieleman, F.M. and R.J. Schouw (1989), ‘Divorce, mobility and housing demand’, European Journal of Population, 5, 235–52. Duncan, G.J. and S.D. Hoffman (1985), ‘A reconsideration of the economic consequences of marital dissolution’, Demography, 22 (4), 485–97. Eurostat (2003), ‘ECHP UDB Manual. European Community Household Panel Longitudinal Users’ Database. Waves 1 to 8. Survey Years 1994 to 2001’, DOC. PAN 168/2003–12, Luxembourg: European Commission. Fahey, T., B. Nolan and B. Maître (2004), ‘Housing expenditures and income poverty in EU countries’, Journal of Social Policy, 33 (3), 437–54. Feijten, P. (2005), ‘Union dissolution, unemployment and moving out of homeownership’, European Sociological Review, 21 (1), 57–71. Hoekstra, J.S.C.M. and A.A. Reitsma (2002), ‘De zorg voor het wonen. Volkshuisvesting en verzorgingsstaat in Nederland en België’, Volkshuisvestingsbeleid en woningmarkt 33, Delft: DUP Science. Holden, K.C. and P.J. Smock (1991), ‘The economic costs of marital dissolution: why do women bear a disproportionate cost?’, Annual Review of Sociology, 17, 51–78. Holmans, A.E. (1990), ‘Housing demand and need generated by divorce’, in P. Symon (ed.), Housing and Divorce, Glasgow: University of Glasgow, Centre for Housing Research, pp. 52–76. Jarvis, S. and S.P. Jenkins (1999), ‘Marital splits and income changes: evidence from the British Household Panel Survey’, Population Studies, 53, 237–54. Kalmijn, M. (2005), ‘The effects of divorce on men’s employment and social security histories’, European Journal of Population, 21, 347–66. Kemeny, J. (2001), ‘Comparative housing and welfare: theorising the relationship’, Journal of Housing and the Built Environment, 16, 53–70. Kiernan, K. (2003), ‘Cohabitation and divorce across nations and generations’, CASE paper 65, London: Centre for Analysis of Social Exclusion. Kitson, G.C. and L.A. Morgan (1990), ‘The multiple consequences of divorce: a decade review’, Journal of Marriage and the Family, 52 (November), 913–24. McCarthy, P. and B. Simpson (1991), Issues in Post-Divorce Housing, Aldershot: Avebury. McManus, P.A. and T.A. DiPrete (2001), ‘Losers and winners: the financial consequences of separation and divorce for men’, American Sociological Review, 66 (April), 246–68. Murphy, M. (1990), ‘Housing consequences of marital breakdown and remarriage’, in P. Symon (ed.), Housing and Divorce, Glasgow: University of Glasgow, Centre for Housing Research, pp. 1–51. Peterson, R.R. (1996), ‘A re-evaluation of the economic consequences of divorce’, American Sociological Review, 61 (June), 528–36. Poortman, A.-R. (2000), ‘Sex differences in the economic consequences of separation. A panel study of the Netherlands’, European Sociological Review, 16 (4), 367–83. Rose, D. (ed.) (2000), Researching Social and Economic Change. The Uses of Household Panel Studies, London: Routledge.

Divorce and housing

285

Ruspini, E. (2002), Introduction to Longitudinal Research, London: Routledge. Smock, P.J. (1993), ‘The economic costs of marital disruption for young women over the past two decades’, Demography, 30 (3), 353–71. Smock, P.J. (1994), ‘Gender and the short-run economic consequences of marital disruption’, Social Forces, 73 (1), 243–62. Sullivan, O. (1986), ‘Housing movements of the divorced and separated’, Housing Studies, 1, 35–48. Symon, P. (1990), ‘Marital breakdown, gender and home ownership: the owneroccupied home in separation and divorce’, in P. Symon (ed.), Housing and Divorce, Glasgow: University of Glasgow, Centre for Housing Research, pp. 110–38. Torgersen, U. (1987), ‘Housing: the wobbly pillar under the welfare state’, in B. Turner, J. Kemeny and L.J. Lundqvist (eds), Between State and Market: Housing in the Post-Industrial Era, Göteborg: Almqvist and Wiksell. Uunk, W. (2004), ‘The economic consequences of divorce for women in the European Union: the impact of welfare state arrangements’, European Journal of Population, 20, 251–85. Wasoff, F. and R.E. Dobash (1990), ‘Moving the family: changing housing circumstances after divorce’, in P. Symon (ed.), Housing and Divorce, Glasgow: University of Glasgow, Centre for Housing Research, pp. 139–66. Watson, D. (2003), ‘Sample attrition between waves 1 and 5 in the European Community Household Panel’, European Sociological Review, 19 (4), 361–78. Williams, K. and D. Umberson (2004), ‘Marital status, marital transitions, and health: a gendered life course perspective’, Journal of Health and Social Behavior, 45 (March), 81–98.

11.

The economic consequences of partnership dissolution: a comparative analysis of panel studies from Belgium, Germany, Great Britain, Italy and Sweden Hans-Jürgen Andreß, Barbara Borgloh, Miriam Bröckel, Marco Gießelmann and Dina Hummelsheim

INTRODUCTION Research on the economic consequences of partnership dissolution has shown that income losses outweigh income gains both in terms of family and per capita income, but these losses are not absorbed equally by both partners. Women and their (dependent) children are most often the losers, while men experience only moderate income losses or even improve their economic status. Table 11.1 summarizes the results for some selected countries.1 According to these data women’s income changes range between minus 80 per cent in one study for Germany and plus 7 per cent in another study for Belgium. Men’s income changes, on the other hand, show less variability. They range between minus 23 per cent in one study for Germany and plus 6 per cent in the only study for Canada. Furthermore, all studies analysing both men and women clearly demonstrate that women’s income is much more negatively affected than men’s. For example, the study by DiPrete and McManus (2000), which shows the most negative results for men (Germany: 223 per cent, USA: 215 per cent), demonstrates that women’s income decline is still about 10 percentage points larger. Despite these overall regularities, the variance of findings across countries is striking. For a comparativist, these country differences come as no surprise. Besides many individual factors, it can be expected that the institutional context determines the economic consequences of separation and divorce. It makes a difference, for example, whether the institutional 286

The economic consequences of partnership dissolution

287

Table 11.1 Percentage change of adjusted household income due to separation and divorce in selected countries Country

Author(s)

Typea Period

Total n Percentage changeb Women

B D

DK NL I SWE

UK

USA

CAN

Dewilde (2002) Uunk (2004) Burkhauser et al. (1990, 1991) Sørensen (1994) Andreß et al. (2003) DiPrete & McManus (2000) Dewilde (2002) Uunk (2004) Dewilde (2002) Uunk (2004) Poortman (2001) Uunk (2004) Dewilde (2002) Uunk (2004) Sørensen (1994) Gähler (1998) Fritzell (1990) Jarvis & Jenkins (1999) Dewilde (2002) Uunk (2004) Burkhauser et al. (1990, 1991) Sørensen (1994) DiPrete & McManus (2000) Finnie (1993)

Men

2a 2d 2b

1994–96 1994–2000 1983–86

n.a. 63c 101

7 220 244

n.a. n.a. 27

1 3

1984 1984–99

n.a. 808

280 227

n.a. 24

2c

1984–96

n.a.

233

223

2a 2d 2a 2d 3 2d 2a 2d 1 3 3 3

1994–96 1994–2000 1994–96 1994–2000 1984–95 1994–2000 1994–96 1994–2000 1981 1980–90 1973–80 1991–94

n.a. 157c n.a. 41c 349 69c n.a. 69c n.a. 137 n.a. 253

26 225 224 213 231 219 213 23 263 19 136 218

n.a. n.a. n.a. n.a. 4 n.a. n.a. n.a. n.a. 133 152 2

2a 2d 2b

1994–96 1994–2000 1981–85

n.a. 145c 540

227 236 224

n.a. n.a. 26

1 2c

1986 1981–93

n.a. n.a.

271 226

n.a. 215

3

1982–86

5875

240

6

Notes: n.a.: no information or not analysed. a 1 5 comparison of countries based on cross-sectional data, 2x 5 comparison of countries based on longitudinal data, 3 5 national longitudinal data. b The values of these columns can, strictly speaking, only be compared within one and the same study, because the studies use different statistical measures (medians, means, regressionbased estimates) and different operationalizations of income change. c Number of cases for women only.

288

Comparative studies

setting encourages women’s labour force participation, whether the labour market provides enough opportunities for gainful employment of both genders, whether state and legal system offer rights and financial support for dependent family members and how the kinship network operates in case of partnership failure. More generally speaking, it is the national configuration of family, market and state institutions that moderates the economic effects of partnership dissolution. Although such institutional explanations are both plausible and popular, they have not been tested very often with comparative studies applying the same methodology on equivalent longitudinal data in each country. Most previous studies were conducted on a national level and sometimes even use only cross-sectional data (cf. Table 11.1). So far there exist only two longitudinal country comparisons including more than three countries (Dewilde, 2002; Uunk, 2004). Besides that, research on the consequences of partnership dissolution is often plagued by the small number of separations observed in population-wide survey samples. For example, one of the first comparative analyses for Germany (Burkhauser et al., 1990, 1991) is based on 56 women and 45 men. As a consequence, country rankings of separation effects on household income are not very stable and observed differences between countries may also be a result of unreliable estimates. Furthermore, different studies use different methods of analysis, as Table 11.1 notes, which is another source of variation that makes substantive conclusions difficult. Finally, many studies focus on women alone, although inequalities resulting from partnership dissolution are most clearly visible by comparing men’s and women’s income development. In sum, there is still a demand for comparative analyses based on greater numbers of cases than in previous studies for both genders and using the same methodology for each country. Such a comparative analysis is most easily done with a cross-national survey like the European Community Household Panel (ECHP) that uses equivalent instruments in each country. Unfortunately, the current version of the ECHP includes only a small number of separations and divorces for each country as the analyses by Uunk (2004) and Dewilde (2002) show. Instead, this chapter uses a similar approach to Burkhauser et al. (2001), who succeeded in creating a cross-national equivalent file from several independent national household panels. Starting from our typology of family support in the Introduction we select five countries (Belgium, Germany, Great Britain, Italy and Sweden) in which such panel studies exist since several years. Even if each panel wave observes only few separations and divorces, it is possible to collect a sufficient number of cases over time with such long-running surveys. Another advantage of this approach is that each separated individual is represented in the data with a long time-

The economic consequences of partnership dissolution

289

series of observations before and after separation. Using suitable statistical methods this enables us to study not only the short-term (before–after) changes found in most other studies, but also the long-term consequences. It is an important question, both from a practical and a theoretical point of view, whether income losses can be compensated in the years following separation and what factors help to do so. Finally, unlike other studies we do not only concentrate on the economic consequences for women, but have a look at both genders. The chapter is organized as follows. In the next section we refer to the typology of family support which has been explained in the Introduction. We present data about the five selected countries and ask whether these countries are valid representatives of the typology. The third section introduces our hypotheses, and the fourth our data and methods. The results are presented next, in two steps. First, a descriptive overview about income changes before and after separation is given both for men and women in all selected countries. This overview suggests a certain impact function how separation affects individual income trajectories. In a second step, this impact function is included in a multivariate panel model and the hypotheses from the third section are tested sequentially. We conclude with a discussion of the main results in light of the proposed typology of family support.

HOW TO LOCATE THE FIVE COUNTRIES IN A TYPOLOGY OF FAMILY SUPPORT? To what extent families and partnerships are supported and promoted by state and society varies among different countries. In the Introduction, we have distinguished four different models of family support: the male breadwinner, the dual earner, the market and the model of extended family solidarity. We assume that these four types of family support create specific dependencies within the family, which in case of separation or divorce may have negative economic consequences for the weaker partner. Using a term introduced by Orloff (1993), we ask how much autonomy is granted to the weaker family members within each of the four models.2 Obviously, gainful employment and own income are the basis of financial autonomy. On the basis of the proposed typology it can be supposed that there is little gender-specific division of labour in the dual-earner model. Both partners have more or less equal chances to combine work and family life. In the two models of rudimentary family policy (that is, in the market and the family model), a definite form of labour division is not explicitly encouraged by the state. It is rather left to the partners to coordinate employment and

290

Comparative studies

childcare. The male breadwinner model ranks in between, because on the one hand a traditional division of labour is supported by granting tax privileges for marriages with one principal earner and by not offering sufficient full-time childcare facilities. Therefore full-time employment for a married mother is inefficient or utterly impossible. On the other hand, this model affords a compatibility of childcare and work based on part-time employment of the mother. Like every typology these four models of family support are ‘Idealtypen’ in a Weberian sense. They represent hypothetical configurations of state, market and family, while the real world of nation states uses mixtures and combinations of these models. Our comparative analysis is based on data from Belgium, Germany, Great Britain, Italy and Sweden. This selection was partly determined by the availability of adequate longitudinal data and partly because preliminary analyses with aggregate data (Andreß, 2003) indicated these countries as possible representatives of the four models of family support. In the following, we give a brief account of each country’s welfare mix and based on this information allocate each country to our typology of family support.3 Belgium Social policy in Belgium is family orientated, but its main focus is the traditional family with a male principal earner (see also Mortelmans et al., in this volume). There are no special measures that help individuals coping with the economic consequences of separation and divorce or with the situation of being a single parent (Ruspini, 2000: 224). Extensive and generous family benefits contribute to quite an extent to disposable household income – independent of family constellation. Thus, child allowances in Belgium constitute a large part of the income package (see Data Appendix) and social assistance benefits are the highest out of the five selected countries (OECD, 2004). The appreciation of the family, in the way it is expressed by family benefits provided by the state, indicates that the family is a solid institution in Catholic Belgium (Fix, 2000: 308). The social security system also has a strong family orientation in terms of benefits and supplements. Besides financial transfers, the state offers an excellent childcare system, which even includes care for toddlers. About one-third of all children aged up to 3 years and almost all children from 3 to 6 years attend some kind of childcare facility (see Data Appendix). In keeping with the preceding observations it can be said that Belgium has an extensive family policy: on the one hand, with good support in terms of money for the traditional male breadwinner model, which would suggest classifying Belgium as a representative of the male breadwinner

The economic consequences of partnership dissolution

291

model. On the other hand, there is a well-organized childcare infrastructure that enables women to work so that Belgium could be assigned to the dual-earner model (Andries, 1996, 1997). Nevertheless, from an empirical point of view, women are not as often employed as could be expected. From 1985 to 2000 about 37 to 52 per cent of all women were employed (see Data Appendix). This is the smallest female employment rate in comparison to the other countries (besides Italy). Thus, Belgium somehow fits into both models. Germany Germany belongs to those countries with an extensive family policy. It acts on the assumption of traditional labour division and compensates for familial burdens (see also Hummelsheim, in this volume). Although policy orientation has changed in recent years, family policy is focused on the traditional two-parent family with an ‘at-home’ mother caring for the children and, therefore, offers strong support for the male breadwinner model (Daly, 2000: 81). Family policy has been generous in terms of benefits such as child allowances, social assistance benefits and tax reliefs since the mid-1990s. The focus on a male breadwinner family can be seen in different life domains. Joint taxation for married partners and the so called ‘Ehegattensplitting’, in which married couples are allowed to split their incomes equally between both partners for tax purposes, impose strong incentives to combine a large primary income, usually the breadwinner’s or man’s income, with a comparatively small secondary income. Another example of strong support for the male breadwinner model is the limited German care system for children, disabled persons and the elderly. Only a small proportion of children aged up to 3 years attend a childcare facility (see Data Appendix). Traditionally, day care is organized for older preschool children and on a part-time basis, intended only to complement care work within the family. Because of uncoordinated school and work hours and the lack of full-time day schools (Büchel and Spieß, 2002), these problems continue when children enter school. This places severe restrictions on the employment opportunities for – in most cases – women (Ruspini, 1999: 95). Furthermore, the generous parental leave system (in terms of time and money) motivates parents to retreat from the labour market for several years and care for their children (Sainsbury, 1999: 194), at least until they can enter a pre-school childcare facility. Since it is predominantly women who make use of the parental leave system, female employment in Germany is moderate and to a large extent part-time employment (Sainsbury, 1999: 131; see also Data Appendix).

292

Comparative studies

As a result, Germany has the lowest percentage of dual-earner and the highest percentage of single-earner households, especially for families with children (Daly, 2000: 94). In conclusion, Germany is an adequate example for the male breadwinner model. Great Britain A ‘no interference’ policy and a low level of public sector involvement characterize the welfare system in Great Britain (Sims-Schouten, 2002: 270; see also Sigle-Rushton, in this volume). Therefore, there is no explicitly formulated policy on how to support families. Nevertheless, quite a few measures are taken in Great Britain to support families in different living conditions (Schmid, 1996). Because government mainly intervenes in situations of crisis or dysfunction, support is means-tested rather than universal and the extent of support is limited. In most areas, at least during the 1980s, Great Britain showed only substandard support through monetary transfers and tax credits (Neubauer, 1993). In the case of separation or divorce there are no special benefits granted by the state. Financial provision following divorce is a private matter, the amount of maintenance is rather low and often the liable parent does not pay the required sum at all. Public childcare provision in Great Britain is poor, especially for toddlers (see Data Appendix). Besides the low provision, the financial costs of childcare are often high. In spite of insufficient opportunities to combine family and work, female labour market participation is high. Nearly two-thirds of all British women are gainfully employed, which is the second highest employment rate in our comparative analysis (see Data Appendix). The restrictions on mothers’ employment, however, show in high rates of part-time working women, which seems to be the only way to combine family and work without risking a longer career break (Ringen, 1997). During the 1990s more than 40 per cent of all active women were working part-time, which is the highest rate among the five countries (see Data Appendix). Additionally, women have to be satisfied with less income than men. Women’s earnings as share of men’s earnings are lower than in Belgium, Germany and Sweden (see Data Appendix). We conclude that Great Britain has a rudimentary family policy, focusing on private arrangements according to individual market resources. Therefore, Great Britain represents the market model. Italy The Italian welfare state contains elements of different prototypes, with the family as the main source of social security (The Clearinghouse on International Developments in Child, Youth and Family Policies, 2003:

The economic consequences of partnership dissolution

293

3). Provision of child care facilities for toddlers (below three years of age) is poor, but the state provides care facilities for older children (over 3 years of age) and generous parental leave arrangements that at least in principle allow adult family members to participate in the labour market independent of their parental obligations (Dienel, 1993: 249). In this respect, Italian family policy includes elements of the dual-earner model. However, these policy measures are not always available at the local level or not accepted by many families. As a consequence, a traditional gendered division of labour is still very common (Dienel, 1993: 251). One reason for this contradictory result is that modern aspects of family policy are embedded in a traditional welfare system that is based on the idea of extended family solidarity. The Italian constitution obliges all citizens to provide alimony and care for family members in need.4 Thereby, the state determines ‘family’ as the basic unit of the welfare state. These institutionalized interdependencies motivate a (gendered) division of labour within the family (Saraceno, 2000: 346), although provision is made for public childcare facilities and mothers need not stay at home. In addition, by granting tax deductions for dependent spouses and providing access to free health care for families with household incomes below a certain threshold, the state provides monetary incentives for potential secondary earners not to work. Thus, it is harder for secondary earners to establish an autonomous economic basis. After partnership dissolution, women are more likely to benefit from infrastructural efforts: the state privileges employees and single parents in allocating childcare facilities (The Clearinghouse on International Developments in Child, Youth and Family Policies, 2003: 6). Furthermore, data on labour force participation reveal that there is good provision of jobs for divorced women. Whereas only 30 per cent of all married women are gainfully employed, 60 per cent of all divorced women have a job (Eurostat, 2003). Contrary to the divorced women, male breadwinners are able to extend and stabilize their sources of economic independence after partnership dissolution. Furthermore, the system of post-marriage alimony is rudimentary compared with the other countries’. Claims are not determined by the standard of living experienced during marriage (Timoteo, 1995: 287). Italy can be assigned to the model of extended family solidarity, because of its rudimentary family policy and its strong focus on family networks in case of economic problems. Sweden Sweden may be characterized as a country with an extensive family policy (see also Hiilamo, in this volume). Support for children and compensation

294

Comparative studies

for the economic costs of child-rearing have been the focus of Swedish policy for decades. The extent of financial support for families is generally very high. Among the five countries, Sweden (together with Belgium) spends the highest share of gross domestic product (GDP) on cash benefits to families (see Data Appendix). There are no special benefits for separated or divorced individuals except an advanced maintenance scheme for children. Theoretically, the law provides spousal maintenance, but in fact it is unusual to make such claims (Saldeen, 1995: 493). Furthermore, various life transitions, which are critical in terms of income, are financially supported by the state: for example, founding a home, giving birth to a child and educating children. There is a lot of support for extended, paid and job-protected parental leave. Fathers explicitly have the right to paternity leave after the birth of a child (this option was introduced in 1974) and part of parental leave is reserved for each parent, so that both partners have incentives to take part (Ministry of Health and Social Affairs, 2002: 2). The Swedish welfare system encourages the employment of women as well as the reconciliation of work and family life for women and men. The Swedish system traditionally focuses on dual-earner families. Women’s employment is nearly as high as men’s, even among married individuals with children below the age of 6 (Sainsbury, 1999: 217; see also Data Appendix). On the one hand, this is due to a socially accepted consensus of gender equity; on the other hand, to a well-established system of childcare, which provides places for almost half of all toddlers and nearly all infants (see Data Appendix). Nevertheless, at least one-third of all economically active women work part-time (see Data Appendix), which in Sweden means about 30 hours a week. Part-time work is a female domain in Sweden. Furthermore, the Swedish system of individual taxation generally encourages women to enter the labour market, and wage differences between men and women are small. Sweden may be said to fit the dualearner model of our typology.

HYPOTHESES It is no surprise that separation often leads to increased financial strain for the involved partners. On the one hand, they lose the advantage of household sharing and using common resources. Sørensen (1994: 178) has estimated these losses of economies of scale between 27 and 36 per cent. On the other hand, both partners have to deal with new costs including, for example, expenses for legal procedures or for setting up a new home. Furthermore, economic inactivity of one partner can no longer be absorbed by the ‘family-income’ of the main earner. Correspondingly,

The economic consequences of partnership dissolution

295

most empirical research on separation and divorce shows that partnership dissolution is combined with income changes for both genders. In line with this research, we expect that H1a separation or divorce causes income changes for both partners. However, research on separation and divorce has also shown an unequal distribution of these income changes. Especially the economically more dependent partner has to deal with income losses rather than income gains. Women in general or the partner who cares for the children after separation (also usually the wife) can be characterized as the economically more dependent partner. Thus, we assume that H1b women – compared with men – experience more often and more severe income losses. Basically, two mechanisms contribute to these inequalities (Holden and Smock, 1991; see also the Introduction). (1) Income sharing during partnership allows women to participate in men’s market incomes, which are on average higher than women’s market incomes. Maintenance payments after separation, if they exist at all, hardly compensate for the loss of this economic support. (2) If children exist, economic needs are higher for the resident parent after separation and this, in most cases, is the mother. She has to care for more economically dependent household members and, because of this, she is often restricted in her earnings capacities. Besides that, the risk that the resident parent does not receive half of the child maintenance costs from the non-resident liable parent is high. Apparently, the gender gap in post-separation incomes would not exist if women’s employment were as high as men’s, if women would achieve the same level of income as men, if childcare were no restriction for gainful employment (or would otherwise be remunerated) and if the costs of childcare were equally shared by both parents. In such a perfect world, women and men would be economically autonomous. However, from our theoretical discussion we learned that gender-specific roles in partnerships exist in most European countries and therefore women often have fewer resources than men. Besides that, the national context is more or less supportive for women’s economic autonomy. Based on the discussion in the previous section, Table 11.2 presents some stylized facts about the amount of cash transfers to families, the availability of public childcare, and the extent of women’s employment in Sweden, Belgium, Germany, Great Britain and Italy. Table 11.2 ranks the countries with respect to each indicator and the level of overall economic autonomy of women within each country.

Comparative studies

296

Table 11.2

Women’s economic autonomy (country ranking) Sweden Belgium Germany Great Italy Britain

Economic autonomy Cash transfers

3 5 high, 2 5 middle, 1 5 low Family cash benefits Child benefits Social assistance Public childcare 0–3-year-olds 3–6-year-olds Women’s Employment rate employment Low part-time employment

3

2

2

2

1

11

0

2

0

22

0 0 11 0 11 1

11 1 1 1 2 0

2 2 22 0 0 0

2 0 22 22 1 22

22 22 22 1 22 11

Notes: Rank compared to five-country-mean: 22 very low, 2 low, 0 average, 1 high, 11 very high.

Sweden scores highest on this index of economic autonomy because most indicators show above average values. Italy for the same reasons scores lowest, while it is hard to differentiate Belgium, Germany and Great Britain, which range in between Sweden and Italy. Belgian women, for example, show below average employment rates, although the state offers an excellent infrastructure of childcare. If they gain economic independence, they do so because of generous cash transfers. Cash transfers and childcare infrastructure in Germany are lower than in Belgium, but German women are more often gainfully employed and may, therefore, experience the same amount of economic autonomy as Belgian women. The same is true for British women, where employment rates are even higher than in Belgium and Germany, although employment to a much larger degree is part-time, possibly because of the meagre system of childcare. If, as this summary suggests, economic autonomy of women is highest in Sweden, the gender gap in post-separation incomes should be the smallest in Sweden. This should also be true for Swedish mothers, not only because of the extensive childcare infrastructure in Sweden, but also because Swedish men more often than men from other countries live with children after separation, as we will see in our later empirical analysis. H2 In Sweden the gender gap in post-separation income is low, while gender inequality with regard to income is high in Germany, Great Britain and Belgium, and especially high in Italy.

The economic consequences of partnership dissolution

297

Up to now, our hypotheses focus on the average woman and the average man in each country. This is clearly a simplification, because the concomitant income changes during the course of separation depend on many individual circumstances: whether the person is employed before and/ or after separation, whether he or she has to care for children, how he or she copes with the new situation, which may be easier for younger and more educated individuals, and so on. All these individual factors should be controlled for as far as possible within the comparative framework. Since we use independent national household panels, we must limit the following analysis to a few selected control variables, namely education, age, employment status and parental obligations, which we were able to measure equivalently between the different data-sets (for more details see the following section). Even if H2 turns out to be true, the question remains whether this is a temporary or a permanent phenomenon. In other words, does the economic position of women recover to former pre-separation levels or does it remain on a level below pre-separation standards? When a partnership breaks down, support from the extended family and the welfare state may form a safety net securing a minimum standard of living. However, in the long run, only own gainful employment guarantees that separated individuals participate in the income prospects of the general population. Another important factor may be re-partnering. In our empirical analysis we distinguish between short- and long-term income changes for women. Given the already high labour market participation of Swedish women and the generous cash transfers of the Belgian welfare state for families and individuals in need (cf. Table 11.2), we expect the short-term consequences of separation for women in both countries to be comparatively low. The opposite is true for Italy where the safety net at best is built up by the extended family and not by the welfare state with its rudimentary benefits for families. Therefore, concerning negative shortterm consequences, we expect the following country ranking: H3

Italy (highest), Germany, Great Britain, Belgium, Sweden (lowest).

Concerning the long-term consequences, again Swedish and Belgian women seem to be better off, either because their participation rate is already high or because an extensive infrastructure of childcare is available that, at least in principle, allows mothers’ employment. However, high female participation rates before separation in Sweden imply that increased employment after separation will induce only minor income gains. Similarly, it is difficult to make a prognosis for Belgium. On the one

298

Comparative studies

hand, the childcare system clearly provides employment opportunities for mothers. On the other hand, generous family benefits, as they are typical for the Belgian welfare state, may as well be an incentive for mothers to stay at home. In contrast to Sweden and Belgium, average female participation rates in Germany and high part-time employment in Great Britain do make increased employment after separation a rewarding option, especially because cash transfers in both countries are comparatively low (cf. Table 11.2). Therefore, H4

we have no a-priori hypothesis about which country scores best concerning the long-term consequences of separation.

DATA AND METHODOLOGY Sample For a detailed analysis of the economic consequences of partnership dissolution it is essential to use longitudinal data. We therefore created a thorough database of partnership dissolutions composed of different national household panels. The challenge was to make these different data-sets comparable across countries. In many respects this worked out fine, in some respects – due to limitations of the data – we had to be content with less perfect solutions. The Data Appendix describes the main features of the five national household panels used for creating our database. The Belgian, German and British panel study offer a broad range of different income data, household information and socio-demographic characteristics of each interviewed person. They were easily integrated into our cross-national data-set, while the Italian and Swedish data were more problematic. The Bank of Italy Survey of Households’ Income and Wealth, due to the initiator’s interest, contains detailed information on income, but has rather poor data with respect to household composition or socio-demographic characteristics. Besides that, the panel section of this survey, which we use for our study, concerns only a small subsample of individuals, while the main survey consists of a series of repeated crosssections. Furthermore, individuals participating in the panel section do not contribute more than three observations and then drop out of the panel. It seems as if not much emphasis is put on keeping the panel members motivated to take part continuously. The Swedish panel, possibly because of unsteady funding, shows a lack of comparability over different panel waves, so that in some waves income data or data about the situation of the household are missing.

The economic consequences of partnership dissolution

299

Our group of interest are men and women experiencing partnership dissolution. Separation is defined either as separation of a consensual union or separation of a married couple.5 For married couples we focus on separation and not on divorce because previous research (Andreß and Güllner, 2001) has shown that separation is connected to more economic changes than legal divorce, which follows separation sometimes several years later, when the economic situation has already stabilized. Separations were identified across the national panel data using similar criteria. If possible we combined information about marital status, different life events and questions about household composition. In some surveys only one of these types of information is available (for example, the Italian data include only marital status), in others (for example, in Sweden) the information varies across waves, because not all questions are asked each survey year. ‘t’ defines the year in which these separations are observed. Since we want to compare the situation before and after separation, data of at least one partner have to be available for years t 2 1 (before) and t (after separation).6 If an individual fits these criteria it becomes part of our study and all available panel information about this individual from the years before and after separation is included in our data-set (plus data from the year of separation). Furthermore, we restrict the analysis to individuals between 18 and 60 years of age to avoid a commingling of income changes due to separation and income changes due to retirement. Our data include 403 Belgian, 1437 German, 1144 British, 111 Italian and 353 Swedish men and women, each contributing between 2 and 16 observations from their panel biography. Table 11.3 shows some of their characteristics. As expected, we found only few cases of separation in Italy. Because of the few observations per individual in the Italian panel it is difficult to look at income changes over time. Therefore, Italy is not included in our multivariate analyses.7 Another limitation of the Italian data concerns the type of partnership: it is only possible to identify marital separations, but not breakups of consensual unions. For Sweden we include both, but because of the limited information on marital status in the Swedish data we cannot tell these two apart. For the other countries where it is possible to differentiate between consensual unions and marriages, the separations we look at are more often marital breakups than dissolutions of cohabitation. Furthermore, Table 11.3 shows that Italian and Swedish men and women are older on average than separating individuals in Germany, Belgium or Great Britain at the time of separation. For Italy this is probably caused by the fact that separations included in the data-set consist of marital disruptions only, and people are usually older when separating from marriage than when separating from a consensual union. Besides that, British and

Comparative studies

300

Table 11.3

Separated men and women in the national household panels Belgium

Germany

Great Britain

Italy

Sweden

M

W

M

W

M

W

M

W

M

W

Separations (N) 181 of which: Marriages 94 Consensual 87 unions Age at separation 36 (mean) High education 37 before separation (%) Employment before separation (%) Not employed 14 Part-time 4 employed Full-time 82 employed Employment after separation (%) Not employed 14 Part-time 2 employed Full-time 84 employed 31 After separation living with childrena

222

684

753

470

674

60

51

183

170

126 96

405 309

444 309

263 207

383 291

60 0

51 0

–b –

– –

34

35

33

35

33

43

40

40

40

30

17

13

43

28

8

6

17

26

36 17

13 4

28 19

15 29

31 38

7 2

49 10

13 8

26 27

48

84

52

56

31

92

41

80

47

38 12

15 3

28 16

20 27

31 37

10 –

35 12

9 9

18 23

51

83

56

53

33

90

53

82

59

90

28

93

25

94

14

94

44

75

Notes: a Data include only separated fathers and mothers. Percentages add up to over 100%, because parents can have joint custody, because children from a new partner are included and because only one parent can be present in the data (see note in the text). b In the Swedish data marital dissolution cannot be distinguished from separation of a consensual union. Source:

Cross-national data set of separations.

The economic consequences of partnership dissolution

301

Belgian individuals are more highly educated than German, Swedish and, especially, Italian individuals.8 Taking together part-time and full-time employment, Swedish (separated) women have the highest and Belgian and Italian women the lowest employment rates in our data-set. It is striking that Great Britain is the only country where almost one-third of the men work part-time – before and after a separation. Women also show the highest part-time employment in Great Britain. Besides actual employment, it is interesting to see whether men’s and women’s labour supply adapts to their new situation after separation. Comparing employment status before and after separation, we notice a trend towards more employment in Italy and Sweden for women, while in Belgium and Germany more women work full-time after separation. Since most men work full-time before separation, further increases in employment are hardly feasible. On the contrary, a reduction of employment for men is much more probable. However, this event is not observed more often for men than for women. Childcare is largely the wife’s responsibility. This is also the case when a partnership breaks down. Table 11.3 shows whether parents among our sample members live together with children after separation (including own and children from a new partner). As can be seen from these data, most mothers in all five countries live together with children after separation. Sweden is the only country, where a significant proportion (44 per cent) of fathers cares for children after separation. The lowest proportion of fathers (14 per cent) living with children after separation is found in Italy. Income Indicators The central focus of our analysis is the economic situation of the household before and after separation. The national panel studies provide various income indicators for different time periods (year, month) both on the individual and the household level. For our analysis we use data on monthly disposable household income. We define disposable household income as the sum of income from labour, assets, public and private transfers (including child and spousal support) minus taxes and payable maintenance expenses. We have calculated monthly disposable household income for each country as analogously as possible (cf. Table 11A.1 in the appendix to this chapter for details). Nevertheless, our specific operationalization may differ from panel to panel due to data restrictions. One restriction concerns the time period: if income was available only on a yearly basis, we divided the yearly amount by 12 to receive an average monthly income. Another restriction concerns tax payments: in some panels questions on income taxes are asked, in others we only have a tax

302

Comparative studies

simulation to estimate net income.9 Finally, the amount of maintenance paid by liable persons and the amount received by entitled persons do not agree in some cases. But, given different question wordings and possible biased reports about support payments (Schaeffer et al., 1991), no attempt was made to correct the data. All incomes were measured in 1995 prices by using the official consumer price indices as provided by the International Monetary Fund (2002). Finally, incomes were converted from national currencies to euros. Analysing the economic position of individuals within the household context, that is, using household instead of personal incomes, is a difficult matter depending on various methodological assumptions concerning income sharing, economies of scale and different economic needs with respect to age and other socio-demographic variables. We adopt the common presumption that all personal incomes are shared and individual needs of each household member are adequately satisfied by this income pool although this has been questioned, especially by feminist scholars (for example, Millar and Glendinning, 1989). Various equivalence scales are used in the literature to control for economies of scale and varying income needs (for an overview see Buhmann et al., 1988), but a common standard has not yet been found. We have used a simple parametric function. All household incomes have been adjusted with respect to household size according to the formula: adjusted household income 5 household income / size0.5 These adjusted household incomes have then been assigned to each individual in the household and analyses were done on the individual level. The value 0.5 is also called an elasticity parameter, yielding moderate economies of scale with increasing household size similar to equivalence scales applied by the Organisation for Economic Co-operation and Development (OECD) (Buhmann et al., 1988).10 Methods To give a comprehensive overview on the economic consequences of separation we analyse the economic changes by simple comparisons of median household incomes at different points in time as well as by multivariate models of household income as a function of country- and gender-specific characteristics. The longitudinal design of the data-set allows us to model the process of separation on the individual level and to identify changes in the income variable as an effect of the event of separation. The exploratory part of the chapter looks at the absolute value of

The economic consequences of partnership dissolution

303

adjusted household income and how its median changes over time before and after separation for men and women in each country. For a better comparison of the actual impact of separation we also look at relative income changes by calculating the percentage change of income in the first survey wave after separation relative to the income at the last survey wave before separation. For Belgium, Germany and Great Britain the time-lag before separation is one year, for Italy and Sweden it is two or three years. We present the median and the quartiles of the distribution of individual changes. In the confirmatory part of the analysis we test our hypotheses about gender- and country-specific effects on the economic consequences of separation. By using a linear regression model we regard income as a function of time and various covariates. This regression model uses data from Sweden, Belgium, Germany and Great Britain. Italy is excluded because of the few observations per individual in the Italian household panel. Overall, the dataset includes information for i 5 1, . . ., N separated individuals coming from c 5 1, . . ., 4 countries. Depending on each individual’s participation in the respective national panel (starting at time point si and ending at time point ei), each individual contributes ti 5 si, . . ., 0, . . ., ei observations to our data. t is measured in years and ti 5 0 represents the year of separation. This data structure is called an unbalanced panel.11 Our dependent variable is adjusted household income and according to our hypotheses the model includes the following variables: incomeit 5 bc0 1 b1agei 1 b2educi 1 bc4childit 1 bc5emplit 1 b8educi 3 sepit 1 b6agei 3 sepit 3 b7emplit 3 sepit gc 1 g (t, ac0, agc 1 , a2 ) 1 ui

(11.1)

Using appropriate dummy variables and their interactions with the variables child, empl, and time t, several effects are estimated separately for each country c and gender g. To keep our explanation as simple as possible these dummies (and their interactions) are not shown in equation (11.1). Instead, country- and gender-specific effects are indicated by superscripts c and g.12 Basically, the model consists of five parts (cf. the five lines of equation 11.1): 1.

The country-specific constant as well as the effect of age (age) and education (educ) control for different national levels of income and for different socio-demographic characteristics of the national samples of separated individuals. Table 11.3 shows that Swedish, Belgian,

304

2.

3.

4.

Comparative studies

German and British individuals differ significantly with respect to age and education. Besides that, time of separation (ti 5 0) is observed early during the life course for some individuals and late for others. Since household income usually shows an increasing trend during the life course, we should control for the time (that is, age), in which our panel observation window [ si, . . ., ei ] captures this trend. Age is measured as age at separation and for ease of interpretation is centred at the overall mean (35.7 years). Education is included as a dummy variable for higher education (higher education versus secondary or primary education). As discussed in the theoretical parts of this chapter, childcare and gainful employment are two important determinants of each individual’s economic autonomy. National labour markets and public infrastructure for childcare more or less support individuals in earning their living and coping with their duties as parents. Therefore, our model includes information whether the individual is gainfully employed (empl) and lives together with at least one dependent child (child) at time t. However, the effects of both of these variables depend on the national context and, therefore, are estimated separately for each country. In the event of separation it can be expected that individuals, depending on their personal resources, more or less successfully cope with the new situation. We expect that higher educated and younger people have more resources because, for example, it is easier to find a new partner or to (re)enter the labour market at young ages, because better educated people may be more attractive, have better employment prospects, and have greater control over their social environment. Therefore, the model includes interactions between age and education on the one side and time after separation on the other. The corresponding dummy variable sep equals one for all observations with ti $ 0. Preliminary analyses showed that it is also necessary to include an interaction with employment, possibly because changes in family status also change net income gained from employment. Different tax rates depending on marital status that are not captured by our income measurements may be one explanation. In principle, all these interaction effects could be made country-specific, but extensive tests with more refined models showed that this is not necessary. The time trend of the dependent variable income is modelled by a function g (t) , which includes three parameters: a linear and country-specific income trend before separation (ac0) , a gender- and country-specific income change in the year of separation (agc 1 ) (the separation effect for short), and a linear, gender- and country-specific income trend after

The economic consequences of partnership dissolution

5.

305

separation (agc 2 ) . This saw tooth function of time is motivated by our descriptive analyses of income trends (cf. section below).13 Our genderand country-specific hypotheses from the previous section concern gc gc the parameters agc 1 and a2 . a1 measures short-term consequences of separation in absolute amounts for men and women in each country. Correspondingly, agc 2 measures long-term consequences of separation. One long-term statistic that can be derived from these parameters is of special interest: the number of years the average man or woman needs to reach the same level of (adjusted) household income s/he gained in the year before separation. It is estimated as the ratio of the separation gc effect and the income trend after separation: agc 1 /a2 . Finally, equation (11.1) includes a term ui measuring all (unobserved) characteristics of the individual not included in the model. As discussed in the corresponding statistical literature (for example, Greene, 2003), such unobserved characteristics can be controlled quite easily within a panel framework.

Basically, there are two approaches to control for this unobserved heterogeneity. The fixed effects method (FE) models individual effects ui by using a dummy variable for each individual. As a consequence, the effects of time-constant individual characteristics (for example, gender, nationality) are not identified due to their perfect collinearity with these dummies. Nevertheless, FE control for these time-constant individual characteristics. The random effects method (RE) treats individual effects ui as outcomes of a random variable, which is uncorrelated with each explanatory variable in the model. This method allows time-constant characteristics in the model. Furthermore, as the probability distribution of sample coefficients has a smaller variance, RE estimates usually can be tested on higher significance levels than FE estimates. Although RE estimation has certain desirable properties compared with FE, the choice between both approaches depends on the correlation between individual effects ui and covariates: ‘If the ui are correlated with some explanatory variables, the FE method is needed; if RE is used, then the estimators are generally inconsistent’ (Wooldridge, 2003: 473). In our tables we present both FE and RE estimates. Several indicators convinced us that in our case the differences between both approaches are small, at least from a substantive point of view, and, therefore, it is permissible to focus our interpretation on RE estimates.14 Our confirmatory analysis proceeds in a stepwise manner from simple models controlling only for differences in socio-demographic composition of each national sample up to the full model of equation (11.1). For each model we report the proportion of explained ‘within’ variance. This value indicates, how much variation of the individual time series of adjusted

306

Comparative studies

household incomes can be explained by the explanatory variables in the model. As our hypotheses focus on the impact of separation on the individual income trajectory, explained ‘within’ variance corresponds much better to our research question than explained ‘between’ or ‘overall’ variance.

RESULTS In the following we present the results of our empirical analyses. The first subsection starts with some exploratory analyses using simple statistics (medians, quartiles) and graphical displays to get an impression of the effects of separation on adjusted household incomes for men and women in each country. Our regression model is based on these figures and models the time path of household incomes as a linear trend interrupted by the event of separation. By including additional independent variables we are able to control for the multitude of individual living circumstances and in doing so to make ‘fair’15 comparisons between both genders and different national contexts. Results of this multivariate analysis, including tests of our hypotheses, are presented in the subsequent subsection. The Economic Situation Before and After Separation Figure 11.1 shows median adjusted household incomes separately for men and women in each country from five years before to five years after separation. Comparing the two graphs it is obvious that women in all five countries experience a significant income drop at the time of separation, while the economic situation for men does not change so drastically. Men’s adjusted household income decreases only minimally in Germany and Sweden and a little more so in Italy. Men in Great Britain and Belgium seem to be prevented from economic losses: their income remains at the same level or even increases. Furthermore, in the years following separation adjusted household income of men continuously grows, rising even above the level kept up during partnership. Unlike men, women do not recover well from the experienced income drop. After five years only German and British women almost reach the same income level as during partnership. The improvement of the financial situation for Belgian and Swedish women, on the other side, proceeds very slowly. Italian women seem to experience a declining income after separation. While Figure 11.1 focuses on the time trend in absolute incomes, Figure 11.2 gives a more detailed overview of relative income changes comparing two time points shortly before (t 2 1) and after (t) separation. The graph shows selected percentiles (median, quartiles) from the distribution

307

500

1000

1500

2000

0

Men

Great Britain

Sweden

5 –5 Years from separation

Germany

Belgium

Italy

0

Women

Adjusted household income before and after separation for men and women (median)

Cross-national data set of separations.

–5

Figure 11.1

Source:

Household income (euros)

5

308

Sweden

Belgium

Great Britain

Italy

Germany

Italy

Great Britain

Belgium

–50

0 Income changes in %

50

25. percentile median 75. percentile

Percentage change of adjusted household income for separated men and women (t versus t 2 1)

Cross-national data set of separations.

Figure 11.2

Source:

Women

Men

Germany

Sweden

The economic consequences of partnership dissolution

309

of change scores for each gender and country. Overall, negative income changes prevail for most of the women in all countries. German and Italian women experience the highest income drops with median income declines of 32–33 per cent, closely followed by British women: half of them lose at least 28 per cent of their former adjusted household incomes. In Belgium half of all women lose around 24 per cent at a minimum, while half of the Swedish women lose ‘only’ 20 per cent, which is the smallest median change score in this comparison. As already seen in Figure 11.1, men are generally better off than their female counterparts. In three of five countries (Belgium, Italy and Great Britain) more than half of the men even experience income gains or no income changes (cf. Figure 11.2). Only in Germany and Sweden do we find more losers than winners: half of all German men lose at least 2 per cent of their former income and Swedish men are worst off, with half of them losing at least 7 per cent. Overall, one can observe high gender inequalities for all five countries: with respect to adjusted household incomes women are the losers, while men are the winners or at least are better off than their female counterparts. Besides these overall developments in household incomes, it is important to remember the great variability in these findings indicated by the interquartile distance in Figure 11.2. Comparing the quartiles between genders, it becomes obvious that some men also experience losses of income and some women have slight increases in income. The variance of these changes is the result of different living conditions following separation. Therefore, comparing all separated women to all separated men clearly exaggerates gender differences. On average, women experience more risk factors than men. They earn relatively smaller incomes than men, these smaller incomes have to support more people and, often, private or public transfers do not compensate for the disadvantages. Therefore we have to focus especially on the two main factors that particularly influence the financial situation of women after separation: children and employment. Determinants of Short-term and Long-term Economic Consequences To explore how separation changes the income situation and which are the relevant factors responsible for gender differences, we have to compare men and women of similar employment prospects and with similar parental obligations. This is done in the following through a series of regression models as described in the previous ‘Data and methodology’ section. Each of the following relates to one of our hypotheses described earlier in this chapter. The regression estimates can be found in the appendix (cf. Table 11A.2). From these estimates we have calculated absolute separation

310

Comparative studies

effects (measured in euros) on adjusted household income for certain groups (cf. Table 11.4). Examples of these calculations have been included in the endnotes (see below). Table 11.4 makes it easier to compare between different groups and countries. As mentioned before, Italy is excluded from these regression models because of the few observations per individual in the Italian household panel. The impact of separation on household income (H1a, H1b) Figure 11.1 has shown that the event of separation leads to income changes particularly for women. The line plots suggest that adjusted household incomes approximately follow two different linear time trends before and after separation. Our basic regression model of adjusted household incomes (model 1) tests this observation more rigorously by including a specific (sawtooth) function of time for each country and gender (cf. previous section). Besides that, it controls for basic socio-demographic differences between the four samples (that is, age, education). Model 1 confirms our exploratory analysis.16 In all countries women experience a significant negative separation effect. They lose between €217 in Sweden and €431 in Germany, while men have to deal only with minor changes (cf. Table 11.4, panel B.1 versus A).17 German (€91) and Swedish men (€104) undergo moderate losses (although the Swedish value is not statistically significant). British men experience no significant income changes (€6) and Belgian men even slightly improve their financial situation by €103 (cf. Table 11.4, panel A). Hence, H1a does not apply for all countries. In Germany, Sweden and Belgium separation causes income changes for both partners. But H1a does not apply for Great Britain, where separation is connected with changes in income only for women but not for men. H1b can be confirmed without restrictions: In all countries women undergo more often and more severe income losses in comparison to men.18 Gender inequalities in post-separation household income (H2) Our exploratory analyses have already shown differences in post-separation household incomes between men and women. In Figure 11.1 as well as in Figure 11.2 the greater income declines for women in all countries are noticeable. Significant gender differences also show up in the multivariate analyses. The gender gap is estimated at €113 in Sweden, €327 in Belgium, €341 in Germany, and €359 in Great Britain (cf. Table 11.4, panel C.1).19 As hypothesized in H2, the gender gap is low and insignificant in Sweden, and it is high and significant in the other countries. Looking at panel C.2 in Table 11.4, one could even say that gender inequalities are the same in Belgium, Germany, and Great Britain (differences between these countries are not significant), while Sweden is the only country where such inequalities do not exist.

311

2111 100* 293** 220 2186* 2223** 2433** 2353** 37 247** 167 210** 130** 80** 275 2323** 2340** 2333**

2217** 2224** 2431** 2365**

7 214* 148 207** 141** 66*

2113 2327** 2341** 2359**

FE

2104 103* 291** 26

RE

Model 1

278 2252** 2304** 2227**

18 222** 155 204** 137** 67*

2217** 2235** 2439** 2372**

2139 17 2135** 2145**

RE

231 2255** 2315** 2234**

47 253** 173 206** 126** 80*

2188* 2235** 2441** 2361**

2157 20 2126** 2127**

FE

Model 2

2128 2279** 2326** 2310**

26 204* 131 210** 137** 73*

2300** 2294** 2504** 2431**

2172* 215 2178** 2121**

275 2301** 2332** 2310**

30 235* 158 205** 128** 77*

2264** 2294** 2499** 2422**

2189* 7 2167** 2112**

FE

Model 3a RE

Estimated changes in adjusted household income (euros)

A. Men: separation effect Sweden Belgium Germany Great Britain B.1 Women: separation effect Sweden Belgium Germany Great Britain B.2 Women: country differencesb Sweden-Belgium Sweden-Germany Sweden-Great Britain Belgium-Germany Belgium-Great Britain Great Britain-Germany C.1 Gender difference Sweden Belgium Germany Great Britain

Table 11.4

294 2206** 2290** 2187**

5 211* 140 206 135 71

2301** 2306** 2512** 2441**

2207** 2100* 2222** 2254**

RE

232 2233** 2307** 2216**

40 241* 166 201** 126** 75*

2267** 2307** 2508** 2433**

2235** 274 2201** 2217**

FE

Model 4a

291 2278** 2355** 2276**

97 294** 128 197** 31 166**

2273** 2370** 2567** 2401**

2182* 292 2212** 2125**

RE

235 2286** 2403** 2318**

122 327** 154 205** 32 173**

2242** 2364** 2569** 2396**

2207* 278 2166** 278

FE

Model 5ad

312

(continued)

C.2 Gender difference compared between countriesb Sweden-Belgium Sweden-Germany Sweden-Great Britain Belgium-Germany Belgium-Great Britain Great Britain-Germany D.1 Women: income trend after separation Sweden Belgium Germany Great Britain D.2 Women: years to gain pre-separation income Swedenc Belgium Germany Great Britain E. Being employed Sweden Belgium Germany Great Britain

Table 11.4

210 25** 43** 54**

– 9 10 7

–

9 10 7

248* 265* 258* 17 10 7

FE

27 23** 45** 55**

214* 227* 246** 13 32 219

RE

Model 1

11 10 7

–

27 22* 43** 54**

174 226* 149 52 225 77

RE

10 11 7

–

210 23 42** 53**

224 284** 203 60 221 81

FE

Model 2

– 12 12 8 22 137** 87** 241**

–

70 175** 128** 312**

28 24* 42** 53**

226* 257* 235* 31 9 22

FE

15 12 8

27 20* 41** 52**

151 198* 182* 47 31 16

RE

Model 3a

56 169** 124** 281**

15 12 8

–

26 20* 42** 53**

112 196* 93 84 219 103**

RE

6 123** 81** 222**

15 12 8

–

29 21* 41** 51**

201 275* 184 74 217 91*

FE

Model 4a

67 150** 141** 303**

34 18 40

–

26 11 31** 10

187 267** 185* 77 22 79

RE

20 90* 100** 266**

23 17 25

–

29 16 33** 16*

251* 368** 283** 117 32 85

FE

Model 5ad

313

2214** 2233** 2150** 2391**

2249** 2224** 2115** 2309**

2205** 2221** 2146** 2359**

2241** 2221** 2112** 2293**

2208** 2215** 2113** 2326**

2243** 2235** 265** 2260**

Source:

Calculated from the estimates in Table 11A.2 in the appendix.

Notes: ** p , 0.01; * p , 0.05. a Separation effects in model 3, 4 and 5 refer to not employed and not highly educated individuals. These effects are printed in italics and cannot be compared straightforwardly with models 1 and 2. b Reference group in Panel B.2 and C.2 is always the first country. c No interpretable value, because the estimated income trend after separation for Sweden is negative. d Model 5 excludes individuals living with a new partner after separation in Belgium, Germany and Great Britain.

F. Caring for children Sweden Belgium Germany Great Britain

314

Comparative studies

We argued, however, that such an overall comparison is unfair and controls for individual living circumstances should be included. This is done in more refined regression models, in which parental obligations (model 2), gainful employment (model 3) and both factors together (model 4) are included as additional regressors. Not surprisingly, employed individuals on average have higher adjusted household incomes (cf. Table 11.4, panel E), while caring for children is associated with lower household incomes (panel F). Employment and childcare have significant effects in all countries across different model specifications, except the effect of employment in Sweden, which is possibly negligible because labour force participation in general is high in Sweden. Great Britain stands out in the other direction: children seem to be exceptionally expensive in Great Britain (adjusted household incomes are €359 lower when caring for children, cf. panel F, model 4).20 In other words, the costs of children are to a minor degree compensated by the state than in the other three countries. This is a result of rudimentary British family policy: Great Britain awards social benefits only in case of need. On the other side, only in Great Britain employment is linked with a noticeable improvement of the financial situation (€281, cf. panel E, model 4).21 This result confirms our assertion that gainful employment forms the basis of the market model in Great Britain. Employment plays a major role for the improvement of the financial situation. When controlling for parental obligations and employment, the aforementioned gender gap gets smaller, but remains significant. In model 4 it is estimated at (insignificant) €94 in Sweden, €206 in Belgium, €290 in Germany, and €187 in Great Britain (cf. Table 11.4, panel C.1). Obviously, gender inequalities in post-separation household income persist when controlling for parenthood, human capital and other socio-demographic characteristics. One explanation might be that working women in Belgium, Great Britain and Germany earn lower wages in comparison to men, partly because the majority of working women works in lower prestige, female-dominated occupations that are paid less than comparable, maledominated occupations (Treiman and Hartmann, 1981). As discussed in the second section, women’s earnings in Germany, Great Britain and Belgium are considerably lower than men’s earnings, which is not true for Sweden. Given these data it is plausible that women in Belgium, Great Britain and Germany – despite being employed – are more strongly affected by financial losses in case of separation than men. In other words, the risk of losing a higher male income is stronger in Belgium, Great Britain and Germany than in Sweden. Indirectly our stepwise procedure also shows which factor (parental obligations, employment) is more important in explaining gender inequalities. For example, the gender difference in Great Britain decreases much

The economic consequences of partnership dissolution

315

more when controlling for children than in Germany or Belgium (cf. Table 11.4, panel C.1, model 1 versus 2). It follows that in Great Britain inequalities between the genders are associated to a larger extent with the costs of children, which are at the same time to a high degree unequally distributed between men and women after separation. As already mentioned, the market model supports families to a minor degree and, therefore, the person who lives with children after separation has to deal with more economic disadvantages. Furthermore, it can be noticed for Germany that both – parenthood and employment status – explain the gender differences in post-separation income to a lesser degree than in the other countries. Our theoretical assumptions regarding gender inequalities in the analysed countries (H2) have been proven true. The economic autonomy of Swedish women is highest, and no significant gender differences were found. In contrast, gender inequality is high in the other three countries, particularly Germany stands out negatively with an (adjusted) gender gap of €290.22 Short-term consequences of separation for women (H3) We have already discussed the separation effect for women in this chapter. Now we want to turn to the implicit country ranking in these effects and discuss whether these short-term consequences of separation confirm our theoretical expectations (H3). According to the data in Table 11.4 (panel B.1) the separation effect is highest in Germany (€431) followed by Great Britain (€365), Belgium (€224), and finally Sweden (€217). This ranking remains the same even in the final model 4 and comes up to our expectations,23 in which we argued that high labour market participation of Swedish women and generous cash transfers of the Belgian welfare state for families and individuals in need protect separated women from drastic financial losses. It should be noted, however, that depending on the statistical specification some country differences are not significant (cf. Table 11.4, panel B.2). Besides that, pre-separation income level in Germany and Great Britain is considerably higher than in Sweden and Belgium (cf. Figure 11.1). Therefore, higher economic losses in Germany and Great Britain can be partly ascribed to the fact that women in Sweden and Belgium with lower incomes prior to separation will experience a more moderate decline, simply because they cannot lose so much. Long-term consequences of separation for women (H4) Up to now we have shown that women in all countries on average experience a substantial decline of income after separation. This loss is plainest immediately after the event of separation (cf. Figure 11.1). If this is only a temporary effect, income should increase gradually in the years following

316

Comparative studies

separation as legal measures such as maintenance payments take effect and the involved persons more or less successfully learn how to cope with the new situation. Depending on welfare state arrangements and personal resources this temporary financial crisis should be overcome sooner or later. For appraising the long-term consequences we take a look at the income trend after separation (cf. Table 11.4, panel D.1, model 4) and the time required to regain pre-separation income (panel D.2). Sweden is the only country without a positive income trend after separation, which means that Swedish women are not able to regain their prior household income before separation.24 Apparently the event of separation in Sweden has long-lasting financial consequences that cannot easily be compensated for in the future – even though the absolute separation loss is rather low. The strongest increase of household income in the subsequent years – with an average annual increment of €53 – can be observed for British women.25 They are the first to regain their pre-separation income (on average eight years after separation). In Germany the income trend after separation is also strongly positive. German women have an estimated annual increment of €42, but because of the large income decline at the event of separation they regain their pre-separation income not until 12 years later. The income trend for Belgian women is also significantly positive but, with an annual increment of €20, rather moderate. Thus, they regain their pre-separation income even later, that is, after 15 years.26 In sum, separation is connected in all countries with long-lasting economic consequences for women, and the period to recover from this financial crisis varies by national context. We had no specific hypothesis concerning these long-term consequences, but simple explanations apparently do not work. For example, the expectation that Belgian and Swedish women are able to compensate income losses rather quickly while it should be harder for German and British women to reach their pre-separation income level turns out to be wrong. Surprisingly, Swedish women are not able to compensate for the comparatively low income losses in the following years. Perhaps because of high living costs in Sweden and high individual taxation one income is not sufficient to absorb the losses. Also contrary to this expectation, British women compensate their income losses rather quickly in comparison to women in the other countries. Both of these results should be investigated more thoroughly in the future. Different trends of re-partnering after separation in each country are one possible explanation. Unfortunately, we were not able to control for this effect in the Swedish data. But it is possible to identify individuals living together with a new partner in the Belgian, British and German data. Therefore, we re-estimated our final model 4, but excluded all Belgian,

The economic consequences of partnership dissolution

317

British and German individuals living together with a new partner after separation. The results are shown as model 5 in Table 11.4. The main conclusions remain the same, but now Belgian (34 years), British (40 years) and German women (18 years) need much more time to reach their preseparation income level. The income trend after separation is much flatter in all countries, and especially so for men, as can be seen from the regression estimates in the appendix (Table 11A.2). Therefore, if re-partnering is much lower in Sweden than in the other countries, this could explain why Swedish women are not able to compensate for the comparatively low income losses in the following years.27 Besides that, the optimistic picture drawn for British women is partly due to higher rates of re-partnering in Great Britain.

DISCUSSION The event of separation is linked with financial changes for most of the involved persons. But our analyses have shown that especially women and, foremost, mothers are more affected by economic losses, while the financial situation of men either does not or does just slightly change. Thus, the costs of separation are not equally distributed between men and women. Owing to the predominance of the gender-specific division of labour in partnerships in Belgium, Germany and Great Britain, economically dependent women experience the higher and most enduring part of the financial loss caused by separation. An outstanding exception is Sweden, where the economic consequences of separation are equally distributed between men and women, while clear gender inequalities are evident for the other countries. Particularly in Britain, women are more dependent on the labour market. They suffer the biggest loss when having children and when not being employed. Caring for children is accompanied by costs, which seem not to be compensated for by the state. Furthermore, children impede women’s ability to work after separation, and external childcare is again linked with costs. As mentioned above, the British welfare state practices only a rudimentary family policy. Instead, own gainful employment forms the main basis of social protection and its relevance becomes obvious in the results of our analyses. Also in the long run, employment is the best instrument to absorb losses, because the state hardly compensates for the lack of income due to unemployment or for additional costs of children. Another element of protection seems to be the partner market, because British men and women living together with new partners compensate the financial losses rather quickly.

318

Comparative studies

Sweden, focusing on the dual-earner household, explicitly supports the employment of women. Therefore, a good childcare infrastructure is offered which enables Swedish women to work full-time. On the other hand, the living costs seem to be very high in Sweden, in particular when economically dependent children live in the household. Because of high individual taxation one income is not sufficient to absorb the losses – in the short run as well as in the long run. Hence, it is not surprising that the dualearner model dominates. As shown before, gender inequalities in Sweden do not exist. Therefore, men and women have to suffer equally long-lasting losses from separation. Germany, like Sweden, has an extended family policy, but in contrast to Sweden the male breadwinner model dominates. Insufficient external childcare, joint taxation of both spouses and financial encouragement of parental childcare are reasons for low female labour force participation. A gender-specific division of labour has exceptionally negative consequences after partnership dissolution. One measure to reward matrimony is joint taxation. This financial advantage ends with separation. Thus, some of the income losses of German women have to be ascribed to this taxation system. It is also an explanation for the fact that German men, in contrast to men in the other countries, experience financial losses after separation. But German women, and especially mothers, are even more negatively affected. Furthermore, they are not able to improve their financial situation through employment to the same degree as British or Belgian women do, and in the long run they need a comparatively long time to recover from the losses. In Belgium, a gender-specific division of labour is also prevalent, but the results show that Belgian women come off better than German women. Apparently the combination of generous social transfers and a very good childcare infrastructure is able to absorb the losses in both cases: if women are employed by offering a good childcare system (even for toddlers) or if women are not employed by paying generous social benefits. However, this effect is visible only in the short run. The immediate separation effect is comparatively low, but in the long run Belgian women have a much harder time in regaining their pre-separation income level, especially if one controls for re-partnering. Even though Italy has been excluded from the multivariate analyses, the results of our exploratory analyses give an idea how to evaluate the economic consequences of separation in Italy (cf. Figure 11.1 and 11.2). It becomes apparent that Italian women come off badly. Together with German women they experience the highest income drops. In contrast, Italian men seem to do rather well after separation – not only in comparison to women, but also to men in the other countries. The gender gap seems

The economic consequences of partnership dissolution

319

to be highest in Italy. An explanation for the exceptionally severe situation of Italian women and the comparatively comfortable situation of Italian men after separation is primarily the dominance of the traditional genderspecific division of labour and its fatal consequences in case of separation. Italy ranks lowest among the five countries with respect to female employment. Furthermore, the system of post-marriage alimony is rudimentary compared with the other countries’ systems. The combination of both produces negative financial consequences for women who – in most cases – do not participate in the labour market and cannot count on extensive subsistence payments from the former partner. We started our analysis with the assumption that the higher the economic autonomy of women in a country, the more equally income changes are distributed between men and women after separation. This assumption can be confirmed. The example of Sweden shows that in a model of family support, in which both women and men have nearly equal chances to participate in the labour market, gender inequalities after separation will disappear. On the other hand, the results also suggest that in countries in which a gendered division of labour is more prevalent, gender inequalities in post-separation income will only marginally adjust, even if women comply with the requirements of economic autonomy, that is, are gainfully employed. But even though the Swedish model produces gender equalities concerning post-separation incomes, the model is not without a blemish: in the long run, both Swedish men and women have to deal with longlasting financial consequences after separation, which do not appear to the same extent in any of the other countries. This disappointing observation certainly needs further research.

NOTES 1.

2. 3.

Most research on the economic consequences of partnership dissolution has been done in the USA. For an overview see McManus and DiPrete (2001) and review articles by McKeever and Wolfinger (2001), Andreß and Lohmann (2000) and Andreß et al. (2003). We are interested in the variance of such studies between different countries. Therefore, Table 11.1 presents only such US-American studies that include at least one other country besides the USA. Uunk (2004) uses the European Community Household Panel and analyses a total of 14 countries. Only those countries, for which at least one other study is available, are presented in Table 11.1. Orloff (1993) defines autonomy as the freedom not to continue a potentially repressive relationship, for example, because of economic dependency. Unfortunately, little comparative information is available on the practice of family law and on social support given by the extended family network. Legal regulations are measures the state can take in order to grant welfare to divorcees and their children. Of course, different legal regulations exist in all countries (see the corresponding single-country chapters in Part II of this volume), but we do not have any comparative information on how they are actually applied. We do not know, for example, whether

320

4.

5.

6.

7. 8.

9. 10. 11. 12.

13. 14.

Comparative studies dependent family members claim for maintenance payments. We also do not know whether liable persons charge their maintenance payments regularly and completely. Therefore, we have to ignore this aspect of governmental family policy, when typifying the countries. Similarly, there are no comparative studies concerning financial support by family members in case of separation and divorce. Paugam and Russell (2000) focus on spells of unemployment. They show that primary (nuclear family) and secondary types of sociability (relatives, neighbours, friends) are considerably stronger in Italy, Spain, Portugal, and Greece than in other European countries. Consequently, in the Southern European countries the risk of social isolation in case of unemployment is much lower and financial support from outside the nuclear family is much higher. Whether the same is true in case of partnership dissolution, remains an open question. In Italy, family members (even those who are not living together) are the main providers of alimony for an individual in need, and not the state (Timoteo, 1995: 265). The Italian care system, for example, is widely based on mutual responsibility within family ties: professional, state financed at home care for the elderly or invalid persons is provided only if family resources are insufficient. Accordingly Saraceno (2000: 347) asserts that ‘[social] services are defined as surrogates for an inadequate family, rather than as services granting individual social rights irrespective of family membership, they are likely to be perceived as creating dependency’. If individuals experience several separations, only the first one is included in the sample. Preliminary analyses were conducted separately for consensual unions and marriages (in each country except Sweden) and are available from the authors on request. Results were slightly different between both groups. But from a substantial point of view the differences were of no practical importance, especially when controlling for age, education and parental obligations (characteristics on which both groups are most different). Therefore, the following statistical model – for the sake of parsimony – does not differentiate between both groups. Given this selection rule and different attrition rates for men and women before and after separation, it is possible that only one and not both partners are included in the analysis. As the number of separations is low in some countries, we refrained from including only complete couple data to maximize sample sizes for men and women. As a consequence, however, average household incomes before separation are not identical for men and women, because some of their partners are missing. Table 11A.2 in the appendix at the end of this chapter shows the number of individuals and panel observations available for our multivariate analyses. According to United Nations Educational, Scientific, and Cultural Organization (Unesco’s) International Standard Classification of Education (ISCED, 1997) we distinguished between primary, secondary, and higher education. A dummy for higher education is used in all our analyses. The tax simulation for the Swedish panel was kindly provided by Roger Wahlberg (University of Gothenburg, Sweden), for the German and British tax simulation cf. Schwarze (1995) and Bardasi et al. (1999). For example, an elasticity parameter of 1 yields per capita incomes, a parameter of 0 total household income. One might add ‘a pooled unbalanced panel’ because unbalanced panel data are pooled from different countries. More specifically, dummy variables were used for Belgium, Germany and Great Britain such that a main effect is estimated for Sweden and each country-specific effect for the other countries is expressed as a deviation from the corresponding Swedish effect. The same applies for gender, where men constitute the reference category (cf. Table 11A.2 in this chapter’s appendix). gc More specifically, the function g(t) is defined as g (t) 5 ac0 3 t 1 agc 1 3 sepit 1 g2 3 c gc t 3 sepit, which yields the parameter agc 2 5 a0 1 g 2 . Table 11A.2 in the chapter’s appendix shows estimates of the correlation between individual effects and covariates in each model. All of these estimates are rather small.

The economic consequences of partnership dissolution

15.

16.

17. 18. 19.

20. 21. 22. 23. 24. 25.

26. 27.

321

Admittedly, these correlation statistics are not a powerful test between both approaches. A Hausman specification test should be used instead. In fact, for all estimated models the Hausman test rejects the hypothesis of no differences between FE and RE estimates. On the other hand, if the number of covariates is large as in our case, this test reacts overly liberal in rejecting the null hypothesis of no differences. Therefore, its result should be interpreted with care. If the differences in the coefficients are not theoretically meaningful and consequently the bias can be regarded as small, RE estimates can be used. We also tested for serial correlation of the individual effects ui, which turned out not to be large (,.3). Furthermore, we don’t expect their variances to differ significantly from individual to individual. In sum, the usual assumptions for RE and FE models seem to be true and unbiased and efficient estimators are to be expected. These comparisons are ‘fair’ in a statistical sense: individuals in similar living situations are compared. It should be stressed, however, that fathers’ low participation in childcare after separation or wage discrimination against women in the labour market is not fair from the viewpoint of gender equality. The following interpretation focuses on RE estimates. FE estimates are reported for the interested reader, but do not yield any different results of practical importance. See also the ‘Data and methodology’ section for a discussion of the methodological differences between both approaches. For example, the separation effect for Swedish women (€217) is calculated as the sum of the RE estimates –104 (separation effect men) and −113 (additional effect for women) in Table 11A.2 (model 1). For the reasons mentioned, Italy could not be included in the multivariate analyses, but the exploratory analyses in ‘The economic situation before and after separation’ subsection for Italy do not contradict H1a and H1b. The Swedish gender gap equals the RE estimate −113 (additional effect for Swedish women in Table 11A.2, model 1). The Belgian gender gap is calculated as the sum of the RE estimates −113 (additional effect for Swedish women) and −214 (additional effect for Belgian women). German and British gender gaps are calculated likewise. The ‘costs’ of caring for children (€359) are calculated as the sum of the RE estimates −205 (child in Sweden) and −154 (child in Great Britain) in Table 11A.2 (model 4). The ‘gain’ because of employment (€281) is calculated as the sum of the RE estimates 56 (employed in Sweden) and 225 (employed in Great Britain) in Table 11A.2 (model 4). Again, the explorative results for Italy in ‘The economic situation before and after separation’ subsection do not contradict H2. It seems as if gender inequality is even higher than in Germany. From the explorative results in ‘The economic situation before and after separation’ subsection Italy nicely fits into the theoretically expected ranking. Italian women probably experience the highest income drops (together with German women). The same is true for Swedish men (cf. the regression estimates in the appendix, Table 11A.2). The income trend after separation for Swedish women (€−6) is calculated as the sum of the RE estimates −8 (trend after separation for Swedish men) and 12 (additional trend for Swedish women) in Table 11A.2 (model 4). The income trend after separation for British women (€53) is a more complicated sum adding all necessary interaction parameters: −8 plus 171 (trend after separation for British men) plus 12 plus −12 (additional trend for British women). For Italy we cannot draw conclusions on the long-term consequences of separation from our exploratory analyses, because the number of observations in the years following separation is too small. We are sceptical whether this speculation turns out to be true, because data from the Fertility and Family Surveys show that Sweden is among the countries with the highest rates of repartnering (Prskawetz et al., 2002).

322

Comparative studies

REFERENCES Andreß, H.-J. (2003), ‘Die ökonomischen Risiken von Trennung und Scheidung im Ländervergleich: ein Forschungsprogramm’, Zeitschrift für Sozialreform, 49 (4), 620–51. Andreß, H.-J. and Güllner, M. (2001), ‘Scheidung als Armutsrisiko’, in E. Barlösius and W. Ludwig-Mayerhofer (eds), Die Armut der Gesellschaft, Opladen: Leske und Budrich, pp. 169–97. Andreß, H.-J. and Lohmann, H. (2000), Die wirtschaftlichen Folgen von Trennung und Scheidung, Band 180 der Schriftenreihe des Bundesministeriums für Familie, Senioren, Frauen und Jugend. Stuttgart: Kohlhammer. Andreß, H.-J., B. Borgloh, M. Güllner and K. Wilking (2003), Wenn aus Liebe rote Zahlen werden. Die wirtschaftlichen Folgen von Trennung und Scheidung, Wiesbaden: Westdeutscher Verlag. Andries, M. (1996), ‘The politics of targeting: the Belgian case’, Journal of European Social Policy, 6 (3), 209–23. Andries, M. (1997), ‘De Belgische sociale zekerheid in de typologie van de welvaartsstaaten’, www.ufsia.ac.be/CSB/csbberichten/berweb/97610401.doc, accessed 10 January 2005. Banca d’Italia (2000), Supplements to the Statistical Bulletin – Methodological Notes and Statistical Information, Rome: Banca d’Italia. Bardasi, E., S.P. Jenkins and J.A. Rigg (1999), ‘Documentation for derived current and annual net household income variables, BHPS waves 1–7’, ISER working papers 1999–25. Büchel, F. and C.K. Spieß (2002), Form der Kinderbetreuung und Arbeitsmarktverhalten von Müttern in West- und Ostdeutschland, Band 220 der Schriftenreihe des Bundesministeriums für Familie, Senioren, Frauen und Jugend, Stuttgart: Kohlhammer. Buhmann, B., L. Rainwater, G. Schmaus and T.M. Smeeding (1988), ‘Equivalence scales, well-being, inequality, and poverty: sensitivity estimates across ten countries using the Luxembourg Income Study (LIS) database’, Review of Income and Wealth, 34, 115–42. Burkhauser, R.V., B.A. Butrica, M.C. Daly and D.R. Lillard (2001), ‘The crossnational equivalent file: a product of cross-national research’, in I. Becker, N. Ott and G. Rolf (eds), Soziale Sicherung in einer dynamischen Gesellschaft. Festschrift für Richard Hauser zum 65. Geburtstag, Frankfurt and New York: Campus, pp. 354–76. Burkhauser, R.V., G.J. Duncan, R. Hauser and R. Berntsen (1990), ‘Economic burdens of marital disruption: a comparison of the United States and the Federal Republic of Germany’, Review on Income and Wealth, 36, 319–33. Burkhauser, R.V., G.J. Duncan, R. Hauser and R. Berntsen (1991), ‘Wife or frau, women do worse: a comparison of men and women in the United States and Germany after marital dissolution’, Demography, 28, 353–60. Casman, M.T. and V. Lenoir (2004), ‘Voorstelling van de PSBH’, in D. Mortelmans, M.T. Casman and R. Doutrelepont (eds), Elf jaar uit het leven in België. Socioeconomische analyses op het Gezinsdemografisch Panel PSBH, Gent: Academia Press, pp. 21–43. Daly, M. (2000), The Gender Division of Welfare. The Impact of the British and German Welfare States, Cambridge: Cambridge University Press. Dewilde, C. (2002), ‘The financial consequences of relationship dissolution for

The economic consequences of partnership dissolution

323

women in Western Europe’, in E. Ruspini and A. Dale (eds), The Gender Dimension of Social Change. The Contribution of Dynamic Research to the Study of Women’s Life Courses, Bristol: Policy Press, pp. 81–110. Dienel, C. (1993), ‘Italien’, in E. Neubauer (ed.), Zwölf Wege der Familienpolitik in der Europäischen Gemeinschaft, Band 2, Stuttgart: Kohlhammer, pp. 243–68. DiPrete, T.A. and P.A. McManus (2000), ‘Family change, employment transitions, and the welfare state: household income dynamics in the United States and Germany’, American Sociological Review, 65, 343–70. Eurostat (2003), ‘Datenbank NewCronos’, www.eudatashop.de/datenba/DE/ thema3 /bereiche.htm, accessed 16 September 2003. Finnie, R. (1993), ‘Women, men and the economic consequences of divorce: evidence from Canadian longitudinal data’, Canadian Review of Sociology and Anthropology, 30, 205–41. Fix, B. (2000), ‘Church–state relations and the development of child care in Austria, Belgium, Germany and the Netherlands’, in A. Pfenning and T. Bahle (eds), Families and Family Policies in Europe: Comparative Perspectives, Frankfurt a. M.: Lang, pp. 305–21. Flood, L. and P. Olovsson (1998), Household Market and Nonmarket Activities (HUS): Survey Description, Vol. 8, Göteborg: Department of Economics, Göteborg University. Fritzell, J. (1990), ‘The dynamics of income distribution: economic mobility in Sweden in comparison with the United States’, Social Science Research, 19, 17–46. Gähler, M. (1998), ‘Life after divorce. Economic, social and psychological well-being among Swedish adults and children following family dissolution’, Stockholm: Swedish Institute for Social Research, Dissertation Series No. 32. Greene, W.H. (2003), Econometric Analysis, 5th edn, Upper Saddle River, NJ: Pearson Education. Holden, K.C. and P.J. Smock (1991), ‘The economic costs of marital dissolution: why do women bear a disproportionate cost?’, Annual Review of Sociology, 17, 51–78. International Monetary Fund (2002), International Financial Statistics Yearbook, Washington, D.C.: IMF. International Standard Classification of Education (ISCED) (1997), International Standard Classification of Education, Unesco, www.uis.unesco.org/TEMPLATE/ pdf/isced/ISCED_A.pdf, accessed 29 March 2005. Jarvis, S. and S.P. Jenkins (1999), ‘Marital splits and income changes: evidence for Britain’, Population Studies, 53, 237–54. McKeever, M. and N.H. Wolfinger (2001), ‘Reexamining the economic costs of marital disruption for women’, Social Science Quarterly, 82, 202–17. McManus, P. and T. DiPrete (2001), ‘Losers and winners: the financial consequences of separation and divorce for men’, American Sociological Review, 66, 246–68. Millar, J. and C. Glendinning (1989), ‘Gender and poverty’, Journal of Social Policy, 18, 363–81. Ministry of Health and Social Affairs (2002), ‘Swedish family policy’, Fact Sheet No. 14, www.sweden.gov.se/content/1/c6/02/42/15/43b9ad0a.pdf, accessed 20 January 2005. Neubauer, E. (1993), ‘Vereinigtes Königreich’, in Bundesministerium für Familie und Senioren (ed.), Zwölf Wege der Familienpolitik in der Europäischen

324

Comparative studies

Gemeinschaft. Eigenständige Systeme und vergleichbare Qualitäten? Band 2, Schriftenreihe des Bundesministeriums für Familie und Senioren, Kohlhammer: Stuttgart. Organisation for Economic Co-operation and Development (OECD) (2004), Benefit and Wages. OECD Indicators, Paris: OECD. Orloff, A.S. (1993), ‘Gender and the social rights of citizenship: the comparative analysis of gender relations and welfare states’, American Sociological Review, 58, 303–28. Paugam, S. and H. Russell (2000), ‘The effects of employment precarity and unemployment on social isolation’, in D. Gallie and S. Paugam (eds), Welfare Regimes and the Experience of Unemployment in Europe, Oxford: Oxford University Press, pp. 243–64. Poortman, A.R. (2001), ‘Sex differences in the economic consequences of separation: a panel study of The Netherlands’, European Sociological Review, 16, 367–83. Prskawetz, A., A. Vikat, D. Philipov and H. Engelhardt (2002), ‘Pathways to stepfamily formation in Europe: results from the ffS’, Working Paper 2002-046, Rostock: Max-Planck Institut für Demographische Forschung. Ringen, S. (1997), ‘Great Britain’, in S.B. Kamerman and A.J. Kahn (eds), Family Change and Family Policies in Great Britain, Canada, New Zealand, and the United States, Oxford: Clarendon Press, pp. 29–102. Ruspini, E. (1999), ‘Social rights of women with children: lone mothers and poverty in Italy, Germany and Great Britain’, South European Society and Politics, 4 (2), 89–121. Ruspini, E. (2000), ‘Lone mothers’ poverty in Europe: the cases of Belgium, Germany, Great Britain, Italy and Sweden’, in A. Pfenning and T. Bahle (eds), Families and Family Policies in Europe: Comparative Perspectives, Frankfurt a.M.: Lang, pp. 221–44. Sainsbury, D. (1999), Gender and Welfare State Regimes, Oxford: Oxford University Press. Saldeen, A. (1995), ‘Family law in Sweden’, in C. Hamilton, K. Standley and D. Hodson (eds), Family Law in Europe, London, Dublin and Edinburgh: Butterworth, pp. 472–510. Saraceno, C. (2000), ‘The ambivalent familism of the Italian welfare state’, in R. Goodin (ed.), The Foundation of the Welfare State, Vol. 2, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 344–66. Schaeffer, N.C., J.A. Seltzer and M. Klawitter (1991), ‘Estimating nonresponse and response bias. Resident and nonresident parents’ reports about child support’, Sociological Methods and Research, 20, 30–59. Schmid, J. (1996), Wohlfahrtsstaaten im Vergleich. Soziale Sicherung in Europa: Organisation, Finanzierung, Leistungen und Probleme, Opladen: Leske und Budrich. Schwarze, J. (1995), ‘Simulating German income and social security tax payments using the GSOEP’, Cross-National Studies in Aging Program Project Paper No. 19. Sims-Schouten, W. (2002), ‘Child care services and parents’ attitudes in England, Finland and Greece’, in A. Pfenning and T. Bahle (eds), Families and Family Policies in Europe: Comparative Perspectives, Frankfurt: Lang, pp. 270–86. SOEP Group (2001), ‘The German Socio-Economic Panel (GSOEP) after more than 15 years – overview’, in E. Holst, D.R. Lillard and T.A. DiPrete (eds), Proceedings

The economic consequences of partnership dissolution

325

of the 2000 Fourth International Conference of German Socio-Economic Panel Study Users (GSOEP2000). Vierteljahreshefte zur Wirtschaftsforschung, 70 (1), 7–14. Sørensen, A. (1994), ‘Women’s risk and the economic position of single mothers’, European Sociological Review, 10, 173–88. Taylor, M.F. (ed.) with J. Brice, N. Buck and E. Prentice-Lane (2004), British Household Panel Survey User Manual Volume A: Introduction, Technical Report and Appendices, Colchester: University of Essex. The Clearinghouse on International Developments in Child, Youth and Family Policies (2003), Comparative Policies: Country: Italy, New York: Columbia University, www.childpolicyintl.org, accessed 22 May 2003. Timoteo, M. (1995), ‘Family law in Italy’, in C. Hamilton, K. Standly and D. Hodson (eds), Family Law in Europe, London: Butterworth, pp. 264–98. Treiman, D.J. and H.I. Hartmann (1981), Women, Work and Wages. Equal Pay for Jobs of Equal Value, Washington, D.C.: National Academy Press. Uunk, W. (2004), ‘The economic consequences of divorce for women in the European Union: the impact of welfare state arrangements’, European Journal of Population, 20, 251–85. Wooldridge, J.M. (2003), Introductory Econometrics, Cincinnati, OH: SouthWestern.

Comparative studies

326

APPENDIX Table 11A.1

Income indicators in the national household panels

Country

Income indicator

Belgium Germany

Monthly disposable net household income Household annual gross labour income 1 household annual asset income 1 household private transfer income 1 household private retirement income 1 household social security income 1 household annual public transfer income 2 net household annual taxes 2 maintenance expenses for ex-spouse or children outside of household / 12 5 monthly disposable household income Household annual gross labour income 1 household annual asset income 1 household private transfer income 1 household private retirement income 1 household social security income 1 household annual public transfer income 2 net household annual taxes 2 maintenance payments for ex-spouse or children outside of household / 12 5 monthly disposable household income Household annual net labour income 1 household annual asset income 1 household annual private transfer income 1 household annual property income 1 household annual retirement income 1 household annual public transfers and social security income 1 household annual income from letting a house/appartment 2 payable interest 2 depreciation /12 5 monthly disposable household income Household monthly net labour income 1 monthly household income from capital 1 monthly household public transfers 1 monthly household housing allowance 1 monthly household other non-labour income

Great Britain

Italy

Sweden

The economic consequences of partnership dissolution

Table 11A.1 Country

327

(continued)

Income indicator 1 support payments received 2 household taxes 2 support payments made 5 monthly disposable household income (For 1988 no income variables available; for 1991 only limited income variables and no information on support payments.)

Sources: Banca d’Italia (2000); Casman and Lenoir (2004); Flood and Olovsson (1998); SOEP Group (2001); Taylor et al. (2004).

328

2104 207*

27** 218 1 5

29**

1223** 8 270** 239** 12**

RE

FE

2111 211*

27* 219 3 8

29**

1470**

Model 1

2139 156

26** 216 1 5 2157 177

25* 216 4 9

212**

2249** 25 134 260

2214** 219 64 2177**

212**

1577**

FE

1331** 30 239** 350** 13**

RE

Model 2

2172* 157

24* 219 1 4

70 105 58 242** 28** 93** 83**

1113** 2117 223* 29 11** 350**

RE

2189* 196*

27* 222 2 4

27 110 60 214** 29** 95** 70**

1354**

FE

Model 3

Determinants of adjusted household income (RE and FE estimates)

Constant (Sweden) Belgium Germany Great Britain Age High education Child (Sweden) Child*Belgium Child*Germany Child*Great Britain Employed (Sweden) Employed*Belgium Employed*Germany Employed*Great Britain Age*Separated Education*Separated Employed*Separated Before separation Trend (Sweden) Trend*Belgium Trend*Germany Trend*Great Britain Separation effect: men Separated (Sweden) Separated*Belgium

Effect

Table 11A.2

2207* 107

23* 217 1 4

1232** 2100 187* 106 12** 322** 2205** 216 59 2154** 56 113 68 225** 211** 102** 83**

RE

FE

2235** 161

25* 219 3 5

2241** 20 129 252 6 117 75 216** 211** 101** 72**

1467**

Model 4

2182* 90

23** 216 7 0

1229** 293 166* 11 11** 311** 208** 27 95 2119* 67 83 73 236** 212** 89** 61**

RE

2207* 129

26** 220 9 21

2243** 8 178** 217 20 70 80 246** 212** 89** 47**

1430**

FE

Model 5b

329

25 6 40** 58**

26 10 44** 60** 25 29 13 6 .042

275 2248* 2265* 2258*

2113 2214* 2227* 2246**

21 20 8 2 .042 2.06 3364 29 228

18 91

13 98

31 30

22 17 8 27 .055

28 16 44** 70**

231 2224 2284** 2203

.01 3364 29 228

0 10 3 212 .054

27 19 47** 73**

278 2174 2226* 2149

4 26

22 77

25 29 12 5 .053

23 3 38* 56**

275 2226* 2257* 2235*

.03 3364 29 228

0 21 5 0 .053

27 6 43** 59**

2128 2151 2198* 2182*

26** 51 **

Source:

Cross-national data set of separations.

34 18

22 18 7 28 .065

27 12 43** 68**

232 2201 2275** 2184

.01 3364 29 228

2 12 1 212 .064

28 14 47** 71**

294 2112 2196* 293

215 247

Notes: ** p , 0.01; * p , 0.05. a Additional effect for women (women’s absolute effect 5 sum of men’s and women’s effect). b Model 5 excludes individuals living with a new partner after separation in Belgium, Germany and Great Britain.

Separated*Germany Separated*Great Britain Separation effect: womena Separated (Sweden) Separated*Belgium Separated*Germany Separated*Great Britain After Separation: men Trend (Sweden) Trend*Belgium Trend*Germany Trend*Great Britain After Separation: womena Trend (Sweden) Trend*Belgium Trend*Germany Trend*Great Britain R-Square (within) Corr (u_i, Xb) n (individuals) n (observations)

41 129

23 33 20 25 .084

26 27 22 30

235 2251* 2368** 2283**

.09 2028 15 172

1 23 16 214 .082

28 25 23 32*

291 2187 2264** 2185*

230 56

12.

Marital disruption and economic well-being: poverty, income and lifestyle deprivation Arnstein Aassve, Gianni Betti, Stefano Mazzuco and Letizia Mencarini

INTRODUCTION As discussed in the introductory chapter, household structures across Europe are changing and evolving, one particular feature being the significant increase in marital breakdowns. As a result the number of children living in single-parent households, most of which are female headed, has also increased. Most of the debate around the economic consequences of divorce is focussed on gender inequalities, and the most consistent finding from the literature is a rather sharp gender difference in terms of financial outcomes following a marital disruption. Early longitudinal research from the USA and Europe showed that women experiencing a divorce tend to suffer a substantial loss of income, whereas men’s economic circumstances seem rather unaffected or even improve slightly in some cases (Burkhauser et al., 1991; Fritzell, 1990; Jarvis and Jenkins, 1999; Manting and Bouman, 2004, Poortman, 2000; Smock, 1993, 1994). There are many reasons behind this pattern. One reason is that women tend to have lower labour market attachment, and therefore face lower earnings. Another is that children tend to stay with the mother following a divorce, in many cases imposing a major strain on the single female-headed household. Finally, lack of state support is why many divorced women suffer financially. Another consistent finding is strong country differences in terms of the economic penalty associated with a marital dissolution (Burkhauser et al., 1991; Duncan and Hoffman, 1985; Finnie, 1993; Fritzell, 1990; Jarvis and Jenkins, 1999; Poortman, 2000; Smock, 1993, 1994; Smock et al., 1999). The general pattern is that divorced women in Scandinavian countries, with their generous welfare provision, are much better off than divorced women in the UK, a country characterized by poorer welfare provision. Andreß et al. (in this volume) comparing Belgium, Germany, Italy, the UK 330

Marital disruption and economic well-being

331

and Sweden analysing the three main providers of individual welfare – (1) the family, (2) the market and (3) the state – show that the configuration of these providers to a large extent determines the economic outcome of marital dissolution. It is also shown that due to limited welfare provision UK mothers are particularly vulnerable and considerably more dependent on the labour market as a means of maintaining a reasonable level of economic self-sufficiency. As expected, the UK setting is quite different to Scandinavian countries, but also different with respect to continental countries such as Germany. The Swedish welfare state is not only generous in terms of levels, but also provides strong support in terms of extensive childcare infrastructure, a system which enables Swedish mothers to work full-time to a much greater extent than in other European countries, and especially in the UK. However, there is no clear consensus on these findings, especially concerning the issue of gender differences. Many maintain that the gender bias is overestimated and that the actual trend constitutes an increasing number of men who are subject to economic strain following separation (McManus and DiPrete, 2001). Indeed, there are many reasons to believe that men also experience economic problems following separation: alimony payments and the need to find another dwelling (usually the conjugal house is assigned to the woman especially if there are children) may relevantly and negatively alter the lifestyle of divorced men (see also Dewilde, in this volume). Thus it may seem hard to believe that men are better off after marital dissolution. One of the key problems underlying this debate is the definition and measurement of the rather vague concept of ‘economic well-being’. Many use income or poverty status as an overall indicator of economic wellbeing, but these measures suffer from many drawbacks. Poverty status as a measure of well-being is criticized because it divides the population into a simple poor/non-poor dichotomy, based on sometimes arbitrarily chosen thresholds (Cheli and Lemmi, 1995). Of course, the dichotomy is easily overcome by using income as a measure of economic well-being. However, this measure is problematic as it is difficult to assess to what extent an income loss brings about a real drop in living standards, especially in a comparative perspective. Moreover both income and poverty status are only monetary measures of well-being, whereas it is well recognized that well-being itself has many more dimensions, often non-monetary in nature (Atkinson, 2003; Bourguignon and Chakravarty, 2003). Another drawback is that poverty status and income depend on the choice of equivalence scale, which is essentially an adjustment for household composition, acknowledging that within a household there are economies of scale in expenditure. Given that a marital breakdown inevitably modifies the household composition, the equivalence scale becomes of great consequence. However, it

332

Comparative studies

is not clear which equivalence to use, especially in comparative analysis. Thus, it is beneficial to consider measures of well-being in which the use of equivalence scales is not imperative. Differently from previous studies, we analyse here the effect of marital separation on economic well-being using a range of different measures. We show that the estimates and conclusions differ depending on which measure is being applied. Another key issue in assessing the role of marital dissolution on economic well-being concerns selection bias. This is driven by the fact that couples experiencing a marital separation may be qualitatively different from couples not doing so. For example, women who are strongly dependent on their partner’s income might be less likely to separate from them as they are aware of the strong economic distress they would experience were they to split from their partner (Becker, 1991). One way to tackle this issue is to implement a propensity score matching (PSM) technique which nets out the impact of separation from the confounding effects driven by other observed covariates. Obviously, many other unobserved covariates may influence the estimate of the effect of marital dissolution. As a result we combine the propensity score matching approach with a difference-indifferences estimator (DD-PSM) as suggested by Heckman et al. (1998). In this way we control for the effect caused by unobserved variables, provided these are time-invariant. Heckman et al. (1998) show that the assumption underlying the DD-PSM estimator is less restrictive than the simple Conditional Independence Assumption (CIA) on which PSM estimates rely, but of course, in so far as unobserved factors vary over time, the approach will not fully eliminate such bias. The analysis is implemented using data from the European Community Household Panel (ECHP; see Data Appendix), which offers a unique scope for comparability at the European level (Eurostat, 1999). Uunk (2004) shows that welfare state arrangements tend to influence the economic consequences of divorce for women. Income-related arrangements alleviate the economic strains most, followed by employment-related arrangements. His findings underpin the importance of welfare regimes, and show that differences in terms of economic strains associated with divorce are not simply a result of country differences. The country classification by welfare regimes as proposed by Esping-Andersen (1990, 1999) is well established. However, as pointed out by Andreß and Hummelsheim (in this volume), this classification is contested and many alternatives have been proposed. The key issue is that country classifications need to take into account the domain concerned, especially when it comes to family policies. This applies to the case of marital disruptions as well. Of particular interest is to consider how different family policies are able to alleviate the possible economic disadvantages associated with marital disruptions. Andreß and

Marital disruption and economic well-being

333

Hummelsheim consider two main types of policies, each depending on the scope and generosity of welfare provided. The two are appropriately termed: (1) extensive and (2) rudimentary family policies. Within the first type two further classifications are made: (a) the male breadwinner model and (b) the dual-earner model. The second type is distinguished by: (a) the market model and (b) the model of extended family solidarity (the family model, for short). Our analysis provides information about the possible effects of different family policies in European countries, with respect to consequences associated with marital disruption. We recognize the importance of the presence of children in the couple, so we make separate estimates for couples with children only. These estimates are compared with the cases where we include couples with and without children. The chapter is organized as follows: the next section explains how we measure economic well-being; the third section gives details of data and estimation strategy; the fourth section presents the results and finally we conclude.

WELL-BEING DEFINITION AND MEASUREMENT Measuring Well-being: The Conventional Approach A simple approach to measuring an individual’s well-being is to consider whether he or she is poor or not. The poverty threshold is normally defined as 60 per cent of median net household equivalized income. Individuals within the household are deemed poor if the income falls below this threshold. Poverty is consequently a relative measure, and takes into account the individual’s position in the income distribution relative to others within his or her own country. An important feature of this approach is that it overcomes the fact that countries will differ in terms of per capita incomes and their purchasing power parity. A drawback, however, is that it is not clear what constitutes an appropriate poverty threshold. Often 60 per cent of median net equivalized household income is chosen, but many use alternative poverty thresholds of 50 and 70 per cent (Whelan et al., 2001). When assessing economic well-being, any measure of household income must be adjusted to reflect the needs of the people living within the household. Larger households need more income than smaller households to attain the same standard of living; adults have different needs than do children. Additionally, there are economies of scale, meaning (for example) that two adults can live together more inexpensively than they could if living separately. Correction for household composition is conventionally done by calculating an equivalence scale, which is a number

334

Comparative studies

reflecting the needs of the household. The adjustment is done by dividing total household income by this equivalence scale. We apply the commonly used Organisation for Economic Co-operation and Development (OECD) modified equivalence scale, which gives a weight of one for the first adult, 0.5 for other adults than the household head, and 0.3 for children. The original OECD scale was based on empirical studies during the 1980s and proved a ‘common’ and simple scale for Western countries. In this case the first adult was given a weight of 1, other adults 0.7, and children given a weight of 0.5. Hagenaars et al. (1994) revised this scale by means of new empirical findings, noting that the original OECD scale did not properly take into account economies of scales. For this reason the ‘modifiedOECD’ scale was proposed and officially adopted by Eurostat as the common scale in the ECHP. Well-being as a Matter of Degree: The Relative Income Measure Dividing the population into a simple dichotomy of ‘poor’ and ‘non-poor’ is clearly unsatisfactory. An individual’s well-being is not a single attribute that characterizes an individual or household in terms of its presence or absence (Cheli and Lemmi, 1995). Instead we propose a measure treating poverty as a matter of degree: in principle all individuals are subject to poverty, but to varying levels. That level, say 1 for the poorest to 0 for the richest, is determined by the individual’s rank in the income distribution, and the individual’s share in the total income received by the population. The state of poverty is thus seen in the form of ‘fuzzy sets’ to which all members of the population belong, but to varying degrees (membership functions). A number of authors have used the concepts of fuzzy sets in the analysis of poverty and living conditions (Lemmi and Betti, 2006) and the present contribution represents a continuation and further development of the work of Cerioli and Zani (1990), Cheli and Lemmi (1995) and Betti and Verma (1999). There are several advantages to treating poverty in this way. Most important is that it utilizes the whole distribution directly as a measure of economic well-being, as opposed to dividing the population by a dichotomous category and hence avoids the specification of a poverty line. Equally important is the potential of this approach in studying poverty (or more generally, deprivation in multiple dimensions) in the longitudinal context. The conventional approach measures mobility simply in terms of movements across some designated poverty line, and does not reflect the actual magnitude of the changes affecting individuals at all points in the distribution. Consequently, the degree of mobility of persons near to the chosen line tends to be overemphasized, while that of persons far from that line is largely ignored (Verma and Betti, 2005).

Marital disruption and economic well-being

335

The degree of income poverty associated with each individual as specified below was first proposed by Betti and Verma (1999), who termed it as ‘Fuzzy Monetary’ (FM). The approach was later officially adopted by Eurostat (Giorgi and Verma, 2002). The approach can be explained as follows. Let yi be the net equivalized household income associated with individual i. An income index is defined as the sum of the incomes of individuals richer than the individual concerned (numerator), divided by the sum of the incomes of all n individuals in the sample (denominator): a yj 0 yj . yi Vi 5

j

n

(12.1)

a yj

j51

This index can be seen as the share of the total income that belongs to individuals richer than the individual concerned. It is easy to see that this index is (almost) 1 for the individual with the lowest income and equals 0 for the one with the highest income. The degree of income poverty (FMi) is defined as a monotonically increasing function of the income index Vi defined as: FMi 5 (Vi) a/Hc

(12.1b)

where Hc is the proportion of individuals who fall below the country poverty threshold, defined as 60 per cent of median net equivalised household income. The parameter a is determined such that for the European population as a whole, the mean of the index FMi is equal to the proportion of poor individuals. The latter is imposed to better facilitate comparison with the conventional approach. A Multi-dimensional and Comparative Perspective: The Deprivation Index Measures The relative income measure FMi overcomes the simplistic poor and non-poor categorization of the population. However, relative income considers deprivation only in its monetary dimension, disregarding other non-monetary aspects which may be important for individual well-being. This calls for a measure which considers deprivation in its multiple dimensions (Tsui, 1985). Certainly, in our application of consequences of marital disruption, we expect that individuals’ experiences of well-being go beyond a simple drop of income: some can experience a dramatic rise in monthly expenses (for example alimony payments) with a substantial change of

Comparative studies

336

lifestyles. Moreover, a marital disruption is likely to change, sometimes dramatically, the housing situation of the individuals involved. Just as in the FM approach described above, we define here the concept of multiple deprivation as a matter of degree. In doing so we select a list of items indicating non-monetary deprivation in the households (see the appendix at the end of this chapter). These items often take the form of simple ‘yes/no’ dichotomies (such as the presence or absence of enforced lack of certain goods or facilities), whereas other items may involve more than two ordered categories, reflecting different degrees of deprivation. At the first step these items are grouped into five different dimensions of deprivation. Thus we want to analyse not only an overall deprivation index but also the deprivation indices for each dimension of well-being. Approaches of this kind applied to poverty analysis of European countries are becoming increasingly common (Aassve et al., 2005; Eurostat, 2002). The five dimensions d 5 1, 2, . . ..5 are identified from factor analysis and are as follows: (1) basic non-monetary deprivation; (2) secondary nonmonetary deprivation; (3) lack of housing facilities; (4) housing deterioration, and (5) environmental problems (see Whelan et al., 2001 for details). The second step consists of creating a deprivation score for every item. Consider the general case of item k with m 5 1 to M ordere d categories, with m 5 1 representing the most deprived and m 5 M the least deprived situation. Let mik be the category to which individual i belongs with respect to item k. As in Cerioli and Zani (1990) we assume that the rank of the categories represents an equally-spaced metric variable, and adopt the deprivation score: dik 5

Mk 2 mik , Mk 2 1

1 # mik # Mk

(12.2)

The third step involves determining weights to be assigned to each item of the deprivation index. The weighting procedure we propose here is a variant of the procedure developed by Betti and Verma (1999) and incorporates crucial dimensions of how the items are distributed in the population. First, the weight is determined by the item’s power to differentiate among individuals in the population, that is, by its dispersion. We take this into account by letting the weight be directly proportional to the coefficient of variation of deprivation score dik. Thus items that affect only small proportions of the population – which can be expected to be considered more critical for the affected individuals (Filippone et al., 2001) – are given a larger weight. Secondly, in order to avoid redundancy, it is necessary to limit the influence of those characteristics that are highly correlated with the others within each index for the five dimensions. This means that the weight of item k in deprivation dimension d is taken as the inverse of an

Marital disruption and economic well-being

337

average measure of its correlation with all the variables in that dimension. There are many examples where items within a dimension can be correlated. One is the two items relating to possession of a television and a video recorder. It is unlikely that a household will possess a video recorder unless they possess a television set as well, thus inducing a positive correlation. Similarly, different items describing the conditions of the dwelling may also be correlated. For instance, a dwelling plagued by rot in window frames or floors is also more likely to report damp walls, floors and foundations (see the appendix to this chapter for a detailed description of the items). However, a household reporting both items should not be counted as being two times worse off than a household reporting none of these items. The final weight is proportional to the product of the two factors: the coefficient of variation of the deprivation score, and the inverse of the average of the correlations. In the fourth step we define, for each dimension and for each individual i, the deprivation score Sd, i only considering the items belonging to dimension d: a wk (1 2 dik) Sd, i 5

k

(12.3) a wk k

where wk are the weights defined above in the third step. Note that (12.3) defines a ‘positive’ score indicating lack of deprivation. We can also consider an overall deprivation score which is a simple average of the dimension-specific scores defined above: 1 5 a Sd, i 5 d51

Si 5

(12.4)

The final step is to create the non-monetary indicators of deprivation. As in the monetary approach, we define the individual’s degree to non-monetary deprivation (FS) as the share of the total ‘non-deprivation’ assigned to all individuals less deprived than the person concerned. It varies from 1 for the most deprived, to 0 for the least deprived individual: a Sj 0 Sj . Si FSi 5

j

n

.

(12.5)

a Sj

j51

The same formulation is applied within each of the five dimensions in order to derive the corresponding degrees of deprivation.

338

Comparative studies

DATA AND ESTIMATION STRATEGY Data and Definition of Marital Breakdown The European Community Household Panel is a set of comparable largescale longitudinal studies implemented by the European Union. The first wave of the ECHP was collected in 1994 for the original countries in the survey: Germany, Denmark, the Netherlands, Belgium, Luxembourg, France, the UK, Ireland, Italy, Greece, Spain and Portugal. Three countries were late joiners to the project: Austria joined in 1995, Finland in 1996 and Sweden in 1997. All countries except Luxembourg, Sweden and Germany are included in the analysis. Luxembourg is omitted because of small sample size, Sweden because the data do not form a panel and Germany is dropped because the information necessary to construct the deprivation indices is not available. Eight waves of the ECHP were collected in total, the last in 2001. Different institutional contexts are operationalized with the country typology proposed in the introductory chapter (Andreß and Hummelsheim, in this volume). We aggregate the data as follows: the United Kingdom and Ireland are classified as members of the market model, Finland and Denmark as members of the dual-earner model, Belgium, the Netherlands, France and Austria as members of the male breadwinner model and, finally, Italy, Spain, Portugal and Greece as members of the family model. As pointed out by Andreß and Hummelsheim the four different policy models should be considered as ideal types. Any country-specific welfare provision will necessarily be a mix of the four models. Thus, dividing the countries the way we have done here can be disputed. Nevertheless, the classification is not unreasonable if we consider the dominant features of the countries included in the analysis. In any case, our classification simplifies the interpretation of the estimates considerably. The assumption is that there is little gender-specific division of labour in the dual-earner model and correspondingly fewer negative consequences for women in these countries. In the two rudimentary family policy models, a definite form of labour division is not explicitly encouraged by the state. Rather, it is left to the partners to coordinate employment and childcare and, hence, it is much more difficult to make a-priori statements about the consequences for men and women. However, market forces and missing private safeguards possibly have a much stronger negative effect for women than for men. Finally, the male breadwinner model ranks somewhere in between. On the one hand, a traditional division of labour is supported by the institutional context. On the other hand, this model often affords a compatibility of childcare and work based on part-time

Marital disruption and economic well-being

339

employment of the mother and it usually provides generous cash transfers for (single) parents with children. The event of interest is marital dissolution that is defined by separation or a divorce. Normally a marital dissolution is initiated by a separation and followed by a formal divorce. The ECHP reports self-reported marital status where separation and divorce are mutually exclusive categories. However, laws and regulations on separation and divorce vary across European countries. Thus, the duration between separation and the divorce will not only differ between individuals within a country, but will also differ systematically between countries. A problem arises from the fact that individuals might put different meaning to the term separation. In general, one would expect that a separation implies that the couple move to separate dwellings. However, in many cases a physical separation may not take place until the couple is formally divorced. This means that many couples report separation despite living in the same dwelling, sometimes for several time periods. Consequently it is problematic to use the selfreported marital status to define the time when a marital split occurred. In this analysis a couple is not formally recorded as separated unless they also reported that they lived in separate households. We make this distinction since in this situation they cannot benefit from economies of scale of the household, nor can they share the burdens of rearing children. In estimating the impact of marital splits on poverty status we exclude those couples who are already poor before the split. By ‘poor’ we mean those individuals whose equivalent household income is below the poverty line. Consequently, the samples differ according to the poverty line used. By using propensity score matching (to be explained in next section) we estimate the differential poverty entry rate among separated and not separated individuals, that is, the difference of the percentages of individuals entering poverty in the two groups. Including also those defined as poor prior to divorce will make interpretation more difficult since this would potentially mix individuals entering poverty with those exiting poverty. Note, however, that when estimating the effect of marital split on relative income or on deprivation indices, the complete sample is used. Propensity Score Matching In estimating the effect of marital disruption on economic well-being, we face the potential problem of selection bias. That is, couples experiencing a marital separation may be qualitatively different from couples not separating. For example, women who are strongly dependent on partner’s income are probably less likely to separate as they are aware of the strong economic distress they would experience in the case they split from their

340

Comparative studies

partner (Becker, 1991). Here we tackle this issue by implementing a propensity score matching technique. Applications of this kind are growing in literature (see, among others, Blundell et al., 2005; Dehejia and Wahba, 1999; Lechner, 2002). In our setting we assume that each individual i has two potential outcomes, Y1i in the case he or she experiences a marital split (the treatment) and Y0i in the case he or she does not (the controls). In the case of a marital split Y0i or conversely Y1i in the case of no split, is also referred to as the ‘counterfactual’. The causal impact is given by the comparison between Y1i and Y0i. Obviously, only one of these two outcomes is observable for each individual making a direct comparison impossible, a problem often referred to as the ‘fundamental problem of causal inference’ (Holland, 1986). Let Di be the treatment variable taking the value 1 if individual i receives the treatment (marital split) and 0 otherwise. One characteristic of interest is referred to as the average treatment effect on the treated (ATET) and is expressed as: ATET 5 E (Y1i 2 Y0i 0 Di 5 1) 5 E (Y1i 0 Di 5 1) 2 E (Y0i 0 Di 5 1) (12.6) In order to estimate ATET we need to identify E(Y0i | Di 5 1), which can be done by imposing assumptions on the selection process. An easy solution is to use a naive estimator of ATET consisting of observed difference between treated and control groups, which assumes ATET 5 E (Y1i 0 Di 5 1) 2 E (Y0i 0 Di 5 0)

(12.7)

But (12.7) assumes that there is no selection bias, which means that the group of treated is randomly selected from the total population so that in all other relevant respects apart from receiving the treatment the two groups may be regarded as comparable. It is well known that in observational studies this assumption is overly strong and treated and control groups are systematically different, implying that an estimate based on (12.7) would be biased for true ATET. It is important to understand the nature of the bias that arises. Heckman et al. (1998) propose writing the bias B as a function of a set of pre-treatment observed covariates X: B 5 3 E (Y0 0 X, D 5 1) dF (X 0 D 5 1) 2 3 E (Y0 0 X, D 5 0) dF (X 0 D 5 0) (12.8) S1X S0X where S1X and S0X are the supports of X for D 5 1 and D 5 0 respectively. These are the sets of values of X we observe for the treated group (D 5 1)

Marital disruption and economic well-being

341

and for the control group (D 5 0). Based on (12.8) Heckman et al. (1998) derive a decomposition of B into three terms they refer to as B1, B2, and B3. B1 arises when the supports of the observable X for the treated and the control group S1X and S0X are not overlapping, that is, among the treated group we observe values of X that are not observed in the control group and vice versa. This implies that for treated individuals whose values of X lie out of the common area of S1X and S0X (the common support) we are unable to find an equivalent (that is, same values of X) individual in the control group to match with. Term B2 depends on mis-weighting of E(Y0i | Di 5 0) in the common support. It arises when the distribution of X is different between the treatment and the control group. Finally, term B3 refers to the bias that arises if the distribution of unobserved variables is different between the treated and untreated (see Heckman et al. 1998, for a more detailed discussion on B1, B2 and B3). The way in which the biases in (12.8) are eliminated by our proposed matching procedure is explained below. The matching method is based on the critical assumption termed conditional independence assumption (CIA) stating that treatment status is random conditional on a given set of X. The CIA is formally expressed as: Y0 ' D 0 X

(12.9)

and means that conditional on X the potential outcome in case of nontreatment (that is, Y0) is independent of the treatment status. Whereas (12.9) impose full independence, identification of ATET requires a less restrictive condition. As Smith and Todd (2005) point out it suffices that we impose mean independence, that is, E(Y0i | Xi, Di 5 0) 5 E(Y0i | Xi, Di 5 1). Thus the ATET can be written as: ATET 5 E (Y1i 2 Y0i 0 Di 5 1) 5 EX 0 D51 { E (Y1i 2 Y0i 0 Xi, Di 5 1) } 5 EX 0 D51 E (Y1i 0 Xi, Di 5 1) 2 EX 0 D51 E (Y0i 0 Xi, Di 5 1) } 5 EX 0 D51 E (Y1i 0 Xi, Di 5 1) 2 EX 0 D51 E (Y0i 0 Xi, Di 5 0) } (12.10) The first two lines in (12.10) expand the ATET defined in (12.6), whereas in the third line we apply CIA in its less restrictive form by substituting the second (unidentified) term EX 0 D 51 { E (Y0i 0 Xi, Di 5 1) } by EX 0 D 51 { E (Y0i 0 Xi, Di 5 0) } . In (12.10) the ATET is now fully identified, and is a direct consequence of CIA. Though theoretically appealing, the matching approach is in practice difficult to apply when the dimension of X is high because of the computational difficulties in estimating the conditional expectations in (12.10). Instead of matching on the basis of X

Comparative studies

342

one can equivalently match the treated units to control comparison units on the basis of a balancing score. A particular form of this is a ‘propensity score’ (Rosenbaum and Rubin, 1983), which is the conditional probability of receiving the treatment given the values of X: p(Xi) 5 Pr(Di 5 1 | Xi). The propensity score is usually estimated with either a logit or a probit model. This result reduces the dimensionality problem of computing the conditional expectation, as we now only need to condition on a one dimensional variable (i.e. the propensity score) and ATET can be written as: ATET 5 Ep(Xi0) { [ E (Y1i 0 Di 5 1, p (Xi)) 2 E (Y0i 0 Di 5 0, p (Xi)) ] } (12.11) Several matching procedures can be used to estimate (12.11) (see, for example Becker and Ichino, 2002; Smith and Todd, 2005), but all of them can be seen as generated by the following formula: ^

ATET 5

1 c Y1i 2 a wij 3 Y0j d n1 i[ {a Di 51} j[ {Dj 50}

(12.12)

where the weight wij is defined according to the matching method used and n1 is the number of treated individuals. Here we implement a nearest neighbour matching consisting of pairing every treated unit with the closest control unit in terms of their propensity score. Thus, a treated unit i is paired with the control unit j that gives the smallest value of |p(Xi) 2 p(Xj)|, meaning that for every i the weight wij is one for unit j that is closest, otherwise the weight is zero (see Caliendo and Kopeinig, 2005; Smith and Todd, 2005). Of course in some cases k control units (with k > 1) may satisfy the matching rule (that is, there are more than one control with the minimum distance from the treated). If so we use all these k controls with weight 1/k. All three sources of bias in (12.8) are now eliminated. B1 is eliminated by allowing matches only in the common support region, that is, treated and control units whose values of X lie outside the common area of S1X and S0X are ruled out from estimation of ATET. B2 is eliminated because the control units are re-weighted according to the value of p(X), leading to balance of X between treated and control units. B3 is the only component of (12.8) that is not eliminated by matching but is assumed to be zero by CIA. The matching procedure described above can be implemented on crosssectional observations by recording both treated (that is, those who experienced a marital split) and the controls (those who did not split). However, our data source is longitudinal which means that measurements of the outcomes of interest are available both before and after treatment. This is a highly useful feature since we are able to compare the mean change of well-being from one time period t to another, t 1 1, of treated, with the

Marital disruption and economic well-being

343

mean change of well-being for the same time period for controls. It means that we can define a difference-in-difference (DD) estimator as follows: 2 Yt1i) 2 E (Yt11 2 Yt0i) 5 E (D1i) 2 E (D0i) (12.13) DD 5 E (Yt11 1i 0i An important advantage of the DD estimator is that it allows us to control for selection into the treatment group caused by unobserved variables. To see this clearer we can define the point-wise bias at X at time t as Bt (X ) 5 E (U t0i 0 Xi, Di 5 1) 2 E (U t0i 0 Xi, Di 5 0) , where U0i is the value of unobserved variables (Heckman et al. 1998). Whereas the CIA assumes that B3, which is a weighted average of B(X), is zero, the critical identifying assumption for the DD estimator is that: Bt 11 (X) 2 Bt (X) 5 0.

(12.14)

In this way the CIA is relaxed since we no longer assume B(X) to be zero, rather we only assume its value does not change from wave t to wave t 1 1. In other words, provided unobserved heterogeneity is time-fixed, its effect will be netted out by taking first difference. As a result it has been argued that the DD-PSM estimator is more robust since it eliminates temporarilyinvariant sources of bias (for example, Dehejia and Wahba, 1999, 2002; Smith and Todd, 2005). Of course, even this assumption might be violated if some time-varying source of bias among the unobserved variables exists. For this reason the selection of matching variables remains a crucial part of estimation. By including all the variables correlated with both the outcome and the treatment in the model, estimating the propensity score makes (12.14) more likely to hold. The final estimator of the impact of marital split on well-being is given by: DD 2 PSM 5 Ep(Xi) { [ E (D1i 0 Di 5 1, p (Xi)) 2 E (D0i 0 Di 5 0, p (Xi)) ] } (12.15) Estimation of (12.15) is done using the estimator defined in (12.12), where the values of D1i and D0i substitute Y1i and Y0i. The DD-PSM estimator is implemented throughout the analysis. However when we estimate the effect of separation on poverty status, DD-PSM and cross-sectional estimators are equivalent given that all those who are poor before the marital split are ruled out from analysis. This means in the analysis of marital split on entering poverty Yt 5 0 for all individuals (that is, non-poor individuals are not included in the sample). In all samples the variables which are suspected to confound the effect of marital split on poverty (or deprivation) are included in the estimation of the

344

Comparative studies

propensity score: year of survey, age, number of children, well-being level prior the event (measured both in terms of income and in terms of deprivation), education and employment status (see the appendix to this chapter). Though estimation results predicting participation into treatment might be of interest, it must be kept in mind that the main purpose of propensity score estimation is to ensure that the distribution of observed covariates is the same between treated and the matched controls. If this is the case then all covariates in X are balanced and satisfy what is termed the balancing property (Augurzky and Schmidt, 2001). Clearly this needs to be tested, since if the balancing property is not satisfied, we would potentially match units with quite different values of X despite their propensity scores being close. In this scenario we would need to correct our logit model underlying the propensity score estimation. Following Dehejia and Wahba (2002) and Smith and Todd (2005) we use a t-test for equality of means for each covariate X, before and after matching. This is likely to suffice considering that almost all the variables in our application are binary. Accepting the null hypothesis means that control units are not different from the treated except for the treatment status. Other tests have been suggested (see, for instance, Dehejia and Wahba, 2002; Sianesi, 2004) but there is no clear consensus on which is the most powerful. Becker and Ichino (2002) argue that using a t-test for equality of means with 0.05 significance level is a relatively conservative approach, especially since this level applies to each single variable used in the propensity score model. The balancing property is satisfied in all of our estimates. The estimation of standard errors of ATET is not a trivial exercise – the main problem being that the estimated variance of ATET should also include the variance due to estimation of the propensity score. The common solution to this problem is bootstrapping (see, for example, Blundell et al., 2005; Lechner, 2002). This is the solution we adopt, using the module developed by Leuven and Sianesi (2003) for STATA. Reported t-values are the ratio between the estimated ATETs and the bootstrapped standard errors. For formal significance testing these ratios may be referred to as critical values using the usual standard normal approximation.

RESULTS Entering Poverty Taking the sample of individuals defined as non-poor in the previous period, Table 12.1 presents the effects of experiencing a divorce/separation event on entering poverty using different poverty thresholds. Note that the estimate refers to what is called the average treatment effect on the treated,

345

Source:

0.819 2.803 2.574 1.735 1.636 0.540

0.009 0.043 0.065

0.038 0.045 0.016

ECHP (1994–2001), own computations.

1.372 1.444 1.372

1.250 0.518 0.000

t-value

0.029 0.047 0.057

0.030 0.016 0.000

ATET

All couples

0.057 0.007 0.042

0.009 0.024 0.039

0.019 0.071 0.064

0.020 0.045 0.011

ATET

1.982 0.225 1.106

0.570 1.183 1.555

0.766 1.909 1.451

0.623 1.024 0.250

t-value

Couples with children

Males

0.217 0.245 0.295

0.123 0.210 0.227

0.110 0.250 0.276

0.335 0.389 0.432

ATET

9.008 7.430 8.012

5.401 7.593 8.066

2.978 5.987 6.612

7.328 8.543 9.086

t-value

All couples

0.215 0.284 0.294

0.147 0.217 0.242

0.103 0.204 0.296

0.365 0.414 0.509

ATET

6.741 7.016 8.445

7.058 8.557 9.353

3.418 4.926 5.223

7.262 7.017 8.755

t-value

Couples with children

Females

Average treatment effect on the rate of poverty entry at different poverty thresholds

Market model 50% threshold 60% threshold 70% threshold Dual-earner model 50% threshold 60% threshold 70% threshold Male breadwinner model 50% threshold 60% threshold 70% threshold Family model 50% threshold 60% threshold 70% threshold

Table 12.1

346

Comparative studies

and reflects therefore the difference between the rate of entering poverty for married couples and individuals experiencing a marital breakup. The results confirm that women are considerably more likely to enter poverty as a result of divorce compared to men. This is the case independent of countries and the poverty threshold used. Moreover, the effects are largely consistent with the hypothesized institutional effects. Especially with the 50 per cent threshold, the ranking of the four policy models is perfectly in line with our argumentation in the ‘Data and definition of marital breakdown’, the dual-earner model having the smallest effect followed in ascending order by the male breadwinner model, the family model and, finally, the market model that presents the highest effect. However, this ranking does not remain perfectly consistent if we consider higher poverty thresholds. By using 60 or 70 per cent of median income, the effect of marital disruption increases dramatically for the male breadwinner and the dual-earner models. In fact the dual-earner countries reach in this case the levels of the countries classified as belonging to the family model. Thus, divorce clearly affects women in the dual-earner model as well in that they are considerably more likely to enter ‘mild’ poverty, and they are more likely to do so than divorced women in the male breadwinner model. Women in the market model clearly experience the strongest effect, independent of the poverty line used. Note that the sample mainly consists of individuals from the United Kingdom, as the number of separations and divorces is rather low in Ireland. As expected the effect for men is far lower and only in the male breadwinner model do we find significant effects (when the poverty line is 60 per cent or 70 per cent of median income). The market model also has the largest gender difference. This gender difference is slightly larger than what we find in the family model. When we consider only couples with children the effect of marital disruption is even stronger: For women in the market model the rise of the poverty entry rate is beyond 0.5 when the poverty threshold is set at 70 per cent of median income. For men the figures are not significantly different when we consider only those with children. Fuzzy Monetary Indicator The results reported in Table 12.2 are the estimates of the average treatment effect for the fuzzy monetary indicator, namely, the relative income. These estimates reflect a decline or a rise in terms of the ranking of incomes within a certain country. In other words, a positive effect means a decline in the income ranking due to marriage dissolution, whereas a negative effect means a rise. Therefore for the market model, for instance, women tend to experience a strong decline whereas men’s ranking remains approximately

Marital disruption and economic well-being

347

Table 12.2 Average treatment effect of marital dissolution on relative income Males All couples

Couples with children

Females All couples

Couples with children

ATET t-value ATET t-value ATET t-value ATET t-value Market model 0.014 0.624 –0.011 –0.508 Dual-earner 0.022 1.178 0.035 1.716 model –0.000 –0.031 –0.001 –0.541 Male breadwinner model Family model 0.011 0.472 –0.015 –0.696 Source:

0.298 0.141

9.039 6.106

0.346 0.077

8.268 3.918

0.144

9.464

0.152

7.528

0.205

9.561

0.199

7.697

ECHP (1994–2001), own computations.

the same after separation. Whereas the results for the market model are consistent with the estimated poverty entry rates presented in Table 12.1, the situation is less straightforward for the other countries. In contrast to the poverty entry rate for men, where some significant effects are noted, declines for men are generally small and statistically insignificant in all institutional contexts. In the family model, declines for women are weaker than in the market model but stronger than in the other two country groups. Women from the dual-earner and the male breadwinner models experience approximately the same decline when we consider the whole sample. However, when considering only couples with children in the dualearner model declines are lower, and lower than in the whole sample for that group. Thus for this group, the effect of separation on own income ranking is milder if they have children. Interestingly we find a reversed trend for the market model: women with children experience a stronger effect compared to all women (with or without children). This is largely in line with results of the divorce effect on the poverty entry rate reported in Table 12.1. The male breadwinner and the family model show no relevant difference between the whole sample and women with children. This is again consistent with the results in Table 12.1. Deprivation Indices We now move to the effect of marital dissolution on total household deprivation. We consider first the change in the total deprivation index due to

Comparative studies

348

Table 12.3

Average treatment effect of marital dissolution on deprivation index Males All couples

Females

Couples with children

All couples

Couples with children

ATET t-value ATET t-value ATET t-value ATET t-value Market model Dual-earner model Male breadwinner model Family model Source:

0.124 0.023

3.100 0.723

0.093 0.073

2.203 2.054

0.138 0.106

4.166 3.646

0.123 0.097

3.094 2.736

0.041

2.688

0.044

2.341

0.058

3.682

0.075

4.795

0.034

1.137

0.036

1.112

0.115

4.860

0.105

3.831

ECHP (1994–2001), own computations.

separation. We then consider in more detail the effect of separation on the five dimensions of deprivation as defined earlier. Here we are showing only the estimated average treatment effect on treated for the overall deprivation index, the basic lifestyle deprivation index, and the secondary lifestyle deprivation index. The estimates for the remaining indices (housing facilities, housing deterioration and environmental problems) are omitted as in none of these cases did we find significant effects of marital separation. Total deprivation The results reported in Table 12.3 show a somewhat different picture than the analysis of poverty entry rates. The effect for women in the market model is still the highest, but now the dual-earner and the family models show quite similar figures for both men and women. We find the lowest impact in the male breadwinner model. Importantly, the effects are now significant also for men, and though the magnitude of the effects is always lower than women, there is less of a gender gap. The effect for men in the market model is particularly striking. Compared with poverty entry and change in relative income (Tables 12.1 and 12.2), the effect on deprivation for men is considerably stronger, and not too different from that of women. The gender bias in the market model is considerably smaller in so far as well-being is measured in terms of overall deprivation. Men in the male breadwinner model suffer a significant rise of deprivation after separation as well, but this is consistent with the figures we reported for poverty entry. As with the market model, the male breadwinner model

Marital disruption and economic well-being

349

Table 12.4 Average treatment effect of marital dissolution on basic lifestyle deprivation index Males All couples

Females

Couples with children

All couples

Couples with children

ATET t-value ATET t-value ATET t-value ATET t-value Market model Dual-earner model Male breadwinner model Family model Source:

0.114 0.033

2.785 0.850

0.136 0.100

2.178 2.251

0.224 0.166

4.541 3.646

0.194 0.104

3.303 2.173

0.086

4.840

0.089

3.904

0.127

6.010

0.145

5.881

0.025

0.809

0.024

0.613

0.126

4.374

0.118

3.988

ECHP (1994–2001), own computations.

now shows a quite narrow gender gap. Thus by measuring well-being in terms of total deprivation the geographical pattern of gender differences changes dramatically. Now the dual-earner and the family models have the largest gender differences out of the four policy models. This time the separation effect changes when considering only couples with children: The effect for males is milder in the market model and stronger in the dualearner model, whereas it does not change much for the other policy models. For women we observe a smaller effect of having children in the market model and a higher one in the male breadwinner model. Basic lifestyle deprivation If we focus on the first dimension of deprivation, that is, deprivation on basic lifestyle, we find results shown in Table 12.4 which are relatively consistent with results for the total deprivation index. Again the market model shows the strongest effect both for men and women, but this time the effect for women is about twice as high compared with men, reflecting a strong gender bias. The weakest effect is found in the family model, and it is particularly low for men. Moreover, for men in the family model it does not matter much whether children are present or not. Whereas women in the family model suffer considerably more than their male counterparts, they are less badly off compared with women in the other institutional contexts. Again for the dual-earner model we notice a relatively high effect for women and a significant gender gap. Finally, we register as before a significant effect for men also in the male breadwinner model.

Comparative studies

350

Table 12.5

Average treatment effect of marital dissolution on secondary lifestyle deprivation Males All couples

Females

Couples with children

All couples

Couples with children

ATET t-value ATET t-value ATET t-value ATET t-value Market model Dual-earner model Male breadwinner model Family model Source:

0.149 0.069

3.311 2.147

0.148 0.119

2.734 2.625

0.147 0.157

3.067 4.976

0.077 0.129

1.647 3.179

0.052

2.750

0.046

1.987

0.086

4.840

0.109

4.938

0.049

1.578

0.042

1.008

0.134

4.393

0.149

4.826

ECHP (1994–2001), own computations.

The presence of children seems to negatively influence the effect for men: apart from the family model, almost everywhere the effect of a marital split is stronger when we consider only couples with children. Conversely, the effect for women is almost everywhere weaker, with the exception of the male breadwinner model. Secondary lifestyle deprivation Finally, in Table 12.5 we look at the effects of marital disruption on the level of secondary lifestyle deprivation. Surprisingly, we find a rather strong effect for women in the dual-earner model and for the estimates concerning all couples they suffer more than women from the other policy models, though the estimate for women in the market model is at a similar level. Once we consider only women with children, we see that women in the family model are worst off, but it is interesting that women from the dual-earner model are worse off than women from both the market and the breadwinner models. Another interesting feature of these results is the effect of separation for men in the market model, which is now quite close to deprivation for women, that is, the gender gap is reduced when considering secondary lifestyle deprivation. The role of children is particularly interesting for the market model. The estimates for couples with children reveal that for men the effect remains the same, whereas it is lower for women, though the estimate is not significant at the 5 per cent level. A substantial drop is registered also for women in the dual-earner model combined with an increase for men. Conversely, we observe a small

Marital disruption and economic well-being

351

increase of the effect for women in the other two policy models, whereas there is no relevant change for men in these models.

CONCLUDING REMARKS The present work is concerned with the economic consequences of marital disruption for both members of the separating couples. Most of the literature on this topic assesses whether there is a large gender bias, with women being exposed to high poverty risks in the aftermath of separation whereas men seem not to experience any dramatic drop in their income and are sometimes even better off after separation and divorce. Some authors (McManus and DiPrete, 2001) challenge this evidence, suggesting that the gender bias is less strong than generally acknowledged, and that men also suffer economically after marital disruption. Here we suggest two issues that are essential to this debate. First, the conventional measures of wellbeing (that is, income and poverty status) are not entirely satisfying. Poverty status creates a distinction between ‘poor’ and ‘non-poor’, but it is not clear which poverty line should be considered appropriate and why. Moreover, income and poverty status do not encapsulate all the dimensions underlying poverty and social exclusion – only the monetary one. We may expect that men are not suffering in monetary terms in the aftermath of separation but they experience an increased deprivation in lifestyle standards all the same because of a rise in expenses due to alimony payments, new dwelling costs, and so on. The second issue concerns selection. This is driven by the fact that men and women who are at high risk of entering poverty may be more likely to avoid separation. By using a propensity score matching procedure combined with a difference-in-differences estimator we control for such a selection bias. We expected that by using different measures of well-being, we are able to observe that both men and women experience an economic deprivation after separation, women being more deprived in monetary terms and men in non-monetary terms. The results conform largely to our expectations: it is confirmed that the definition of the poverty threshold is an important issue. Results differ considerably depending on whether we use a 50 per cent, a 60 per cent or 70 per cent poverty line. Moreover when we use monetary measures (that is, poverty status and relative income) it is unquestionable that women suffer a disproportionately larger negative effect than men. Also important is that, by using monetary measures, we find that most of the results are consistent with the hypothesized institutional effects. However, the non-monetary measures (that is, deprivation indices) provide a different picture. Women are still found to suffer significantly

352

Comparative studies

more than men, but it is also clear that men’s level of deprivation also increases, and in some cases there is no significant difference between the average treatment effect estimated for men and women. This is the case in the market model when using the overall deprivation index and the secondary lifestyle deprivation index. Children play an important role in explaining gender differences. If there are children in the conjugal dwelling, mothers are much more likely to be granted custody following divorce. Thus the divorce event will for many women imply reduced income (poorer access to the husband’s income) and higher relative expenditures. Men on the other hand, are likely to live alone or with parents, and are much less likely to experience poverty and financial strain. Considering couples with children only in the analysis of entering poverty, we notice that in the market model and the family model the gender gap is even larger, in the dual-earner model it is smaller, and in the male breadwinner model it remains more or less unaltered. However, in terms of deprivation, men do suffer significantly. Many of the items used to compute the deprivation index refer to characteristics of the dwelling. If it is the case that men normally have to leave the dwelling after a divorce, they will, in the short run at least, lose out on many of the goods and services that the household would provide. So although men are not worse off financially, they are worse off in terms of consumer durables and certain expenditure goods. It also seems likely that the new dwelling is often of poorer quality than the original dwelling (see also Dewilde, in this volume), which is consistent with our estimates. The gender difference is clearly smaller when children are not present in the dwelling. With no children, the effect on lifestyle deprivation among men becomes higher, whereas it is slightly smaller for women. One important factor here is that it is less clear which of the spouses will remain in the conjugal dwelling if the couple has no children.

REFERENCES Aassve, A., S. Mazzuco and L. Mencarini (2005), ‘Childbearing and well-being: a comparative analysis of European welfare regimes’, Journal of European Social Policy, 15, 283–99. Atkinson, A.B. (2003), ‘Multidimensional deprivation: contrasting social welfare and counting approaches’, Journal of Economic Inequality, 1, 51–65. Augurzky, B. and C. Schmidt (2001), ‘The propensity score: a means to an end’, Institute for the Study of Labour (IZA) Discussion Paper number 271, March, Bonn. Becker, G.S. (1991), A Treatise on Family, Cambridge, MA: Harvard University Press.

Marital disruption and economic well-being

353

Becker, S. and A. Ichino (2002), ‘Estimation of average treatment effects based on propensity scores’, The STATA Journal, 2 (4), 358–77. Betti, G. and V. Verma (1999), ‘Measuring the degree of poverty in a dynamic and comparative context: a multi-dimensional approach using fuzzy set theory’, Proceedings of the Sixth Islamic Countries Conference on Statistical Sciences, Vol. 11, Lahore: pp. 289–301. Blundell, R., L. Dearden and B. Sianesi (2005), ‘Evaluating the effect of education on earnings: models, methods and results from the National Child Development Survey’, Journal of the Royal Statistical Society, Series A, 168 (3), 473–512. Bourguignon, F. and S.R. Chakravarty (2003), ‘The measurement of multidimensional poverty’, Journal of Economic Inequality, 1, 25–49. Burkhauser, R.V., G.J. Duncan, R. Hauser and R. Berntsen (1991), ‘Wife or frau, women do worse: a comparison of men and women in the United States and Germany after marital disruption’, Demography, 28, 353–60. Caliendo, M. and S. Kopeinig (2005), ‘Some practical guidance for the implementation of propensity score matching’, IZA Discussion Paper n. 1588, Bonn. Cerioli, A. and S. Zani (1990), ‘A fuzzy approach to the measurement of poverty’, in C. Dagum and M. Zenga (eds), Income and Wealth Distribution, Inequality and Poverty, Studies in Contemporary Economics, Berlin: Springer Verlag, pp. 272–84. Cheli, B. and A. Lemmi (1995), ‘A totally fuzzy and relative approach to the multidimensional analysis of poverty’, Economic Notes, 24, 115–34. Dehejia, R. and S. Wahba (1999), ‘Causal effects in nonexperimental studies: reevaluating the evaluation of training programs’, Journal of the American Statistical Association, 94 (448), 1053–62. Dehejia, R. and S. Wahba (2002), ‘Propensity score matching methods for nonexperimental causal studies’, The Review of Economics and Statistics, 84 (1), 151–61. Duncan, G.J. and S.D. Hoffman (1985), ‘A reconsideration of the economic consequences of marital dissolution’, Demography, 22, 485–97. Esping-Andersen, G. (1990), The Three Worlds of Welfare Capitalism, Princeton, NJ: Princeton University Press. Esping-Andersen, G. (1999), The Social Foundations of Post-industrial Economies, Oxford: Oxford University Press. Eurostat (1999), ECHP Data Quality, 108/99, Luxembourg: European Commission. Eurostat (2002), European Social Statistics: Income, Poverty and Social Exclusion: 2nd Report, Luxembourg: Office for Official Publications of the European Communities. Filippone, A., B. Cheli and A. D’Agostino (2001), ‘Addressing the interpretation and the aggregation problem in totally fuzzy and relative poverty measures’, ISER Working Paper, no. 2001-22, University of Essex. Finnie, R. (1993), ‘Women, men and the economic consequences of divorce: evidence from Canadian longitudinal data’, Canadian Review of Sociology and Anthropology, 30, 205–41. Fritzell, J. (1990), ‘The dynamics of income distribution: economic mobility in Sweden in comparison with the United States’, Social Science Research, 19, 17–46. Giorgi, L. and V. Verma (2002), European Social Statistics: Income, Poverty and

354

Comparative studies

Social Exclusion: 2nd Report, Luxembourg: Office for Official Publications of the European Communities. Hagenaars, A.J.M., K. de Vos and M.A. Zaidi (1994), Poverty Statistics in the Late 1980s: Research Based on Micro-data, Luxembourg: Eurostat. Heckman, J.J., H. Ichimura, J. Smith and P.E. Todd (1998), ‘Matching as an econometric evaluation estimator: evidence from evaluating a job training programme’, Review of Economic Studies, 64, 605–54. Holland, P.W. (1986), ‘Statistics and causal inference’, Journal of the American Statistical Association, 81, 945–70. Jarvis, S. and S.P. Jenkins (1999), ‘Marital splits and income changes: evidence from the British Household Panel Survey’, Population Studies, 53 (2), 237– 54. Lechner, M. (2002), ‘Some practical issues in the evaluation of heterogeneous labour market programmes by matching methods’, Journal of the Royal Statistical Society, Series A, 165 (1), 59–82. Lemmi, A. and G. Betti (2006), Fuzzy Set Approach to Multidimensional Poverty Measurement, New York: Springer. Leuven, E. and B. Sianesi (2003), ‘PSMATCH2: Stata module to perform full Mahalanobis and propensity score matching, common support graphing, and covariate imbalance testing’, University of Amsterdam and Institute for Fiscal Studies, London. Manting, D. and A.M. Bouman (2004), ‘Short and long term economic consequences of union dissolution: the case of the Netherlands’, paper presented at the Third European Conference of the Research Network on divorce, Keulen, December. McManus, P.A. and T.A. DiPrete (2001), ‘Losers and winners: the financial consequences of separation and divorce for men’, American Sociological Review, 66, 246–68. Poortman, A. (2000), ‘Sex differences in the economic consequences of separation: a panel study of the Netherlands’, European Sociological Review, 16, 367–83. Rosenbaum, P.T. and D.B. Rubin (1983), ‘The central role of the propensity score in observational studies for causal effects’, Biometrika, 70 (1), 41–55. Sianesi, B. (2004), ‘An evaluation of the active labour market programmes in Sweden’, The Review of Economics and Statistics, 86 (1), 133–55. Smith, J.A. and P. Todd (2005), ‘Does matching overcome Lalonde’s critique of nonexperimental estimators?’, Journal of Econometrics, 125 (1–2), 305–53. Smock, P.J. (1993), ‘The economic costs of marital disruption for young women over the past two decades’, Demography, 30, 353–71. Smock, P.J. (1994), ‘Gender and the short-run economic consequences of marital disruption’, Social Forces, 73, 243–62. Smock, P.J., W.D. Manning and S. Gupta (1999), ‘The effect of marriage and divorce on women’s economic well-being’, American Sociological Review, 64, 794–812. Tsui, K. (1985), ‘Multidimensional generalisation of the relative and absolute inequality indices: the Atkinson-Kolm-Sen approach’, Journal of Economic Theory, 67, 251–65. Uunk, W. (2004), ‘The economic consequences of divorce for women in the European Union: the impact of welfare state arrangements’, European Journal of Population, 20, 251–85. Verma, V. and G. Betti (2005), ‘Sampling errors for measures of inequality and

Marital disruption and economic well-being

355

poverty’, invited paper in Classification and Data Analysis 2005 – Book of Short Papers, pp. 175–9, CLADAG, Parma, 6–8 June. Whelan, C.T., R. Layte, B. Maitre and B. Nolan (2001), ‘Income, deprivation and economic strain: an analysis of the European Community Household Panel’, European Sociological Review, 17, 357–72.

356

Comparative studies

APPENDIX Variables for Calculating Deprivation Indices Dimensions and items of non-monetary deprivation 1. Basic non-monetary deprivation – these concern the lack of ability to afford most basic requirements: Keeping the home (household’s principal accommodation) adequately warm. Paying for a week’s annual holiday away from home. Replacing any worn-out furniture. Buying new, rather than second-hand clothes. Eating meat, chicken or fish every second day, if the household wanted to. Having friends or family for a drink or meal at least once a month. Inability to meet payment of scheduled mortgage payments, utility bills or hire purchase instalments. 2. Secondary non-monetary deprivation – these concern enforced lack of widely desired possessions (‘enforced’ means that the lack of possession is because of lack of resources): A car or van. A colour television. A video recorder. A microwave. A dishwasher. A telephone. 3. Lacking housing facilities – these concern the absence of basic housing facilities (so basic that one can presume all households would wish to have them): A bath or shower. An indoor flushing toilet. Hot running water. 4. Housing deterioration – these concern serious problems with accommodation: Leaky roof. Damp walls, floors, foundation and so on. Rot in window frames or floors. 5. Environmental problems – these concern problems with the neighbourhood and the environment: Shortage of space. Noise from neighbours or outside. Dwelling too dark/not enough light.

Marital disruption and economic well-being

357

Pollution, grime or other environmental problems caused by traffic or industry. Vandalism or crime in the area. Estimation of the Propensity Score of Marital Disruption: Matching Variables Dummies for each year (1994–2001, reference: 1994) Age Number of children Country dummies Log household income (t 2 1) Log personal income (t 2 1) Deprivation index (t 2 1) Employment status Employed (reference) Unemployed Student Out of labour force Education Less than secondary level (reference) Secondary level Higher than secondary level Shortage of space in household

PART IV

Conclusion

13.

When marriage ends: results and conclusions Hans-Jürgen Andreß and Dina Hummelsheim

We started our research with the assumption that the social and economic consequences of separation and divorce in Western Europe cannot be understood without a thorough understanding of each country’s institutional context. In this final chapter, we want to discuss the most important empirical findings with respect to this institutional explanation. Before doing so, we first summarize the results that are common to all the countries represented in our study. We are especially interested in the question of whether men and women are affected differently by separation and divorce. Having established an idea of the divorce effect, we can then turn to the country-specific results. This is a two-step process. First, we ask whether the different national contexts can be summarized in terms of four different models of family support, as proposed in the Introduction. Assuming that our typology turns out to be valid, we then attempt to draw some general conclusions about the social and economic consequences of divorce as they are typically found in different institutional settings. We conclude with a discussion of the methodological and practical implications of this study.

WHAT ARE THE SOCIAL AND ECONOMIC CONSEQUENCES? One overarching result obtained from all of the country studies and comparative analyses in this volume is the observation that separation and divorce is a financial disaster – certainly for women and in most countries for men as well. Only the British and the Belgian country studies report an increase in equivalized household income for men after a breakup (and this is confirmed by the comparative study on household incomes; see Andreß et al., in this volume). An especially vulnerable group is women with children. Throughout all the countries in this study, women are the resident parents in most cases. The economic situation is less clear when 361

362

Conclusion

children live with their fathers. First, it is much less common for children to stay with their fathers when a marriage ends and, therefore, it is quite probable that this is only the case if either the financial situation or childcare is less of a problem (for example, because children are older and less dependent on a carer). Hence, given men’s higher incomes, it is no surprise that fathers living with their children are much better off than mothers, even if the majority of them experience a decline in equivalized household income (see the German country study, Hummelsheim, in this volume). Interestingly, the Danish country study shows that men who have taken custody of their children after a divorce are a relatively successful group of divorcees. The unemployment rates for single fathers went down immediately after divorce, and welfare dependency was less frequent than it was for single mothers. But according to Hussain and Kangas, Danish single fathers are rather privileged in terms of income and education. Therefore, possible selection effects should be controlled for more thoroughly before we can draw any general conclusions (for example, about the possible positive effects of parenting on divorced men). A second overall result is the large gender gap in economic well-being. In relative terms, women are the economic losers in partnership dissolution. Certainly, gender differences in equivalized household incomes vary by institutional context (as we discuss below), but it is a striking observation that even the generous Danish welfare state is not able to balance the economic losses for men and women. An example presented by Hussain and Kangas illustrates the problem (see Table 4.3 in the Danish country study). The example starts with the most optimistic assumptions: during marriage and after divorce, both spouses are gainfully employed, although they have two children; after divorce, the non-resident father transfers the full amount of maintenance payments to his former wife, who takes care of both children. These assumptions are then combined with the generous cash transfers for children and single parents in Denmark, which are paid exclusively to the caregiving mother. Nevertheless, the single mother loses one-third of her previous (marriage) income, while her former husband gets by with a 10 per cent loss in income. Given these assumptions, the economic loss is solely a result of changes in household composition and differences in the market incomes of men and women. During marriage, incomes are shared and expenses are managed for a household of four, but after separation the wife participates only partially in her former husband’s higher income (as a result of the maintenance payments), and both adults have to manage the expenses for their own households. Even the most generous welfare state will have problems compensating for the incomepooling effect and the economies of scale experienced in dual-earner households. Even though this is a very simple example, it communicates

When marriage ends: results and conclusions

363

an important result; namely, that differences in market incomes between men and women and economies of scale in multi-person households have such significant effects that it is no surprise that the gender gap in postseparation incomes abounds in all Western European countries. Hence, equal employment and pay as well as a balanced sharing of the costs of children may be much more effective than any kind of (compensatory) cash transfers from the welfare state. Although this conclusion sheds a rather pessimistic light on the effectiveness of political interventions, it still leaves a lot of things for the welfare state to do; for example, providing a childcare infrastructure that enables all parents to be gainfully employed, guaranteeing men and women equal opportunities on the labour market and securing adequate maintenance payments for the resident parent, either through law enforcement or through advanced maintenance schemes. Unfortunately, little is known about the practice of maintenance law in each of the countries represented in our study. While most country studies report a stunning similarity in legal regulations concerning spousal and child maintenance, our knowledge about their application is mostly anecdotal. Only the German country study reports data on maintenance payments because there have been specialized surveys of German divorcees which collected information about the regularity and amount of maintenance payments. These data confirm the rather pessimistic view in the literature that private transfers from the former spouse are an insufficient and unreliable resource for the recipients (see Hummelsheim, in this volume). This is not to say that they are irrelevant. On the contrary, for small household incomes they may be an important contribution that elevates the household above the poverty line. But certainly they do not fully compensate for the true costs of raising children alone. In addition to the similarities, the discussion of family law in our singlecountry studies also revealed different legal traditions. While in Great Britain, family matters tend to be relegated to the private sphere and have to be settled in private court agreements (see Sigle-Rushton in this volume), other countries specify at least minimum rights (for example, concerning child maintenance) and sometimes even provide advanced payment schemes that serve as a sort of contingency insurance (see also Data Appendix). Furthermore, while Great Britain – at least historically – focused on a ‘clean break’ divorce with few obligations towards the former spouse and children (assuming a quick remarriage), most continental European countries focus on the lifelong promise of marriage that has to be compensated for if it breaks down unexpectedly. Finally, Sweden interferes very little into relationships between adult individuals, but the system encourages parents to share care and financial responsibility for their children both before and after partnership dissolution. Thus,

364

Conclusion

Swedish family law and policy focus on parent–child relations and put more emphasis on cooperation than on compensation. Unfortunately, it is difficult to use these subtle and qualitative differences between different legal regimes in statistical analyses of representative survey data. Problems of operationalization and the lack of information on maintenance payments and other consequential matters of divorce in large multi-purpose surveys are the main reasons why we know so little about the effects of different legal regimes on the consequences of divorce. On the other hand, the little evidence that is presented in our country studies does not make us very optimistic that the present legal regulations will practically offset the aforementioned gender gap. If neither the state nor the former partner fully compensates for the economic loss women experience due to divorce, it is up to the women themselves to improve their economic standing. Thus, there is a strong incentive for separated and divorced women to increase their labour supply. However, according to the comparative analyses by van Damme and Uunk, female labour force participation increases only modestly when a partnership ends. The average employment rate among women before separation was almost 63 per cent for the 13 countries in their study; after separation it was 68 per cent, a modest increase of 8 per cent. Overall, 13 per cent of all women entered the labour market after a divorce and almost 8 per cent left it. Furthermore, women who were already working hardly changed their working hours; the overall increase was insignificant. Although these figures are averages and may be quite different in specific countries (see van Damme and Uunk in this volume), the overall impression is not the expected rise in female employment. Van Damme and Uunk suggest two explanations that are also vividly discussed in the literature: anticipation and selection. As the average pre-divorce employment rate of 63 per cent indicates, women who are at risk of divorce have a much higher propensity to be gainfully employed than the average married woman. Obviously, they have increased their labour supply (long) before the event of separation, either because they have anticipated the financial problems they will experience after separation or because they have been much more economically independent from the very beginning. In any case, this economic independence allows them to end unhappy partnerships more easily than economically dependent women can, but it also means that it is difficult on a practical level to improve their financial situation, because they have already exhausted their earnings potential. We have to be cautious when drawing general conclusions from these averages because they do not characterize the situations of all separated women in every country. But they remind us that increased employment is not an option for many divorced women, because they have already been

When marriage ends: results and conclusions

365

employed full-time before their partnerships break down. If, in addition, they still have to bear the brunt of childcare, and if they do not work in comparable jobs with similar earnings to those of their former husbands, then there is no chance of equalizing the gender gap in post-separation income. As the data from the German country study show, employed women do not fare better than employed men – neither in absolute nor in relative terms (see Hummelsheim in this volume: Table 2.2). In sum, there seems to be no doubt that women are the economic losers in relative terms. Looking at relative incomes, that is, at each man’s and each woman’s position within the national income distribution, women tend to experience a more or less strong decline whereas men’s ranking remains approximately the same after separation (see Aassve et al., in this volume: Table 12.2). A similar picture arises if we look at the vulnerable groups at the bottom of the income distribution. The analyses by Aassve et al. confirm that women are considerably more likely to enter poverty as a result of divorce than men are. When we consider only couples with children, the effect of marital disruption is even stronger. But as we noted in the Introduction, focusing only on monetary indicators of well-being may paint an unbalanced picture of all the social and economic consequences of partnership dissolution; and within a more comprehensive analysis, the situation may appear less rosy for men. In two comparative analyses, we have studied the effects of separation and divorce on residential mobility and lifestyle. Dewilde shows that divorce substantially raises the risk of leaving owner-occupation for both men and women. However, this risk seems to be smaller for women in those countries where the economic consequences of divorce are more severe (see below). For both genders, divorce also increases the risk of entering shared accommodation. Overall, divorce initiates a process of downward mobility with respect to housing status. The effects on housing quality are less clear. The corresponding indicators (size of the dwelling, quality of the dwelling and neighbourhood characteristics) show only small changes or even indicate a small improvement in housing quality.1 Dewilde interprets this finding by pointing out that housing conditions are also an important determinant of divorce. Housing deprivation significantly increases the risk of divorce, and hence, the unclear results may also be due to the fact that some individuals escape out of a situation of housing deprivation. What about gender differences with respect to housing and residential mobility? In fact, there is one obvious difference, but it is related to the aforementioned gender gap in income. After separation, each partner has to pay for his or her own home – which is, of course, much more expensive than sharing one family home before separation. Given women’s (relatively) smaller incomes, it is no surprise that women experience a much

366

Conclusion

larger increase in housing costs than men do. But apart from this wellknown income effect, there is no other significant disadvantage for women. On the contrary, Dewilde found that in some countries women stay in the former family home more often than men do. This may be an indication that women in countries with less formalized maintenance regulations might be compensated for the economic consequences of divorce by advantageous procedures of property settlement. Nevertheless, as Dewilde says, other explanations are equally possible. Outright ownership may be more common, which makes it easier to stay in the marital home. Furthermore, alternative forms of housing may be limited, so that divorced men and women might be forced to stay in owner-occupation even if they cannot afford it. Whatever the causal mechanisms behind this observation, it demonstrates that the consequences of divorce may be more balanced between the genders once we take other dimensions (besides income) into account. In a similar vein, Aassve et al. analyse several non-monetary lifestyle indicators. They distinguish between basic and secondary non-monetary deprivation – that is, the inability to afford the basic requirements of life (for example, keeping the home warm, replacing worn-out furniture, paying regular bills, and so on) and the enforced lack (because of insufficient resources) of widely desired possessions (for example, a car, a telephone, a colour television, and so on). Again, women are found to suffer significantly more from lifestyle deprivation than men do. But it is also clear that men’s level of deprivation increases as well, and in some cases there is no significant difference left between men and women. Furthermore, the presence of children seems to negatively influence the effect for men. With the exception of the Southern European countries, when considering only couples with children, the effect of a marital split is stronger for men almost everywhere. Conversely, the effect for women is weaker in almost every case. Aassve et al. explain this observation in terms of different residential mobility following divorce. If it is the case, as they assume, that men normally have to leave the family dwelling after a divorce, they will – in the short run at least – lose out on many of the goods and services that the household would provide. So although men are not worse off financially, they are worse off in terms of certain lifestyle indicators. Additional insights into the gender gap can be obtained from the Danish country study, which because of its extensive database is able to analyse how men and women cope with the event of divorce over a longer period of time. This longitudinal study shows that women seem to be hit harder by divorce initially, but they appear to be able to cope with their lives as divorcees better than men do. For example, long-term unemployment seems to be a characteristic that applies to men, especially to those men who never remarry and/or have no children. Kangas and Hussein assume

When marriage ends: results and conclusions

367

that the increase in unemployment among divorced males without children may result from the fact that divorced men without children suffer from a substantial loss of social contacts. A similar trend is visible in the case of welfare dependency. Beginning in the fourth year after divorce, figures on the receipt of social assistance begin to decrease for women, whereas there is an upward trend among men. Kangas and Hussein also analyse subjective indicators of well-being and again, men do not always perform better than women. For example, with respect to trust (social capital), single women seem to be better off than their married sisters, while divorced single males perform worst. Couples enjoy the happiest life; and divorced men are again at the bottom. Finally, with respect to health, the most sickness-stricken group seems to be divorced males.

DOES THE INSTITUTIONAL CONTEXT MATTER? Having summarized the overall consequences of separation and divorce, we can now turn to the question of how these divorce outcomes vary between countries. We started with the assumption that the national configuration of family, market and state institutions moderates the consequences of partnership dissolution. According to this hypothesis, economic and social consequences of partnership dissolution are the result of individual risk factors (for example, low earnings capacity, parenting and lost network contacts) and context effects due to the institutional setting in each country. To this end, we proposed a typology of family support in industrialized countries (see the Introduction). This typology starts with the impact of state institutions on individual welfare and distinguishes between extensive and rudimentary family policies. In the first case, the state plays a significant role in welfare provision. In the second case, either the market or the family has to step in. Accordingly, the typology distinguishes between four models of family support: the male breadwinner, the dual-earner, the market and the family model. The typology implies that the institutional context may support separated individuals and their dependent children in different ways: through income support from the state (for example, child benefits, tax allowances, state-enacted support payments), through employment support from the state (for example, public childcare provisions), through family support (for example, financial support via the family network) and through the market, which provides more or less extensive opportunities for men and women to earn a living. We also suggested how the countries in this study could be classified with respect to the typology: Belgium and Germany were classified as

368

Conclusion

representatives of the male breadwinner model; Denmark, Finland and Sweden as representatives of the dual-earner model; Great Britain as a representative of the market model and, finally, Greece and Spain as representatives of the family model. The Introduction stressed, however, that this allocation is tentative and should be checked against the detailed institutional information found in the single country studies. Therefore, the following discussion summarizes the results with respect to two questions: is the tentative classification of countries a valid operationalization of the proposed typology and, if so, how do the different institutional contexts modify the economic and social consequences of divorce? How Should We Classify the Countries Represented in this Study? Figure 13.1(a) shows the amount of public spending that the countries represented in our study invested into families and children in 2003. These data from the Organisation for Economic Co-operation and Development (OECD) Family Database reveal three distinct clusters of countries: (1) Denmark, Finland, France and Sweden, which spend a comparatively large amount on family services; (2) Austria, Belgium, Germany, Ireland and the United Kingdom, which spend a comparatively large portion on cash and tax benefits, and (3) Greece, Italy, the Netherlands, Portugal and Spain, which spend comparatively less money in both dimensions. According to these data, family policies seem to be of little (‘rudimentary’) importance in the third group of countries, while the two other groups of countries seem to practise more extensive family policies. According to our typology, the first group of countries should support the dual-earner model because family services (for example, childcare facilities) usually increase employment opportunities. The second group of countries should support the male breadwinner model, because cash and tax benefits are usually implemented to ease the familial burdens of single earners. In this section, we use the context data from the single country studies to present the reallife situation behind these dry statistics. Following the order of the single country studies in Part II, we start with those countries that were assumed to have an extensive family policy. As the analysis by Mortelmans et al. and the OECD data show, Belgium is indeed such a country, where the state assumes a strong responsibility for the welfare of families. According to Mortelmans et al., Belgian public policies for families are characterized by a combination of high social transfers and an elaborate public childcare infrastructure – although, in a comparative perspective, the latter focus does not show up in the expenditure data (see Figure 13.1(a)) because a large proportion of childcare facilities is church-related and thus financed by the church (cf. Fix, 2001).

When marriage ends: results and conclusions

369

But with almost one-third of all toddlers and nearly 100 per cent of all children aged 3–6 years in publicly funded childcare, Belgium has similar enrolment rates to those in some of the Scandinavian countries (see Data Appendix). In other words, Belgium performs quite well both in terms of income and employment support from the state. As we already mentioned, it is a common assumption that income support decreases work incentives while public childcare increases them (see, for example, van Damme and Uunk, in this volume). In this respect, Belgium is a borderline case between the male breadwinner and the dual-earner models. However, despite the elaborate childcare infrastructure, overall female employment is low compared with the other European countries represented in our study; only the percentage of working mothers with children under the age of 6 amounted to 68.8 per cent in 2005 (see Data Appendix). Thus, at least the high labour market participation of mothers – especially younger women – proves that being a solo parent and working full-time is a tough, though feasible combination in Belgium. Historically, Belgium is a typical example of the conservative welfare regime as described by Esping-Andersen (1990), with work-related social security measures and a strong accent on the family as the ultimate provider of welfare. It is also important to remember that the origin of the elaborate system of public childcare was not a public interest in women’s employment. It dates back to the beginning of the twentieth century2 and was the result of a conflict between church and state about the proper education of children (Fix, 2001). At that time, employment support for women was not a topic of political discourse. A recent analysis of representative survey data from the European Value Study (EVS) and the International Social Survey Programme (ISSP) also shows that Belgians have much more conservative attitudes towards female employment and gender equality than Germans do (Hummelsheim, 2009). Hence, Belgian public policies are characterized by a sort of ‘optional familialism’, where the caregiving family is supported, but at the same time families are also given the option of being (at least partially) unburdened from childcare responsibilities. Dual-earner families are an option, but the actual participation rates of Belgian women tell a different story. Thus, in our context, Belgium can still be classified as a member of the male breadwinner model (although not a typical one) and measures of income support may be much more important to the characterization of the social and economic consequences of divorce in Belgium than they are in other countries. If Belgium is a hybrid case, Germany is a prototype of the male breadwinner model, as the analysis by Hummelsheim shows. Family policy is an important part of German politics. There is a dedicated Federal Ministry of Family Affairs, and in the European context, German public expenditures

370

Services in % of GDP

.5

1

1.5

2

2.5

.5

ES IT

GR

PT

1

1.5 Cash + tax benefits in % of GDP

NL

FI

SE

DK

a) By type of expenditure

2

DE

BE

FR

IE 2.5

AT

UK

371 0

1

2 3 Total expenditures in % of GDP

4

Public spending on family benefits in cash, services and tax measures (percentage of GDP, 2003)

OECD Family Database; Adema and Ladaique (2005).

Figure 13.1

Source:

Spain (ES)

Greece (GR)

Italy (IT)

Portugal (PT)

Netherlands (NL)

Ireland (IE)

Germany (DE)

Finland (FI)

Austria (AT)

Belgium (BE)

United Kingdom (UK)

Sweden (SE)

France (FR)

Denmark (DK)

b) Total expenditures

372

Conclusion

for families are comparatively high, with a strong focus on cash and (especially) tax benefits (see Figure 13.1).3 Thus, there is no doubt that Germany is characterized by an extensive family policy. Hummelsheim, however, shows that most existing political measures support the traditional gender division of labour and compensate primarily for the resulting familial burdens. This overall policy orientation can be seen in the system of joint taxation, which favours the wife as the homemaker or the additional (marginal) worker as well as in the many financial incentives for parents (that is, mothers) to take over childcare by themselves. Even though there are many services in support of families, the main intention is to ease rather than to substitute for the mother’s traditional role as the homemaker and caregiver. This is especially evident for the system of public childcare. At least in the Western part of Germany, it has traditionally been assumed that women will take full-time responsibility for their dependent children until their third birthdays. For somewhat older children, there is a legal claim to attend nursery schools but, until recently, many childcare facilities were open only on a half-day basis. This situation continues for school-age children because of the few full-time schools that are available in Germany. Consequently, female participation rates are much lower in Germany than in other countries in our study (but not as low as they are in Belgium). Because of the meagre system of public childcare, the child effect on female employment is much more negative than it is in other countries (see the activity rates of German mothers in the Data Appendix). This is different from Belgium, where public childcare facilities are much more developed. The three Scandinavian countries in our study, Denmark, Finland and Sweden, were tentatively classified as members of the dual-earner model. According to the OECD statistics, the welfare state in these countries invests at least as much as Germany does into families and children, however with a strong focus on family services (see Figure 13.1). Hussain and Kangas describe Denmark as a comprehensive welfare state including well-developed public childcare facilities, active and flexible labour market programmes and generous income transfers for families. On a more general level, as they point out, the Danish case is a good example of how the ‘male breadwinner’ model of the past has been transformed into a ‘dual-earner’ model; it is no surprise that participation rates of Danish women are the highest among the countries in our study (see Data Appendix). Concerning separation and divorce, there is also a strong built-in equalization mechanism in maintenance payments that attempts to combat the economic hardships caused by divorce and single parenthood. From the statistical data presented in the Data Appendix, it is obvious that Finland and Sweden share a similar policy orientation to that of Denmark. Both countries invest at least 3 per cent of their GDP into

When marriage ends: results and conclusions

373

families and children, and more than half of this sum goes into family services. Hence, there is strong employment support in both countries and accordingly, female participation rates are well above the corresponding figures in the other countries (see Data Appendix). Therefore, Finland and Sweden, together with Denmark, are well-fitting representatives of the dual-earner model in our typology. Nevertheless, a more detailed analysis by Hiilamo also reveals some important differences, especially when it comes to the social and economic consequences of divorce. In his comparison, he refers to two different policy orientations: the egalitarian model, which stresses the government’s role in family welfare and gender equality, and the familial model, which emphasizes the role of the family in guaranteeing the well-being of individuals. An important measure which illustrates this difference is the Finnish home care allowance that is granted to parents with children under 3 years of age. According to Hiilamo, Finland uses this and other policy instruments to emphasize parental choice between home care and day care, while Sweden supports only public day care. In this respect, Finnish family policy tends to be more familial than egalitarian. Accordingly, female participation rates are lower in Finland than they are in Sweden (see Data Appendix), and Hiilamo shows that during the economic recession, employment rates for partnered mothers decreased relatively significantly. This could be a result of the home-care allowance, which allows these mothers to rely on their partners’ income while on home-care leave (see also the comparatively low participation rates for Finnish mothers in the Data Appendix). Finally, looking at divorce and maintenance regulations, it seems that the Swedish system not only avoids building barriers to separation and divorce, but that it also encourages parents to continue sharing the financial and care responsibilities for their children after partnership dissolution. This is also an indication that Sweden has a strong egalitarian orientation built into its family policies. It seems correct to assess Denmark, Finland, and Sweden as representatives of the dual-earner model in our typology. However, recent neo-familial tendencies in Finnish public policies may challenge that country’s assignment. Finally, we have Great Britain, Greece and Spain, which have tentatively been assigned to the category of countries with rudimentary family policies. Looking at the expenditure data in Figure 13.1, this seems to be a good designation for Greece and Spain, but definitely not for the United Kingdom.4 However, in her country report, Sigle-Rushton describes Great Britain as a country that is liberal in its approach to welfare and individualistic in its approach to family matters. Historically, she says, family policy has been minimal in the United Kingdom, reflecting the country’s liberal, individualistic and market-oriented ideology. On the other hand,

374

Conclusion

the United Kingdom was unique (along with the Netherlands) in that it expected solo mothers to remain at home and care for their children full-time. The British welfare state provided low levels of income support so that this was a feasible option for solo mothers. In recent years, by implementing above-inflation increases in child benefits and introducing various types of tax credits, British welfare policies have made a significant move towards greater governmental support for families with children, especially when those families are engaged in (low) paid work. All these measures explain why the United Kingdom has a comparatively good standing in terms of income support for families and children (see Figure 13.1). But much of the government’s support for families with children continues to be limited to those on low incomes, and because eligibility for tax credits is calculated at the household level, these measures may work to discourage second earners – usually mothers – from entering the labour market. Sigle-Rushton also notes that many of these changes took place so recently that their effects will not be reflected in many estimates of the economic consequences of divorce. In 1998, the government also launched a National Childcare Strategy, which resulted in a significant rise in available childcare places, often, however, with limited opening hours. For many dual-earner families, this means that they must rely on a complex mix of formal, informal and back-up support. These logistical requirements still make full employment a difficult task for many mothers, as Sigle-Rushton points out. According to the Data Appendix, participation rates among British mothers are at the same level as those in Germany, our representative of the male breadwinner model. British mothers continue to reconcile work and family by engaging mostly in part-time work, and often only after their children have entered school. Sigle-Rushton notes that part-time workers in Britain face a good deal of occupational segregation and an extremely large part-time pay penalty. In case of separation, these women fall back on rather small incomes and, if maintenance payments are insufficient, they will have a hard time making ends meet. Historically, British families have been expected to make choices and bear the financial (and emotional) costs of those decisions privately. In recent years the government has sought to intervene to a greater extent, especially with respect to child support, but Sigle-Ruston reports that this more interventionist approach has been far from successful. Hence, given the traditional liberal and individualistic policy orientation of the British welfare state, it is certainly legitimate to assign Great Britain to the market model in our typology. On the other hand, however, the British welfare state provides a fair amount of income support for families in the lower income brackets, which makes Great Britain similar to the male breadwinner model.

When marriage ends: results and conclusions

375

Greece and especially Spain rank very low on public expenditures for families and children (see Figure 13.1(b)). Compared to the other countries in the study, both countries are definitely characterized by rudimentary family policies. According to Symeonidou’s country report, Greece is a representative example of the Southern European welfare model. The family plays a crucial role in substituting for the state’s inability to satisfy social needs, and women are the agents of care for older and younger family members. Female participation rates are the lowest in our country sample (see Data Appendix), because women – due to homemaking and caregiving – are often impeded from participating in the labour force, or they work in low-status, low-paid jobs – very often as unpaid family members in their fathers’ or their husbands’ businesses. Although Greece has a rather liberal divorce law and quite a few regulations concerning spousal and child support, the divorce rate in Greece is extremely low (see Data Appendix). Symeonidou explains this seeming contradiction based on women’s weak economic position and their multiple roles in the family, which make divorce a difficult step to take in Greece. She also notes that the practice of the law is quite often very different from the legal regulations. Similarly, Flaquer and Garriga describe how the limited development of the welfare state in Spain goes hand in hand with the strength of family solidarity. One of the most prominent characteristics of the Spanish system of social protection is its low level of provisions for the well-being of families, which has its roots both in the desire to distance itself from former Francoist social policies and in the focus on labour market problems during the 1980s and 1990s. The Spanish welfare state is much more oriented towards retired people than it is towards families and children. Spain has a low level of spending not only on child benefits (mostly directed at very poor families), but on family services as well (see Data Appendix). Measures intended to reconcile work and family life are poorly developed and, correspondingly, female participation rates are as low as they are in Greece. Divorce is clearly not an acknowledged event in the Spanish welfare state. Spain is one of the few countries in Europe where a recent rapid and drastic liberalization of access to divorce has not been compensated for by more effective regulation and supervision of the paying of maintenance on the part of the authorities and by increased public provisions to mitigate the economic consequences of marital disruption. Thus, one of the ways in which many people try to face the adverse consequences of divorce – especially in the lower strata of the population – is by moving in with relatives, particularly parents. In sum, Spain and Greece are clear representatives of the family model in our typology. In addition to the eight countries discussed thus far, Figure 13.1 also shows data for those countries that are represented in the comparative

376

Conclusion

analyses in Part III: Austria, France, Ireland, Italy, the Netherlands and Portugal. With the exception of Austria (male breadwinner model), Italy and Portugal (both in the family model), their assignment to the four models of family support is less clear. France is treated as a member of the male breadwinner model in all comparative analyses, although the data in Figure 13.1 make it a borderline case between the male breadwinner and the dual-earner models. In two comparative analyses, Ireland is classified in a similar way as the United Kingdom. Both are Anglo-Saxon countries, and the data in Figure 13.1 strongly resemble the British case, but van Damme and Uunk assign Ireland to the family model because in Catholic Ireland, as they point out, family ideology is more in line with that of the Southern European countries than with that of the market model in the United Kingdom. Finally, the Netherlands is always a controversial country. On the one hand, the Netherlands has an extensive welfare state and generous social welfare programmes, although this is not evident in public expenditures for families and children (see Figure 13.1). But this is why many scholars (including Aassve et al.) follow Esping-Andersen (1990) and assign the Netherlands to the cluster of Northern European countries: the dual-earner model in our case. On the other hand, when it comes to supporting female employment, as van Damme and Uunk argue, Dutch institutions lag behind and make the Netherlands a representative of the male breadwinner model. In all three borderline cases, more detailed institutional information would be helpful in order to make a more sound assignment. Until then, sensitivity checks with different assignments can indicate how the empirical results depend on the chosen classification (see van Damme and Uunk, in this volume). What Institutional Effects Do We See? Having discussed the validity of our typology of family support across the 14 countries represented in our study, we now come to the question of how these different institutional contexts modify the social and economic consequences of partnership dissolution. In the following section, we summarize the main results from the comparative country analyses in Part III of this volume, which focused on four dimensions of divorce consequences: (1) income and poverty, (2) labour force participation, (3) residential mobility and housing and (4) lifestyle deprivation. Income and poverty In the first section we discussed the gender gap in divorce outcomes, most prominently the different economic consequences for men and women. Therefore, a natural starting point for the following discussion is the

When marriage ends: results and conclusions

377

question of how the gender gap in incomes is modified by the institutional context. The chapters by Andreß et al. and Aassve et al. start from the concept of economic autonomy and how much autonomy is granted to the weaker family members within different institutional contexts. Obviously, a fundamental source of economic autonomy is gainful employment and the availability of a personal income. Since there is little gender-specific division of labour in the dual-earner model, the vast majority of women are gainfully employed and have personal incomes at their disposal. Hence, the authors assume that there will be fewer negative economic consequences for women in these countries. In the two rudimentary family policy models, a definite form of labour division is not explicitly encouraged by the state. Rather, it is left to the partners to coordinate employment and childcare, and hence it is much more difficult to make a priori statements about the economic autonomy of women in the market and the family models. However, market forces and a lack of private safeguards possibly have a much stronger negative effect for women than they do for men. Finally, the male breadwinner model ranks somewhere in between. On the one hand, the institutional context supports a traditional division of labour, which increases economic dependency for women. On the other hand, the male breadwinner model often affords a compatibility of childcare and work based on the part-time employment of the mother, and it usually provides generous cash transfers for (single) parents with children. Thus, economic dependencies may be less strong. This hypothesis is broadly confirmed in the analysis by Andreß et al., who use panel data on divorcees from five countries in our study. They find that the gender gap in post-separation income is low in Sweden, a representative of the dual-earner model, while gender inequality with regard to income is high in the countries of the other family policy models – that is, in Germany, Great Britain, Belgium and especially in Italy. With respect to short-term income losses, Andreß et al. observe the following ranking for women: Italy (highest), Germany, Great Britain, Belgium, Sweden (lowest). As expected, the country from the family model (Italy) performs worst, while the country from the dual-earner model (Sweden) performs best. The relatively good performance of Belgian women is no surprise, given our former characterization of Belgium as a hybrid case between the male breadwinner and the dual-earner models. A similar statement applies to Great Britain, where our previous discussion showed that the United Kingdom has a comparatively good standing in terms of income support for families and children. These additional incomes may ameliorate the economic decline for British women after separation. The ranking of the four policy models is also confirmed by an analysis of relative incomes and poverty entry rates using a larger country sample

378

Conclusion

(Aassve et al., in this volume). Especially at the 50 per cent poverty threshold, the ranking is perfectly in line with our former expectation – the dual-earner model having the smallest effect on poverty entry, followed in ascending order by the male breadwinner model, the family model and, finally, the market model, which shows the highest effect. A similar picture appears when we analyse each individual’s relative position in the income distribution. Women in the market model experience the largest declines in income after separation, while the income decline is less marked for women in the dual-earner model. Labour force participation Owing to their economic losses following separation, women should have strong incentives to increase their labour supply. In principle, this is a feasible option for inactive women or for women who do not work fulltime. Since the majority of men work full-time, this is less of an option for their male partners, irrespective of the fact that their economic needs may be much lower. Thus, female labour participation after separation is an interesting topic for studying the effects of the institutional context. Van Damme and Uunk analyse three characteristics of the context that they expect to influence women’s employment decisions. These are (1) employment support through family services, (2) income support through cash benefits and (3) financial and caregiving support from the extended family. Employment support is expected to have a positive effect on female labour force participation, income support should have a negative effect, and the effect of the extended family is less clear. On the one hand, financial resources from the extended family unburden women from the necessity of earning a living on their own. On the other hand, additional caregivers within the extended family relieve women with children from their familial obligations so that they can be (more) active on the labour market. With the characteristics of the four models of family support in mind, van Damme and Uunk expect that the dual-earner model is the most employment-friendly context for women after separation because of its elaborate childcare facilities and the absence of large family networks that could provide alternative economic resources. The least employmentfriendly context is expected to be the male breadwinner model, which provides disincentives to own employment through generous income support programmes, and where women are forced to take over large portions of the caregiving work due to insufficient public and private alternatives. The market and the family models rank in between. These expectations turned out to be only partly true. Women in the dual-earner model did indeed experience the most positive outcomes after separation. They had the highest odds of employment entry after

When marriage ends: results and conclusions

379

separation, the lowest odds of employment exit and the highest increase in working hours for those working. However, divorced women from the male breadwinner model did not fare as badly as expected. They did not differ from the dual-earner model with respect to employment exit and working hour changes, and with respect to employment entry they ranked between the dual-earner and the market models. Contrary to the expectations, the worst labour market outcomes were observed for the market model. Women from these countries experienced the lowest odds of entry, the highest odds of exit, and a decrease in working hours for those working. But the typology worked quite well in explaining the (negative) child effect on women’s labour supply. In the dual-earner and family models, the presence of young children posed less of an obstacle to separated women increasing their labour supply than it did in the market and male breadwinner models. This observation is perfectly in line with the previous conclusion that childcare is less of a problem in both contexts, either because of the elaborate system of public childcare in the dual-earner model or because of the extended family network in the family model. The results also show which public policies matter. The male breadwinner, market and dual-earner models have comparable levels of income support for separated women, as van Damme and Uunk demonstrate using different policy indicators (see Table 9.1 in van Damme and Uunk in this volume), yet these countries differ substantially in labour market outcomes for separated women (and mothers in particular). The authors attribute these differences to varying degrees of employment support for women. For instance, women in the dual-earner model enter employment after separation more often than women in other regime types. Furthermore, the results for women in the family model indicate that support from the extended family seems to be a substitute for employmentsupportive policies. Residential mobility and housing Another important consequence of divorce is the residential mobility that accompanies and follows it. However, as Dewilde notes, it is much more difficult to link these processes to the institutional context. One obvious hypothesis is the assumption that the structure of the housing market provides different opportunities for residential mobility on the part of divorcees. But it is less clear how to typify national housing policies, and – if this were possible – to explain how they relate to our typology of family support. Data on public expenditures for housing indicate that this policy field is of lesser importance than family policies, at least in the countries represented in our study.5 Housing policy is therefore seldom incorporated into comparative welfare state research.

380

Conclusion

But Dewilde’s comparative analyses of tenure status reveal some marked differences for the working-age population between our types of countries. For instance, a social rental sector with possibly cheap housing exists to a considerable extent only in those countries belonging to the market, male breadwinner and dual-earner models. In these countries, 11–15 per cent of the working-age population live in council housing. Countries belonging to the family model, on the other hand, are characterized by high percentages (60 per cent and more) of outright owners. Dewilde notes that housing policies in the family model countries mainly encourage owner-occupation, with only a small social housing segment. Furthermore, rigid systems of rent control have resulted in a gradual shrinkage of the private rental sector. Thus, for many households, owner-occupation is the only available tenure. In addition, it is often an important asset of the extended family. As we mentioned in the first section, separation is associated with a great deal of downward mobility on the housing market. For example, it substantially raises the risk of leaving owner-occupation for both men and women. Interestingly, this risk seems smaller for women in those countries where the economic consequences of divorce are more severe – that is, in the market and family models (see ‘Income and poverty’ subsection). Dewilde suspects that women in these countries might be compensated for the economic consequences of divorce by advantageous procedures of property settlement. Whether this different behaviour is a result of the institutional context is an open question. On the one hand, one could argue that a lack of public provisions in the family and the market models, where family policies are rather rudimentary, make it necessary to develop private solutions. In family model countries, this overall tendency may be further intensified by strong family ties. On the other hand, alternative explanations are possible as well. As we mentioned in the previous paragraph, outright ownership is more common in the family model, which makes it easier to stay in the marital home – sometimes even necessarily, because alternative forms of housing are limited. Lifestyle deprivation Finally, Aassve et al. have analysed the lifestyles of men and women after separation. More specifically, they look at the inability to afford the basic requirements of life (for example, keeping the home warm, replacing worn-out furniture, paying regular bills, and so on) and the enforced lack of widely desired possessions (for example, a car, a telephone, a colour television, and so on). Here it is less obvious how the institutional context modifies the effect of separation on the lifestyles of men and women. On the one hand, lifestyle and consumption decisions are indirectly influenced by available economic resources, and thus we should observe similar effects

When marriage ends: results and conclusions

381

to those found in the analysis of income and poverty. On the other hand, the analysis of lifestyle deprivation often focuses on goods and activities that are consumed in the household context (see the items in the appendix to Aassve et al., in this volume). This makes it quite probable that in cases of need, the extended family will help out. The latter argument suggests that separation and divorce have a lower negative effect on individuals’ lifestyles when a strong family network exists – that is, in countries of the family model. The results presented by Aassve et al. partially support the latter explanation. The weakest separation effect with respect to basic lifestyle deprivation is found in the family model, and it is particularly low for men. Moreover, for men in the family model, it does not matter much whether children are present or not. Whereas women in the family model suffer considerably more than their male counterparts, they are still less badly off compared with women in the other institutional contexts. The strongest separation effect is observed in the market and the dual-earner models. This is also in line with the family network argument, because according to the data presented in van Damme and Uunk (see their Table 9.1), countries from both the market and the dual-earner models score particularly low on an index of family support (Vogel, 2003). On the other hand, it is quite surprising that men and women in the dual-earner model experience such negative changes in their lifestyles, because they were the ones with the smallest income decreases in our comparative analyses (see ‘Income and poverty’ subsection). One possible explanation is that life is much more expensive in the dual-earner model, not because prices are higher, but simply because the dual-earner family has become the new income norm, such that living on a single income is only feasible by reducing one’s standard of living. These different explanations definitely need further investigation. But they also cast some shadows on the rosy picture of the dual-earner model that emanates from the other dimensions of divorce consequences.

CONCLUSION In our view, explaining the social and economic consequences of partnership dissolution in terms of a typology of countries based on family support has been fruitful. But typological approaches in comparative research are not without problems, as Stier (in this volume) notes in her methodological chapter. Our typology, like many others, assumes that there are more or less coherent clusters of institutional arrangements that embody distinct values and rationales. This implies that it is often not

382

Conclusion

possible to disentangle which specific policy measure (or which combination of measures) is responsible for a specific outcome. For instance, is it the significant amount of child benefits or the generous welfare payments for single mothers that ease economic life for divorced mothers in some male breadwinner countries? Moreover, with a simple categorical variable like our typology it is impossible to estimate the precise magnitude of the institutional effects. Finally, typologies suffer from classification problems: Assigning countries to clusters is difficult (see the hybrid countries in the second section) and may influence the estimates of the institutional effects (see the sensitivity analyses in van Damme and Uunk, in this volume). Classifying countries into four clusters also ignores within-cluster variation, and certainly, institutions and policies change over time (Great Britain and Germany are prominent examples; see Sigle-Rushton and Hummelsheim, both in this volume). In principle, questions about the precise mechanisms and magnitude of institutional effects can be better answered in analyses with quantifiable institutional indicators (so-called macro–micro analyses; see Stier, in this volume). But this is a difficult task in our case, because macro indicators characterizing the institutional support of families are scarce and often available only for certain countries and time periods. Further research is needed to provide more precise and comprehensive measures of the relevant institutional parameters that influence the well-being of families and children. This will increase our possibilities for analysing the consequences of partnership dissolution in greater detail. From a substantive point of view, the chapters in this book make clear that the main economic and social consequences of divorce are pretty much the same across Western Europe, although separation and divorce occur in equally diverse institutional, economic and socio-demographic settings. Women are the economic losers, but men are not better off when we look at other divorce outcomes such as housing, consumption, lifestyle and social contacts. The presence of children and labour force participation are key variables in explaining the gender differences and the magnitude of the divorce consequences. The extended family is an important resource of material and social support. Furthermore, public policies are able to cushion the negative effects of children by compensating for the additional economic needs (cash transfers) or by enabling both parents to work (services: childcare). But in all countries, the economic consequences are predominantly negative, even in the generous Scandinavian welfare states with high female employment. Hence, whatever the institutional context is like, partnership dissolution is an important new social risk, and people have to find their own ways and strategies to deal with the changes that come along with it.

When marriage ends: results and conclusions

383

NOTES 1. See also Aassve et al. (in this volume), who report no significant divorce effect with respect to housing facilities, housing deterioration or environmental problems. 2. Almost 50 per cent of all Belgian children visited a public childcare facility in the 1920s (Fix, 2001: 120). 3. Tax benefits for families make up 1 per cent of the GDP in Germany, the highest percentage among our 14 countries. With tax benefits excluded, Germany would fall back into the third group of countries (Greece, Italy, the Netherlands, Portugal, and Spain), with comparatively low public expenditures for families. 4. Statistics from the OECD Family Database refer to the United Kingdom as a whole, that is, Great Britain and Northern Ireland. Sigle-Rushton’s country report mostly refers to Great Britain, that is, the contiguous regions of England, Wales and Scotland. We use the terms ‘United Kingdom’ and ‘Great Britain’ accordingly. 5. In all our countries (except the United Kingdom), public expenditures for housing make up less than 1 per cent of the GDP (1.4 per cent in the UK; OECD, 2008).

REFERENCES Adema, W. and M. Ladaique (2005), ‘Net social expenditure, 2005 edition’, Social, Employment and Migration Working Papers, No. 32, OECD, Paris. Esping-Andersen, G. (1990), The Three Worlds of Welfare Capitalism, Cambridge: Polity. Fix, B. (2001), Religion und Familienpolitik. Deutschland, Belgien, Österreich und die Niederlande im Vergleich, Wiesbaden: Westdeutscher Verlag. Hummelsheim, D. (2009), Die Erwerbsbeteiligung von Müttern – Institutionelle Steuerung oder kulturelle Prägung? Eine empirische Untersuchung am Beispiel von Belgien, West- und Ostdeutschland, Wiesbaden: VS Verlag für Sozialwissenschaften. Organisation for Economic Co-operation and Development (OECD) (2008), ‘Social expenditure data base’, http://stats.oecd.org/wbos/Index. aspx?usercontext5sourceoecd, accessed 6 June. Vogel, J. (2003), ‘The family’, Social Indicators Research: special issue, 64 (3), 393–435.

PART V

Data Appendix

387

Further reading

Time period Panel design Number of waves Target population

Institution

Name

Table A1

Institute for Social and Economic Research 1991–2001 Annual 11 All household members at time of interview

All household members at time of interview

1990–99 Annual 10

Statistics Denmark

All household members at time of interview SOEP Group Taylor et al. (2001) (2004)

British Household Panel Survey (BHPS)

German SocioEconomic Panel Study (GSOEP) German Institute for Economic Research 1984–99 Annual 16

Administrative panel data

European Community Household Panel (ECHP)

Household Market and Nonmarket Activities (HUS) University of Gothenburg

Eurostat

European Union

Sweden

Banca d’Italia (2000)

All household members present on 31 December of interview year

Household head, partner, and max. one adult household member

All household members at time of interview in 15 European Union member states Flood and Wirtz and Olovsson (1998) Meyer (2002)

1987–2002 1984–98 1994–2001 Every 2–3 years Every 2–3 years Annual 7 7 8

The Bank of Italy Survey of Households’ Income and Wealth (SHIW) Banca d’Italia

Great Britain Italy

Germany

Denmark

Representative sample of 1% of the Danish population aged 30–48 years in 1990 Casman and Statistics Lenoir (2004) Denmark (2008)

1992–2002 Annual 11

Panel Studie van Belgische Huishoudens (PSBH) Universities of Antwerp and Lüttich

Belgium

Main features of national and cross-national surveys used in this volume

Data appendix

388

9.2 35.7 22.9 28.4 3.9 2.0 6.5 1.62 11.61

68.4 41.0 64.4 2.0 25.8 4.1 10.6

69.1 37.2

1.8 21.1 6.5 16.7

1990

7.0 (1980) 40.8 (1980) 24.8 (1980) 23.2 (1980) 4.2 (1980) 1.9 5.8 1.51 7.07

1985

Selected statistical indicators for the single country studies

Family structure Composition of households (% of the population) Single parents Couples with children Couples without children One person households Other households Crude divorce rate Crude marriage rate Total fertility rate Live births outside marriage (% of all live births) Labour market Employment rate (persons aged 15–64 years) (%) Men Women Women (with children less than 6 years) Part-time employment (% of total employment) Men Women Unemployment rate (%) Men Women

Belgium

Table A2

7.6 12.7

3.0 30.5

66.9 45.0

3.5 5.1 1.55 17.28

1995

5.6 8.5

5.5 37.4

69.5 51.5 69.5 (1999)

11.1 30.2 22.2 31.2 5.3 2.6 4.4 1.66 26.0

2000

7.6 9.5

7.6 40.5

68.3 53.8 68.8 (2002)

3.0 (2004) 4.10 (2004) 1.64 (2004) 31.0 (2003)

2005

389

Women’s earnings as share of men’s earnings in % (wage earners in manufacturing) Poverty rate Single parent with dependent children Two adults with 1 child Two adults with 3 or more children Two adults younger than 65 years Welfare state Public expenditures for families (% of GDP) Family cash benefits (% of GDP) Child benefit as share of monthly income (%) Advanced child maintenance Maternity leave (in weeks) Parental leave (in weeks) – without career break Children in publicly funded childcare (%) 0–3 years 3–6 years No 14 –

No 14 – 30 95

2.3 2.1

74.5

2.6 2.5

74.3

30 97

2.6 1.8 8.7 (1999) No 15 12

27 7 9 10

34 10 17 12 2.3 2.1 9.5 No 14 –

79.2 (1998)

79.1

2.7 (2003) 1.7 (2003) 8.4 (2004) No 15 12

36 (2004) 10 (2004) 18 (2004) 11 (2004)

390

(continued)

Family structure Composition of households (% of the population) Single parents Couples with children Couples without children One person households Other households Crude divorce rate Crude marriage rate Total fertility rate Live births outside marriage (% of all live births) Labour market Employment rate (persons aged 15–64 years) (%) Men Women Women (with children less than 6 years) Part-time employment (% of total employment) Men Women Unemployment rate (%) Men Women

Denmark

Table A2

4.1 19.7 26.5 34.0 15.6 2.7 6.13 1.67 46.60

83.5 71.6

10.4 38.4 7.0 8.4

85.4 69.3

8.4 43.9 5.8 8.6

1990

3.5 (1980) 25.9 (1980) 25.8 (1980) 29.1 (1980) 15.6 (1980) 2.8 5.73 1.45 42.96

1985

5.6 8.1

10.8 35.4

79.9 66.7

4.1 18.3 26.8 35.9 14.9 2.5 6.64 1.80 46.47

1995

3.9 4.8

10.2 34.1

80.8 71.6

4.0 18.4 27.5 37.0 13.1 2.7 7.19 1.77 44.57

2000

4.4 5.3

12.7 33.0

79.8 71.9 74.3 (2002)

2.9 (2004) 7.00 (2004) 1.78 (2004) 45.4 (2004)

2005

391

Women’s earnings as share of men’s earnings in % (wage earners in manufacturing) Poverty rate Single parent with dependent children Two adults with 1 child Two adults with 3 or more children Two adults younger than 65 years Welfare state Public expenditures for families (% of GDP) Family cash benefits (% of GDP) Child benefit as share of monthly income (%) Advanced child maintenance Maternity leave (in weeks) Parental leave (in weeks)a Children in publicly funded childcare (%) 0–3 years 3–6 years Yes 18 10 1 2

Yes 18 10 1 2 48 (1993/94) 82 (1993/94)

3.3 1.4

84.6

2.6 0.9

85.6

3.8 1.8 7.0 Yes 18 10 1 2

84.6

64 (1998/2000) 91 (1998/2000)

3.9 (2003) 1.6 (2003) 7.0 (2004) Yes 18 32 1 2

16 (2004) 4 (2004) 14 (2004) 5 (2004)

12 (2001) 3 (2001) 13 (2001) 4 (2001) 3.7 1.5 7.1 (1999) Yes 18 12 1 4

86.9 (2003)

84.4

392

(continued)

Family structure Composition of households (% of the population) Single parents Couples with children Couples without children One person households Other households Crude divorce rate Crude marriage rate Total fertility rate Live births outside marriage (% of all live births) Labour market Employment rate (persons aged 15–64 years) (%) Men Women Women (with children less than 6 years) Part-time employment (% of total employment) Men Women Unemployment rate (%) Men Women

Finland

Table A2

79.1 72.4 64.3 5.8 13.5 3.7 2.9

6.2 17.2 6.1 6.0

2.6 5.01 1.78 25.24

1990

78.2 72.3

7.9 32.0 22.1 31.7 6.7 1.8 5.25 1.65 16.39

1985

15.7 15.1

8.2 15.4

64.2 59.0

7.8 28.4 22.5 35.2 6.1 2.7 4.65 1.81 33.12

1995

9.1 10.6

8.0 17.0

70.1 64.2 58.8 (1999)

7.6 25.4 24.4 37.7 5.3 2.7 5.05 1.73 39.21

2000

8.2 8.6

9.2 18.6

70.3 66.5 49.4 (2002)

2.5 (2004) 5.60 (2004) 1.80 (2004) 40.80 (2004)

2005

393

Women’s earnings as share of men’s earnings in % (wage earners in manufacturing) Poverty rate Single parent with dependent children Two adults with 1 child Two adults with 3 or more children Two adults younger than 65 years Welfare state Public expenditures for families (% of GDP) Family cash benefits (% of GDP) Child benefit as share of monthly income (%) Advanced child maintenance Maternity leave (in weeks) Parental leave (in weeks) Children in publicly funded childcare (%) 0–3 years 3–6 years Yes 17.5 26

Yes 12 18 32 (1993/94) 59 (1993/94)

3.2 1.9

77.3

2.6 1.5

76.8

4.0 2.6 10.8 Yes 17.5 26

8 (1996) 3 (1996) 4 (1996) 5 (1996)

79.4

22 (1998/2000) 66 (1998/2000)

3.1 1.7 10.0 Yes 17.5 26

22 5 5 8

80.6

3.0 (2003) 1.6 (2003) 8.7 (2004) Yes 17.5 26

16 (2004) 5 (2004) 12 (2004) 7 (2004)

83.9 (2004)

394

(continued)

Family structure Composition of households (% of the population) Single parents Couples with children Couples without children One person households Other households Crude divorce rate Crude marriage rate Total fertility rate Live births outside marriage (% of all live births) Labour market Employment rate (persons aged 15–64 years) (%) Men Women Women (with children less than 6 years) Part-time employment (% of total employment) Men Women Unemployment rate (%) Men Women

Germanyb

Table A2

78.1 53.6 41.1 2.6 33.8 4.0 5.9

2.0 29.6 6.2 8.7

6.0 31.5 25.7 35.0 4.6 1.9 6.50 1.45 15.32

1990

78.1 48.5

30.2 (1980) 3.0 (1980) 2.3 6.39 1.28 16.22

6.3 (1980)

1985

5.8 10.9

3.6 33.7

73.7 55.3

6.8 30.0 27.5 34.9 5.1 2.1 5.27 1.25 16.06

1995

6.0 8.7

5.0 37.9

72.9 58.1 51.1 (1999)

6.0 28.0 28.8 36.0 1.2 2.4 5.09 1.38 23.41

2000

8.9 10.3

7.8 43.8

71.2 59.6 57.1 (2002)

2.6 (2004) 4.8 (2004) 1.36 (2004) 28.0 (2004)

2005

395

Women’s earnings as share of men’s earnings in % (wage earners in manufacturing) Poverty rate Single parent with dependent children Two adults with 1 child Two adults with 3 or more children Two adults younger than 65 years Welfare state Public expenditures for families (% of GDP) Family cash benefits (% of GDP) Child benefit as share of monthly income (%) Advanced child maintenance Maternity leave (in weeks) Parental leave (in weeks) Children in publicly funded childcare (%) 0–3 years 3–6 years Yes 14 48

Yes 14 24 2 (1993/94) 85 (1993/94)

1.7 1.2

72.7

1.8 1.3

72.7

10 (1998/2000) 78 (1998/2000)

1.9 1.2 9.9 (1999) Yes 14 144

44 7 17 9

55 11 23 10 1.9 1.2 4.3 Yes 14 144

73.5

73.7

1.9 (2003) 1.2 (2003) 10.8 (2004) Yes 14 144

38 (2004) 14 (2004) 24 (2004) 8 (2004)

74.0 (2003)

396

(continued)

Family structure Composition of households (% of the population) Single parents Couples with children Couples without children One person households Other households Crude divorce rate Crude marriage rate Total fertility rate Live births outside marriage (% of all live births) Labour market Employment rate (persons aged 15–64 years) (%) Men Women Women (with children less than 6 years) Part-time employment (% of total employment) Men Women Unemployment rate (%) Men Women

Greece

Table A2

75.9 38.5 42.9 2.2 7.6 3.9 10.8

2.8 10.0 5.0 10.6

6.0 49.1 23.8 14.6 6.5 0.6 5.81 1.39 2.17

1990

79.0 37.4

0.8 6.41 1.67 1.80

1985

6.2 14.1

2.7 8.4

72.5 38.1

1.0 6.02 1.32 3.04

1995

7.5 17.2

2.6 7.8

71.5 41.7 48.6 (1999)

7.3 41.7 22.5 20.5 8.0 1.0 4.48 1.29 4.02

2000

6.1 15.3

2.3 9.3

74.2 46.1 49.5 (2002)

7.3 37.9 23.5 25.0 6.3 1.1 (2004) 4.20 (2004) 1.29 (2004) 4.90 (2004)

2005

397

Women’s earnings as share of men’s earnings in % (wage earners in manufacturing) Poverty rate Single parent with dependent children Two adults with 1 child Two adults with 3 or more children Two adults younger than 65 years Welfare state Public expenditures for families (% of GDP) Family cash benefits (% of GDP) Child benefit as share of monthly income (%) Advanced child maintenance Maternity leave (in weeks) (private/public sector) Parental leave (in weeks) (private/public sector) Children in publicly funded childcare (%) 0–3 years 3–6 years No 16/18 12/12

No 15/18 12/— 3 (1993/94) 70 (1993/94)

0.8 0.4

78.4

0.4 0.3

77.6

3 (1998/2000) 70 (1998/2000)

1.1 0.7 2.5 (1999) No 17/20 14/14

23 9 26 20

32 11 17 17 1.2 0.9 2.5 No 17/18 12/12

82.1 (1998)

80.4

1.3 (2003) 0.9 (2003) 2.5 (2004) No 17/20 14/14

38 (2004) 15 (2004) 32 (2004) 14 (2004)

398

(continued)

Family structure Composition of households (% of the population) Single parents Couples with children Couples without children One person households Other households Crude divorce rate Crude marriage rate Total fertility rate Live births outside marriage (% of all live births) Labour market Employment rate (persons aged 15–64 years) (%) Men Women Women (with children less than 6 years) Part-time employment (% of total employment) Men Women Unemployment rate (%) Men Women

Spain

Table A2

69.0 31.1 36.1 1.6 12.1 11.9 24.1

2.4 13.9 20.2 25.1

0.8 19.1 16.5 9.9 53.7 0.6 5.68 1.36 9.61

1990

63.8 25.2

0.5 5.20 1.64 7.97

1985

14.8 24.6

2.9 16.4

62.5 31.7

0.7 17.4 16.9 10.9 54.1 0.8 5.10 1.18 11.09

1995

7.9 16.0

2.8 16.8

71.2 41.3 41.8 (1999)

7.0 51.6 17.3 11.3 12.8 1.0 5.39 1.24 17.74

2000

7.0 12.2

4.5 24.2

75.2 51.2 51.0 (2002)

2.1 5.00 1.32 23.2

2005

399

Women’s earnings as share of men’s earnings in % (wage earners in manufacturing) Poverty rate Single parent with dependent children Two adults with 1 child Two adults with 3 or more children Two adults younger than 65 years Welfare state Public expenditures for families (% of GDP) Family cash benefits (% of GDP) Child benefitc as share of monthly income (%) Advanced child maintenanced Maternity leave (in weeks) Parental leave (in weeks)e Children in publicly funded childcare (%) 0–3 years 3–6 years No 16 No

No 14 No 2 (1993/94) 84 (1993/94)

0.3 0.2

0.3 0.2

5 (1998/2000) 84 (1998/2000)

0.9 0.3 3.0 (1999) No 16 No

57 16 35 15

37 14 31 13 0.4 0.3 3.5 No 16 No

15

13

1.0 (2003) 0.4 (2003) 3.2 (2004) No 16 No

40 (2004) 14 (2004) 39 (2004) 12 (2004)

15 (2004)

400

(continued)

Family structure Composition of households (% of the population) Single parents Couples with children Couples without children One person households Other households Crude divorce rate Crude marriage rate Total fertility rate Live births outside marriage (% of all live births) Labour market Employment rate (persons aged 15–64 years) (%) Men Women Women (with children less than 6 years) Part-time employment (% of total employment) Men Women Unemployment rate (%) Men Women

Sweden

Table A2

4 22 30 40 5 2.3 4.73 2.13 47.00

84.9 80.7 85.0 7.4 41.8 1.7 1.7

83.4 77.1

6.8 46.6 3.0 2.8

1990

4 24 31 36 5 2.4 4.59 1.74 46.35

1985

9.7 7.8

7.3 35.8

73.1 68.8

5.6 21.2 24.1 45.4 3.9 2.6 3.81 1.73 52.96

1995

5.9 5.3

8.2 32.3

75.1 70.9 77.8 (1999)

6.2 20.4 23.3 46.5 3.7 2.4 4.50 1.54 55.33

2000

7.9 7.7

11.5 39.6

74.4 70.4 77.5 (2002)

5.7 20.1 24.2 46.4 3.6 2.2 4.80 1.75 55.40

2005

401

Women’s earnings as share of men’s earnings in % (wage earners in manufacturing) Poverty rate Single parent with dependent children Two adults with 1 child Two adults with 3 or more children Two adults younger than 65 years Welfare state Public expenditures for families (% of GDP) Family cash benefits (% of GDP) Child benefit as share of monthly income (%) Advanced child maintenance Maternity leave (in weeks) Parental leave (in weeks) Children in publicly funded childcare (%) 0–3 years 3–6 years Yes 8.5 56

Yes 8.5 56 33 (1993/94) 72 (1993/94)

4.5 2.1

88.9

4.1 1.8

89.8

3.8 2.0 9.5 Yes 8.5 56

90.0

48 (1998/2000) 80 (1998/2000)

3.5 (2003) 1.6 (2003) 9.1 (2004) Yes 8.5 60

19 (2004) 8 (2004) 14 (2004) 6 (2004)

13 (2001) 5 (2001) 8 (2001) 4 (2001) 3.3 1.5 8.2 (1999) Yes 8.5 56

91.0 (2004)

91.3

402

(continued)

Family structure Composition of households (% of the population) Single parents Couples with children Couples without children One person households Other households Crude divorce rate Crude marriage rate Total fertility rate Live births outside marriage (% of all live births) Labour market Employment rate (persons aged 15–64 years) (%) Men Women Women (with children less than 6 years) Part-time employment (% of total employment) Men Women Unemployment rate (%) Men Women Women’s earnings as share of men’s earnings in % (wage earners in manufacturing)

United Kingdom

Table A2

10.0 33.0 28.0 27.0 2.0 2.9 6.52 1.83 27.90

82.1 62.6 42.5 5.3 43.2 7.4 6.6 68.0

77.3 55.0

4.4 44.8 11.8 11.0 66.5

1990

9.0 (1980) 39.0 (1980) 26.0 (1980) 22.0 (1980) 4.0 (1980) 3.1 6.94 1.79 18.89

1985

9.9 6.8 70.8

7.8 44.4

75.1 61.7

2.9 5.50 1.71 33.57

1995

5.8 4.8 75.6

8.9 44.3

77.8 64.7 55.8 (1999)

9.0 29.0 29.0 29.0 4.0 2.6 5.12 1.64 33.48

2000

5.1 4.3 79.5 (2003)

10.4 42.7

77.6 65.9 57.0 (2002)

2.8 (2004) 5.1 (2004) 1.74 (2004) 42.30 (2004)

2005

403

No 0/29f 0

No 0/29f 0 3 (1993/94) 70 (1993/94)

1.9 1.5

2.3 1.8

2.4 1.9 6.4 No 14/28g 0

60 11 36 9

34k (1999) 60k (1999)

2.6 1.9 7.1 (1999) No 18/29h 13j

57 12 32 7 2.9 (2003) 2.2 (2003) 6.9 (2004) No 26/52i 13j

40 (2003) 13 (2003) 27 (2003) 11 (2003)

a

Notes: Parental leave 1 parental leave for fathers. b Data for 1985 and 1990 refer exclusively to West Germany. Subsequent years refer to West and East Germany. c 1995–1999: child benefit taxable. d Beginning in 2008 a scheme of advanced child maintenance payments has been set up in Spain. It is a means-tested measure, intended for lowincome families, with quite restrictive conditions of eligibility and monthly payments up to €100 per child. e Subsequent to maternity leave there is no paid parental leave, but an unpaid parental leave, until the child’s third birthday. f Mothers with 2 years (16 and more hours per week) or 5 years (8–16 hours per week) continuous employment with the same employer into the eleventh week before expected week of birth could return to work at any time within 29 weeks of confinement. g All pregnant women were entitled to 14 weeks of job protected leave. Mothers with 2 years of continuous service were entitled to 28 weeks of leave. h All pregnant women are entitled to 18 weeks of job protected leave. Mothers with 1 year of continuous service were entitled to 29 weeks of leave. i All pregnant women are entitled to 26 weeks of job protected leave. Mothers with 1 year of continuous service are entitled to 52 weeks of leave. j Each parent has an individual entitlement of 13 weeks before the child’s fifth birthday, but an individual parent cannot take more than 4 weeks in any given year. k Data for England only.

Poverty rate Single parent with dependent children Two adults with 1 child Two adults with 3 or more children Two adults younger than 65 years Welfare state Public expenditures for families (% of GDP) Family cash benefits (% of GDP) Child benefit as share of monthly income (%) Advanced child maintenance Maternity leave (in weeks) Parental leave (in weeks) Children in publicly funded childcare (%) 0–3 years 3–6 years

404

Data appendix

DATA SOURCES AND DEFINITIONS OF STATISTICAL INDICATORS IN TABLE A2 Advanced child maintenance Additional benefit available to single parents in the case the liable parent fails to make his/her maintenance payments. Child benefit as share of monthly income in percentage Cash benefit for a couple household with two dependent children (2 and 7 years of age) as share of an average production worker wage (APW). Source: own calculations based on OECD, 2004, 1999, 1995. Children in publicly funded childcare (percentage) Sources: OECD, 2001: 144; Daly, 2000; Gornick et al., 1997. Composition of households (percentage of the population) Sources: UN Statistics Division Database; Greece: Eurostat; Sweden: Statistics Sweden. Crude divorce rate (per 1000 inhabitants) Ratio of the number of divorces during the year and the average population in that year. Source: Eurostat. Crude marriage rate (per 1000 inhabitants) Ratio of the number of marriages during the year and the average population in that year. Source: Eurostat. Employment rate (proportion employed as percentage of the population in the corresponding age bracket) Sources: 1985–1990: European Commission (1998); 1995–2005: European Commission (2006). Family cash benefits (percentage of GDP) Family cash benefits include expenditure which supports families (that is, excluding one-person households). This expenditure is often related to the costs associated with raising children or with the support of other dependants. Source: OECD.stat Database, 2008. Live births outside marriage (share of all live births in percentage) Source: Eurostat. Maternity leave for the mother (in weeks) Sources: MISSOC, 2001; Leira, 2002; OECD, 2001. Parental leave (in weeks) Only those parts of the parental leave that can be taken either by the mother or the father. Sources: Moss and Deven, 1999; OECD, 2001; The Clearinghouse on International Child, Youth and Family Policies at Columbia University; Leira, 2002. Part-time employment (percentage of total employment) Sources: 1985–1990: European Commission (1998); 1995–2005: European Commission (2006). Poverty rate by household type Percentage of persons with an equivalized

Data appendix

405

disposable household income below the risk-of-poverty threshold, which is set at 60 per cent of the national median equivalized disposable household income (after social transfers). Source: Eurostat. Public expenditures for families (percentage of GDP) Social expenditure is the provision by public (and private) institutions of benefits to, and financial contributions targeted at, households and individuals in order to provide support during circumstances which adversely affect their welfare, provided that the provision of the benefits and financial contributions constitutes neither a direct payment for a particular good or service nor an individual contract or transfer. Such benefits can be cash transfers, or can be the direct (‘in-kind’) provision of goods and services. Source: OECD.stat Database, 2008. Total fertility rate (mean number of children) The mean number of children that would be born alive to a woman during her lifetime if she were to pass through her childbearing years conforming to the fertility rates by age of the given year. Source: Eurostat. Unemployment rate (proportion unemployed as percentage of the labour force [employed and unemployed] in the corresponding age bracket) Sources: 1985–1990: European Commission (1998); 1995–2005: European Commission (2006). Women’s earnings as share of men’s earnings in percentage (wage earners in manufacturing) Spain: difference between men’s and women’s average gross hourly earnings as percentage of men’s average gross hourly earnings (for paid employees). Source: ILO (own calculations); Eurostat (2006).

REFERENCES Banca d’Italia (2000), Supplements to the Statistical Bulletin – Methodological Notes and Statistical Information, Rome: Banca d’Italia. Casman, M.T. and V. Lenoir (2004), ‘Voorstelling van de PSBH’, in D. Mortelmans, M.T. Casman and R. Doutrelepont (eds), Elf jaar uit het leven in België. Socioeconomische analyses op het Gezinsdemografisch Panel PSBH, Gent: Academia Press, pp. 21–43. Daly, M. (2000), The Gender Division of Welfare, Cambridge: Cambridge University Press. European Commission (1998), Employment & Social Affairs, Luxembourg: Office for Official Publications of the European Communities, http://ec.europa.eu/ employment_social/employment_analysis/eie/1998_en.pdf, accessed April 2008. European Commission (2006), Employment in Europe 2006, Luxembourg: Office

406

Data appendix

for Official Publications of the European Communities, http://ec.europa.eu/ employment_social/employment_analysis/eie/eie2006_annex_en.pdf, accessed April 2008. Eurostat (website), http://epp.eurostat.ec.europa.eu/portal/page?_pageid51090, 30070682,1090_30298591&_dad5portal&_schema5PORTAL, accessed June 2008. Eurostat (2006), http://ec.europa.eu/employment_social/employment_strategy/ pdf/compendiumjuly2006_en.pdf, accessed April 2008. Flood, L. and P. Olovsson (1998), Household Market and Nonmarket Activities (HUS): Survey Description, Vol. VIII, Gothenburg: Department of Economics, University of Gothenburg. Gornick, J.C., M.K. Meyers and K.E. Ross (1997), ‘Supporting the employment of mothers: policy variation across fourteen welfare states’, Journal of European Social Policy, 7, 45–70. International Labour Organization (ILO) (website), http://laborsta.ilo.org/ (B8 wages in manufacturing), accessed June 2008. Leira, A. (2002), Working Parents and the Welfare State. Family Change and Policy Reform in Scandinavia, Cambridge: Cambridge University Press. MISSOC (website), http://ec.europa.eu/employment_social/missoc/2001/index_ en.htm, accessed June 2008. Moss, P. and F. Deven (eds) (1999), Parental Leave: Progress or Pitfall? Brussels: NIDI/CBGS Publications. OECD.stat Database (2008) (website), http://stats.oecd.org/wbos/Index. aspx?usercontext5sourceoecd, accessed June 2008. Organisation for Economic Co-operation and Development (OECD) (2001), ‘Employment outlook’, www.oecd.org/document/37/0,3343,en_2649_33927_31 736485_1_1_1_1,00.html#EMO01, accessed June 2008. Organisation for Economic Co-operation and Development (OECD) (2004, 1999, 1995), ‘Benefits and wages’, www.oecd.org/document/0/0,2340,en_2649_201185 _34053248_1_1_1_1,00.html#country, accessed June 2008. SOEP Group (2001), ‘The German Socio-Economic Panel (GSOEP) after more than 15 years – overview’, in E. Holst, D.R. Lillard and T.A. DiPrete (eds), Proceedings of the 2000 Fourth International Conference of German Socio-Economic Panel Study Users (GSOEP2000), Vierteljahreshefte zur Wirtschaftsforschung, 70 (1), 7–14. Statistics Denmark (2008), www.dst.dk/asp2xml/puk/udgivelser/get_file. asp?id512752&sid5entire, accessed June 2008. Statistics Sweden (website), www.scb.se/default____2154.asp, accessed June 2008. Taylor, M.F. (ed.), with Brice, J., Buck, N. and Prentice-Lane, E. (2004), British Household Panel Survey User Manual Volume A: Introduction, Technical Report and Appendices, Colchester: University of Essex. The Clearinghouse on International Child, Youth and Family Policies at Columbia University (website), www.childpolicyintl.org/, accessed June 2008. UN Statistics Division Database (website), http://unstats.un.org/unsd/default.htm, accessed April 2008. Wirtz, C. and L. Meyer (2002), ‘The European Community Household Panel (ECHP)’, Schmollers Jahrbuch. Zeitschrift für Wirtschafts- und Sozialwissenschaften, 122 (1), 143–54.

Index Aassve, A. 336 Abrahamsson, P. 100 advanced child maintenance payment 62, 189, 400 Alberdi, C. 187 Alberdi, I. 186, 190, 194 Aliaga, C. 207 alimony, see spousal maintenance payments Allison, P.D. 245 Amato, P.R. 2, 5, 17, 194, 263 Andreß, H.-J. 2, 3, 4, 5, 6, 13, 15, 18, 22, 59, 66, 69, 70, 71, 287, 290, 299, 319 Andries, M. 88, 291 Anttonen, A. 141 Arber, S. 29 Arts, W. 13 Ashenfelter, O. 10 ATET (average treatment effect on the treated) 340 Atkinson, A.B. 94, 331 Augurzky, B. 344 Austria housing 270 male breadwinner model 238, 338 post-divorce labour supply 256 residential mobility 275, 276 average treatment effect on the treated (ATET) 340 Bahle, T. 132 Bank of Italy Survey of Households’ Income and Wealth 298 Barbagli, M. 187 Bardasi, E. 320 Barlow, J. 268 Batty, D. 164 Beck, N. 9 Becker, G.S. 8, 64, 211, 235, 332, 340 Becker, S. 342, 344 Behr, A. 243, 271

Belgian Household Panel (PSBH) 90 Belgium 78−9, 93−4 data appendix 388−9 demographic changes 79–81 divorce risks and consequences 89−92 dual-earner model 241 economic autonomy of women 296 family law 83−7 family support model 290−91 housing 270 labour market 81−3 male breadwinner model 87−8, 238, 338−9, 368−9 residential mobility 275, 276−7, 282 welfare state arrangement 87−9 see also income changes after separation, comparative analysis benefits, see child benefits; family benefits Bentolila, S. 185 Bernhardt, E. 137 Betti, G. 334, 335, 336 BHPS (British Household Panel Survey) 271 Bianchi, S.M. 89 Birkelund, G.E. 237 Björnberg, U. 134, 178 Blaiklock, O. 164 Blakemore, K. 37 Blau, F.D. 40, 235 Blossfeld, H.-P. 29, 31, 214, 218 Blundell, R. 340, 344 Böttcher, K. 52 Bouman, A.M. 6, 233, 234, 330 Bourguignon, F. 331 Bradbury, K. 3, 233, 234 Bradshaw, J. 132, 163 Brewster, K.L. 37 Brinton, M.C. 32 407

408 British Household Panel Survey (BHPS) 271 Bröckel, M. 2, 3, 4, 5, 6, 59, 71 Broese van Groenou, M.I. 89 Bryant, W.K. 7 Büchel, F. 291 Budig, M.J. 29 Budzt, H. 103 Buhmann, B. 302 Bumpass, L.L. 212 Burkhauser, R.V. 2, 66, 212, 263, 287, 288, 330 Bynner, J. 157 Caliendo, M. 342 Cantillon, B. 86, 88 case-study approach data selection 18−36 multiple country 32 single countries 30−32 Casman, M.T. 327 Casper, L.M. 38 Castles, F.G. 133, 270 Cerioli, A. 334, 336 Chakravarty, S.R. 331 Charles, M. 29 Cheli, B. 331, 334 Cherlin, A.J. 22, 211, 212 child allowance Belgium 290 Germany 60−62 child benefits Belgium 88 Finland and Sweden 139 Great Britain 166−7 Spain 182 see also family benefits child custody 4−5 Belgium 86 Germany 58−9 Greece 217 Spain 188−9 Sweden 137 see also men, with custody child gap 190 Child Home Care Allowance, Finland and Sweden 142−3 child maintenance payments Denmark 106−7 Finland and Sweden 137−8

Index Germany 62, 70 Spain 189 see also advanced child maintenance payment Child Support Act (Great Britain, 1991) 163−4 Child Support Agency (Great Britain) 164 childcare arrangements 40 childcare facilities Belgium 88, 90, 290, 368−9 Finland 140−41 Germany 62−3, 291, 371 Great Britain 167−8, 292 Greece 220 Italy 293 Sweden 140−41, 294 see also National Childcare Strategy children, and gender differences 352, 366 CIA (conditional independence assumption) 341 Cigno, A. 7 Clark, W.A.V. 266 class and divorce 179, 193 Spain 197 cohabiting couples 17 Belgium 81, 83, 87, 93 Finland and Sweden 134−6 Germany 54−5, 58 Great Britain 157 Greece 213 Coleman, J.S. 71 comparative research design 17−19, 42−3 compulsory altruism 220 conditional independence assumption (CIA) 341 Cooke, P.L. 32 coping with partnership dissolution 9−11 Corden, A. 41 Corijn, M. 81 Cornelißen, W. 56 country typologies 33−4 Cousins, C. 181, 184, 185 Crompton, R. 33, 156 Daly, M. 291, 292 Darton, D. 160

Index data appendix Table A1 Main features of national and cross-national surveys used in this volume 387 Table A2 Selected statistical indicators Belgium 388−9 Denmark 389−90 Finland 390−92 Germany 392−3 Greece 393−4 Spain 394−6 Sweden 396−7 UK 397−8 data problems 15−16, 263, 288 selection bias 332 data sources 400−401 Bank of Italy Survey of Households’ Income and Wealth 298 BHPS (British Household Panel Survey) 271 ECHP (European Community Household Panel) 221, 243, 271, 288, 338 EU-SILC (EU survey of Statistics on Income and Living Conditions) 220−21 GSOEP (German Socio-Economic Panel) 271 PSBH (Belgian Household Panel) 90 Datta Gupta, N. 106 Davies, K. 160 De Corte, R. 85 De Decker, P. 266 De Lathouwer, L. 82 Dehejia, R. 340, 343, 344 Del Boca, D. 31 Denmark 124−6 data appendix 389−90 determinants of consequences of divorce 113−21 divorce laws 102−3 dual-earner model 104, 238, 338, 372−3 economic consequences of divorce 362−3 family policy 106−7 family structure and divorce patterns 104−6

409

housing 270 post-divorce labour supply 250, 251−2, 258 residential mobility 275, 276 social consequences of divorce 122−4 study data and methods 107−9 results 109−13 welfare state 100−101, 106−7 deprivation, see lifestyle deprivation deprivation index measures 335−7 Dewilde, C. 6, 12, 66, 81, 89, 90, 170, 212, 233, 237, 264, 287, 288 Diekmann, A. 9, 52 Dieleman, F.M. 4, 266, 267 Dienel, C. 293 DiPrete, T.A. 2, 32, 66, 264, 286, 287, 319, 331, 351 divorce democratization 179 divorce laws application of 363 Belgium 79, 83−7 Denmark 102−3 Finland and Sweden 135−6 Germany 57−8 Great Britain 162−3 Greece 215−17 Spain 178, 185−9 divorce penalty 3, 11, 264; see also marriage premium/bonus divorce rates Belgium 78, 80 Denmark 99 Finland and Sweden 134−5 Germany 52 Great Britain 155 Greece 213 Spain 190−92 divorce vs separation 16−17 income levels, Spain 199−200 divorced women, post-divorce labour supply 250, 257−9 Dnes, A.W. 163 Dobash, R.E. 4, 266, 267, 280 Dolado, J.J. 185 Drake, R. 37 Drenth, A. van 165 Drew, E. 31 Dronkers, J. 194

410

Index

dual-earner model 33, 238, 289, 338 Belgium 241 Denmark 104, 238, 338, 372−3 deprivation 348−51 economic well-being impacts of marriage dissolution 346−9 Finland 238, 338, 372−3 Fuzzy Monetary Indicator 347 Netherlands 376 post-divorce labour supply 250, 256−8 residential mobility 275 Sweden 293−4, 372−3 theoretical expectations 241−2 theory 14 Dumon, W. 218 Duncan, G.J. 3, 212, 233, 234, 263, 330 Duncan, S. 132, 268 Düsseldorfer Tabelle 59

Engstler, H. 52, 53, 54, 55, 57 equivalence scales 331, 333−4 Ermisch, J. 157, 171 Esping-Andersen, G. 12, 13, 32, 33, 37, 88, 93, 146, 184, 190, 237, 332, 369, 376 Esping-Andersen typology 241 EU-SILC (EU survey of Statistics on Income and Living Conditions) 220−21 EU survey of Statistics on Income and Living Conditions (EU-SILC) 220−21 European Community Household Panel (ECHP) 221, 243, 271, 288, 338 Evers, A. 21 extended family solidarity model (family model) 14−15 extensive family policy 14

ECHP (European Community Household Panel) 221, 243, 271, 288, 338 economic well-being gender gap 362−3 literature review 2−3 measurement conventional approach 333−4 deprivation index measures 335−7 relative income measure 334−5 overview 330−33, 351−2 study data and method data 338−9 propensity score matching 332, 339−44 study results deprivation indices 347−8 entering poverty 344−6 Fuzzy Monetary indicator 346−7 lifestyle deprivation, basic 349−50 lifestyle deprivation, secondary 350−51 total deprivation 348−9 economies of scale 7−8, 64−5 egalitarian model 133, 135 Emery, R.E. 17 Emmerling, D. 53 employment-supportive policies 37−8 England, P. 29

Fahey, T. 270 familial model 133, 135 family benefits Germany 60−64 see also child benefits family law application of 363−4 Belgium 83−7 Germany 57−9 Great Britain 162−4 Greece 215−17 see also divorce laws family model (of family support) 238, 338 deprivation 348−51 economic well-being impacts of marriage dissolution 346−9 Fuzzy Monetary Indicator 347 Greece 241, 338, 375 Ireland 241, 376 Italy 241, 292−3, 338 Portugal 241, 338 post-divorce labour supply 251−2, 256−8, 257 residential mobility 275−6 Spain 241, 338, 375 theoretical expectations 242 see also extended family solidarity model

Index family policy 14 Belgium 290 Denmark 106−7 Finland and Sweden 141−3 Germany 60−64, 291 Great Britain 164−8 Greece 218−20 family support classifying countries studied 368−76 typology 14−15, 237−41, 333, 367−8 family/kinship solidarity model 181 Feijten, P. 4, 5, 6, 266, 267 female-supportive policies and women’s employment after divorce study data 243−4 hypotheses 241−3 labour supply changes, description 247−52 labour supply changes, multivariate analyses 252−7 literature review 234−5 method 244−7 overview 233−4, 258−9 theoretical background 235−41 Ferrarini, T. 98, 107 Ferrera, M. 13, 181, 182, 218, 225, 270 fertility Belgium 80−81 Denmark 104 Germany 55−6 Great Britain 157−8 Greece 212−13 Spain 190 Sweden and Finland 135 Filippone, A. 336 Finch, N. 134, 135, 143 Finland data appendix 390−92 dual-earner model 238, 338, 372−3 moving towards neo-familialism 133 post-divorce labour supply 251 residential mobility 275, 276−7 taxation 139 see also Nordic welfare model countries Finnie, R. 2, 3, 233, 234, 235, 287, 330 Fix, B. 290, 368, 369, 383 Flaquer, L. 181, 182, 186, 190, 196, 197, 199, 202

411

Flood, L. 327 Flora, P. 132 FM (Fuzzy Monetary) indicator 335, 346−7 Fokkema, T. 233 forced residential mobility 4 Forssen, K. 98 France male breadwinner model 238, 338, 376 residential mobility 276−7, 282 Francesconi, M. 156, 157, 159, 160 Fraser, N. 132 Friedberg, L. 192, 208 Fritzell, J. 6, 99, 107, 287, 330 fundamental problem of causal inference 340 Furstenberg, F. 22 Fuwa, M. 35, 41 Fuzzy Monetary (FM) 335, 346−7 Gähler, M. 6, 99, 146, 287 Gallagher, M. 3 Gallie, D. 13 Garib, G. 206 Gartner, H. 57 Gauthier, A.H. 13, 132, 133 Gelissen, J. 13 gender differences 8−9, 65−6, 99−100, 346−52 Denmark 106, 110−21 Spain 179, 184−5 Gerlo, J. 85, 86, 87, 94 German Socio-Economic Panel (GSOEP) 271 Germany data appendix 392−3 economic autonomy of women 296 empirical study assumptions 64−6 evidence 66−71 family law 57−9 family policy 60−64, 291 family structure changes 52−5 housing 269−70 labour market trends 55−7 male breadwinner model 51−2, 56, 60, 73−4, 238, 291−2, 369−70 post-divorce labour supply 257

412 residential mobility 276−7, 282 see also income changes after separation, comparative analysis Geurts, V. 266 Ginn, J. 29 Giorgi, L. 335 Glendinning, C. 302 glossary 400−401 Goode, W.J. 179, 186, 187, 193, 197, 252 Gornick, J.C. 9, 13, 29, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 104, 112, 132, 233, 237 Gosling, A. 156, 159, 160 Grainger, H. 166 Great Britain 155−6, 172−3 anti-poverty measures 169 demographic changes 156−8 economic autonomy of women 296 family law 162−4 family policy 164−8 focus on ‘clean break’ divorce housing after divorce 266−7 labour market 159−61 market model family support 292, 373−4 partnership dissolution consequences 169−72 poverty 161−2 Working Families Tax Credit 3 see also income changes after separation, comparative analysis; UK Greece 224−6 data appendix 393−4 family law 215−17 family model family support 241, 338, 375 family structure 212−14 housing 270 labour market 214−15 partnership dissolution consequences 220−24 post-divorce labour supply 256 residential mobility 275, 276 welfare state 218−20 Greene, W.H. 305

Index GSOEP (German Socio-Economic Panel) 271 Güllner, 299 Gysi, J. 54 Haataja, A. 133, 140, 147, 148, 149, 150 Hagenaars, A.J.M. 94, 334 Hakim, C. 29, 31, 37, 155, 214 Han, S.-K. 29 Hansen, M.N. 42 Hantrais, L. 37 Hardarson, O. 207 Harkness, S. 32, 159, 161 Härkönen, J. 134, 194 Härpfer, J. 66 Hartmann, H.I. 314 Hartmann, J. 9 Haskey, J. 157, 172 Hatland, A. 132 Haurin, D.R. 3, 233, 234 Heckman, J.J. 332, 340, 341, 343 Henau, J. de 167 Henshaw, D. 164 Henz, U. 99 Herlyn, I. 71 Heyvaert, A. 84, 85 Hiilamo, H. 139, 141, 147, 150 Hinnfors, J. 142 Hinz, T. 57 Hobson, B. 132 Hoekstra, J.S.C.M. 268 Hoffman, S.D. 3, 212, 233, 234, 263, 330 Holden, K.C. 8, 65, 99, 263, 295 Holland, P.W. 340 Holmans, A.E. 4, 266, 281 Holt, H. 166 Hook, J. 35 Houle, R. 194 Houmann, A.L. 99 household complexity 179, 197−9 housing after divorce comparative study analysis 274−81 before and after separation 306−9 consequences for women 315−17 data 271 determinants of economic consequences 309−10

Index gender inequalities postseparation 310−15 impact on household income 310 method 271−2 operationalizations 272−4 overview 268−70, 281−3, 317−19 institutional effects 379−80 literature review 266−7 overview 264−5, 365−6 housing allowances, Finland and Sweden 140 housing regimes 268 Hummelsheim, D. 369 Hurrell, K. 160 Ichino, A. 342, 344 Iglesias de Ussel, J. 186 income, as measure of economic wellbeing 331, 333−4 income changes after separation Belgium 90−92, 286 comparative analysis data sample 298−301 family support types for countries involved 289−94 hypotheses 294−8 income indicators 301−2 methodology 302−6 overview 286−9 results 306−17 Denmark 109−13 Finland and Sweden 145−50 Germany 66−71, 286 Great Britain 169−72 Greece 220−24 Spain 199−204 women as experiencing more severe income loss 295−6, 310 institutional context 11−15 effects income and poverty 376−8 labour force participation 378−9 lifestyle deprivation 380−81 residential mobility and housing 379−80 measurement 36−8 significance of 367−82 involuntary residential mobility 4−5 Ireland economic well-being impacts 346

413

family model family support 241, 376 housing 270 market model family support 338 post-divorce labour supply 257 residential mobility 275 Italy economic autonomy of women 296 family model family support 241, 292−3, 338 post-divorce labour supply 250, 256 residential mobility 275, 276 see also income changes after separation, comparative analysis Jacobs, J. 237 Jalovaara, M. 134 Jarvis, S. 2, 3, 6, 170, 212, 233, 234, 250, 263, 264, 275, 287, 330 Jenkins, S.P. 2, 3, 6, 170, 171, 172, 212, 233, 234, 250, 263, 264, 275, 287, 330 Jensen, A. 102 Johnson, W.R. 3, 9, 233, 234, 235, 252 Kahn, A. 132, 141 Kahn, L.M. 40 Kalleberg, A.L. 32, 233 Kalmijn, M. 3, 6, 11, 89, 211, 264 Kamerman, S.B. 39, 132, 141 Kampmann, C. 66 Kangas, O. 98, 106, 121, 141 Katz, J. 3, 233, 234 Katzenstein, P.J. 132 Kaufmann, F.X. 13, 21, 132, 133, 135, 145, 218 Kaurala, M. 139 Kautto, M. 132 Keith, B. 2, 17 Kelly, J.B. 17 Kemeny, J. 268 Kenny, L.W. 9 Kiernan, K.E. 212, 218, 272 kinship support Germany 71 Kitson, G.C. 263 Klammer, U. 56 Klein, T. 52 Klenner, C. 56

414 Klijzing, F.K.H. 212 Koch-Nielsen, I. 103 Konietzka, D. 54 Kopeinig, S. 342 Kopp, J. 52 Korpi, W. 38, 104, 132 Kreyenfeld, M. 54 Kvist, J. 141 labour market Belgium 81−3 gender equality, measures of 40 Germany 55−7 Great Britain 159−61 Greece 214−15 Spain 182−5 labour supply 3, 10−11 Land, H. 220 Layard, R. 10 Lechner, M. 340, 344 Leibfried, S. 13, 181 Leira, A. 104 Leitner, S. 13, 181 Lemmi, A. 331, 334 Lenoir, V. 327 ‘les misérables’ 120−21, 125 Lesthaeghe, R. 78, 80, 156 Leuven, E. 344 Lewin-Epstein, N. 35, 40 Lewis, J. 13, 33, 51, 56, 104, 112, 132, 164, 237 lifestyle deprivation basic 349−50 institutional effects 380−81 secondary 350−51 Litmala, M. 137 Liu, G. 134 Lohmann, 319 low-income men residential mobility 5 Magdalinos, M. 215, 218 Mahon, R. 133, 145 maintenance payments, see child maintenance payments; spousal maintenance payments male breadwinner model 14, 33, 238, 243, 290, 338−9 Austria 238, 338 Belgium 87−8, 238, 338−9, 368−9

Index deprivation 348−51 economic well-being impacts of marriage dissolution 346−9 France 238, 338, 376 Fuzzy Monetary Indicator 347 Germany 51−2, 56, 60, 73−4, 238, 291−2, 369−70 housing regimes 270 Netherlands 238, 241, 338, 376 post-divorce labour supply 251−2, 256−8 poverty 346 residential mobility 275 Spain 181 see also familial model Mandel, H. 34, 35, 36, 38, 40, 41, 42, 43, 237 Manfred, A. 39 Manning, A. 155, 159, 166 Manting, D. 6, 330 marital disruptions, definitions 180 marital dissolution, definitions 339 marital stability 211−12 market model (of family support) 238, 265, 338 deprivation 348−51 economic well-being impacts of marriage dissolution 346−9 Fuzzy Monetary Indicator 346−7 Great Britain 292, 373−4 housing regimes 270 Ireland 338 post-divorce labour supply 251, 256, 259 residential mobility 275 theoretical expectations 243 theory 14 UK 238, 338 marriage, as utility maximization 211 marriage premium/bonus 3, 11, 100, 264 marriage rates Belgium 80 Finland and Sweden 134 Germany 53 Great Britain 157 Greece 213 Spain 190 married couples, policy support for, Germany 60

Index Marsh, A. 170 Martín-Casals, M. 186, 187 Martiny, D. 71 material quality of life 2 maternity leave Belgium 89 Great Britain 165 Greece 219 Spain 182 see also parental leave Matthijs, K. 78 Mayhew, E. 147 McCarthy, P. 267, 280 McKay, S. 158 McKeever, M. 2, 319 McLanahan, S. 212 McManus, P.A. 2, 32, 66, 264, 286, 287, 319, 331, 351 Meil Landwerlin, G. 200 Melby, K. 102, 103 men with custody 120−21, 361−2 see also single fathers economic well-being 2−3, 263−4 health 264, 367 labour supply 3 presence of children 366 see also low-income men Menning, S. 53, 54, 55, 57 methodological approach 15−19, 29−30 Meyer, D. 54 Meyer, D.N. 41 Meyers, M.K. 9, 33, 34, 36, 37, 38, 39, 40, 41, 42, 112 micro−macro analysis 18, 19 country typologies 33−4 direct account of macro-level effects 34−6 Millar, J. 163, 165, 302 Mitsopoulos, P.G. 212, 221, 225 Moen, P. 29 Montanari, I. 132 Moore, S. 233, 234 Morgan, K.J. 37, 39 Morgan, L.A. 3, 263 Mosteller, F. 94 mothers residential mobility 5 Moussourou, L.M. 218

415

Mueller, R.E. 3, 233, 234 Müller, R. 212, 218 Müller, W. 31 Münder, J. 57, 58, 59, 62 Murphy, M. 157, 266, 267, 280, 281 Myers, M. 104 Naldini, M. 181, 186 National Childcare Strategy (UK) 374 Natti, J. 32 Neels, K. 78, 80 Netherlands dual-earner model 376 male breadwinner model 238, 241, 338, 376 post-divorce labour supply 250, 251, 257, 258 Neubauer, E. 292 New Deal programmes, Great Britain 168−9 Nielsen, L. 103 Niilola, K. 143 Nordic welfare model countries empirical analysis of partnership dissolution 145−50 institutional context 135−45 overview 132−3, 150−51 socio-demographic trends 134−5, 143−5 normal allowance (normalbidrag, Denmark) 103 Nyberg, A. 133, 135, 145, 147, 148, 149, 150 Nygaard Christoffersen, M. 104, 121 occupational quality of life 2 O’Connor, J. 237 O’Donoghue, C. 164 Olovsson, P. 327 Olsen, B.M. 106 Oppenheimer, V.K. 211 Orloff, A.S. 15, 22, 37, 237, 289, 319 Ostner, I. 13, 51, 56 Ott, N. 8 Page, M.E. 171 Palme, J. 38, 98, 106, 107 Palmer, G. 160, 161, 174 Palmer, T. 166 Papadopoulos, T.N. 31

416

Index

parental leave 400 Belgium 89 Denmark 106 Finland 140 Germany 61−2, 291 Great Britain 165−6 Spain 182 Sweden 140, 294 parental support, see kinship support parents, policy support for, see family benefits partnership dissolution consequences 361−7 consequences vs causes 1−2 literature review 2−6, 99−100 significance of institutional context 367−82 paternity leave Belgium 89 Greece 219 Spain 182 Sweden 294 see also parental leave Paugam, S. 13, 320 Perrons, D. 166 Peterson, R.R. 2, 3, 99, 124, 263 Petrongolo, B. 155, 159, 166 Pfau-Effinger, B. 33, 44, 56 Pfenning, A. 132 Pintens, W. 84, 85 Plantenga, J. 39 Pong, S.L. 22 Poortman, A.R. 2, 6, 9, 89, 211, 233, 235, 236, 252, 263, 287, 330 Portugal family model family support 241, 338 residential mobility 275, 276 poverty risks, and life phases 98−9 poverty status, as measure of economic well-being 331−2, 333 propensity score matching (PSM) 332, 339−44 property division 10 Prskawetz, A. 321 Pryor, J. 17 PSBH (Belgian Household Panel) 90 PSM (propensity score matching) 332, 339−44 Putnam, R.D. 129

Raeymaeckers, P. 81, 89 Rainwater, L. 12, 162 Rake, K. 167, 168, 169 Randall, V. 164 Rasmussen, J. Vostrup 103 re-partnering 66 Germany 66−9 Great Britain 171, 173 Spain 179, 196−7 Reitman, F. 100 Reitsma, A.A. 268 remarriage Belgium 81 Great Britain 163 Spain 179, 196−7 residential mobility 4−6, 10, 365−6 institutional effects 379−80 Reskin, B.F. 29 Rindfuss, R.R. 37 Ringen, S. 292 Ritakallio, V.M. 107, 121 Roca, E. 186, 187, 188, 207 Rodgers, B. 17 Rodgers, W. 212 Rogers, S.J. 9 Roos, P. 29 Rose, D. 271 Rose, H. 220 Rosenbaum, P.T. 342 Rosenfeld, R.A. 29, 32, 233, 237 Ross, H. 211 Rostgaard, T. 104 Rowlingson, K. 241 Rowntree, S. 98 Rubin, D.B. 342 rudimentary family policy 14 Ruiz Becerril, D. 196 Ruspini, E. 69, 271, 290, 291 Russell, H. 320 Sainsbury, D. 51, 56, 237, 291, 294 Sakaranaho, T. 142 Saldeen, A. 294 Saraceno, C. 181, 187, 293, 320 Sawhill, I.V. 211 scale economies 7−8, 64−5 Schaeffer, N.C. 302 Scheiwe, K. 57, 58, 59, 60, 62 Schluter, C. 171, 172 Schmid, J. 292

Index Schmidt, C. 344 Schneer, J.A. 100 Schoen, R. 211 schooling arrangements, Germany 63 Schouw, R.J. 4, 266, 267 Schwarze, J. 66, 320 Second Demographic Transition 78, 80, 156, 190, 212 Semyonov, M. 35, 36, 38, 41, 42, 43, 237 Senaeve, P. 79, 83, 84, 85, 86, 87 separated women, post-divorce labour supply 250−51, 257−9 ‘separation from bed and board’ 84 separation vs divorce 16−17 income levels, Spain 199−200 Shavit, Y. 31 Sianesi, B. 344 Siaroff, A. 13 Sigle-Rushton, W. 22, 173 Simó, C. 194, 196 Simpson, B. 267, 280 Sims-Schouten, W. 292 single fathers 80, 120, 125, 362; see also men, with custody single/solo mothers 3, 38, 259, 382 Belgium 80 Denmark 362 East Germany 54 Finland 148−50 Great Britain 158−9, 161, 163−5, 168−9, 171, 374 Sweden 148−50 Sipilä, J. 141 Skinner, C. 168 Skinner, J. 3, 9, 233, 234, 235, 252 Smart, C. 163 Smeeding, T. 162 Smith, I. 155 Smith, J.A. 341, 342, 343, 344 Smock, P.J. 8, 65, 89, 99, 193, 212, 233, 263, 295, 330 Snoeckx, L. 90 social assistance, Germany 64 Solsona, M. 194, 196 Sørensen, A. 8, 65, 66, 287, 294 Spain 178−80, 204−6 data appendix 394−6 divorce laws 178, 185−9

417

economic consequences of divorce at-risk-of-poverty rates 202−4 capacity to face unexpected expenses 204 factors underlying 192−9 median equivalized disposable income 199−202 educational attainment of divorced population 194 evolution of marital disruption 189−92 family model family support 241, 338, 375 housing 270 labour market 182−5 male breadwinner model 181 post-separation labour market participation 196 residential mobility 275, 276 welfare state 181−2 Spieß, C.K. 291 spousal allowance (ægtefællebidrag, Denmark) 103 spousal maintenance payments 9 Belgium 86−7 Denmark 103 Germany 59, 70 Greece 216 Spain 189 Sweden 294 Standing, K. 166 Steele, F. 22 Stevens, A.H. 171 Stevens, J. 160 Stier, H. 29, 33, 34, 35, 37, 38, 40, 42, 233, 237 Sullivan, O. 4, 5, 266 Sundström, M. 138 support payments, see spousal maintenance payments Sutherland, H. 159, 161, 162, 164, 167, 168 Sweden commitment to egalitarian model 133 data appendix 396−7 dual-earner model 293−4, 372−3 economic autonomy of women 296 joint child custody 137

418

Index

taxation 138 see also income changes after separation, comparative analysis; Nordic welfare model countries Symeonidou, H. 212, 213, 214, 215, 218, 220, 221, 225, 226 Symon, P. 5, 266, 267

Van Dijk, L. 29, 30, 31, 35, 36, 233, 236, 237 van Ham, M. 4 Vanhove, T. 78 Vaskovics, L.A. 56, 57 Verma, V. 334, 335, 336 Vikat, A. 134 Vogel, J. 381

Takala, P. 142 tax credits, Great Britain 167 Taylor-Gooby, P. 99 Taylor, M.F. 327 Teachman, J.E. 134 Thornton, A. 212 Timoteo, M. 293, 320 Todd, P. 341, 342, 343, 344 Torgersen, U. 268 Treiman, D.J. 314 Treviño, R. 194, 197 Troedsson, I. 151, 152 Tsui, K. 335 Tukey, J.W. 94

Wagner, M. 1, 52 Wahba, S. 340, 343, 344 Waite, L. 3, 100 Waldfogel, J. 32 Wallerstein, J.S. 22 Walters, P. 29 Wang, D. 157 Warren, T. 29, 38 Wasoff, F. 4, 266, 267, 280 Watson, D. 271 Wegener, M. 102, 103 Weiß, B. 1, 52 Weitzman, L. 2, 99, 124 welfare production 12 welfare regimes typology 13−15 well-being indices, construction of 129 Wennemo, I. 139 Whelan, C.T. 333, 336 Williams, K. 264 Wolfinger, N.H. 2, 319 women economic autonomy of 296 economic consequences in comparative study 315–17 economic well-being 2, 99, 211 employment-supportive policies 37–8 as experiencing more severe income loss 295–6, 310 having children and participating in labour market 236, 256−7 Denmark 104 Germany 56−7 Great Britain 155, 159−60 Greece 215 institutional arrangements contributing to economic independence 38−42 labour force participation 3, 364−5 and divorce anticipation 235

UK data appendix 397−8 economic well-being impacts of marriage dissolution 346 housing 270 market model family support 238, 338 poor welfare provision 330−31 post-divorce labour supply 250−52, 257 residential mobility 275, 276 see also Great Britain Umberson, D. 264 utility maximization, as applied to marriage 211 Uunk, W.J.G. 6, 12, 18, 29, 33, 35, 37, 41, 66, 81, 89, 90, 91, 94, 99, 146, 170, 194, 212, 224, 233, 235, 236, 237, 242, 275, 287, 288, 319, 332 Uyttenhove, A. 86 Valiente, C. 181 Valkama, E. 137 Valtakari, M. 143 Van de Kaa, D.J. 78 Van der Lippe, T. 29, 30, 31, 35, 36

Index institutional effects 378−9 and risk of divorce 134, 211 residential mobility 4−5 see also female-supportive policies and women’s employment after divorce study; mothers women-friendly policies, see employment-supportive policies

Wooldridge, J.M. 305 working-time regulations 39−40, 81 Yeandle, S. 166 Zani, S. 334, 336 Zapf, W. 12 Zippel, K. 37, 39

419