Social Vulnerability in Europe: The New Configuration of Social Risks

Social Vulnerability in Europe The New Configuration of Social Risks Edited by Costanzo Ranci Social Vulnerability i...

13 downloads 1012 Views 2MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Social Vulnerability in Europe The New Configuration of Social Risks

Edited by

Costanzo Ranci

Social Vulnerability in Europe

This page intentionally left blank

Social Vulnerability in Europe The New Configuration of Social Risks

Edited By

Costanzo Ranci Polytechnic of Milan, Italy

Selection and editorial matter © Costanzo Ranci 2010 Individual chapters © their respective authors 2010 All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No portion of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, Saffron House, 6-10 Kirby Street, London EC1N 8TS. Any person who does any unauthorized act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The authors have asserted their rights to be identified as the authors of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2010 by PALGRAVE MACMILLAN Palgrave Macmillan in the UK is an imprint of Macmillan Publishers Limited, registered in England, company number 785998, of Houndmills, Basingstoke, Hampshire RG21 6XS. Palgrave Macmillan in the US is a division of St Martin’s Press LLC, 175 Fifth Avenue, New York, NY 10010. Palgrave Macmillan is the global academic imprint of the above companies and has companies and representatives throughout the world. Palgrave® and Macmillan® are registered trademarks in the United States, the United Kingdom, Europe and other countries. ISBN-13: 978–0–230–58091–6 hardback This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. Logging, pulping and manufacturing processes are expected to conform to the environmental regulations of the country of origin. A catalogue record for this book is available from the British Library. A catalog record for this book is available from the Library of Congress. 10 9 8 7 6 5 4 3 2 1 19 18 17 16 15 14 13 12 11 10 Printed and bound in Great Britain by CPI Antony Rowe, Chippenham and Eastbourne

Contents

List of Figures and Tables

vii

List of Abbreviations and Acronyms

xii

Notes on the Contributors

xiii

Preface

xv

Acknowledgements

xix Part I Concepts and Methodology

1

Social Vulnerability in Europe Costanzo Ranci

3

2

Bringing Territory Back in Social Comparative Research Costanzo Ranci

25

Part II Aspects of Social Vulnerability 3

Beyond the Male Breadwinner Family Model Emmanuele Pavolini and Costanzo Ranci

37

4

Income Vulnerability in Europe Salvatore Curatolo and Guglielmo Wolleb

58

5

Unstable Employment in Western Europe: Exploring the Individual and Household Dimensions Ivana Fellini and Mauro Migliavacca

88

6

Housing Deprivation and Vulnerability in Western Europe Pietro Palvarini and Emmanuele Pavolini

126

7

Disability and Caregiving: A Step Toward Social Vulnerability? Giuliana Costa and Costanzo Ranci

159

Part III Multidimensional Analysis 8

The Vulnerability of Young Adults on Leaving the Parental Home Giuseppe A. Micheli and Alessandro Rosina

189

9

Social Vulnerability: A Multidimensional Analysis Costanzo Ranci and Mauro Migliavacca

219

v

vi

Contents

10 Explaining Social Vulnerability Costanzo Ranci, Brunella Fiore, Emmanuele Pavolini

250

General Conclusions Costanzo Ranci

279

References

283

Index

295

List of Figures and Tables

Figures 1.1 1.2 1.3 1.4 1.5 2.1 3.1 3.2 3.3

3.4 4.1 4.2 5.1 5.2 5.3 5.4 6.1 6.2 6.3 6.4

Share of families in temporary, recurrent or permanent poverty in the EU-15, 1995–2001 Share of families dealing with housing deprivation in the EU-15, by geographical areas – 2001 Share of temporary workers and unemployed workers in EU-15, 1983–2006 Female activity rates for women aged 25–49 years old by number of children in the EU-15, 1995–2005 Proportion of people aged 65–79 years and aged 80 years and more in the EU-15, 1977–2008 Macro-regions in the EU-15 Demographic structures of households in the EU-15, percentage by area Household income structures, percentage by area Changes in the income structure of households caused by the presence of children (in relation to households without children), percentage by area Household typology, EU-15 Equivalent median income by macro-region, 2001. Index number (EU-15 = 100) Gini index for the distribution of disposable equivalent household incomes – EU-15 countries – 2000 Share of temporary employees as a proportion of total employees – EU-15 Main income source of unstable workers, by country – 2001 Work family models Share of households with children aged less than 12 years by family/work patterns Home- ownership and variation in home- ownership over time (1980–2003) in the EU-15, by country Percentage of households with overcrowding, by macro-region (2001) Percentage of households with poor housing quality, by macro-region (2001) Percentage of households living in neighbourhood with poor quality, by macro-region (2001) vii

6 8 9 10 12 31 43 48

51 54 61 62 99 109 112 118 129 139 141 143

viii

7.1 7.2 7.3 7.4 8.1 8.2 9.1 9.2 9.3 10.1 10.2

List of Figures and Tables

Average income of families with 1+ dependent person/average income of families without dependents, EU-15 Percentage of families with dependent adult receiving DSBs and average amount of DSBs in the EU-15 Effect of DSBs on the poverty risk of families with dependent people – 2001 Gender and age composition of dependent persons and of caregivers – 2001 Flow chart of the vulnerability in the transition to adulthood Five European macro-regions with distinct demographic and anthropological patterns Share of vulnerable population in 13 Western European countries, by macro-regions Geographical clusters of social vulnerability Map of vulnerability clusters GDP level (PPT) per capita by macro-region (in euros) Level of welfare protection against new and old social risks of the macro-regions

165 170 171 175 200 212 237 239 242 262 263

Tables 1.1 3.1 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9

4.10

Vulnerability and precipitation/protection factors for different level of functioning Position of individuals over 20 years of age in families, by age groups Population shares by country, region and income class – 2001 Population shares by country and income class – 2001 Distribution of the European population by income class Poverty persistence rate by country and macro-region 2001 Income class provenance of the poor, by country Income class destination of the poor, by country Proportions of people experiencing from 0 to 5 episodes of poverty over the 5-year period 1997–2001 by country Longitudinal and cross-sectional poverty: estimated number and share of people Longitudinal poverty and 2001 income class: absolute values, percentages and class probability of experiencing poverty Demographic and social factors connected with vulnerability and persistent poverty (difference in percentage relative to the whole population)

20 41 64 66 67 68 71 71 74 76

77

79

List of Figures and Tables

4.11

4.12

5.1

5.2 5.3

5.4 5.5 5.6 5.7

5.8 5.9 5.10

6.1 6.2 6.3 6.4 6.5 6.6 6.7

Economic factors connected with vulnerability and persistent poverty (difference in percentage relative to the whole population) Structural factors connected with vulnerability and persistent poverty (difference in percentage relative to the whole population) Share of unstable employment as a proportion of total employment, percentage composition of unstable workers and share of temporary employment on total employment – 2001 Total and unstable employment by sex, age, education and sector of activity – 2001 Ratio between the share of temporary workers and the share of total employment for women, for aged 15–24 and aged 24–39, for workers with ISCED (0–2) and for workers in the service sector – 2002 and 2007 Unstable workers in 2001: labour market positions in the previous 3 years Unstable workers in 1998: labour-market positions in the following 3 years Different inclusion/exclusion patterns of unstable workers over the period 1998–2001 Household members/employed members ratio and relation to total employment rates, female activity rates, and household size. Year 2001 Share of work/family models by macro⫺region 2001 Work/family models below the national poverty line. Weighted deviations. Year 2001 Share of households with children aged over 12 years by family/work patterns (with the woman’s age max. 35) National differences from the European average. Year 2001 Percentage of owner-occupied housing by country 1980–2003 The housing situations of poorer households (incomes < 60 per cent of the national average) Percentage of households with different degrees of housing affordability, by macro-region (2001) Percentage distribution of households by housing situation measured on three indices (EU 2001, %) Percentage of households with housing problems, by macro-region – three deprivation indicators (2001) Percentage of households with housing problems by type of occupancy Percentage of households with housing deprivation in different geographical areas, by type of tenure (2001)

ix

80

82

94 97

99 103 104 106

111 114 117

120 130 131 136 145 147 149 149

x

List of Figures and Tables

6.8 Percentage of households with housing problems, by type of household (2001) 6.9 Percentage of households with housing deprivation in different geographical areas, by type of household (2001) 7.1 Percentage of households with a dependent member by care regime 7.2 Levels of economic compression in households with or without dependent members by care regime, percentage of the total of households with a dependent 7.3 Characteristics of households with dependent members easily able to satisfy their material needs, by care regime 7.4 Share of families with dependents consisting of single individuals or of cohabiting adults of different generations, by care regime – 2001 7.5 Effects of the family situations of dependent persons on poverty, restrictions on consumption and on access to DSBs – 2001 7.6 Percentage of caregivers with basic education by age – 2001 7.7 Activity rate of caregivers by age and gender – 2001 7.8 Economic conditions of families with a dependent by intensiveness of the caregiving – 2001 7.9 Participation of caregivers in various forms of sociality by intensiveness of caregiving – 2001 8.1 Average age at first marriage and at birth of first child, by country. Year 2003 8.2 Percentage of young adult home-leavers with specific characteristics (measured in the year following exit), by country 8.3 Percentage of young adult home-leavers according to type of occupancy (in the year following exit) by country 8.4 Vulnerability indicators. Income poverty and housing deprivation, by gender and country 8.5 Vulnerability indicators. Percentage of young adult home-leavers unable to save in the past year, or with difficulties in ‘making ends meet’, by gender and country 8.6 Multipliers (M) and reducers (R) of vulnerability when leaving home: susceptibility to a lower second-order criticality (no saving) 8.7 Multipliers (M) and reducers (R) of vulnerability when leaving home: susceptibility to a lower second-order criticality (drift under poverty line)

150 152 164

167 168

173

174 177 178 179 180 191

195

196 197

198

203

204

List of Figures and Tables

8.8

8.9

8.10 9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8

10.1

10.2 10.3

Typology of multipliers and reducers of vulnerability, by level of severity of the second-order critical state (less frequent factors in brackets) Shares of young people age 25–29 and 30–34 who have left home, have formed an union, and had a child, by gender (year of survey 2001) Areas of influence of the vulnerability multipliers Share of households and individuals in social vulnerability conditions, by risk factor Multi-dimensional index of social vulnerability (percentage of the total of households and individuals) Logistic regression of the individual risk factors on the multiple vulnerability index Logistic regression of the individual risk factors on the severe distress index Vulnerability index by family type Mean values of vulnerability indices by territorial cluster Distribution of vulnerability by deciles of household income Distribution of the vulnerability index by social class position of households (occupational category of the head of household) Principal Component Analysis of the distribution of 12 policy measures addressing social risks in the macro-regions: communalities and extracted components matrix Probability of vulnerability by family type (logistic regression) Multi-level determinants of vulnerability in 27 macro-regions; random intercept model

xi

205

213 215 224 227 231 233 235 240 245

247

260 267 268

List of Abbreviations and Acronyms DSB ECHP EHI ESPON EU-15

Disability/sickness benefit European Community Household Panel Equivalent household income European Spatial Planning Observation Network the ‘Europe of the 15’: Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweden, United Kingdom GDP Gross Domestic Product ISCED International Standard Classification of Education, ISCO88 International Standard Classification of Occupations ISPL International Standard Poverty Line MBW Male Breadwinner (family) NUTS Nomenclature of Territorial Units for Statistics OAH One and a half (Family)

xii

Notes on the Contributors Giuliana Costa is Assistant Professor in Sociology at the Polytechnic of Milan. She teaches courses on Sociology and Social Policy. She is member of the Social Policy Research Laboratory at the Polytechnic. Her main research topics are social policy and elderly care, on which she has published several books and articles. Salvatore Curatolo is Assistant Professor in Applied Economics at the University of Parma, Faculty of Economics. He teaches courses on Industrial Organization and Welfare Economics. His main research topics are labour economics, institutions and public goods provision and welfare economics. Ivana Fellini is Assistant Professor of Economic Sociology at the University of Milan – Bicocca. Her main research interests cover labour market analysis and trends with particular focus on flexibility, gender and immigration. Brunella Fiore took her Ph.D in Applied Sociology and Research Methods in 2008 and she is currently a research fellow at the University of Milan – Bicocca. Giuseppe A. Micheli is currently full Professor of Demography at the University of Milan – Bicocca. He has published extensively on labour market and urbanism, ageing and mental disorders, as well as on mathematical models of population dynamics. His recent research includes both demographic (exploring the logic of the choices of transition to adulthood) and social policy issues (studying mechanisms of entrapment in the processes of impoverishment). He is a member of the scientific board of Genus, and of the Italian Council of Social Sciences. Mauro Migliavacca, Ph.D. in Sociology and Social Research, teaches and researches at the Catholic University of Milan, at the Polytechnic of Milan and at the Ca’ Foscari University in Venice. His main interests include social vulnerability, with special attention paid to changes in the labour market and in family organization. He recently wrote Famiglie e lavoro. Trasformazioni ed equilibri nell’Europa mediterranea, (Family and work. Transformations and stability in Mediterranean Europe), Mondadori, 2008. Pietro Palvarini took his Ph.D. in Urban and Local European Studies (URBEUR) at the University of Milan – Bicocca. His main research interests include social stratification in urban areas, multi- dimensional approaches to social inequalities, housing deprivation and disadvantaged neighbourhoods. xiii

xiv Notes on the Contributors

Emmanuele Pavolini Ph.D. is Professor in Economic Sociology and Social Policy at the University of Macerata. He has been an adviser and researcher in welfare policies for different private and public institutions in Italy. He has published articles and books on social policy and the welfare state, especially on the third sector and on long-term care policies. He has recently published with C. Ranci an article in the European Journal of Social Policy on ‘The transformations in the Social Care Markets for Long-Term Care in Europe’ (2008). Costanzo Ranci is Ph.D. Professor in Economic Sociology and Director of the Social Policy Research Laboratory at the Polytechnic of Milan. He has served as adviser and researcher in welfare policy for many private and public institutions in Italy, including the Italian government. He has published numerous articles and books on social policy and the welfare state, the third sector and voluntary organizations, new social risks and social vulnerability. With U. Ascoli, he has edited Dilemmas of the welfare mix. The new structure of the welfare in an era of privatisation, New York: Kluwer/Plenum Publ. (2002). Alessandro Rosina is Associate Professor in Demography at the Catholic University of Milan. He teaches in courses on Demography and Population Dynamics Modelling. He is author of a number of international articles on the issues of family formation and inter-generational relationships. Guglielmo Wolleb is Professor of Economics at the University of Parma. He teaches Micro- economics and European regional policies and is Director of the post-graduate course ‘Local Development and International Cooperation’. His current research focuses on local development policies in Italy and Europe. His most recent work has been as co-editor of the Handbook of Social Capital (2008) for the Oxford University Press. He has collaborated on European projects and policies and has acted as an evaluator of European programmes on social and economic cohesion.

Preface

The so- called ‘European social model’ was based for many decades on a close association between permanent employment, stability in the division of roles within the nuclear family and the progressive extension of welfare guarantees. All these conditions seem currently to be lacking because of greater job insecurity and the consequent discontinuity in incomes, the organizational tensions taking place in the family and the weak capacity for innovation of European welfare systems. Taken together, these transformations and inaction are causing a spreading of social vulnerability in the population. This book is primarily aimed at describing the dimensions and characteristics of social vulnerability in Western Europe. Although the literature on the ‘new social risks’ is now quite broad, there has so far been no analysis that shows the actual diffusion and intensity of such risks. It is our purpose to fill this gap by taking the paradigm of the ‘new social risks’ as a hypothesis to be verified through an empirical analysis that covers a broad spectrum of aspects. The central notion of the book is social vulnerability. Different from the notion of poverty, it includes aspects that are not connected exclusively with income level, but also with housing conditions, employment, the management of care for children and dependent persons, the models of family organization and the difficulties arising in the transition through different phases of the life- course. It does not concern individuals facing severe hardship or social exclusion, but people who are permanently exposed to instability and weak integration into society. The aim of the book is to show to what extent the emerging risks in postindustrial societies are combining to compromise the normal functioning of a substantial proportion of the Western European population: how material conditions and the standard of living attained are or are not guaranteed, how well or badly the principal changes in the life-course are managed and how care for children and dependent people is provided. The analysis maintains a comparative purpose: the aim is in fact to identify the different profiles of vulnerability that exist in a number of the different territorial areas that make up Western Europe. Since these profiles depend on a complex set of social, economical and institutional characteristics, one of the main contributions of the book is to develop comparisons on a sub-national scale, starting with the identification of 28 ‘macro-regions’. xv

xvi Preface

Structure of the book The book is structured in three parts and 10 chapters. Part I presents the main theoretical concepts and the analytical tools that have been adopted in the study. Chapter 1 illustrates the principal hypotheses of the book, by looking at the changes that have occurred in the last 10 years in Western Europe and their effects on the risk profile of the population. Five new phenomena are identified: the spread of income instability, the rise of new forms of housing deprivation, the growth in temporary work, the difficulties in reconciling care and working, and the explosion of a group of dependent persons because of the ageing population. The notion of social vulnerability is then introduced and its connection with the concepts of ‘new social risks’ and poverty is explained. Finally, the possible policy implications of such perspective are illustrated. Chapter 2 presents the division of Western Europe into 28 ‘macro-regions’. First it is explained why a comparative analysis, focused on changes occurring in the social structure, should consider the existence of important sub-national differences through Europe. This discussion is also set in the context of the trend toward regionalization of welfare policies. The method used for the comparative analysis is also explained and discussed. Part II of the book describes the different dimensions of social vulnerability that have been under scrutiny. Chapter 3 is focused on the transformation in the household structure as a consequence of the crisis of the so- called ‘male breadwinner model’ and the spread of new family models. The following dimensions are considered: the demographic composition of the household (who lives with whom), the household income structure and the organization adopted by families with children in order to solve the problems of reconciling care and working. Chapter 4 presents an analysis of income inequalities through Western Europe. The focus is on a population characterized by transitory or recurring poverty, conditions that indicate more an unstable income than an actual risk of poverty. The differences between European countries are analysed, looking at the income position of individuals before and after episodes of poverty. Once the vulnerable population in terms of income has been identified, the chapter describes the principal social and economic characteristics, distinguishing this profile clearly from those of the more traditionally analysed people in permanent condition of poverty. Chapter 5 presents a ‘work instability index’, which considers the presence and the characteristics of temporary workers while taking account of the different labour market regulations that exist in the EU-15. This index is used to compare the magnitude of employment instability and to analyse current trends. One specific section is dedicated to understanding to what extent temporary work constitutes an element of being segregated and trapped in a marginal labour market. The last part of the chapter analyses

Preface

xvii

the impact of temporary work on families’ incomes and their propensity to have babies. The shift in the focus from the individual to the family level makes it possible to assess the potential of temporary work for increasing vulnerability in different countries. Chapter 6 focuses on conditions of housing deprivation that do not imply inability to gain access to housing, but the presence of problems concerning the quality and the cost of housing. Four dimensions are considered: affordability, overcrowding, the physical quality of dwellings and the quality of the surrounding area. Subsequently the presence of accumulation, as well as interconnections, among the different dimensions is considered. The analysis is performed by considering the variations that have occurred over time and the differences existing between the macro-regions of Western Europe. Finally, Chapter 7 describes the living conditions of the European population in conditions of severe dependency and of the informal caregivers who care for these people. The basic hypothesis is that dependency constitutes a significant risk factor that considerably increases the probability of these persons falling into situations of social vulnerability. What emerges is that dependency causes a strong compression in the use of economic resources rather than any substantial risk of poverty. The final part of the chapter is dedicated to an analysis of informal caregivers; their social and economic characteristics are reconstructed and it is considered whether and to what extent caregiving constitutes a factor of social vulnerability. Part III of the book is intended to present analyses that take the whole set of vulnerability dimensions into consideration in order to study emerging problems and provide a general interpretation of the current situation. Chapter 8 analyses the potential for vulnerability in a specific phase of the transition to adult life: the situation of individual leaving his or her parents’ household. In fact, European regions show extremely varied trends in the timing and sequence of this process. This chapter discusses these differences and concentrates on a longitudinal analysis of factors that facilitate or hinder the living conditions of young people in the transition period. The second part of the chapter identifies the existence of different geographical areas in Western Europe characterized by specific transition models associated with particular types of cohabitation and particular configurations of life- courses. The aim of the last two chapters is to bring back together the different dimensions of social vulnerability that have previously been addressed in a single analysis. In Chapter 9 the different dimensions of vulnerability are therefore considered in the way they combine and in how they are composed to identify complex and diversified profiles of vulnerability. The distribution of vulnerability profiles in households and in different regions of Europe is then investigated. Furthermore, the chapter analyses the extent to which emerging vulnerability profiles intersect with the traditional

xviii

Preface

measures of inequality used in comparative analysis, ranging from analysis of poverty to analysis of class stratification. Chapter 10 has a more interpretative aim. It attempts to verify the extent to which the vulnerability profiles previously identified are influenced in their depth and distribution by a plurality of social, economic and institutional factors. Both family-related and territorial variables are simultaneously considered in a multi-level statistical model. Moreover, the analysis considers the role played by welfare systems in covering specific areas of vulnerability and vice versa in leaving specific needs without protection. The role played by regional socio- economic development is also taken in account. Useful indications emerge from here for the analysis and policy making of social policies. COSTANZO R ANCI

Acknowledgements

The analysis of this plurality of aspects was performed by a multi-disciplinary group composed of sociologists, political scientists, demographers and economists who worked together for a long time following the same analytical perspective. The scholars engaged in the present research agreed to participate in a quite complex project, which implied many collective meetings and discussions, and the use of the same ECHP database and specific analytical tools. This book is not, therefore, a presentation of specific studies. It is, instead, the result of a collective research, even if the final responsibility for the aggregate outcomes belongs to the editor. Like many other comparative studies, this project needed quite a significant level of financial resources. This was provided by Compagnia San Paolo, a bank foundation operating in Turin that gave great economic support to this research and allowing us to work with no strict time constraints: a rare situation that all researchers would like to enjoy and that usually is difficult to find. Compagnia San Paolo funded the project through the active involvement of URGE, the Research Unit on European Governance, a research centre dedicated to the study of European governance. I am very grateful to Maurizio Ferrera and Stefano Sacchi, current director of URGE, for the help and interest shown in the project. A major part of this research was carried out at the Social Policy Research Laboratory (LPS), operating at the Polytechnic of Milan. I want to thank all LPS researchers, who wonderfully assisted and supported me during this work in so many ways. A special thanks to Laura Di Maria, who serves as technical associate at LPS, for the attention and competence used to solve the thousands of troubles that empirical comparative research always presents to researchers. I am sure that without her assistance this book would not have been in your hands now. Many different people have helped to make this book possible, and the contributors and I would like to acknowledge all of them. A special help came from Maurizio Ferrera, who believed in the project and supported me with great solicitude in many ways, always giving his time and attention to my ideas and problems. Patrick Le Gales was also ready to discuss with me a preliminary draft and to encourage me to submit it to a prestigious publisher. I want to thank, too, my colleagues at the LPS who supported the preparatory and final stages of the research: Giuliana Costa, Fabio Manfredini, Mauro Migliavacca, Francesco Minora and Rossana Torri. Other colleagues gave me their precious comments on preliminary drafts on the book: among them I would like to thank Manuela Naldini and Antonio Tosi. xix

xx

Acknowledgements

Preliminary versions of specific chapters have been presented and discussed in innumerable academic meetings and international conferences. Among them I would like to remember the informal seminars held at the State University of Milan and at the Polytechnic of Milan, the 2007 International Conference of SASE (Society for the Advancement of SocioEconomics) in Copenhagen and the Yearly Conference of ASA (American Sociological Association) held at Boston in 2008. These and other meetings gave me the opportunity to test my ideas and research results, and confront them with the knowledge and sometimes the scepticism of my colleagues. This research is the final outcome of many years of work filled not only with plenty of practical and intellectual efforts, but also with passion and enthusiasm. The result is a book that is full of figures and numbers. This was part of the project, of course. But I will be very satisfied if the reader, going through the huge amount of numbers presented in the book, finds not only interesting results, but also challenging ideas. COSTANZO R ANCI

Part I Concepts and Methodology

This page intentionally left blank

1 Social Vulnerability in Europe Costanzo Ranci

The new social question The second half of the twentieth century will be remembered as an era characterized, in advanced capitalist European countries, by the creation of huge systems of welfare protection. Supported by substantial economic growth and by a relatively peaceful international situation, Western European countries developed economic and institutional mechanisms that guaranteed a high standard of living for the very broad majority of the population. The main risks encountered by people during their lives – illness, unemployment, disability and old age – were insured against by generous systems of state protection. At the beginning of the twenty-first century, the great majority of the western European population is in possession of social guarantees and is basically protected against the major threats that hung very heavily over the heads of previous generations. Nevertheless, although social protection systems have now reached huge numbers of people, the changes that have occurred over the last two decades have created new forms of insecurity and instability, which are hitting a large proportion of European citizens. The new century has inherited a strange paradox from the previous century: uncertainty and instability have been growing constantly since the capacity of social systems to offer protection against social risks reached its maximum historically. How can this paradox be interpreted? What explanation is there to offer? One possible answer is to consider uncertainty as the inevitable effect of a transition phase in which industrial society is destructuring with the passage toward a new form of social organization. The simultaneous presence of maximum security and great insecurity would reflect the ambivalences and contradictions that are typical of transition phases, when elements of disorganization are more prevalent than aspects of organization. It is a transitory phenomenon that reflects the crisis of organized capitalism (Lash and Urry, 1987) more than the construction of a new social order. According to Crouch (1999), the fragmentation and disorganization of society that 3

4

Social Vulnerability in Europe

characterize the current period in history are the product of the strong individualization of social relations, to the point where crisis and anomie prevail over the chance for innovation. According to the perspective opened up by the analysis of Beck about risk society, uncertainty is not, however, a transitory syndrome but a permanent trait of post-industrial society. Since industrial societies were oriented toward future development, the propensity to risk was high: in industrial societies the logic of the production of wealth dominates over the logic of the production of risks (Beck, 1992). The latter were considered as latent side effects. However, risks cease to constitute a side effect in post-industrial societies and they move increasingly to the centre of the stage. Confidence in the ability to keep risks under control is replaced by the idea that risks are not fully predictable and controllable. According to Beck, ‘a utopia of security’ with a peculiarly negative and defensive character is growing. It is no longer a question of obtaining something good, but just of avoiding the worst: the dominant purpose is self-limitation (Beck, 1992). But what are the new risks that are emerging in post -industrial societies today? The so-called ‘European social model’, which has characterized the development of industrial societies since the post-war period, has rested on three basic foundations (with the partial exception of Scandinavian countries where the transition to post-industrial society occurred much earlier): high employment stability, broad and generous welfare programmes and the persistence of strong family ties based on a gender division of roles. The development of welfare systems has made a substantial contribution to the bond between the dominant organizational model in the sphere of production and the pattern dominating in the family sphere, offering protection against what has been considered the most serious social risk: losing a job (Esping-Andersen, 1999). In the course of a few decades, the three foundations (work, family and welfare) on which post-war European societies rested have progressively lost their capacity to provide for the well-being and security of many citizens. According to Esping-Andersen, these institutions are today the principal sources of danger (1999). Those threatened are both citizens at the extremes of the age bands (children and the elderly) and the middle class. It is a process that has been defined as a progressive erosion of intermediate positions (Castel, 1995). The first form of erosion regards the organization of work. The fundamental break with the industrial wage-earner model lies in the weakening of the labour market to function as the principal mechanism of social integration. It is increased job insecurity that lies at the origin of this process, founding working relationships not on skills made continuously available, but on the immediate performance of specific tasks. There is nothing marginal in this trend: increasing job insecurity is a mainstream process, determined by the new technological and economic demands connected with the evolution of

Social Vulnerability in Europe

5

modern capitalism. According to Castel (1995) increasing insecurity consists of three processes: the destabilization of workers who were previously stable – which lowers their living standards, the growth of a large mass of long-term unemployed and the creation of a ‘supernumerary’ population of persons who are permanently excluded from the labour market. Even in countries where the silent revolution of work is not generating great unemployment, it is nevertheless undeniable that we are now a long way from the ‘wage-earner society’ that characterized Western Europe until the 1990s. The second form of erosion consists of the gradual weakening of kinship support networks as a consequence of new demographical trends and of the reorganization of households. New forms of households have developed, while established family models have experienced profound internal reorganization. Single-person households and single-parent households have become common. Male breadwinner families have become increasingly less numerous because of the progressive increase in female employment (see Chapter 3 for details). At the same time, new types of household are emerging where adult children remain in their parents’ home long after they have reached economic independence (see Chapter 8). While on the one hand these phenomena are a sign of the progressive individualization of social life, on the other hand they compromise the family capacity for collecting and redistributing resources to the benefit of its weakest members: children, the elderly and people unable to work. The functioning of families faced with the difficult task of reconciling different duties is in fact being heavily overloaded. There is a strong contrast between these changes and the great rigidity of welfare systems. They are undergoing a third process of erosion. Welfare systems have remained trapped in a model that is no longer in harmony with the emerging risk profiles in European societies (Taylor-Gooby, 2004a). The social protection systems existing in many European countries offer an extensive social protection only to citizens who are fully integrated in the labour market. Hit by the fiscal crisis of the welfare state, and as a consequence of the changes in the labour market, today this protection is granted to a smaller proportion of citizens and with less generosity than in the past. However, what is even more important is that at the same time new risk profiles are emerging for which the existing welfare state is not organized to provide an adequate response (Esping-Andersen, 1999). Therefore the welfare state requires a general rethinking of its financial and organizational architecture if new risk profiles are to be adequately protected. While protection against risks was guaranteed in post-war society by the association between secure jobs, a stable division of roles within the nuclear family and a progressive extension of the guarantees furnished by the welfare state, today new risks are arising precisely because of the combination

6

Social Vulnerability in Europe

of increasing job insecurity, decreasing caring capacity of families and the institutional inertia of social welfare systems. New categories of social risk have therefore emerged and the purpose of this book is to define and to describe them.

The new social risks New social risks arise at the point where job insecurity, income instability, increasing fragility of family support and inertia of welfare institutions intersect. Five principal critical problems can be identified. The first concerns the spread of ‘integrated poverty’ (Paugam, 1997), which includes a large group of European citizens who temporarily or sporadically face a situation of relative poverty. An area of ‘transient poverty’ or ‘recurrent poverty’ (Layte and Fouarge, 2004) affects a total of 20–25 per cent of the population in Western European countries (according to ECHP (European Community House Panel) data), with peaks in Southern Europe and in the UK and lower levels in continental Europe and Scandinavian countries. Layte and Fouarge (2004; see also Layte and Whelan, 2005) show that people affected by temporary poverty are much more than people in a condition of persistent poverty. Data presented in Chapter 4 confirm the same fact (see Figure 1.1), indicating that cumulated poverty is less diffuse than financial fragility and income instability. 0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

Luxemburg The Netherlands Finland Germany Austria Belgium Denmark France Italy UK Spain Portugal Greece Irland Temporary poverty

Recurrent poverty

Permanent poverty

Figure 1.1 Share of families in temporary, recurrent or permanent poverty in the EU-15, 1995–2001 (no data for Sweden) Source: ECHP, authors’ own calculations.

Social Vulnerability in Europe

7

Another sign confirming the considerable expansion of economic vulnerability is the number of people in the European population who live just above the standard poverty line. According to Forster and Mira d’Ercole (2005), about 6 per cent of the European population have incomes that lie between 50 and 60 per cent of the median income, while 10.6 per cent are below the 50 per cent of median disposable income threshold. These data show the relevant dimension of a population group that is not poor, but economically fragile, and that accounts for more than 50 per cent of the population that lies below the poverty line in one year. What therefore emerges is a very broad area of income instability in Europe. Nearness to the poverty line and crossing it at times does not necessarily determine a drop into permanent poverty. Income instability indicates a condition of life characterized by strong economic stress and a marked reduction in the standard of living that is accompanied by an uncertain financial position. This condition of fragility increases the probability of social exclusion when negative events occur (illness, unemployment, family breakups, etc.). Moreover, it constitutes in itself a condition of difficulty that has effect on life conditions. A second problem concerns the diffusion of housing deprivation related to affordability difficulties or housing inadequacy: situations that expose people to social instability and financial strain, but do not translate directly or necessarily into severe hardship or homelessness. Western European countries have seen a great improvement in the housing conditions of their citizens in the last three decades. About three quarters of the Western European population today owns the house where they live. Even traditional difficulties, such as overcrowding, have been partially overcome. Quite paradoxically, however, this growth has come with increasing tensions and difficulties, mainly as a result of rising housing costs and changes in the housing market. The co-presence of these two phenomena – increased home ownership and higher housing costs – has made access to a dwelling much more difficult for some sections of the population. The paradoxical finding is therefore that in Europe (and especially in some areas of it) relative housing deprivation has grown. According to the data provided in this book (see Chapter 6) European people who fulfil at least one condition of housing deprivation make up between 5 and 8 per cent of the total, with marked geographical differences (see Figure 1.2). Today access to housing constitutes one of the harshest difficulties met by the young in their transition toward adulthood (see Chapter 8 for details), especially in metropolitan areas. The huge costs of accessing housing contribute not only to lowering the living standards of households, but also to delaying life projects and to depressing expectations for the future. A third problem concerns the spread of jobs and careers in which work is insecure and temporary. Temporary employees above 25 years of age in the EU-15 as a percentage of total workers rose from 8 per cent in 1996 to

8

Social Vulnerability in Europe

Small areas in Central-Northern Europe

Almost all Central-Northern Europe

Physical quality Over-crowding Affordability

Iberian peninsula. Central-Northern Italy. Ireland

Multiple problems

Greece and Southern Italy

0

5

10

15

20

25

30

35

40

Figure 1.2 Share of families dealing with housing deprivation in the EU-15, by geographical areas – 2001 Source: ECHP, authors’ own calculations.

11 per cent in 2006. The current percentage for women is 12 per cent. In recent years the percentage of temporary workers has exceeded the rate of the unemployed, which fell from 10 per cent in 1996 to 7 per cent in 2006 (see Figure 1.3). This happened in most European countries except for continental countries in the German area and for the UK and Ireland. The spread of temporary work has been considered a factor in increasing the risk of acquiring a lower wage, of workers becoming trapped in jobs that are constantly insecure and of exclusion from the labour market. These risks appear to be particularly high for workers with low levels of education and few occupational skills. As a consequence of the progressive increase in the level of education of workers over the last decade, a reduction occurred in the relative proportion of temporary workers with low education (down from 39 to 33 per cent between 1996 and 2005 in the EU-15). In the same period job insecurity became more frequent in both the low-waged service industries and highly skilled and very professionalized industries, contributing therefore to the polarization of the labour market and affecting social groups that were traditionally considered protected against the risks of temporary employment. Temporary employment increases the risk not only of impoverishment, but also to the general living conditions of workers. This situation does not only affect the young population, but also, and increasingly, the adult population: only 34 per cent of people in insecure jobs in the EU-15 are under 24, while 41 per cent are between 25 and 39 years of age.1 The probability that a young person under 24 years had a temporary job was 34 per cent in 1996 and over 10 years it has increased by 20 per cent to reach 41 per cent. In the population between 24 and 39 years old the growth of temporary workers

Social Vulnerability in Europe

9

12 10 8 6 4 2

Unemployed workers Temporary workers

0 1984 1986 1988 1990 1993 1994 1996 1998 2000 2002 2004 2006 Figure 1.3 Share of temporary workers and unemployed workers in EU-15 1983–2006 Source: Eurostat Statistics (2007).

was 34 per cent in 10 years (from 16 to 21 per cent). Progressive shifts in job insecurity beyond the phase of access to the labour market, as well as a progressive extension of job insecurity over a longer period of careers, have therefore taken place in the last decade. As job insecurity spreads in the population of between 30 and 40 years of age, the tendency of the young generations to remain economically dependent on their family of origin and to postpone plans for autonomy also increases. Studies also show that job flexibility for women correlates negatively with fertility rates (Del Boca and Wetzel, 2007). The fourth critical area is reconciliation of working and childcare. The spread of this problem depends on the constant increase in female employment2 and the growing need for families to have two earners to maintain a satisfactory income. The female activity rate increased in the EU-15 countries by 10 per cent over the decade and in 2006 reached 63 per cent of the total female population of working age. The gender gap narrowed in the same period by a third (from –23 to –16 per cent). The most evident consequence of increased female participation in the labour market is the spread of dual earner households and the parallel reduction in traditional male breadwinner households. This nevertheless triggers strong tensions around reconciling working with childcare. If only the period of life in which maternity and work activity most easily coincide (25–49 years) is considered, the gender gap is around 15 per cent in the EU-15 (with 2 exceptions: it falls below 10 per cent in Scandinavian countries and increases by more than 20 per cent in Southern European countries). Most of this gap (approximately two thirds according to Eurostat

10 Social Vulnerability in Europe

data) is determined by the parallel assumption of family responsibilities. The presence of children in preschool age (1–5 years) pushes the female activity rate down further: the activity rate falls by 8 per cent in the EU-15 for women with one child, by 14 per cent for women with two children and by 30 per cent for women with three or more children (see Figure 1.4). The higher gender gap for women with children is caused not only by mothers working less frequently, but also by fathers working more frequently than their peers without children.3 Therefore, while the female activity rate of women between 25 and 50 years old has increased continuously in recent years in all the EU-15 countries (one point more per year), the gap between activity rates for women with children and women without children has narrowed only in some countries (in the UK, Spain and France, but not in Italy and Germany) and at a rate that is in any case very low (only 2 per cent in 10 years). The constant increase in female participation in the labour market is making the risk of not reconciling working and childcare more common. It is a problem that has negative effects not only on female employment and gender inequalities in the labour market, but also on two other aspects: the exposure of families with small children to the risk of poverty and the increasing demographic unbalance of the European population. Problems of reconciliation are therefore to be considered in relation not exclusively to gender opportunities, but also to the diffusion of other social problems. To consider the first aspect, according to most observers, households organized along the lines of the traditional male breadwinner model are at the greatest risk of poverty (Forster and Mira d’Ercole, 2005). During the 1990s the spread of dual earner households seemed to be the best defence

85 80 75

82.2 77.7

75.3

70 65 60

69

68.2

62

55 50 45

0 child 1 child 2 children 3+ children

51.5 45.6

40 1995

1997

1999

2001

2003

2005

Figure 1.4 Female activity rates for women aged 25–49 years old by number of children in the EU-15, 1995–2005 Source: Eurostat Statistics (2007).

Social Vulnerability in Europe

11

against the risk of poverty and in fact counterbalanced the tendency for economic inequalities to increase (Forster and Mira d’Ercole, 2005). According to Esping-Andersen et al. (2002), an increasing polarization is in fact emerging between dual income and male breadwinner families. This risk mainly affects families in which women have a low level of education and occupational skills. The second problem concerns the relationship between female employment and fertility rates. These two factors do in fact seem to correlate positively as a result of public measures aimed at supporting the cost of children and the diffusion of childcare services (Del Boca and Wetzels, 2007). This, together with cultural factors, explains why the correlation is actually negative for countries in Southern Europe and very low for countries in continental Europe. A substantial aspect of the problem regards the meaning that European citizens attribute to maternity and family responsibilities. According to radical individualism, the reduction in the gender gap marks a clear tendency toward the individualization of social relations and of maternity. From this perspective the presence of generous childcare policies constitutes a strong support for the pursuit of individualistic strategies and for gender equality. Various signals, however, indicate that, together with the quest for greater career parity and economic autonomy by women, there remains a certain continuity in care practices. According to Lewis and Giullari, ‘obligations and practices of mutual support and care continue to bind people together’. This occurs because ‘care cannot be fully de-familialized or commodified because [...] it is emotional and relational, because the pressure for women to care is stronger than it is for men and is part of a gendered identity formation’ (Lewis and Giullari, 2005). According to Crouch (1999), the increase in female employment has had strong impacts on families and has heaped extra pressures on women who work. Despite these greater difficulties in reconciliation, Crouch nevertheless remarks the diffuse tendency to preserve family relations by limiting numbers of children, postponing their birth, involving kinship networks in caring and combining recourse to childcare services and maintaining family relationships. In other words, Crouch sees the reduction in the number of children per woman and the delay in the decision to have children not only as a strategy aimed at reducing the costs of mothering and at lessening stress, but also as an attempt to preserve the quality of family relationships (see Chapter 3 for a further discussion of this point). The last critical area concerns the living conditions of the elderly. The most important implication concerns the increase in the number of dependent persons who require long-term care assistance. Dependency is destined in future to become more common as a result of life expectancy becoming progressively longer. Furthermore it will be concentrated increasingly in the population over 80 years of age. There were approximately 17 million

12

Social Vulnerability in Europe

persons over this age in the EU-15 population in 2007, accounting for 4.5 per cent of the population. The absolute number is nevertheless destined to increase rapidly, partly as a result of more and more people reaching this age. While the overall population will remain stable, the population over 80 years of age is destined to increase by about 3 per cent per year, doubling in approximately 30 years, accounting for 7 per cent of the total population in 2025 (see Figure 1.5). Even if the percentage of dependency will be lower than at present, it is calculated that the dependent population will in any case increase exponentially in the coming decades (Jacobzone, 1999). The family constitutes the primary care resource for dependent persons in all European countries. Oesterle (2001) estimates that informal care covers around three quarters of total care for the disabled in western European countries: it is a percentage that indicates the very secondary role of public protection. In addition to the disparities between poor and rich determined by the residual nature of the public provision of long-term care, new problems are emerging because of a progressive reduction in the informal care provided by families. The reasons for this fact are various. First, the transformations observed in the forms of households, connected with the growing individualization of social life, help to increase the demand for care that cannot be satisfied by members of the same household: the numbers of elderly living alone are in fact increasing in all European countries, while the numbers of the elderly living with their children are decreasing. For example in the UK 55 per cent of persons over 80 years of age live alone, 60 per cent in Germany and

20 18 Proportion of population

16 14 12 % 65–79 % over 80

10 8 6 4 2 0 1977

1980

1984

1988

1992

1996

2000

2004

Year

Figure 1.5 Proportion of people aged 65–79 years and aged 80 years and more in the EU-15, 1977–2008 Source: Eurostat Statistics (2007).

Social Vulnerability in Europe

13

42 per cent in Italy. All countries show an increase in the rates of elderly people living alone. This is set against a progressive decrease in the potential for support from kinship networks. This is determined primarily by the worsening of the old age dependency ratio4 (the statistic currently varies between 25 and 28 per cent depending on the country) as a consequence of the reduction in the effect of the generation turnover. The rise of the numbers of elderly in a situation where the adult population remains stable therefore has the effect of increasing the care load on families. Furthermore, the progressive reduction in the number of children per woman (which fell from 2.4 children for women born in the 1930s to 1.8 children for women born in the 1960s) is reducing the availability of family caregivers. Europe is destined to become the area of the world where the potential for support from kinship networks is most reduced. A second factor that is weakening the caregiving capacity of informal networks is the increase in female participation in the labour market in a context where no significant advances in public homecare services have taken place. The effects of higher female employment on informal care for the elderly are not clear. According to a report on caregiving in six European countries (Lamura et al., 2003), the increase in female activity rates is not significantly reducing caregiving activity when this is for a few hours per week, while it has an appreciable effect on caregiving to persons who require continuous assistance, making home care or institutionalization necessary. Generally, while there is uncertainty over the decline in the number of informal caregivers, it is likely that the total time spent on caregiving is decreasing. These tendencies indicate the need to develop new public long-term care programmes. Some countries have already developed reforms in this direction making long-term care one of the policy fields that is most open to innovation and experimentation. The impact of dependency will be strong not only because of the number of persons in need, but also because of the complexity of the potential negative indirect effects. Ageing is destined to create considerable pressures on traditional patterns of care, putting the organization of families and the integrity of relations between generations under pressure. The presence of a dependent person in-low income families may increase the risk of poverty. The number of elderly dependent persons living alone will increase, causing new problems for community care services. Difficulties in developing public protection systems will increase the need for private services and this will expose the poorest groups in the population to further risk. In other words a social reorganization will take place in the future around dependency. Dependency will simultaneously challenge the integrity of relations between generations within families and the capacity of public policies to provide protection for the most disadvantaged. It will demand the construction of a care system in which it will be

14

Social Vulnerability in Europe

particularly difficult to guarantee social equity and quality standards at the same time.

From risk to vulnerability In industrial societies it was recognized that conditions like unemployment or illness did not depend on individual responsibility, but on factors beyond the control of the individual that had important negative consequences for the whole of society. Since the fundamental element of protection was provided by work, the events that for various reasons prevented a person from working – sickness, accident, unemployment and old age – assumed the status of a ‘social risk’ and involved recognition of the right to public protection. Because these negative events and their frequency were clearly identifiable, they could be analysed, predicted and protected through insurance mechanisms. In the most recent decades the changes mentioned above caused the progressive appearance of ‘new social risks’. According to Taylor-Gooby (2004a), these risks are new in two senses: on the one hand they are spreading progressively, even if they were already present in industrial societies, while on the other hand it is only in recent years that their social, rather than individual, dimension has been recognized. However, the characteristics of these new risks are actually so different from the ‘old social risks’ that they require a redefinition of the notion of social risk itself if they are to be recognized properly. In risk analysis risk is defined as the possibility of experiencing a negative outcome or a significant damage as a consequence of one (or more) factors (called ‘risk factors’). The negative outcome clearly identified in industrial societies was the loss of a permanent job and, as a consequence, of the chance of receiving a secure wage. Social protection against this risk was basically aimed at reintegrating that income which was not guaranteed by an employment position. However, as the cases previously described show, a broad share of the present population is exposed to negative outcomes that do not primarily consist in the loss of a job. New social risks concern a much broader spectrum of negative outcomes that cannot be reduced to the loss of a job and a wage. Rather than on position in the labour market, the new risks depend on the difficult connections between the labour market, household organization and public welfare. As a consequence, while ‘old social risks’ were connected mostly with middle or old age, most of the new social risks affect persons at the start of their working life. They are primarily related to the difficulty of finding a stable position in the labour market and/or assuming caregiving responsibilities in the initial phases of family life. While the old social risks are mostly related to income problems, the new ones, even when they regard problems of

Social Vulnerability in Europe

15

health and ageing as in the case of dependency, do not affect only personal incomes, but also more complex aspects such as housing conditions, solitude and isolation, the integrity of relations between generations and the reorganization of families around caregiving activities. Even new risks arising in the labour market, such as job insecurity, affect more complex aspects than income, such as the possibility of making investments in training and careers, cooperation between family members with different positions in the labour market and the support provided by previous generations to the next ones. It is precisely their positioning in the gap between labour market, family and welfare system that makes public recognition of new social risks very difficult (Taylor Gooby, 2004a). They concern areas of social life that have long been considered a private sphere. It is only in the more individualized countries of Northern Europe that the problems connected with caregiving responsibilities have been considered and treated as ‘social risks’, while they have been heavily under-represented in social policy in other European countries. Neither have they been easily recognized by trade unions and reformist political parties. A second peculiarity of new social risks is that the relationship between causes and negative outcomes is complex and multidimensional. The causes that triggered negative outcomes in industrial societies were clearly identified. They could be reduced down to four basic factors around which the main mechanisms of social protection were constructed: sickness, old age, adult disability and unemployment. The relationship between risk factors and negative outcomes was clear because the labour market was the main mechanism for the distribution of social resources and risks were therefore identified in the points where the labour market malfunctioned. In post-industrial societies individuals participate in the distribution of resources through a number of different channels. A very high percentage of the income of individuals comes from participation in the distribution of public resources. Welfare systems have gradually broadened the range of their beneficiaries beyond the social group of workers and have introduced mechanisms for the distribution of resources that are to a large extent independent of rules that apply in the labour market (Esping-Andersen, 1999). The ageing of the population is also increasing the percentage of people who owe their material survival to public programmes for the distribution of resources. Moreover, changes in the labour market and the increase in temporary employment have weakened the social protection mechanisms grounded on the occupational position of people. As the material conditions of people today depend on a number of different social mechanisms, then there are a number of different possible negative outcomes, which are difficult to correlate with specific causes. This explains why the social impact of risk factors, such as disability, can differ considerably. Since people participate simultaneously in a number of different

16

Social Vulnerability in Europe

resource distribution systems, compensatory mechanisms often come into play, making it possible to rely on a large range of resources when difficulty arises. Social situations in which a number of negative outcomes accumulate are very limited, as is shown by the statistics already reported on the spread of poverty in Europe, while situations in which both negative outcomes and compensatory mechanisms come together are much more common. As a consequence of this fact, new social risks basically arise from the difficulty of coordinating the different mechanisms of resource distribution. For example, the problem of reconciling caring and working emerges from the difficulty of coordinating labour market rules with family organization. Social problems related to dependency often arise from difficulties in combining family support and welfare benefits. Multiple participation in a plurality of resource distribution systems creates problems of coordination among systems that follow very different logic and regulation. The basic social organization that most has guaranteed this coordination has been the family, founded on a gendered division of roles that has facilitated the task of dealing with various social risks. In many situations today, however, the reorganization of families that is in progress makes combined management of different problems more difficult. This causes greater exposure of individuals to the negative outcomes produced by specific risk factors. It is in this aspect that the inadequacy of the traditional notion of social risk is found. In a system characterized by the participation of individuals in a number of different resource distribution mechanisms, the degree of the negative outcomes depends not only on specific risk factors, but also on the functioning of a plurality of resource distribution mechanisms and the capacity of individuals to organize and manage complex situations. Some individuals and families are more exposed than others to greater damage as a consequence of the same risk factor. Disability puts some individuals deeper into difficulty than others. Temporary work has negative outcomes for some workers and not for others, and temporary poverty implies the passage to a condition of permanent poverty for some people, while for many others it constitutes only a passing condition. Risk analysis has introduced the concept of vulnerability to explain how the effect of the same risk factor can be different for equally exposed individuals. According to Vatsa (2004), vulnerability explains the distribution of a negative outcome on a population in relation not to the cause (the risk factor) that determined it, but to the greater or lesser exposure of the population to suffering the consequences of this cause.5 In other words, vulnerability identifies a situation that is characterized by a state of weakness which exposes a person (or a family) to suffering particularly negative or damaging consequences if a problematic situation arises. Vulnerability does not necessarily identify trajectories of impoverishment or social exclusion, but rather a high degree of exposure to serious damage: dependent persons may suffer severe impoverishment if they are alone or have no access to

Social Vulnerability in Europe

17

care services; temporary workers may suffer serious damage if they become sick; a temporarily poor family may fall into a condition of permanent poverty if a member of the family loses his/her job, or if a child is born and obliges the woman to stop working. Social risk in contemporary societies therefore includes two aspects: hazard (the probability of a potential negative situation occurring) and vulnerability (the degree of exposure to damage that may result from the situation). The more the risk factors diversify and the more difficult it becomes to predict the negative outcomes, the more central the dimension of vulnerability becomes in understanding the areas of social disadvantage. New social risks show a final characteristic. Old social risks consist of a well-defined range of events considered as ‘damaging’ or undesired with relative certainty. Exposure to these risks identifies precisely how near individuals are from events that are identified and classified as potentially dangerous. The definition of a risk coefficient corresponds to the attribution of a precise social identification: any person is either ‘normal’ or ‘at risk’, or alternatively in a situation of evident hardship (when the damaging event has already occurred). Since the ‘old risks’ concern individuals in the labour market, it is clear that the exposure of individuals depends crucially on their position in the class system. The same operation seems more difficult to perform for new social risks. In fact it is instability, precisely, that is the peculiar trait in the critical situations we have identified. Consider the examples of temporary workers, people hit by chronic invalidity and families floating above and below the poverty line. These are situations characterized by few social guarantees, by instability in the fundamental mechanisms for acquiring resources and by the fragility of social and family relations. What they have in common is that their position within the main systems of social integration (work, family and the welfare system) is characterized by uncertainty. The effect of social class on these positions appears difficult to assess and will be verified empirically further on (see Chapter 10). It is from the instability of the social position occupied that the notion of vulnerability draws its relevance. Exposure to the risk of serious negative outcomes depends not only on class, but also on a broad set of situations in which people fluctuate (Castel, 1995). Fluctuation occurs in various ways: horizontal mobility between different jobs, flexibility in work and family roles, uncertainty over the position occupied, absence of welfare guarantees and difficulty in reconciling and coordinating different roles and responsibilities. While on the one hand such fluctuation opens up the possibility for many individuals of ‘building their own biography’ (Beck, 1992), on the other hand it contributes to social instability and difficulties in being independent. To summarize, the notion of social vulnerability identifies not only specific risk profiles, but also the nature of the risks themselves. They have

18 Social Vulnerability in Europe

shifted from situations in which the relationship between causes and negative outcomes was clearly identifiable into conditions characterized by unforeseeable varying degrees of exposure to possible damage depending on a complex set of risk factors. They have changed from situations that were clearly identifiable by observing the employment position of individuals into situations that are characterized by a multiplicity of resource distribution mechanisms. Finally they have transformed from relatively stable situations into situations characterized by uncertainty. The spread of new social risks therefore brings out the importance of social vulnerability. This is characterized by an uncertain access to fundamental material resources (a wage and/or welfare benefits) and/or by the fragility of the family and community social networks. It is characterized not only by a resources deficit, but also by an exposure to social disorganization, which reaches such a critical level as to put the stability of everyday life in danger. It takes the form of a life situation in which autonomy and the capacity of individuals and families for self determination are threatened by the introduction of uncertainty into the main systems of social integration. The instability of the social position does in fact translate into a reduction of opportunities in life and of possibilities for choice. It is characterized not so much by the scarcity of resources tout court, as by the instability of the mechanisms used to obtain them.

The crisis of household functioning Vulnerability can be described by referring to the notion of functioning (Sen, 1985; 1987). According to Sen functioning relates to the ways people have freely adopted of pursuing a state of well-being to which a specific value is attributed. Capabilities are the various combinations of functioning that a person can achieve. They are not the means needed for well-being (e.g., the availability of certain primary goods, to use Rawls’ notion), but the ‘things which constitute well-being’, ‘freedoms actually enjoyed’ (Sen, 1985), which can vary from basic things like being properly fed and in good health or escaping preventable ill health and premature death to more complex things like being happy, having self-respect, taking part in community life (Sen, 1985). From this perspective social vulnerability may depend on the scarcity of means and/or on the difficulty experienced in converting available means into capabilities. According to Sen it is precisely this last problem that explains the paradox of the large numbers of people in poverty in affluent societies: it does not originate solely from the unequal distribution of resources, but also from the fact that some functionings in advanced societies are very complex to manage. From this perspective social vulnerability constitutes a situation characterized by the presence of objective obstacles to the conversion of available

Social Vulnerability in Europe

19

resources into basic capabilities: even before it becomes a deprivation factor, the absence of stability is a factor that depresses the functioning of persons, limiting their freedom to achieve or their chances of converting resources into capabilities. The most appropriate scale of observation at which to reconstruct social vulnerability seems to be the household. The priority given to the household in this book depends not on a theoretical assumption but on an empirical convenience. The household is in fact the basic unit for collecting and distributing resources and converting them into well-being. First, the activities of collecting fundamental material resources (income, housing) change according to how households are composed and how roles are divided within them. Second, redistribution of resources occurs on the basis of family roles and the consideration given to individual needs. The household structure is therefore crucial in determining both the amount of resources available to individuals and their degree of economic protection. It must also be considered that the household constitutes the main channel for access to many welfare benefits in a number of European countries. Third, the household is the principal social channel through which people in need are provided with care. Again the performance of this activity depends on the role structure and the internal organization of households. Even though Sen defines ‘functioning’ as a specific way of using resources that can vary from person to person, it is nevertheless possible to identify some functionings on which to concentrate our attention. Our hypothesis is that three fundamental functionings of households can be identified.6 These functioning mechanisms operate on increasingly complex levels of household organization, in the sense that by moving to successive levels the problems connected with the previous level appear again in a new form: functioning 1: acquisition and use of basic resources that are necessary for the material survival of household members; functioning 2: management of major life events (job seeking, forming a family, birth of children, old age) that preserves the material survival of the household; functioning 3: provision of social care for dependent members (children of pre-school age or disabled persons) while material survival (and/or management of major life events) has to be guaranteed. Specific situations of vulnerability can be identified for each of the three levels of functioning (see Table 1.1). The first level of functioning regards obtaining the basic acquisitions that are needed in order to guarantee the material survival of a family. Fundamental needs to satisfy concern housing, income and work. It is not only a question of ensuring an adequate flow of resources (as many analyses of poverty assume). The material survival of a family over time does not imply only

20

Social Vulnerability in Europe

Table 1.1 Vulnerability and precipitation/protection factors for different level of functioning Family functioning

Situations of vulnerability

Precipitation/ protection factors

Material survival

Economic vulnerability, scarcity and fluctuations in income (Chapter 4) Job instability, lack of career continuity, unemployment (Chapter 5) Housing deprivation (Chapter 6)

Position in the class system

Management of major life events

Difficulty in the transition to adulthood (Chapter 8) Family organization during course of life (Chapter 3)

Family ties

Reconciliation of caregiving and work responsibilities

Reconciliation of childcare and work (Chapter 3) Caring for a dependent person (Chapter 7)

Extent and degree of welfare protection

acquiring an adequate income (combining incomes from different sources), but also creating a minimum stock of resources that makes it possible both to procure the minimum resources necessary (a home, a car and fundamental services) and to protect against future misfortune. From this viewpoint stable employment has been longly a key factor. In short, the acquisition of a stable income, a stable job and permanent housing constitute crucial achievements not only for the financial survival of a household but also for its organizational stability. As a consequence the principal factors of vulnerability consist of income instability (including the possibility of falling, even temporarily, into poverty), job precariousness and housing insecurity. The stability of material acquisitions is the basis allowing individuals and households to successfully manage the different stages of their course of life.7 Nevertheless there are critical phases that often require a basic reorganization of the household. The second level of functioning regards household organization designed to find a balance between activities that are necessary for material survival (the first level of functioning) and activities that are needed for managing critical life events, such as the transition of young people to adult life and the ageing of the members of the household. This second level of functioning appears more complex than the previous one. It is in fact a question of maintaining organizational stability over time, which allows both material survival and the management of relevant changes to be made at the same time. The most problematic phases in the course of life today seem to be the transition to adult life on the one hand, and the organization

Social Vulnerability in Europe

21

of everyday life at a very advanced age on the other: two critical situations that reflect the demographic pressures (low birth rate and delayed childbirth; high rate of ageing) to which many European countries are subject. Finally the third level of functioning regards the organization of the household that is adopted to care for children or dependent persons. The organization of care puts the household under strong pressure and implies adaptation both in the basic mechanisms guaranteeing the household’s survival (first level of functioning) and in the management of specific life events (second level of functioning). The main problem concerns the caregiving burden, which in most European countries is taken by households, to the point that strong financial and organizational tensions undermine the stability of families (see Chapter 7 for further details on this point). The recent increase in public measures supplying payments for care and home care services still leaves many responsibilities with households. Furthermore, the ageing population and the weakening of family ties increase the number of dependent people who cannot rely on a stable and intensive family support. Therefore both the presence of households that are able to take on caring duties and their absence constitute situations of high social vulnerability.

The role of social policies The role of public policies in structuring social vulnerability is the last key point in this book. The current debate on new social risks has led to two principal acquisitions over which there is a general consensus. A deeper analysis of these statements, however, shows a certain ambiguity in the interpretation. On the one hand it is underlined that traditional welfare systems are not properly equipped to provide adequate protection against new social risks. In this sense new social risks show the limits to growth and adaptation of modern social protection systems. On the other hand, it is underlined that the appearance of new risk profiles indicates the presence of some chances for innovation in these welfare systems (Taylor-Gooby, 2004a). According to this perspective, a better awareness of the limited capacity of current welfare systems would solicit a reformist position, which would be oriented not toward welfare retrenchment but on a restructuring strategy aimed at better tackling new social risks. A corollary (not secondary) of this position is that Scandinavian welfare systems are better equipped to respond to the new risk profiles thanks to their universalistic principles, while corporatist and familistic systems are in the most difficulty. Therefore greater attention to new social risks would imply a marked change of direction for these systems and the construction of universalistic systems that are far removed from the meritocratic approach typical of continental welfare systems (Esping-Andersen et al., 2002). Two issues, which are to be considered separately, emerge from this discussion. The first question regards the causal effect of current welfare systems

22

Social Vulnerability in Europe

on social vulnerability: the diffusion and character of social vulnerability varies in fact as a function of the degree of development and the differences in national and regional welfare systems. The second question concerns the capacity of welfare systems to react to new forms of social vulnerability. In this case the question is not so much the degree of cover for social vulnerability as the capacity of current systems to innovate and adapt to current social and economic changes. The first point in question is addressed in this book through a specific analysis of the factors that contribute to defining geographical (and sometimes national) profiles of social vulnerability. In this perspective welfare policies constitute one of the main factors in protection from vulnerability, together with other factors such as family and class structures. The hypothesis supporting this analysis is that, other factors remaining constant, the welfare system contributes to the structuring of peculiar vulnerability profiles through selectivity in the access to benefits and the generosity of the benefits distributed. It is therefore hypothesized that different models of public support contribute, together with other social and economic factors, to creating different levels of social vulnerability in different geographical areas.8 The second point in question regards the capacity of policies to react to the spread of social vulnerability. The questions posed are: by contributing to the creation of differentiated vulnerability profiles, are European welfare systems able to develop responses that are also adaptive? Which systems seem to be able to give adequate responses most easily? Which problems are addressed most adequately by which welfare systems? This research is able to give only indicative answers to these questions. They concern the capacity of current welfare systems to innovate. It is innovation that is not driven solely by the demands of austerity and financial compatibility, but also by the internal limits in adequacy and effectiveness of current social protection systems. Continental and Southern European welfare systems seem to be the most challenged by the issue of social vulnerability. According to Taylor Gooby (2004a) their corporatist structure does not acknowledge the new social risks and makes them unable to provide adequate measures to protect against them. The presence of strong coalitions to defend insider interests also makes political representation of new social risks and the inclusion of new needs on the political agenda difficult. Therefore innovation must develop outside the traditional ‘world of welfare’, through the introduction of new welfare measures that remain on the margins and are tailored to meet the needs of very specific social targets (Palier, 2000). Quite apart from institutional factors that may explain the greater or lesser ability of welfare systems to adapt to social vulnerability, the structure of relations between the public and the private spheres is also in question. Social vulnerability emerges from problems that are connected with the

Social Vulnerability in Europe

23

critical transition to a post-industrial society. Its development sets the problems of the connection of the labour market with household organization, life transitions, care needs and the spread of social instability at the centre of the problem. These are fields where welfare systems have historically intervened only with residual programmes, delegating the solution of many problems to the spontaneous market or to the family. Even today, many emerging problems in these areas are perceived predominantly as problems of market functioning and/or private solidarity. When state intervention is in place, it assumes the logic that is typical of activation polices and active integration in the labour market: personal responsibility and activation are the key words in a perspective that delegates fundamental responsibilities to individuals and to market functioning. The main limitation of these policies is that the responsibility of providing answers to social vulnerability is entrusted entirely to the supposed inclusive capacities of the labour market and/or of the family. The following reconstruction of the complex and differentiated profiles of vulnerability may help in an understanding of how policies designed to strengthen participation in the labour market of vulnerable population, or to support the caring capacity of families, must be combined in future with a broader set of social policies, including social care, housing, health and family policy.

Notes 1. The proportion increases for Spain (52 per cent) and for Italy (46 per cent) while it comes to less than one third for Germany, Sweden and the United Kingdom. 2. The female employment rate in the EU-15 countries was 50.2 per cent in 1996 and 57.4 per cent in 2005, with constant growth during the entire period. The growth rate in Northern Europe appears to be lower in recent years, while it is faster in Southern European countries. 3. According to Spiess et al. (2004) the correlation between the presence of children and working mothers does not clarify the causal relationship between having children and working, because the employment situation of mothers could predate the child’s arrival. By developing a longitudinal model they discovered that the effect of the presence of a child is found only in the Germanic continental area. 4. While the old age dependency ratio (which measures the ratio of the elderly population in relation to the population of working age) improved in the 1980s and 1990s when the baby-boom generation entered active life, it started to worsen rapidly in 2000 when longevity increased and the birth rate fell. In continental and Southern Europe it increased from 23 per cent to 27 per cent in a decade (from 1995 to 2005). The consequence is that it is increasingly more common to find four generations active simultaneously. This will have increasingly greater negative effects on the allocation of time and economic resources and on satisfying the care needs of the elderly population. 5. The clearest way to explain the meaning of the concept of vulnerability is the case of a population exposed to a natural disaster (like an earthquake or a flood). The

24 Social Vulnerability in Europe negative outcome caused by the event is distributed across the population hit not on the basis of the probability of being exposed to the cause, but on the basis of social and economic variables, as was illustrated by the tragic case of Hurricane Katrina which hit New Orleans in 2005. 6. It must be borne in mind that households are treated here as homogeneous units, without considering them as systems of personal relationships exposed to internal and external tensions. 7. Various studies now demonstrate the utility of taking into account the life histories of individual cases (Leisering and Leibfried, 1999) when analysing situations of hardship (poverty, unemployment or insecure employment, etc.). This can be done both by adopting longitudinal analysis and by considering specific crucial phases in the course of life. 8. The analysis in Chapter 10 not only compares the importance of policy variables with respect to socio-economic structure variables in determining the different vulnerability profiles, but also the extent to which the classification of EU welfare systems into ‘welfare regimes’ (Esping-Andersen, 1990) provides a useful heuristic model of social vulnerability in Europe.

2 Bringing Territory Back in Social Comparative Research Costanzo Ranci

Introduction This book analyses the configuration of social vulnerability in the area of Western Europe that is conventionally known as the ‘Europe of the 15’ (EU-15). Most of the comparative studies on inequality and poverty have taken individual countries as the unit of analysis, on the basis of a number of often implicit assumptions. Comparative analysis between countries based on these assumptions has also developed supranational units (like ‘welfare regimes’) that include whole nation states regardless of their internal differences. This chapter is aimed at discussing this nation-statecentred approach and at presenting the concepts and methodology that are proposed and used in this book in order to overcome the main shortcomings of the traditional comparative approach.

Regional disparities in Europe and comparative research European comparative analysis has almost exclusively considered nationstates, even though for several decades – under the impetus of decentralization policies in the 1970s and the ensuing construction of a European political space (Ferrera, 2005) – the autonomy and the institutional recognition of European regions have grown. While the European regions have acquired increasing importance in relation to both the European Commission and intra-national institutional relations, and despite the fact that several European countries have federalist structures of government, comparative research has remained strongly anchored to the nation-state as its principal unit of analysis, with a few exceptions. This situation is substantially unchanged despite the increasing availability of information and data that are collected on a regional or sub-regional scale.1 The main reason why the comparative approach has focused on comparison among countries is the central role performed by the nation-state in the institutional construction of citizenship and political representation, and 25

26 Social Vulnerability in Europe

in the development of the public regulation of economy and society. In the analyses of Rokkan (1970), as well as those that have subsequently identified the existence of diverse European ‘regimes’ (Esping-Andersen, 1990), the political-institutional dimension has always been of central importance in explaining the territorial differentiation of European societies. It is in fact mainly the result of the regulatory action of the nation-state and of the institutional definition of national boundaries if European countries have constituted territorial entities endowed with an internal social and economic homogeneity that makes differences within countries less marked than those between countries. According to Ferrera (2005), the building of the nation-state (and therefore of the principal systems of public protection against social risks) came about through a structuring process that defined national boundaries in both spatial and social terms. The outcome was the construction of national systems united by a bounded solidarity: ‘the fusion between territorial control and identity, mass democracy, and the welfare state produced very solid and highly integrated political systems, functioning according to distinct internal logics’ (Ferrera, 2005: 25). In the past two decades several phenomena have directed attention more closely to the regional dimension. The first factor has been political-institutional in nature. The process of European integration, together with the more general dynamics of globalization, has given greater importance to regional governments (Keating, 1998). The growth of regions has been driven by two processes of institutional adjustment. First, the growth of ultra-national regulatory powers (since European integration) has restricted from the top downward the room to manoeuvre available to national governments. Second, rescaling processes have circumscribed the governance of social policies to more limited territorial areas (Brenner, 2004). Therefore regions have acquired new responsibilities for contrasting reasons: on the one hand, the new regionalism has been the direct consequence of imperatives intended to increase the fiscal responsibility of sub-national governments and accordingly to reduce the national fiscal burden (Pierson, 2001; Ferrera 2005); on the other, the local and regional level has often seemed more appropriate in developing social policies able to deal with new social risks than the state level (Taylor-Gooby, 2004a). These two processes have generated not only the relative autonomy of regional institutions but also the possibility of more marked territorial differences in paths of social and economic development. While the dominance of the nation-state coincided with a historical phase of decreasing territorial differences, the new regionalism has obstructed the homogenizing thrust of the nation-state and given rise to new social and economic inequalities (Brenner, 2004). The emergence of marked territorial inequalities is a second factor that solicits closer attention to the regional scale. Although social inequalities are still analysed and discussed within national boundaries, it is increasingly

Bringing Territory Back in Comparative Research 27

evident that they also depend on factors connected with the international division of labour. At the global level, inequalities concern not individuals but regions inasmuch as the territory is the crucial basis for the definition and expression of economic interests. On the other hand, it is precisely the enlargement of the EU that requires this analysis, given that today it must address an increase in regional inequalities within nation-states matched by a gradual decrease in inequalities among them (Jesuit et al., 2002; Kangas and Ritakallio, 2004; Heidenreich and Wunder, 2008). According to Heidenreich and Wunder (2008), the regional inequalities within states in the enlarged Europe have increased by 15 per cent over the last 8 years, while the between-nation inequalities in Europe have fallen by 45 per cent. The homogenizing action of European integration is being resisted not only by political-institutional factors but also by the structural characteristics of territories connected with social and economic specificities: ‘different regional employment patterns, industrial structure and the regions’ location within the European space’ (Heidenreich and Wunder, 2008: 19). There are progressively more studies, made possible by the availability of large regional databases, reporting the persistence of pronounced regional differences among inequality and poverty indicators. Berthoud (2004a) observes that the level of inequality within European countries is higher for countries with a lower median income. It thus seems that inequality and poverty are explained less by the traditional distinction among different welfare regimes than by the degree of economic development and the general level of well-being of the population. The countries of Southern Europe exhibit marked regional differences in levels of well-being, while countries like Germany and France are much more homogeneous.2 Differences among nation-states in levels of well-being and inequality are therefore mainly caused by variations among rich and poor regions within the countries themselves (Stewart, 2005). These analyses highlight the shortcomings of conventional comparative analysis centred entirely on variations among countries. If variations within countries are considered, in fact, certain differences among nation-states traditionally attributed to the existence of diverse welfare regimes are better explained by the existence of marked regional differences. These appear to result more from structural social and economic factors than from the different regulatory systems of welfare policies. These studies suggest some final considerations. First, comparative analysis among nation-states (and the consequent identification of supranational clusters) is based on the assumption that the central state has played, and still plays, a fundamental role in the social and economic homogenization of the territory. A large body of empirical evidence, however, shows that this homogenizing action has not eliminated territorial differences and that, rather, territorial differences have recently predominated over the standardizing action of the nation-state. This suggests that attention should

28

Social Vulnerability in Europe

be paid to territorial differences within countries – without taking it for granted, as has often happened in European comparative analysis, that the national level is best suited to grasping the specificities of economic and social structures. Second, in more analytical terms, attention to the territorial dimension requires more careful testing of the assumption that territorial differences result mainly from the diversity of political-institutional regimes rather than from the structural characteristics of the territories themselves – including cultural aspects connected with language, religious identity and geographical location.3 Hence there are several reasons for ‘bringing territory back’ into comparative analysis. The persistence over time of regional differences confirms how strong is the influence exerted on these local systems by their structural characteristics. Nevertheless the regional dimension has received little attention in analyses of poverty and material deprivation in Europe. So far, social indicators suitable for monitoring the living standards of the European population have not been developed on a regional scale, with the sole exception of the unemployment rate. Nonetheless, Atkinson et al. (2002) have recommended that all the social indicators relevant to European policies should also be developed on a regional scale. The necessity to consider the presence of territorial differences in the distribution of poverty and vulnerability stems from the problem that these phenomena are concentrated in specific areas, and not only in individuals or households. According to Stewart (2003), traditional analyses of poverty and vulnerability have focused on people poverty rather than on place poverty: People poverty occurs where individuals are poor due to personal or household characteristics and not area conditions. If they are found in clusters, this is because their incomes affect where they live, rather than the reverse. In contrast, place poverty is what emerges when area circumstances, such as the quality of public services, contribute to poverty outcomes; that is, when people are poor because of where they live [...] it seems little short of absurd to maintain that area conditions are irrelevant in driving deprivation, particularly once we consider a wider view of exclusion, covering health and education indicators as well as income poverty. (Stewart 2003: 337–338) Paying attention to the territorial distribution of social vulnerability serves three main purposes. First, it is aimed at determining in what territorial areas social vulnerability is most concentrated. A territorial map of vulnerability emerges from this analysis detecting regional contexts where a peculiar concentration of social risks is caused by specific contextual factors. Second, specific territorial patterns can be identified in which distinct risk factors are interconnected by their common territorial origin. This

Bringing Territory Back in Comparative Research 29

analysis is not only aimed at ranking territories, but also at bringing out the specificities of each of them, as well as identifying the principal reasons for specific territorial patterns of social vulnerability. Third, such analysis would make it possible to determine the extent to which territorial patterns of social vulnerability reproduce differences among welfare regimes in the EU-15. While the substantial permanence of the welfare regimes identified by Esping-Andersen seems confirmed (Castles and Obinger, 2008), one may wonder whether the structuring of the new social risks corresponds to those models.

How to split nation-states in macro-regions: the methodology There are two main methodological obstacles to comparative analysis at a scale lower than the national one. The first is the difficulty of identifying a sufficiently homogeneous territorial level below the national level. Having questioned the homogeneity of the territory included within national boundaries, the problem recurs in the same terms at a more reduced scale. The term ‘region’ in fact denotes a mere administrative border that does not always correspond to a homogeneous social and economic structure. The criteria (including those related to size) used to mark out regional territories, like those developed by Eurostat (through the NUTS 1, NUTS 2 and NUTS 3 levels), are rather variable,4 as well as being purely administrative. Finally, administrative regions often have structures very similar to those of adjoining regions, owing to broad areas of hybridization and overlap. The second obstacle concerns the comparability of regional areas rather different in terms of their geographical extent and population sizes.5 Analyses conducted at sub-national level have often adopted administrative boundaries without any particular concern about their significance, treating them as simple statistical units. Nevertheless, it is evident that the final results differ according to the geographical scale used (Berthoud, 2004a), and that there are no ‘true’ criteria with which to define the appropriateness of that scale. Berthoud (2004a), for example, analysed regional poverty levels by considering the regions identified by the Eurostat NUTS 1 code, aggregating only contiguous areas with fewer than a million households, and avoiding the breakdown of the smaller European countries to a sub-national level. Stewart (2002) used the NUTS 1 level to overcome problems of sample size, even if the NUTS 2 level is considered more appropriate in general. Heidenreich (2003), although he defined a region as a sub-national territorial unit characterized by relatively homogeneous demographic, economic and socio-cultural patterns and a governance structure, adopted the NUTS 2 level without particular verification of internal homogeneity, and consequently identified 265 European regions.

30 Social Vulnerability in Europe

It is therefore clear that a homogeneous sub-national area cannot be defined on the basis of objective criteria. Administrative definitions are manifold and diverse and they follow criteria which do not always concern structural homogeneity. Moreover, the notion itself of territorial homogeneity is problematic. While on the one hand this concept refers to a historical continuity marked by specific economic and social structures, on the other hand it is also the result of public regulation, both local and supra-local or national, determining differentiated access to protection against social risks. Rather than using an ‘objective’ notion of a homogeneous area, therefore, it is more appropriate to rely on specific analytical features. The issue, therefore, is how to identify the most appropriate geographical scale on which to reconstruct uniform social conditions that enable comparable social vulnerability profiles to be ascertained. Since it is the relationship between the labour market, household structure and the organization of welfare systems that generates social vulnerability, it is through these aspects that the existence of homogeneous geographical contexts must be verified. For the reasons already given, both the national and the regional scales identified by Eurostat are inadequate. A possible third way is based on the notion of ‘territorial social formation’, by which is meant a geographical area characterized by a social structure with a distinctive configuration. A good example of this approach was developed by Bagnasco to analyse economic and social development in Italy. Bagnasco (1977) identified three ‘territorial social formations’ with high internal homogeneity and a distinctive socioeconomic structure. These social formations consist of aggregations of Italian regions created on the basis of socio-economic indicators and historical and anthropological investigation. According to Bagnasco, a territorial social formation is the expression not only of a collective identity but also of a contrast with alternative and often geographically contiguous systems. This book has applied the concept of ‘territorial social formation’ on a European scale. The main criteria on which the analysis was based were as follows. First, ‘territorial social formations’ on a scale intermediate between internal regions and nations were identified only for the five largest European countries (France, Germany, Italy, Spain and the United Kingdom), while the national scale was adopted for the other countries. This solution made it possible to have a similar number of survey cases for all the macro-regions. In addition, the division of the five largest countries was performed within the national borders of each nation, the purpose being to determine whether or not there was nationwide social homogeneity. Second, macro-regions were identified by agglomerating the internal regional units (NUTS 1 level except for some states) within each nationstate. For each of the latter, a cluster analysis was performed among the NUTS 1 areas in order to identify the areas most homogeneous in terms of their socio-economic structures (see Appendix for details). The criteria used for the cluster analysis were identical in all the countries considered. They

Bringing Territory Back in Comparative Research 31

Figure 2.1

Macro-regions in the EU-15

identified four different dimensions of social and economic structure: the degree of economic development (per capita GDP), the employment structure and family organization (female activity rate), the demographic structure (old age dependency ratio) and the urban dimension (population density). Finally, 28 territorial social formations were identified (see Figure 2.1) and this new geography of Western Europe was subsequently used to reconstruct broader geographical configurations of social vulnerability. By identifying the social distance between different European macroregions, this analysis sought to overcome the traditional limitations of comparative analysis anchored to a vision of national membership, which the birth of a ‘European space’ makes less necessary for both social analysis and public policy (Grasland, 2005).

32 Social Vulnerability in Europe

Appendix: Notes on the construction of macro-regional clusters The macro-regional clusters for the five largest countries of Western Europe were constructed on the basis of the following variables: ● ● ● ●

the level of urbanization (density) the degree of socio-economic development of the area (per capita GDP) the demographic structure of the population (old age dependency ratio) the occupational structure (female activity rate).

The clusters represented territorial aggregations of NUTS 1, and they took the form of macro-regions. Some constraints were introduced which forced the analysis: a. b. c.

NUTS 1 were aggregated on a national base: it was not possible to aggregate NUTS 1 belonging to different countries; geographical proximity constraint: NUTS 1 not geographically contiguous could not be aggregated; the households in the ECHP database for each macro-region had to number at least 1,000 cases.

The constraints on geographical proximity and on the number of households for each cluster were verified a posteriori. Then a cluster analysis was conducted for each country: this fixed the number of clusters so as to comply with the criterion on the number of cases, selecting the clusters for which the contiguity condition was fulfilled. In other words, it was the result of the cluster which suggested the aggregation of adjacent NUTS 1 with an adequate number of households, not vice versa. Two parallel analyses were conducted to construct aggregations respecting these parameters. The first produced hypothesis A on the clustering of European countries. This analysis involved, after standardization of the data, the construction of a proximity matrix. The latter was constructed by summing the squared deviations among the standardized values of each NUTS 1 with those relative to the other NUTS 1 for every variable. In other words, all four variables and all the NUTS 1 of each country were made to interact simultaneously. For example, NUTS 1 in Italy are: Lombardy, North-West, North-East, Emilia, Centre, Lazio, Abruzzo, South, Campania, Sicily and Sardinia. The standardized value of Lombardy for the variable ‘density’ was subtracted from the standardized value of the North-West for the same variable; the square of the difference was added to the squared difference between the standardized value of Lombardy for the variable ‘GDP’ and that for the North-West, and likewise for the other two variables. Thus obtained was

Bringing Territory Back in Comparative Research 33

a synthetic index representing the distance between Lombardy and the North-West. The same operation was performed between Lombardy and the other Italian NUTS 1s, and subsequently for every NUTS 1. The result was a similarity matrix based on the quadratic Euclidean distance between the variables of each NUTS 1 with those of all the others. The NUTS 1s were then aggregated, starting with the lowest cells, constantly maintaining the territorial contiguity constraint, and ensuring that the aggregation enabled the clustering of at least 1,000 cases in the ECHP file. This procedure produced a first map of Europe containing 24 clusters. The second analysis was conducted by producing a hierarchical cluster and visualizing the results with a cartographic software program (hypothesis B). Since the size of NUTS 1s is not homogeneous, some NUTS 1s are never clustered (this being the case, for instance, for metropolitan areas like London, Paris, Berlin, Bremen and Madrid). To overcome this drawback, it was decided to perform an analysis at a European scale, considering all of the final outcomes with a range from 20 to 30 clusters. Thus highlighted were the NUTS 1s that proved particularly difficult to cluster with others. A cluster analysis was then performed at the country level. The behaviour of the previously unclusterable NUTS 1s was observed: while some joined other NUTS 1s in the same country to constitute a cluster, others remained isolated even in the country-level analysis. In this latter case, it was decided to exclude these cases from the analysis and to form a single group of unclusterable NUTS 1s. The final map contained 28 clusters. Finally, the two solutions were compared. The clusters that resulted identically in both the analyses were confirmed. When different clusters emerged, it was decided to keep the clusters identified by each analysis distinct. No NUTS 1 was attributed to two different clusters by the two analyses. The attribution choices adopted were the following. Italy: analysis A identified three clusters, distinguishing among NorthWest, Centre North-East, and South, while analysis B identified only North (including the Centre North-East) and South. The final choice was to keep distinct the three clusters produced by analysis A. Germany: analysis A identified a North cluster (not identified by analysis B), while analysis B identified a North-West cluster (not identified by analysis A); both of these clusters were added to an East cluster and a South cluster, identified in both the analysis. Spain: both analyses identified a North-East cluster and a cluster corresponding to the Madrid area. For numerical reasons, the two clusters were merged, thus breaching the territorial contiguity constraint in this case. Finally, a South cluster and a North cluster were identified. France: analysis A aggregated Mediterranean France with Central-Eastern France, while analysis B aggregated it with West France. It was decided to

34

Social Vulnerability in Europe

distinguish a North-East France, North-West France, and South France (as well as the Île-de-France, which was identified by both analyses). United Kingdom: analysis A identified four clusters, distinguishing Greater London, North, South-West and Centre-West areas; analysis B identified Greater London, a small cluster in the Centre-West area and the rest of England. The results of analysis A were chosen, in that they entirely comprised the results of analysis B. The final outcome consisted of 28 clusters.

Notes 1. Eurostat, for example, has long made available a large number of economic and social indicators on different territorial scales, ranging from the national scale (NUTS 1) for some countries to the regional (NUTS 2) or sub-regional (NUTS 3) ones. 2. The existence of marked regional differences within the countries of Southern Europe is confirmed by Stewart (2002), whose study of a large number of well-being indicators shows clear territorial demarcations in Italy and Spain. See Förster et al. (2002) about regional disparity in Central and Eastern Europe. 3. According to Castles and Obinger (2008), in past decades there has been no convergence within the EU among the main national welfare regimes, as one would expect to have happened following the process of European integration. Rather, ‘within the new boundaries of the EU, all the previously existing families of nations are present, but are now joined by a still more distinctive post-Communist family’ (Castles and Obinger, 2008: 23). While on the one hand this can be explained by the evanescence of the EU’s social programmes, on the other, ‘the persistence of policy clusters is, to a significant degree, a function of the persistence of underlying structural characteristics’ (Castles and Obinger, 2008: 23). 4. As exemplified by the differences in size between the British counties and the French departments, on the one hand, and the German Länder and the Italian regions on the other. 5. This was already problem for comparative analyses among nation-states, where countries with 70 million inhabitants like Germany were compared with countries of 8 million inhabitants like Denmark.

Part II Aspects of Social Vulnerability

This page intentionally left blank

3 Beyond the Male Breadwinner Family Model Emmanuele Pavolini and Costanzo Ranci

Introduction The structure of the family has gone through a phase of reorganization in most European countries (with the exception of Scandinavian countries1) in the last 15 years. The fundamental point of tension concerns the crisis of the family model based on a single wage earner (known as the male breadwinner model, MBW) which has brought about the emergence of new family models, discussed and catalogued in the literature by various authors. The crisis and decline of MBW families is now clearly documented at European level by various research studies (Lewis, 1993; Sainsbury, 1996b; Pfau-Effinger, 2005). Even if a certain persistence of MBW family models has been observed in Mediterranean countries (Blossfeld and Drobnic, 2001), the dynamics in recent years clearly show that these countries are also taking the path of the progressive decline and crisis of this type of family, although with some delay (Naldini, 2003). However, this does not mean that the model of family organization based on gender division of working and domestic roles is disappearing. It is more of a transformation than a disappearance. The theoretical point in question here concerns precisely how the crisis of the MBW model can be interpreted. The aim of this chapter is to document the forms of organization that European households are adopting in response to the changes in progress, considering the existence of very different historical traditions and systems of national regulation. It attempts to highlight the differences between macro-areas, to identify the main trends and to discuss the main points of tension. Attention is focused on three fundamental dimensions in which the reorganization of households is occurring: a.

the demographic composition of households: various factors, including demographic changes (ageing, the reduction in birth rates) and the de-institutionalization of the family (increase in cohabiting couples and 37

38

b.

c.

Social Vulnerability in Europe

children born outside marriage, growing instability of marriages), have determined both the diffusion of new forms of families and the transformation of traditional forms; the income structure of households and their position with respect to the labour market: there is a progressive differentiation in the income structure of households caused by changes occurring in the labour market and in welfare systems; numbers of single earner families are falling, while there is a parallel diffusion of dual-earner families, with differing organizational models; the gender arrangements adopted by households with children in order to reconcile caring with working: the organizational models vary basically according to the working role of the woman and to the ‘gender contract’ (Pfau-Effinger, 1998a) existing between the partners; determining factors here include the original cultural models and the institutional structure of opportunities (welfare measures, economic incentives for part time work and so on) within which organizational decisions are made.

Explaining family changes: between individualization and family solidarity A very widespread explanation in the theory sees individualization as the fundamental process that has driven the reorganization of households beyond the MBW model (with the general outcome being a decline of the social functions of the family). It is claimed that this process determines specific outcomes for family organization: ●

●

●

a strong spread of single-person households and the spread of families that are composed by members of the same age (only young people, only old people); the increase in the female activity rates and the tendency toward gender equality in employment give rise to the spread of dual income families and, more generally, of an adult-worker model family (Lewis, 2002); an increase of defamilization and commodification of care as a consequence of the search for reconciliation between working and childcare.

An alternative model proposed by Bengtson (2001) for the USA (see also Micheli, 2004 for Europe) is the ‘family solidarity approach’, based on the idea that intergenerational links are becoming increasingly strong, giving families great resilience and capacity to adapt, to the point that an increase in familism seems to be spreading in the USA as a useful strategy for dealing with ageing, the reconciliation of work with care and the crisis of singleearner families. It is claimed that the consequences of this possible trend

Beyond the Male Breadwinner Family Model 39

are as follows: ●

●

●

the spread of bi- or multi-generational families, going with the maintenance of intense inter-generational ties (e.g., as a consequence of the tendency of the elderly to return to live with their children), and a halt in the previous trend toward the growth of one-member families; as a consequence of the increased working activity for women, families based on a ‘one and a half’ (OAH) model, or based on a combination of one permanent job and one flexible job, are increasing; there is an intensive use of the resources of the extended family to reconcile childcare and working.

Comparative research has already shown that family structures in EU-15 countries have a wide variety of different starting points (Saraceno et al., 2005). It is clear in this respect that Northern European countries are closer to an individualistic model, while Southern European countries are heavily characterized by stronger family solidarity. It is nevertheless less clear what changes have taken place in recent years and what the impact of the crisis in MBW families is in continental and Southern European countries. The idea of a progressive individualization in European society brings about the parallel thought that a convergence of the ‘Atlantic model’ and the ‘Mediterranean model’ is taking place, while other interpretations (Micheli, 2004) stress the strong opposition and persistence of multiple, locally and culturally based family models. A broader interpretation is proposed by Crouch (1999). According to Crouch, the mid- century social compromise consisted of some basic facts: a high female presence in the home, young marriages and a high marriage rate, high fertility rate (higher than 2:1) and low divorce and illegitimate birth rates. Today this model is becoming obsolete with a trend of weakening of most of the traditional family properties: reduced presence of females in the home, increase in marriage age, low marriage and fertility rates and high divorce rates. The increase in the rate of female activity also highlights the importance of the dual role of women. Despite this increased vulnerability of families, Crouch observes a general tendency to preserve family activities: the tendency is not just to compress family duties and to delegate them externally, but also to preserve the quality of family relations (and the standard of life) through new reproductive strategies (such as having smaller families or delaying birth decisions). Crouch’s analysis does not therefore identify an alternative between the two models (individualization versus family solidarity), so much as the existence of different overlaps between them that vary according to different social and institutional structures. According to this perspective, a careful analysis must be made to identify which kinds of combination between individualization and family

40 Social Vulnerability in Europe

solidarity are emerging. This chapter sets three objectives: ●

●

●

to focus on which forms of organization are adopted in the three crucial dimensions of family organization: the structure of the household, the income structure, and the gender contracts adopted by families with children; to identify the geographical distribution of the models and the overlaps existing between them – the aim is to go beyond the well-known differences between Northern and Southern Europe and reconstruct the family models that exist in the different geographical areas of Europe and to locate those areas along a theoretical continuum between individualization and family solidarity; and to observe short-term trends, that is to say changes that coincide within the historical period in which the decline of the MBW family took place – changes that occurred in the period 1995–2001 (for which directly comparative data are available).

Emerging geographical clusters: beyond state borders As already mentioned in Chapter 2, the basic geographical unit used in this study is not the national level but sub-national levels for the five larger European countries. The purpose of this attempt, as compared to the more traditional approach of grouping by states, is to see whether important variations and common factors exist between geographical areas of similar size and between internally homogeneous geographical areas. Starting with the 28 European macro-areas that were preliminarily identified in Chapter 2, the analysis conducted at the level of single geographical areas constituted the basis of an attempt to group the 28 regions within macro-areas on the basis of similarity in terms of the two dimensions considered. The results of the analysis showed that there was no national unity for three (Germany, Italy and Spain) of the largest five countries, while it remained firm in the two remaining countries (France and the United Kingdom). Even the smaller countries were aggregated together with the macro-regions of which the larger countries are composed on the basis of a special model as shown by the results of the cluster analysis given below, which identified the following macro-regions2: ●

●

●

an Anglo-Dutch area consisting of the United Kingdom and the Netherlands; a Germanic area consisting of Southern and Western Germany, Austria and Luxembourg; an extended Scandinavian area, composed of Sweden, Finland and Eastern Germany;

Beyond the Male Breadwinner Family Model 41 ●

●

●

a Mediterranean area composed of Southern Italy and Spain except for Madrid and Eastern Spain; a second Mediterranean area composed of Central and Northern Italy, of Portugal, Greece, Madrid and Catalonia, together with Ireland; and a French-Danish area composed of France, Belgium and Denmark.

Within each of these areas households with consistent and relatively homogeneous structures are found that are sufficiently different to be distinguishable from those observed in other regions of Europe.

The demographic composition of families The analysis of family composition is performed in three steps. First, population distribution is observed (by stage of life) in family situations characterized by individualization or family solidarity, with a particular focus on the young and the elderly. The demographic composition of European families is then analysed by distinguishing five specific types, and focusing the attention on inter-generational relations. Attention is finally focused on current trends. Whom do Europeans live with? The strength of individualization and family solidarity processes can be observed by analysing the distribution of European adult individuals in four typical situations: living alone, living in a couple without children, being a parent (single or not) living with children and being a child living with parents. What emerges from this analysis (see Table 3.1) is that living alone is a peculiarity of two extreme age groups: young adults and the elderly. Living with parents is nevertheless also quite widespread for young adults (21–35 years), just as elderly persons remaining within their children’s family are also very common (almost one fifth). If we consider young adults (21–35 years), it can be noticed that Europe is an area in which the individualization of young people varies greatly. Table 3.1

Position of individuals over 20 years of age in families, by age groups

Age

Alone

21–35 36–50 51–65 over 65

12.0 8.2 11.7 31.5

In a couple without children 15.2 9.6 37.3 46.6

Source: ECHP, authors’ own calculations.

Parent with children

Child with parents

Total

30.3 74.3 48.2 21.6

42.5 7.9 2.8 0.3

100 100 100 100

42

Social Vulnerability in Europe

Almost 60 per cent of individuals in the 21–35 age group live away from their parents and more than half of them are already parents. Nevertheless 42 per cent of young Europeans aged 20–35 years old live with their parents (or with one parent). Differences between macro-regions are very great. The percentage of young adults living with parents reaches 60 per cent in the Mediterranean area, while it remains between 30 and 40 per cent in continental countries and falls right down to 20 per cent in the United Kingdom and to 10 per cent in Scandinavian countries. Data about the percentage of young adults living alone are perfectly complementary to these. A similar situation can be observed if we consider the over 65 population. Ageing exposes a large percentage of the population to solitude at a late stage in life: almost one third of Europeans over 65 live alone, almost one third live as a couple, while approximately one fifth of them live with their children (see Table 3.1). There is also a great variety of situations. A significant proportion of the elderly in Mediterranean countries (more than 40 per cent) live in families with their adult children. The percentage of the elderly residing with children is around 20 per cent in the French-Danish area, while it falls to below 15 per cent in the Germanic area, to become marginal (under 10 per cent) in Anglo-Dutch and Scandinavian countries. There are therefore two opposing polarized models. On the one hand there are regions in which the tendency toward solitude is countered by the presence of family models based on two generations of adults living together, and on the other hand there are regions in which the individualized model prevails and a large majority of the elderly population lives alone. The position across the continental macro-regions seems to lie in an intermediate position.3 Generally, the great variety of family situations in both the young adult and elderly populations is located along a continuum running between a strongly individualized model and a model based on the preservation of strong family solidarity. There is also a very strong similarity between young and elderly situations in the same area (the correlation for geographical areas is 0.85). Different models of cohabitation therefore emerge with at one extreme a model based on the cohabitation of two or more generations of adults (a ‘strong family’ model, constructed around the extended family and the centrality of inter-generational ties; see Reher, 1998) and at the other extreme a family model characterized by strong individualization and the marginality of families characterized by inter-generational ties. Continental European regions fall into an intermediate position with a broader and more heterogeneous range of family models. European family types Europe is therefore characterized by the simultaneous presence of different family models in which the balance between inter-generational ties and

Beyond the Male Breadwinner Family Model 43

individualization varies enormously. Many of the most popular classifications in the socio- demographical analysis suffer from the limitation of not bringing out this fundamental distinction. The simplified classification presented in Figure 3.1 takes account of the different degree of individualization versus family solidarity that is present in family models. Taken together single-person households and couples without children account for more than half of European families (52 per cent). Their distribution varies greatly geographically. Families with children under 25 years of age account for two fifths of total families, but for more than half of families if the population over 65 is excluded. There is a greater presence in Southern Mediterranean (54 per cent) and Central-Northern Mediterranean countries (47 per cent), while distribution falls appreciably in the Scandinavian, Germanic and Anglo-Dutch areas. Crisis and the break-up of couples shown by the number of single-parent families only appear in significant percentages in the Anglo-Dutch area and partially in Scandinavian and Germanic countries, while the figures is less in Francophone and Mediterranean countries.

(%) 0

20

40

60

80

100

Extended Scandinavian area Germanic area Anglo-Dutch area French-Danish area Central-Northern Mediterranean area Southern Mediterranean area CentralSouthern Northern Mediterranean Mediterranean area area

FrenchDanish area

Anglo-Dutch area

Germanic area

Extended Scandinavian area

Parent(s) with children > 25 years

15.1

13.3

7.0

4.4

7.9

4.2

Single-parent with children < 25 years

3.9

4.6

4.4

5.8

4.0

4.6

Parents with children < 25 years

49.9

42.8

35.1

27.6

31.7

23.5

Couples

16.4

19.8

28.4

29.4

22.1

26.3

Single

14.7

19.4

25.1

32.8

34.3

41.4

Figure 3.1 Demographic structures of households in the EU-15, percentage by area Source: ECHP, authors’ own calculations.

44

Social Vulnerability in Europe

The last type of family could be called ‘long families’. This term is used to refer to families in which one or two parents live with a child (or more than one child) who is (are) over 25 years of age.4 The presence of these families undoubtedly constitutes a strong factor of diversification among European countries. These families account for 9 per cent of the total, but almost double (accounting for 15–13 per cent) in Mediterranean countries. The percentage in the French-Danish and Germanic areas is around average, while in Northern countries it is below 5 per cent. These lines of differentiation are influenced not only by the dissimilar timing with which new families are formed in the different areas, but also by the different propensity for maintaining strong inter-generational ties between adults. Around half of these families have an elderly relative within them. It is therefore a diffuse model which helps counter the tendency of elderly populations to live alone. A significant proportion of these families (one third approximately) consist of single-parent families, in which adult children live together with a single parent.5 To sum up, this analysis found a tendency toward the formation of singlegeneration households in some areas and the resistance of others within a model that is still strongly bi-generational. The single-generation component in the Southern Mediterranean area is very limited accounting for 31 per cent. It constitutes 40 per cent in the Central-Northern Mediterranean area. A substantial proportion of ‘long families’ was found in these countries. Taken together, multi-generation households in most of the Southern Mediterranean area account for two thirds of the total, while in areas farther north they account for 60 per cent. The opposite model is found in the Scandinavian and the Anglo-Dutch areas. Single-generation households are clearly prevalent here, accounting for more than 60 per cent of families. It is therefore an important and recognizable model, consisting mostly of young and adult people located at the centre of social and working life. ‘Long families’ account for less than 5 per cent, an almost marginal model. Finally, the situation found in the FrenchDanish and in the Germanic areas is more balanced. Single-generation households in these countries account for approximately half of the total. These areas are characterized by a strong component of families with children below 25 years of age, while ‘long families’ account for 7–8 per cent. They generally show a more balanced structure. The changes observed in the period 1995–2001 reflect the different paths traced by the groups identified. In the period considered, ‘long families’ increased while families with children under 25 decreased. While the increase in ‘long families’ affected all areas (with greater intensity in Mediterranean areas), the decrease in households with children under 25 did not affect the French-Danish or Germanic areas. In the Germanic and the Scandinavian areas the increase in ‘long families’ occurred in parallel with a process of further individualization of households, shown by the

Beyond the Male Breadwinner Family Model 45

increase of single-person households. In Mediterranean areas the increase in ‘long families’ constitutes the only substantial positive trend. Generally these data suggest that on the one hand the process of individualization and the progressive decrease in families composed of different generations has not been the dominant one in the last 10 years, and that on the other hand numbers of families composed of more than one generation of adults are increasing. In any case, the hypothesis that there is a progressive convergence at European level between family models is not confirmed by this evidence.

The income structure of families: what is beyond the crisis of the MBW model? A look at long term trends and interpretations The second point of analysis regards the income structure of European families. We want to know which impact the increase in the female activity rate and the tendency toward equal gender conditions in employment have had on the division of labour within families composed of couples. More specifically we ask whether these trends have led to the spread of an adult-worker family model, or whether the increase in working activity for women is reconciled by working shorter hours than men, by more flexible employment conditions or more variability of the employment position of women during the course of family life, to produce therefore the diffusion of the OAH model or the adoption of different family models. We also ask to what extent and in which geographical areas ‘resistance’ of the MBW model or ‘modified’ variations of the MBW model has been reported. The studies and statistics available clearly show a trend toward a more complex family income structure, which is progressively moving away from the classic MBW model. An increase has in fact been found in the number of income earners as a proportion of family members, thanks to the growth in female activity and, mostly in Central and Southern Europe, to the importance of work performed by young adults still living with their parents. A generalized increase in the female activity rate started in the 1960s in Western Europe, but this was very limited in some specific countries (in Southern European countries and partly in continental countries). Female employment has increased moderately at more or less the same pace in all Europe since 1980: the relative positions were established in the 1960–1980 period, with strong growth in Scandinavian countries and in the United Kingdom. Female employment started to increase greatly in Germany in the middle of the 1970s compared to Southern Europe, while the historical delay for Southern European countries (with the exception of Portugal) does not appear to be increasing in the last two decades (Moreno Minguez, 2005).

46

Social Vulnerability in Europe

The long-term changes therefore offer a picture of an MBW model that is becoming increasingly obsolete. The current question is what the alternative models are. The data for current trends are less clear on this point. According to some scholars (Lewis, 2002), the emerging model is the adultworker family model, with the presence of two earners. The growth of this model seems determined by factors that are both economic (it is the most effective against the risk of poverty) and cultural (it assumes complete gender equality in the labour market and progressively also for caring). According to Del Boca (2002) this is explained above all by the increase in job opportunities and childcare services and less by the economic needs of families: this is shown by the fact that the increase in female employment is occurring primarily for women with high levels of education and with high salaries. Nevertheless, a large part of female employment is part-time6 in many EU countries (Rubery et al., 1998). More than one third of women (36.3 per cent) were employed in part-time jobs in 2005. According to many analyses the wages paid for part-time work do not give economic independence in the large majority of cases. Its justification instead lies in the fact that it constitutes one way of reconciling work and family. On a theoretical continuum between complete economic independence with complete absorption of time in working activity and total investment of time in family caring activity, part-time employment seems closer to the second extreme with part-time work associated with the existence of a ‘modified male breadwinner role’ (Pfau-Effinger, 1998b). It could therefore be hypothesized that, rather than a dual-earner model, what is spreading is the OAH earner model, with female income that supports the male income rather than increasingly equalizing it. This fact would mean that the present situation basically shows a tenuous move away from the MBW model, if not actually a tendency to adapt the same model to the new circumstances. Family organizations Let us consider the income structure of working European households.7 In 2001 MBW households (one income, one job) still constituted the most common income model: they represented more than one third of working couples (35.4 per cent). In addition to the pure form, a substantial proportion of families (10.1 per cent) constitute an ‘extended MBW model’, where households have at least two incomes but only one comes from the main couple in the family, while the second is earned by young adults or other family members. In this case the underlying gender arrangement (Pfau-Effinger, 1999) assigns gender roles according to the traditional MBW model. These are in fact ‘long families’ that were originally organized on a MBW model. Taken together, pure and modified MBW households account for 45 per cent of the households considered here. In the first stage of family life (with the couple under 40 years of age) MBW households account for only one third and are much less common than dual- earner households. In the subsequent stage of

Beyond the Male Breadwinner Family Model 47

family life (with couples aged between 40 and 65 years) they represent the most prevalent model. Dual- earner households account for 40 per cent of the total. This is the most common model among the youngest families (aged up to 40 years), where it exceeds 50 per cent. According to Lewis (2002), attention must be paid even in dual- earner households to the amount of the economic contribution of working women, which is often much less than that of the men. If only income from work is considered, the average percentage of income generated within families by women is 42 per cent. To summarize, 12 per cent of couples consist of ‘weak dual earners’, in which the income of one of the two partners, very often the female,8 is at least half of the income of the other partner, while in 28.3 per cent of families (defined as ‘strong dual’) male and female incomes are basically equal. The OAH income structure (one full time + one part time) accounts for only 15 per cent of all working couples: two couples in every thirteen. It is therefore a minority model in Europe. The percentage does not vary as a function of the stage of family life. On average, female income accounts for 30 per cent of income from work. To sum up, there is generally a considerable persistence of the MBW model. While dual-earner families constitute the majority of young couples, many of these couples constitute a weak dual model in which the contribution made by the women remains markedly less than that of their male partners. Finally, the total proportion of OAH income families remains low at the EU level. Geographical differences The general configuration presented so far contains considerable geographical differences (see Figure 3.2). There is a strong diffusion of households composed of one person. Dualearner households are widespread in Scandinavian countries (38 per cent), in the French-Danish area (44 per cent) and to a significant, but much lesser extent, in the Central-Mediterranean area (32 per cent). If they are divided into ‘strong’ and ‘weak’, the latter are less evident in the Southern Mediterranean and in the Anglo-Dutch areas. The MBW and extended MBW families are spread above all in Mediterranean areas (68 per cent in the South and 51 per cent in the centre), while the OAH income models are found in the Anglo-Dutch area (24 per cent). The Germanic area has generally the most balanced mix. A detailed picture of the different regional characterisations that are identifiable in Western Europe can be drawn from this analysis. Generally four principal situations emerge: ●

a situation based on a strong presence of dual- earner couples: this situation is typical of the French-Danish and Scandinavian areas (Ellingsaeter, 1998). It is also diffuse in the Central Mediterranean area. The percentage of strong

48

●

●

●

Social Vulnerability in Europe

dual-earner couples is close to (or higher than) 30 per cent in these areas and is markedly higher than the percentage of weak dual earner couples; a situation based on a balanced mix ofMBW, dual- earner and OAH income couples: this is typical of the Germanic area as well as of Central-Northern Mediterranean area. It is a situation that combines a variable mix of the different forms of family organization; a situation characterized by the widespread rise in OAH income couples, the strong version of which is the Anglo-Dutch situation and the weak version is that of the Germanic area; and a situation hinged on the prevalence of MBW households: peculiar to the Southern area of Mediterranean Europe, where as many as two thirds of families present this model (which also includes the ‘extended MBW model’ typical of ‘long families’).

Apart from the two areas that are located in the far North and in the far South of Western Europe, which are most polarized along the individualization/ family solidarity axis, the rest of the areas present a very varied configuration, with many different mixes. French-Danish area Central-Northern Mediterranean area Southern Mediterranean area Extended Scandinavian area Germanic area Anglo-Dutch area 0

Strong dual Weak dual Oah Extended mbw

10

20

30

40

50 (%)

60

70

80

CentralExtended Southern Northern Scandinavian Mediterranean Mediterranean area area area

Anglo-Dutch area

Germanic area

17.1

14.4

28.6

12.1

22.1

90

100

FrenchDanish area

33.6

7.0

8.7

10.2

6.8

10.0

11.0

23.7

14.4

6.1

6.0

7.3

11.5

4.2

8.5

1.3

16.3

14.6

4.4

Mbw

25.5

27.8

19.4

51.4

36.2

21.6

Single working

22.6

26.2

34.4

7.4

9.9

17.9

Figure 3.2 Household income structures, percentage by area Source: ECHP, authors’ own calculations.

Beyond the Male Breadwinner Family Model 49

The situation presented for 2001 is the outcome of the trends that occurred in the previous period, in which a diffuse crisis of the traditional MBW model took place. The spread of this model fell by 12 per cent between 1995 and 2001 (to more than 34 per cent). The family types that increased in the same period were the ‘extended MBW’ households (+ 4.7 per cent), the ‘strong dualearner’ households (+ 3.4 per cent) and to a lesser extent OAH income households (+ 2.3 per cent) and ‘weak dual-earner’ households (+ 2.1 per cent). If we observe the trends for the period 1995–2001, we see that the crisis of the MBW model is widespread in all areas of Europe: it is therefore a generalized process. The responses to it are nevertheless varied and run along the lines of the emerging models that we have already identified in the different areas. There is strong growth in dual- earner households and in the ‘strong’ version in particular in the French-Danish area. The same thing also occurs in the countries of the Mediterranean area, above all in the Central-Northern part, a phenomenon that is also accompanied by an increase in extended MBW households. The increase in extended MBW households is the most visible phenomenon in the Southern Mediterranean area. The Scandinavian area experienced only very small changes (the overall changes in mixes amounted to 5 per cent) that, however, were in the direction of the individualization of young adults: a process that is also emerging in the Anglo-Dutch area where it is accompanied by increases in OAH income families. Finally, the Germanic area is experiencing more complex development of all family models with the exception of the MBW model. However it is interesting to see how the slight increase in dual- earner households in the last three areas is found exclusively in the ‘weak’ forms of this type. The analysis shows that there is no current convergent trend toward the spread of a single, uniform model of dual- earner families, but that there is a great variety of family organization formulas. If any convergence can be identified, then this is exclusively negative and concerns the decrease in ‘pure’ MBW households. Nevertheless, a great variety of different paths away from this starting point were found, varying from recycled editions of ‘new MBW households’ (which include the second wage earned by an adult child still living with parents) to the spread (moderate and concentrated in a small number of countries) of the OAH model and of families with two earners present, but with strong economic dependence on the male breadwinner. The crisis of the ‘pure’ MBW model has therefore been met with a wide and varied range of family models that can only be superficially identified with the spread of the dual- earner family.

Reconciling childcare and work: problems and solutions A look at long-term trends and interpretations The problem of reconciling caring and work is of growing importance as the model of the continuous career woman becomes more popular in all

50 Social Vulnerability in Europe

European countries and is much more widespread than the sequential model in which women’s careers are interrupted for childbirth. One key to interpreting current processes of greater female participation in the labour market is based on the growing individualization taking place within it and the consequent commodification of care. According to this interpretation, care only constitutes an exogenous constraint on female employment to be resolved either by commodification (through direct purchase by families and the public distribution of cash payments that support the purchase of care on the market) or by the direct public supply of services. According to some analyses, the rate of female employment varies with childbirth as a function of the number of children. In some countries (Germany, the Netherlands and the United Kingdom), the fall in the employment rate occurs more with the change from one child to more children, than in the move from no children to one child (Naldini, 2006). According to Del Boca (2002) the gender gap increases by 10–14 per cent for women with one child and by 20 per cent for women with two children. This gap remains basically unchanged for all European countries, even though the point of departure is naturally different with regard to the level of the female activity rate. According to Iacovou (2004) the correlation between the presence of children and of working mothers does not, however, clarify the causal relationship for working mothers because the specific employment situation of the mother could predate the child’s birth. By inserting a longitudinal model, they discovered that the child effect seems to be present in only a small number of countries (Germany and the Netherlands). Generally, it is nevertheless observed that, regardless of the timing and causal sequence of events, the problem of finding a balance between childcare and working exists for all women and every couple with a child or children. According to Lewis (2002), obligations and practices of mutual support and care continue to bind people together. Care cannot be fully de-familialized or commodified as it expresses lively reciprocal obligations within the family. The multiple solutions that are actually found to the problem must be therefore inscribed in a complex model that simultaneously takes account of the employment constraints and expectations of women, and of the dominant cultural model in couples. Existing situations Given these considerations, the approach adopted here is understanding which family model is employed when there are children in a family.9 A comparison is made between couples with children and couples without children, with the analysis limited to families in which the head of family is not more than 45 years old.10 The percentage of MBW households increases appreciably in the presence of children (it rises from 24 to 33 per cent and to 42 per cent where

Beyond the Male Breadwinner Family Model 51

there is more than one child). The reduction in the dual- earner model is even stronger (–16 per cent) and is also very pronounced for the change to more than one child (–29 per cent). Finally OAH income families double in the presence of children, but nevertheless remain scarce in relative terms (7–10 per cent). Generally the tendency for the rate of MBW households to increase in the presence of children is already pronounced in the passage from no children to one child and it increases almost proportionally in the passage to two children. Substantial differences emerge among the different areas (see Figure 3.3). In the French-Danish and Scandinavian areas childcaring does not produce strong changes in family income structures. In these areas not only is the dual- earner model very widespread, but it doesn’t drop significantly for families with children. There is also a high degree of equality between mothers and women of the same age without children as far as their position on the labour market is concerned. The Germanic and Anglo-Dutch areas show the largest changes in households with children. In the Germanic area the presence of children is associated with the widespread adoption of the MBW model and a substantial and widespread abandonment of the dual- earner model, leaving the OAH model in a marginal position. It is therefore a situation that seems to create

French-Danish area Central-Northern Mediterranean area Southern Mediterranean area

Extended Scandinavian area

Germanic area

Anglo-Dutch area

−40

−60

Anglo-Dutch area

−20

Germanic area 16

0 Extended Scandinavian area 1.1

20 Southern Mediterranean area −1

40

60

Central-Northern French-Danish Mediterranean area area 3.4 8.3

oah

35

dual

−48.3

−54.5

−13.5

−23.5

−24.2

−13.7

mbw

13.3

38.5

12.4

23.5

20.9

5.3

Figure 3.3 Changes in the income structure of households caused by the presence of children (in relation to households without children), percentage by area Source: ECHP, authors’ own calculations.

52

Social Vulnerability in Europe

a polarization between mothers and women of the same age without children. In the Anglo-Dutch area the presence of children is associated with the growth of the OAH model. Finally, in Southern Europe there is a weak tendency for households with children to favour the adoption of an MBW model compared to households without children, as opposed to the dual-earner model and the OAH model. The MBW model is widespread in countries where the presence of MBW households is very dominant even in the absence of children. It is therefore an organizational model that is still very well established and is not adopted only because of the difficulties in reconciling mothering with working. Comments Four conclusions can be drawn from the above analysis. The first is that the European areas most polarized around the dualearner model (the Scandinavian countries) or the MBW model (the areas of Southern Europe) reproduce these models even in households with children. Their presence does not produce qualitative changes, but only a further radicalization of the already dominant model. In these areas the problems of reconciling childcare with working are therefore dealt by using the most traditional form of family organization that exists. In both these scenarios families with children seem to be pushed even more strongly toward the dominant family model than families without children. The second conclusion is that the problem of reconciling childcare with working in continental countries is addressed very differently: while in the Germanic area the gap for the dual-earner model coinciding with the presence of children affects a large proportion of households, in the FrenchDanish area dual- earner households with children are just as widespread as households without children. These two large continental areas therefore follow two opposing lines of action to address the problems of balancing childcare with working. The last conclusion concerns the scarce use of the OAH model to reconcile childcare and working. No general confirmation was found for the tendency described by Del Boca,11 according to whom women in Southern Europe leave the labour market more frequently in the absence of part-time work (or because it is highly undesirable) and women in Central Europe prefer part-time work as a form of compromise, while full-time employment is preferred in Northern Europe because of the presence of childcare services. In reality the picture that emerges shows a large exodus of mothers from the labour market in Germanic countries, with a preference for part-time work in the Anglo-Dutch area, and a decrease in the dual- earner model in Southern Europe. Finally, highly diversified geographical models of family structure are present in Western Europe with macro-areas positioned at different points along theoretical axis that runs from individualization to family solidarity.

Beyond the Male Breadwinner Family Model 53

This axis is intersected by other axes that identify different models of combination and overlap between individualization and family solidarity: the Anglo-Dutch model, strongly oriented toward the OAH model; the Germanic model characterized by a mixed and heterogeneous configuration (with widespread return to the MBW model during the reproductive phase); the French-Danish model heavily oriented toward a dual-earner structure, with a strong capacity within the system for maintaining childcare in line with this family organization; a Southern Mediterranean model still firmly anchored to options typical of a familist model, hinged on total internalization of caring functions and on the persistence of the traditional MBW model.

Family models Finally the analysis leads to the identification of a complex family classification designed to describe and summarize the variety and complexity of the family structures existing in Europe today. Family structures have been analysed by jointly considering two dimensions: the demographic structure (including the presence of children) and the income structure of the household. For the first dimension analysis led to the identification, beyond the classic division into four classes (single person households, couples without children, couples with children aged under 25 years and single-parent families), of a fifth type, very widespread in many European countries, consisting of households in which at least two generations of adults with kin ties are present. One specific characteristic of these households is that they have a life that extends beyond the period of time in which the simultaneous presence of two generations is caused mainly by educational demands because the younger generation is not yet fully economically self-sufficient. This is a family type that is different from the traditional households with children, because the members continue to live together even when the children have attained a certain material independence, at least in potential terms.12 This type accounts for 9 per cent of European households and reaches much higher levels in many EU-15 countries. The second dimension concerns the income structure of households. The analysis considered the spread of the three most consolidated models of family income organization consisting of adults of working age: singleincome families (generally defined as MBW because the large majority of the incomes are earned by men), dual- earner households (in which both partners work) and OAH income households (two incomes, one from fulltime work and one from part-time work). Another model was added to these, termed an ‘extended MBW’ model, with two incomes, one from an adult (usually an offspring) who is not one of the two partners: it is an income model that is common among the ‘long families’ already described. If account is taken of other family models that are characteristic of elderly persons (who therefore have an income from pensions) and persons

54

Social Vulnerability in Europe

who live alone, 10 family models can be identified on the joint basis of demographic and income structure. The final classification of family types therefore describes, although necessarily in summary form, a large part of the variation present in Europe today in the different forms of family organization and it is able to constitute a basis for identifying differences and factors common to different European regions (see Figure 3.4). Households composed of pensioners constitute 30 per cent of European households. The percentage increases in some regions of Germany and the United Kingdom, while it falls considerably in Spain, Southern Italy, Greece, Portugal and Ireland. In all, 59 per cent of these households consist of single persons, while others are composed of elderly couples. Households composed of adults without children constitute one quarter of European households. They amount to 30 per cent in Scandinavian countries, in some areas of Germany and the United Kingdom, while they remain at around 10–15 per cent in Southern European countries and in Ireland. Around half of these household are composed of one person, while the remaining households consist of couples with a predominance of dualearner couples (60 per cent). 20 18 16 14 12 10 8 6 4 2

Figure 3.4

Household typology, EU-15

Source: ECHP, authors’ own calculations.

Parent(s) with children > 25 years

Single-parent with children < 25 years

Parents with children < 25 years OAH

Parents with children < 25 years Dual

Parents with children < 25 years MBW

Couple dual or OAH

Couple MBW

Single working

Couple welfare

Single welfare

0

Beyond the Male Breadwinner Family Model 55

Households composed of adults with children aged less than 25 years represent almost 40 per cent of households: approximately two households in five. This percentage increases in Southern European countries and in Ireland (close to 50 per cent of households) and drops considerably (down to only a quarter of households) in the Scandinavian countries and in the Germanic area. The majority of these households adopt an MBW model. The MBW model is predominant in 17 geographical areas out of 28. The dual- earner model predominates in only nine areas (in the Scandinavian countries and the French area). The Anglo Saxon area has a high presence of single-parent households (one fifth of all households with children under 25 years), while the Netherlands have developed a particular OAH income model. Finally, households composed of parents and adult children (over 25 years) account for 8 per cent. The percentage reaches 12–14 per cent in the macroregions of Southern Europe. As has already been shown, this is a new model, which reflects the protraction of the transition to adult life of young generations and a certain propensity to maintain household composition based on kinship and on individual choice.

Conclusions At the start of this chapter questions were posed, using which it was intended to analyse different aspects of family organization in Western Europe. Which models are most popular in this period, in which the MBW model is becoming more obsolete? To what extent is an analysis based on individualization processes able to explain what is happening in European families? Are there convergence processes currently in progress between geographical areas? To what extent is the national variable still relevant, at least for large countries, in explaining family strategies and structures compared to divisions based on macro-regions that cut transversely across different countries? The study proposed here offers some indications for these questions, indications that, however, often do not identify clear and unequivocal processes. The processes of change in progress, as they relate to demographic and labour market changes, are reflected in families in the search for new equilibria in which a balance is sought between individualization and family solidarity, without the appearance of any clear prevalence of one trend over any other. On the one hand there was an increase in extended MBW households and ‘long’ households in the 1990s, as well as in the adoption of one and a half income models as a reconciliation strategy, above all, for families with children. On the other hand trends also emerged that indicated a fall in the proportion of MBW models and a widespread increase in the dual- earner model in both single-person households and in young couples without children.

56 Social Vulnerability in Europe

The abandonment of the MBW model is therefore occurring in an anything but linear manner: in many cases this abandonment is more apparent than real (consider for example the phenomenon of the ‘extended MBW’ households, in which the second income is earned not by the partner but by other family members outside the couple), while in others its effect is more limited than expected with the change to households based on more than one wage (e.g., the phenomena of ‘weak’ dual- earner couples in which one income, usually the female one, is much lower and of one and a half income households). Furthermore there are no clear processes of convergence between European regions: they often follow along the same historical paths. Even if the time period is relatively short (6 years, from 1995 to 2001), the picture that often prevails is that of stability, if not at times of a slight increase in differences in the adoption of diverse forms of family structure (except for a convergence for households with children under 25 years). It is therefore important to reflect on European models and trends. European models are developing along different lines, in which the balance between individualization and family solidarity is found through a very diversified mix. These models, and this is the last interesting result, do not always, however, run along national lines: three of the five main European countries (Italy, Spain and Germany) are in fact characterized by the presence within them of macro-regions that are grouped in different clusters, while countries with a stronger centralist tradition (France and the United Kingdom) remain more unified.

Notes 1. In Scandinavian countries the tensions on which this paper focuses arose in an earlier period. As is shown in the analysis in this chapter, it seems that new tensions are being generated in these systems too. 2. The cluster analysis performed was of the hierarchical type and it was carried out on the spread of the 10 family models described below in this paper for each of the 28 regions on which this analysis is based. 3. This distinction is not affected by the different composition of the population over 65; even if the over-75s are considered separately or a distinction is made between men and women, the sequence of the macro-regions remains basically the same. 4. The choice of 25 years of age as the threshold for adult offspring used to identify ‘long families’ is because at that age more than 50 per cent of children who live with their parents have their own income and the median age at which European children leave home is 24 years. 5. In two thirds of these households one parent lives with one child. In the remaining third there are two children. It is therefore a very small family type. There is one elderly parent in half of these households. 6. By using the definition adopted by the Eurostat ECHP survey, part-time work is intended to mean any job of less than 30 hours per week.

Beyond the Male Breadwinner Family Model 57 7. The focus in this section is on those family income structures that depend on the present position of their components in the labour market. Households receiving welfare benefits and households with a mix of wage and welfare incomes are therefore excluded, together with lone-parent families and those consisting of one adult only. The total proportion excluded from the analysis amounted to 55 per cent of European households. 8. A total of 15 per cent of weak dual-earner couples consisted of a woman with an income at least twice as large as that of her male partner. In 85 per cent of cases the situation was the opposite. 9. It is, however, not assumed that childcare influences working, but that there is reciprocal conditioning between these two aspects. 10. This choice was made in order to be able to compare family strategies in couples who, with or without children, are at a stage of life in which children will be either minors or in any case still economically dependent on parents. 11. Del Boca estimates that the presence of small children has the effect of increasing part-time work in Northern and Central European countries in particular, but not in Italy and France where part-time work is more heavily determined by demand factors (compulsory part-time working, discrimination against women in the labour market). Shorter working hours and the recourse to part-time work are therefore signs of discrimination and of disadvantage in the labour market in Southern Europe and in France, while it is a sign of compromise between working and childcare in the Northern EU and Scandinavian countries. Generally, it is claimed that the Scandinavian way is preferred by women from a working viewpoint, because it avoids the disadvantages and vulnerability often associated with part-time work. 12. Naturally this analysis does not take any account of cultural considerations. In many countries coming of age is not considered a crucial moment that identifies the attainment of material independence. It is, however, accepted in virtually all of Europe that officially coming of age marks the start, if not the completion, of the transition to adult life characterized by economic and housing autonomy.

4 Income Vulnerability in Europe Salvatore Curatolo and Guglielmo Wolleb

Introduction Economic vulnerability is defined in this chapter as the exposure of a household to the likelihood of suffering economic hardship or material deprivation as a result of negative events affecting its ability to ensure that each of its members enjoys the living standards prevalent in the country of residence. The concept of economic vulnerability is closely associated with the notion of economic insecurity. It refers to a situation where there is a high risk of income difficulties, which may remain potential or become effective but which create psychological malaise and objective difficulties in exploiting the opportunities offered by life. It therefore concerns a wider aggregate than poverty, which consists exclusively of persons who are poor in a given year. The notion that there is a category of persons exposed to the risk of economic hardship broader than those strictly defined as ‘poor’ has also been advanced in the literature on new social risks. Some scholars have argued that long-term economic, demographic and cultural trends (Taylor- Gooby, 2004a) in Europe have generated new social risks that replace, or add to, the old ones. These new social risks affect a larger number of people than the old ones, and they do not properly fit into the existing class structure of postindustrial societies. Some commentators have spoken of the ‘individualization’ of poverty risks (Leisering and Leibfried, 1999: 9). These risks do not necessarily lead to permanent poverty or deprivation, but rather generate recurrent difficulties distributed across the life- course (Taylor- Gooby, 2004a, 2004b; Whelan and Maitre, 2004). Empirical evidence in support of this idea has been furnished by poverty studies that use a longitudinal approach rather than a traditional crosssectional one (Leisering and Leibfried, 1999; Apospori and Millar, 2003; Layte and Whelan, 2003; Whelan and Maitre, 2007b). The main finding of these 58

Income Vulnerability in Europe

59

studies is that the number of people affected by episodes of poverty over a period of time is much larger than the number of people affected in a single year.1 This is because the persistently poor are flanked by a very large number of the temporarily poor affected by isolated episodes, and by the recurrently poor, who repeatedly exit and enter poverty (Layte and Fouarge, 2004). The aggregate of temporarily and recurrently poor is what we describe in this chapter as vulnerable. The high number of the vulnerable has been interpreted as indicating the existence of a large group of households, distinct from persistent poor households, living in a state of fragility and likely to lapse into economic hardship if hit by negative but typical events of family life. The aim of this chapter is to investigate the heterogeneous world of poverty distinguishing between cross-sectional poverty and longitudinal poverty and, within longitudinal poverty, between persistent poor and vulnerable. We shall assess the size of the different typologies of poverty; examine entries to and exits from poverty and upward and downward movements along the income scale; analyse the demographic, social and employment characteristics of the persistent poor and vulnerable. The subject of study will be the EU-15 countries in both their national and macro-regional dimensions. The chapter begins by briefly describing the concepts, definitions and methods used in the analysis. The next three sections reconstruct the context of our analysis of income poverty and highlight the wide differences across Europe in terms of development levels, degrees of inequality and income distributions. The fifth section analyses the phenomenon of persistent poverty. The next two sections examine the direction and scale of movements by households up and down the income scale, the purpose being to determine the origins and destinations of the poor. The eighth section distinguishes the different types of poverty between structural, recurrent and transient. The results are used in the two final sections to estimate the number of persistent poor and vulnerable and to identify their characteristics.

The definition and measurement of income Income is defined here as the equivalent disposable income of households measured in terms of purchasing power parity. We specify disposable income because the value of an income comprises both primary and secondary distribution. It therefore consists of all income earned from dependent and self- employment in the labour market, as well as capital income deriving from the stock of wealth. It includes, as income, all public cash transfers to the household budget and, as outgoings, all taxes and social contributions paid by the household.2 We specify household income because the unit of analysis is the individual living in the household who determines his/her standard of living by participating equally in the household’s total resources.3 Each individual

60

Social Vulnerability in Europe

is therefore attributed the income level of the household to which he/she belongs. We specify equivalent income because, in order to compare the incomes of families of different sizes and compositions, we use a modified OECD (Organization for Economic Cooperation and Development) equivalence scale, which allocates a coefficient of 0.5 to each additional person in the household and a coefficient of 0.3 to each child. Therefore, compared with a single-person family, a two-person family needs to increase its income by 50 per cent to maintain the same standard of living, and a family with a child needs a 30 per cent increase in income to reach the same standard.4 Finally, we specify income as measured in purchasing power parity in order to take account of differences in price levels among countries.5 The period considered is from 1997 to 2001, the years covered by the data furnished by the European Household Panel.6

Income differences among European countries and macro-regions It is useful, before examining the structure of income distribution for the various countries, to focus on the median income level7 in order to highlight the wide disparities among the countries and regions that make up the EU, even before enlargement (see Figure 4.1). It is useful because the adoption of a relative method on a national scale to assess the wealthy and poor in each country is likely to obscure the fact that being poor in Portugal and being poor in Denmark do not have exactly the same significance, given that the latter country records a median income almost double that of the former. In 2001, the equivalent median income in the EU-15 countries was 14,014 euros at purchasing power parity. Lying below this value were Portugal, Greece, Spain, Italy, Sweden, Finland and Ireland. The width of the gap with respect to the European median income was particularly marked in Portugal and Greece, where it stood at around 40 per cent; the gap was somewhat smaller (20 per cent) in Spain and Italy; and 10 per cent or less for the other three countries. This picture matches expectations as regards the position at the bottom of the scale occupied by Portugal and the first two Mediterranean countries; it confirms the intermediate position of Italy and the strong performance of Ireland in the 1990s. Perhaps the surprise is the shortfall of Sweden, whose relative position worsened in the second half of the 1990s.8 The magnitude of disparities is much smaller at the top of the income scale, with the exception of Luxembourg. Denmark, Austria, Belgium, Germany and the United Kingdom are, for example, separated by differences of less than 5 per cent. Even the gap between the richest, Denmark, and the least rich of the high-income countries, the Netherlands, is less than 12 per cent. This confirms the existence in Europe of a core group of

Income Vulnerability in Europe

61

countries that are very wealthy and very homogeneous in terms of income and living standards.9 Regional disparities exhibit different characteristics and intensities in Italy, Spain, France and the United Kingdom. In Italy the gap replicates the country’s traditional North/South duality, with the regions of the South recording a 27 per cent gap with respect to the national average. Within the Centre-North the differences are much smaller, and the area as a whole can be associated to the richer central core of Europe. In Spain, the gap is widest between Eastern Spain (Catalonia and Valenciana), which includes the region of Madrid, and the rest of the country, both Centre-North and South. The less advanced areas fall below the national average by about 20 per cent.

Legend < 75 75 – 95 96 – 115 > 116

Figure 4.1 Equivalent median income by macro-region, 2001. Index number (EU-15 = 100) Source: ECHP, authors’ own calculations

62 Social Vulnerability in Europe

In France and the United Kingdom, regional inequalities instead manifest a marked polarization between the capital region and the rest of the country. This polarization is particularly evident in France, where the Paris region has an income 23 per cent higher than that in the rest of the country. In the United Kingdom, income in Greater London is 14 per cent higher than the national average. Regional differences are much less pronounced in Germany, even after inclusion of the new Länder: as expected, the Eastern Länder has a median income below the national average, although the gap is less than 10 per cent.10

National inequalities in the distribution of income The European countries also exhibit marked differences in the degree of inequality and in the structure of income distribution (Atkinson, 1996; Forster and Mira d’Ercole, 2005). Figure 4.2 shows the values of the Gini index11 for the EU-15 countries for the years around 2000. The countries can be grouped into three separate blocks. The first comprises countries with high degrees of inequality, with a Gini value above 0.3 and consisting of the United Kingdom, Portugal, Greece, Italy and Spain. The second includes Ireland and the countries of Central Europe, Belgium, Germany and France, with intermediate degrees of inequality. The third block comprises the Nordic countries, Denmark, Sweden, Finland, together with Austria, the Netherlands and Luxembourg, which have low degrees of inequality. One explanation for such differences in the degree of inequality among countries refers to the theory of welfare-regime types based on the well-known 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05

DNK

SWE

AUT

LUX

FIN

DEU

NLD

FRA

BEL

IRL

ESP

ITA

GRC

PRT

GBR

0

Figure 4.2 Gini index for the distribution of disposable equivalent household incomes – EU-15 countries – 2000 Source: OECD, 2000.

Income Vulnerability in Europe

63

classification by Esping-Andersen (1990) of social democratic, liberal and corporate regimes, which was eventually modified with the addition of the Mediterranean regime category (Arts and Gellissen, 2002; Ferrera, 2005). According to this explanation, social-democratic welfare regimes have lower levels of inequality than corporate regimes, and especially than liberal and Mediterranean regimes. The data above do not contradict this argument as a whole, but there are some exceptions for individual countries.12 Differences in inequality, however, may also result from strictly economic variables. There is indeed a strong correlation between inequality and level of development. With the exception of the United Kingdom, the five countries with the highest degrees of inequality are also countries with low levels of development and with marked internal territorial disparities.13 Table 4.1 provides information on the structure of distribution. For each country, it details the proportions of the population distributed according to income classes based on national median income values, with intervals above and below 10 per cent. The median income is situated at the upper threshold of the tenth income class. The population is therefore distributed according to distance from the median income.14 The various classes have been grouped in order to provide an overview. The bottom five income classes comprise the ‘poor’ population, with incomes amounting to less than 50 per cent of the median income.15 The three income classes immediately above the poverty threshold contain families ‘at risk’, which are more likely to become poor because of their contiguity to the threshold. The next four classes, from the ninth to twelfth, comprise the ‘middle class’, with incomes around the median. The three higher classes include the section of the population with incomes 20–50 per cent above the median and which we may term ‘affluent’. Finally, the remaining classes of income, representing the top of the distribution, comprise very highincome families, which we term ‘wealthy’. The Nordic countries and the countries of Central Europe are characterized by a more egalitarian income distribution, with a lower proportion of poor families and a greater concentration in the central income brackets. Belonging to this group are Denmark, the Netherlands, Belgium, France, Austria, Finland, Sweden, Germany and Luxembourg. The proportion of the poor is around 6–7 per cent in almost all countries, and slightly higher in France. The share of the middle class varies from 35 to 40 per cent. Internal differentiation within this group is recorded only for the proportions relating to affluent or wealthy families. In the Scandinavian countries and Austria, well-being is more attenuated, with a larger number of affluent families and fewer wealthy ones, contrary to the Netherlands, Belgium, France and Germany. The other Mediterranean countries, Portugal and Ireland, exhibit a distribution with more poor families, less concentration in the middle

64

Social Vulnerability in Europe

Table 4.1 Population shares by country, region and income class* – 2001

Country or macro-region

Poor

At risk

Denmark Netherlands Belgium France Centre-East South West and South-West Île- de-France Ireland Italy North-West Centre and North-East South Greece Spain Centre and North-West East and Madrid South Portugal Austria Finland Sweden Germany North Centre and East West South Luxembourg United Kingdom North Centre Greater London West and South-West EU-15

6.5 3.7 6.2 8.0 7.1 10.3 11.1 4.5 16.7 12.1 3.9 5.6 22.4 16.7 13.0 19.1 6.8 18.2 12.6 6.8 6.0 6.6 6.0 4.7 7.5 4.7 6.4 3.9 9.2 11.4 10.9 6.5 9.3 9.0

22.3 24.9 24.4 23.1 26.3 24.9 26.6 9.7 22.4 23.7 15.5 18.9 32.6 22.5 23.6 25.4 19.2 31.3 21.0 23.5 25.4 20.3 20.7 23.0 26.8 17.0 18.4 26.2 23.6 23.0 27.3 17.9 27.9 22.8

Middle class

Affluent

Wealthy

Total

39.8 35.0 35.2 34.1 36.4 32.3 33.9 29.9 27.4 30.0 32.0 33.5 26.2 26.7 29.1 31.0 27.5 30.5 27.6 37.1 35.6 40.1 38.6 43.2 43.9 36.7 35.1 35.4 30.0 32.8 28.7 28.1 30.9 33.2

17.9 17.8 15.3 15.4 14.7 13.8 14.7 18.8 16.4 16.4 23.2 20.8 8.6 12.3 14.1 12.1 16.0 12.5 13.6 17.6 17.2 18.6 16.5 14.5 11.7 19.7 18.1 16.6 14.8 16.8 11.1 15.3 15.3 15.7

13.5 18.6 18.9 19.4 15.5 18.7 13.7 37.2 17.1 17.8 25.4 21.2 10.2 21.8 20.2 12.4 30.5 7.5 25.2 15.0 15.8 14.4 18.2 14.6 10.1 21.9 22.0 17.9 22.4 16.1 22.0 32.2 16.6 19.3

100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100

Notes: * Share of people by income class according to the national median. Source: ECHP, authors’ own calculations.

class and more wealthy families. The share of the poor exceeds 10 per cent and is highest in Greece and Ireland, where it reaches 16.7 per cent. The relative weight of average income classes is around or below 30 per cent. The percentage of wealthy families instead exceeds 20 per cent in Greece, Spain and Portugal. Close to the countries of South Europe is the United Kingdom, which has shares of poor and wealthy families above the

Income Vulnerability in Europe

65

European average, and a concentration in the middle classes lower than that of the most egalitarian countries. Also to be noted is that in all countries the range immediately above the poverty threshold, and therefore vulnerable from the point of view of income, comprises between 20 and 25 per cent of the population. Distributive inequalities amplify considerably when macro-regions are considered, and they qualify our previous results. Italy, for example, has the lowest proportion of poor people in Europe in its North-Western and NorthEastern regions, but the absolute highest amount in the South.16 Spain exhibits a similar pattern with very low rates of poverty in Eastern Spain and Madrid and rates three times higher in the rest of the country. In France, the most marked territorial disparity is between the Île- de-France, which has very few poor inhabitants and many affluent and wealthy ones, and the other macro-regions. A similar division exists in the United Kingdom between Greater London and the rest of the country. In Germany, although the gap is less wide, there are fewer poor families and more wealthy ones in the Western and Southern macro-regions of the country. When this structure of income distribution is observed at the country level, it is compatible with the theory of welfare regimes, which predicts a larger proportion of poor families in the liberal and Mediterranean systems, both of which are characterized by the weaker redistributive capacity of the welfare state. But it is also compatible with theories on late development, which predict wider internal distribution inequalities at regional level, too, in countries that began industrialization later than others, such as Greece, Ireland, Portugal, Italy and Spain (Fuà, 1980).

Inequalities in the distribution of income from a European perspective How does the structure of income distribution appear if we treat the EU-15 countries as a single political entity rather than as a group of nations? So far we have identified wealthy and poor families in various countries and macro-regions in relation to national median income. This approach is justified on the grounds that, even today, the subjective and socially accepted perception of poverty or wealth is formed with reference to a national standard of living, and because social policies are still mainly an area of national competence within the EU. But the question is nevertheless of interest for at least two reasons. The first is that the increasing mobility of workers and other persons within Europe may in the near future produce a European standard of living that would become an essential benchmark for all European citizens. A person would be, or would feel him/herself to be, poor or well off in relation to this standard and not to the national level. A second reason concerns the legitimacy of an EU-level social policy that requires all European citizens to be assessed according to the same standard,

66

Social Vulnerability in Europe

Table 4.2 Population shares by country and income class* – 2001

Country

Poor

At risk

Middle class

Affluent

Wealthy

Total

Denmark Netherlands Belgium France Ireland Italy Greece Spain Portugal Austria Finland Sweden Germany Luxembourg United Kingdom Europe

4.1 5.2 3.5 6.2 17.3 19.7 36.2 20.5 38.0 4.5 9.7 9.2 4.2 0.3 7.7 11.2

15.4 20.1 18.7 19.9 22.6 27.6 31.1 29.2 32.4 15.4 28.6 29.5 14.5 4.3 19.6 21.5

33.0 35.5 40.2 32.1 32.2 32.6 20.6 28.4 16.4 34.5 37.5 38.3 37.5 15.6 28.3 32.1

23.6 18.0 15.8 18.3 14.7 11.0 6.7 9.0 5.1 20.3 13.4 14.0 18.1 18.7 15.8 15.1

23.9 21.2 21.8 23.5 13.2 9.1 5.4 12.9 8.1 25.3 10.8 9.0 25.7 61.1 28.6 20.1

100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100

Notes: * Share of people for income class according to the European median. Source: ECHP, authors’ own calculations.

and targets to be set in terms of European values, as has already happened with the Lisbon Agenda. To answer the above question, Table 4.2 relating distribution of income among classes has been constructed by using the European rather than the national median.17 The consequence of this different way of constructing the distribution is inevitably an increase in the numbers of poor and at-risk persons in countries with median national incomes lower than the European median income, and vice versa an increase in the numbers of the affluent and wealthy in countries with median incomes higher than the European median income. However, what is most striking is the magnitude of these changes. By adopting the European criterion, 70.4 per cent of the population of Portugal is poor or at risk, 67.3 per cent in Greece, almost 50 per cent in Italy and Spain and, somewhat surprisingly, 38.7 per cent in Sweden. At the other extreme is the concentration of the affluent or wealthy in some countries, such as Denmark or Austria, where higher income brackets cover more than 45 per cent of their citizens. The results obtained when the European median is used to determine the national composition of the poor by weighting the specific risk of poverty of each country with its demographic size are somewhat unexpected (see Table 4.3). In this case, the largest number of poor Europeans is Italian (27.3 per cent), mostly in South Italy, and Spanish (18.6 per cent). There are more

Income Vulnerability in Europe

67

Table 4.3 Distribution of the European population by income class*

Country

Poor

At risk

Middle class

Affluent

Wealthy

Total

Denmark Netherlands Belgium France Ireland Italy Greece Spain Portugal Austria Finland Sweden Germany Luxembourg United Kingdom EU-15

0.5 2.0 0.9 8.7 1.6 26.5 8.9 18.7 8.7 0.8 1.2 1.8 8.6 0 11.1 100

1.1 4.0 2.4 14.4 1.1 19.4 4.0 13.8 3.8 1.5 1.9 3.0 15.2 0 14.4 100

1.5 4.7 2.8 15.6 1.0 15.4 1.8 9.1 1.3 2.3 1.6 2.6 26.2 0.1 14.0 100

2.3 5.1 3.4 19.0 1.0 11.0 1.2 6.1 0.9 2.8 1.3 2.0 27.0 0.2 16.7 100

1.8 4.5 3.7 18.2 0.7 6.8 0.7 6.6 1.0 2.6 0.7 1.0 28.6 0.4 22.7 100

1.5 3.6 2.8 15.6 0.9 15.6 2.7 10.2 2.5 2.1 1.4 2.0 23.1 0.1 15.9 100

Notes: * According to the European median. Source: ECHP, authors’ own calculations.

British poor than Greek, and more German poor than Portuguese. In light of these data, therefore, the problem of poverty is not an exclusively regional problem restricted to countries or regions of late development; rather, it is a Europe-wide problem. A hypothetical European anti-poverty policy would therefore not just benefit citizens of the less advanced countries. Conversely, however, the wealthy are also concentrated in certain countries of considerable demographic size: 70 per cent of the wealthy in Europe are in fact British, French or German.18

The persistence of poverty In the background to our analysis, therefore, is a Europe highly differentiated by level of income and by the degree and structure of inequality. This should be borne in mind when the focus of analysis moves to poverty defined according to national standards. The world of poverty captured at a specific moment in time is highly diversified. It includes both individuals and families who are structurally poor and individuals and families for whom poverty is a transitory or recurrent situation. Longitudinal analyses of poverty have brought these differences to light and enabled estimation of the weight of the different types of poverty. A similar estimation for the years 1997–2001 has

68

Social Vulnerability in Europe

Table 4.4 Poverty persistence rate by country and macro-region 2001* Country or macro-region

1997/99

1999/2001

31.9 39.6 44.4 41 42.7 28.8 46.1 37.4 51.7 50.7 21.1 31.3 60 55.1 41 42 26.3 50.6 54.5 51.9 48 44.4 33.2 41.4 39.2 54 39 47.4 53.8 49.8 33.9 49.6 46.3

43.1 27.7 41.5 48.4 43.2 58.4 58.5 30.1 68 56.8 42.2 52.8 59 61.9 48.6 59.2 38.1 45.2 60.6 49.7 46.2 51.8 39.1 59.1 39.5 57 47.2 50.7 57.7 48.7 51.8 45.5 51.9

Denmark Netherlands Belgium France Centre-East South West and South-West Île- de-France Ireland Italy North-West Centre and North-East South Greece Spain Centre and North-West East and Madrid South Portugal Austria Finland Germany North Centre and East West South Luxembourg United Kingdom North Centre Greater London West and South-West Europe

1997/2001 24 20.5 40.9 36.7 34.6 34.8 47.2 24.8 55.5 42.8 22.4 32.8 48.4 50.7 35.1 38.1 25.1 40 45.4 36.2 32.2 35.6 23.9 42.1 27.3 40.6 28.7 36.7 42.4 45.2 24.4 33.1 38.5

Notes: * Poor people at the beginning of the period who are still poor at the end of the period over the poor of the beginning of the period. Percentage shares. Source: ECHP, authors’ own calculations.

been made here. Table 4.4 shows the rates of persistence19 of poverty in the two periods 1997–1999 and 1999–2001, and in the longer period 1997–2001. The rate for Europe as a whole confirms the findings in the literature on structural poverty. About half of all the poor in the initial year of observation are no longer poor 2 years later. The persistence rate measured over 2 years assumes a value of around 50 per cent.

Income Vulnerability in Europe

69

The range of variation of this average value for Europe is very wide when analysis is conducted by country. The minimum values are recorded by Denmark with 31.9 per cent in the first 2-year period, and by the Netherlands in the second with 27.7 per cent. The maximum values are in Greece and Portugal, with around 55 per cent in the first period, and in Ireland, with 68 per cent in the second. With regard to the persistence rate, European countries can be grouped into two blocks. The first consists of countries with rates below the European average in observations in both 1997–1999 and 1999–2001. It comprises Denmark, Spain, the Netherlands, Belgium, France, Luxembourg and Germany. The second block consists of the countries with persistence rates above the European average in both sets of observations and comprises Ireland, Italy, Greece and Portugal. Countries with higher or lower values according to the year of observation are Austria, Finland and the United Kingdom. The countries with the lowest poverty rates are also those with the least persistence of poverty. The correlation is not perfect however. There are two exceptions. Although Spain has a high poverty rate, it has low rates of poverty persistence, i.e. a larger number of people experience poverty but only for a short time. Finland, by contrast, although it has a low proportion of poor citizens, still has a high persistence rate: there are fewer poor people but they remain in poverty for longer. These cases of non-correlation can be explained by referring to differences in the mobility of income distribution and to distinctive features of the welfare system. A country may have low rates of poverty but high rates of persistence, because socio- economic mechanisms foster the formation of poverty traps that make upward mobility difficult, because of a rigid and segmented labour market or because the welfare system is not designed to combat chronic poverty. Likewise, a country may have a large proportion of poor citizens but a low persistence rate if economic mechanisms foster upward mobility, if the labour market is highly flexible or if the welfare system intervenes effectively in extreme situations. Finally, the persistence rate was measured over the entire period from 1997 to 2001. One would expect the persistence rate measured over such a longer term to be lower. Families and individuals would have more time to overcome their income difficulties and rise above the relative poverty threshold. The figure for Europe as a whole confirms this expectation. In fact, the persistence rate is lower, standing at 38.5 per cent. This also holds in almost all cases if the data are analysed by country: the maximum value is now in Greece, equal to 50.7 per cent, while the minimum value is in the Netherlands, at 20.5 per cent. An overview of macro-regions once again brings out territorial disparities within the European countries. The phenomenon of the persistence of poverty is much more intense in less advanced macro-regions. Focusing on the period 1997–2001, the South of Italy and Spain, West and South-Western

70 Social Vulnerability in Europe

France, Central-Eastern Germany and the Central United Kingdom suffer from much more intense structural poverty than do their respective advanced regions. It is evident from these cases that welfare regimes calibrated on a national scale find it difficult to deal with the specific nature of regional problems. However, closer analysis of persistence rates measured over the longer period gives rise to other and more detailed considerations. In some countries, such as the Netherlands, Finland and Austria, the decline in poverty over 4 years is much more marked than over 2. In other countries, however, the decrease in the persistence measured over a long period is instead small. For example, out of 100 poor people in Belgium in 1997, only 44 were poor in 1999 but 42 of them were still poor in 2001. Ireland is the limiting case where those poor in 1997 were still poor in 2001 to a greater extent than in 1999. These marked differences in the downward trend of poverty may result from two different factors. The first is the different speed of exit from poverty. The economic mechanisms and the welfare systems of countries enable rapid exit from poverty in some cases, while in others exits take longer. The second likely reason for the high rates of persistence in the medium term is that a proportion of persons poor in 1997, but not in 1999, may have lapsed back into poverty in 2001. The structural and transient components of poverty are therefore flanked by a third one, which has elsewhere been termed recurrent poverty (Layte and Whelan, 2003: 175–180; Layte and Fouarge, 2004: 206–212). These are individuals and families who ‘navigate’ through life around the poverty threshold, rising above or falling below it according to events that affect either their capacity to produce income or the quantity and quality of their needs.

Where do the poor come from? Having established that the majority of people who are poor in 1 year are no longer poor in the next, we want now to determine where the poor come from, and where they go. We need to understand whether they concentrate in the income brackets immediately contiguous to poverty, as envisaged by a model of risk polarization, or whether they are distributed more widely across the entire income scale, as envisaged by a model of risk diffusion. To this end, we analysed the composition of entries into and exits from poverty by class of income, with the result as shown in Tables 4.5 and 4.6 for the period 1997–2001. We first focus on entry into poverty. At the European level, 38 per cent of the poor in 2001 had been poor in 1997. This is the ‘hard core’ of poverty: families who, over 5 years, have either been unable to exit poverty or have returned to it. An equal proportion of 38 per cent fell within the income brackets that we have called ‘at risk’. They are the families who 5 years before

Income Vulnerability in Europe

71

Table 4.5 Income class provenance of the poor, by country* Middle class

Affluent

35.1 40 29.4 40.9 45.4 33 30.3 39.1 33 29.4 49.2 36.7 53.6 38.5

32.9 16.8 16.8 14.6 9.6 11.6 14 15.3 10.9 21 21.6 17 8.6 14.3

8.5 7.8 3.8 1.5 3.1 2.3 3.5 2.8 2.9 3.7 6.4 3.5 0.7 6.9

7 3.3 4.8 4.8 2.6 1.8 2.9 6 0.9 5.3 4.2 3 0.9 5.4

100 100 100 100 100 100 100 100 100 100 100 100 100 100

38.0

16.0

4.0

4.1

–

Country

Poor

At risk

Denmark Netherlands Belgium France Ireland Italy Greece Spain Portugal Austria Finland Germany Luxembourg United Kingdom Europe**

16.5 32.1 45.2 38.2 39.3 51.3 49.3 36.8 52.3 40.6 18.6 39.8 36.2 34.9 37.9

Wealthy

Total

Notes: * Income class in 1997 of those who are poor in 2001. Percentages shares. ** Unweighted average. Source: ECHP, authors’ own calculations.

Table 4.6 Income class destination of the poor, by country*

Country

Poor

At risk

Middle class

Denmark Netherlands Belgium France Ireland Italy Greece Spain Portugal Austria Finland Germany Luxembourg United Kingdom Europe**

29.5 24.5 40.7 37.8 51.8 44.5 48 32.3 43.3 34.7 37.2 38 30.7 38.6 38

40.6 31.2 41.1 36.3 26.4 27.5 28.6 29.8 34.4 41.1 28 38.5 52.3 37.1 35.2

15.3 22.9 11.9 16.2 19.9 18.4 13.4 25.1 18.1 17 27.8 16.7 14.3 15.6 18

Affluent 9.7 13 1.9 5.8 1.8 4.2 5.1 6.2 2.4 4.1 4.1 5.4 0.5 3.1 4.8

Wealthy 4.9 8.4 4.4 3.9 0.1 5.4 4.9 6.6 1.8 3.1 2.9 1.4 2.2 5.6 4

Notes: * Income class in 2001 of those who were poor in 1997. Percentages shares. ** Unweighted average. Source: ECHP, authors’ own calculations.

Total 100 100 100 100 100 100 100 100 100 100 100 100 100 100 –

72 Social Vulnerability in Europe

had already been economically vulnerable and had since moved downward on the income scale. Some 16.1 per cent originated from central income classes, and less than 10 per cent from all the other classes. These are families who 5 years before seemed to enjoy relative security but were then affected by negative events and lapsed into poverty. These data show that past poverty is still the main source of future poverty, but they also show that the majority of the poor of 2001 were not poor 5 years before. The shift from sections of the population at risk to the poor class is particularly marked, whereas the transition from the middle class to poverty, although less pronounced, is not negligible. There is therefore a concentration of risk at the income levels immediately above the poverty threshold, but also an extension of this risk to part of the middle class, while few households descend into poverty from the affluent or wealthy classes. However, the degree of mobility in income distribution, or its stickiness, differs greatly among countries. Denmark, Finland and Austria have a more widespread contribution to poverty by income bracket. At least 30 per cent of the poor originate from medium or higher income classes. At the other extreme stand Italy, Ireland, Portugal and Luxembourg, where the contribution to poverty is more closely concentrated in specific income classes. Less than 20 per cent of the poor come from the middle class or from the higher income classes; all other poor of 2001 were either also poor in 1997 or originated from income classes immediately above the poverty threshold. Thus a clear inverse association between income mobility and the degree of income inequality is apparent. Denmark, Finland and Austria are countries with low degrees of inequality, while Italy, Portugal and Ireland are marked by considerable distributive inequalities.

Where do the poor go? We now focus on the destinations of the poor. In Europe, 38 per cent of the poor in 1997 were still poor in 2001, while 35.2 per cent had moved into the three income levels immediately above; 18 per cent into the central income brackets and 8.8 per cent into the other higher income classes (Table 4.6). The picture that emerges is very similar to that depicted by Table 4.5. These data also exhibit the characteristics of, and limits to, upward income mobility. There is a structural component, of relatively large size, that remains in poverty for several years. There is also a large component that exits poverty, but not necessarily a situation of economic vulnerability and strong exposure to the risk of lapsing back into poverty. More than 54 per cent of those leaving poverty move into these income brackets, which can be considered ‘communicating vessels’ with poverty. Finally, a significant minority makes income ‘jumps’ into the middle class or, more rarely, affluent or wealthy income classes.

Income Vulnerability in Europe

73

There is wide national variance around this picture based on the European aggregate. The countries characterized by the greatest upward mobility, where the poor achieve the highest income jumps, are the Netherlands, Spain, Finland and Denmark. Almost 50 per cent of the poor in the Netherlands are able to move to levels of income security in the space of only 5 years. Lower values, but still above the European average, are recorded by Finland and Denmark, where 30 per cent or more of the poor in 1997 jumped at least four income classes by 2001. In all these countries, high upward mobility is associated with an egalitarian income distribution. An exception is Spain, where high upward mobility is associated with inequality in income distribution. At the opposite extreme stands a bloc of countries comprising Belgium, Ireland, Greece, Portugal, Austria, Germany and the United Kingdom. In Ireland, Greece and Portugal, low upward mobility is associated with a high degree of income inequality. Yet there is also low upward mobility in Austria, Belgium and Germany, although these are countries in which there is no marked degree of income inequality.

The profiles of poverty The longitudinal analysis of flows into and out of poverty has shown that the latter comprises the following: a component of persistent poverty, a component that moves only marginally, continuing to be economically fragile and a component that moves to income brackets representing relative security where poverty is an occasional episode in the life-course. This requires us to conduct deeper investigation into the various profiles assumed by poverty across countries. To this end, the population is divided according to the number of episodes of poverty experienced in the 5 years observed (see Table 4.7). For the sake of simplicity, we shall use never poor to stand for persons who have never experienced an episode of poverty, transient poor for those who have experienced a single episode of poverty, recurrent poor for those who have experienced two or three episodes, and persistent poor for those who have experienced four or five episodes (Layte and Fouarge, 2004). The European population that has experienced at least one poverty episode over the 5-year period – so- called longitudinal poverty – amounts to just over 20 per cent: a percentage about double that of episodes of poverty measured in a single year. The decidedly more common phenomenon is a single episode, which was experienced by 9.4 per cent of the European population, confirming the predominance of the transient component of poverty. The percentage of the population that has experienced two or three poverty episodes is much lower –4.6 and 2.5 per cent respectively – and which probably comprises the recurrent component of poverty. The proportion of persons who experience four or five poverty episodes is even lower – equal to 4.1 per cent. These are people who remain in permanent hardship, suffering severe material and psychological distress as a consequence.

74 Social Vulnerability in Europe Table 4.7 Proportions of people experiencing from 0 to 5 episodes of poverty over the 5-year period 1997–2001 by country Number of poverty episodes Country

0

1

2

3

4

5

1/(2+3+4+5)

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

84 87.7 83.7 81.4 72.9 74.8 66.8 70.7 73.8 83.9 86.9 85.5 89.4 79.4

9.4 6.9 8.9 9 10.4 10.1 12.3 13.5 10.2 7.5 5.4 6.6 4.5 8.6

3.5 2.8 2.6 4.2 6 5.1 6.1 7 4.7 3 3.6 3.5 2.6 5.1

1.8 1.5 2 2.2 3.3 3.4 5 4.2 2.9 1.8 1.9 1.4 1.2 2.6

0.9 0.7 1.4 1.6 3.1 3.6 3.9 2.8 3.6 1.9 1.6 1.5 1.3 2.1

0.3 0.3 1.4 1.6 4.3 3 6 1.7 4.7 1.8 0.7 1.5 1 2.1

1.4 1.3 1.2 0.9 0.6 0.7 0.6 0.9 0.6 0.9 0.7 0.8 0.7 0.7

Total

79.4

9.4

4.6

2.5

2.1

2

0.8

Source: ECHP, authors’ own calculations.

This breakdown of poverty into its transient, recurrent or persistent components produces specific profiles for different countries – as is clear from the last column in Table 4.7, which shows the values of the ratios between those experiencing one poverty episode and those experiencing from two to five. In a first block of countries the recurrent and persistent components are relatively large. This block consists of Portugal, Greece, Italy, Ireland, Luxembourg, Finland and the United Kingdom. In the two Mediterranean countries, Portugal and Ireland, the large weight of this component is associated with a high poverty rate: many people are poor and many of them permanently poor. Luxembourg and Finland exhibit a different pattern, with few poor people but with characteristics of high stability. At the other extreme is a second block of countries, comprising Denmark, the Netherlands and Belgium, with a component of transient poverty that is very large compared with chronic and recurrent poverty. These are countries with relatively egalitarian income distributions and low poverty rates. The third block comprises France, Austria, Germany and Spain, which occupy intermediate positions. Spain has the distinctive feature of associating a high poverty rate with a large transient component. In regard to macro-regions (not reported in the table), the persistent and recurrent poverty components assume substantial proportions in Southern Spain, Central and North-Western Spain, and Southern Italy. There is an evident association between the high rates of poverty in these regions and

Income Vulnerability in Europe

75

its persistence. A large component of non-transient poverty is also recorded in the Northern United Kingdom.

Longitudinal and cross-sectional poverty The analysis of the number of episodes of poverty makes possible an estimate and a comparison of longitudinal and cross-sectional poverty rates. Two caveats are important in interpreting the estimate of longitudinal poverty. The first caveat is that this estimate depends on the length of time considered. The longer the period, the higher the absolute value of the longitudinal poor is. The estimate should then be considered as an order of magnitude but should not, however, bias the comparison among countries. A second caveat is that interpretation of the estimated figures has some margins of ambiguity. A large number of longitudinal poor, in fact, may be the result of both a high number of poor each year and a high poverty turnover, cases which must be kept distinct. The ‘longitudinal poverty rate’ in Europe20 stands at about 20 per cent, which is just over twice the mean cross-sectional poverty rate in the same period (see Table 4.8). This result matches the findings of similar research (Layte and Whelan, 2003: 176). The differences between countries are marked, ranging from minima of 10.8 and 12.5 per cent in Luxembourg and the Netherlands and maxima of 33.1 and 29.0 per cent in Greece and Spain. These differences broadly reflect the pattern found for cross-sectional poverty rates. In the Mediterranean countries, Portugal and Ireland, more than one person in four has been poor for at least 1 year in 5. In the Scandinavian countries, the Netherlands, Austria, Belgium and Luxembourg the same figure is 1 in 6.25 or more. The correlation between longitudinal poverty and cross-sectional poverty is imperfect because countries also differ in their poverty turnover rates. This is evident from the ratios between the two estimates of poverty reported in the last column of Table 4.8. In countries where the annual turnover of poverty is high, such as the Netherlands and Denmark, the number of longitudinal poor is also relatively large compared with the number of cross-sectional poor. In other countries, such as Greece, Ireland, Italy and Portugal, the reverse is the case, because high poverty rates are associated with low turnover rates. There are countries with low poverty rates and low turnover rates, like Finland and Austria, but also countries with high poverty rates and high turnover rates like Spain. These differences obviously result from structural differences in national economies and welfare systems and do not necessarily depend on the degree of distribution inequality. Turning to the examination of macro-regions, a striking pattern of territorial longitudinal poverty emerges. The strongest contrast is exhibited by Italy, which contains the area with the second highest European rate but also the area with the absolute lowest rate. In the 5 years of our analysis,

76 Social Vulnerability in Europe Table 4.8 Longitudinal and cross-sectional poverty: estimated number and share of people

Country or macroregion Denmark Netherlands Belgium France Centre-East South West and South-West Île- de-France Ireland Italy North-West Centre and North-East South Greece Spain Centre and North-West East and Madrid South Portugal Austria Finland Germany North Centre and East West South Luxembourg United Kingdom North Centre East and South-East West and South-West Europe

Longitudinal Population* poverty

Crosssectional poverty

Population*

Longitudinal poverty/ Crosssectional poverty

856,782 1,936,093 1,685,114 10,661,086 4,734,762 1,459,665 3,020,457 1,446,202 1,031,556 14,103,083 1,855,151 2,545,945

16.14 12.48 16.34 18.54 17.91 21.14 22.05 13.83 27.09 25.15 12.55 14.03

290,360 666,075 670,080 4,425,637 1,980,898 575,024 1,345,786 523,929 478,989 6,884,025 642,665 995,471

5.47 4.29 6.50 7.70 7.49 8.33 9.83 5.01 12.58 12.28 4.35 5.49

2.95 2.91 2.51 2.41 2.39 2.54 2.24 2.76 2.15 2.05 2.89 2.56

9,701,987 3,310,073 11,403,702 4,434,053

41.93 33.14 29.07 37.91

5,245,889 1,695,339 4,760,738 1,938,840

22.67 16.97 12.13 16.57

1.85 1.95 2.40 2.29

3,364,282 3,605,367 2,610,330 1,232,261 688,395 11,542,815 1,884,909 2,980,192 2,300,658 4,377,056 47,764 12,366,320 3,074,253 2,839,009 3,006,127 3,446,931 73,475,372

17.67 42.44 26.66 16.05 13.18 14.56 15.40 16.97 13.51 13.50 10.81 20.76 24.17 22.76 16.38 21.53 20.43

1,204,790 1,617,108 1,333,140 544,994 292,421 4,974,174 718,586 1,257,673 911,322 2,086,594 20,706 5,454,572 1,441,205 1,368,339 1,145,981 1,499,047 32,491,249

6.33 19.03 13.61 7.10 5.60 6.28 5.87 7.16 5.35 6.44 4.69 9.16 11.33 10.97 6.24 9.36 9.03

2.79 2.23 1.96 2.26 2.35 2.32 2.62 2.37 2.52 2.1 2.31 2.27 2.13 2.07 2.62 2.30 2.26

Notes: * Shares on average population over the period 1997–2001. Source: ECHP, Authors’ own calculations.

almost 42 per cent of the population of Southern Italy experienced at least one poverty episode, compared with less than 13 per cent of the population in the North-Western part of the country. A difference of similar intensity is apparent between the South of Spain and the Eastern part of the country and Madrid. In the United Kingdom and France, the disparities are principally

Income Vulnerability in Europe

77

Table 4.9 Longitudinal poverty and 2001 income class: absolute values, percentages and class probability of experiencing poverty

Absolute values

Percentage

Class probability of experiencing poverty

Poor At Risk Middle Classes Affluent Wealthy

32,985,515 23,509,773 12,321,670 3,330,574 2,931,823

43.9 31.4 16.4 4.4 3.9

100 28.3 10.3 5.9 4.2

Total

75,079,355

100

20.7

Income class

Source: ECHP, authors’ own calculations.

between Greater London and Île-de-France, on the one hand, and the remaining British and French macro-regions on the other. The disparities recorded in Germany are insignificant. The area of longitudinal poverty comprises subjects with very different levels of income. Table 4.9 shows the distribution of longitudinal poor by 2001 levels of income. Among those in Europe who experienced at least one poverty episode in the 5 years considered, in 2001 43.9 per cent were in poverty, 31.3 per cent in the at-risk groups, 16.4 per cent in the middle classes and 8.3 per cent in affluent and wealthy classes. There is therefore a concentration of longitudinal poverty in classes contiguous to the poverty threshold, which extends to include a part of the middle class but only marginally affects the higher income classes. This finding is confirmed by analysis of the probability of experiencing poverty in specific regard to each income bracket.21 Members of the affluent or wealthy classes have an approximately 5 per cent chance of experiencing poverty; those in the middle class a chance of around 10 per cent, and those in the at-risk brackets a 28 per cent chance. The curve of a hypothetical graphic of the distribution of probability by income classes would assume a convex shape with a steep negative slope in correspondence to low incomes, and a flatter one in correspondence to medium-to-high incomes. In this case, too, the general findings must be qualified when one moves to domestic cases (not reported in Table 4.9). The Netherlands and Spain, for example, exhibit patterns of a wider spread of poverty risk with a presence of longitudinal poor in medium or high classes greater than 30 per cent. Ireland is the opposite case, with about a 15 per cent presence in the same classes.

The vulnerable and the persistent poor We now examine the features of individuals defined in our analysis as ‘longitudinal poor’. What type of people are they? In what families do they live?

78

Social Vulnerability in Europe

What is their relationship with the labour market and the welfare system? What other distinctive characteristics do they have? (EC, 2008; OECD, 2008b). The results of this analysis are set out in Tables 4.10–4.12 with reference to 14 Western European countries. The people experiencing poverty episodes over the 5-year period 1997– 2001 have been divided into two categories according to the number of poverty episodes: the vulnerable, who experienced from one to three episodes of poverty; and the persistent poor, who experienced four or five episodes. In total, the examined population represents 20.4 per cent of the European population, the vulnerable 16.2 per cent, and the persistent poor 4.2 per cent. The three tables present the socio-demographic, economic and structural factors most associated with vulnerability and persistent poverty. The socio- demographic factors are sex, age, marital status and type of family. The economic factors are the number and sources of income, employment status and type of employment contract. Finally, the third category comprises features that are permanent and might barely change over the short term. These are level of education, social class approximated by type of employment, social capital or the presence of loans. The figures refer to the deviations in percentage points between the share of vulnerable individuals or the permanently poor in respect of the total 1997 population. The ‘plus’ sign signifies that a particular condition – being a woman rather than being unemployed – is over-represented among these groups; and the ‘minus’ sign – belonging to a family with two income earners rather than one fulltime worker – denotes that this condition is under-represented. A first feature clearly apparent in all three tables is the significance of the distinction between vulnerability and persistent poverty. The variables associated with vulnerability are the same, but the magnitudes of shifts to positive and negative are systematically greater for the persistently poor. There may be various reasons for this difference. One is that the same negative events have different impacts if they affect families with different structural characteristics. The capacity to react or to obtain support is greater, ceteris paribus, among families with higher educational levels, belonging to higher social classes or with greater endowments of social capital. For members of these families, a negative event in life may produce vulnerability, from which they are able to emerge by relying on their own individual or kinship resources. For those who do not possess such resources, a poverty episode may turn into a permanent trap. A second explanation is that persistent poverty occurs in cases where several hardship factors accumulate. Recent research on the causes of poverty has shown that individual factors of hardship become decisive in generating poverty when they are compounded by others or, alternatively, when another compensatory factor does not intervene (EC, 2008). Even one of the conditions of greatest hardship – such as

Income Vulnerability in Europe

79

being a separated woman with dependent children – may not be decisive if the woman is able to find and keep a job. We now turn to the main associations between vulnerability or poverty and the characteristics of a person and his/her family (see Table 4.10). Being female is a disadvantage for women living alone whether they are separated or divorced or widowed or simply single. Living alone is, however, a relevant risk factor for all. This might depend on the inability to share the costs of collective goods or to combine incomes or it might result from the difficulty of reconciling caring with paid employment. Being single becomes particularly problematic when the sole source of income is the welfare state, a condition closely associated with persistent poverty. Couples, even those with children, can more easily maintain an adequate standard of living, provided that they do not have more than four children. Table 4.10 Demographic and social factors connected with vulnerability and persistent poverty (difference in percentage relative to the whole population)

Categories of vulnerability ==>

Vulnerable people

Persistent poor

Variables Gender Marital status

Female Separated Divorced Widow/widower Never married

3.6 0.6 0.2 0.7 6.3

5.2 1.2 2.1 4.3 2

Number of family components

1 2 3 4 5 6 7 or more

2 4.9 2.4 1.4 3.5 1.7 1.5

9.9 5.9 8.5 6.2 2.9 4.5 3.3

Family type

Single welfare Single work Couple welfare Couple work Couple with children > 18 Couple one worker with children < 18 Couple two workers with children < 18 Single parent

2 0 1.4 4.9 2.4

11.5 1.6 2.1 8.1 4.3

5.7

8.9

7.0

10.3

1.8

2.7

Source: ECHP, Authors’ own calculations.

80

Social Vulnerability in Europe

Families with children are at risk when the household is large and only one member is in employment. In this case the evident discrepancy arising between income and family needs makes these households highly exposed to vulnerability. Table 4.11 Economic factors connected with vulnerability and persistent poverty (difference in percentage relative to the whole population) Categories of vulnerability ==>

Vulnerable

Persistent poor

Variables Sources of family income

No income One work income Two + work incomes One welfare income Two + welfare incomes Mixed incomes

2.5 4.1 12.1 6.7 0.2 1.4

2.5 1.7 19.1 22.8 5.6 13.5

Number of income receivers

1 2 3 4 or more

10.6 7.9 2.0 0.7

24 12.7 7.6 3.6

Main income source

Wages or salaries Self- employed Pensions Unemployment benefits Other social benefits Rents

18.8 6.9 0.3 2.5 6.4 2.7

36.1 2.2 10.6 6.4 13.6 3.3

Main activity

Paid work 15 hours or more Self- employed 15 hours or more Student Unemployment Retired Unpaid family work 15 hours or more Others inactives Total inactives and unemployed

20.7

31.8

3.6

0.7

4.6 5.3 1.7 5.6

1.5 11.1 4.9 10.9

1.2 14.8

1.5 29.9

Full-time or part-time

Full-time work Part-time work

16.0 1

28.6 0.7

Type of work contract

Permanent contract Temporary contract

21.7 0.5

30.5 1.5

Experiences of unemployment in the last 5 years

Yes

Source: ECHP, authors’ own calculations.

5

9.4

Income Vulnerability in Europe

81

In contrast, families with children but with two income earners are relatively less vulnerable. We will now concentrate on the characteristics of the vulnerable and the poor with reference to the sources of income and the relationship with the labour market and the welfare system (see Table 4.11). The first confirmed facts are that the presence of two family members in employment or perceiving incomes is a greater guarantee of security and that work incomes reduce risks much better than incomes from welfare. Households most at risk are those living on a single welfare income. They are over-represented mainly in the category of the persistently poor. The presence of a second welfare income drastically reduces exposure to risk. Likewise, households with a single income are decidedly more vulnerable than those with two incomes. Households living on wages and salaries are generally much better off than those with other sources of income; the situation of households relying on unemployment benefits or other social benefits is particularly problematic. But self- employment and agricultural work are also associated with vulnerability. This results from the presence in these groups of a population with marginal positions in the labour market, and the greater instability of their income. The close association of pensioners with persistent poverty, but not with vulnerability, is notable. It is likely that persons receiving pensions deriving from a regular career are not particularly exposed to poverty, unlike those who receive a social pension. With regard to the main activity, there is an evident high concentration of inactive and unemployed persons among both the vulnerable and the persistently poor. Unemployed workers, persons still in education and those carrying out unpaid family work are over-represented among the vulnerable. In contrast, paid employment as the main activity ensures income security, especially if the employment is full-time and on a permanent contract. Finally, Table 4.12 groups a set of structural conditions that do not assume the nature of events. Citizenship, education level, sector of employment and position in a profession tend to define the income brackets within which a person can be located, but they are not conditions that can be changed easily or rapidly. Education matters a great deal (EC, 2008: 22). A very high proportion of the vulnerable have only basic qualifications; conversely the proportion with higher and advanced qualifications is very small. The difference between the persistently poor and the overall population is particularly marked. In addition, the type of work is important. Although almost all types of work are under-represented among the vulnerable or the poor, owing to the large weight of the inactive and the unemployed, the risk level is particularly high for all types of unskilled labour, as well as for workers in fishery

82 Social Vulnerability in Europe Table 4.12 Structural factors connected with vulnerability and persistent poverty (difference in percentage relative to the whole population) Categories of vulnerability ==>

Vulnerable

Persistent poor

Variables Education level accomplished

Still in education 3° level (Isced 5–7) 2° level (Isced 3) Lower than 2° Level (Isced 0–2)

1.8 9.3 8.5 15.5

0.8 13 16.2 27.3

Membership of a club or organization

Yes No

9.2 8.9

16.0 15.8

Present occupation

Legislators, senior officials, corporate managers Professionals Technicians and associated professionals Office and customer care clerks Services and sales clerks Agriculture, forestry and fishery workers Craftsmen Machine and plant operators Unskilled occupations

1.3

3.0

4.6 5.1

6.1 7.4

4.0

6.4

0.4 2.3

-3.9 3.8

1.3 1.2

4.7 3.0

Economic sector

Private sector (not-forprofit included) Public and semi-public sector

0.6

1.3

7.7

19.8

8.7

10.6

Industrial sector

Agriculture Industry Services

2.6 5.4 12.0

3.9 10.6 23.1

State of health

Very good Good Normal Bad Very bad

0.1 2.6 0.7 1.1 0.8

3.9 7.4 4.9 5.1 1.6

Home loan

Yes No

10.6 4.3

18.0 10.3

Source: ECHP, authors’ own calculations.

and agriculture. Professionals, managers, technicians, entrepreneurs and office workers are all more secure. Being the member of a club or an association is linked with lower income vulnerability. This finding signals the importance of social capital in the

Income Vulnerability in Europe

83

formation of income. However, that the causation operates in reverse cannot be excluded: people with greater income security have greater propensities or opportunities to cultivate social relationships. In conclusion, these findings confirm what several scholars maintain (Whelan and Maitre, 2008): that social class, even if it is classified under the heading of ‘old’ social risks, is still an important factor of inequality in income distribution. In addition, the state of health is a factor that sets apart the vulnerable and the poor. But this difference is particularly apparent among the persistently poor, whose health might suffer during a prolonged phase of income poverty or might, on the other hand affect their capacity to earn an income and meet the costs of health care. Finally, people who took out a mortgage are clearly under-represented among the vulnerable. The purchase of a house, even if with a bank loan, implies a capacity for saving that is clearly not associated with economic vulnerability.

Conclusions We can now draw the main conclusions from this inquiry into the economic dimension of vulnerability. Even before enlargement, Europe was made up of countries that differed profoundly, not only by levels of development but also by degree of inequality, structure of distribution and rates of poverty. The relative proportions of the poor, the people at risk, the middle class and the affluent and wealthy classes are very different between countries. Some of the latter are also traversed by deep territorial disparities that impact upon inequality and the structure of the distribution of national income. The presence of these differences prompts the question of whether a European ‘social model’ is a reality or rather a large-scale socio-political objective. The explanation furnished by a consolidated body of studies is that such differences result from different forms of welfare regimes. But the close association between the level of development and inequality, evident at both national and macro-regional levels, suggests that strictly economic variables, linked to historical lags in industrialization, are an equally important part of the explanation. The analysis of rates of persistence has confirmed the findings of the most recent literature on the heterogeneity of poverty. The component of permanent poverty that persists over time stands, in Europe as a whole, at around 40–50 per cent of the poor population, depending on the time period over which the persistence is measured. This component comprises people bereft of the human, social and financial resources with which to escape the poverty trap. In this case, too, the national and regional variance in poverty turnover rates is so wide as to indicate not only diversity in the

84 Social Vulnerability in Europe

level and composition of social spending, but also substantial differences in the mechanisms of socio- economic mobility. The constant turnover of the poor population raises the issue of the provenance and the destination of the poor. The results of our inquiry show that the majority of people who exit poverty remain in classes of relatively low income close to poverty, in what we have called the income classes ‘at risk’, and, moreover, that the majority of people entering poverty originate from these same classes. A small but significant proportion accomplish income ‘jumps’ from poverty to the middle class, and vice versa, while income mobility to or from high income brackets is very low. Those belonging in the at risk brackets have around a 30 per cent probability of falling into poverty, compared with a 10 per cent probability of the members of the middle class, and a less than 5 per cent probability for those in the medium-high brackets. These results do not support the hypothesis advanced in the literature that social risk is distributed across the entire income scale. Instead, they seem more congruent with the hypothesis that risk is polarized in low and medium-low income classes. This seems to hold for the majority of the countries examined, although vertical income mobility varies greatly across countries. This polarization of risk in the low income classes shows the existence of a heterogeneous population group that rotates around poverty, living in situations of uncertainty and exposure to risk. Besides the persistently poor, this group comprises people who experience poverty as a short episode in their lives, and by people who are instead recurrently poor because they enter and exit poverty intermittently. All together, the structural, transitory and recurrent components of poverty constitute a heterogeneous world including both people who are actually poor and those at high risk of becoming poor. We have estimated this area by using longitudinal poverty in the 5 years from 1997 to 2001 as a proxy. The dimension resulting from this estimation is 20.4 per cent of the European population with a national range from 10 per cent to around 30 per cent, and a macro-regional range from 10–42 per cent. The significance of the distinction among the various components of this mixed aggregate is matched by the analysis of its composition. The transitory or recurrent component exhibits features different not only from the section of the population that has never experienced poverty, but also, partly, from the population in permanent poverty. The most important factors determining the composition of this vulnerable and poor population, considered as a whole, relate to household type, relationship with the labour market and some structural characteristics. The relationship with the labour market is a crucial factor in the greater or lesser incidence of vulnerability and poverty in the population. This area comprises a relatively large proportion of persons who are unemployed, inactive or dependent on welfare. All of these social categories receive lower incomes

Income Vulnerability in Europe

85

on average than persons in employment. However, full-time workers with permanent contracts, whose wages are higher and more stable than those of other workers, are also under-represented. Moreover the nature of employment assumes even greater importance when it is associated with the household type. Singles and large families are generally over-represented in this area, especially singles living only on welfare incomes and large families with only one worker or income earner. Finally, this area consists largely of persons belonging to low social classes, with low educations and weak endowments of social capital. These distinctive features are much more accentuated in the persistent poor component than in the vulnerable group that experiences transitory or recurrent poverty episodes. It is likely that the persistence of poverty results from the accumulation of several factors or the greater impact of negative events on persons suffering from particular structural weaknesses.

Notes 1. It is inevitable that it should be larger; otherwise there would be a poverty turnover equal to zero. But what is surprising is the magnitude of this turnover. 2. The monetary income thus defined might in some circumstances not be a good indicator of the household’s purchasing power because it omits self-consumption, imputed income deriving for example from use of a self-owned dwelling, goods and services furnished by the state for free or at subsidized prices, gifts, accumulated wealth or, on the other hand, a debt to be repaid. Significant differences in the presence of some or all of these elements might cause distortions in comparisons among various types of family members or among families in different countries or macro-regions. 3. The hypothesis of equal participation in the household’s total income by each of its members is commonly adopted in studies on income poverty. Although extreme and not applicable to individual cases, it is considered sufficiently realistic in most family situations. 4. The adoption of equivalence scales acknowledges the different needs of the various members of the household and the presence of a significant share of non-rival goods in family consumption. There is a certain degree of arbitrariness in the choice of a specific equivalence scale. This is because there is no general consensus on the elasticity of needs to household size, i.e., how much additional consumption is needed in order to maintain the well-being of the family unaltered when the latter increases in size or changes in composition. The hypothesis commonly used in trans-national comparisons that the values of this elasticity are equal among countries is equally arbitrary. Our choice of the modified OECD equivalence scale is justified only by its widespread use in the literature. For more thorough discussion of equivalence scales see Boarini, Johansson and Mira d’Ercole (2006: 21–22). 5. Net disposable household income, stated in the national currency, has been transformed into equivalent household income (EHI) at purchasing power parities through the following three steps: (a) Conversion into euro-equivalents with the annual average exchange rates for 1997-98-99-2000-01 (Eurostat).

86

6. 7.

8.

9.

10.

11.

Social Vulnerability in Europe (b) Conversion into GDP purchasing power parity using the Eurostat conversion rates. (c) Conversion into per capita values through the modified OECD equivalence system. The use of purchasing power parities reduces, with respect to the exchange rate, the distortions relative to differences in the prices of non-tradeable goods among countries. For the sake of consistency, the values of all the variables used in the text refer to these years, even when more recent values are available. The choice of the median, instead of mean, makes the index of the central trend less sensitive to the very high values of income distribution. In particular, this distribution, as well known, is heavily skewed leftward, so the value of the median, sensitive to the concentration of the population in the medium-low levels, is less than the value of the mean. In the income distributions of Western countries the level of income used to define the poverty threshold (50 per cent of average income) is not dissimilar to 60 per cent of the median income (see Baldini and Toso, 2004: 100). However, it should be borne in mind that monetary income is an imperfect indicator of the standard of family life. In particular, families in Sweden and in the other Nordic countries receive very high amounts of public in-kind transfers estimated by the OECD at around 40 per cent of monetary income, compared to the much lower values, often half, of the other European countries (see OECD, 2008b: 223–251). Note that the hierarchy of income among countries described here may differ significantly from the hierarchy that would result if other indicators were used. In fact, alternative measures may be employed to estimate and compare the level of well-being of individuals in various countries on the basis of income. At the aggregate level of the entire economy the one most frequently used is gross domestic product (GDP) at factor cost per capita. More correct for the purposes of measuring the income of residents are Gross National Income (GNI), which takes account of incoming and outgoing transfers of foreign income, and net national income (NNI), which also takes account of depreciation. In general, these three income indicators refer to the entire economy and do not alter the hierarchy of the countries measured in terms of purchasing power standards proportional to the resident population. More functional measures with which to estimate living standards, which is the focus of the analysis here, refer to household consumption or alternatively to household income. Consumption is the value of goods and services actually purchased by households, while disposable income is the capacity of current consumption and saving to finance future consumption. Their use, because these aggregates do not refer to the entire economy and have components in part different from GDP, inevitably gives rise to different hierarchies of income among countries. Finally, it should be noted that household incomes can be estimated on the basis of national accounts or sample surveys. These two very different methodologies produce different results (see Boarini, Johansson and Mira d’Ercole, 2006: 9–17). It should be borne in mind that purchasing power parities are calculated nationwide and do not take account of differences in price levels at macro-regional level, which are likely to cause real differences in living standards. The Gini index is the indicator of inequality most widely used by studies on the distribution of income. It should be borne in mind, however, that the degree of inequality is sensitive to both the indicator and the scale of equivalence adopted.

Income Vulnerability in Europe

12.

13.

14.

15.

16. 17.

18.

19.

20. 21.

87

Recent studies have shown that the particular combination adopted makes little change to the overall hierarchy of countries in regard to the degree of inequality, but it may produce large shifts of individual countries (see EC, 2008: 7–17; OECD, 2008b: 24–26). A typical cases would be the Netherlands and Belgium, with overly high indexes of inequality, and Austria, which would be too low in relation to the classification in terms of welfare systems (see Muffels, Frick and Uunk, 2004: 176). For empirical verification of the validity of the theory of welfare regimes see also Maitre, Nolan and Whelan (2005:157–171). A similar association between the degree of inequality, according to the Gini index, and the level of development, measured as per capita GDP, emerges for the 27-country Europe in 2004 (EC, 2008). Equivalent household income (EHI) was used to calculate the national and the EU-15 median and to classify individuals into the 21 EHI classes, which are defined as follows: • class 1 = EHI from 0 to 10 per cent of the median EHI value • class 2 = from over 10 to 20 per cent of the median EHI value • class 5 = from over 40 to 50 per cent of the median EHI value (the upper extreme of class 5 was used to set the poverty threshold) • class 6 = from over 50 to 60 per cent of the median EHI value (this is the first class of the non-poor) • class 21 = from over 200 per cent of the median EHI value to the maximum EHI of the individuals in the population. To define poverty, we adopt a concept of relative poverty based on the distance from the country’s median income. This method is widely used in the literature, and it is based on recognition of the historically determined character of a socially accepted standard of living. It is not without drawbacks, however, especially when employed in dynamic analysis. For a brief description of the various methods used to measure poverty and their shortcomings see Nolan and Whelan, 1996, pp. 10–28. Note that the data on regions are not strictly homogeneous with national data because they are calculated on the basis of national medians, not regional ones. A similar exercise was carried out for 2004 using EU-SILC data on the EU-27 countries in European Commission, Directorate-General “Employment, Social Affairs and Equal Opportunities,” 2008, pp. 70–79. This conclusion changes only partly if one considers the composition of the poor by country after enlargement. Using 50 per cent of the median income as the poverty threshold, 60 per cent of the poor live in the new member-states, 35 per cent in Poland alone; 40 per cent still live in the EU-15 countries, of which 10 per cent in Spain, and 8 per cent in Italy (see European Commission, DirectorateGeneral “Employment, Social Affairs and Equal Opportunities,” 2008, p. 75). The persistence rate is the ratio between the number of those who were poor in both observations and the number of all those who were poor in the first observation. The number of people experiencing at least one poverty episode between 1997 and 2001 in the average population during the same period. The probability is calculated as the ratio between the number of the longitudinal poor and the number of members of each 2001 income class. By definition, 100 per cent of those who were poor in 2001 had experienced at least one poverty episode.

5 Unstable Employment in Western Europe: Exploring the Individual and Household Dimensions Ivana Fellini and Mauro Migliavacca

Labour flexibility and social vulnerability in a comparative perspective In recent decades the diversification of forms of employment has been a leading mechanism in transforming national labour markets and social structures in Western Europe. From a comparative perspective, the spread of atypical, flexible and/or unstable forms of work, driven by the growing demand for flexibility and the pressures of international competition, shows significant differences among countries in labour market structures, employment regulation systems (OECD, 1999; EC, 2006), service sector development (Reyneri, 2005) and the quantitative- qualitative characteristics of the labour supply (OECD, 2002).1 In the early 1990s, the debate on new forms of employment in Western European countries focused on the relationship between flexibility and employment growth, exploring whether the increase in non-standard employment would enhance the matching of labour demand with supply (OECD, 1999; EC, 1999). Empirical findings have highlighted the ambivalence of the process of labour-market flexibilization and generated debate on the individual and social sustainability of flexible work. In particular, the exposure of non-standard workers to a higher risk of discontinuous employment and income greater discrimination in training and career opportunities, as well as to likely closer constraints on family formation and other course- of-life choices, has highlighted that employment flexibility may be an important dimension in social exclusion patterns (Purcell and Purcell, 1999; TaylorGooby, 2004a; Paci, 2005) where risk is assuming new features (Beck, 1998; Schmid and Shömann, 2004). Socio- economic research is still exploring which dimensions of flexibility are connected with greater risks of occupational and social entrapment, exclusion and marginalization, with an especial focus on the characteristics 88

Unstable Employment in Western Europe

89

of flexible/atypical workers and the consequences of labour market instability on life- courses (Barbieri, 2007), and the specificities of national labour markets (adopting a meso and micro sociological perspective). From a comparative perspective, instead, macro-institutional approaches that focus on external flexibility are more developed and have been particularly important for policy design. They explore employment protection strictness (including the regulation of non-standard contracts) and the trade-off between regulatory systems and labour-market efficiency, 2 generally distinguishing between ‘job security’ and ‘employment security’ (Wilthagen and Tros, 2004). The social vulnerability approach is considered less in comparative analysis when dealing with labour-market flexibility, although its theoretical and analytical suggestions are often (explicitly or implicitly) used. Indeed, trans-national studies on the characteristics of temporary workers and their chances of entering stable employment consider that non-permanent positions expose workers to greater risks of instability in the labour market and to significant consequences on their life- courses; thus they seek to identify those dimensions that are more likely to foster socio- economic exclusion. In the comparative analysis of labour flexibility, it is instead the dynamic/ interaction among the individual, familial and social dimensions, distinctive of the social vulnerability approach, that has been little explored, although the role of institutional variables and of household structures has been widely considered (Esping-Andersen, 2002). This chapter analyses non-standard employment in Western European countries, combining a comparative approach that explores the individual characteristics of flexible/unstable workers with an analysis of the familywork system. Thus, analysis of household structures goes along with analysis of the employment positions of household members, exploring the role of unstable employment and its consequences on access to economic resources. The chapter is organized as follows. The next section presents the definition of instability used in the analysis: comparative studies usually measure flexible employment through fixed-term employees whereas here other forms of dependent employment with no guarantee regarding the stability of working conditions are considered, as well as forms of self- employment presumed to be unstable.3 The following five sections analyse the individual dimension of non-standard employment, exploring the characteristics of unstable workers and their short-term labour-market histories. The ensuing three sections focus on households and explore the different work-family models when considering unstable employment as a peculiar form of employment, while the final section draws together our conclusions. As regards the definition of unstable employment, a preliminary note may be useful. The literature gives different labels to non-standard employment relations: alternative work arrangements, non-traditional work arrangements,

90 Social Vulnerability in Europe

flexible working/staffing practices, atypical employment, vulnerable work, precarious employment, new forms of employment and several others (Kalleberg, 2000). In this analysis we mainly use ‘unstable employment’ because we focus on different forms of non-standard employment that do not afford any guarantees as to the continuity of working conditions. We also use ‘non-standard employment’ because we consider these forms of employment to be different from standard dependent employment or from traditional self-employment; we also often use ‘flexible work’, because all the forms of employment considered here are forms of external flexibility (Wilthagen and Tros, 2004).

The instability index One of the assumptions of this analysis is that many of the new forms of inequality comprised in the concept of social vulnerability range along the employment stability/instability continuum. To forestall a simplistic correspondence between employment flexibility and social vulnerability, one must identify a criterion with which to discriminate between the different forms of employment. Descriptive indicators can then be used to identify the vulnerabilizing potential of different employment situations. For this reason, in what follows an indicator of stability is proposed: this is not intended to be an indicator of work vulnerability in the strict sense (an unstable worker is not necessarily a vulnerable one), but a means by which to discriminate between conditions so that exposure to different vulnerability dynamics can be evaluated. Stability – by which is meant the guaranteed continuity of employment (and therefore of income) – is a discriminatory dimension in keeping with the more consolidated tradition of socio- economic studies on labour flexibility (Reyneri, 2005). From this perspective, the type of contract is a direct indicator of the extent to which employment continuity is guaranteed, although the complexity of national labour-market regulatory systems should be borne in mind. Employment stability, and consequently instability as well, are determined by contractual forms indicative of the degree of flexibility that a socio- economic system institutionalizes. The contractual dimension is therefore one of the main elements in the operationalization of the concept of stability, particularly as regards forms of dependent employment. Yet it is not the only one, for this analysis also seeks to verify the concept of stability with reference to self- employment, for which contractual information is obviously not available.4 On the basis of these considerations, in what follows we use information taken from the ECHP database to develop a set of indicators with which to construct an index of job stability.5 But what does make work unstable? What are the factors responsible for its stability or instability? The principal feature of stability is the temporal

Unstable Employment in Western Europe

91

continuity of employment, and therefore also of income. As regards dependent employment, in this study we consider as stable all workers on open- ended contracts. Using the same criterion, we consider as unstable all workers employed on fixed-term contracts of various kinds (fixed-term or short contracts, different forms of paid apprenticeship and casual work with no contract). Part-time work is consequently considered stable if it exhibits the characteristics of a stable job (an open- ended contract) and unstable when it does not (fixed-term contracts of various kinds). As regards the self- employed, because it is not possible to use contractual information, we have had to formulate a different definition. While in the literature and in European directives (Barbieri, 2001; Eurostat, 2003b; Arum and Müller, 2004), self- employment is often treated as an undifferentiated whole, here we identify the self- employed according to their position along the stability/instability continuum. Given the information available, we use the number of hours worked per week as an indirect proxy for the degree of work continuity (OECD, 2000).6 On this basis, we have decided to set a threshold that, taking national differences into account, discriminates between the stable selfemployed and the unstable self- employed. Among the many levels possible, this threshold has been fixed at 30 hours a week, because this is a good compromise among national specificities. Considered to be stable, therefore, are self- employed workers who work 30 hours or more a week; and considered as unstable are those who work fewer than 30 hours a week.

Unstable workers: the individual dimension The instability index enables the identification of dependent and selfemployed workers with no or few guarantees on the continuity of their working conditions. In the next four sections descriptive indicators are used to highlight: a.

b.

the importance of unstable employment in Western European countries. The analysis also deals with the characteristics of unstable employment and its recent trends, considering that the strengthening (or conversely the weakening) of well-known segregation factors (gender, age and educational attainment) may be indicators of the greater vulnerability of specific population/labour-force groups; short- and medium-run changes in the labour-market position of unstable workers. The hypothesis is that in labour markets where movements among statuses (stable employment, unstable employment, unemployment and inactivity) are easier and faster workers are less vulnerable than in those where it is less easy to exit unstable employment and/or enter stable employment. Employment instability is not

92

c.

Social Vulnerability in Europe

a vulnerabilizing factor in itself; rather, it is vulnerabilizing where it is more difficult to exit instability or where the connection between instability and fragmented labour-market careers is stronger. the generalized limited access to economic resources experienced by unstable workers, and the different roles played by the welfare state in supporting their incomes.

Comparative analysis of the characteristics of non-standard workers, of their chances of entering stable employment (or unemployment and inactivity) and of the availability of economic resources will make it possible to identify those Western European countries where unstable employment is more likely to be connected with individual vulnerability.

Incidence and main characteristics of unstable workers across Western Europe According to ECHP data, in 2001 unstable workers represented approximately 17 per cent of the employed in Western Europe, a value about 5 percentage points higher than that registered, again for 2001, when considering only temporary employees and according to the European Labour Force Survey data (Table 5.1). Hence, the incidence of unstable employment significantly increases with respect to the usual figures when casual workers with no contracts, workers with work arrangements other than permanent contracts and the presumably discontinuous self- employed are considered as well as workers on temporary contracts. The situation differs across countries, with values ranging from 11 per cent in Austria to 30 per cent in Spain (Table 5.1). The ranking of countries according to the incidence of unstable employment is rather different from the ranking based on fixed-term contracts (OECD, 1999), although the latter represent the largest component of unstable employment. The following situations can be identified: ●

Spain, Greece, Portugal, Germany and Ireland are the countries with the highest incidence of unstable work, with figures above the European average. Spain is an exceptional case in Europe owing to its very significant growth of temporary (dependent) employment in the 1990s. The high incidence of non-standard employment in Greece and Portugal mainly results from the importance of informal work (e.g. forms of family employment) and of unstable self- employment. Germany exhibits a significant incidence of unstable employment as a result of both fixedterm contracts supporting the labour-market entry of young people (e.g. paid apprenticeships and training under special schemes related to employment, which account for more than one-third of unstable employment) and to casual work. In Ireland, the importance of non-standard

Unstable Employment in Western Europe

●

●

93

employment combines a significant weight of occasional workers (casual workers with no contract represent around one out of four unstable Irish workers) with a significant weight of apprenticeships and work-training contracts (about one out of five unstable workers). The United Kingdom, Denmark, Finland, Italy, France and the Netherlands are countries with significant incidences of non-standard employment, with figures ranging from 13 to 16.5 per cent. In this group of countries, too, the incidence of unstable employment depends on the structural features of their national labour-market regulatory systems, so it is quite difficult to deal with them as a homogeneous aggregate. In the United Kingdom people working very few hours a week account for significant share of non-standard employment, the rate of which would otherwise be significantly lower7; Denmark, in contrast, has a larger proportion of apprenticeships and in Italy unstable self-employment accounts for a quite substantial share of unstable employment.8 Austria and Belgium are the Western European countries with the lowest incidence of non-standard employment, which in both cases accounts for around 10 per cent of total employment.

When macro-regions are considered, some national cases show a territorially differentiated situation, but no transnational regional clusters are clearly identifiable. Indeed, in North-Western Italy around 10 per cent of workers are in non-standard employment, and the figure is more similar to those for Austria and Belgium than to those of around 20 per cent in Central and Eastern Italy. The Île- de-France records a less significant 11 per cent when compared to the 15–16 per cent of the other French macro-regions. In Spain, where the rate of unstable work is exceptionally high, the Southern regions have an unstable employment share close to 40 per cent, a much larger figure than the 26–30 per cent recorded in the other Spanish macro-regions. Taking account of unstable employment as derived from our instability index instead of the more usual (and restrictive) fixed-term employment rate, the European picture of flexibility partly changes (Table 5.1). On the one hand, Spain and Portugal remain at the top of the ranking of European countries where non-standard employment is more significant; on the other, in the other Mediterranean countries (Italy and Greece) the importance of non-standard employment increases significantly. Also the Anglo- Saxon countries show a significant increase in the share of unstable employment, becoming countries where non-standard employment is of great importance. This results from unstable self- employment and underemployment in the United Kingdom, and from casual work in Ireland. Once the Anglo-Saxon countries are repositioned, Austria and Belgium are still the countries with the lowest shares of non-standard employment, while the Scandinavian and the continental countries (in particular France and

Table 5.1 Share of unstable employment as a proportion of total employment, percentage composition of unstable workers and share of temporary employment on total employment – 2001 of which (% of unstable workers) Unstable workers as a share of total employment (%) Denmark Netherlands* Belgium France* Centre and East South West and South-West Île- de-France Ireland Italy North-West South Centre and NorthEast Greece Spain Centre and NorthWest East and Madrid South Portugal Austria Finland Sweden ***

Temporary employees

On trainingwork contracts/ schemes

Casual or no contract

Self- employed and family workers

Temporary employment (LFS)**

Other

13.4 15.7 (12.7) 11.9 14.3 (11.5) 14.8 15.7 15.7 10.8 17.2 13.9 9.9 20.5 11.7

28.4 17.8

26.5 61.1

34.4 6.7

9.2 12.5

1.5 1.9

8.6 12.6

66.2 79.5

8.8 0.0

6.1 14.8

10.1 3.6

8.8 2.1

7.5 13.2

76.0 86.7 75.7 91.4 17.8 40.2 39.6 36.8 43.9

– – – – 40.5 23.7 – – –

– – – – 24.7 13.4 – – –

– – – – 13.2 18.9 – – –

– – – – 3.8 3.8 – – –

– – – – 3.8 6.9 – – –

19.7 29.8 29.6

33.2 67.9 63.3

37.2 15.1 15.4

1.0 4.6 –

27.2 4.9 –

1.4 7.5 –

8.1 25.6 –

26.4 40.2 19.5 11.1 14.4 14.3

71.9 64.9 42.4 35.8 76.3 73.7

12.7 19.8 26.5 15.9 7.9 0.0

– – 4.2 33.0 2.2 8.4

– – 21.0 11.4 11.6 10.3

– – 5.9 3.9 2 7.6

– – 14.4 7.0 15.6 14.2

Germany North Eastern and Central West South Luxembourg* United Kingdom* North Centre Greater London West and South-West Total *

19.3 19.1 19.2 17.5 20.4 10.1 (8.1) 13.0 (7.1) 13.7 13.4 12.0 13.6

32.6 37.1 46.3 26.8 27.3 66.5

30.7 – – – – 0.0

30.7 – – – – 25.3

6.0 – – – – 8.2

0.0 – – – – 0.0

11.0 – – – – 4.0

66.1

0.0

1.4

31.7

0.8

5.9

79.2 66.9 61.0 66.1

– – – –

– – – –

– – – –

– – – –

– – – –

16.8 (15.8)

49.7

22.2

11.0

13.1

4.0

11.3

Notes: The figures in brackets unstable may not be significant. * The overall share of unstable workers also considers people working less than 15 hours a week and for whom no other information about their job is available. Figures at the net of this component are provided in bracket. Calculations on the composition of unstable workers were made not considering this component. ** This is the ratio between employed on fixed- or short-term contracts, including training contracts, and total employment as from Eurostat Labour Force Survey data. *** Data for Sweden do not provide information on training-work contracts. Source: ECHP, authors’ own calculations.

96

Social Vulnerability in Europe

Finland), which had the most significant share of temporary employees, move to intermediate positions. Whether considering the share of unstable employment by means of the instability index or the share of fixed-term employees, the picture of nonstandard employment in the Western European countries does not clearly reflect the well-known European welfare-state geography (Esping-Andersen, 1999). However, the figures derived using the instability index differentiate better between Northern and Southern Europe. Indeed, the Mediterranean countries show a greater importance of unstable work, although Italy, owing to differences between its Northern and Southern macro-regions, is far from the figures recorded for Spain, Greece and Portugal. Rather the figure for Northern Italy is closer to those recorded for countries of continental Europe. Moreover, on the one hand the United Kingdom and Ireland exhibit quite differentiated characteristics of the phenomenon, while on the other, the range of values recorded both in the continental and Scandinavian countries is narrow enough to make it difficult to aggregate them into homogenous groups of countries. Recent institutional analyses show that European countries can be grouped according to different degrees of labour-market rigidity, measured also by considering the regulation of fixed-term contracts (OECD, 2004). A study by the European Commission (EC, 2006) has analysed the rigidity of employment protection legislation, the regulation of fixed-term contracts, unemployment benefits and financial resources devoted to active labour-market policies in Europe and has identified five groups of countries. The AngloSaxon countries are characterized by high labour-market flexibility (looser employment protection legislation), low security (intermediate to low spending on labour-market policies) and low taxation. The continental countries (Germany, Belgium, Austria and France) have medium-low labour-market flexibility, medium-high employment security and medium-high taxation. The Mediterranean countries, with the exception of Italy, have a low degree of labour-market flexibility, relatively low employment security and diversified spending on labour policies. The Eastern European countries and Italy are characterized by a medium-high degree of labour-market flexibility, a high degree of insecurity and medium-high taxation. Finally, the Nordic countries (Denmark, Sweden, Finland) and the Netherlands combine high security, medium-high labour-market flexibility and medium-high taxation. However, these dimensions are only weakly connected with the share of non-standard or flexible employment.9 This is probably because external flexibility is linked to the characteristics of labour demand and of productive assets (dominant sectors, size and organization of firms and productivity levels). Thus, the characteristics of national and/or local economies (more or less tertiarized, more or less fragmented and/or ‘delocalized’, etc.), too, can define the demand for flexibility. Rather than the extent of the

Unstable Employment in Western Europe

97

phenomenon, the characteristics of flexible/non-standard employment, its persistence over time and the chances of exiting it (also resulting from the role played by the welfare model) can depict a clearer European geography of unstable employment (as shown below when the labour-market careers of unstable workers are discussed). Comparative analysis of the characteristics of unstable employment (Table 5.2) shows features that are common to most Western European countries, although figures may differ significantly across countries. First, unstable workers are relatively younger than the overall employed. This over-representation of young people is a feature shared by all the

Table 5.2 Total and unstable employment by sex, age, education and sector of activity – 2001 (%)

Under 35

Isced 3–4**

Service sector

Unstable workers

Women

Under 35

Isced-3–4

Service sector

Denmark The Netherlands* Belgium France* Ireland Italy Greece Spain Portugal Austria Finland Sweden Germany Luxembourg* United Kingdom* EU-15

Of which

Women

Total employment

46.1 40.6 44.6 44.1 40.4 37.6 37.6 38.1 45.0 43.8 47.4 47.1 42.5 39.4 45.4 42.5

36.0 41.5 37.9 36.9 48.8 37.0 36.7 46.0 43.1 44.6 32.1 34.7 35.0 40.8 37.6 39.1

53.1 na 34.8 11.8 44.3 45.2 38.1 20.1 13.3 72.8 45.4 50.2 57.4 37.3 20.2 33.7

68.8 77.3 72.4 70.6 61.9 65.0 63.7 61.8 52.3 61.0 65.1 72.8 60.3 78.5 73.1 65.8

13.4 15.7 11.9 14.3 17.2 13.9 19.7 29.8 19.5 11.1 14.4 14.3 19.3 10.1 13.0 16.8

43.7 51.8 65.3 48.6 46.6 50.4 53.9 46.1 54.5 56.1 60.4 70.8 52.4 54.1 54.5 53.2

63.2 61.3 53.1 68.5 53.1 58.1 50.9 64.3 58.4 70.2 58.7 49.7 55.5 57.0 44.0 58.1

43.0 na 33.8 20.6 41.7 40.2 38.0 18.3 11.5 45.7 51.2 44.5 47.9 30.7 28.3 30.1

65.3 78.8 84.0 68.7 66.5 59.9 61.2 60.6 50.3 64.7 78.7 86.2 67.2 83.1 85.5 66.1

Notes: na = not available. * Not considering people working fewer than 15 hours a week and for whom no information on their job is available. ** The ISCED - International Standard Classification of Occupations allocates secondary qualifications to level 3–4. Source: ECHP, authors’ own calculations.

98 Social Vulnerability in Europe

countries considered, although in the Anglo Saxon countries the gap between the share of under-35s in unstable work and in total employment is less significant. Conversely, the rate of under-35s in unstable employment as a proportion of total employment is particularly high in Denmark (63.2 versus 36 per cent), France (68.5 versus 36.9 per cent), Austria (70.2 versus 44.6 per cent), Finland (58.7 versus 32.1 per cent), Germany (55.5 versus 35 per cent) and Italy (58.1 versus 37 per cent). This is also because temporary contracts combining training and work are more likely to be addressed to younger workers. Second, unstable employment is also relatively more feminized than overall employment, but gender differences are less pronounced than age differences. This holds for all the countries considered, with the sole exception of Denmark, where stable female (part-time) employment in the public sector is likely to support female employment in permanent contracts: 43.7 per cent of Danish unstable workers are women, compared to 46.1 per cent of all workers in employment. France, Spain, Ireland and the United Kingdom show less marked gender differences than do other countries, whereas the gaps are much wider in Belgium, Austria, Italy and Greece. Third, the educational attainment of unstable workers is lower than that of the overall employed. Comparison between the educational attainment of non-standard workers and that of all workers should be conducted with more caution, because differences among national education systems may be not adequately captured by the classification of educational attainment levels.10 However, when secondary qualifications (ISCED 3–4) are considered, some countries show a greater share of non-standard workers with intermediate educational attainments compared with all workers; in particular, unstable workers in France, Finland and the United Kingdom have higher educational profiles.

Recent trends in the incidence of temporary workers From a dynamic perspective, data for the 2000s on fixed-term contracts only11 (see Figure 5.1) show a steady and general growth in the incidence of temporary workers in all Western European countries. After a 2-year stable trend (2001–2003), fixed-term contracts resumed their growth between 2004 and 2005 rising to 14 per cent of dependent employment and accounting for almost half of the additional employment created, albeit with national specificities (EC, 2006). In 2006–2007 this growth continued and fixed-term employment reached 15 per cent (see Figure 5.1 and Table 5.3). In absolute terms, from 2002 to 2007 in the EU-15 countries as a whole temporary employment grew by approximately 20 per cent, compared to the 7–8 per cent growth in both dependent and total employment over the same period. The growing incidence of temporary work has been accompanied by a further, though slight, growth in the share of temporary female employment

Unstable Employment in Western Europe

99

45 40 35 30 Total Women 15–24* Under 39*

25 20 15 10 5 0 2001

2002

2003

2004

2005

2006

2007

Figure 5.1 Share of temporary employees as a proportion of total employees – EU-15 Source: ECHP, authors’ own calculations. Eurostat LFS – Labour Force Survey.

Table 5.3 Ratio between the share of temporary workers and the share of total employment for women, for aged 15–24 and aged 24–39, for workers with ISCED (0–2) and for workers in the service sector – 2002 and 2007

Women

Aged 15–24 Aged 24–39 ISCED (0–2)

Service sector

2002 2007 2002 2007 2002 2007 2002 2007 2002 2007 Belgium Denmark Germany Ireland Greece Spain France Italy Luxembourg The Netherlands Austria Portugal Finland Sweden United Kingdom EU-15

1.39 1.23 1.05 1.32 1.25 1.14 1.17 1.33 1.10 1.21 1.02 1.10 1.24 1.21 1.17 1.23

1.30 1.19 1.02 1.33 1.30 1.10 1.12 1.31 1.14 1.13 1.04 1.05 1.28 1.20 1.17 1.19

4.04 3.02 4.72 3.02 2.78 2.31 3.79 3.14 4.05 2.69 5.28 2.70 3.14 3.55 2.11 3.19

Source: ECHP, authors’ own calculations.

4.09 2.75 4.36 3.10 2.93 2.23 3.97 3.69 5.37 2.70 4.40 2.91 2.93 3.52 2.47 3.19

1.00 0.99 0.78 0.65 1.20 1.16 1.06 1.12 1.00 0.89 0.52 1.21 1.10 0.99 0.87 1.03

1.03 1.07 0.93 0.77 1.26 1.17 0.99 1.15 1.03 0.99 0.72 1.30 1.15 1.01 0.85 1.09

0.97 1.51 2.58 0.99 0.97 1.10 1.11 0.99 1.11 1.32 3.03 0.89 0.97 1.25 0.74 1.34

1.29 1.34 2.45 1.07 1.02 1.14 1.15 0.98 1.12 1.38 2.73 0.90 1.10 1.44 0.73 1.38

1.13 1.11 1.07 1.22 1.15 0.89 1.05 0.99 1.10 1.01 0.98 1.17 1.15 1.14 1.12 1.01

1.09 1.07 1.06 1.22 1.13 0.90 1.02 1.00 1.07 1.03 1.05 1.14 1.15 1.12 1.11 1.01

100 Social Vulnerability in Europe

(from 14.4 per cent in 2001 to 15.6 per cent in 2007), but with no substantial changes in the gender gap, which has stood at around 1 percentage point throughout the 2000s. The gap widens considerably when age is considered: for workers aged under 40 the incidence of temporary work rose from 19 per cent in 2001 to 22 per cent in 2007 and for the youngest age brackets (15–24 years), the incidence of temporary workers rose from 38 to 42.4 per cent. Thus, age, and to a limited extent gender, have remained the distinctive features of unstable work during the 2000s, at least when employees on fixed-term contracts are considered. According to Eurostat Labour Force Survey data, the share of involuntary temporary workers has increased more markedly than that of such workers on fixed-term contracts: from 52 per cent in 2002 to more than 57 per cent in 2007. For women, the values rise from 51 per cent in 2002 to almost 60 per cent in 2007. In Western European countries as a whole, the relative increase in involuntary temporary workers has gone hand in hand with a change in the length of fixed-term employment. Very short-term contracts have increased in frequency, to the detriment of contracts of more than 2 years’ duration: in 2007 more than 46 per cent were for less than six months against 35 per cent in 2002. As seen, non-standard employment is differentiated by age and gender in all Western European countries. Recent trends in age and gender concentration can be used to understand whether flexibility increasingly involves specific labour-force groups or whether the phenomenon is assuming more transversal features. According to ECHP data (which allow analysis only of the second half of the 1990s), the feminization rate of unstable employment12 grew between 1995 and 2001 in many of the countries considered, indicating a greater concentration of women in unstable work. This growth has been generally greater than that recorded for women in stable employment and especially in dependent employment.13 With regard to age, between 1995 and 2001 in most countries the incidence of under-35s diminished with respect to both unstable employment and overall employment. The Mediterranean countries are distinctive in this regard because they have an increased share of under-35s in unstable employment combined with the decrease or stability of the incidence of under-35s in total employment. However, over a 5-year period, the age composition of unstable employment has changed significantly in most of the countries considered, with a relative growth in adult workers. The only exceptions are the Mediterranean countries, where the age composition of unstable employment does not exhibit this trend. The most recent Eurostat data on temporary employment confirm (Table 5.3) that fixed-term employment is age- and gender- differentiated in all of the countries considered. However, the data do not show a greater concentration trend by individual characteristics over the 2000s, although some interesting country differences can be found.

Unstable Employment in Western Europe

101

The coefficient of female representation for fixed-term employment in 2007 (i.e. the ratio between the share of women in fixed-term employment and the share of women in total employment) is greater than 1 in all countries. This confirms that women are everywhere over-represented in fixedterm employment. This is true to a lesser extent in Austria and Germany and to a very significant extent in Belgium, Finland, Greece and Italy. Between 2002 and 2007 the coefficient of representation diminished in most countries, with the exceptions of Ireland, Greece and Finland. This is probably because the growth of female employment during the 2000s14 has also concerned permanent employment to a significant extent. The very marked over-representation of young people aged between 15 and 24 was highest in 2007 in Belgium, Germany and Austria, where, as we have seen, work-training contracts represent an important share of fixed-term employment. In contrast, although Spain, the United Kingdom, the Netherlands and Denmark record considerable over-representations of younger workers in fixed-term employment, they have lower concentration indices. Between 2002 and 2007, the majority of countries recorded a widening gap between the share of young people aged 24 and under in temporary employment and in total employment. The trend is less marked in Germany and Austria – the Western European countries with the most well- established work-training systems – and in the Nordic countries. The growth in the share of younger workers in temporary employment is instead greater in the Mediterranean countries, although in these countries worktraining contracts are less widely used. The coefficient of representation for the 25–39 age group15 is everywhere much lower than for the under-25s, but it is higher in the Mediterranean countries (Portugal, Greece, Spain and Italy), while Finland has figures similar to these. This is the indicator that increases in the great majority of countries, highlighting an increase in fixed-term employment among older workers as well. As regards education, the index shows a similar trend for the employed with the lowest educational attainments, with some exceptions: it diminishes in Austria and Germany, which previously registered very high values, while in Italy and the United Kingdom it shows substantial stability, despite an under-representation of low educational profiles in fixed-term employment in comparison with the other countries. Hence, during the 2000s, unstable work, with specific reference to fixedterm contracts, has not undergone greater age and gender concentration trends, even though age and gender remain its distinctive features.

Mobility in the labour market The risk that non-standard employment, far from representing a stepping stone to stable employment, may be a dead end consisting of occasional

102

Social Vulnerability in Europe

jobs, discontinuous income and difficult life- courses has been and remains a widely debated issue. In this section the analysis addresses unstable workers’ labour-market histories to highlight the national differences in the patterns and timing of transitions to stable employment or to other labour-market conditions. The analysis is based on ECHP data, which although they are somewhat dated, make it possible to reconstruct a panel and to follow it over time focusing on employment rather than jobs, as well as to consider a broader range of unstable workers than workers on temporary contracts. Whether the previous labour market-position of unstable workers in 200116 (in the 3 previous years; see Table 5.4), or the subsequent position of unstable workers in 1998 (in the following 3 years17; see Table 5.5) is considered, in the very short term (1 year), unstable workers exhibit a significant persistence in their condition in all of the countries considered, albeit with some differences. If the panel of workers unstable in 2001 is considered, more than half of them were already unstable in 2000 in most countries, with the Netherlands and the United Kingdom recording a larger share of unstable workers in 2001 who in the previous year had been stable (more than one third) or inactive (about a quarter). Moreover, the Netherlands, the United Kingdom and Denmark are the countries with the largest proportions of unstable workers who had been inactive in previous years, while the Mediterranean countries – in particular Italy, Spain and Greece together with Finland – record a larger proportion of unstable workers who were previously unemployed. Overall, the continental countries, with the exception of Germany, and the Mediterranean countries together with Finland, have larger shares of unstable workers who, in the previous 3 years, were already unstable (around one third) compared with the Netherlands, Denmark and the United Kingdom. The condition of workers unstable in 1998 a year later highlights that the Netherlands remains the country with the lowest level of persistent instability, while Portugal has the highest level. Analysis of the transition to stable work indicates the Netherlands, the United Kingdom and Belgium as the countries with the highest rates of transition from unstable to stable employment, with values above 40 per cent, followed by Ireland and Austria, with values above one third. Finland (12.5 per cent), France (11.9 per cent), Italy (11.7 per cent), Spain (10.5 per cent), Greece (9.9 per cent) and, at a distance, Germany (7.4 per cent) record, in the short term, non-negligible shares of unstable workers who instead enter unemployment (see Table 5.5). In the short term (after 2 years), the proportion of unstable workers achieving stability increases everywhere, but the hierarchy of countries where there are greater chances of entering stable employment remains unchanged. In the Netherlands, the United Kingdom, Belgium and Austria, over half of workers unstable in 1998 were in stable employment 2 years later. By contrast, in Finland, Italy, Greece, Portugal, Spain and Germany, the proportion of unstable workers still in unstable employment after

Table 5.4

Unstable workers in 2001: labour market positions in the previous 3 years (%)

Unstable workers in 2001*

DK

NL

B

FR

One year before (2000) Non-permanent/unstable Permanent/stable Unemployed Inactive

52.6 21.3 (6.2) 19.9

38.6 34.7 (3.5) 23.2

50.1 23.9 (5.2) 20.8

53.9 14.9 12.4 18.9

Two years before (1999) Non-permanent/unstable Permanent/stable Unemployed Inactive

46.3 23.1 (7.2) 23.4

24.6 34.6 (7.1) 33.8

41.1 26.0 (8.1) 24.9

Three years before (1998) Non-permanent/unstable Permanent/stable Unemployed Inactive

27.7 26.7 (8.4) 37.2

23.9 33.6 (6.2) 36.4

34.8 29.3 (9.3) 26.7

IR

IT

GR

SP

PT

AU

FI

G

LU

UK

50.2 24.6 (6.5) 18.8

53.6 22.6 11.6 12.1

53.1 23.0 11.9 11.9

50.2 19.5 14.0 16.4

48.7 32.7 7.0 11.6

50.0 31.4 (6.8) (11.9)

52.6 14.0 18.6 14.8

52.2 23.7 7.1 17.0

39.6 24.2 (5.9) (30.3)

33.5 34.8 (5.3) 26.4

40.9 15.0 15.0 29.0

34.3 28.4 (10.9) 26.5

41.4 22.8 17.7 18.1

45.7 23.8 12.4 18.0

42.6 15.1 18.2 24.1

47.2 23.1 8.7 21.1

43.2 30.9 (6.3) 19.7

46.7 17.2 15.8 20.4

27.2 42.1 7.0 23.7

28.1 26.7 (3.8) 41.5

25.7 39.3 (2.5) 32.5

39.3 13.6 13.5 33.6

34.6 25.5 (9.8) 30.1

34.4 22.8 19.6 23.2

41.9 31.1 10.5 16.5

35.6 16.5 20.6 27.4

38.1 25.7 7.9 28.3

34.9 34.3 (9.7) 21.1

37.7 17.9 17.5 26.9

19.7 39.4 10.3 30.7

16.8 25.1 (14.0) 44.1

26.4 37.4 (4.0) 32.3

Notes: DK = Denmark, NL = the Netherlands, B = Belgium, FR = France, IR = Ireland, IT = Italy, GR = Greece, SP = Spain, PT = Portugal, AU = Austria, FI = Finland, G = Germany, LU = Luxembourg, UK = United Kingdom. The figures in brackets may be not significant. * Panel data for Sweden are not available. Source: ECHP, authors’ own calculations.

Table 5.5

Unstable workers in 1998: labour-market positions in the following 3 years (%)

Unstable workers in 1998* One year later (1999) Non-permanent/unstable Permanent/stable Unemployed Inactive Two years later (2000) Non-permanent/unstable Permanent/stable Unemployed Inactive Three years later (2001) Non-permanent/unstable Permanent/stable Unemployed Inactive

DK

NL

B

FR

IR

IT

GR

SP

PT

54.7 26.3 (6.4) (12.7)

40.7 46.2 (4.3) 8.8

49.8 43.3 (4.7) 2.2

52.2 22.3 14.1 11.4

48.7 33.4 (6.6) 11.4

49.6 28.5 11.7 10.3

48.4 27.7 9.6 14.3

50.9 30.9 10.5 7.7

63.1 25.6 5.3 6.0

36.7 41.3 (6.7) 15.2

27.3 61.0 (1.8) 9.9

37.7 54.0 (3.2) (5.2)

37.4 39.3 12.2 11.1

36.3 47.2 (4.9) 11.6

42.2 35.7 9.9 12.2

36.4 40.0 9.4 14.2

42.6 38.2 10.2 9.1

21.9 51.9 (4.6) 21.6

19.5 67.2 (4.8) 8.6

29.8 61.4 (3.1) (5.7)

32.2 44.9 13.1 9.8

30.0 50.7 (7.3) 12.0

35.0 43.0 8.9 13.2

40.8 39.1 5.0 15.2

36.8 45.6 10.1 7.6

AU

FI

G

LU

UK

53.6 34.2 (3.8) 8.4

45.9 29.2 12.5 12.4

50.8 30.1 (7.4) 11.6

42.3 48.0 (5.3) (4.4)

32.3 45.6 (2.3) 19.8

43.5 41.0 7.6 7.9

32.8 53.9 (4.6) 8.7

36.8 39.2 12.1 12.0

41.7 40.1 7.6 10.6

27.4 64.4 (2.1) (6.0)

22.1 52.1 (4.4) 21.4

35.6 51.1 4.7 8.7

23.0 62.1 (5.5) 9.4

30.9 45.4 11.2 12.6

28.9 49.6 9.4 12.2

20.0 68.4 (1.9) (9.7)

22.1 52.8 (3.3) 21.8

Notes: DK = Denmark, NL = the Netherlands, B = Belgium, FR = France, IR = Ireland, IT = Italy, GR = Greece, SP = Spain, PT = Portugal, AU = Austria, FI = Finland, G = Germany, LU = Luxembourg, UK = United Kingdom. The figures in brackets may be not significant. * Panel data for Sweden are not available. Source: ECHP, authors’ own calculations.

Unstable Employment in Western Europe

105

2 years ranged from 30 to 40 per cent. The share of unstable workers who enter unemployment, although diminishing, still stands at significant levels in these countries. In the medium term (after 3 years), the situation of workers unstable in 1998 confirms the greater inclusive capacity of the Dutch, Belgian and Austrian labour markets, where almost two out of three unstable workers enter stable employment. Denmark, the United Kingdom, Ireland, Portugal and Germany are countries in which one out of two unstable workers enter stable employment over the 3-year period, while France, Italy, Greece, Spain and Finland show lower chances for transition into stable employment.18 As for gender differences, women are more disadvantaged then men in the transition to stable employment in all countries: the proportion of women unstable in 1998 who had become stable workers in 1999 is lower everywhere than the overall value. Vice versa, the share of women who remain in unstable employment and the share of those entering unemployment or inactivity are everywhere larger over the 3 years considered. As regards age, the situation differs across countries. While in some countries (Denmark, Belgium, Spain, Finland and Germany) the share of under-35s in unstable employment who remain in the same condition is higher than for overall employment, some other countries (the United Kingdom, the Netherlands) show more highly favourable chances of entering stable employment for workers aged under 35. Over the 2 and 3 year periods considered, however, unstable workers aged under 35 everywhere show more favourable rates of transition to stable employment when compared to the employed overall.

Distinctive patterns of labour-market inclusion and exclusion Different degrees of labour-market inclusiveness can be detected across Western European countries when consideration is made of the distinctive labour-market histories into or out of employment experienced by unstable workers (Table 5.6). Patterns of full inclusion, such as transitions into stable employment (unstable workers entering permanent employment),19 are especially evident in the Dutch labour market (where the share of unstable workers is quite significant) and in the Austrian and Belgian labour markets (where unstable work is less widespread). In these countries more than 45 per cent of unstable workers enter and remain in stable employment; they are followed with slightly lower values by the British, Danish, Finnish, German and French labour markets (above one third of the sample). In these countries, the risks connected with instability are reduced by the greater chance of leaving unstable work. Unstable employment is different in the Mediterranean countries, where patterns of full inclusion in stable employment are less likely to occur (the values are below 30 per cent), despite the significant share of unstable workers in overall employment.

Table 5.6

Different inclusion/exclusion patterns of unstable workers over the period 1998–2001 (%)

1998–2001* Included (enter stable employment) Settled (always unstable, never unemployed/inactive) Excluded (alternate instability and unemployment) Relapsed (achieve stability and then exit it) Expelled (become chronically unemployed or inactive) At least one unemployment spell Women Included (enter stable employment) Settled (always unstable, never unemployed/inactive) Excluded (alternate instability with unemployment) Under–35s Included (enter stable employment) Settled (always unstable, never unemployed/inactive) Excluded (alternate instability and unemployment)

DK

NL

B

FR

34.5 12.2

53.7 8.8

49.3 19.1

33.1 20.1

25.4

13.2

12.1

(6.6) (11.3)

AU

FI

38.9 28.5 28.7 30.1 30.5 16.0 21.4 22.1 17.4 19.2

45.5 15.9

26.2

19.5 25.4 22.4 24.8 16.2

12.4

10.4 (4.4)

11.5 6.7 (3.2) 12.5

14.2

8.9

(7.1) 22.9

33.7 (11.0)

44.9 (11.2)

43.4 19.4

28.9

19.3

17.6

40.2 (9.4)

62.3 55.6 39.1 (5.8) (16.3) 15.7

22.9

9.7

7.4

IR

IT

GR

SP

PT

G

LU

UK

33.9 17.3

33.5 18.3

59.1 13.2

41.6 9.8

29.0

21.2

12.4

21.5

10.4 10.2 13.7 14.1 10.7 (6.1) 12.2 11.1 6.5 6.0

7.5 6.5 (5.6) 13.1

8.7 5.8

13.3 19.7

11.2

21.4

16.8

31.3 19.0

41.4 23.1 20.0 30.9 31.4 40.5 31.2 17.2 27.2 24.3 14.2 22.7 (19.8) 19.7

38.6 17.4

52.1 36.5 (13.0) (11.2)

29.7

21.6 29.0

19.1

20.9

41.6 15.9

31.9 19.5

57.2 48.5 (11.4) (2.6)

17.4 24.3 18.4 24.5 12.2 (10.0) 25.9

22.8

(10.1)

20.1

16.0 22.9 13.4

(3.0) 10.4 (3.1) 10.8

31.9 30.1 19.7 (16.7) 34.8

48.8 32.2 34.2 31.2 36.2 (6.6) 16.3 20.7 16.9 16.4

42.9 16.9

(8.8)

8.6

22.4

21.2

Notes: DK = Denmark, NL = the Netherlands, B = Belgium, FR = France, IR = Ireland, IT = Italy, GR = Greece, SP = Spain, PT = Portugal, AU = Austria, FI = Finland, G = Germany, LU = Luxembourg, UK = United Kingdom. Sum in column might not equal 100 because the patterns identified are not exclusive. The figures in brackets may not be significant. * Panel data for Sweden are not available. Source: ECHP, authors’ own calculations.

Unstable Employment in Western Europe

107

Unstable workers might, however, experience continuity in employment even though they do not enter stable employment. This is the case of unstable workers who, although they never achieve stable employment, are continuously unstable20 as they do not experience unemployment or inactivity episodes. Italy, Greece and France, with values above 20 per cent, are the countries with the highest occurrence of this situation, while the Netherlands and the United Kingdom record the lowest proportions of unstable workers who are continuously unstable (below 10 per cent) and this is associated with the lowest proportions of unstable workers experiencing unemployment at least once over the 3-year period (below 9 per cent compared to above 20 per cent in France, Italy and Spain). A third pattern is drawn by transitions from unstable employment into unemployment and again into unstable employment. The Netherlands, Belgium and Austria are the countries where this recurrent pattern of exclusion from stable labour market is less common.21 The other countries, both those with more inclusive (Denmark, Ireland, France, Germany) and those with more segmented (Italy, Spain, Greece) labour markets, show similar values (around 20–25 per cent). This may suggest that instability generates a very similar share of particularly vulnerable workers who experience very discontinuous and fragmented labour-market careers and this occurs regardless of the incidence of unstable work and of the inclusiveness of national labour markets. Other discontinuous labour-market histories are distinctive to unstable workers who entered stable employment but then returned to unstable employment (entering standard employment does not mean employment stabilization). This situation varies considerably across countries: Greece and Spain show the largest share (around 14 per cent), followed by the Netherlands, Ireland, Italy, Portugal and the United Kingdom, with values of around 10 per cent. Workers who, after a spell of unstable work, irreversibly enter unemployment or inactivity are minor yet important cases because they represent patterns of actual exclusion from employment and are likely to be connected with situations of particular risk. The largest proportions of unstable workers experiencing irreversible exclusion from employment over the 3-year period are to be found in Italy, Greece, France, Finland and the United Kingdom, with values above 10 per cent. At the opposite extreme, Belgium and the Netherlands record values slightly below 5 per cent. Thus the labour-market histories of unstable workers give clearer coordinates to the geography of flexibility in Western European countries. Indeed, on the one hand, Southern European countries increase the vulnerability of unstable workers by forcing them to a larger extent into unstable employment or discontinuous labour markets careers where spells of instability, unemployment and inactivity follow each other. On the other hand, the countries where unstable employment is less common (Belgium and Austria)

108 Social Vulnerability in Europe

reduce the risk of unstable workers being trapped in a ‘parallel’ labour market through the greater rapidity and fluidity of labour-market transitions. Denmark and the Netherlands also exhibit this pattern, but within the framework of a more significant incidence of unstable work and of explicit labour policies promoting the rapidity and fluidity of labour-market transitions. Conversely, the United Kingdom, which bears some similarities to the Netherlands, is characterized by the greater occurrence of labour-market histories of alternating instability and unemployment, and transitions leading to unemployment and inactivity. The situation in the countries of continental Europe such as France and Germany is more complex. In these cases, different degrees of segregation in unstable employment mingle with different chances of discontinuous careers consisting of unstable work, unemployment and inactivity, with France coming closer to the pattern of the Southern European countries owing to the greater frequency of continuous instability patterns and of labour-market exclusion histories.

The economic vulnerability of unstable workers Differences between stable and unstable workers in the availability of economic resources, especially those deriving from work, represent an important indicator of vulnerability. The ECHP data22 show that in all Western European countries the median annual income from work of unstable workers ranges from 60 to 70 per cent of the median annual income from work of overall workers, and that it is even lower in Ireland, Greece and Italy (around 50 per cent) where unpaid family work is more frequent among unstable workers. When only adult workers (aged 25–49) are considered, all countries still show a disadvantage for unstable workers, but the differences are narrower and the ratio between the median work income of unstable workers and the median work income of overall workers is more favourable, ranging from 65 to 75 per cent. Yet when the first quartile of work income is considered, the ratio decreases considerably, especially in the Mediterranean and the Nordic countries. Data on the current monthly wage23 show that the ratio between the median income of unstable workers aged between 25 and 49 and the income of overall workers is slightly more favourable to unstable workers in all countries except the United Kingdom and Ireland. This confirms that the current wages of unstable workers are on average lower than those of the stable employed, but also shows, given that the differences increase when yearly earnings are considered, that discontinuity in employment may play a major role in reducing the annual income.24 It is more useful for our purposes here to analyse differences between countries in the formation of economic resources by unstable workers (see Figure 5.2). According to ECHP data, around two thirds of unstable workers in 2001 declared that their main source of personal income was earnings

Unstable Employment in Western Europe

109

100

80

No Income from any socurce Work Any form of social transfer Other

60 (%) 40

20

0 DK NL BE FR IE IT GR SP PT AU FI SW G LU UK

Figure 5.2 Main income source of unstable workers, by country – 2001 Source: ECHP, authors’ own calculations.

from work, compared to 95 per cent of stable workers. In the Mediterranean countries, with the exception of Spain, up to 15 per cent of unstable workers reported that they had no source of personal income, while in Denmark, Finland, Ireland and the United Kingdom unstable workers more frequently declared that their main source of personal income was a welfare payment (values between 18 and 25 per cent). In general, therefore, these data suggest that the income of unstable worker is supplemented by social welfare programmes differently across Western Europe.

The household dimension of job instability The analysis of individual profiles has shown that employment instability in Europe is an important phenomenon, but that it is difficult to draw a clear map at continental level because it is strictly correlated with the specific economic policies and labour-market arrangements of countries. Despite the difficulty of depicting the spread of the phenomenon, it is here hypothesized that unstable employment is associated with forms of social vulnerability. The varying degrees of protection provided by local labour policies, the differing structures of national economies and the resulting differences in occupational trajectories are among the main determinants of these situations. In this regard, the difference between unstable workers who experience vulnerability and unstable workers who are not vulnerable is closely connected with the type of household to which they belong. Having a partner in stable employment, rather than one in casual work, or having one or

110

Social Vulnerability in Europe

more adult children unemployed or in unstable employment generates different dynamics and impacts. We shall analyse these aspects by considering households according to the employment situations of their members, in order to show the distribution of different work/family patterns. The situations of households will be analysed by means of specific indicators in order to determine the effects of employment instability. The work-based structure of households has changed profoundly in the past decade. The increased labour-market participation of women is undoubtedly one of the factors responsible for these changes (Scherer and Reyneri 2008; Employment Outlook, 2008). Further changes have taken place in the demographic structure, such as the growth of the elderly population, delayed transition to adulthood – which particularly concerns the countries of Southern Europe (Heath and Miret, 1996; Cherlin et al., 1997; Holdsworth, 2000; Rosina, Mencarini and Rettaroli, 2005) – and an increasing number of problematic households (such as lone-parent families). Aggregate analysis of the relationship between work and household structure is therefore necessary to understand the outcomes of the most recent changes.

Work/family models Analysis of the data shows that the pattern of one employed person per household (excluding households consisting of single persons), although still present in all countries and thus confirming the persistence of this traditional model, reaches percentages above the European average of 40 per cent in several countries. This is especially the case of Southern Europe (Italy, Greece and Spain), where the male breadwinner (MBW) model is still deeply rooted in Mediterranean culture (Reher, 1998; Micheli, 2000; Villa, 2004; Moreno and Crespo, 2005), but also of Ireland, a country with a strong Catholic culture; of Germany, whose industrial tradition has been historically characterized by low female employment; of the United Kingdom and the Netherlands. In this regard, analysis of the relationship between household size and the number of its members in employment yields some interesting findings. If we consider a ratio between these two numbers which is at least equal to two as favourable (so that each employed household member has no more than one unemployed or inactive member as a ‘dependant’), we can identify four different patterns across Europe (see Table 5.7). The first is represented by countries with high numbers of households with an unfavourable ratio (greater than two), such as Spain, Ireland, Italy and Greece. A second group, with a less unfavourable ratio, comprises Luxembourg, Belgium, Germany, France and Portugal. Moreover we have two groups, the first consisting of the Netherlands, Austria, the United Kingdom and Finland and the second consisting of Denmark and Sweden, where very few households are in

Unstable Employment in Western Europe

111

Table 5.7 Household members/employed members ratio and relation to total employment rates, female activity rates, and household size. Year 2001 Family size/members employed ≥2

Household size Employment rate

Female activity rate

Spain Ireland Italy Greece Luxembourg Belgium France Germany Portugal

41.3 40.1 39.5 38.3 29.8 28.4 27.6 27.5 27.2

Italy Greece Spain Belgium France Luxembourg Germany Ireland Finland

54.8 56.3 57.8 59.9 62.8 63.1 65.8 65.8 68.1

Netherlands Austria

24.5 24.1

Austria Portugal

United Kingdom Finland Denmark Sweden

23.2

United Kingdom Sweden Netherlands Denmark

68.5 Portugal 69.0 United Kingdom 71.4 Netherlands

22.4 16.5 13.6

Italy Greece Spain Luxembourg Belgium Ireland France Germany Austria

74.0 Finland 74.1 Denmark 76.2 Sweden

41.1 41.5 43.1 50.9 51.0 54.9 56.0 58.7 60.7

Portugal Ireland Spain Austria Italy Germany Greece France United Kingdom 61.3 Finland 65.0 Belgium

3 ⫹ members 24.4 21.7 19.9 19.8 15.1 13.5 13.3 11.7 10.2 9.0 8.8

65.2 Sweden

8.5

65.4 Denmark 72.0 Luxembourg 72.3 Netherlands

8.2 6.5 5.4

Source: ECHP, authors’ own calculations.

situations of heavy overload. These data are closely correlated with total employment rates and female activity rates, but less so with household size (see Table 5.7). The correspondence between the indices concerns in particular those countries – Spain, Greece and Italy – where the total household members/employed members ratio is more unfavourable. Ireland differs slightly from this group as it shows one of the highest average numbers of components per household but simultaneously with more substantial employment rates than those of the countries of Southern Europe. It should be pointed out, however, that the share of households with unfavourable ratios between their employed and unemployed members has diminished in many European countries in recent years. There has been a particular improvement in this indicator for Spain, because of its favourable economic growth over the past decade, and for the Netherlands, whose growth can be explained by the massive use of part-time work (which rose by more than 10 percentage points between the 1990s and the 2000s) (Eurostat, 2003b). But how are households distributed according to the different employment situations of their members? To answer this question we broke down households on the basis of the employment situations of their individual components and then recombined the possible work situations to be found within a household – the states of stability, instability and unemployment – as

112

Social Vulnerability in Europe

Stable employed

Unstable employed

Unemployed

Figure 5.3 Work family models

shown in Figure 5.3. The different compositions of these three states define what we call ‘work/family models’. A first reaggregation of the different combinations produced a typology consisting of three types of households: households with only stable workers, households with only unstable and/or unemployed workers and mixed households with stable and unstable and/or unemployed workers. To gain a better understanding of the specific features of the different household models, we distinguished these three types by taking account of the number of members in employment. This process allowed the definition of a more detailed work/family typology with which to capture differences and similarities among macro-regions. In detail (see Table 5.8), we split out households consisting of only stable workers, households comprising both stable workers and unstable and/or unemployed workers, households with only unstable workers (alone or with unemployed members) and households composed of only unemployed workers. Stable households consisting of single persons are considered in isolation, given their role in the total of European households. As already mentioned in regard to the analysis of the ratio between employed and inactive household members, work/family models are strongly influenced by the number of household members. While households with only stable workers and those with only unstable workers are characterized by a small number of components, this number increases in the case of mixed households, where couples with children aged over 18 represent the dominant model. Stable families (all the active members in the household are permanent workers) represent the majority of cases, accounting for 68.4 per cent of households. Within this large aggregate we identify one-member households (13.6 per cent), dual- earner households (31.8 per cent) and singleearner households (22.8 per cent). This last group is represented by families

Unstable Employment in Western Europe

113

with a traditional gendered division of labour. Mixed families are divided into two subgroups: households with two permanent workers and one temporary worker represent 5.6 per cent, while households combining one stable worker with an unstable one make up 13.2 per cent. Finally, unstable families are made up of households with only temporary workers (7.6 per cent) and households whose active members are all unemployed (4.4 per cent) or partly temporary workers (1.2 per cent). Households touched by temporary work or unemployment are therefore estimated in this analysis as representing approximately one third of all European families. More than half of these situations (19 per cent) are characterized by a mixed model, where unstable workers and stable workers are combined together. Only 12 per cent of households, therefore, are characterized by the presence of principal earners who are in a temporary position or unemployed. The distribution of these models through Europe is very unequal. Households including temporary workers, either in a mixed model or in an unstable model, are dominant in Spain, South Italy, Portugal and East Germany, where they account for more than 40 per cent of cases. In North-Western Italy, Central and North-Eastern Italy, the United Kingdom and in Denmark these households make up less than 26 per cent of the whole. Therefore a strong concentration of unstable and mixed families occurs in some areas of Western Europe, cutting across traditional geographical distinctions. However, even in areas most affected by temporary work there is a high presence of mixed models, where instability and stability in the labour market are joint. If we analyse the structure of the different work/family models by considering the number of household members in employment, some interesting features emerge. In Denmark, Austria, Belgium and Portugal, for example, there is a marked prevalence of stable couples, while Scandinavia and Germany have a large presence of stable singles. The MBW model is still predominant in Italy, Greece and Ireland, as opposed to a low proportion of employed singles. Overall, it is possible to identify three groups of countries: ●

●

●

the first comprises countries and macro-regions with large numbers of households with two stable workers as a result of high female employment, associated with a marked presence of dual-worker households and a pronounced individualization of household models (Denmark, Belgium, North Italy, the Île- de-France, Portugal and Sweden) shown by a large number of singles; the second group comprises countries and macro-regions (Southern Italy, Spain, Greece and Ireland) with a high presence of MBW households with one stable worker and a small number of singles; the third group comprises countries with high levels of both MBW and in single-person households: a mixed combination that is peculiar to Germany.

Table 5.8

Share of work/family models by macro-region 2001 Stable Family

Mixed Family

1 Stable ⴙ 2 stables 1 stable ⴙ 1 Stable inactive ⴙ inactive unstable/ single members members unemployed Austria Belgium Germany North Germany Central-East Germany West Germany South Denmark Spain Centre and North-West Spain East and Madrid Spagna South Finland France Centre-East France South

2 stable ⴙ unstable/ unemployed

Unstable Family Only Only Unstables ⴙ unstable unemployed unemployed Total

18.3 11.0 22.3 17.5

21.4 21.3 25.6 12.0

37.7 38.4 19.5 25.2

8.0 12.0 14.7 15.5

6.6 3.7 3.4 6.2

4.4 5.4 8.2 8.1

3.1 7.7 5.0 12.4

0.5 0.5 1.3 3.1

100 100 100 100

23.0 18.4 10.5 4.1

29.3 25.5 15.1 26.1

18.5 20.8 47.8 20.8

11.3 13.7 10.9 17.2

4.8 6.8 4.6 8.3

8.9 10.5 7.6 13.0

3.7 3.7 3.1 7.1

0.5 0.6 0.4 3.6

100 100 100 100

8.4

25.4

24.3

17.4

6.8

13.0

3.5

1.2

100

2.2 21.3 11.9 12.4

19.9 16.9 23.1 24.2

13.5 32.4 32.3 28.8

21.9 12.0 15.6 15.5

7.7 2.4 5.0 4.0

17.5 7.6 5.4 6.5

10.4 6.3 5.2 6.7

6.9 1.1 1.5 1.9

100 100 100 100

France West and South-west France Île- de-France Greece Ireland Italy North-west Italy South Italy Centre and North- east Luxembourg The Netherlands Portugal Sweden Uk North Uk Centre Uk Greater London Uk West and South-West EU-15

13.2

22.5

29.5

16.9

6.5

4.2

6.2

1.0

100

15.9 5.5 9.1 10.4 4.1 8.9

19.6 30.5 27.6 29.0 31.5 28.5

37.1 29.1 26.2 40.7 17.6 36.8

13.4 16.1 13.3 8.8 17.8 11.6

5.1 4.9 8.8 4.7 6.2 6.2

2.8 9.3 7.6 4.6 8.1 4.2

5.2 3.6 6.1 1.5 10.2 2.9

0.9 1.0 1.3 0.3 4.5 0.9

100 100 100 100 100 100

17.7 17.9 3.0 29.1 12.5 12.2 18.4

34.7 25.4 17.8 6.2 26.0 27.1 29.4

30.6 28.3 39.3 38.1 37.1 35.7 32.1

10.5 11.8 17.7 10.0 8.9 9.0 7.3

2.3 2.2 11.6 6.7 2.4 1.6 3.3

2.8 10.0 7.0 5.4 9.4 10.1 7.5

1.2 3.7 2.1 4.2 2.6 3.8 1.5

0.2 0.7 1.5 0.3 1.1 0.5 0.5

100 100 100 100 100 100 100

13.8

28.8

34.8

7.3

2.5

10.4

1.9

0.5

100

13.6

22.7

31.7

13.2

5.6

7.6

4.4

1.2

100

Source: ECHP, author’s own calculations.

116

Social Vulnerability in Europe

Furthermore, analysis of macro-regional details reveals some interesting internal differences within countries. Specifically, in some countries the weight of instability is often determined by a strong internal macro-regional disparity, as shown by Table 5.8. In the cases of Spain and Italy, for example, the disparity is to the detriment of the Southern regions, while for Germany it weighs most heavily on the Länder of former East Germany and for the United Kingdom on the highly industrialized areas of the centre-north. Analysis of the trend between the mid-1990s and early 2000s shows a growth of unstable work and a simultaneous reduction in unemployment. In fact, considering the households belonging to the area of instability, those with only unemployed members fell by 8.7 per cent and those with both unemployed and unstable members decreased by 2.2 per cent, compared with a 10.9 per cent increase in households with only unstable members. This phenomenon is also apparent in the case of mixed households: those with both stable and unemployed members diminished by 11 per cent. In contrast, during the same period, mixed households of stable and unstable workers increased by 10.3 per cent.

The impact of job instability on households This analysis so far has concentrated on the relationship between household structure and the work/family pattern, highlighting the households structurally most at risk of instability and unemployment. To conclude the analysis, we now focus on two household features largely influenced by the work-based structure of the household: exposure to poverty and choices on whether or not to have children. The aim is to show the relationship between these dimensions and the employment conditions of households, on the hypothesis that households with unstable or unemployed members are the most vulnerable. It is obvious that job stability ensures a steady income. When considering the frequency of job instability, therefore, it is important to determine its impact on the economic circumstances of the households affected by this phenomenon. Moreover, several studies have shown that adequate assessment of the economic impact of work conditions requires the use of household income as the benchmark (Lucifora, 2000). Using the data on national poverty lines (calculated as 50 per cent of the median national income) as a proxy for possible exposure to income-related problems, and subsequently calculating the weighted deviations for each work/family model, it emerges that, in all countries, households with only unemployed members are more exposed to income poverty than are stable households, mixed households and households with only unstable workers (see Table 5.9). It is worth mentioning the fact that unstable workers living in a mixed households do not show a higher poverty risk in respect to stable workers, while temporary workers in unstable families do.

Table 5.9

Work/family models below the national poverty line. Weighted deviations. Year 2001 Stable Family

1 Stable single Austria Belgium Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Sweden United Kingdom

0.1 0.2 1.3 0.2 ⫺0.4 ⫺0.7 ⫺0.3 ⫺0.5 ⫺0.6 ⫺0.3 ⫺0.2 0.2 ⫺0.7 ⫺0.1 0.0

1 Stable ⴙ inactive members 0.6 ⫺0.2 ⫺0.6 ⫺0.6 0.2 0.2 ⫺0.1 ⫺0.2 ⫺0.1 0.3 0.0 0.2 ⫺0.1 ⫺0.4 0.0

Source: ECHP, authors’ own calculations.

Mixed Family

2 stables ⴙ inactive members ⫺0.6 ⫺0.8 ⫺0.9 ⫺0.5 ⫺0.8 ⫺0.8 ⫺0.5 ⫺0.9 ⫺0.8 ⫺0.8 ⫺0.8 ⫺0.4 ⫺0.7 ⫺0.8 ⫺0.8

1 stableⴙ unstable/ unemployed 0.1 ⫺0.2 ⫺0.7 ⫺0.8 ⫺0.1 ⫺0.6 0.1 ⫺0.5 0.3 0.9 ⫺0.1 0.2 ⫺0.5 0.0 ⫺0.1

2 stable + unstable/ unemployed 0.1 ⫺0.6 ⫺1.0 ⫺0.6 ⫺0.8 ⫺1.0 0.1 ⫺0.7 ⫺0.5 ⫺0.5 ⫺0.3 ⫺0.3 ⫺0.3 ⫺0.9 ⫺0.5

Unstable Family

Only unstable 0.0 0.8 4.5 2.6 2.6 1.7 0.7 2.3 1.0 1.6 1.6 1.3 0.9 3.3 2.1

Only unstables + unemployed unemployed 1.1 4.5 5.6 2.3 3.1 4.0 2.7 5.0 3.8 6.7 2.6 0.9 2.5 3.3 2.9

8.0 ⫺1.0 ⫺1.0 0.8 1.2 1.4 2.9 0.9 3.4 ⫺1.0 3.4 1.0 0.9 4.2 6.3

118

Social Vulnerability in Europe

In this case, too, it is difficult to identify a precise geography of the phenomenon, which affects not only unemployed households but also households with unstable members to a significant extent. As is to be expected, households with unemployed members are heavily penalized with respect to those with stable members. To a lesser extent, unstable worker households are also exposed to income vulnerability, and in some cases this is quite severe. Data show that the economic impact of temporary employment can be very different according to the country. Instability seems to be more disadvantageous in countries of Northern Europe (Denmark, Finland, Sweden, the United Kingdom and Ireland) and in France. In continental and Mediterranean countries poverty mostly affects households with unemployed members. While poverty is one of the possible effects on households of employment instability, a further and no less important impact concerns the birth choices of couples. One of the reasons why young couples decide to postpone having babies is the job insecurity of one of the household members. If we analyse the distribution of work/family patterns only for households in which the woman has a maximum age of 35, we find that the presence of children diminishes markedly from households with only stable members to families with also (or only) unstable members (Figure 5.4). Mixed households have a proportion of women aged under 35 with children that is less than a half of households with two permanent workers. Just 5 per cent of households with only unstable workers have children aged less than

0

5

10

15

20

25

30

35

40

45

50

1 stable 2 + stable 1 stable + unstable/ unemployed 2 stable + unstable/ unemployed Only unstable Only unemployed Unstable + unemployed

Figure 5.4 Share of households with children aged less than 12 years by family/work patterns (woman’s age max. 35) Source: ECHP, authors’ own calculations.

Unstable Employment in Western Europe

119

12 years: a very high gap in respect of the 47 per cent of two-stable-worker households with children. Mixed families show an intermediate propensity for having babies, confirming their better condition in respect of households with only temporary workers. In this case, too, there are significant geographical differences (see Table 5.10). In particular, in some countries with strong familist traditions (Italy, Greece and Ireland), as well as in Germany, the presence of children is relatively concentrated in households with only one stable worker: a model very close to the traditional MBW model. In these countries the presence of two earners is not a strong incentive to fertility, as happens in many other continental and Nordic countries (the Netherlands, Denmark, Belgium, Sweden and Austria). In France the propensity for having children in mixed households is relatively high. Instability and unemployment are associated with higher fertility only in Spain, Portugal, Greece and Finland.

Conclusions In recent decades, forms of work different from the working-life-long permanent contracts typical of Fordist society have arisen in all Western European countries. Labour flexibility has invaded various spheres: working hours, job tasks and above all contractual arrangements. This chapter has analysed the importance and features of unstable employment considering all the forms of employment that do not give workers guarantees on the continuity of their employment. The analysis has been carried out considering not only workers on fixed-term contracts, as the available statistics usually do when dealing with labour flexibility, but also exploring both individual and the household dimensions in order to highlight whether employment instability generates higher levels of vulnerability for individuals and households in Western European countries. Although the incidence of unstable employment in total employment does not chart a clear European geography of the phenomenon, it does partly change the ranking of countries where flexibility is most frequent when only the fixed-term employment rate is considered. This analysis has shown that, when measuring unstable workers by means of the instability index, in the Mediterranean countries (with the exception of Spain, where fixed-term employment is exceptionally common) and the Anglo- Saxon countries unstable employment is important and even crucial. Age and gender have been, and still are, the distinctive features of unstable employment in all European countries. However, there is no evidence of a further concentration of the phenomenon among younger workers and women in most recent years. Rather, the incidence of unstable employment has increased among older workers as well. Only the Mediterranean countries have recently registered a trend of an even stronger concentration of the phenomenon among young people aged 24 and under.

Table 5.10 Share of households with children aged over 12 years by family/work patterns (with the woman’s age max. 35) National differences from the European average. Year 2001

1 stable Denmark Netherlands Belgium France Ireland Italy Greece Spain Portugal Austria Finland Sweden Germany Luxembourg United Kingdom EU 15 average

2 stable or more

⫺4.6 0.4 ⫺13.6 ⫺2.1 13.8 11.0 13.7 ⫺0.4 ⫺15.8 ⫺0.2 ⫺0.7 ⫺10.6 6.2 2.5 6.0 24.2

Source: ECHP, authors’ own calculations.

11.7 4.5 16.7 ⫺1.3 ⫺10.3 ⫺11.0 ⫺13.4 ⫺16.1 6.0 3.9 ⫺5.1 9.4 ⫺7.8 13.2 6.1 47.4

1 stable ⴙ unstable/ unemployed ⫺4.4 ⫺0.5 ⫺1.5 3.8 ⫺1.9 ⫺1.7 0.7 3.4 9.4 ⫺4.0 4.9 0.2 ⫺0.8 ⫺7.2 ⫺9.5 18.3

2 stable ⴙ unstable/ unemployed ⫺0.9 ⫺0.9 ⫺0.9 ⫺0.9 ⫺0.9 0.2 ⫺0.7 ⫺0.4 2.0 1.6 ⫺0.9 4.3 ⫺0.7 ⫺0.2 ⫺0.4 0.9

Only unstable ⫺1.7 ⫺1.6 ⫺2.9 ⫺0.6 ⫺2.2 0.9 2.1 8.8 ⫺0.7 0.5 2.4 ⫺4.7 1.9 ⫺4.8 ⫺1.8 5.5

Only unemployed 0.9 ⫺1.1 1.9 0.5 0.7 0.8 ⫺1.4 1.9 ⫺2.2 ⫺1.1 0.0 1.9 1.3 ⫺2.0 0.0 2.2

Unstable ⴙ unemployed ⫺1.0 ⫺1.0 0.3 0.5 0.9 ⫺0.1 ⫺1.1 2.8 1.3 ⫺0.7 ⫺0.6 ⫺0.4 ⫺0.1 ⫺1.5 ⫺0.5 1.5

Unstable Employment in Western Europe

121

Clearer coordinates for a geography of unstable employment in the European countries can be inferred from the labour-market histories of unstable workers over time. These show a greater vulnerability of unstable workers in the Southern European countries caused by greater difficulty in entering stable employment, longer time spent in unstable work and the greater chances of chronic unemployment and/or inactivity. Some countries in Northern and Continental Europe (the Netherlands, Denmark, the United Kingdom, Austria and Belgium) make unstable workers less vulnerable by enabling more rapid transitions into stable employment and by reducing the number of situations where workers are continuously unstable or experience chronic unemployment and inactivity. There are some national differences, however. On the one hand, the greater inclusiveness of Austria and Belgium is connected with the relatively low amount of unstable work in those countries. On the other hand, greater labour-market inclusiveness in the Netherlands and Denmark is connected with more rapid and more fluid transitions likely supported by the explicit flexicurity systems adopted in those countries. The United Kingdom is somewhat similar to the Netherlands, but it has higher risks of exclusion for unstable workers, given the greater preponderance of labourmarket histories that alternate unstable employment and unemployment, and of labour-market histories that lead to unemployment and inactivity. The situation in continental Europe is more diversified, owing to different combinations of the risk of segregation into unstable employment and the risk of discontinuous labour-market histories alternating unstable work, unemployment and inactivity. In this framework, France more closely resembles the pattern of the Southern European countries. As well as the presence of a proportion of unstable workers, ranging from 20 to 25 per cent, who experience discontinuous labour-market careers alternating employment instability, unemployment and inactivity, a feature shared by all countries is the systematic female disadvantage in access to stable employment. As for economic resources, the generalized disadvantage of unstable workers highlights that they are particularly vulnerable in the countries of Southern Europe. The median work income of unstable workers varies between 60 and 70 per cent of that of workers overall, with an even more unfavourable ratio in the Mediterranean countries, where, moreover, income support measures for unstable workers are much less widespread. Shifting from the individual to the familial dimension is crucial for grasping the real vulnerabilizing potential of work (OECD, 2008a). In this regard the growth of female activity rates – growth which has been one of the main junctures in transformation of European labour markets – has been decisive. Our analysis has corroborated the findings of the analysis on individual profiles by showing that – amid the diversified frequency of work

122 Social Vulnerability in Europe

instability – it is possible to identify specific work/family patterns. Two findings are relevant in this respect. First, unstable workers are split into two very different work/family models: a mixed model, in which they are joined by stable workers, and an unstable model, in which they often share a critical situation with an unemployed person. Second, this distribution in work/family models cuts across the traditional distinction in welfare regimes, showing a high concentration of the unstable model in specific areas of Mediterranean Europe. In some of these areas, however, there is also the highest concentration of stable families. Italy emerges as the country most polarized between these two opposite models. The analysis of the exposure to economic vulnerability of the various work/family models has highlighted that families consisting solely of unemployed workers are the most hit by poverty, while unstable families have less probability of suffering economic stress. Unstable families are, however, strongly affected in their fertility strategies. The low propensity for having babies is peculiar not only to unstable households, but also of the mixed model, signalling that temporary employment discourages fertility even in situations where its financial impact is not relevant and is counterbalanced by cohabiting permanent workers. Finally, the propensity for having babies is higher in dual- earner stable families than in single-worker stable families. This is a further confirmation that, if the dynamics of the labour market influence family projects, family organization is also a necessary aspect to take into account in order to gain an understanding the social impact of temporary employment.

Notes The first, third, fourth, fifth, sixth, seventh and eighth sections were written by Ivana Fellini; the second, ninth, tenth and eleventh by Mauro Migliavacca. The final section was written jointly. 1. In addition, labour policies for specific targets, such as women and young people, affect the extent to which flexible forms of employment are used: consider, for example, training/work contracts and the public resources allocated to them (Chung, 2005). 2. Important examples are: (i) the Transitional Labour Markets approach (Schmid and Gazier, 2002) inspired by the Danish model of flexicurity, which combines low employment protection, generous support for the unemployed and large investments in active labour market policies. It focuses on (individual) transitions in the labour market rather than on job security. (ii) Institutional analysis, which synthesizes national labour-market rigidities with appropriate indicators (for example the OECD employment protection index) by considering the regulatory framework of employment relations and its connections with employment features (overall employment levels, share of temporary work, turnover rates, employment tenure, labour-market segmentation and perceived job security) (EC, 2006).

Unstable Employment in Western Europe

123

3. The main form of non-traditional self-employment is ‘economically dependent self-employment’ – that is pertaining to persons who are formally self-employed (they usually have some sort of service contract with the employer) and depend on a single employer for their income (or large part of it) (EIRO, 2002), and who therefore have no guarantees on employment continuity. The European Commission began to deal with this issue in the early 2000s; the OECD (2000) found that economically dependent self-employment might hide forms of false self-employment. A comparative study by EIRO – European Industrial Relations Observatory (2002) tried to estimate the phenomenon for countries in which it was presumably most widespread: Italy, Denmark, the Netherlands and Austria, followed by Portugal and Greece. The study also reported a significant incidence of the phenomenon in Germany and the United Kingdom. Unfortunately, the ECHP data do not allow identification of this kind of self-employment. We shall anyway attempt to consider also some ‘unstable’ self-employed workers among unstable workers. 4. For preliminary discussion of these issues and a first version of the stability index see Migliavacca (2008). 5. The indicators used to construct the stability index are the following: statements by the interviewees concerning their main work activity (information that can be used to isolate respondents undertaking paid work of different kinds), the types of contract of employed respondents (which makes it possible to discriminate between open-ended contracts with guaranteed continuity of employment and fixed-term contracts, or similar, without job security) and the number of hours worked in a week, which makes it possible to specify stability in cases (self-employment for example) where contractual information cannot be used. This latter information is essential for countries like Italy, Greece and Spain, where self-employment has a high incidence in overall employment and where it is necessary to identify differences with respect to the concept of stability. 6. Other information that could have been useful for differentiating the self-employed, such as the presence or absence of employees, was not present in the ECHP database. 7. As discussed, the stability index is built on the self-defined activity status of interviewees. For some countries (the United Kingdom, the Netherlands, France and Luxembourg) no information on the work positions of the majority of people defining themselves as ‘working fewer than 15 hours’ is available. The ECHP database reclassifies most of them as economically inactive when using the ILO – International Labour Office classification of the activity status. Since self-definition is used to classify workers as stable or unstable, in Table 5.1 both values (accounting and not accounting for people working fewer than 15 hours and for whom no other information on the working condition is collected) are reported. 8. This is consistent with the importance of economically dependent self-employed workers in the Italian labour market (EIRO, 2002), although the instability index cannot take account of information on economic dependence (presence of employees or an exclusive relationship with only one employer). 9. Comparative studies on the relationship between flexible work and the overall level of employment in European countries (EC, 1999), albeit considering only temporary employees, have already shown that the share of flexible employment is not significantly correlated to higher levels of overall employment. This finding has recently been corroborated by a European Commission study (EC, 2006).

124 Social Vulnerability in Europe 10. The ECHP data consider three levels of educational qualification based on the ISCED-International Standard Classification of Education. The high level (ISCED 5–7) corresponds to tertiary qualifications, the medium level to secondary school qualifications (ISCED 3–4) and the low level to pre-secondary qualifications or no qualifications (ISCED 0–2). 11. The aggregated data from the European Labour Force Survey do not allow use of the instability index as with ECHP data. 12. The feminization rate of unstable employment is the percentage share of women in unstable employment, and it is the simplest measure of female concentration. 13. It should be borne in mind that the 5 years considered here were a particular period for European labour markets: after the employment crisis of 1994–1995, in 1998 employment resumed substantial growth, giving a major boost to stable dependent employment (OECD, 2000). 14. At EU-15 level, the female employment rate grew from 52 per cent in 2002 to 54 per cent in 2007. 15. The Eurostat data do not allow a break-down into more specific age brackets. 16. The analysis is based on a panel reconstructed for the years 1998–2001. It focuses on the labour market condition of workers unstable in 2001 during the previous three years and on those in unstable employment in 1998 1, 2 and 3 years later, with the purpose verifying the consistency of certain ‘trajectories’, such as the transition to stability, exclusion from employment and persistence in instability. In the following analysis we focus on the forward trajectories of unstable workers in 1998, rather than on backward trajectories of unstable workers in 2001. 17. Unstable and stable workers were defined using the instability index. In the panel analysis, workers who had exited the labour force because of age (retirement) over the period of time considered were excluded. 18. A similar exercise carried out by the OECD (2006a) on temporary dependent workers only is consistent with these results. It shows that more than half of temporary employees in 1998 were in permanent jobs a year later in Austria and the United Kingdom. After 3 years, Belgium and the Netherlands became the best performers, with 70 per cent of temporary employees in 1998 in a permanent position in 2001, against less than 40 per cent in Greece. Between one third and two thirds of temporary workers move into permanent jobs within two years. The data also show that temporary workers are much more likely to move to unemployment than the permanent employed. Moreover, while the majority of temporary workers who were not employed had moved into unemployment in most countries, in particular Finland, France, Germany, Greece and Spain, transitions from temporary employment to inactivity predominated in Denmark, Ireland and the United Kingdom (a trend that has become more pronounced over the years) (OECD, 2006a: 172–73). 19. Unstable workers entering stable employment are those who, unstable in 1998 were stable over the next 3 years, or if still unstable in 1999, were stable in 2000 and 2001. 20. Continuous unstable workers are those who qualified as unstable workers over the 3 years. 21. Unstable workers excluded from the labour market are those who have not experienced stable employment over the 3 years and combine spells of unstable work with spells of unemployment/inactivity.

Unstable Employment in Western Europe

125

22. For annual income, the analysis considers that the income declared in wave eight (2001) refers to 2000. Thus, the labour-market position in the previous year has been reconstructed for the panel, in order to match the employment condition with income for the same year. 23. In this case, the data on income refer to the time of the interview, not to the previous year. Consequently, all unstable workers in the 2001 wave and their monthly work wages have been considered. 24. This summary information on incomes should of course be treated with caution, given that no information about the discontinuity of employment over the course of the year was used, neither are the months of actual work, etc. considered. Moreover no data on specific characteristics of unstable workers and their occupations that may account for differences are considered. The data on income provided here are only intended to highlight the limited income from work of unstable workers.

6 Housing Deprivation and Vulnerability in Western Europe Pietro Palvarini and Emmanuele Pavolini

Introduction The debate on social disadvantage has evolved considerably in recent decades. The traditional conception of the problem – largely one- dimensional and focused on economic poverty – has gradually been replaced by a variety of theoretical perspectives that address the issue in multi- dimensional terms.1 Nevertheless, until quite recently, discussion at the European level on the indicators of social disadvantage has concentrated mainly on factors such as income and occupational status. Also, in the political arena, decisions within the European Union have moved in this direction during the past decade, with the recognition that deprivation indicators should be developed on dimensions other than income and employment, such as, for example, ones relating to housing (Laeken Council of 2001). Only in most recent years has theoretical consensus on the need for a composite approach shifted to the empirical level with the devising of research programmes for the multi- dimensional study of social disadvantage. Much of this research has recognized housing as one of the main areas in which deprivation occurs (Eurostat, 2000, 2003a; Whelan et al., 2001; Atkinson et al., 2002). In keeping with this strand of analysis, this chapter examines housing deprivation as a specific form of social vulnerability. The definition of housing deprivation will be examined in particular.2 To be stressed, however, is that the following analysis does not consider the most extreme forms of housing deprivation. The focus will be on situations of housing inadequacy rather than housing exclusion or homelessness. These, in fact, are the forms of deprivation best suited to interpretation through the ‘lens’ adopted in this book, namely that of social vulnerability. Vulnerability springs from precarious access to essential material resources and from the fragility of a person’s social network (see Chapter 1). It is evident that housing is of major significance in both these dimensions. First, the dwelling is an economic resource of prime importance: it is generally 126

Housing Deprivation and Vulnerability in Western Europe

127

the main item in the household budget and, together with income, the principal factor determining the resources available to a household and, therefore, its standard of living. At the same time, the home, especially if owner- occupied, is one of the principal forms of wealth possessed by families. It is so in regard to both the implicit income that it produces and its redistributive effects within the family life- course and inter-generationally (Kurz and Blossfeld, 2004). Thus, housing deprivation affects individuals and households principally in terms of their economic resources. However, the role of the home is not restricted to the economic domain alone. It is also the core of people’s social space: its location conditions access to other types of resources (work, welfare and leisure) and it simultaneously impacts on the relational and informal support networks upon which individuals and households can rely (Forrest and Kearns, 2001; Friedrichs et al., 2005). Moreover, the home is replete with socially constructed meanings, in that it represents stability, cognitive order and existential security (Dovey, 1985), as well as privacy and freedom from social control (Saunders and Williams, 1988). Given the complexity of the object ‘home’ and its manifold connections with diverse aspects of its ‘occupants’ social lives, housing deprivation can have serious consequences. First, it may combine with a shortage of economic resources and a poor standard of living. Second, it may complicate familial and social relationships. Third, it may restrict the creation and maintenance of spaces for privacy. And fourth, it may create a discrepancy between a person’s self-perception and his/her socially recognized identity, thus giving rise to cognitive insecurity and uncertainty about the future. Thus, although housing deprivation does not constitute outright exclusion, it falls wholly within the field of social vulnerability, as it may cause stresses and difficulties in the organization of everyday life. Against this interpretative background, the aim of this chapter is to map housing deprivation along two dimensions: the territorial dimension and the household dimension. As regards the former, the intention is to determine – besides differences among countries – the extent to which significant differences exist at the sub-national level. Rather than the more standard comparison among countries, the purpose here will be to verify whether significant similarities and differences exist, first between geographical areas of similar size, and second among internally homogeneous territorial areas.3 In other words, the intention is to ascertain whether countries of large size (France, United Kingdom, Italy, Germany and Spain) exhibit marked sub-national variations and how these match the characteristics of smaller countries. Besides the territorial dimension, analysis will be focused on households, with the purpose of understanding whether and to what extent family structure correlates with different forms of housing deprivation. Finally, the chapter will construct a framework based on both these dimensions. The overall aim of the chapter is to answer these research questions by means of an empirical analysis that moves through the following stages.

128 Social Vulnerability in Europe

The second section will provide an overview of housing conditions at European level based on data drawn from a variety of institutional sources. The third section conducts comparative analysis of housing deprivation in the different European macro-regions using ECHP 2001 data. The various dimensions of deprivation will be analysed separately, maintaining a distinction between severe and moderate forms in order to take account of the above-mentioned gradualism of the phenomenon. The fourth section constructs a typology of housing deprivation for investigating the overlaps and interweavings among the various dimensions considered. This phase of the analysis clarifies whether the various problems connected with housing tend to concentrate or whether, conversely, they tend to arise independently. The fifth section uses the typology of deprivation constructed in the preceding section to conduct a cluster analysis of the European macro-regions that allocates them among four groups with similar housing characteristics. The sixth section analyses housing deprivation by type of household, in order to determine which households are most vulnerable to the phenomenon and what kinds of household are most frequently associated with particular types of housing difficulty. The seventh section summarizes the main results of the analysis. The methodological appendix sets out the operational definitions of the indices used in the chapter.

An analysis of the available official statistics The key issues in the current debate and the assumption underlying the international literature can be stated as follows: although the large majority of households in Western Europe are home- owners, European society has not solved its housing problems. Paradoxically, in the recent decades of absolute growth in affluence, tensions resulting from housing difficulties have grown more acute. On the one hand, the housing conditions of most European households have improved; on the other hand, certain situations have deteriorated to the point that they resist solution by persons with insufficient resources to afford their own home (Tosi, 1994). Before the data from the ECHP survey are illustrated, there follows a general overview of housing conditions based on figures issued by national and European statistical institutes. The figures concern three main aspects of the ‘housing question’: home- ownership, overcrowding (particularly among more economically disadvantaged groups) and the cost of housing. As regards ownership, the information available covers a sufficiently long time span – the period 1980–2003 – for conclusions to be drawn. Figure 6.1 and Table 6.1 enable two analyses: (i) of the positions of individual countries in terms of the incidence of home- ownership and its trend from 1980 to 2003 and (ii) with this long period broken down into two shorter time spans (1980–1990; 1991–2003) of the extent to which the growth of homeownership occurred in one decade rather than the other.

Housing Deprivation and Vulnerability in Western Europe

15,00

129

Italy Netherlands United Kingdom

Var % ownership: 1980–2003

10,00

Spain

Belgium

France

Luxembourg Germany (excl. Ex-GD) Portugal

5,00

Austria

Sweden Ex-GDR Ireland Finland

0,00

Greece Denmark

−5,00 30,00

40,00

50,00

60,00

70,00

80,00

90,00

% Ownership: 2003

Figure 6.1 Home- ownership and variation in home- ownership over time (1980– 2003) in the EU-15, by country Source: Authors’ own calculations using data from national statistical institutes; Ireland: Department of the Environment, Heritage and Local Government (2006).

The first analysis identifies four large clusters of countries in Europe: ●

●

countries with at least two thirds of households being owner-occupiers and which have reached this level because of strong increases (above 5 per cent) in the relative incidence of owner-occupation during the period 1980–2003; this concerns the two larger Mediterranean countries (Spain and Italy) together with the United Kingdom, Luxembourg and Belgium; countries in which at least two thirds of households are owner- occupiers and which maintained that level without significant variations during the period 1980–2003: this concerns Greece and Ireland, countries which in 1980 already had very high proportions of home- owning households (around 75 per cent);

130 Social Vulnerability in Europe Table 6.1 Percentage of owner- occupied housing by country 1980–2003

Austria Belgium Denmark Finland France Germany (excl. former East) Germany (former East) Greece Ireland Italy Luxembourg Netherlands Portugal Spain Sweden United Kingdom

Var. 1980–2003

Var. 1980–1990

Var. 1990–2003

1980

1990

2003

52 59 55 63 47 39

55 67 54 72 54 42

58 68 53 63 56 45

6 9 ⫺2 0 9 6

3 8 ⫺1 9 7 3

3 1 ⫺1 ⫺9 2 3

31

26

34

3

⫺5

8

75 76 59 60 42 52 73 42 58

76 79 68 64 45 67 78 39 65

74 77 73 67 55 58 82 46 69

⫺1 1 14 7 13 6 9 4 11

1 3 9 4 3 15 5 ⫺3 7

⫺2 ⫺2 5 3 10 ⫺9 4 7 4

Note: Own calculations using data from national statistical institutes; Ireland: Department of the Environment, Heritage and Local Government (2006).

●

●

countries in which fewer than two thirds of households are owneroccupiers (the percentage generally ranges between 50 and 60 per cent) and which recorded marked increases in the period 1980–2003; this concerns (West) Germany, France, Austria, Portugal, Sweden and the Netherlands (in this last country the variation over almost 25 years has been especially pronounced, being +13 per cent); countries in which less than two thirds of households are owner- occupiers and which also have a lower growth in ownership; this concerns Finland and Denmark, the latter country indeed being characterized by a 2 per cent decrease in the incidence of home- owning households.

Thus, although it is possible to divide the European countries into four large groups, it should be emphasized that, during the period 1980–2003, only some countries recorded progressive home- ownership growth of the same intensity in both the 1980s and the 1990s: ●

●

Spain, Austria, Germany, Luxembourg and Denmark maintained the same rate of growth: between 3 and 5 per cent in both decades; Italy, the United Kingdom, Belgium, France and the Netherlands recorded growth, but at very different rates in the 1980s (the decade when it was

Housing Deprivation and Vulnerability in Western Europe

●

131

much higher, except in the Netherlands) and the 1990s (with relatively smaller positive variations); Ireland, Greece, Portugal, Finland and Sweden instead exhibit trends of different directions in the 1980s (positive, except for Sweden) compared with the 1990s (negative).

Overall, therefore, the 1980s were a period of generally strong growth in home purchases (almost 5 per cent on average at European level), with some exceptions (Sweden and Finland). But during the 1990s the growth subsided (+1.5 per cent at European level). Either the intensity diminished or the trends of the previous decade reversed in the majority of countries, although with some exceptions. The growth of home- owning households in the second half of the 1990s also affected the more disadvantaged social groups (those with incomes amounting to less than 60 per cent of the national average) in almost all European countries (see Table 6.2). In the same period, this type of household seems also to have achieved qualitatively better housing conditions if these are measured in terms of the incidence of overcrowding, which Table 6.2 The housing situations of poorer households (incomes ⬍ 60 per cent of the national average) (% of poor households)

Home- owning households

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

Households with overcrowding problems*

1995

2000

Var. 1995–2000

1995

2000

40.6 52.2 35.5 na 42.1 23.5 89.1 57.3 64.2 45.7 27.4 59.2 74.6 na 48.0

46.5 59.9 38.7 42.6 45.5 32.5 87.3 67.6 65.1 44.9 32.1 61.7 83.4 40.6 51.0

5.9 7.7 3.2 na 3.4 9 ⫺1.8 10.3 0.9 ⫺0.8 4.7 2.5 8.8 na 3

15.8 13.0 5.5 na 16.7 16.8 27.1 37.8 42.1 19.1 2.9 22.1 24.7 na 7.2

9.3 12.1 8.8 8.1 15 13.3 25.4 19.1 38.3 23.5 3.9 21.5 19.2 5.2 5.3

Notes: na: data not available. * Criterion: more than one person per room. Source: Authors’ own calculations on Eurostat (2007) data.

Var. 1995–2000 ⫺6.5 ⫺0.9 3.3 na ⫺1.7 ⫺3.5 ⫺1.7 ⫺18.7 ⫺3.8 4.4 1.0 ⫺0.6 ⫺5.5 na ⫺1.9

132 Social Vulnerability in Europe

diminished almost everywhere (except for some countries, in which, however, the incidence was relatively minor). Thus housing conditions have improved, at least in terms of overcrowding, and home- ownership has increased also among lower-income groups. However, throughout Europe expenditure for home rentals increased by around 20 per cent between 1995 and 2001. The largest increases, generally equal to or above 25 per cent, were recorded in most of the Mediterranean countries and, above all, in Ireland. They were lower in continental Europe (France and Germany especially) and Scandinavia (with the partial exception of Finland). The co-presence of the two above-mentioned phenomena (increased home- ownership and high housing costs) suggests that, for some sections of the population, access to a dwelling has become more costly. The paradoxical finding is that in Europe (or in some areas of it) relative deprivation has grown. In the next sections, the ECHP data are analysed to test this hypothesis more thoroughly.

The dimensions of housing deprivation Housing deprivation is a complex phenomenon, and there is no consensus among scholars on how it can be defined. The specialist literature abounds, but it is contradictory in its conclusions. Two features in particular make the study of this phenomenon particularly complex. The first is its gradualism, or the fact that deprivation can occur with different levels of severity (Edgar et al., 2002). The second is its multi- dimensionality, or the fact that the range of possible housing problems is very broad and diversified (Murie, 1983). In what follows, this complexity is addressed through separate analyses of the different dimensions of housing deprivation, and gradation of the phenomenon according to increasing levels of severity. The term ‘gradualism’ refers to the fact that housing deprivation comprises situations of differing degrees of intensity: (i) extreme situations of outright housing exclusion (homelessness); (ii) forms of severe housing deprivation, less visible but more widespread than the previous ones, in which the problem is not the lack of a dwelling but its inadequacy; and (iii) situations of moderate housing deprivation, characterized not by outright hardship but disadvantage relative to normal housing conditions. There are three main approaches to the issue of housing deprivation in the literature. The first studies situations of type (i), namely phenomenon of housing exclusion. This approach also appears in a large body of specialist literature on homelessness (for a recent comparative review see Fitzpatrick and Christian, 2006). A second approach includes the less extreme forms of deprivation within the same interpretative framework as homelessness, so that housing inadequacy and relative housing

Housing Deprivation and Vulnerability in Western Europe

133

disadvantage are viewed as factors provoking the risk of outright homelessness (Springer, 2000; Brousse, 2004; Edgar and Meert, 2005; Clanché, 2008). This approach has been criticized because it furnishes little empirical evidence of continuity among the different levels of housing deprivation (Marpsat, 2005; Tosi and Torri, 2005). A third approach focuses on housing inadequacy, considered in this case not as a preliminary stage toward homelessness, but as one possible form of social disadvantage understood in a multi- dimensional sense (Dale et al., 1996; Whelan et al., 2001; Moisio, 2004; Navarro and Ayala, 2008). This last approach will be used in the following analysis, which will concentrate on the second and the third levels of the scale just outlined – that is, on the situations defined as severe and moderate housing deprivation respectively. At the base of this choice is the theoretical assumption, set out in the introduction to this chapter, that such states of non-extreme deprivation represent one of the possible forms of social vulnerability. The second source of complexity in the notion of housing deprivation is its multi- dimensionality. It should be borne in mind that the social sciences have long addressed issues to do with housing on the basis of physical indicators that classify dwellings according to their structural features (Dale et al., 1996). This perspective tends to identify housing deprivation with the presence in the residence of structural problems or overcrowding. This approach to housing deprivation depicts a one- dimensional phenomenon closely correlated with economic poverty. These indicators were indubitably suitable for describing the phenomenon in a historical period (up until the mid-twentieth century) when the housing issue in Europe was essentially a matter of poverty, and attention concentrated on the hygiene and salubriousness of the domestic environment. However, during the period of intense economic growth which followed World War II, average housing conditions in Europe improved progressively, also by virtue of housing policies able to furnish a large number of affordable dwellings of a good standard (Balchin, 1996). In this phase, housing deprivation was regarded as a transitory problem bound to be remedied by economic development. This forecast was mistaken, however, because in recent decades housing deprivation has not disappeared. Rather, it has assumed new features and proved to be a persistent phenomenon connected with long-period structural factors, such as the shift to the post-Fordist economy and the crisis of the welfare state (Harloe, 1991). The so- called ‘new housing shortage’ has resulted from profound transformations in both housing supply and demand, such as changes in production systems and their impact on social stratification (increased inequalities, impoverishment and greater social vulnerability), the spread of neo-liberal policies within the housing sector, demographic shifts and changes in the family structure and relational and solidarity networks (Hallet, 1993).

134 Social Vulnerability in Europe

Today, the traditional approaches to the study of housing seem no longer able to yield adequate understanding of such complex relations. The most recent theoretical and empirical studies use a broader, multi- dimensional concept of housing deprivation that also emphasizes, besides the usual physical and structural aspects, new ones: insecurity of tenure, the stress caused by the costs of mortgages and rents and the difficulties caused by the neighbourhood in which the dwelling is situated (Avramov, 2002, 2006; Dewilde, 2004; Till, 2005; Tomlinson et al., 2008). It is no longer correct to identify a single form of housing deprivation. Rather, diverse features represent housing deficiencies. Five main phenomena are of interest to housing studies, and these represent the same number of dimensions of housing deprivation: ●

●

●

●

●

Unaffordability : the pressure on households because housing costs (rent or mortgage) take up too large a proportion of the household income (for a review of the concept of affordability see Hancock, 1993). Overcrowding: the inhabitants have insufficient space for their needs, or not enough according to given standards (for a review of definitions and measures of overcrowding see Gray, 2001). Housing inadequacy : structural deficiency or a lack of housing facilities (see Kutty, 1999 for an overview of the reference literature). Quality of the neighbourhood: it is not the dwelling, but its setting which is perceived as a problem (on the role of the neighbourhood in urban studies see Kearns and Parkinson, 2001). Insecurity of tenure: tenure conditions do not allow the household to make stable and legitimate use of the home (eviction, squatting, unregistered leases or without guaranteed renewal on expiry) (see Crane and Warnes, 2000).

This study adopts a multi-dimensional approach that analyses, first separately and then jointly, the various components of housing deprivation, depicting the complexity of the phenomenon as faithfully as possible. It accordingly constructs four indices of housing deprivation corresponding to four of the five dimensions just identified: an index of affordability, an index of overcrowding, an index of the dwelling’s physical quality and an index of the quality of the neighbourhood. It has not been possible to construct an index of tenure insecurity, because the ECHP database does not contain information on evictions, squatting or irregular leases. Each of the following sections will analyse one of the four dimensions of housing deprivation just identified. Housing affordability Affordability concerns the economic sustainability of a household’s housing costs. One of the most useful definitions of this concept is provided by

Housing Deprivation and Vulnerability in Western Europe

135

MacLennan and Williams (1990: 9): ‘affordability is concerned with securing some given standard of housing (or different standards) at a price or rent which does not impose, in the eye of some third party (usually government) an unreasonable burden on household incomes’. Although the concept is quite intuitive, its operationalization is rather complex, and there is no consensus within the scientific community on how to measure it. There are two main problems: how to measure the weight of housing expenditure on income, and what threshold can be fixed as ‘unreasonable’ for households. The approach with the longest tradition, and most widely used, is the ‘ratio approach’, so called because it measures affordability as the ratio between housing expenditure and the household’s total income. The condition of unaffordability arises when a household spends more than a given threshold percentage of its income on housing. This threshold – generally defined a priori on the basis of conventions rather than precise theoretical assumptions – has been varyingly fixed at between 20 and 40 per cent of income (for a review of this approach see Hulchanski, 1995). Although this approach is widely used, it has some major drawbacks: its main weakness is that it does not take account of the differences among different types of household, or among households with different income levels. For example, income remaining equal, a smaller household may spend more on housing than a large one which must satisfy the non-housing needs of several people. Likewise, household composition being equal, a high-income household can afford to allocate a larger proportion of its budget to housing expenditure. It is evident that a single threshold valid for all situations is an oversimplification. Indeed, some researchers propose abandoning this approach for the so- called ‘residual income approach’ (Stone, 1993, 2006a, 2006b). In this case it is not the ratio between housing expenditure and income that is considered, but the difference between them, which is the household’s residual income after housing expenditure. The residual income is then compared to a monetary sum assumed as the minimum amount necessary to meet the household’s non-housing needs. Two methods have been proposed. The first is to construct a basket of goods and minimum services necessary for the maintenance of the various types of household, and then to calculate its monetary value: if the residual income after housing expenditure falls below this value, the household has an affordability problem (Stone, 1993 speaks of ‘shelter poverty’). On the other hand, the adoption of a more common threshold that is easier to calculate than the standard basket has been proposed. Kutty (2005) sets the affordability threshold at two thirds of the poverty threshold: if the residual income after housing expenditure is less than this amount, the household has an affordability problem that puts it in a situation of ‘housing-induced poverty’. In our analysis we used Kutty’s version of the residual income approach. We therefore constructed an affordability index as the ratio between residual household income after housing expenditure and the relative poverty

136 Social Vulnerability in Europe

threshold calculated on a regional basis (see the methodological appendix for details). The index was coded into the following four categories corresponding to conditions of increasing well-being in regard to housing: unaffordability, risk, equilibrium comfort. As Table 6.3 shows, affordability problems are anything but marginal in Europe: 6.5 per cent of European households have unaffordable dwellings, in that they spend more than they can afford on housing. This group’s income net of housing expenditure is in fact less than two thirds of the relative poverty threshold, and it is not therefore enough to meet non-housing needs. More indirectly, this dimension of housing deprivation affects another broad group of households as well, those termed ‘at risk’. This concerns

Table 6.3 Percentage of households with different degrees of housing affordability, by macro-region (2001)

Southern/South-Eastern United Kingdom Western Germany France Île- de-France Northern Germany Central United Kingdom Southern Germany Denmark Netherlands Eastern Germany Southern Italy Finland Mediterranean France Greece Western/South-Western United Kingdom Northern United Kingdom Austria Central/North-Western Spain Southern Spain Eastern Spain and Madrid Portugal Western, South-Western France Central-Eastern France North-Western Italy Northern- Central Italy Ireland Belgium Luxembourg Total

Unaffordability

Risk

Equilibrium

Comfort

Total

9.40

11.10

39.80

39.70

100.00

9.30 9.10 9.00 8.40 8.10 8.10 7.40 7.20 7.20 7.10 7.10 7.10 7.10

9.30 9.40 11.10 11.90 11.10 11.70 11.20 10.90 7.70 10.10 10.80 9.30 10.30

45.90 43.40 50.80 38.80 44.60 48.80 47.60 48.80 38.90 44.20 40.60 36.10 42.70

35.50 38.10 29.10 40.90 36.20 31.40 33.80 33.10 46.20 38.60 41.50 47.50 39.90

100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00

7.00 6.50 6.50

14.10 8.00 7.70

37.00 43.90 39.90

41.90 41.60 45.20

100.00 100.00 100.00

6.20 6.00 5.90 5.80

7.40 9.00 12.10 9.90

41.40 39.80 37.60 46.50

45.00 45.20 44.40 37.80

100.00 100.00 100.00 100.00

5.60 5.50 4.70 4.40 3.70 2.90

9.50 6.90 8.00 15.60 8.90 9.80

46.90 43.20 41.80 36.50 46.80 46.00

38.00 44.40 45.50 43.50 40.60 41.30

100.00 100.00 100.00 100.00 100.00 100.00

6.50

10.10

42.90

40.50

100.00

Source: ECHP, Authors’ own calculations.

Housing Deprivation and Vulnerability in Western Europe

137

the 10 per cent of households able to rely on residual incomes just sufficient to cover their non-housing expenditures. In their case, an unexpected event that reduces income or increases housing costs may cause the onset of poverty. Then there are the 43 per cent of households whose housing expenditure does not imbalance the household budget, while the 40 per cent at the top of this affordability scale are in a comfortable situation because, after they have paid their housing costs, they have a household budget two times higher than the poverty threshold. Unaffordability depends both on housing costs disproportionate to income and on a very low base income: 77 per cent of the cases classified in the unaffordability category have incomes below the poverty threshold (calculated on a regional base) – a proportion that diminishes to 44 per cent for households ‘at risk’. If regional differences in the distribution of affordability problems are then analysed, one finds that they are more widespread in countries such as United Kingdom, Germany, Denmark and the Netherlands and in the regions containing the largest metropolitan areas (particularly London and Paris). In these zones, where there is greater demand for housing and the prices of dwellings tend to be higher, the affordability problem arises with greater frequency, above all for lower-income households. As regards Italy, the pattern differs somewhat between the South and the rest of the country: while in the North and the Centre the frequency of affordability problems lies below the European average, the South has a high concentration of both unaffordability situations and lack of housing comfort. This pattern is also to be found in Greece, where it highlights a ‘polarization’ between households placed under excessive strain by housing costs and households that are instead well protected (probably benefiting from extremely widespread home- ownership). Crowding The second dimension of housing deprivation analysed was crowding, which is one of the aspects most frequently examined by studies on housing problems. In this case too, however, some preliminary definitions are necessary. Usually meant by the term ‘crowding’ is a housing situation in which the space available to each member of the household falls short of a standard considered acceptable. A distinction should be drawn between ‘density’ and ‘crowding’: density is an objective measure referring to the number of persons who occupy a particular space, and it therefore has neither positive nor negative connotations. ‘Crowding’ refers to an excessive density that may lead to a lack of privacy, an increase in unwanted interactions and psychological distress (Gray, 2001). Definition of crowding or of overcrowding entails a normative judgement that decides the degree of density deemed acceptable in a particular society.

138 Social Vulnerability in Europe

The standard of ‘one person per room’ is widely used in the literature, even if it raises a number of obvious problems. First, the criteria for its selection are rarely made explicit, and they generally reflect an ethnocentrism largely inspired by norms and values typical of the middle classes (Myers et al., 1996). Moreover, the use of the number of rooms as an indicator of the space available to a household is questionable: it would probably be more correct to consider the surface area of the dwelling. Again, the one-person-per-room ratio excludes a priori that people living alone can suffer from overcrowding. Finally, this threshold systematically penalizes the largest households with children, for which the one-to- one ratio is unrealistic. Bearing these problems in mind, for the purposes of our analysis we nevertheless decided to use one person per room as the threshold of overcrowding. There were two main reasons for this decision. The first concerned data availability: it was not possible to refer to housing surface area, because data on this aspect are not collected by ECHP. The second reason was substantive in nature. Once it had been established that the indicator was to be based on the number of rooms, the overcrowding threshold could be fixed in either absolute terms (that is by establishing a standard value) or relative ones (i.e. by setting a threshold value in function of the average number of rooms per capita in the sample). Given that this latter value was very high in the sample (around three rooms for a single person and four for a couple), a relative threshold would have identified as overcrowded situations not commonly perceived as such (e.g., a single person in one room or a couple in two rooms). It was therefore decided to use the standard of one occupant per room, which, despite all the above-mentioned drawbacks, undoubtedly has the advantage of having a long and consolidated tradition in the literature (it was proposed as the standard as early as 1958 by Cullingworth). Nevertheless, in order to remedy at least the problem of over- estimation of crowding in households with children, the standard criterion was modified. Account was taken of the differences in the space needs of adults and children by attributing different weights to each category (see the methodological appendix for details). European households were therefore classified on the base of an overcrowding index with three modalities that represented increasing well-being in housing conditions: overcrowding, normality and under-utilization. Around 8 per cent of households live in overcrowded housing, 72 per cent fall within the intermediate band and and 20 per cent live in dwellings overlarge compared with the size of the household, given that there are three or more rooms for each household member. However, these European averages are derived from the composition of highly diversified regional situations that tend to configure three distinct macro-regions (see Figure 6.2). The first consists of Southern Italy and Greece, where overcrowding affects 25 and 21 per cent of households respectively, and where the percentage of under-utilized accommodation is

Housing Deprivation and Vulnerability in Western Europe

139

Legend < 4% 4.1%–8% 8.1%–12% > 12.1%

Figure 6.2 Percentage of households with overcrowding, by macro-region (2001) Source: ECHP, authors’ own calculations.

half of that of the rest of Europe. There is then a second group of countries and regions, formed by Portugal, Spain, Central-Northern Italy and Ireland, where overcrowding is above the European average. In the macro-area of Central and Northern Europe, instead, the phenomenon grows increasingly marginal until it is almost non- existent in the case of the Netherlands. Physical quality of the dwelling Numerous studies stress that the traditional forms of housing deprivation – structural unsuitability, lack of services and physical degradation – are declining in all the Western countries, now that generally very good housing standards have been achieved (Power, 1993; Balchin, 1996; MacLennan et

140

Social Vulnerability in Europe

al., 1997; Giorgi et al., 2001). It is evident that the physical quality of the dwelling can no longer be taken as the sole indicator of housing deprivation. Nevertheless, its importance should not be underestimated. For this type of problem has not entirely disappeared: a proportion, albeit small, of citizens still live in sub-standard dwellings and this problem tends to increase where the housing standards are higher and the expectations with regard to the home are greater. The physical dimension was considered in our analysis as one of the possible vectors of housing deprivation and it was investigated using a relative approach. Deprivation was not defined in an absolute and context-free manner, but rather in relation to average conditions in the geographical region concerned. Quantitative and qualitative differences among the housing stocks of countries, normative standards, cultural preferences in regard to types of dwelling and climatic features are all factors that operate on a local scale and make the use of the same scale of evaluation at a Europe-wide level inadvisable. For each geographical region, therefore, we constructed a specific index of dwelling quality based on nine indicators concerning physical amenities: bath or shower; indoor flushing toilet; hot running water; a place to sit outside; insufficient light; lack of adequate heating facilities; leaky roof; damp walls, floors or foundations; and rot in window frames or floors. In every region, a different weight was assigned to each indicator according to its frequency in the context of reference, following a procedure well-established in the literature on deprivation in general and on housing deprivation in particular (Betti and Verma, 1998; Giorgi et al., 2001; Avramov, 2002) (see the methodological appendix for details). This approach considers the types of deprivation most rarely found in the population as the most severe and the most widespread types as the least severe. For example, the lack of a bathroom in a context where the vast majority of the population has access to one causes greater discomfort than the absence of central heating, where this amenity is less necessary because of climatic reasons. The continuous index thus constructed was then divided into three classes characterized by increasing levels of housing quality defined as poor, belowaverage and comfortable. Overall 70 per cent of European households live in dwellings with high- quality services and structures. Around a quarter of households have dwellings of below-average quality, while those suffering from outright housing deprivation amount to 5.5 per cent. The distribution of this index has a strong negative skew, that is the great majority of the cases have got high scores on it, which denotes the good physical quality of dwellings across Europe. Consideration of the number of deprivation indicators reported by the interviewees shows, in fact, that households in comfortable dwellings declared on average 0.1 problems out of the nine considered, while those with below-average quality dwellings cited 1.5 problems. In contrast, households with poor housing quality

Housing Deprivation and Vulnerability in Western Europe

141

referred on average to 3.5 problems. This is a rather large number – especially if one considers that today the level of housing quality is decidedly high in Europe. As regards regional differences (see Figure 6.3), a certain variability is apparent among the various areas of Europe, with levels of deprivation ranging from 1 to 8.3 per cent. The problem of low-quality housing tends to be more widespread in the Netherlands and the United Kingdom, while it is marginal in Finland, Belgium, Ireland and, especially, Germany. However, to be noted regarding this last country is a sizeable difference between the regions of former West Germany, which rank among the highest positions

Legend < 4% 2.6%–5% 5.1%–7.5% > 7.6%

Figure 6.3 (2001)

Percentage of households with poor housing quality, by macro-region

Source: ECHP, authors’ own calculations.

142

Social Vulnerability in Europe

in the dwelling quality table, and those of former East Germany, which record some of the highest rates of physical deprivation in Europe. These geographical disparities are not found in Italy, where the three macro-regions under consideration record similar shares of poor housing quality compared with the European average. The housing situation in the South differs from that of the rest of the country because of the greater occurrence of below-average quality dwellings. The quality of the neighbourhood So far we have analysed forms of deprivation with reference to the dwelling: its physical quality, size and cost. The fourth component examined does not have directly to do with the dwelling itself, but rather with its surroundings. The dwelling’s local setting is certainly one of the ways in which housing deprivation can be manifest, because people do not live solely within their houses but instead in a complex system of local interactions and relationships. Consequently, the neighbourhood has been analysed both by housing studies and research on deprivation in general (Avramov, 2002; Betti and Verma, 2004; Whelan and Maitre, 2007a). We investigated deprivation resulting from the neighbourhood by analysing four indicators present in the ECHP database: the presence of noisy neighbours; noise from outside the home; the presence of environmental problems, pollution or dirt; and the presence of crime or vandalism in the area. These were combined into a synthetic index of neighbourhood quality, which comprised three levels: poor neighbourhood quality, normality and good quality (see the methodological appendix for details). Unfortunately, as information on neighbourhood quality is not collected in all European countries, Germany and Luxemburg had to be excluded from the analysis. The frequency of poor neighbourhood quality is similar to the frequencies of the other dimensions of housing deprivation already discussed. It is around 7 per cent on the continental scale, 29 per cent of households fall within the second class characterized by intermediate neighbourhood quality while 64 per cent live in good- quality neighbourhoods with few or no significant problems. Figure 6.4 shows the distributions of this index by macro-region. Regional differences are also sizeable on this indicator, with poor neighbourhood quality ranging from 1.7 to 12.8 per cent. The areas with most widespread deprivation are Italy (above all North-Western and Southern), Portugal, France and the region comprising Eastern Spain and Madrid. The characteristics of these regions suggest that there may be a relationship between a region’s urban density and neighbourhood quality. This hypothesis is confirmed by analysis of the level of poor neighbourhood quality in relation to the degree of urbanization. In fact, the rate is 1.7 per cent in low density housing zones, rises to 4.4 per cent in those of average urbanization and

Housing Deprivation and Vulnerability in Western Europe

143

Legend < 4% 4.1%–6.5% 6.6%–9% > 9.1%

Figure 6.4 Percentage of households living in neighbourhood with poor quality, by macro-region (2001) Source: ECHP, authors’ own calculations.

reaches 10.8 per cent in densely urbanized areas. The regions with the best neighbourhood quality are Austria, Denmark, Ireland, Central Spain and the West of the United Kingdom, where deprivation is below 4 per cent, while the percentage of households reporting high neighbourhood quality is above 70 per cent.

A typology of housing deprivation Now that the four dimensions of housing deprivation have been examined, it is appropriate to analyse whether they combine in the experience of

144

Social Vulnerability in Europe

households. Do the distinct vectors of deprivation operate independently or do they tend to combine and give rise to forms of composite hardship? Although constructing an aggregate index of housing deprivation would have enabled us to arrange household situations on a scale of increasing severity, we believed it more insightful to construct a typology of housing conditions. This we did by cross-referencing the various dimensions examined above. There were two main reasons for this decision. First, constructing a synthetic index of housing deprivation would have compelled us to select a mathematical function with which to correlate its various dimensions. An operation of this kind could not be neutral; on the contrary, it would be biased by theoretical assumptions on relationships among the dimensions of housing deprivation and the weight of each of them. Although some attempts have been made to construct multi- dimensional indices of housing deprivation (see for example Marsh et al., 2000; Navarro and Ayala, 2007), they are largely based on physical indicators such as overcrowding and the presence of essential amenities. Thus, because a single housing deprivation index synthesising the four dimensions examined could not be backed by well- established findings, it would have been largely artificial and arbitrary. We therefore decided against developing such an index. But there was another reason for our decision: we wanted to evaluate not so much the intensity of housing deprivation as a whole as to see how its various dimensions combine, thereby highlighting their cumulative or compensatory effects. In this regard, a first important element emerges from the analysis of the correlations among the four dimensions considered. The four continuous indices previously constructed correlate only very weakly,4 which means that the four dimensions are likely to act independently. This first general finding can be expanded by describing how the typology of housing conditions was constructed. First, the three dimensions for which information was available on all countries were used: the physical quality of dwellings, the space available and affordability. The ordinal variables (with three or four categories) examined in the previous section were re- coded into dichotomous variables on the basis of the presence or absence of deprivation conditions.5 The three dichotomous variables were then cross- correlated to obtain the eight possible situations shown in Table 6.4. Although dichotomization always implies a loss of information, it was necessary in order to reduce the complexity of the classification and to obtain a typology that immediately combined the situations of deprivation in the various dimensions. The largest groups were considered independently, while the residual situations were aggregated. As shown by Table 6.4, the broadest group, which comprised three quarters of European households, is the one without housing problems. There are then three groups of different sizes formed by

Housing Deprivation and Vulnerability in Western Europe

145

Table 6.4 Percentage distribution of households by housing situation measured on three indices (EU 2001, %) Space available Affordability

Physical quality

Deprivation

Deprivation

Deprivation Non- deprivation

0.3 1.3

1.4 13.6

1.7 14.9

Non- deprivation

Deprivation Non- deprivation

0.5 6.1 8.2

3.3 73.5 91.8

3.8 79.6 100

Total

Non- deprivation

Total

Source: ECHP, authors’ own calculations.

households with only one of the possible deprivation conditions: those with only problems of affordability (13.6 per cent), only problems of crowding (6.1 per cent) and only problems of housing quality (3.3 per cent). There is finally a minority of multi-problem households (3.5 per cent) that experience different combinations of two problems, or the co-presence of all three conditions of deprivation. These results show the low level of accumulation among different forms of housing deprivation. In contrast, these forms seem to be substantially independent of each other. They affect different types of household and only in rare cases are they concentrated in the same one. It is now possible to develop the analysis further by adding the fourth dimension of housing deprivation – the neighbourhood – to the typology. However, neighbourhood quality is not recorded in all countries, so Germany and Luxembourg are excluded from the analysis. If this omission is acceptable, a more complete typology is obtained, in which the territorial dimension is combined with the others to produce new categories. The overall pattern does not change, but two new categories are formed: households that suffer only from a poor quality neighbourhood (4.4 per cent) and those with neighbourhood problems together with one or more further elements of deprivation (2.4 per cent). The fact that aggregate problems affect only a minimal share of the population suggests a further hypothesis: namely that deprivation in one dimension may be associated with better quality in the other dimensions. For example, one can imagine the existence of a trade-off between costs and housing standards whereby households pay for the good physical quality of the home, sufficient space and a satisfactory neighbourhood in terms of affordability. However, this hypothesis is not borne out by the empirical evidence. Analysis of the correlation between the indices of housing quality and housing expenditure shows that greater expenditure does not necessarily correspond to better housing quality. Both the index measuring the physical quality of the dwelling and the index measuring available space exhibit

146

Social Vulnerability in Europe

only a slight positive correlation with the amount of housing expenditure (the Pearson correlation index is equal to 0.091 and 0.122 respectively). In the case of the index of neighbourhood quality, this does not exhibit a statistically significant correlation with housing expenditure. Household income has only a slightly more intense relationship with housing quality (the correlation between income and the physical quality index is 0.182, between income and available space 0.151 and between income and quality of the neighbourhood 0.024). Consequently, expenditure on housing does not increase with housing quality, or, to be more precise, it increases to a lesser extent than the household income. Thus, households do not pay for better levels of physical quality of the dwelling, available space and quality of the neighbourhood in terms of affordability. Put otherwise, there is no trade- off between quality and housing costs. To conclude, the analysis of the correlations and the typological analysis yield consistent results. These substantially refute the hypothesis of widespread accumulation effects among the various dimensions of housing deprivation, as well as the hypothesis of a trade- off between housing quality and affordability. The diverse forms of housing deprivation tend to act independently of each other. Consequently, the various dimensions should be analysed separately, with no attempt to construct a cumulative index of housing deprivation that would arbitrarily combine largely independent dimensions.

Vulnerability and housing between housing regimes and geographical macro-regions The typology described in the previous section was used for analysis of the situation in the various areas of Western Europe. The intention was to determine whether it was possible to identify regions with homogeneous characteristics from the point of view of housing difficulties. Hierarchical cluster analysis was conducted for this purpose and it classified the various European regions into four groups, with a clear dividing line between the North and South. As shown in the previous sections, it is possible to construct two composite typologies of housing deprivation that differ in terms of the numbers and kinds of problems considered (three or four). This distinction was necessary because not all countries (particularly Germany and Luxembourg) furnish information on neighbourhood quality. The analysis reported in this section was based mainly on the more restricted indicator comprising three dimensions (affordability, physical quality of the dwelling and overcrowding) and this verified the clusters of regions that would emerge. The procedure was then repeated using the synthetic index with four dimensions (and therefore including neighbourhood

Housing Deprivation and Vulnerability in Western Europe

147

Table 6.5 Percentage of households with housing problems, by macro-region – three deprivation indicators (2001) No problems Greece and Southern Italy Iberian peninsula, Central-Northern Italy, Ireland Almost all of CentralNorthern Europe Small areas in CentralNorthern Europe (Lux., Bel., Au., W/E Fr.) Total

Physical Overquality crowding Affordability

Multiple problems Total

63

2.5

18.1

10

6.4

100

70.6

3.2

10.4

11.9

3.9

100

75.1

3.7

2.1

16.5

2.6

100

80.5

3.3

1.9

11.4

2.9

100

73.5

3.3

6.1

13.6

3.5

100

Note: Lux = Luxembourg, Bel = Belgium, Au = Austria, W/E Fr = Western and Eastern France. Source: ECHP, authors’ own calculations.

problems). The purpose was to see whether the previous classification was confirmed, or whether it changed markedly. Using only three dimensions of deprivation (see Table 6.5), the most problematic area appeared to be the Mediterranean countries (excluding France) and Ireland. However, there are two very different situations within this group: 1. Greece and South Italy are the areas in which housing deprivation is most widespread and intense: in fact, more than one third of households have at least one housing problem (37 per cent against an average in the rest of Europe of around 25 per cent) and, indeed, 6.4 per cent of them have several sources of deprivation (compared with less than 3 per cent in the other geographical regions). The problems seem particularly to concern overcrowding, which affects almost one household in every five, while unaffordability is a relatively less common problem; 2. the rest of Italy, Ireland and the Iberian peninsula as a whole form the second area, which, given its problematic features, is basically a replication of the previous pattern albeit with lower criticality: 30 per cent of the households in this area have at least one (and almost exclusive) housing problem, mainly concerning both the cost of housing and overcrowding. The latter phenomenon, albeit much less markedly present than in Greece and Southern Italy, is anyway much more widespread than in the rest of Central-Northern Europe. The rest of Central-Northern Europe shows a much better situation than Southern Europe, even though it has different types of housing problem,

148

Social Vulnerability in Europe

which concern affordability more and overcrowding less: 3. the majority of the countries in Central-Northern Europe have very similar characteristics, with around three quarters of households with no housing problems and dominance affordability problems (16.5 per cent); overcrowding is not a significant problem (2.1 per cent) and few households have multiple problems (2.6 per cent); 4. the fourth pattern appears to be a continuation – though better in terms of housing conditions – of the previous one, and it mainly concerns regions and countries of Central-Northern Europe – Austria, Belgium, Luxembourg and parts of France – where the band of deprivation (20 per cent) is narrower, mainly because here problems of affordability are less frequent than in the rest of Northern Europe (11.4 per cent of households compared with 16.5 per cent). This situation does not change significantly when the fourth dimension of deprivation – the quality of the neighbourhood – is introduced: the only difference is represented by Ireland, which moves from being an area of the Mediterranean in less difficulty to the area of countries of Central-Northern Europe. Given this pattern, it is therefore likely that – although it cannot be confirmed because of the lack of data mentioned above –Germany and Luxembourg should also remain in the clusters identified above. To what extent, therefore, does the regional dimension matter, compared with the national one, in differentiating the distribution of housing deprivation profiles? It seems that it generally counts to only a limited extent. However, two exceptions should be borne in mind: Italy and France. Italy seems particularly significant in light of the differences among its various areas. Southern Italy differs strikingly from the Centre-North, being far more characterized by situations of deprivation. France, in contrast, has relatively few situations of housing stress, which in some areas of the country – generally the non-metropolitan areas (the East and the West) – are even less common. A further dimension along which to investigate the relationship between geographical regions and housing deprivation is home- ownership (see Table 6.6). Households in rented dwellings – around 30 per cent of the total – suffer more frequently from housing deprivation (41.5 per cent), mainly because of rental costs, than households with their own homes (20 per cent), which represent two thirds of the households in the EU-15. In this latter case, the stress may sometimes be financial or provoked by overcrowding. No marked differences are apparent among tenant households in the relative distributions: the incidence of households with housing problems is practically the same among households leasing their homes, whether from public authorities, not-for-profit organizations or private landlords. Only the relative incidence of the various types of problem changes somewhat,

Housing Deprivation and Vulnerability in Western Europe

149

Table 6.6 Percentage of households with housing problems by type of occupancy Only Only No physical overOnly Multiple problems quality crowding affordability problems Rented

58.60

5.20

4.50

25.30

6.40

Free occupancy Owneroccupied

68.40

4.90

9.00

12.20

5.50

80.00

2.50

6.60

8.70

2.20

Total

73.50

3.30

6.10

13.60

3.50

Total 100.0 (29.3)* 100.00 (3.7)* 100.00 (67.0)* 100.00 (100.0)*

*

The percentages in parentheses refer to the overall distribution of households by type of occupancy. Source: ECHP, authors’ own calculations.

Table 6.7 Percentage of households with housing deprivation in different geographical areas, by type of tenure (2001) Owner- occupancy Southern Mediterranean area Iberian-Italian area ⴙ Ireland Central-Northern Europe Small areas in Central-Northern Europe

34.9 26.5 13.3 11.4

Tenancy 47.4 44.5 41.9 36.5

Difference T- O 12.5 18 28.6 25.1

Source: ECHP, authors’ own calculations.

with a larger proportion of households with economic difficulties in the social housing sector. The distribution of home- ownership is quite diversified among the various areas of Europe: it is particularly high in Southern Mediterranean countries (81.8 per cent) and tends to decrease moving northward, especially in Central-Northern Europe (58.9 per cent). Home- ownership in itself only partly ensures greater coverage against housing risks, compared with tenancy. Moreover, such greater coverage is conditioned by latitude (one third of home- owning households in the Mediterranean South have housing problems, compared with 11–13 per cent in Central-Northern Europe). Besides these features, Table 6.7 highlights two further interesting, and somewhat surprising, findings: ●

the percentage of tenant households with housing problems does not vary significantly across Europe, ranging from 41.9 per cent in CentralNorthern Europe to 47.4 per cent in the South Mediterranean area;

150 Social Vulnerability in Europe ●

however, the gap between households with and without housing deprivation according to whether they are owner- occupiers or tenants is much wider in the South than in the North of Europe; it is 12.6 per cent in the South Mediterranean area, 18 per cent in the Northern Mediterranean and 25–28 per cent in Central-Northern Europe.

These two phenomena can be given a twofold interpretation. On the one hand, in the areas in which home- ownership is most frequent, this situation does not ensure an increase in housing quality that is as marked as in the North of Europe. On the other, where home-ownership is less common, that is Central-Northern Europe, social inequalities related to housing quality are much more evident: home- ownership makes much more of a difference, in terms of reduced housing deprivation, than in South Europe. Some explanations of this phenomenon can be put forward on the basis of Table 6.7. It is immediately clear that, whereas in Southern Europe there are no particular differences in housing deprivation between owner- occupiers and tenants, in Central-Northern Europe being a tenant entails much greater exposure to housing deprivation, and especially to economic pressures, as we have already shown in the second section. This situation may be connected to the rental market, and in particular to the part of the rental market that it subject to public regulation and support.

Vulnerability and housing by household type The literature on housing issues emphasizes that the incidence of the forms of deprivation also differs by household type (Harding et al., 2004). Our Table 6.8 Percentage of households with housing problems, by type of household (2001)

Lone parent Single on welfare Couple ⴙ children ⬍ 25 MBW Working single Extended family Couple on welfare Couple ⴙ children ⬍ 25 dual- earner Couple ⴙ children ⬍ 25 OAH MBW couple Dual ⴙ OAH couple

No problems

Only physical quality

Only overcrowding

57.90 60.60 61.70

3.70 4.60 1.80

5.60 – 18.60

26.40 30.30 11.10

6.40 4.50 6.80

100.00 100.00 100.00

77.40 81.00 81.90 82.80

5.70 3.80 3.20 1.50

– 8.00 0.70 10.50

15.10 5.20 12.60 3.70

1.80 2.00 1.60 1.50

100.00 100.00 100.00 100.00

83.30

2.80

6.70

4.60

2.60

100.00

83.90 92.00

3.40 3.90

1.10 1.10

10.10 2.40

1.50 0.60

100.00 100.00

Source: ECHP, authors’ own calculations.

Only Multiple affordability problems

Total

Housing Deprivation and Vulnerability in Western Europe

151

calculations on ECHP data clearly indicate which types of households are in most difficulties, and of which kind (see Table 6.8). They identify four household types differentiated according to their level of vulnerability to housing deprivation: ●

●

●

●

non-working single persons, generally elderly pensioners, lone-parent households and single-income couples with children aged below 25 are the households most at risk of housing problems: around 40 per cent have at least one type of difficulty and they are often multi-problematic (4–7 per cent). The various household types are differentiated by the kind of problem suffered: on the one hand, couples with children are mainly affected by overcrowding (18.6 per cent), on the other, elderly single persons and lone parents with small children suffer markedly (26–30 per cent) from affordability problems; single workers form a distinct group characterized in around 23 per cent of cases by problems of affordability (15 per cent), but also to a significant extent by physical quality (5.7 per cent); almost all the other household types are in a better situation (extended families, elderly couples, dual-income couples with children and single-income couples without children), given that in these cases fewer than one fifth of households are in difficulty, although they are subject to some forms of hardship (households with dependant children or elderly relatives have more marked problems of space while single-income couples or elderly couples have greater affordability difficulties); dual-income couples without children are those with the lowest housing deprivation (7.9 per cent). However, it should be borne in mind that this result is, at least in certain areas of Southern Europe, closely bound up with the difficulty of young people in starting a family in the absence of sufficient financial resources. Consequently, the low level of housing deprivation may indicate that these are childless couples already wellestablished in terms of income and employment.

The European situation just outlined changes significantly when the macro-regions described in the previous section are introduced. Given that the differences in housing conditions between areas, such as those of the Mediterranean South and those of Central-Northern Europe, have already been illustrated, we now examine two aspects in particular: the classification by macro-regions of the most deprived households (i.e. the ranking of the household types most at risk in each region), and the level of exposure to risks in the various regions for the same type of household (see Table 6.9). As regards the first point, the rankings of the various household types according to housing difficulties differ substantially: ●

in Mediterranean Europe in general, it is single-income households with children that are most vulnerable to housing deprivation, but they are not

152 Social Vulnerability in Europe Table 6.9 Percentage of households with housing deprivation in different geographical areas, by type of household (2001)

Couple with children ⬍ 25 MBW Lone parent Couple with children ⬍ 25 dual- earner Single on welfare Couple with children ⬍ 25 OAH Extended family Couple on welfare MBW couple Working single Dual ⫹ OAH couple

Small areas in CentralNorthern Europe

Southern Mediterranean area (Greece South Italy)

IberianItalian area CN+ Ireland

53.4

44.3

28.6

23.7

39.7 35.5

37.4 21.0

49.5 12.0

36.5 11.4

32.8 32.1

42.4 23.9

42.3 14.7

31.4 10.7

31.3

19.4

15.9

11.2

27.0 23.2 16.0 10.7

20.5 18.1 17.4 6.9

16.9 14.4 25.8 8.0

12.7 14.0 18.2 7.0

CentralNorthern Europe

Source: ECHP, authors’ own calculations.

●

●

among those most at risk in Central-Northern Europe. While for a couple in Southern Europe having children and a single income entails very high exposure to housing risk (53 per cent in the South Mediterranean area and 44 per cent in the Iberian-Italian and Irish area), such vulnerability radically diminishes in Central-Northern Europe, where it concerns around a quarter of households belonging to this category; in Central-Northern Europe, by contrast, it is lone parents with children who are faced by the greatest difficulties, but although they are among the most disadvantaged household types in Southern Europe, they do not occupy first place. Interestingly, there is a higher proportion of loneparent families in difficulties (49 per cent) in Central-Northern Europe than in Southern Europe (40 per cent). This is the only case – apart from that of the working singles discussed below – in which vulnerability to housing deprivation is higher in Central-Northern Europe than in the South; the single elderly form a third category in difficulty in all areas, although in this case the differences between Northern and Southern Europe are less marked, especially if the largest areas of these two macro-regions are compared (on the one hand, the Iberian peninsula, Central-Northern

Housing Deprivation and Vulnerability in Western Europe

●

●

153

Italy and Ireland with, on the other, the largest areas of Central-Northern Europe); by contrast, dual-income childless couples everywhere (and largely in the same way) have fewer housing difficulties; all the other household categories occupy intermediate positions. Two of these categories warrant particular attention, given the distinctive nature of their housing condition. Working singles tend to have worse housing conditions in most of Central-Northern Europe compared with their counterparts in the South. This finding is explained by the fact that young people in Mediterranean Europe tend to live at home longer (see Chapter 8 on this aspect). On the one hand, this causes more stress than elsewhere for families with children, as illustrated above for single-income couples with children; on the other, it reduces the number of singles with housing difficulties. Finally, in the South Mediterranean area (Southern Italy and Greece) housing difficulties, mainly caused by overcrowding, also affect dual-income couples with childrens; a phenomenon that is less marked in other areas.

Concluding remarks This chapter has addressed two sets of questions: (i) the intensity of housing deprivation in Europe and the various forms that it takes and (ii) how these phenomena differ according to the geographical context and type of household. The first step in the analysis was the identification of four dimensions of housing deprivation: affordability, overcrowding, physical quality of the dwelling and quality of the neighbourhood. Then indices were constructed with which to determine the level of deprivation experienced by households in each of the dimensions considered. It was thus possible to furnish an overview of housing deprivation both in Europe as a whole and at the level of geographical macroregions. The analysis showed that, although European households generally enjoy high housing standards, certain difficulties still persist: indeed, on each of the dimensions investigated, between 5 and 8 per cent of European households suffer housing deprivation, albeit with marked geographical differences. Joint analysis of the four vectors of housing deprivation was then conducted and this enabled a multi-dimensional typology of the phenomenon to be constructed. When this typology was used as the basis for analysis, the area of vulnerability expanded: around 30 per cent of European households were shown to have problems in at least one of the four dimensions considered. It was also evident that the four forms of deprivation act independently of each other, with no significant compensatory or cumulative effects. The geographical distribution of housing deprivation tends to structure itself along supranational and not sub-national lines, and often even broader ones. Our analysis also confirms the dividing line between Central-Northern

154

Social Vulnerability in Europe

and Southern Europe (plus, in many respects, Ireland), with the former being less characterized by housing deprivation than the latter. Within these two large groups we observed, on the one hand, ‘downward’ differentiations, with a further worsening of conditions in some areas of the South (Greece and Southern Italy), and on the other, ‘upward’ ones, with even more positive situations in some regions (generally non-metropolitan) and small countries in Central-Northern Europe. Only two countries – France and Italy – have regions belonging to different clusters. While in the case of France this results from more satisfactory housing conditions in the East and West compared with the rest of the country, Italy exhibits a ‘cleavage’ between disadvantaged areas in the South and ones with much better housing conditions in the Centre-North. It has been found that households are subject to different forms of housing stress. While some types of household are particularly vulnerable everywhere in Europe (generally, lone parents with children, elderly people living alone, single-income couples with children and, to a lesser extent, working singles), it is interesting that the severity of their housing deprivation differs according to the geographical context. In Mediterranean Europe, single-income families with children are the most vulnerable to housing deprivation, but these are not among the families most at risk in CentralNorthern Europe. In this latter area, it is lone parents with children who have the greatest housing difficulties. Moreover, lone-parent families (together with working singles) in Central-Northern Europe record a higher incidence of deprivation than their counterparts in Southern Europe.

Methodological appendix I Affordability index The traditional approach to affordability uses the ratio between housing expenditure and income, and it calculates RIR indexes (‘Rent to income ratio’ or, more generally, ‘Expenditure to income ratio’).6 The analysis reported in this chapter has instead used the ‘residual income approach’ (Stone, 1993, 2006a, 2006b; Kutty, 2005), which considers the household’s residual income net of housing expenditure and compares it against a monetary sum assumed as the minimum amount of money necessary to meet non-housing needs. The affordability index used in this chapter is written as follows:

IAF

RF SA 1 n SP

where: RF ⫽ net annual household income; SA ⫽ annual housing expenditure;

Housing Deprivation and Vulnerability in Western Europe

155

ñ ⫽ number of adult equivalents according to the modified OECD equivalence scale (which assigns a weight of 1 to the first adult; 0.5 to each of the other adults; and 0.3 to each child); and SP ⫽ regional relative poverty threshold (50 per cent of the median of equivalent incomes). The continuous affordability index thus obtained was then divided into four classes: Unaffordability: IAF ⬍ 0.66; Risk: 0.66 ⱕ IAF ⬍ 1; Equilibrium: 1 ⱕ IAF ⬍ 2; and Comfort: IAF ⱖ 2. II Index of available space The crowding index is the tool most widely used in the literature to measure the adequacy for the household of its dwelling’s size. It is calculated as the ratio between the number of household members and the number of rooms. A ratio of more than one person per room is usually defined as overcrowding. A different index has been used for the analysis conducted in this chapter: the index of available space. This has two advantages over the crowding index: first, it is less penalizing for large households because it assigns lower weights to children compared with the adults in calculation of household size, given that the former need less space than the latter; second, it is coherent with the indices constructed for the other dimensions of housing deprivation, where low values denote hardship and increase with comfort. The formula for the index of available space is as follows:

ISD

s a 0.5 b

where: s ⫽ number of rooms (excluding the kitchen); a ⫽ number of adults in the household; and b ⫽ number of children ⬍ 14 years in the household. The index was then divided into the following three classes: Overcrowding: ISD ⬍ 1; Normality: 1 ⱕ ISD ⬍ 3; and Under-utilization: ISD ⱖ 3 III Index of the dwelling’s physical quality The variables set out in Table 6.1 were used to calculate this index. All the variables were coded with the same direction (1 = presence of the service; 2 = absence of the service). For questions phrased in negative terms (from HA016 to HA020), a semantic inversion was performed.

156

Social Vulnerability in Europe

Variables used to calculate the index of the dwelling’s physical quality Name of the variable

ECHP Label Does the dwelling have a bath or shower? Does the dwelling have an indoor flushing toilet? Does the dwelling have hot running water? Does the dwelling have a place to sit outside, e.g. terrace or garden? Is the accommodation too dark/not enough light? Does the accommodation lack adequate heating facilities?7 Does the accommodation have a leaky roof? Does the accommodation have damp walls, floors, foundations, etc? Does the accommodation have rot in window frames or floors?

X1 X2 X3 X4 X5 X6 X7 X8 X9

Name of the ECHP variable HA009 HA010 HA011 HA013 HA016 HA017 HA018 HA019 HA020

Two items were excluded that, although they concern housing quality, gave rise to problems: the question about the presence of a separate kitchen (which has more to do with differences in housing types than with quality of the dwelling) and the question on the presence of central heating (too closely bound up with climatic factors and made redundant by variable HA017). In each region, a different weight was assigned to each item in order to take account of both their differing contributions to housing quality and the broad geographical variability resulting climatic factors and disparities among housing stocks. The weights were calculated as follows: for each item and for each region, the proportion of households declaring the presence of the service was calculated. This value was then compared to the sum of the scores in the region on all the items (⫽100). In formal terms:

wij

p1ij 9

¦ p1ij

100

i 1

where: wij ⫽ weight assigned to variable Xi in region j; and p1ij ⫽ percentage frequency of the modality ‘1’ for variable Xi in region j. The weight assigned to each variable therefore represented its percentage contribution to definition of housing quality in that specific region. For example: weight of WC in Austria = 97,043/853,367 * 100 = 11,372

Housing Deprivation and Vulnerability in Western Europe

weight of outside space in Eastern Germany 17,200

⫽

157

79,284/460,945 * 100 =

In Austria, therefore, the presence of an indoor toilet had an approximately 11 per cent weight in definition of housing quality; in Eastern Germany, where a smaller number of items were collected, outside space had a value of around 17 per cent. The synthetic index of the dwelling’s physical quality for region j was: 9

IQFj

¦X

i

wij

i 1

The index varied from 0 (absence of all items) to 100 (presence of all items). The continuous index IQFj was then divided into three classes. Each household was assigned to one of the classes according to the difference between its score and the average of the scores for the same region: Deprivation: IQFj ⫽ less than two standard deviations from the regional average; Below-average quality: IQFj ⫽ between -2 and 0 standard deviations from the regional average; and Comfort: IQFj ⫽ above the regional average. IV Index of neighbourhood quality The ECHP database has four indicators for this dimension: 1. 2. 3. 4.

RV: noise from neighbours (question HA015A); RE: noise from outside (question HA015B); PA: environmental problems, pollution, dirt (question HA021); and CR: presence of crime or vandalism in the area (question HA022).

Information on these indicators is not collected in Germany and Luxembourg, which therefore had to be excluded from the analysis on this dimension. The variables were coded in two categories (0 = presence of problem; 1 ⫽ absence of problem). Then an additive index that combined the perceptions of noise, pollution and crime was calculated. Since noise was present in two items, these were combined by assigning a weight of 0.5 to each. The formula for the index of neighbourhood quality IQT was:

IQT

RV RE PA CR 2 2

158 Social Vulnerability in Europe

The index of neighbourhood quality varied from 0 (presence of all problems) to 3 (absence of all problems). The index was then divided into three classes: Neighbourhood deprivation: IQT ⱕ 1; Normality: 1 ⬍ IQT ⱕ 2; and Good quality neighbourhood: IQT ⬎ 2.

Notes The present chapter is the result of a joint endeavour by the two authors. However, individual sections may be attributed as follows: the first and seventh sections written by both authors; the third and fourth, as well as the methodological appendix, were written by Pietro Palvarini and the second, fifth and sixth by Emmanuele Pavolini. 1. The theoretical contributions with perhaps greatest currency are the following: the theory of three capitals (Bourdieu, 1979), the concept of relative deprivation (Townsend, 1979), the capabilities approach (Sen, 1985, 1992, 1999), the paradigm of social exclusion (Lenoir, 1974; Paugam, 1996) and the concept of social vulnerability (Castel, 1995). 2. For an analysis of the concept of housing deprivation and its operationalization see the section on the dimensions of housing deprivation. 3. The analysis of the ECHP data concerns 14 countries, excluding Sweden, given that there is a great deal of missing information on various indicators of housing deprivation for that country. 4. The mean Pearson coefficient of correlation among the five indices of housing deprivation was 0.073. 5. As regards the variable relative to affordability, the modality ‘unaffordability’ also incorporated the modality ‘risk’, since this also denotes a problematic situation. 6. For criticisms of the ratio approach see the sub-section on housing affordability. 7. For Germany and Luxembourg and for gaps in the other countries the variable ‘HA012: DOES THE DWELLING HAVE HEATING OR ELECTRIC STORAGE HEATERS?’ was used.

7 Disability and Caregiving: A Step Toward Social Vulnerability? Giuliana Costa and Costanzo Ranci

Introduction In the EU-15 there are around 25 million people who need assistance with the basic activities of daily living: 40 per cent of them are aged between 20 and 60, and 60 per cent are aged over 60 (Grammenos, 2003). These figures include persons with various levels of functional difficulties because of mental or physical disability. According to estimates by Eurostat (Eurostat, 2003c), one third of such persons were in need of substantial help. Serious disability is one of the most critical events that a human being can experience. The loss of the capacity to take care of oneself and the need to rely on the constant help of another person to perform the most basic of everyday activities compel profound redefinitions of material, organizational and symbolic aspects of life. As Anderson and Bury (1988) point out, chronic diseases attack various areas of a person’s functioning, and not only his/her body. Using Amartya Sen’s terminology, dependency (meaning one person’s reliance for constant help from another in order to lead a dignified life) diminishes both the functionings and the capabilities of individuals and families. Severe chronic diseases, disabilities and the forms of dependency that derive from them restrict opportunities for everyday living and reduce people’s capacities in numerous spheres of their lives (Lyons and Sullivan, 1998). Although they do not necessarily give rise to individual and social breakdown, they nevertheless entail a reorganization of material and symbolic resources that requires support from public policies. The thesis of this chapter is that physical and/or mental dependency is a social risk likely to disrupt the organizational routine of households and to decrease the life chances of people. However, dependency is only one among several risk factors, for dependency may have very different impacts according to the type of household, the economic and social resources on which people draw and the generosity of welfare programmes designed to protect dependent persons. The first part of this chapter will deal with these aspects. 159

160

Social Vulnerability in Europe

The right to be cared has been called an ‘incomplete social right’ (Leira, 1999) because it is rarely granted as an enforceable right. In countries where care provision is a right, it mainly takes the form of an entitlement to economic benefits paid to the caregiver or the care recipient. The impact of welfare policies on the vulnerability of the dependent person must therefore be carefully evaluated, taking both the quantity and types of benefits into consideration. When dependency is severe and of long duration, it involves not only the dependent person, but also the persons who care for him or her, primarily family members (both cohabitants and otherwise). As a large body of research has shown (Österle, 2001; EUROFAMCARE Consortium, 2006), albeit with significant differences among countries, care is still – at least in Europe – a ‘family matter’, requiring a reorganization that sometimes involves several households. Care was for a long time confined to the sphere of intimacy and of private solidarity. Only recently, with the explosive growth of the elderly population, has it moved into the public domain. The recent focus on informal care has been the result of, in fact, demographic forecasts on support for the elderly and the trend in dependency rates. ‘Who will care?’ was the title of a research report published more than 10 years ago. It was accompanied by a subtitle significant in how it positioned the entire issue: ‘Future Prospects for Family Care of Older People in the European Union’ (European Foundation for the Improvement of Living and Working Conditions, 1995). The problem of how, and with what resources, communities can deal with the increase in the dependent population has arisen as a consequence of two processes: the extraordinary increase in life expectancy and the progressive demise of the family as the primary caring institution for the most fragile. With regard to the former phenomenon, it should be remembered that – because the increased life expectancy in the EU-15 countries has been accompanied by greater ‘hope for a life without disabilities’ – different scenarios have been put forward on the link between ageing and increased dependency. The majority of researchers agree that there will be ‘more sick but less disabled’ (Freedman and Martin, 2000), and that the care needs of the over-85s will increase (both in absolute value and as a percentage of the total population), this being the age group with the highest disability rate and long-term needs (Jacobzone et al., 1999). The second part of the chapter will therefore examine whether and in what circumstances informal care is not only a robust element in support of the dependent but also an activity socially sustainable by their caregivers. The main factors responsible for the decline of informal care have been women’s increased labour-market participation, a markedly gendered distribution of care responsibilities and the progressive raising of the retirement age because of economic-financial difficulties of pensions systems. Two of the goals expressly set by the Lisbon Agenda in 2000 – increased rates of activity by women and the population aged between 55 and 64– presage a

Disability and Caregiving 161

decrease in the reservoir of care provided by middle-aged women forming the so-called ‘sandwich generation’ caught among their manifold care roles toward children, grandchildren and the elderly. Accordingly, we wonder whether dependency compromises the ‘normal’ functionings whereby people self-determine their lives (see Chapter 1). Vulnerability as a result of dependency will be seen in the fact that many families with dependents, because of their care responsibilities and relative costs, experience a marked decline in their living standards. The analysis will maintain a dual focus on dependent care recipients and their caregivers. The comparative analysis will rely largely on the distinction drawn by Anttonen and Sipilä (1996) among different care regimes. These authors identified five care regimes in Western Europe: a Scandinavian model, two continental models (one Francophone and the other German-speaking), an Anglo-Saxon model and a Mediterranean one.1

The notion of dependency and its working definition in the ECHP survey Disability and care needs are not adequately represented in national statistical systems. Information on these aspects is collected on the basis of logical schemes, sampling techniques and variables selected in accordance with rather diversified objectives (Fujiura and Rutkowski-Kmitta, 2001). Defining what is meant by ‘dependent person’ is anything but straightforward and free of ambiguity. The operational definition of ‘disability’ used in this chapter, on the basis of ECHP data, raises several problems and entails choices which should be explained in detail. Before providing a working definition of dependency – inevitably conditioned by the limited data available – we shall discuss the concepts correlated with it. The concept of disability has been subject to constant discussion on its definition with many controversies (Barnes, Mercer and Shakespeare, 1999). Disability has been defined by the WHO International Classification of Impairments, Disabilities and Handicaps (ICIDH) – which was used until 2001 – as ‘any restriction or lack of ability (resulting from an impairment) to perform an activity in the manner or within the range considered normal for a human being’ (1980). Whilse impairment is the physical damage incurred by a person through illness or accident, disability is the incapacity to perform the normal activities of daily life consequent on impairment. Handicap is instead defined as the social disadvantage caused by having a disability. Although this classification has been superseded, it still informs the majority of the instruments used by national statistics offices. 2 Yet it still leaves the concept itself of disability undefined. Three main new interpretative frames have recently been introduced in this discussion. The first, based on a medical perspective, connects disability with the presence of an illness or impairment. The second ties disability to

162

Social Vulnerability in Europe

functional limitations in performing everyday activities. The third takes an environmental perspective and regards disability as the result of interaction between the individual (with his/her specific characteristics) and the social and physical context (Burchardt, 2004; Fujiura and RutkowskiKmitta, 2001). The second frame most frequently informs European statistical surveys. According to it, limitations in the performance of everyday activities (walking, seeing, listening and speaking) determine a more or less pronounced level of disability. The majority of studies use batteries of questions relating to two indices: one developed by Katz, the other by Lawton and Brody.3 These indices are respectively based on two typologies of activity: ADL (Activities of Daily Living) and IADL (Instrumental Activities of Daily Living). ADL and IADL are often combined, but they yield definitions of disability that vary considerably according to the criteria used to establish cut-off thresholds in the levels of impairment that is considered disabling. The international literature shows a clear connection between the presence of substantial health problems and problems of severe disability.4 Data on disability are generally constructed by verifying the presence of physical or mental anomalies, chronic illnesses or disorders and, consequently, the presence or absence of limitations on the ability to perform everyday activities, or a loss of functions. This is also the approach taken by the ECHP, which first asks ‘do you have any chronic physical activity or mental health problem, illness or disability?’ It then asks whether, because of the chronic condition, the respondent is ‘hampered in daily activities’. This filter question is intended to establish the presence of a persistent disability or illness, with the purpose of distinguishing long-term problems from acute ones, which may be severe and limiting but are contingent or with rapid remission. Moreover, the reference to the severity of impediments to daily life (the possible answers to the specific question in the ECHP are ‘no’, ‘to a certain extent’ and ‘severely’) comes very close to the phenomenon of serious disability. In 2001, 21.4 per cent of respondents defined themselves as hampered to a certain extent in the performance of daily activities, while only 8.4 per cent regarded themselves as severely hampered. These values are much more congruent with national data collected by surveys conducted specifically to measure states of health and disability (Eurostat 2003c; Lafortune et al., 2007). However, this definition of dependency, too, requires caution because not all severe limitations in daily life necessarily give rise to a need for constant help and care. The notion of dependency in this context comprises some specific types of limitation: long-term dependency with a substantial need for help in terms of intensity and/or frequency. In the absence of specific information on the need for help, in what follows below dependency is defined on the basis of a twofold criterion: those defined as dependent are persons who, besides being ‘severely hampered’, also declare that they suffer from bad health (those replying ‘very bad’ or ‘bad’ to the question ‘how is your health in general?’). For these persons, the

Disability and Caregiving 163

perception of the severity of their activity limitation is heightened by the fact that they do not enjoy good health. Here, therefore, dependency is construed as a long-term condition connected to problems of illness or chronic disability, which restrict the performance of daily activities and have a negative impact on the person’s perception of his/her state of health. It is therefore tied to three conditions: chronicity, disability and poor health.

The characteristics of the dependent population According to ECHP data for 2001, on the basis of the working definition given above, 5.7 per cent of the population of the EU-15 countries were dependent, 3 per cent if related to the adult population and 13.3 per cent in relation to the population aged over 65. Excluded from this estimate are persons in residential facilities – an important component of the population definable as ‘dependent’.5 Overall, the sample of the dependent population comprises 5730 individuals, divided in half between adults and elderly persons. The distribution of this sample in the 14 countries for which figures are available (with the exception of Luxembourg) is uneven because of various factors, such as different linguistic and cultural interpretations of the terms used in the survey, the different demographic compositions of populations and differing perceptions of dependency as a public issue and therefore a declarable matter. In all countries the dependency rate increases with age, and there are more dependent women than men. The joint effect of these two features is the presence of a large component – approximately one third of all dependent persons – consisting of elderly women. Overall, therefore, the dependent population is characterized by a high degree of old age and feminization.

Profiles of vulnerability connected with dependency Dependency is a factor that indubitably places those who experience it in a situation of vulnerability. However, adequate assessment of this effect requires consideration of the situation of the whole household in which the dependent person lives. In Europe, almost one household in every 10 (9.4 per cent) has a dependent member, 5.4 per cent in the case of households with an adult dependent and 18 per cent in case of households with a dependent elderly person (see Table 7.1). These are extremely high proportions, and this reflects the effects of the general ageing of the population. The share of households with at least one dependent person is greater in Southern Europe (11.5 per cent) and substantially lower in the Scandinavian area (6.3 per cent), with intermediate levels in continental countries (9.9 per cent in the German-speaking area and 8.4 per cent in the Francophone area) and in the Anglo-Saxon area (8.6 per cent). These differences mainly result from the fact that dependent persons are more frequently placed in residential facilities in the countries of North Europe.

164

Social Vulnerability in Europe

Table 7.1

Percentage of households with a dependent member by care regime

Care regimes

% of households with 1+ adult dependent person (aged under 65)

% of households with 1+ dependent elderly person (aged over 65)

% of households with 1+ dependent person (total)

Scandinavian area Continental German-speaking area Continental Francophone area Anglo-Saxon area Mediterranean area

3.9 6.5

13.8 18.3

6.3 9.9

4.8 6.6 5.5

16.8 13.1 21.5

8.4 8.6 11.5

Total

5.4

18.0

9.4

Source: Our calculation on ECHP database.

The compression of living standards Families with dependent members often experience a ‘compression’ of their living standards due to two main factors. The first is the reduced capacity to work of dependent individuals. The activity rate among dependent adults is only 30 per cent (in the rest of the adult population it is 74 per cent). Their share of part-time work (9.5 versus 4.3 per cent) and of unemployment rate (9.8 versus 69 per cent) is much higher than in the rest of the population. The loss of work coincides with the onset of the state of dependence: the share of dependent people who have never worked (11 per cent) is in fact largely identical to that of the non-dependent population. Scandinavian and continental German-speaking countries offer greater employment opportunities to adult dependent persons than do the other regimes. In these areas, in fact, the activity rate rises to 36–37 per cent, compared with 27 per cent in the Francophone continental area, 26 per cent in the Anglo-Saxon one and 23 per cent in Mediterranean countries. These last are also the countries that offer fewer work opportunities for adult dependents. The work difficulties of dependent persons significantly affect both the organization of their families (the household members who assume caregiving responsibilities must reconcile paid work with care by accepting jobs with reduced hours or low wages), and the household’s overall income. Moreover, in countries where welfare benefits are linked to the contributive capacity of citizens, the low activity rate of dependents and their family members, or the presence of family members with low wages, also reduces welfare payments. Therefore, lost earnings resulting from dependence are scarcely compensated by welfare measures targeted on persons with diminished work capacity or on their caregivers.

Disability and Caregiving 165 100 95 90 85

Families with 1 + dependent adult Families with 1 + elderly dependent

80 75 70 65 60 Scandinavian Anglo-Saxon Francophone Mediterranean area area area area

Germanic area

Figure 7.1 Average income of families with 1+ dependent person / average income of families without dependents, EU-15 Note: The income considered is the equivalent household income (modified-OECD scale). Source: ECHP, authors’ own calculations.

The second compression factor is related to the fact that dependency entails increased expenditure on routine activities (using public transport, shopping, home cleaning, preparing food, etc.), which are more costly for a disabled person, and for the care activities made necessary by the state of dependency (help with physical mobility and care of the person, rehabilitation, etc.).6 The first step in our analysis concerns the income differentials between the dependent and non-dependent populations. The average per capita income of households with at least one dependent member is 20 per cent less than the income of households without dependents (see Figure 7.1). The gap diminishes slightly when the dependent person is aged over 65 (16.5 per cent). In Mediterranean countries the differential increases substantially (reaching 30 per cent for adults and 21 per cent for the elderly), while it is markedly smaller in Scandinavian countries (around 15 per cent for both adults and elderly persons) and in the German-speaking area (13 per cent for adults and only 5 per cent for elderly dependents). Income differentials for families with adult dependents increase sizeably in the areas in which the activity rate of dependent people is lower. The possibility of finding and keeping a job is therefore an important factor in the defence against the economic vulnerability caused by dependency. The elderly population exhibits – except in the Mediterranean area – smaller income differentials, owing to the presence of generous income support programmes. These wide income differentials to the detriment of the dependent population have several consequences. The first is the greater exposure of these

166

Social Vulnerability in Europe

households to the risk of poverty. In general, dependency determines a 30 per cent increase in a household’s poverty risk. The probability of being poor substantially increases in the continental Francophone area (by 60 per cent), in Mediterranean countries (by 37 per cent) and in the AngloSaxon area (by 30 per cent). But it does not increase in the German-speaking area, and indeed substantially diminishes (–20 per cent) in Scandinavian countries. As we shall see, this variation depends closely on the extent and generosity of public welfare programmes. Nonetheless, the incidence of poverty among households with dependents is fairly low, given that it affects 13 per cent of European households. Much more frequent, instead, are less severe situations of material compression, such as being unable to make any savings or having to cut back consumption. More than a factor of impoverishment, therefore, dependency is a vulnerability factor that decreases economic resources, restricts possible options, reduces protection against other negative events and compresses the overall standard of living. The ECHP data show that households with a dependent member are notably financially compressed (see Table 7.2). Some 70 per cent of these households declare that they are unable to save any money, a 30 per cent increase compared with households without dependents. Moreover, 50 per cent of them report problems of affordability regarding two or more of the following goods and services: a week of holiday away from home, replacing broken household appliances, eating meat, chicken and fish every second day, entertaining friends or family for a meal at least once a month. The proportion is twice as high in households with dependents as in households without dependents. Taken as a whole, these data suggest that dependency, in mature welfare systems, strongly reduces the use of economic resources, while the poverty risk affects a relatively small proportion of the dependent population. In the Scandinavian area and in German-speaking countries the economic compression of households with dependents is less negative than in other areas. One of the factors responsible for the economic compression of households with dependent members is that they incur additional costs in their everyday activities. An indirect way to assess their need for additional income is to consider income differentials solely for households which declare that they can easily ‘make ends meet’. These represent only 11 per cent of all European households with at least one dependent member, while they represent 22 per cent of households without dependents. The territorial distribution is markedly uneven: such households exceed 20 per cent in the Scandinavian and German-speaking area, but amount to only 4 per cent in the Mediterranean area (see Table 7.3). It can be presumed that the average income of these ‘satisfied’ households indicates the average economic threshold above which households with dependents are able to meet all their expenses, including those relative to care needs and the transport of dependents. Unfortunately, the ECHP data

Table 7.2 Levels of economic compression in households with or without dependent members by care regime, percentage of the total of households with a dependent Cutbacks in consumption* (cannot afford at least two of the following: holiday, furniture replacement, meat or meals with friends)

Impossibility to make savings* Households without dependents Scandinavian area Continental German-speaking area Continental Francophone area Anglo-Saxon area Mediterranean area Total Note: *

Households with dependents

Households Difference without dependents

Households with dependents

Difference

50.1 33.2

61.3 46.6

+ 11.2 13.4

14.4 15.3

29.6 31.5

+ 15.2 + 16.2

56.1

71.8

15.7

15.3

37.1

+ 21.8

47.1 71.5

64.4 85.6

+ 17.3 + 14.1

18.9 41.4

41.3 69.0

+ 22.4 + 27.6

54.2

69.9

+ 15.7

25.1

49.4

+ 24.3

Data for Sweden missing.

Source: ECHP, authors’ own calculations.

168

Social Vulnerability in Europe

on these ‘satisfied’ households exclude important countries like Germany and Sweden from the analysis.7 Nevertheless, they furnish some interesting insights (see Table 7.3). With Germany and Sweden excluded from the analysis, the equivalent annual average income of these ‘satisfied’ households is around 18,000 euros. Compared with the average income of all households with dependents, the percentage increase required to satisfy all material needs is 70 per cent (for households without dependents, the increase is instead 46 per cent). It increases enormously in Mediterranean countries (where income must double for all material needs to be easily satisfied), while it is very low in Scandinavian countries (only 14 per cent). In the latter countries the abundant availability of services and forms of public support therefore brings a net improvement to the quality of life of low-income dependent persons. In absolute terms, a household with a dependent member in the Mediterranean area needs additional income three times higher than that required by a Scandinavian household. To sum up, dependency strongly squeezes the material circumstances of households, and this compression in its turn drastically reduces their living standards. More than increasing the poverty risk, dependency produces a general situation of economic vulnerability generated by the decrease

Table 7.3 Characteristics of households with dependent members easily able to satisfy their material needs, by care regime

Average Percentage annual income share (in €) Scandinavian area Continental Germanspeaking area Continental Francophone area Anglo-Saxon area Mediterranean area Total

Percentage Average increase in income increase income with necessary to respect to the satisfy material average income needs ‘easily’ of all households (in €) with dependents

21.9

15,960

2006

14

24.5

18,625

4502

32

13.6

18,996

5254

38

17.4

22,071

6688

43

4.0

13,796

6604

92

11.1

18,014

7372

69

Source: ECHP, authors’ own calculations.

Disability and Caregiving 169

in income consequent on the reduced activity rate of dependent persons and the higher expenses incurred by their families. Only a minority of households have sufficiently high incomes to be able to afford specific care services and maintain a satisfactory standard of living. However, marked differences are apparent in the economic conditions of households with dependents in different care regimes. The next section therefore examines the specific role performed by policies to reduce the economic vulnerability of dependent people. The role of policies in reducing economic compression The income level of the dependent population is strongly determined, in all the EU-15 countries, by public monetary transfers. What, therefore, is the role of welfare benefits in reducing the risk of economic deprivation for households with dependents? In this section we concentrate on two questions: the coverage of support programmes, and their ability to reduce poverty risks and attenuate economic compression. The focus will be on disability/sickness benefits (DSB): allowances implemented and paid on the basis of eligibility criteria that differ greatly across the European countries. Two types of benefits can be distinguished: benefits intended to make up for the lack of income resulting from dependency and those provided to pay for care services. The former are disability pensions, highly standardized in European countries and compensating for a continuous inability to work. The latter are care allowances paid in order to support the direct provision or the purchase of daily home assistance.8 Unfortunately, the information available from the ECHP survey cannot be used to distinguish between the two types of benefits, so these will be considered jointly here. Figure 7.2 shows that almost half of households with a dependent adult receive some benefits.9 The coverage is wide in the Scandinavian area and Anglo-Saxon countries, and narrower in Southern Europe and in the German-speaking area. As regards the generosity of benefits, their amounts (weighted by the purchase power parity) are particularly high in the German-speaking area and more modest in the Francophone and AngloSaxon ones. In Mediterranean countries the amount of benefits is close to the European average. While Germany, therefore, has a model of economic support based on generous but strongly selective programmes, the AngloSaxon area exhibits the reverse model based on the broad coverage of less generous measures. In Scandinavian countries ample coverage of measures combines with high amounts, while the Francophone area has low amounts and a limited coverage of programmes. Now we will look at the effect of DSBs on the poverty risk and on the compression of the living standards of households with dependent members. Altogether, DSBs represent 36 per cent of the total income of households with dependents in the EU-15 countries. This percentage does not change substantially across the areas considered. Benefits therefore make a

170 Social Vulnerability in Europe 70

10,000 9,000

60

8,000 50

7,000 6,000

40

5,000 30

4,000 3,000

20

Percentage of families with dependent adults Average amount of DSB (in euro)

2,000 10

1,000 Mediterranean area

Anglo-Saxon area

Continental Francophone area

Continental German-speaking area

0 Scandinavian area

0

Figure 7.2 Percentage of families with dependent adult receiving DSBs and average amount of DSBs in the EU-15 Source: ECHP, authors’ own calculations.

significant contribution to the incomes of these families. The role of these measures in countering poverty is also significant, as shown by Figure 7.3. Around 45 per cent of households with dependents whose incomes would fall below the national poverty thresholds are able to rise above them because they receive a DSB. The share is particularly high in Scandinavian countries, where the average amount of DSBs is larger: around two thirds of potentially poor households lie above the poverty threshold because they receive a DSB. In Southern Europe and in the Francophone area the capacity of DSBs to combat poverty is weakened by their lower amounts: in these areas the households that avert poverty thanks to DSBs are respectively 31 and 45 per cent. The effect of DSBs on the degree of household economic compression is less clear. In general, the distribution of DSBs is unable to reduce to any significant degree the proportion of households finding it difficult to satisfy their material needs. Receipt of DSBs reduces such households by 3 per cent overall and by 8 per cent in the countries of the Anglo-Saxon and Germanspeaking areas. Households with dependent members are therefore strongly exposed to economic vulnerability. Only a very small proportion of these households have a standard of living enabling them to satisfy all their essential material needs. For the majority of households, instead, the presence of a dependent member significantly reduces their living standards. Welfare

Disability and Caregiving 171 40 Poor households 35

Poor households if without DSBs

30 18.5

25

7.7

20

9.9

9.9 15 11.9

9.1

10

16.4

17.2 12.7

6.2

7.4 Continental Germanspeaking area

12.4

Scandinavian area

5

EU 15

Mediterranean area

Anglo-Saxon area

Continental Francophone area

0

Figure 7.3 Effect of DSBs on the poverty risk of families with dependent people – 2001 Source: ECHP, authors’ own calculations.

benefits nevertheless reduce significantly the poverty risk for these families, even if they seem insufficient to prevent dependency from being associated frequently with some degree of material deprivation. Within this general picture, marked differences emerge among care regimes. In the Scandinavian area, the activity rate of adult dependent people is higher. This gives rise to better income levels and living standards compared with the rest of Europe; in additionDSBs act as significant shock absorbers by protecting a large proportion of households against the risk of poverty (but without significantly weakening the economic compression). The situation is similar in the German-speaking area, where the income level of households with dependents is particularly high, and DSBs perform a significant redistributive function. The Anglo-Saxon area exhibits an analogous pattern. The situation is more difficult in the Francophone and Mediterranean areas: here households with dependents suffer from a higher income differential and from a tighter economic compression, which is only slightly lessened by lower and less widely available state income support.

The support provided by families We have already said that the loss of self-sufficiency is a ‘family matter’ that involves not only dependent persons but also their family networks. Throughout Western Europe caring for dependents is entrusted primarily to

172

Social Vulnerability in Europe

family solidarity (Attias-Donfut et al., 2005; Österle, 2001; Alber and Köhler, 2004; Eurobarometer, 2007; Hantrais, 2004). The persistence of the central role played by family in care delivery has been explained in two different ways. In the first interpretation care is founded on norms regulating the relationships between individuals and their families (Finch, 1989). In the second interpretation the persistence of family care responsibility depends mostly on the architecture of social protection systems (Leira, 1999; Leira and Saraceno, 2002). According to this second view, disparities in the levels of informal care reflect differences among care regimes in their provision of residential facilities, the level and extension of disability pensions and delivery of home care services. The impact of dependency therefore varies substantially according to the type of family and to the quantity and type of support that the family is able to provide. State support remaining equal, whether or not dependent persons have families able to look after them largely determines the quality of their life. Forms of cohabitation Unfortunately, the ECHP data cannot be used to reconstruct relations between care recipients and caregivers in detail. The analysis must therefore be restricted to the types of family structure in which dependent people live, without consideration of the support structure among non-cohabiting family members (Attias-Donfut, 1995; Cioni, 1999). Dependent people more frequently live with other relatives than do nondependent people of the same age. Compared with families without dependents, the proportion of families with children under 18 years diminishes (from 31.9 to 16.6 per cent), while the proportions of couples (from 23.1 to 31.7 per cent) and families with adult children (from 16.8 to 23.8 per cent) increase. The percentage of dependent persons living alone remains largely the same (around 28 per cent). Marked differences emerge according to the care regime considered (see Table 7.4). In Scandinavian countries single dependent people form the overwhelming majority – exceeding 50 per cent among households with elderly dependents – while cohabitations are marginal. In Mediterranean countries and in the Francophone area, by contrast, the proportion of single dependents is small, while cohabitation accounts for the majority of household forms: this is indicative of the resilience of family solidarity in those countries. In the Germanic and Anglo-Saxon areas the model is more individualized and resembles the Scandinavian pattern. Table 7.4 shows that loneliness is very widespread among the dependent elderly. One third of elderly dependent people, in fact, live alone. Loneliness increases up to 42 per cent among individuals aged over 75 years. The absence of cohabitant family members compels care

Disability and Caregiving 173

Table 7.4 Share of families with dependents consisting of single individuals or of cohabiting adults of different generations, by care regime – 2001 (row percentages) Families with dependent adults

Families with elderly dependents

Singles

Cohabitation between parents and adult children

Singles

Cohabitation between parents and adult children

Scandinavian area German-speaking area Francophone area Anglo-Saxon area Mediterranean area

37.9 26.5 12.1 18.1 9.0

11.5 18.7 30.0 15.4 45.8

54.0 42.4 32.6 46.8 21.5

7.1 14.2 19.3 12.1 31.3

Total

20.0

26.8

33.0

21.9

Care regimes

Source: ECHP, authors’ own calculations.

arrangements to be made by resorting to external resources, both informal and formal. Indeed, 74 per cent of dependent single persons have daily contact with other people, a frequency compatible with receiving significant help from non-cohabitants. The remaining 26 per cent declare that they meet other people once a month (14 per cent) or less than once a month/never (12 per cent). Loneliness has a striking impact on the life quality of families with dependents (see Table 7.5). Dependent persons living alone are much more frequently poor than are dependents living with a partner or with adult children, and they experience a slightly stronger compression of their living standards. Moreover, they have less frequent access to DSBs. Families with dependents in the household and children under 18 years, are in an equally critical situation because the financial load resulting from disability tends to grow alongside the costs of maintaining the children. Cohabitation between parents and adult children mainly reduces the poverty risk. Finally, it needs to be stressed that most of the care problems of dependent persons arise at the point of intersection between forms of family cohabitation and state support. In an institutional context with abundant services the quality of life may continue to be relatively high even though family lacks. In Southern Europe, on the contrary, where the endowment of public care services is lower, the living arrangements of dependent persons are more frequently based on cohabitation among different generations of adults, a solution that appears functional as regards both provision of care and the avoidance of excessive material compression.

174

Social Vulnerability in Europe

Table 7.5 Effects of the family situations of dependent persons on poverty, restrictions on consumption and on access to DSBs – 2001

Family situation

% poor families

% families with severe % families restrictions on consumption receiving DSBs

Single Couple Cohabiting parents and adult children Couples with children aged ⬍ 18

15.0 9.8 9.4

55.4 41.0 51.9

19.2 30.3 40.5

18.6

52.0

42.4

Total

12.6

49.3

31.6

Source: ECHP, authors’ own calculations.

Who are the caregivers? The ECHP survey identifies caregiving to dependent people by asking interviewees whether their daily activities include giving free care to a person requiring help because of age, illness or disability. This care activity is then specified with information about the weekly number of hours devoted to it. In the analysis that follows the persons who assume the caregiver role are defined by the fact that they devote at least 4 hours a week to it. This definition is assumed so as to concentrate on quite intense and continuous caregiving activity. The care of children is not analysed. Using this definition, the overall caregiving rate in Europe is 3.8 per cent (Sweden and Luxembourg are not considered because of lack of data). It decreases slightly among adults (3.5 per cent) and instead increases among people aged over 65 years (5.1 per cent). The range of variation among countries is between a minimum of 2.8 per cent for Germany and a maximum of 6.5 per cent for Finland. The only countries in which caregiving is more frequent among adults than among the elderly are Italy and Ireland. As expected, a marked asymmetry between men and women in the assumption of care tasks is found: the incidence of caregiving among women is twice that among men (5.2 against 2.4 per cent). This asymmetry is even greater in the Mediterranean area: in Portugal, Greece and Spain the incidence of caregiving by women is three times higher than that by men. While informal caregiving is a crucial resource for dependent people, it nevertheless has problematic features that may generate specific forms of social vulnerability. In order to specify the exposure to vulnerability of caregivers, we first consider their socio-demographic characteristics. Caregivers are aged 55 on average, with some variations between geographical areas. The central age band (40–64 years) comprises 58 per cent of caregivers, while caregivers aged 65 and over represent 27 per cent of the total. More than two thirds of

Disability and Caregiving 175 100 90 80 Axis title (%)

70 Elderly female Female adults Elderly males Male adults

60 50 40 30 20 10 0 Dependent people

Caregivers

Figure 7.4 Gender and age composition of dependent persons and of caregivers – 2001 Source: ECHP, authors’ own calculations.

caregivers (70 per cent) are female. However, the proportion of women varies according to the geographical area: it is high in Southern Europe (where it reaches 78 per cent) and lower in the German-speaking area (61 per cent). Moreover, 20 per cent of caregivers are adult men, 10 per cent are elderly men, 53 per cent are adult women and 17 per cent are elderly women. The largest group comprises therefore women belonging to the so-called ‘sandwich generation’. This group accounts for around 45 per cent of female caregivers in the various European areas and reaches 64 per cent in Mediterranean countries. Comparing the structure by age and sex of caregivers with the demographic composition of dependent persons highlights the relationship between generations and genders as regards care arrangements (see Figure 7.4). The large elderly component in the dependent population (52 per cent) is matched by the large proportion of female caregivers of adult age (over 53 per cent). The care relationship therefore redistributes time and help to the advantage of the elderly male population; and the care burden is mainly assumed by adult women. Caregiving is therefore characterized by a high degree of feminization and it is delivered mainly by people in central ages likely to be active in the labour market. Moreover, caregiving by adults is often (in 50 per cent of cases) organized within complex families comprising adults belonging to different generations. Predominant among elderly caregivers are people who live in couples (53 per cent), although the proportion of single caregivers (20 per cent) is substantial.

Care activities exposing to vulnerability Family caregiving has long been neglected by policy-makers, especially in the countries of Southern Europe. This neglect has been, paradoxically, a result

176

Social Vulnerability in Europe

of the large-scale delegation of care responsibilities to families, which has contributed to confine the issue of care in the private sphere. A contributory factor is that caregiving is often undertaken at an advanced adult age, and therefore at a stage of life when serious problems of compatibility between caregiving and work do not arise. Consequently, the impact of caregiving on the life quality of caregivers has only recently been studied, superseding research that focused almost exclusively on the link between caregiving and psycho-physical stress (Lazarus and Folkman, 1984). Gender studies have analysed the significance assumed by caregiving in gender contracts, the aim being to de-naturalize the allocation of care tasks solely to women (see for example Graham, 1983; Ungerson, 1983; Land and Rose, 1985; Lewis and Meredith, 1988; Lewis, 1993; Schulz and Beach, 1999; Thomas, 1993). Therefore less attention has been paid so far to the negative social impact of caregiving for caregivers themselves and their families. Considered here is the extent to which caregiving may cause vulnerability. A crucial aspect in this regard is the amount of time invested in carework. A hypothesis put forward in studies on caregiving is that very intensive carework makes it more difficult to reconcile the constraints and opportunities connected with other social activities. The lives of intensive caregivers are ‘colonized’ by their carework, and this frequently leads to their social isolation (Costa, 2004, 2007; Twigg, 1993). According to the ECHP data, caregiving is provided for an average of 23 hours a week, although the median value is 15 hours. Some 43 per cent of caregivers look after their dependents for a maximum of 14 hours a week (an average of 2 hours a day for 7 days); but 29 per cent have very intensive care responsibilities which occupy them for more than 28 hours a week (an average of 4 hours for 7 days). The intensity of caregiving varies markedly according to gender: the average weekly amount of care time for women (25 hours) is one third more than the amount for men (19 hours). Moreover, 33 per cent of female caregivers look after their dependents intensively (more than 28 hours a week), while the same figure for men is only just over half as much (18 per cent). The intensity of caregiving also differs greatly according to the geographical area: it is substantially lower in the Anglo-Saxon countries (where average care time is 13 weekly hours) and much greater in the Mediterranean area (where the average amounts to 30 hours a week). In all, 42 per cent of caregivers in Mediterranean countries devote at least 28 hours a week to their carework, while the figure for the rest of Europe is around 20 per cent, and less than 10 per cent for the Anglo-Saxon countries.10 We consider three impacts of caregiving: (a) on work, (b) on income and consumption and (c) on social relationships and sociality. The analysis will compare caregivers and non-caregivers and, with regard to the former, compare those who provide care intensively with medium caregivers. Differences between the EU-15 countries will also be examined.

Disability and Caregiving 177 Table 7.6

Percentage of caregivers with basic education by age – 2001 Less than second stage of secondary education (ISCED 0–2)

Non-caregivers Age 16–39 40–64 65–74 Over 74

40.5 55.0 75.2 80.4

Non-intensive caregivers 42.7 50.5 65.6 70.5

Medium caregivers 57.7 69.5 84.2 84.1

Intensive caregivers 60.4 76.9 88.5 85.0

Source: ECHP, authors’ own calculations.

Before analysing these aspects, we would point out that the education levels of medium or intensive caregivers are considerably lower than the education level of the average population of the same age, as well as that of non-intensive caregivers (see Table 7.6). One explication is that scant social and cultural value is placed on intensive caregiving: it is hardly comparable with a highly skilled professional job and is therefore undertaken by people with modest education. A second explanation points to the conditions in which caregiving becomes competitive with respect to work in the market. When the work income is low because of poor qualifications, caregiving is a rational response. Caregiving and work Our analysis starts with the nexus between caregiving and work. This has been an issue in the European policy agenda since the Lisbon Summit, because of widespread concerns that increasing the activity rates of women and people aged between 55 and 64 will shrink the reservoir of help so far almost unconditionally delivered by middle-aged women (Lewis, 2005). The factors influencing decisions about work and caregiving are not only personal moral convictions but also the characteristics of the labour market and of the welfare system. The ECHP data can be used to verify to what extent caregiving is a deterrent against entering the labour market (see Table 7.7). Indeed, the activity rate of working-age caregivers significantly and progressively decreases with the increase in the time devoted to caregiving. For intensive caregivers the activity rate falls by 40 per cent among men aged over 40 years (from 82 to 50 per cent) and by 40–50 per cent for women. For caregivers who devote fewer than 14 weekly hours to care, instead, the impact of caregiving on the activity rate is almost zero, for both genders and all age groups. Light caregiving therefore appears to be compatible with working, while the trade-off between work and care visibly increases as care time extends beyond 14 hours a week.

178 Social Vulnerability in Europe

Table 7.7

Activity rate of caregivers by age and gender – 2001 Activity rate

Non-caregivers

Light caregivers

Medium caregivers

Intensive caregivers

Men 16–39 40–64

81.0 81.7

86.8 79.6

77.1 64.6

86.3 50.0

Women 16–39 40–64

65.7 55.7

64.0 56.3

57.5 48.4

38.4 28.8

Total 16–39 40–64

73.4 68.7

72.4 64.6

49.4 41.8

45.1 31.8

Source: ECHP, authors’ own calculations.

However, this does not clarify whether it is caregiving that discourages paid employment, or whether caregiving is most frequent among people who did not work prior to assuming their care responsibilities. Sarasa (2008) has shown that eligibility criteria for welfare benefits have the greatest impact on the occupational behaviours of adult female caregivers. Two findings from the ECHP data are useful in answering this question. First, contrary to the idea that caregiving mainly involves people who have never been in the labour market, the great majority of caregivers inactive at the time of the ECHP survey (including intensive caregivers) had previously worked (around 70 per cent of the women and 80 per cent of the men). Almost all caregivers, therefore, have been in paid employment, even if for many of them it is not simultaneous with caregiving. Second, almost one third of currently inactive female caregivers consider caregiving to be a factor that prevents them from working (among the women with intensive care duties, the value rises to 40 per cent). A quarter of the women simultaneously working and providing care report that they are hampered in the labour market by their caregiving obligations, while the same perception is not present among male caregivers. We may therefore conclude that caregiving has a discouraging effect toward paid work mainly for women, and particularly for intensive caregivers. This effect is less marked in Scandinavian and Francophone countries, while it is more pronounced in the Mediterranean and Anglo-Saxon areas. Caregiving and household income The relationship between caregiving and income may assume various features. It is generally assumed that, because caregiving prevents or restricts

Disability and Caregiving 179

access to the labour market, it diminishes the household’s standard of living. On the other hand, it can be supposed that this activity is undertaken by people who, as beneficiaries of pensions or care allowances, have sufficient income to satisfy their material needs. Moreover, the caregiver may be part of a household whose income is assured by other members who work or receive high pensions. The impact of caregiving on income varies substantially according to the intensity of the caregiving, and to gender (see Table 7.8). Caregivers who devote fewer than 14 hours a week to looking after a dependent generally have personal incomes significantly above the average, whether or not they are in paid employment. The incomes of male caregivers are always above the average, while the incomes of women dramatically decrease as the intensity of caregiving increases. The average income of women intensive caregivers decreases by 40 per cent if they work and by 37 per cent if they do not work. The impact of caregiving on household incomes displays the same pattern. In families with light caregivers, the equivalent per capita income is 8 per cent higher than the general average; it decreases by 16 per cent if the caregiver engages in carework to a medium extent; and it reduces by one third (33 per cent) when the caregiver delivers intensive care. In addition indices related to the economic compression of households (see Table 7.8) show that intensive caregiving is a major cause of economic vulnerability, while light caregiving does not have a significant economic impact, and indeed is very often delivered in households with medium-high living standards.

Table 7.8 Economic conditions of families with a dependent by intensiveness of the caregiving – 2001 Without caregivers Average income of families with 1+ dependent person/ average income of families without dependents* Below the poverty threshold Inability to save Strong restrictions on consumption Difficulty in ‘making ends meet’

With light caregivers

With medium caregivers

With intensive caregivers

100.0

107.8

84.0

66.8

10.3

7.2

9.9

12.5

55.4 27.0

53.3 25.5

60.2 33.4

71.8 48.0

18.3

15.5

20.1

32.1

Source: ECHP, authors’ own calculations.

180

Social Vulnerability in Europe

In short, caregiving is a social activity that, if performed for a limited amount of time, does not have significant effects on the income of caregivers, nor on the economic conditions of their families. When care work intensifies and is undertaken by a woman, the income of the caregiver and her family diminishes drastically. Therefore caregiving, in certain circumstances, exposes those undertaking it (and their families) to considerable economic vulnerability. Caregiving and sociality According to a large body of studies (Lyons, Sullivan and Ritvo, 1995; Kittay, 2001), caregiving is very likely to cause social isolation and the loss of meaningful social relations. The relationship between caregivers and care recipients confines the former to an exclusive dyadic relationship that absorbs large part of their time, with a consequent negative impact on their social capital. In this regard, we will verify the extent to which the intensity of caregiving restricts the ability of caregivers to maintain frequent social contacts outside the family circle. We consider three social domains essential for the development of sociality: relations with neighbours, relations with friends and participation in associative activities. Figures in Table 7.9 show that caregiving, even when intensive, does not restrict informal relations with neighbours and friends. Quite surprisingly, informal relations intensify for people looking after dependents, to show the strong sociability that seems to be characteristic of caregivers. Conversely, formal participation in associations is strongly restricted by highly intensive care commitments. Overall, caregiving does not necessarily entail a loss of sociality. Rather, it induces the caregiver to set greater value to informal relations with neighbours and friendships. But caregiving is associated with limited participation in associative activities requiring an involvement that extends beyond primary relationships into Table 7.9 Participation of caregivers in various forms of sociality by intensiveness of caregiving – 2001

Frequent contacts with neighbours Frequent meetings with friends and not-cohabitant relatives Participation to associations

Without caregivers

With light caregivers

44.0

46.6

57.7

63.7

83.0

87.5

81.9

76.5

32.4

43.2

32.0

19.9

Source: ECHP, authors’ own calculations.

With medium caregivers

With intensive caregivers

Disability and Caregiving 181

the public domain: such involvement is difficult to manage when the caregiver has binding and time-consuming care responsibilities. Caregiving as a vulnerability factor It may be concluded that caregiving fits without problems into the organization of the daily lives of people and families only when it is non-intensive. In this case, in fact, caregiving is compatible with an average income, with working and with an active social life. Caregiving becomes a significant factor of vulnerability when it takes up more than 28 hours a week and thus becomes the caregiver’s dominant social commitment. In these cases, the entire organization of the caregiver’s life suffers: it becomes difficult to reconcile care with work, a career is difficult to pursue, income diminishes markedly and the chances of social participation are restricted (although primary relationships remain close). Intense caregiving mainly involves women with low levels of education, and they are subject to multiple risk factors. However, our analysis does not make it possible to clarify the extent to which intensive caregiving is a matter of choice or imposed by the lack of viable alternatives furnished by the labour market or the welfare system. The strong economic and cultural deprivation of many intensive caregivers shows that external constraints are strong and cogent. Comparative analysis among care regimes confirms the strength of these constraints. In general, Scandinavian and German-speaking countries enable easier reconciliation between care and work, compared with Mediterranean countries and the Francophone area. In these last countries the substitution effect between work and care is strong. Here the assumption of heavy care responsibilities is a rational strategy in face of the few opportunities made available by the labour market and the scant generosity of welfare benefits.

Concluding remarks This chapter has argued that dependency (as a specific combination of severe difficulties in performing everyday activities and poor health) is a cause of vulnerability, which compromises the ‘normal’ functionings whereby people determine their own life-courses. The chapter has therefore investigated how dependency destabilizes family arrangements and significantly diminishes the life-chances of their members. In Western Europe, 9.4 per cent of families have a dependent member, 5.4 per cent having at least one adult dependent person, and 18 per cent a dependent elderly person. Dependency is therefore rather widespread in Europe. The two features shared by all countries are that the dependency rate increases with age, and that dependency is more frequent among women than men.

182

Social Vulnerability in Europe

One of the aspects analysed has been the reduction in income. In all the European countries dependency is associated with a decline in the living standards of the dependents and their families. Rather than increasing the risk of poverty, dependency generates a general situation of material compression as a result of increased spending on care and a lower activity rate among dependent people. These factors depress household consumption and ability to save. However, the economic impact of dependency displays marked geographical differences. Countries able to create greater work opportunities for dependent persons exhibit smaller differentials in average income between families with and without dependents, compared with countries unable to create such opportunities. DSBs have a significant capacity for reducing the risk of poverty: indeed, almost half of families with dependents receiving such benefits rise above the poverty threshold. The proportion is particularly high in Scandinavian countries while it is lower (but still sizeable) in the other areas of Europe. The effect of DSBs on the degree of economic compression of families is instead less evident. A second aspect considered has been the protection furnished by families. Throughout Europe, care is still a ‘family matter’: families deliver the bulk of care to dependent people. The impact of dependency therefore varies substantially according to the type of household in which the dependent person lives and the support that the family is able to furnish. Dependent people more frequently live with other family members that do non-dependent people. However, there are significant differences among care regimes. Scandinavian, German-speaking and Anglo-Saxon countries exhibit a more individualized model, with a large proportion of elderly dependent people living alone. In contrast, in Mediterranean countries, and to some extent in the Francophone area, the proportion of dependent persons living alone is low, while cohabitation between parents and adult children is widespread. In these countries, the frequency of this type of family organization reduces the poverty risk connected with dependency. Two specific vectors of social vulnerability emerge. The first concerns the material circumstances of life and the second has to do with family organization. Dependency creates economic vulnerability with pronounced effects on the living standards of families, notwithstanding generous public benefits that often prevent the onset of poverty, if not severe material compression. Since this vulnerability is closely correlated with the worklessness of dependent people, it has been least prevalent in those countries (particularly Scandinavian and Germanic countries) with greater labour-market participation by dependent adults. In any case, throughout Europe dependency brings about economic risks that are insufficiently recognized and protected against. The second focus of the analysis was the characteristics of caregivers. These have been defined as persons who devote at least 4 hours a week of unpaid carework to persons requiring help because of age, illness or

Disability and Caregiving 183

disability. Assuming this definition, the caregiving rate in Europe is calculated at 3.8 per cent – 3.5 per cent among adults and 5.1 per cent among the elderly. A marked gender asymmetry emerges, because twice as many women than men are caregivers. The most frequent caregivers are women aged between 40 and 64 years, belonging to the so-called ‘sandwich generation’. This group represents around 45 per cent in the various European areas (in Mediterranean countries it reaches 64 per cent). Our analysis of the vulnerability profiles connected with caregiving started from the assumption that coping with dependency takes place mainly within the family. Family care is therefore a crucial resource for dependent people in all the European countries. Nevertheless various forms of complementarity and/or substitution among informal care, labour-market participation and welfare services are apparent. To what extent is caregiving both a response to a social problem and a cause of vulnerability? Answering this question has highlighted the crucial role performed by the amount of time devoted to caregiving. In fact, very intensive care work makes it much more difficult to balance the constraints and opportunities connected with other activities. According to our data, caregiving is provided for 23 hours a week on average, with a median value of 15 hours. Three levels of caregiving have been defined according to the amount of time devoted to care: light (up to a maximum of 14 hours a week: 43 per cent of caregivers), medium (from 15 to 27 hours to week: 28 per cent of caregivers) and intensive (more than 28 hours a week: 29 per cent of caregivers). Vulnerability arises in three aspects. The first is the impact of caregiving on work. The structures of the labour market and of the welfare system play a decisive role in increasing or reducing opportunities to reconcile work and care. Increasing the hours devoted to care decreases the possibility of balancing carework and paid employment. While light caregiving appears to be compatible with maintaining a job, caregiving that exceeds 28 hours a week causes great difficulties in reconciling working and caring. A second vulnerability profile concerns the effects of caregiving on the production of income and living standards. The impact of caregiving on income and, in general, on the level of economic compression differs substantially according to the intensity of the carework and gender. It affects women much more than men, and exclusively impacts on intensive caregivers. We examined the relations between caregiving and sociality. Caregiving is associated with enhanced informal relationships with neighbours and friends. However, it reduces social participation outside primary relationships, which is particularly difficult for persons intensively engaged in care tasks. Once again caregiving, although crucial and indispensable, often gives rise to vulnerability when it is very intensive and may become excessively burdensome. In these cases, carework often takes the place of paid employment and exposes the caregiver to the risk of poverty and notable

184

Social Vulnerability in Europe

material compression, as well as stunting his/her social life. Overall, and in regard to all the profiles, caregivers in the South of Europe are in a much worse position than are caregivers in the other European areas. Finally, this analysis highlights major challenges for research and for social policy-making. Dependency – for both people experiencing it and caregivers – still has only a weak currency in public discourse and in public policies. Yet the analysis has shown that dependency can have a major impact on people’s lives. This is a problem that requires a new approach by policy-makers and a radical change in welfare systems. Of course, there are structural dynamics that clamour for such change and that increasingly manifest the gap between the growing care needs of the population and the weakening care capacity of contemporary society. Becoming a dependent person is an increasingly widespread experience. But caring will, in turn, be more difficult to combine with other social activities. Moreover, it will be more difficult to deliver care with the intensity required by dependent people. The problems of caregiving require the concept of well-being to be enlarged, both in theoretical terms and in the definition of social-policy goals. This entails redefining the constituent components of living (Sen, 1992) and the ‘basic entitlements’ (Nussbaum, 2000) that underpin Western societies and their welfare systems.

Notes 1. Supranational aggregates are used for the comparative analysis because of the scarcity of data. Unlike in the other chapters of the book, this prevents analysis at the national and macro-regional levels. 2. The WHO has recently adopted a new instrument and a new conceptual scheme to describe and measure the population’s health and disability: the International Classification of Functioning, Disability and Health (ICF). It has thus superseded the conception of disability centred on the individual, that neglected environmental and social factors. Although the concept of disability is extended in this new conceptual scheme, it is often used as ‘an umbrella term covering any or all of the following components: impairment, activity limitation and participation restriction, as influenced by environmental factors’ (Lafortune et al., 2007). 3. This applies mainly to elderly people. Eurostat, for instance, when collecting data for its Health Status Statistics, uses a disability indicator for adults of working age constructed by combining three variables relating to restrictions on the type of work performed, the amount of work that can be done and mobility to and from the workplace. 4. There is substantial agreement on their underlying causes: mainly musculoskeletal problems, arthritis, cardiovascular problems and respiratory disorders (Freedman and Martin, 2000). 5. According to OECD data (2006b), in 2004, the rates of institutionalization of the over-65s in residential facilities were as follows: Italy 1.5 per cent, Ireland 1.8, Germany 3.4, Austria 3.6, the United Kingdom 4.2, Denmark 4.4, Finland 4.9, the Netherlands 5.6, Norway 5.8, France 6.3 and Sweden 7.5.

Disability and Caregiving 185 6. As information on expenditure by the families of dependent persons is not available in ECHP data, this aspect can only be investigated indirectly. 7. For this reason the figures in Table 7.3 are not directly comparable with those in Figure 7.1. 8. These schemes generally belong to the category of long-term care measures (such as, for instance, the sums paid by the insurance scheme for the non self-sufficient introduced in Germany in 1995). Unfortunately, the ECHP survey does not furnish information on the receipt of services in kind as alternatives to monetary transfers. For this reason, given the variety of mixes between cash and care in the European countries, we have not considered here the economic benefits deriving from social assistance measures, because it is not possible to estimate the benefits delivered through services in kind. 9. Only benefits paid to dependent people of working age are considered here, given that above this age in many European countries disabilities benefits are converted in old age pensions. 10. The great majority of very intensive caregivers (more than 28 hours a week) are women (83 per cent), especially of adult age (61 per cent). Thus three in every five intensive caregivers are women of working age; the other two are elderly women and men of varying age.

This page intentionally left blank

Part III Multidimensional Analysis

This page intentionally left blank

8 The Vulnerability of Young Adults on Leaving the Parental Home Giuseppe A. Micheli and Alessandro Rosina

Introduction Transitional phases are always those in which people are most at risk, defenceless and vulnerable to external dangers. One of the most crucial changes in human life is the transition to adulthood. Although the timing of such transition has changed, and although the ancient rites of passage have disappeared, crucial stages still persist in the process of acquiring adult autonomy and responsibilities. One of them is leaving the parental home to live on one’s own or to form one’s own family. Young people leave the safe haven of their parents’ home to tackle life at first hand relying on their own resources. However, the greatest risks arise at the beginning of independent life. The first years are the true test for young adults entering public life. Starting off well builds self- confidence and the willingness to accept new challenges. Starting off badly, or hesitantly, may foster a sense of insecurity with harmful consequences. In post-industrial societies the capacity to control and manage risks has diminished, and the transition to adulthood has grown more complex. Globalization has created new opportunities, but it has also brought new risks for broad sections of the population. The scientific and political debate on the transformation of welfare systems required by the onset of these new risks highlights that the situations of young people are particularly problematic. The new risks are largely connected with difficulties of labourmarket entry and the acquisition of a stable and suitable job (Taylor- Gooby, 2004a). In countries with weak public welfare systems and national institutions that are more rigid in their responses to the potentially harmful effects of the great changes in progress (economic, demographic and social), in particular, young people tend to encounter greater difficulties in finding work, attaining economic independence and stabilizing their work careers. This situation generates insecurity and uncertainty about the future, with a consequent tendency to postpone choices that entail the assumption of responsibilities (Blossfeld et al., 2005). 189

190

Social Vulnerability in Europe

But the heterogeneity in the timing and patterns of exit from the family of origin in Europe also stems from cultural and anthropological factors that interact with the great forces of change. This interaction has unforeseen consequences on the system of constraints and opportunities encountered by young people as they build their adult lives, on their expectations, on the type of risks to which they are subject and on how they cope with them. This chapter deals specifically with vulnerability on exit from the parental home. Besides the importance of structural and institutional factors, the analysis expressly considers the family’s protective and propulsive role, which is closely connected with the nature of the tie between parents and children. This leads to exploration of inter- and intra-national similarities and differences, with results that confirm previous findings while also affording new insights.

The background Changes in the timing and patterns of leaving the parental home In the final decades of the last century, throughout the Western world, the process of transition to adulthood changed profoundly (Corijn and Klijzing, 2001). In all the advanced societies, family formation, and especially marriage, was postponed. Average age on first marriage in the countries of the EU increased by twoand-a-half years between 1990 and 2003, rising to almost 30 years for men and 27.5 for women. The highest values were recorded in the Scandinavian countries (particularly Denmark and Sweden), where age on marriage was over 30 for women and over 32 for men (Table 8.1). However, this delay in marriage does not correspond in those countries to a relatively older age when independence is attained and family formation begins. Instead, Sweden and Denmark are precisely those countries of the EU-15 in which early exits from the parental home are most widespread (European Foundation for the Improvement of Living and Working Conditions, 2003). There is therefore a marked difference between the countries of the North and the South of Europe in regard to the median age on leaving home. In 2003, this age was highest in Italy, Greece and Spain (with the highest values in Europe, according to Eurostat: 30 years for men and 27 for women) and lowest in Denmark and Finland (the lowest values in Europe: respectively 20 and 21 years). In general, these figures show that in countries where young people leave the parental home later, they more often directly form a family through marriage. In the Mediterranean countries, especially, a very brief amount of time (around 1 year) elapses on average between age on exit from the family of origin and age of first marriage. This is a consequence of the traditional close synchronization of the two events in those countries, where the majority of young people leave home only when they are about to get

Vulnerability of the Young on Leaving Parental Home

191

married (Billari et al., 2002; Rosina, 2004). This also applies to Belgium: according to de Beer and Deven (2000), ‘the fact that Belgian people are older when they leave home stems from the fact that the majority leave home only to marry’ (p. 54). Where young people instead leave home earlier, they tend to live for long periods as singles before they form a couple. Moreover, the first union consists much more frequently of cohabitation. For instance, more than 90 per cent of women in Sweden begin partnerships with informal arrangements (Kiernan, 2002). More than being the initial event in the process of forming a family, consequently, marriage sets the seal on the reciprocal commitment made by the partners when children arrive (Bernhardt, 2004). The lapse of time between exit from the parental home and marriage is the longest in the Northern European countries (more than 10 years for Danish males). Moreover, in Denmark, Sweden and Norway, it is common for the first child to arrive before marriage (Table 8.1). Therefore, although marriage is significantly postponed in those countries, childbearing is not particularly delayed. However, not all countries conform to a schematic opposition between the North and South of Europe. Portugal, for example, differs from the other Mediterranean countries of Southern Europe because the various events in the transition to adulthood occur at a younger age. Portuguese young Table 8.1 Average age at first marriage and at birth of first child, by country. Year 2003 First marriage Countries Austria Belgium Denmark Finland France Germany Greece Ireland Italy Norway Netherlands Portugal United Kingdom Spain Sweden

Difference

Women (A)

Men (B)

Women (C)

CA

27.4 27.1 30.1 28.5 28.2 28.1 27.3 25.5 27.4 29.1 28.4 26.1 27.2 28.3 30.5

29.9 29.3 32.3 30.4 30.4 30.6 31.0 ... 30.4 31.6 30.8 28.0 29.3 30.2 32.9

26.9 ... 27.8 27.9 ... 28.8 27.9 28.3 ... 27.5 28.8 27.1 29.3 29.2 28.5

0.5 ... 2.3 0.6 1.0* 0.7 0.6 2.8 1.6* 1.6 0.4 1.0 2.1 0.9 2.0

Note: * Estimated figures (on the basis of previous years). Source: Edurostat.

First child

192

Social Vulnerability in Europe

people leave the parental home later than do Italians and Spaniards, but at a decidedly older age than Scandinavians. In Portugal, the ages at marriage and the birth of the first child are the lowest of the EU-15 countries. France and the United Kingdom (but also the Netherlands and Germany) share with Italy and Spain a marked delay in the transition to motherhood. However, unlike in the countries of South Europe, it is much more common in the North-Western area of the continent for the first union not to be a marriage, and for births to take place outside marriage: in 2003, more than 40 in every 100 births in France and the United Kingdom were to unmarried parents, against a European average of less than 30 per cent (Eurostat data). More generally, in the Central-Northern European countries the changes brought about by modernization and globalization have been matched by a greater ‘de-standardization’ of the sequencing of the transitional events to adulthood. Features common to those countries are that young people leave the parental home while they are still in full-time education, they form their first unions even before they have stable jobs, they have children before getting married, etc. Vice versa, in countries where the sequence is still rigidly fixed, young people take longer to achieve independence (Billari and Liefbroer, 2007). Welfare regimes and different demographic patterns The wide variety in the timing and form of leaving home in Europe is matched by equal variety in the constraints and opportunities and institutional forms of support, experienced by young people in their transition to adulthood. The European Union can be seen as a natural laboratory, where labour market performance (providing jobs and earnings), welfare support (social services, transfers, labour-market policies, family policies), and family support systems (family formation and structure) should explain the timing of transitions to adulthood (Vogel, 2002: 125). Two polar models are usually identified in the ‘welfare mix’ (here understood as comprising the labour market): the social protection system and family ties (Vogel, 1999). One is a North-European model characterized by high employment rates, greater social spending on young people, low levels of poverty and few income disparities. The opposite is the Mediterranean model, with lower youth employment rates, greater social spending on the elderly, higher levels of poverty and income disparity and the important role of family solidarity (Ferrera, 1996). However, types of welfare regime cannot be reduced to simple polar opposites. An intermediate situation occurs in the countries of Central Europe, where (as in France and Germany) there is a system of guarantees

Vulnerability of the Young on Leaving Parental Home

193

and protections similar to the Northern model, but with social activity centred more on the family than the individual. There is also the more liberal model represented by the United Kingdom (with many similarities to the USA), where there are significant levels of inequality and poverty (EspingAndersen, 1990). Nevertheless, the dimension of family ties – crucial for understanding the difficulties, risks and opportunities of young people in their transition to adulthood – highlights the importance of the North/South axis. Various recent studies, in fact, show that the presence of solid and enduring relations between parents and children is anthropologically deep-rooted in the communities of Mediterranean Europe, compared with the weaker bonds typical of the rest of continent (Dalla Zuanna and Micheli, 2004). These strong vertical ties are expressed especially in the substantial support given by parents to their offspring. This support continues beyond the attainment of majority and the conclusion of full-time education until the attainment of paid and satisfactory employment, and thereafter until marriage, when the family provides crucial assistance with purchase of a dwelling (Rosina and Fraboni, 2004). But parental support continues even after the children have formed their own families. Strong ties induce children to live close to their parents and this proximity favours a constant flow of material and emotional resources in subsequent life-phases (Barbagli, Castiglioni, and Dalla Zuanna, 2003; Tomassini, Wolf, Rosina, 2003). The crucial role of the family of origin in Mediterranean countries can be considered both the cause and effect of the under- development of public welfare systems to assist young people in the crucial events of their transition to adulthood (Saraceno, 1994). Moreover, since the Mediterranean welfare model almost exclusively tasks the family with assisting young people in difficulties, it is also highly inequitable (Livi Bacci, 2005) because it penalizes young people from families of lower socio-cultural status with fewer economic resources. It is therefore a system that is less fair and less able to meet the challenges raised by the great changes now ongoing. As Taylor- Gooby writes (2004a: 8): Successfully managing new risks is increasingly important, particularly for the more vulnerable groups [...] New risks tend to affect people at younger stages of their lives than do old social risks, since they are mainly to do with entering the labour market and establishing a position within it, and care responsibilities primarily at the stage of family building. The fact that social welfare spending in the Mediterranean countries largely concentrates on the elderly and is ungenerous toward the younger generation – particularly as regards unemployment and housing benefits and support for family formation (Pizzuti, 2006) – renders young people

194

Social Vulnerability in Europe

more vulnerable when they leave home, and heightens their dependence on the family of origin. Young people in the Scandinavian countries leave home at an earlier age. They are encouraged to cope with difficulties by themselves and to live on their own even when unemployed (Vogel, 2002). Consequently, in the countries of Northern Europe, young people tend to leave home even with scant economic resources and only provisional work and housing arrangements. They also tend to have a better capacity to adapt. Although they are more exposed to risks, they live in countries with generous public welfare systems. The central area of Europe exhibits situations and features midway between those of the North and South. This applies, as we have seen, to the timing of exit from the parental home, but also pertains to the public measures and benefits available to young people, which are substantially more generous and efficient than in the Mediterranean area. The countries belonging to the liberal model have rather different features. In the United Kingdom and Ireland, marked inequality and modest public welfare provisions make young home-leavers vulnerable if they have low wages, especially if they have small children (Aassve et al., 2006). The presence of these intermediate patterns within the North/South axis therefore requires a more detailed analysis of national differences, which is conducted in the next section.

From a first interpretation of national differences to some basic questions To untangle the complex and geographically multi-faceted relationship between exit from the family and vulnerability using ECHP data, we have selected only interviewees aged between 16 and 34 years. Among these, we have selected only young people who left home during the period 1995– 2001, whose circumstances before and after the exit are known. Finally, a further selection has been made by considering as ‘non- exits’ people who left home for reasons of study, because this is an event decidedly different from family exit to form a union or for work (the majority of these subjects, in fact, tend to be economically dependent on their parents; a similar procedure has been followed by Iacovou, 2002).1 Since our aim is to study the vulnerability of young people during separation from the family of origin, we concentrate on the first year after exit, when home-leavers are particularly likely to find themselves in difficulties.2 Characteristics and conditions of home-leaving in Europe The ECHP data confirm the existence of marked differences between European countries in the times and modes of leaving the parental home, and these substantially replicate the North/South axis (see Table 8.2). For

Vulnerability of the Young on Leaving Parental Home

195

Table 8.2 Percentage of young adult home-leavers with specific characteristics (measured in the year following exit), by country Age 25

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

In a union

With children

Men

Women

Men

Women

Men

Women

85.1 63.3 41.8 61.3 52.7 17.6 30.6 24.2 45.3 52.1 80.7 43.2 17.4 69.4

93.8 78.7 64.4 74.5 66.1 34.1 49.6 36.4 59.0 64.1 89.8 74.0 46.1 75.9

37.9 44.7 66.7 52.9 50.0 69.9 73.3 74.1 89.2 59.3 42.2 57.9 64.1 53.0

46.5 56.0 66.1 58.4 42.4 79.6 82.7 80.1 90.3 61.4 45.7 63.0 61.8 55.8

1.2 0 1.2 0.5 1.8 4.3 4.7 4.7 6.9 6.4 2.6 4.6 8.2 1.8

1.4 1.3 5.1 4.3 17.8 8.0 9.1 11.8 17.9 22.2 1.6 5.1 9.1 15.1

Source: ECHP, authors’ own calculations.

instance, the great majority of young people in Denmark leave home before the age of 25, but only a small proportion of them form unions during the year following exit. Conversely, in Mediterranean countries, young people leave home later, and much more often tied with the formation of their first union. In these latter countries, therefore, the proportion of home-leavers already with children in the year after exit is also high; but the proportion is also high in Austria, the United Kingdom and Ireland, where there is a greater propensity to have children outside formalized unions. The situations of young people in regard to housing are highly diversified (Table 8.3). In some countries, large use is made of rented accommodation, as a result of a particular type of housing market, greater mobility (for work) and looser family ties (and therefore less help and proximity). Many countries of continental Europe (particularly France, Germany and Austria) and the Nordic area (Denmark and Finland) belong to this group. Renting homes is also common in the United Kingdom. On the contrary, in the Netherlands and Belgium the dwelling is more frequently purchased, because banks grant mortgages to young people on easier terms. In the countries of Southern Europe, home-leavers are instead divided almost equally between those who move into rented property and those who purchase or already own a dwelling, with the prevalence of the latter. The majority of owner-occupied dwellings in Spain and Portugal have been acquired with mortgages, while in Italy and Greece home- ownership more often derives from transfers made by parents or direct purchase (often made

196

Social Vulnerability in Europe

Table 8.3 Percentage of young adult home-leavers according to type of occupancy (in the year following exit) by country Mortgage

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

Rent

Self- owned

Men

Women

Men

Women

Men

Women

22.4 60.7 45.5 8.1 32.6 16.5 4.3 45.0 31.8 19.3 11.1 11.0 21.7 35.7

14.6 56.7 34.5 7.9 29.2 15.4 3.5 42.6 33.4 21.6 7.5 6.7 29.7 37.9

72.7 31.3 17.0 89.5 16.5 45.0 53.0 44.2 50.8 70.0 83.3 87.8 31.0 60.5

84.7 32.7 19.8 90.4 17.4 48.5 48.0 40.0 48.2 68.6 89.3 90.7 33.3 60.1

5.0 8.0 37.6 2.5 50.9 38.5 42.7 10.8 17.4 10.7 5.6 1.2 47.3 3.8

0.7 10.7 45.8 1.7 53.4 36.1 48.4 17.4 18.4 9.8 3.2 2.6 37.0 1.9

Source: ECHP, authors’ own calculations.

possible by financial assistance from parents and by the postponement of home-leaving until sufficient funds have been accumulated). Ireland is a particular case in that it has the highest level of direct home- ownership without a mortgage. Geographical differences in levels of vulnerability When conducting our analysis, we decided to use various indicators of vulnerability. We selected first, as a specific aspect of vulnerability, a diminished capacity to save (net of expenditure). This does not necessarily denote an actual situation of difficulty, although it indicates exposure to one. It therefore seems a particularly suitable indicator of moderate vulnerability. It can be flanked with another indicator – the ‘ability to make ends meet’ – which, because it refers to a self- defined ability, is more closely influenced by subjective perceptions and mixes personal expectations with self- efficacy. The no-savings indicator can be appropriately used to assess the difficulty, immediately after the formation of the new household, of building capital through saving. However, because the indicator has certain drawbacks,3 we shall use a further indicator of material deprivation that combines income with housing.4 As is to be expected, the geographical distribution of vulnerability differs according to the indicator used. If we consider the material deprivation indicator (see Table 8.4), worse conditions are apparent where young people are

Table 8.4 Vulnerability indicators. Income poverty and housing deprivation, by gender and country Men

Denmark Netherlands Belgium France Ireland Italy Greece Portugal Spain Austria Finland Germany Luxembourg

Women

Neither deprivation nor poverty

Housing deprivation

Income poverty

Cumulated deprivation

Neither deprivation nor poverty

Housing deprivation

Income poverty

Cumulated deprivation

36.9 73.9 63.8 41.3 66.5 58.5 46.9 76.6 56.6 72.7 23.8 44.7 70.3

33.5 6.5 26.3 34.6 8.8 20.1 32.6 11.0 33.1 18.0 47.6 37.6 21.6

7.8 14.5 4.6 7.6 15.0 13.9 8.1 8.0 4.7 7.3 5.8 8.3 3.4

21.8 5.1 5.3 16.5 9.8 7.5 12.4 4.5 5.6 2.0 22.9 9.5 4.7

25.9 75.0 62.0 41.3 56.7 56.1 47.0 72.4 55.6 62.7 14.6 39.5 57.3

35.3 5.8 28.0 32.2 11.9 22.9 31.4 11.5 29.4 23.5 47.4 36.6 29.8

11.2 16.7 6.7 6.9 17.4 12.6 10.1 11.9 7.3 8.4 7.5 8.0 3.8

27.7 2.5 3.3 19.7 13.9 8.4 11.5 4.2 7.8 5.4 30.5 15.8 9.2

Note: Information not available for the United Kingdom. Source: ECHP, authors’ own calculations.

198

Social Vulnerability in Europe

more likely to live on their own even when underemployed or unemployed. This happens in Scandinavian countries, the Netherlands, Germany, the United Kingdom and France, where the greatest risks are housing deprivation and the combination of poor housing with low income. Conversely, in the Southern European area, only Greece has levels comparable with those of the Northern countries. Data not reported here relative to the material deprivation indicator also show that it is significantly tied to age: in all countries, the absence of housing deprivation and poverty is lower for young people aged less than 25. Consequently, there is less vulnerability in countries where young people leave home at an older age. Contrary to the material deprivation indicator, the one based on inability to save instead records potentially worse conditions in the Southern European countries (see Table 8.5). The results obtained using the subjective indicator (‘making ends meet’) are consistent with this finding. The countries in the worst situations are Greece, Portugal and Italy. Finland follows, although on the subjective indicator it is overtaken (in the negative sense) by Spain. In considering this dimension of vulnerability, therefore, the difficulties encountered by young home-leavers in the Mediterranean (in terms of low skills, lower entry wages and less generous welfare benefits) are only partly offset by delayed exit and the greater support provided by the family of origin. Table 8.5 Vulnerability indicators. Percentage of young adult homeleavers unable to save in the past year, or with difficulties in ‘making ends meet’, by gender and country

Unable to save

Denmark Netherlands Belgium France Ireland Italy Greece Spain Portugal Austria Finland Germany Luxembourg

Difficulty in making ends meet

Men

Women

Men

Women

50.3 33.3 45.7 59.6 50.5 71.8 90.1 42.6 84.9 30.2 67.0 46.1 43.7

52.1 34.7 43.2 63.4 54.3 71.1 92.1 41.2 84.2 43.8 70.8 41.6 43.6

11.8 8.0 12.1 15.0 18.8 21.7 41.7 25.7 35.9 15.0 21.5 n.a. 5.5

17.4 10.7 7.4 15.7 16.7 21.9 41.0 26.4 35.1 18.3 19.4 n.a. 6.7

Notes: Information not available for the United Kingdom. na: not available. Source: ECHP, authors’ own calculations.

Vulnerability of the Young on Leaving Parental Home

199

These indicators, too, show that deprivation declines with age in all countries. An exception is Ireland, where there is an increased incidence of difficulties among people who leave home after the age of 30. A somewhat similar pattern is apparent in the Iberian peninsula, but the differences by age are very small. This first descriptive analysis therefore confirms the presence of situations of more acute income deprivation where inter-generational solidarity is weaker, where young people more commonly leave the parental home earlier despite the likelihood of encountering an initial phase of severe difficulties to which they will have to adapt and where a public welfare system helps young home-leavers maintain their independence even when they are unemployed. In countries where it is instead more common to leave home at an older age and then form a family straight away (with the help of the parents), it is also more common for home-leavers to try to build solid but costly exit conditions. Hence, the substantial investments required in this first phase away from home reduce the chances of saving.

Four questions What factors condition the differing degrees of vulnerability in the various regions of Europe? The results of the foregoing statistical analysis point to some main explanatory factors. Consideration of them raises the following four questions. [a] With regard to human or economic capital, it is likely that higher educational qualifications, a more stable job or a higher salary – which gives access to a mortgage – generally, other conditions remaining equal, afford better protection against vulnerability. In what countries does the lack of these material resources generate greater vulnerability among young home-leavers? This question comprises – and generally conceals – a second one. It is customary to think of the home-leaver’s age as, not a variable influential on subsequent life- events, but only as a proxy for other processes that develop with age. For example, postponing the age of exit from the parental home enables young people to accumulate resources and thereby reduce their vulnerability. Yet, might it not be the case that breaching the ‘norm’ on family exit (leaving home at an older or younger age than ‘normal’5) produces more marked vulnerability? And in which European regions – in which welfare and market contexts, but also in which cultures of transition to adulthood – does this happen? [b] As regards housing, having to pay rent or repay a mortgage heightens the difficulties of young adults in the period after they have left home. It

200 Social Vulnerability in Europe

is likely that this exacerbation of difficulty varies with the welfare system, the characteristics of the housing market and cultural aspects concerning the nature of the ties between parents and children and inter-generational transmission. In what countries does the type of occupancy (whether the dwelling is owner- occupied, rented or mortgaged) more strongly influence the vulnerability of young adults? [c] As regards living arrangements after family exit, it is likely that having a partner gives protection against deprivation, whereas having a partner without an income, or with one or more dependent children, generates greater vulnerability. Is this hypothesis applicable to all of the European countries, or does some partition of the continent contradict it? [d] The family’s resources may provide young people with protection when they leave home, endowing them with greater start-up funds or larger human capital to spend on finding employment. Then, after the young person has left home, the family’s resources can continue to avert or resolve difficulties. In what countries does the absence of family capital – the young person’s economic or human capital remaining equal – increase the homeleaver’s vulnerability? The next section seeks to give systematic answers to these questions. To this end, we shall try to identify – for every country or macro-region – the individual or contextual markers that significantly increase (multipliers) or decrease (reducers) the vulnerability of individuals coping with critical situations of varying severity. More specifically (see Figure 8.1), we shall measure the statistical significance (by means of a suitable logistic regression model)6 of a range of factors reducing or increasing the vulnerability of an individual coping with a firstorder critical state, that is leaving the parental home, and consequently, though not necessarily, incurring a second- order criticality of varying severity (a lower one, like the inability to save, or a higher one, like downdrift below a poverty line).7

First order criticality (leaving family)

Vulnerability

Multipliers / Reducers (factors increasing or decreasing vulnerability)

Second order lower criticality (no saving) Second order higher criticality (poverty)

Figure 8.1 Flow chart of the vulnerability in the transition to adulthood

Vulnerability of the Young on Leaving Parental Home

201

National frames of vulnerability on leaving home The array of factors influencing vulnerability Above, we described trans-national differences in the frequencies of certain vulnerability conditions of young adults in the first year after they leave home. The piecemeal nature of that description requires us to seek more systematically for factors – familial or economic, cultural or structural or tied to life-events or to conditions in the labour and housing markets – that most forcefully induce vulnerability. Because the range of possible explanatory factors – as well as the range of possible forms of vulnerability to be explained – is ample and indeterminate, we shall restrict the modelling in two ways. First, if light is to be shed on geographical differences in the vulnerability of young adults, it is essential to control for the severity of the second-order critical situation. This is because having to cut down on non- essential expenditure and settling for ‘dignified poverty’ is different from lapsing, even if temporarily, into poverty. We must therefore explore the reasons for vulnerability by considering two distinct, yet hierarchical, levels of criticality. As previously, the first indicator consists in the difficulty of saving; the second, which denotes a condition of greater economic criticality, is an income below the International Standard Poverty Line (ISPL), this being calculated according to the income distribution in the country concerned. While an inability to save indicates a lower level of criticality,8 falling below the ISPL indicates a higher one. However, it should be borne in mind that – in a life- course phase characterized by exit from home and by recently acquired economic independence – poverty may not be necessarily a chronic condition, but may be also one of the acceptable and accepted costs of becoming adult, like those connected to the birth of a child (Micheli, 1999). Second, the factors potentially influencing the vulnerability of young adults are selected and aggregated into four groups corresponding to the questions posed in the preceding section. In the statistical models that follow we hypothesize that distinct roles are performed by: a. events in the educational and work career associated with the age stages marking the transition to adulthood, in addition to the schedule of leaving home if this is at odds with the norm in the country concerned (early or delayed), that might have effects on vulnerability different from those of work and education; b. labour and housing market mechanisms shaping the transition to adulthood that might hamper its stabilization; c. crucial – but not unavoidable – events in family formation (forming a stable relationship with a partner, a child’s birth); and

202 Social Vulnerability in Europe

d. the family’s economic (income) and human (father being a graduate) capital. Tables 8.6 and 8.7 summarize the factors statistically significant in the production of vulnerability, distinguished by gender, country (or small block of countries) and the two criticality levels selected: one lesser (no savings) and one more severe (income below the poverty line). Before geographical differences in vulnerability patterns are analysed in detail, a preliminary finding drawn from Tables 8.6 and 8.7 is synthesized in Table 8.8: the main factors multiplying or reducing vulnerability that affect the two levels of criticality are, at least partly, different. Consider compliance with the correct timing of the transition to adulthood, which is one of the most incisive and widespread factors in reducing susceptibility to strong vulnerability: the older the age on leaving home, the lower the risk of falling below the poverty threshold. Age is instead much less influential on vulnerability to a less critical state because it is presumably mediated by other and more specific factors, such as good income or higher educational qualification. Hence, postponement of these stages toward acquisition of economic independence is a strategy useful for avoiding states of slight criticality because such postponement consolidates the career and augments resources. By contrast, age is decisive in the formation of poverty: compliance with the life- events schedule is essential for keeping economically solvent, while precocious transition is the decisive factor in impoverishment. The life- events schedule therefore has very important weight in producing vulnerability. But some of these events, those connected with family formation, are decisive regardless of their scheduling. Forming a couple reduces vulnerability at both levels of criticality (but with an ambivalent effect – to which we shall return – when only the male is employed). Equally evident is the vulnerability multiplication effect exerted by the birth of a child.9 Like income, education also affords protection. But with an interesting distinction: while higher household income (and greater savings) is influenced by possession of a degree, the structure of the labour market makes upper-secondary education level (such as A-levels in the United Kingdom or the baccalauréat in France) valuable in avoiding poverty. Confirming this is the fact that problems in job-seeking are among the prime factors responsible for the onset of poverty, while they have much less weight in the formation of a state of slighter criticality. Housing costs instead play a role at both levels, and in close connection with local markets. A final important set of vulnerability-reducing factors consists in the economic and human capital transmitted by the family of origin. The family income very significantly protects home-leavers against falling below the

Table 8.6 Multipliers (M) and reducers (R) of vulnerability when leaving home: susceptibility to a lower second-order criticality (no saving) (Statistical significance: R/M: .10; RR/MM: .01) Men UK

Life events

Partner Partner no income Children

Market factors

Older age Personal Income Degree R Upper secondary qualification

Atypical job* No job* Rented home Mortgaged home

Home

Age normed

Sca

Parental income Graduate father

R

Irl

Fra

Women

BNe

DAu

Por

Spa

Ita

R

RR

R

R

R

Gre

Sca

UK

Irl

Fra

BNe

DAu

R

RR

R

R

Por

Spa

Ita

Gre

R

R

M

RR

R

RR

RR

RR R

R

R

M

M

R

RR

RR

R

RR

R RR

RR

R

MM

M

R

M

M

MM

M M

R M

M MM R

R R

Note: * Compared with the reference situation of a permanent job. Source: Our calculation on ECHP database.

R

RR

M R

R

M MM M R

RR

Table 8.7 Multipliers (M) and reducers (R) of vulnerability when leaving home: susceptibility to a lower second-order criticality (drift under poverty line) (Statistical significance: R/M: .10; RR/MM: .01)

Life events

Partner Partner no income Children

Market factors

Older age Personal income Degree Upper secondary qualification

Atypical work* No job* Rented home Home mortgaged

Home

Age normed

Men

Parental income Graduate father

Sca

UK

Irl

Fra

BNe

DAu

RR

RR

R

RR

RR

R

Women Por

Spa

Ita

Gre

Sca

UK

R

RR

Irl

Fra

BNe

R

RR

R R

RR

R

RR R M

RR RR

R

RR

RR

M

MM

MM

M

RR

RR

RR

M

Gre

R R

RR

M

RR

R

RR

RR

M

MM

M

MM

M

M

M

MM

MM

M

RR MM

Note: * Compared with the reference situation of a permanent job. Source: ECHP, author’s own calculations.

Ita

M

MM

RR

Spa

M

M M RR

R

Por

R

R RR

R

R MM

DAu

M

RR

RR M

RR

R

R

RR

RR

Vulnerability of the Young on Leaving Parental Home

205

Table 8.8 Typology of multipliers and reducers of vulnerability, by level of severity of the second-order critical state (less frequent factors in brackets) Second-order criticality Factors Reducers

Lower Income Degree Partner (Parental income)

Multipliers

Children (Rented home)

Ambivalent

Mortgaged home

Higher Older age Upper secondary education Partner Parental income (Graduate father) Children No job (Rented home) Partner no income

Source: ECHP, authors’ own calculations.

poverty threshold. It is flanked (though less frequently and intensely) by human capital, as if the father is a graduate. The protective role of the family income is instead weaker in the case of lower- order criticality (indeed, it is likely that the family of origin intervenes mainly to deal with the severest difficulties, but much less so for routine problems). Inspection of Tables 8.6 and 8.7, however, shows a factor that can influence vulnerability in two opposite directions: by exacerbating susceptibility to a second- order criticality, or by reducing it, according to the context (economic, cultural or type of welfare system). This is the case of the presence of an income-less partner, which exposes French males to a greater risk of falling below the poverty threshold. And it is also the case of the ambiguous meaning of the effect of a mortgage. As a rule, this has a strong multiplicatory function. But in some cases (in Table 8.6 for Spain, for both males and females, and only with reference to the production of slight criticality), the surprising finding is that repayment of a mortgage performs a protective role. It may be that, for some young people, having a mortgage is a proxy for a good income that enables them to save.10 But the precise location of this reversal of effect in the Iberian region – where (see Table 8.4) it is commonplace for parents to purchase a dwelling for their children when they leave home by taking out a mortgage – suggests a different explanation, one which interprets the protective effect of a mortgage as a proxy for the strong support of the family of origin. But this requires us to go more deeply into the geographical differences among vulnerability factors. Geographical and gender differences in the incidence of vulnerability factors The factors reducing or multiplying susceptibility to a second- order critical state do not operate uniformly across Europe. Within the dense thicket of statistic significance signs shown in Table 8.5 – which cluster differently among the countries and regions of the continent – it is possible to identify

206 Social Vulnerability in Europe

some demarcation lines separating distinct patterns of vulnerability by gender and area. These are now briefly described. A first and not obvious result is that the role of age, taken on its own, in the production of poverty is significant – for males even more than females – almost solely in the countries of Central and Northern Europe (Scandinavia, the United Kingdom, France, Belgium and the Netherlands) while it is of no importance, for either males or females, in Mediterranean countries (Spain, Portugal, Greece and, for males only, Italy). It may be said that the risk connected with moving the life schedule forward is more marked in the ‘strong’ and competitive segments of the European continent: in the liberal and social democratic welfare regimes more than elsewhere, and for men more than women. We shall return to this point in the next section. For less critical states, the direct influence of age in exacerbating susceptibility to poverty is replaced by the direct and widespread influence of the variables ‘income’ and ‘education’, even though they are age dependent.11 But the two indicators almost never exert a joint influence (with the exception of females in the Benelux countries and in the German-speaking area), which is indicative of their mutual substitutability, context by context. The protective role of income is particularly significant among males and females in Germany and Austria, Belgium and the Netherlands, among females in the United Kingdom and, to a lesser extent, in Italy. But among French, Irish and Scandinavian males the decisive role is performed by education. Surprisingly, neither a degree nor personal income has a direct influence on the onset of poverty. An exception is the protective function performed in some countries by possession of an upper-secondary school qualification, which is a valuable facilitator of the job matching process. Apart from this exception, however, education and income affect the likelihood of poverty only indirectly. They do so in two ways: one is the already-mentioned age on leaving home; the other (and crucial) one is the possession of a job: a condition that is essential almost everywhere, with few exceptions (among them Spain and Portugal, as regards both males and females). Two different recipes for the economic equilibrium of young people leaving home are differently mapped across Europe. Countries – and gender segmentations – in which such a qualification performs a protective role largely overlap with the ones in which a precarious job – and joblessness all the more so – is an insurmountable obstacle to economic well-being. A secondary-school qualification and steady employment, or at least an atypical job, protect young Scandinavian, Italian and Irish males, and Belgian, Dutch and Greek females, against the poverty risk. By contrast, a job alone suffices to protect French, English and German-speaking males and females. The two significant forms of protection against poverty are differently but haphazardly distributed across Europe. We can imagine – but without developing the point further – that these two forms of protection and their two

Vulnerability of the Young on Leaving Parental Home

207

mappings stem from different workings of the relative labour markets, as well as from conjunctural dynamics. However, the crucial role of job precariousness in generating poverty among young males – apparent everywhere but statistically most significant in Germany, Austria, Italy and Ireland – has also to do with the persistence of the MBW model. Linked to this model is the marked vulnerabilizing effect for French males of having an income-less partner. Thus delineated is an area of Central- Southern Europe comprising Italy, France and Germany, where the vulnerability of young people transiting to adulthood seems linked with the weakness – and irreversible crisis – of the traditional gender division of labour. In the countries of Northern Europe, in contrast, and particularly Scandinavia and the United Kingdom, it is ‘non-couple formation’ that leads to exposure to the risk of poverty. The outlines of these two areas seemingly confirm the broad cultural divide stereotypically drawn between Northern and Southern Europe. But this simplification of a single great border between North and South is at the same time contradictory, because the Mediterranean countries (Spain, Greece, Portugal) do not pertain to these European ‘Midlands’. This third segment of Europe exhibits a different pattern, with partly novel features. Here vulnerability to states of high criticality, for young adult males, is defined by a sum of negatives: it does not depend on non- compliance with the age schedule, job precariousness (which instead exerts strong influence in Italy, an adjoining country), the lack of a partner or even on the presence of a child.12 What, then, is it that produces the poverty risk of young adult males in the Southern part of the continent? It is, and strikingly so, the absence of help from the family of origin. A strong mechanism of inter-generational wealth transfer operates mainly through male offspring.13 Further evidence of the dominance of this mechanism is provided by the particular protective effect exerted by a mortgage in this part of Europe. If a ‘mortgage and job’ interactive factor is inserted into the model alongside the ‘job’ and ‘mortgage’ variables, it proves statistically significant and exerts a vulnerability multiplication effect. Although there may be numerous explanations for this finding, one endorsed by the historical and sociological literature (as we shall see in the next section) is that the jobless, in a system with a predominant MBW model, are able to obtain mortgages for which, in these cultures, family support is decisive. Where home-leavers instead have jobs and take out their own mortgages, the latter are more likely to be further obstacles against coping with second-order criticalities. Analysis of Tables 8.6 and 8.7 therefore suggests a tripartite division of Europe. The first part comprises many countries of Central-Northern Europe (Scandinavian countries, the United Kingdom, France), where the vulnerabilizing effect of joblessness is flanked by a second effect concerning further anticipation in the life- event schedule. While on the one hand

208 Social Vulnerability in Europe

the traditional earlier entry into adulthood in these countries is beneficial to their demographic exchange and to the vitality of their economies, on the other it increases, not so much the slight risk of low savings, as the more severe (though temporary) risk of poverty. The second part of Europe comprises the South continental countries (Germany, Austria and Northern Italy), excluding the strictly Mediterranean countries. Here the strong vulnerabilizing impact of the male’s job precariousness or the partner’s non-work indicates a cultural model of gender relations (the traditional one of the single-income household) subject to unsustainable pressures by the new rules of a market by now disembedded from the reciprocity system (Micheli, 2003). The third focus of Europe centres on the Mediterranean countries (and sub- country regions, as we shall soon see), where economic precariousness, the absence of a partner or a dependent child, have less influence on vulnerability, as they are offset by the decisive protection provided by the family. The hypothesis of a North/South divide that accounts for all the different regimes (demographic or welfare ones or those of work or social policy) therefore seems contradicted by the existence of a sort of Midlands region, lying between the two sides of the divide. We shall seek to resolve this apparent contradiction by examining the sub-national boundaries that cut across the cultural models of Southern Europe.

Macro-regional frames of vulnerability on leaving home The sub-national boundaries of the family models of inter-generational transmission Three distinct ‘identikits’ of three segments of Europe aggregate countries with similar parameters into the models explaining vulnerability. But these models lump all the regions of a nation-state together within the same boundaries. But to what extent is it legitimate for the centripetal force of the nation-state to cover and conceal the centrifugal force of its internal regions? And if we let this centrifugal force operate, what is the likelihood that new boundaries, and new territorially compact and culturally homogeneous areas, will emerge? In the past decade in Europe, two distinct approaches have been used to partition Europe into national blocks on the basis of affinities both in their socio- cultural and policy frameworks. The first and predominant approach is to cluster countries according to similarities in their policy regimes, because such regimes are viewed as deriving deterministically from deep-lying and persistent structural processes, generally related to the great demographic, class-based and gender-based coordinates. Esping-Andersen’s (1990) typology of three distinct welfare regimes (liberal, conservative and socialdemocratic), based on the different degrees of commodification of welfare programmes, has been followed by similar attempts to typologize regimes.

Vulnerability of the Young on Leaving Parental Home

209

These differ according to the specification of the welfare segment considered. But in the regime- clustering approach, the typologies of European countries tend to privilege the rules regulating relations among market, social reproduction and the practices of welfare implementation. In this way the pre- conditioning role of the patterns models of primary anthropological relationships stratified over time and constituting the core around which the state-form coalesces is overlooked. The historical dimension of structures, or the structural dimension of history, is neglected. Sensitive to this dimension is the alternative approach to the reconstruction of European geography, which assumes that affinities are associated with the local properties of regions, like a shared language or culture. The ground-breaking study on these matters (Castles and Mitchell, 1993) – which groups national patterns of public policy into three families of nations, English-speaking, German-speaking and Scandinavian – rests on the idea that modern nation-states manifest greater or lesser resemblances in virtue of characteristics deeply embedded in their historical trajectories of development. Reference to Polanyi’s notion of ‘embedding’ gives a sound frame to this line of thought. In fact, the substratum of primary anthropological relationships is essential for explaining, not only practices of social reproduction, but also the ‘policy prescriptions’ designed, with different combinations, to reconcile the production and welfare systems. Not by coincidence, Castles finds a marked longevity of clusters of nations identified throughout the quarter of a century from the 1970s until today: this persistence of policy typologies results from precisely the fact that they are embedded in distinct reciprocity systems. In the presence of family models that have stratified over time, in locally different ways, the evidence for the weight of structural features and welfare regimes in the geographical differentiation of Europe should not be overestimated. Policy matters, no doubt; but the primary anthropological structures influence the impact of policies and market strategies just as much as (perhaps more than) policies and the market influence them in return. Taking due account of the historicity of local family patterns leads to definition of European geographies different from those that have emerged from the comparative political economy debate. It is significant that the main taxonomies developed in accord with Castles and Mitchell’s approach (Leibfried, 1993; Ferrera, 1996; Castles, 1998; Stockard and O’Brien, 2006), converge on a fourfold division of the European patterns, where the three worlds of Esping-Andersen (1990) are flanked by the South-European subcontinent. But it is historical biographic analysis on archival sources for past centuries that furnishes the most persuasive evidence. In a recent study that has rapidly become the benchmark for the historic biographical debate, David Reher (1998: 203) has observed that in Europe ‘it is not difficult to identify areas where families and family ties are relatively

210

Social Vulnerability in Europe

strong and others where they are relatively weak’: The dividing line, in some ways, is actually much simpler, with the Central and Northern part of Europe (Scandinavia,14 the British Isles, the Low Countries, much of Germany and Austria), together with North American society, being characterized by relatively weak family links, and the Mediterranean region15 by strong family ties. (cit.: 203–204) The strength of family ties (i.e. of the type of pact established between generations belonging to the same kinship descent) produces important functional differences in many aspects of family life. In particular, when children leave the parental home, in Northern Europe they do so as soon as they have achieved a minimum degree of economic independence (and only later get married and form a family); in Southern Europe home-leaving tends to coincide with marriage and the attainment of a stable job. Reher proposes the ‘strong’ family as an anthropological pattern that applies to Central-Southern Europe in its entirety. Nevertheless, between Northern and Southern Spain, and between Northern and Southern Italy, the organization of the family exhibits differences that clash with this presumed homogeneity, as we have already seen in the previous section. The divide between the two regions of the South-European family depends on the rule determining the residence of children transiting to adulthood. Whereas in the continental South of Europe it is still the social practice for children to live close to the parental home (a legacy from the nineteenthcentury model of the stem-family), the regions of the Mediterranean basin are instead characterized by the neo-localist organization of newly formed couples (Micheli, 2000). This basic distinction between the two SouthEuropean cultures gives rise to different ways of interpreting the role of the physical place of residence (house), as distinct from the home, which is the locus of family relations and the affections, of the family’s continuity and identity. In the continental regions of South Europe, the two concepts merge into the sole category of ‘home’, which has physical, relational and symbolic dimensions: in this context leaving home, and therefore the family of origin, is a wrench more painful than elsewhere. Instead, the bond between family and home is weaker in the Mediterranean regions. Thus a precise demarcation line is delineated within Southern Europe between the regions of the Southern European orographic ridge and the regions along the Northern rim of the Mediterranean basin. Thus, the boundary between the two distinct patterns of primary anthropological relations sometimes passes within and sometimes across the borders of the nation-states. As Reher writes: The specific boundaries of different family systems are often not crystal clear and there is much sub-regional difference. For example, in some

Vulnerability of the Young on Leaving Parental Home

211

respects Ireland does not fit well into Northern European family patterns, there are indications that Northern and Southern France often walk divergent paths, and the Southern fringes of Spain, Italy or Portugal often show distinct characteristics from the Northern parts of those same countries. Within individual societies, there is also much room for heterogeneity affecting families and family life. This multiplicity of forms and behaviour, however, does not negate the existence of more general regularities affecting large areas of Europe. (Reher, 1998: 204) This requires us to repeat some previous analyses (based on the traditional national clusters) drawing demarcation lines that pass within and across national borders. These demarcation lines are obtained, not by relying on the automatism of cluster analysis algorithms, but by drawing on the stock of knowledge accumulated by historical biographical studies. This knowledge is tentatively extended to North Europe, in order to take account of the differences and similarities among countries described in the previous section. It is thus possible to plot a map (Figure 8.2) of five macro-regions16 where Southern Europe is split into two parts, where there persists a Central continental area (comprising Belgium, the Netherlands and the North-Western sector of Germany) and where a further division of the North – different from the traditional one – groups the Central-Northern regions of France together with the British ones, distinguishing them from Scandinavia and Eastern Germany. It is entirely evident that this partition of Europe is not equally robust in the North and in the South of Europe. Whereas the dividing lines among the three segments of the North continental area rest essentially on statistical similarities, the separation between the two Southern macro-regions finds support in the most recent and authoritative historical research. The findings of such analyses cannot be disregarded, and they form the basis of the discussion in the next section. The European Midlands and two polarities in the schedule for entry into adulthood The simplification of Europe’s geography into five areas makes it possible to identify trends in the timing of transition to adulthood and in their recent evolution (see Table 8.9), trends which are quite similar (with the inevitable exceptions) within each area and different from area to area. Table 8.9 shows a clear hierarchy in schedules for detachment from the family of origin. In the Mediterranean regions and those of the South, young people leave home much later,17 but they then more rapidly (usually immediately) move to the next transitional stage to adulthood, namely the formation of a union. In effect, the differences in the schedules for stable union-formation are much less marked among the five European

212

Social Vulnerability in Europe

Legend North cluster North-West cluster Centre cluster South cluster Mediterranean cluster

Figure 8.2 Five European macro-regions with distinct demographic and anthropological patterns

macro-regions than those for leaving the parental home. What remains highly differentiated, between the North and the South of Europe, is the schedule for completion of the final stage, that is procreation: while in the two South European regions the lapse of time between forming a couple and the birth of a child has continued to lengthen, in the North (especially in the North-Eastern quadrant) early home-leaving and a substantial wait before forming a stable union are followed by a strong propensity to rapidly complete the transition to adulthood with the birth of a child. Only 13 per cent of 30-year- old males have a child in the Mediterranean macroregion, while the figure is twice as high in the Scandinavian and NorthWestern regions; moreover, only 37 per cent of women aged 30 have a child in the former macro-regions, as opposed to 55 per cent in the latter.

Vulnerability of the Young on Leaving Parental Home

213

Table 8.9 Shares of young people age 25–29 and 30–34 who have left home, have formed an union, and had a child, by gender (year of survey 2001) Aged 25–29 Me

S

Aged 30–34

Ce

Nw

Ne

Me

S

Ce

Nw

Ne

Out of 100 males, those who ... have left home [t1] have formed a couple [t2] have a child [t3] CV18 Heterogeneity in t1

26

46

71

69

88

55

70

90

85

93

23

40

56

54

63

55

65

77

74

78

13 0.269 0.77

21 0.298 0.994

20 0.437 0.824

22 0.405 0.854

25 0.441 0.422

37 0.173 0.99

45 0.18 0.84

45 0.268 0.36

49 0.217 0.51

53 0.221 0.26

Out of 100 females, those who ... have left home [t1] have formed a couple [t2] have a child [t3] CV Heterogeneity in t1

49

61

85

78

93

77

80

96

91

98

47

56

71

61

74

75

76

83

76

82

31

35

31

40

44

63

63

61

67

71

0.19 0.999

0.222 0.952

0.367 0.51

0.26 0.686

0.287 0.26

0.086 0.708

0.099 0.64

0.181 0.153

0.127 0.328

0.132 0.078

Note: Me = Mediterranean Cluster; S = South C.; Ce = Centre C.; Nw = North-West C.; Ne = North-East C. Source: ECHP, authors’ own calculations.

The usual interpretation of these findings is as follows: the diversification of when children are born is responsible for the persistent differences between the total fertility rates in the North and South of Europe. But a further interpretation suggests that this gap signals a life script rigidly and persistently normed by age in both the Mediterranean and the North-Eastern regions. The anomaly of the life-schedule in the Mediterranean and (to a lesser extent) Southern continental regions is well known: very late home-leaving is closely associated with union-formation, and a certain propensity to complete (even though not so rapidly) the trajectory by having a child. The rigidity of this schedule stems not so much from its requirement of delayed home-leaving as from its imposition, because of long-standing cultural models, of an overlap mainly between the two first steps. This is confirmed by the fact that the coefficients of variation (CV) calculated on the three rates of transition (home-leaving, couple formation and birth of a child) are again lower in the two regions of the South. But equally rigid and norm- driven – albeit in a different sense – is the age script for NorthEastern Europe, characterized by precocious and generalized home-leaving. Consider females aged under 30: in the Mediterranean macro-region, one in every two has left home (49 per cent), but in the North-Eastern quadrant

214

Social Vulnerability in Europe

area only 7 per cent (a very small minority) have not done so. The Gini index calculated on the frequencies of family exit19 confirms that uniformity of behaviour, at the ages of both 30 and 35, for both males and females, is much greater in the North-East. Paradoxically, the postponement of home-leaving by a large proportion of young Southern European adults has made this choice more flexible, more ‘considered’ and less mandatory. Precisely this rigidity of the North European schedule for home-leaving heightens the vulnerability of those who infringe the norm by leaving home earlier.20 If the distinctive feature of the Southern regions is the influential role of the family heritage, the binding sign of the Northern regions is therefore a kind of ‘intrinsic clock’ that rigidly states the age of entry into adulthood. Table 8.10 synthesizes the results of statistical models similar to those reported in Tables 8.6 and 8.7, calculated not for countries but for the five macro-regions. It substantially confirms the mix of factors influencing vulnerability shown in those tables, at the two different levels of secondorder criticality. Some factors affect susceptibility to less critical situations: income, educational qualifications and career advancements, which add to the pure and simple process of ageing. Other factors instead exacerbate exposure to situations of stronger criticality: difficulties of labour-market entry, the lack of family support resources and a too precocious entry into adulthood. Instead, exposure to both levels of severity is influenced by the formation of a union (which exerts a vulnerability-reducing effect), and by the birth of a child (with a vulnerability-multiplication effect). Moreover, Table 8.10 provides confirmation of the weakness of the proposed partition among three macro-regions lying to the north of Reher’s boundary between weak and strong families. The few differences that emerge – though worthy of some note – suggest that there are no strong distinctions among these three models, but rather gradations along a continuum of culturally homogeneous ones. But Table 8.10 also more clearly recapitulates the distinct features of the three principal European macro-regions divided along the North/South gradient by two virtual lines roughly following the forty-seventh and fortysecond parallels. What distinguishes the three macro-regions of the Centre-North from those of the South is the intrinsic clock regulating the age of entry into adulthood. Any further shift forward of the age of transition, even if slight, heightens exposure to the risk of falling below the poverty threshold. That efficient workfare and welfare-to-work systems may make this criticality temporary and not chronic, so that young home-leavers can more easily cope with it, is another matter. It only confirms that no social model is flexible in all its mechanisms, just as it is not rigid in all of them. Reher’s border between weak-family and strong-family regions is in reality crossed and disrupted by a feature that unites the two macro-regions of the

Table 8.10

Areas of influence of the vulnerability multipliers Lesser criticality Men Ne

Early transition Low income Low education No Partner Children Partner no income Precarious job/ joblessness Rented home Mortgaged home Low family income

Nw

Ce

Women Su

Me

Ne

* ** *

Greater criticality Men

Nw

Ce

Su

**

*

**

** **

**

** **

Me

Ne

Nw

Ce

**

**

**

*

** **

Women Su

Me

Ne

Nw

Ce

Su

**

**

**

**

**

* **

*

**

** **

**

*

**

**

** *

Me

**

* **

**

** **

*

*

*

Significance: ** < .001; * < .01. The symbol indicates a reducing rather than multiplying effect Source: ECHP, authors’ own calculations.

*

**

**

*

**

**

216

Social Vulnerability in Europe

North-West and the continental South; a feature that has to do with the work and family ‘career’ patterns with which people shape their life schedules. The vulnerabilizing impact of the birth of a child is more striking in these European ‘Midlands’ than elsewhere. And here more than anywhere else, the clash between the drastic deregulation of the labour market, on the one hand, and the still widely persisting traditional gender division of roles and the MBW model, on the other, generates the risk of poverty.21 However, the two regions of Southern Europe (continental and Mediterranean) also have a feature in common, and it has to do (consistently with Reher’s bipartite typology) with the crucial role performed by the family heritage in reducing vulnerability – a relational capital that acts mainly through the direct transfer of monetary resources, but also through more indirect channels, like help with mortgage repayments (in Spain) or the transfer of usufruct title to a family-owned property (in Greece and Italy). Finally, there is a feature that distinguishes the Mediterranean regions from those of the continental South: not the support provided by the family (a social practice present in both macro-regions), but rather the fact that such support is almost the only factor reducing susceptibility to poverty, so that it comes to resemble a ‘monopoly’ in social regulation.

Notes The interpretation of the results in this chapter derives from a joint effort by the two authors. The first and fourth sections were written jointly by both authors, the second and third by Alessandro Rosina and the fifth and sixth by Giuseppe A. Micheli. Statistical analysis was performed by Stefano Mazzucco. 1. Unfortunately, the ECHP data do not give information on the reasons for leaving the parental home in the previous year. We have therefore been forced to identify people who left home for reasons of study indirectly, doing so by considering those who stated they were students before and after leaving home. 2. Preliminary analyses have shown that vulnerability is higher in the phase immediately after leaving home and then tends to diminish rapidly. 3. A lack of savings in this phase of the life-course might in some cases also result from a strong propensity to invest, and not necessarily to financial difficulties. Moreover, the indicator gives a ‘snapshot’ of the situation in one specific year without clarifying whether assets already exist. 4. Income is considered by using a dichotomous index that shows whether the family is above or below the poverty threshold, this being defined as 60 per cent of the national median income. Housing deprivation is measured by considering three different aspects of housing: (a) the ratio between the number of rooms in the dwelling (excluding the kitchen) and the number of family members (b) the presence of a bath or shower and (c) the presence of a toilet. These three aspects are aggregated so that if deprivation exists on at least one of these aspects (overcrowding, lack of a toilet or lack of a bath or shower), then the family suffers from housing deprivation. The vulnerability indicator is thus constructed by compounding the indicators of poverty and housing deprivation. It has four modalities: (a) no deprivation (family above the poverty threshold and with no housing

Vulnerability of the Young on Leaving Parental Home

5.

6.

7.

8. 9.

10.

11.

12.

217

deprivation), (b) housing deprivation but income above the poverty threshold, (c) income below the poverty threshold but no housing deprivation and (d) income below the poverty threshold and the presence of housing deprivation. Although it has been emphasized that there is still no measurement method that serves as a standard when using the sociological concept of norms specifically in connection with the life-course (Billari and Micheli, 2001), one cannot but agree with Modell (1997: 282) when he argues that the age norm ‘has proven so attractive that social and behavioural scientists have tended to accept it with inadequate specification and empirical underpinning’. Because the sociological literature conceives age norms as ‘prescriptions or proscriptions about behaviour in the form of “should” and “should not” ’, supported by consensus and enforced through various mechanisms of social control and social sanctions (Settersten and Mayer, 1997), it is important to bear in mind that the greater the uniformity of acceptance of a rule (which may concern either the age of an individual at a specific event, or the sequencing between a series of events), the greater the risks attendant on breach of that norm. The connection between uniformity of acceptance of a rule and that rule’s vulnerabilizing capacity if it is breached, explains why, when describing Table 8.10 the schedule for entry into adulthood in North-Eastern Europe is termed ‘rigid’. The longitudinal nature of the ECHP enables study of the condition of young people in the earliest years after leaving the parental home, relating it to individual characteristics and to those of the family of origin. Analysis is conducted on the sample described at the beginning of the third section. We stress the difference between the terms ‘vulnerability’ and ‘criticality’ (see Figure 8.1). By ‘vulnerability’ we mean the exposure of an individual to secondorder criticalities (such as no savings or temporary poverty) as a consequence of a first-order criticality (like leaving home). By ‘criticality’ we mean a critical condition of an individual, marked by a varying level of severity, that can be considered a symptom of vulnerability, i.e. exposure to some modes of failure. In defining criticality we borrow the technical concept of ‘criticality matrix’ – a representation of failure modes along with their probabilities and severities – from operations research and engineering. Notwithstanding the ambiguity of this indicator, of which we are aware. It is likely that the interaction between age and the presence of a child has a significant multiplicatory effect. A large body of literature reports that precocious pregnancy is a major factor in impoverishment. This explanation is consistent with the fact that personal income does not have a significant direct effect, which may therefore be substituted by the status effect for which the mortgage is a proxy. It is true that the influence of income and education is net of age, but even thus standardized it includes a reference to the individual’s age. We may simplify the difference between the two models of vulnerability by schematizing it as follows: while it is younger age on leaving home that multiplies the risk of poverty, regardless of income and education, slight vulnerability derives precisely from a lower level of income or education at whatever age the person leaves home. This description ‘by a sum of negatives’ is strictly accurate only for young males. In the case of females, it is attenuated by the distinctiveness of the Greek case, where, as we have seen, the ‘job and qualification’ combination strongly conditions the degree of vulnerability.

218 Social Vulnerability in Europe 13. In regard to family formation, we have hitherto dealt only with its effect on poverty. In the case of lesser vulnerability, one can tentatively detect a distinct pattern: the presence of a child seems to penalize Spanish and Italian males, but it above all penalizes young women in the English-speaking and Germanspeaking countries. 14. ‘Including Iceland but not Finland’ (ibid). 15. ‘Here we are referring mainly to Portugal, Spain and Italy, although at times in this text Southern France and Greece are included as well’ (ibid). 16. The macro-regions re-aggregate the regional clusters as follows: the NorthEast comprises Denmark, Finland, Eastern Germany; the North-West includes the United Kingdom, Ireland (included because of geographical contiguity), Central-Eastern France, the Île-de-France; the Centre comprises Belgium, the Netherlands, Luxembourg, West and Northern Germany; the South comprises Eastern and Central-Western Spain, Portugal, South-Western and Mediterranean France, North-Western and Central-Eastern Italy, Austria and Southern Germany; the Mediterranean area comprises Spain, Southern Italy and Greece. 17. With a certain convergence, from 1994 onwards, on European values by males in the South and a further postponement of Mediterranean ones. 18. The coefficients of variation are roughly calculated among the three rates of transition in t1, t2 and t3. 19. The Gini index, normalized between zero and one, is calculated for every macroregion on the Bernoulli variable ‘left home versus not left home’. For values close to one it indicates strong heterogeneity of choices; for lower values it indicates greater homogeneity. 20. There is no contradiction between the deregulation of the sequence of the events in the transition to adulthood (reported for North Europe when describing Table 8.10) and the schedule rigidity deriving from the precocious home-leaving age. If in fact the average age of a transition is very low, being the weighted sum of individual ages, all or almost all of these will be very close to the lower limit: which means limited variability and therefore close uniformity in behaviours (the same applies, note, to a very high average age, which squeezes and uniforms behaviours upward). 21. This applies mainly to the continental South, where the presence of an incomeless partner significantly increases the poverty risk.

9 Social Vulnerability: A Multidimensional Analysis Costanzo Ranci and Mauro Migliavacca

Social vulnerability can be defined as a life-situation characterized by a multi- dimensional combination of factors of disadvantage and advantage, of inclusion and exclusion. Its distinctive feature is that weak and unstable integration in the main mechanisms of resource distribution in contemporary society places people in a situation of uncertainty and high exposure to the risk of poverty and, eventually, of social exclusion. Castel (1995, 2003) describes vulnerability as a dynamic condition marked by the transition from inclusion to exclusion and vice versa. It differs from permanent poverty and social exclusion in that these are static situations characterized by chronicity. Precisely because of its often temporary nature, vulnerability is difficult to capture: it is only apparent when it no longer exists and it has turned into a more severe situation, although most of the time it has been absorbed back into ‘normality’. The analysis that follows describes various aspects of social vulnerability and shows the structuring of the multiple elements of which it is constituted. The focus ranges across the family structure, exposure (temporary or permanent) to poverty, housing conditions, the family/work system (with particular attention paid to the presence of unemployed and temporary workers in the household), the ways in which work and childcare are reconciled and the presence of dependent persons in households in economically compressed situations. All of these aspects will be considered together and in their reciprocal combination and interdependence. We also reconstruct the geographical distribution of social vulnerability across Western European, its distribution by family type and social class, and its association with poverty.

A map of social vulnerability Table 9.1 summarizes the manifold dimensions of social vulnerability. Corresponding to these dimensions are as many new social risks. Three distinct groups of the population are identified with each risk. The first group 219

220

Social Vulnerability in Europe

comprises the population not affected by a risk factor, the second includes the population that, to various extents, is moderately but significantly exposed to the risk factor while the third group, of smaller size, comprises persons for whom the exposure to the risk is so high that it significantly deteriorates their living conditions. Economic vulnerability This is defined as a situation characterized by the presence (with varying occurrence) of episodes of falling below the income poverty threshold.1 The data (see Table 9.1) show that 5 per cent of European households (4.8 per cent of individuals) are permanently poor (being poor at least four times within a time span of 5 years), while 7.8 per cent of them experience recurrent poverty (either two or three episodes of poverty in 5 years) and 9.8 per cent transitory poverty (only one episode in 5 years). 2 Moreover, 77.5 per cent of European households are not affected by economic vulnerability. The transition to vulnerability, and thence to permanent poverty, takes place in parallel with an evident deterioration in living standards. Transitory poverty is associated with a 58 per cent reduction in the average per capita income of family members, with a doubled probability of experiencing serious economic difficulties (while 12.5 per cent of non-vulnerable households find it difficult to reach the end of the month, 27 per cent of vulnerable households do so) and a greater likelihood of experiencing major cutbacks in everyday consumption (21 per cent of non-vulnerable households, 39 per cent of vulnerable households). Permanent poverty is associated with tripled probabilities that all these things will happen, and with a reduction of the average per capita income to 28 per cent of the income received by non-vulnerable households. Therefore, if there is a huge gap in the material conditions between vulnerable families experiencing transitory or recurrent poverty and families in permanent poverty, economic vulnerability is very widespread through Western Europe characterizing 18.5 per cent of families facing a significant reduction in income and repeated episodes of poverty. Housing risk Housing conditions provide a clear indication of people’s living standards. At the same time they constitute an important dimension of social vulnerability in so far as they concern the material bases necessary to develop stable familial and social relationships (Edgar, Doherty and Meert, 2002). Despite the great importance of the home in the lives of individuals and families, research on poverty and equality has so far paid little attention to this dimension. It has at most included some indicators of housing conditions within the broader scales used to measure material deprivation (Whelan and Maitre, 2004, 2007a). Yet the housing conditions of European citizens (see Chapter 6) indicate that situations of vulnerability or serious deprivation

Social Vulnerability: A Multidimensional Analysis

221

are quite widespread. When overcrowding and the physical quality of the dwelling are considered,3 it emerges that 3.1 per cent of European households have serious housing problems (see Table 9.1). Since these households have often many members, the share of European citizens suffering severe housing deprivation rises to 5.5 per cent. Moreover, one tenth of European households (and around 13 per cent of individuals) experience a less serious situation characterized by poor- quality accommodation (standardized to take account of the variability of geographical contexts) or by overcrowding (calculated by considering the number of people per room). Overall, about 13 per cent of European households, and 18 per cent of individuals, live in substandard accommodation. As is to be expected, housing deprivation does not overlap with poverty, although the latter is certainly influential in worsening housing conditions. In fact, while the share of households with serious housing deprivation quadruples when they are permanently poor (rising from 2.4 to 10 per cent), fully 75 per cent of households in permanent poverty do not report housing deprivation. Moreover, situations of overcrowding and physical deterioration of the home do not vary significantly between temporarily and permanently poor households (the proportions remain at around 5 per cent for overcrowding and around 10 per cent for physical quality). On the other hand, while overcrowding is associated with a per capita income that is 20 per cent lower than the income of families with standard dwellings, the per capita income associated with poor physical quality of the dwelling or with serious housing deprivation reduces by almost one half (see Table 9.1). Overall, therefore, while economic vulnerability and housing vulnerability are two interdependent dimensions, in many cases the two risks are not copresent but balance each other out. Work vulnerability The analysis in Chapter 5 reconstructed the family/work system by jointly considering the occupational positions of all the active members of the household. The same analysis (see also Table 9.1) also identified the presence of both unemployed and temporary workers. It emerged that 18 per cent of European households have a critical relationship with the labour market (if individuals are considered, the proportion increases to 21 per cent). In 3.1 per cent of households, all the members active in the labour market are unemployed and 6.2 per cent of households have only temporary workers (in some cases they are co-present with unemployed members). Overall, therefore, 9.3 per cent of households do not have any active member in regular employment. A similar percentage (9.1 per cent) consists of households in which the presence of an active member in precarious employment, or in unemployment, is accompanied by the presence of a stable worker: this situation is characterized by potential vulnerability countered by the presence of a stable and guaranteed worker.

222

Social Vulnerability in Europe

The relationship between work vulnerability and poverty is a complex one. Previous analysis of this issue has considered only the position of individuals in the labour market, examining the features of temporary workers with respect to permanent ones. In this analysis, however, the attention shifts from individuals to households. Our hypothesis is that the impact of unemployment or work precariousness on poverty depends closely on the household structure, and specifically, on the presence of stable workers in the household. First, unemployment increases the risk of poverty much more than does temporary employment. Second, the presence in the household of temporary (and eventually unemployed) workers has a significant effect on its exposure to poverty only in the absence of a stable worker. While the average per capita income in these latter households decreases by around one fifth in comparison with families combining stable and temporary workers, the likelihood of falling into poverty increases from 4.6 to 7 per cent. The same probability increases fourfold for households with only temporary workers, and sixfold for households with only unemployed members, where the per capita income decreases, respectively, by one third and by almost a half (45 per cent). Overall, therefore, the family/work system contributes substantially to varying the exposure of individual workers to poverty. Occupational situations characterized by unemployment, and to a lesser extent by instability, increase the exposure to poverty. Nevertheless, since 50 per cent of these vulnerable positions are co-present in the same household with stable occupational positions, the impact on poverty is much less marked. This phenomenon does not differ significantly among countries. We may assume that in such families the weak position of one member in the labour market is matched by a good economic situation of the household because another member is in permanent employment. The occupational structure of families therefore performs a decisive role in decreasing the negative impact of both unemployment and temporary employment. To sum up, temporary employment exposes individuals to poverty only in the absence of another worker with regular employment in the same household. In light of these results, it is therefore not correct to claim that unemployment or temporary employment are major factors in the impoverishment of families. Although these phenomena are widespread in some areas of Europe, in many cases they are absorbed within family structures and offset by the crucial economic role still performed in numerous households by the male breadwinner. It is the crisis of the latter figure, rather than the unemployment or precariousness of the woman or the children, which constitutes the severest risk factor. Thus, in the future, the more unemployment and precariousness affect breadwinners, the more they will contribute to enlarging the borders of economic vulnerability.

Social Vulnerability: A Multidimensional Analysis

223

Difficulty of reconciling work and childcare The fourth aspect of vulnerability concerns the presence of serious problems in balancing work with caring for children aged less than 13 years. Although reconciling work and childcare has been long a crucial, and mostly unsolved, problem even in industrial societies, it has recently grown in importance owing to the increase in the female activity rate and the concomitant static fertility rates in many European countries (Crompton, 1999; Esping-Andersen, 2002; Jacobs and Gerson, 2004). In the last decade, while the gender activity gap has decreased significantly in most of the European countries (including the Mediterranean ones, where the female activity rate was traditionally very low and persistent), there has been no equivalent reduction in the activity rate gap between women with children and women with no children (see Chapter 1 for further details). The extent of reconciliation problems is generally measured by the fertility gap, which is the ratio between the number of children desired and their actual number (Sleebos, 2003; d’Addio and d’Ercole, 2005). Since this information is not available in the ECHP database, here we use a different indicator applicable only to families with children aged under 13 (an age threshold defined by the survey itself). This indicator considers the situation of these families as regards childcare responsibilities (if assumed by one parent at least, or by persons external to the household) and the employment situations of the two parents (in work/seeking work/inactive). It is assumed that situations characterized by trade- offs are those where childcare is wholly undertaken within the family and simultaneously one of the parents does not work. This is a potentially misleading figure, because this group comprises some families in which the inactive position of one of the parents (in most of cases this is the mother) is a result of moral choices or conditions (opportunities or constraints) not directly connected with looking after the children. Nevertheless, the fact that in many European countries the activity rate of women of fertile age (25–49 years) without children is now very close to the activity rate of men aged 25–49 years suggests that the main factor reducing the female activity rate in this age range is the presence of small children. Regardless of whether or not this is the result of a deliberate choice, the fact remains that such families find it impossible to combine parental responsibilities with work. There is also evidence (see Table 9.1) that difficulty in reconciling work and childcare is a source of poverty. In the ECHP database, families in this situation represent 7.4 per cent of all households, 29 per cent of households with children aged under 13, and almost half (46 per cent) of households that internalize care of the children. A smaller group of families – 1.1 per cent – is characterized by the fact that the non-working partner is in search of employment. The data on the risk of poverty and on average income confirm that families in which one of the parents does not work and looks after the children have greater economic

224 Social Vulnerability in Europe Table 9.1 Share of households and individuals in social vulnerability conditions, by risk factor

Dimensions of social vulnerability No poverty Temporary poverty Recurrent poverty Permanent poverty Standard housing conditions Overcrowding Poor housing quality (below national standard) Dwelling with no basic elements No active members Only permanent workers Permanent and temporary/ unemployed workers Only temporary (with unemployed) workers Only unemployed workers Children aged 12 or less ⫹ both parents working Children aged 12 or less ⫹ one parent not working. not looking after children Children aged 12 or less ⫹ both the parents working and looking after children Children aged 12 or less + one parent not working and looking after children Children aged 12 or less ⫹ one parent searching work and looking after children No dependent persons Hampered persons. No compressed consumption Moderately hampered persons. compressed consumption Severely hampered persons. in very poor health. compressed consumption

Share of households

Share of individuals

Incidence of poverty

Average per capita income 2001

77.4 9.8 7.8 5.0 86.6

78.2 9.5 7.5 4.8 81.8

– – – – 8.4

17.261 9,979 8,072 4.940 15.779

4.1 6.2

7.8 4.9

16.1 21.2

12.168 8,474

3.1

5.5

29.8

8,492

29.8 51.8 9.1

12.1

15.4 4.6 7.0

12.505 17.655 14.140

6.2

6.3

20.4

11.733

3.1 5.9

3.0 –

30.3 2.7

9,742 16.758

10.0

13.070

2.3

5.5

–

5.3

14.472

6.3

–

16.7

10.370

1.1

–

26.2

8.980

85.7 7.1

9.4 6.8

15.202 14.635

2.3

23.2

7.890

4.9

18.4

8,682

Source: ECHP, authors’ own calculations.

Social Vulnerability: A Multidimensional Analysis

225

difficulties: the risk of poverty increases threefold, and the average per capita income diminishes by 29 per cent. Presence of dependent persons in the family The last vulnerability factor concerns the presence of major care problems in the household because of the presence of a dependent member. The ageing of the population, together with advances in healthcare, not only contributes to the general increase (in absolute numbers) of people with disabilities but also extends their lifespans. Yet physical and psychic dependency gives rise to considerable difficulties, both economic (because of the costs of care) and organizational (they require constant and often intensive assistance). Most families with dependent members suffer from strong compression of their living standards because of the difficulty of providing care and earning adequate incomes at the same time. The capacity of families to resolve these problems depends on the level of public welfare support for dependent persons (Österle, 2001) and on the employment conditions of the family members (two aspects that are obviously interdependent). The dependency index used in our analysis combines a subjective evaluations of both dependence and the state of health. Furthermore, we distinguish between households with dependent persons experiencing high compression of their consumption level (no holidays, no inviting friends for dinner, no meat or fish on a weekly basis, inability to afford new furniture) and families with no such compression (see Chapter 7). On the base of this information (see Table 9.1), 14.3 per cent of European households have at least one member who is severely restricted in her/his routine functions, with a high proportion among the elderly population. Only half of families with a dependent person do not have to cope with a high compression of their living standards. In total, households with dependent members and material compression amount to 7.2 per cent. Dependency is a major factor in impoverishment (Grammenos, 2003). The risk of poverty increases by more than 200 per cent for households with a dependent member, while per capita income diminishes by 49 per cent. As shown in Chapter 7, this proportion remains the same even when controlling for the age of the dependent person. Vulnerability emerges from these data as a widespread phenomenon, in a variety of aspects, none of which appears to be more widspread than the others. This map of social risks identifies the principal problems that cause significant material difficulties for a large proportion of European households. It does not show situations of severe hardship, but rather an area of social fragility that generally includes middle- class groups usually considered to be protected against the risk of poverty and social exclusion. Only for a small minority of households, in fact, does this fragility translate into real hardship, as shown by the data on the situations of greater deficit, which invariably affects around 3 per cent of households.

226

Social Vulnerability in Europe

The deficit of resources in families affected by vulnerability, however, should not be considered as an index of material deprivation (Whelan and Maitre, 2004). While, on the one hand, the recurrence of poverty or the shortage of housing space are social handicaps, on the other, the severity of such deficits should be evaluated by considering the whole family situation and time perspective. In many cases, in fact, the weakness of one family member is compensated by the work stability of other members. In others temporary problematic situations arise that are rapidly and definitively overcome, or situations of distress indeed emerge but do not prevent households from maintaining a balance between resources and needs by reducing their consumption. The multiplicity of resource distribution mechanisms typical of Western European societies is reflected here in the manifold arrangements by which households deal with specific risks or troublesome permanent situations. Moreover, the ability to combine multiple resources and different social roles in the same household plays a crucial role in the strategies adopted to face social risks. Evaluating the gravity and magnitude of social vulnerability therefore requires account to be taken of the multiple risk factors considered here. This aspect is examined in the next section.

A synthetic index of vulnerability The European population appears to be marked by a wide range of vulnerability factors. Only in extreme situations are these factors sufficiently strong to produce a situation of social exclusion or severe material deprivation. One specific risk, in fact, may be balanced by a good position along the other dimensions of social vulnerability. It is therefore not only the severity of the risk itself, but also the accumulation or combination of several risks, that determines the degree of social vulnerability. Hence social vulnerability takes the form of a situation characterized by serious deficits on one or more dimensions of social risk. It is precisely this interweaving among the various risk factors that we will now discuss. What are the relations among the various social risks? To what extent do they cumulate or compensate for each other? The analysis that follows highlights a perhaps unexpected finding: there are scant overlaps among the different dimensions of vulnerability considered. A synthetic index of vulnerability was considered by assembling the numerous dimensions of social vulnerability.4 This index was calculated by means of the following two different procedures. ●

A cumulative procedure was used to consider the different degrees of vulnerability on each item and thus construct a multi-dimensional cumulative index,5 from which it emerges that 47.9 per cent of European households

Social Vulnerability: A Multidimensional Analysis

●

227

have at least one occurrence of vulnerability. However, the strongly accumulated component is very circumscribed including only 8.3 per cent of households. This analysis therefore shows that vulnerability is very widespread. But the index does not show how the different dimensions interweave and balance each other out. These linkages were considered by using a categorical index that analysed the extent to which the different dimensions of vulnerability combine with each other. However, analysis of the combinations shows scant interweaving: 10 per cent of households are vulnerable on two dimensions, only 2 per cent are vulnerable on three or more, while 28 per cent are vulnerable on only one. Moreover, 7.8 per cent of households are affected by one or more severe risk factors. Therefore vulnerability only rarely tends to cumulate, while in the majority of cases the pattern is the copresence of one situation of difficulty with numerous other situations of non-vulnerability.

Combining the two criteria identified – the accumulation and combination of the different dimensions of vulnerability – yields a multi- dimensional index that identifies the dimensions of social vulnerability, shown in Table 9.2. Table 9.2 Multi-dimensional index of social vulnerability (percentage of the total of households and individuals)

Non-vulnerable Income vulnerability Work vulnerability Housing vulnerability Dependence with severe material compression No reconciliation of work and childcare Low multi-vulnerability* High multi-vulnerability** Severe hardship*** Total *

Households

Individuals

52.2 8.7 8.1 5.1 2.7

49.5 7.0 9.4 6.9 2.6

2.9

3.2

8.8 3.5 8.0

11.0 4.0 6.4

100

100.0

This category comprises situations with more than one element of vulnerability and totally scoring 3 or less. ** This category comprises situations with more than one element of vulnerability and totally scoring more than 3. *** This category comprises situations characterized by one or more elements of vulnerability scoring 3 points. Source: ECHP, authors’ own calculations.

228 Social Vulnerability in Europe

The following three areas of vulnerability with very different characteristics emerge from Table 9.2. 1. The first area is characterized by moderate vulnerability, which is the presence of households that are vulnerable in only one dimension and compensate their exposure to a single risk with other more favourable factors. In all, 28 per cent of households are in this situation. This slight form of vulnerability is distributed across all the dimensions of vulnerability considered. 2. The second area – low or high multi-vulnerability – is characterized by the presence of interweavings among several risk factors, but without a marked accumulation of disadvantages. The households in this area are exposed to two or more risk factors, but without this giving rise to severe deprivation. The area therefore identifies a situation of social vulnerability that derives from a very wide range of dimensions. This area comprises 12.3 per cent of European households. 3. The third area – severe hardship – is characterized by the presence of one or more very high risk factors: all working members of the household are unemployed workers, chronic poverty for at least 4 years in a 5 years time-span, severely substandard housing, high care needs in the households creating a deep material compression or no reconciliation between working and childcare as one of the parents looks after her/his child/ children and is looking for work at the same time. The presence of one or more of these factors heavily penalizes the household’s standard of living. This group includes 8 per cent of households. It therefore emerges from this analysis that around half (52 per cent) of European households are not affected by any risk factor and can therefore be considered as ‘non-vulnerable’, while households that exceed a threshold of serious vulnerability – characterized by the presence of at least one severe hardship condition – represent 8 per cent of the total. Between these two extremes lie 40 per cent of households, which are exposed (with different intensity and extension) to one or more of the risk factors considered. Around half of European households, therefore, are not affected by any of the social risks considered. All the other households experience difficulties, but they assume the character of severe deprivation for only one household in every eight cases. Vulnerability is therefore distributed across a very broad population. Only in a small number of cases is there a cumulative effect which pushes a minority of this vulnerable population into a situation of severe hardship. The area of social vulnerability is therefore mainly characterized by certain risk factors, balanced by other aspects of material life that are not problematic. It thus identifies life-situations characterized by uncertainty that are unlikely to degenerate into hardship, but that instead appear to be

Social Vulnerability: A Multidimensional Analysis

229

constantly liable to deteriorate. A broad array of problems emerges in this area: an inadequate flow of income, sometimes recurrent; overcrowded or substandard housing conditions; precarious employment or unemployment of one person; difficulty of reconciling work and childcare; or the presence of a dependent person requiring constant care in a family with high compression of consumption. Vulnerable households and individuals seek to achieve an often precarious balance among these interweaving difficulties. Thus delineated is a social area that coincides neither with households with normal living standards nor with households overloaded with deficits. It is a social area that more than others is structurally affected by the economic cycle, by trends in the housing market and changes in welfare policies and that presents weaknesses that, in adverse situations, may generate more severe hardships. Economic and financial crises may in fact have a strong impact on this population, worsening the material conditions of these households, weakening their capacity to deal with difficult situations, and exposing them to a higher risk of poverty and social exclusion. But even in ordinary situations moderate exposure to social vulnerability generates insecurity as a consequence of uncertain access to crucial resources (employment, income or housing) or difficulties in combining tasks (care responsibility with material sustainability): this is a weakened social integration that constitutes a problematic situation in itself and not only a factor responsible for a possible deteriorated social position. Below the area of social vulnerability lies a further area in which the negative outcomes of the fragilization process are apparent. Here the various vectors of vulnerability converge and most of the time accumulate. And here social distress assumes its most evident guise of material deprivation affecting and constraining the entire management of material life: not only the family budget, but also housing and work strategies and care for dependent members. If, on the one hand, this area is the possible (although unlikely) outcome for all households affected by social vulnerability, on the other, it is a threshold below which a qualitative shift takes place: materially vulnerable households become entirely bereft of material resources and the accumulation of disadvantages or the severity of the deficit are their distinctive and often irreducible features.

Interactions among risks One of the crucial issues raised by the analysis above concerns the connections among the various risk factors considered. Table 9.2 highlighted, in fact, the presence of a broad area of moderate vulnerability characterized by the presence of a single risk factor largely compensated by good performances on all the other aspects considered. Moreover, the area of multi-vulnerability, where interweavings and cumulative effects among different vulnerability factors are generated, is considerably smaller than the

230 Social Vulnerability in Europe

first one, although it still comprises 12 per cent of families and 15 per cent of individuals. In this section we ask whether the occurrence of a specific vulnerability factor increases the exposure of households to other vulnerability factors, producing a cumulative effect leading to situations of multi-vulnerability or social hardship. While we consider social vulnerability as a social condition not associated with a high probability of lapsing into a state of serious hardship, it is nevertheless useful to observe whether some vulnerability factors more frequently generate cumulative negative effects than others. A logistic regression analysis was carried out in order to estimate the statistical probability of each vulnerability factor determining other or more severe conditions of vulnerability. We estimated, for example, the probability that a recurrently poor household might experience vulnerability in regard to employment or housing conditions, as well as the probability that the precarious employment of a household member might determine the occurrence of poverty or housing deprivation. Each vulnerability factor was therefore considered as a specific cause that might determine, according to an estimable probability, a condition of vulnerability characterized by the co-presence of other vulnerability factors or by the onset of severe hardship. Social risks were therefore evaluated in terms of two different effects: to what extent they determine other social risks (the horizontal dimension) and to what extent they determine a situation of severe hardship (the vertical dimension). As stated above, 12 per cent of European households experience a multiple vulnerability characterized by the association of two or more risk factors: this is an area characterized by a deterioration of living conditions, which simultaneously affects different life-spheres and different systems of resource acquisition. What vulnerability factors seem most likely to drive households into this situation of multiple vulnerability? The logistic analysis highlighted the strong impact of housing deprivation and difficulties in reconciling work and childcare (see Table 9.3). Households with severely compromised housing conditions are approximately three times more likely than a household with adequate housing to exhibit other features of moderate vulnerability, while those with moderate housing deprivation have a probability of 73 per cent. Similar probabilities also concern households reporting problems in reconciling work and childcare. Precarious employment or unemployment, however, have less strong impact: if a household comprises only unstable (or unemployed) workers, the probability increases by 100 per cent (compared with households of stable workers), but it increases by only 20 per cent if the unstable workers cohabit in the same household with stable workers. Finally, income poverty matters even less, considering that the probability of accumulating other vulnerability factors is 45 per cent for permanently poor households and 40 per cent for temporarily poor households. Overall, therefore, multiple vulnerability is frequently associated with problems related to the low quality of the

Social Vulnerability: A Multidimensional Analysis

231

Table 9.3 Logistic regression of the individual risk factors on the multiple vulnerability index 95 Confidence Interval for Exp(␤)

␤

Std. Error

Sig.

Exp(␤)

Intercept Permanent poverty Temporary poverty Non-poverty

⫺1.199 0.375 0.335 0

0.014 0.064 0.034 –

.000 .000 .000 –

– 1.455 1.398 –

– 1.283 1.308 –

– 1.65 1.494 –

Intercept Only temporary workers

⫺1.008 0.716

0.013 0.042

.000 .000

2.047

1.886

2.223

0.04

.000

1.206

1.114

1.305

–

–

Stable ⫹ temporary workers Stable workers

0.187 0

Lower Bound

Upper Bound

–

–

–

0.013 0.182 0.037

.000 .000 .000

3.735 1.729

–

–

–

–

–

⫺0.877 1.006 0.508 0

0.013 0.132 0.048 –

.000 .000 .000 –

2.735 1.662 –

2.113 1.514 –

3.539 1.824 –

⫺0.882 0.115 0.045 0

0.013 0.04 0.054 –

.000 0.004 0.408 –

1.122 1.046 –

1.038 0.941 –

1.213 1.162 –

Intercept Severe housing deprivation Moderate housing deprivation Standard housing conditions

⫺0.947 1.318 0.547

Intercept Childcare prevents working Childcare, not working Reconciliation of childcare and work Intercept Severe dependency Moderate dependency No dependent family member

0

2.613 1.608

5.338 1.859

Note: The multiple vulnerability index includes all the risk factors with the exclusion of the specific risk factor for which the regression coefficients are calculated; the coefficients referring to diverse risk factors are therefore only indicatively comparable. Source: ECHP, authors’ own calculations.

dwelling and the difficulty of reconciling caring for children with work, while employment and income instability generate cumulative effects less frequently. The main reason for the lower potential for multiple vulnerability of employment precariousness and income instability is the possibility of counterbalancing these difficulties with the presence of another household member in stable employment. In effect, two thirds of European temporary workers cohabit with a household member in stable employment. Since occupational precariousness principally affects young people and women (see Chapter 5), its most relevant effect is their greater economic dependence

232 Social Vulnerability in Europe

on household members in stable employment. This delays the autonomization of young people from their families of origin and creates new gender and age discriminations. Overall, therefore, while, on the one hand, occupational precariousness and income instability are the most widespread social risks, on the other, they are offset within households by a combination of different work incomes; but this hampers the individualization and autonomization of young people and women. In contrast, the most vulnerabilizing risks are housing deprivation and the difficulty of reconciling (female) work with childcare. These two conditions are associated not with the occupational and economic conditions of households, but with their capacity to achieve a stable social situation, confirmed by an adequate home and the presence of children. Vulnerability therefore assumes its most dangerous form in the process of social stabilization, when the professional career and earnings must be converted into a satisfactory standard of living and in social reproduction. To be noted, in fact, is that these two social risks arise in the initial phases of adult life, constituting the two ideal final stages of transition to adulthood: accessing to an independent house and having children. These two transitions appear rather difficult today, as shown in Chapter 8, and they expose numerous young European people to accumulated difficulties, such as strong compression of living standards and a reduced organizational autonomy of the households, as well as of individuals. The difficulties of the housing market and the still scant availability of childcare services in many European countries make this situation of vulnerability particularly complex and difficult to cope with. We now consider the incidence of risk factors on severe hardship (see Table 9.4). Here severe hardship is defined as a condition of life characterized by one or more factors of high deprivation or social exclusion: permanent poverty, unemployment of all members of the family, severely sub-standard housing condition, impossibility of work because of childcare responsibilities and the severe physical or psychic dependence of one household member bringing with it high compression of the household’s consumption level. The risk factors considered in our analysis significantly increase the probability of severe hardship compared with situations in which these risk factors are not present: a fact that is entirely predictable, given that situations of severe hardship are likely to be characterized by cumulative difficulties. Compared with households that never suffer poverty, households in permanent poverty are six times more likely to experience severe hardship on other dimensions of social vulnerability as well. Nevertheless, in this case too, the greatest risk factors are housing deprivation (probability 10 times higher than for households with satisfactory dwellings), and incompatibility of work and childcare (probability 7.6 times higher than for households able to reconcile work and childcare), while permanent poverty (probability 6.6 times higher than for never-poor households) and

Social Vulnerability: A Multidimensional Analysis Table 9.4 index

233

Logistic regression of the individual risk factors on the severe distress

95 Confidence Interval for Exp(␤) ␤

Std. Error

Sig.

Intercept Permanent poverty Temporary poverty Non-poverty

⫺2.164 1.887 1.207 0

0.021 0.057 0.038 –

Intercept Only unemployed workers Stable and temporary workers Stable workers

⫺2.106 1.492 0.495

0.021 0.049 0.056

Intercept Severe housing deprivation Moderate housing deprivation Standard housing conditions

⫺1.712 2.630 1.161

Intercept Childcare prevents working Childcare, not working Reconciliation of childcare and work Intercept Severe dependency Moderate dependency No dependency

0

Exp(␤)

Lower Bound

Upper Bound

.000 .000 .000 –

– 6.602 3.342 –

– 5.907 3.101 –

– 7.38 3.602 –

.000 .000 .000

4.445 1.640

4.041 1.470

4.890 1.829

–

–

–

–

–

0.018 0.166 0.040

.000 .000 .000

13.871 3.192

10.012 2.949

19.217 3.455

–

–

–

–

–

⫺1.408 2.029

0.015 0.119

.000 .000

7.609

6.020

9.616

0.906 0

0.050 –

.000 –

2.474 –

2.243 –

2.730 –

⫺1.728 0.509 0.620 0

0.018 0.048 0.060 –

.000 .000 .000 –

1.664 1.858 –

1.516 1.651 –

1.828 2.092 –

0

Note: The multiple vulnerability index includes all the risk factors with the exclusion of the specific risk factor for which the regression coefficients are calculated; the coefficients referring to diverse risk factors are therefore only indicatively comparable. Source: ECHP, authors’ own calculations.

unemployment (probability 4.4 times higher than for stable workers) have less visible impacts. Overall, the main risk factors are therefore more related to social stabilization (the possession of adequate housing and the presence of children with no gender segregation from the labour market) than to the position of people in the labour market or the level of the household income. In the affluent societies of Western Europe, therefore, it is the process of social stabilization that most distinguishes severe distress from normality. It is crossing this threshold that exponentially increases the probabilities of lapsing into hardship and social exclusion.

234 Social Vulnerability in Europe

Vulnerability and household models The family structure in Western Europe is moving toward a balance between individualization and the maintenance of family solidarity. As shown in Chapter 3, the current transition is characterized by the progressive decrease in male breadwinner households (Crompton, 1999). In this process a profound reorganization of the family structure is taking place, with opposite pulls at work. While, on the one hand, individuals are increasingly disentangled from their original family ties and can create their own biographies more freely than in the past, on the other, there is an evident tendency to strengthen family solidarity in an attempt to solve new problems such as the prolonged work precariousness of the young or the growth of the elderly population living alone. What are the types of households that are today most exposed to social vulnerability? We have already pointed out that the family often provides vulnerable individuals with protection and supplementary resources that weaken the negative impact of specific risk factors. The family therefore constitutes one of the most crucial agencies for the social integration and protection of individuals at risk. However, not only is the incidence of individuals without family protection increasing in contemporary societies, but the protective capacity of families varies greatly according to household type. It is therefore crucial to determine which kinds of families are most vulnerable, and which risks most affect families. The most significant finding in this study, in fact, is the high exposure to social vulnerability of male breadwinner households (see Table 9.5). Only 23 per cent of such households do not fall within the area of vulnerability, while 30 per cent of them experience multi- dimensional vulnerability. Also particularly high, as to be expected, is the incidence of work/care reconciling problems in these families. By contrast, dual-income households with small children are less exposed to vulnerability: the probability of being vulnerable is more than twice as high among male breadwinner households as among dual-income households of the same composition. This finding confirms that the impossibility of earning two incomes in a family with small children (to a large extent because of the inactivity or unemployment of the mother) is currently one of the principal sources of vulnerability in Western Europe. For these households, it is the absence of a double income, besides the fact of having small children, that constitutes the main risk factor. A second group of households with high vulnerability consists of lone-parent households (66 per cent of which are vulnerable), households with couple and adult children (51 per cent), and households composed of adults (one parent and adult children) belonging to different generations (52 per cent). While loneparent households are especially exposed to recurrent or permanent poverty, households with adults of different generations (so-called ‘long households’) have higher levels of vulnerability in housing and work. The co-existence of

Table 9.5 Vulnerability index by family type (row percentages) Vulnerabilty index

Housing deprivation

Dependent persons in material compression

Difficulty in reconciling work and care

Moderate multiple vulnerabilities

Intense multiple vulner abilities

Severe hardship

Total

18.7

3.2 4.4 2.7 3.3

6.2 .4 7.1 2.3

– – – –

4.3 4.1 3.2 5.5

1.2 3.7 1.2 3.7

15.3 2.5 9.4 2.7

100.0 100.0 100.0 100.0

4.7

16.3

3.5

.3

–

4.7

.8

1.3

100.0

23.0

3.9

7.4

3.4

.2

21.9

21.1

8.0

11.1

100.0

54.0

3.7

11.1

8.4

.5

3.0

11.8

4.3

3.2

100.0

33.1

15.5

6.8

4.6

.4

1.6

13.2

6.0

18.8

100.0

48.8

5.7

13.1

8.5

2.6

.0

10.2

3.4

7.7

100.0

47.6

7.5

12.2

4.6

3.8

–

8.4

5.1

10.8

100.0

52.2

8.6

8.1

5.1

2.7

3.0

8.7

3.5

8.1

100.0

No vulnerability

Income vulnerability

Single welfare Single work Couple welfare Couple one income Couple two incomes Couple with children ⬍ 25 one income Couple with children ⬍ 25 two incomes Lone parent with children ⬍ 25 Couple with children ⬎ 25 Lone parent with children ⬎ 25

52.7 65.4 64.8 57.1

17.1 12.0 11.6 6.7

68.4

Total

Source: ECHP, authors’ own calculations.

Work vulnerability

– 7.5

236

Social Vulnerability in Europe

two adult generations in the same household is to be considered as a form of protection against social risks taking place in the labour market. A further type of household especially exposed to vulnerability is made by single pensioners living alone. This type has high coefficients of vulnerability in poverty (double the average in the population), in the incidence of disability (as expected) and in the accumulation of severe factors of disadvantage (the incidence is 15 per cent). Finally, households with a lower level of social vulnerability are composed of working singles or working couples without children. This latter type, however, exhibits a significant degree of vulnerability because of the weak occupational position of one of the partners (16 per cent). To sum up, not only the degree, but also the pattern of vulnerability, varies significantly according to the type of household. In general, the household structures most exposed to moderate vulnerability are different from households that are exposed to social hardship. A clear example is provided by the male breadwinner household: while on the one hand this exposes numerous households to forms of moderate multiple vulnerability, it has scant effect on the more severe forms of hardship. The opposite applies to single pensioners: they are not very exposed to moderate vulnerability, while their risk of severe hardship is much greater. In general, moderate vulnerability mainly affects one-income households and multiple vulnerability affects households with children, while severe hardship mainly afflicts two types of household: pensioners living alone and lone-parent households. In this pattern the diversified impact of the main risk factors can be seen: while the crisis of one-income households is of limited impact, it increases substantially when it concerns households with children of school age. Severe hardship instead appears to be closely associated to specific conditions of isolation and loneliness. It is concentrated in two critical phases of the household’s life-cycle: when the children are born and when members of the family reach elderly age. Vulnerability in both of these phases results from the same factor: in the former the risk increases when the couple breaks up, while in the latter it increases when advanced age and loss of the partner coincide. Severe hardship is therefore characterized by the interweaving between income fragility and social isolation. Besides being characterized by the lack of economic means, it is associated with the loss of social and family bonds.

Social vulnerability across Western Europe How is the vulnerable population distributed across the 28 macro-regions of Western Europe? Significant macro-regions differences repeatedly emerged during the analysis. This is an aspect almost taken for granted in a continent historically marked by different historical and social milieux. In this

Social Vulnerability: A Multidimensional Analysis

237

section we look at these differences in order to obtain an overview of the different configurations of social vulnerability across Europe. This analysis is aimed not only at ranking geographical areas in Europe on the basis of specific indicators of social vulnerability (Stewart, 2003), but also at identifying specific territorial profiles within which peculiar configurations of social vulnerability emerge. In addition to identifying more or less vulnerable areas, the purpose of the analysis is therefore to reconstruct configurations of vulnerability that vary geographically. We will consider the following three questions: ● ●

●

in which European macro-regions is social vulnerability most widespread? which European macro-regions have similar profiles as regards social vulnerability? does the distribution of social vulnerability match the standard division of European countries into three, or four, welfare regimes?

Marked differences emerge on observing the overall amount of the vulnerable population. Vulnerability does not significantly pass the 35 per cent threshold in Belgium, Denmark and Finland or the metropolitan area of Paris.6 It exceeds 40 per cent in Portugal, Greece, Southern Italy and Spain. Around 30 per cent of the overall vulnerable population in Western Europe resides in the macro-regions of Southern Europe, excluding North-West and Central-Eastern Italy, and Central-Eastern Spain. Southern Europe is where 39 per cent of the population with severe vulnerability is concentrated. As shown by Figure 9.1, the variability of vulnerability levels is significantly greater in Spain and Italy. This last country comprises macro-regions 55 50 45 40 35 30

Greece

Spain

Portugal

Irland

Italy

UK

Germany

Netherlands

Austria

Denmark

Finland

France

Belgium

25

Figure 9.1 Share of vulnerable population in 13 Western European countries, by macro-regions Source: ECHP, authors’ own calculations.

238 Social Vulnerability in Europe

belonging both to the area where vulnerability is most widespread, and some of the European macro-regions in which vulnerability is close to the continental average. In France and the United Kingdom, the metropolitan areas of Paris and London exhibit lower levels of vulnerability than in the rest of the country, while in Germany the area of greatest vulnerability is the East. Overall, these findings confirm that comparative analysis should use distinctions subtler than the one based on national differences alone. The presence of marked geographical differences draws attention not only to the general amount of the vulnerable population, but also to the different geographical patterns of social vulnerability. A hierarchical cluster analysis was performed with the aim of identifying specific geographical configurations of social vulnerability across Western Europe. In the analysis each social risk was considered not only separately but also in its possible interweaving with other risks.7 As the attention is focused on vulnerability, situations that are characterized by accumulated high risks or severe hardship were excluded. According to the logic of hierarchical cluster analysis, clusters of macro-regions were identified by aggregating together those macro-areas presenting the most similar profiles of vulnerability. We chose the Wards’s method to determine cluster membership. Euclidean distance was adopted to specify the distance to be used in clustering. Furthermore, in order to identify which variables contribute more to the distinctiveness of clusters, a further mean analysis was undertaken. Figure 9.2 displays the dendogram of the hierarchical cluster, while Table 9.6 reports the mean values of vulnerability indices by cluster. This analysis has identified four clusters, consisting of macro-regions with similar characteristics in social vulnerability (see Figures 9.2 and 9.3). Cluster one comprises 12 macro-regions: Belgium, the Netherlands, Denmark, Austria, all the macro-regions of Germany, Finland, CentralEast France, the Île- de-France (Paris) and Greater London. It is characterized by a very modest level of vulnerability. This cluster includes a territorially quite contiguous, central area of Europe consisting of the continental and Northern countries of small size (Belgium, the Netherlands, Denmark and Finland), the large metropolitan areas of Northern Europe (Paris and London) and the contiguous areas of North-Eastern France and NorthWest Germany. This area comprises the population least exposed – compared with the rest of Europe – to specific forms of social vulnerability. It is also the area of Europe already characterized, according to ESPON surveys (2006), by the highest levels of wealth and productivity. The only difficulty for families living in such areas is related to temporary employment (see Table 9.6). Cluster two includes seven macro-regions: South and West France, three in the United Kingdom (excluding Greater London) and two in Italy (located in North-West and Central-East Italy), that have a vulnerability level much lower than the other Mediterranean areas. The cluster constitutes a sort of ring surrounding cluster one and it is characterized by a slightly higher

Social Vulnerability: A Multidimensional Analysis

239

Rescaled Distance Cluster Combine CASE

0

South France

13

West, South west France

14

North UK

23

Central UK

24

West, South West UK

26

North-west Italy

18

North-east, centre Italy

20

Belgium

2

France Ile de France

15

Denmark

7

Austria

1

UK, Greater London

25

West Germany

5

Finland

11

The Netherlands

21

Centre-east France

12

South Germany

6

North Germany

3 4

East Germany Centre, North-west Spain

8

South Italy

19

South Spain

10

Greece

16

Portugal

22

East Spain, Madrid

9

Ireland

17

5

10

15

20

25

Figure 9.2 Geographical clusters of social vulnerability Note: Sweden and Luxembourg are not included in this analysis Source: ECHP, authors’ own calculations.

proportion of vulnerable households (see Table 9.6). The general level of vulnerability in this area is moderate. This cluster is mainly characterized by income instability: the proportion of households experiencing temporary or recurrent poverty is markedly higher than in all the other clusters. The high incidence of poverty also affects the probability of housing deprivation, which is higher here than in cluster one. Furthermore, multiple vulnerability is here higher than in cluster one, showing that social risks are more often intertwined in these areas than in the best- off areas of Europe. Multiple vulnerability and hardship are generally low in this area. Integration in the labour market is generally more than adequate. Cluster three and cluster four include macro-regions located in Mediterranean Europe and Ireland. They congregate in one single bigger cluster that is clearly separated by the other macro-regions in Europe as displayed by the dendogram in Figure 9.2. This confirms that a clear North-South divide is

Table 9.6

Mean values of vulnerability indices by territorial cluster

Cluster

Total vulnerable families

Temporary or recurrent poverty

Temporary work or unemployment

Housing deprivation

Dependent persons in material compression

Difficulties in re-conciling work and care

Multiple vulnerabilities

1 2 3 4

30.7 34.0 43.5 49.4

8.5 10.5 8.7 9.9

8.5 6.3 9.5 8.7

3.6 5.2 6.0 4.8

2.2 2.8 3.2 3.0

2.7 2.5 4.1 4.7

5.4 6.7 12.1 18.4

Total

35.7

9.2

8.1

4.5

2.6

3.1

8.3

Source: ECHP, authors’ own calculations.

Social Vulnerability: A Multidimensional Analysis

241

present in Western Europe as far as social vulnerability is concerned, with the relevant exception of many parts of Italy that are clustered with other continental and Anglo-Saxon areas. Cluster three includes North-Eastern Spain (including Madrid), Greece, Portugal and Ireland. It is an area characterized by high flexibility in employment and widespread housing deprivation. However, in this area the general level of vulnerability is lower than in other Mediterranean countries as is confirmed by the lower level in recurrent or temporary poverty. Finally, cluster four includes Southern Italy, Southern Spain and Northern and West Spain. This is the area with the highest social vulnerability, with the greatest concentration of social risks influencing the material conditions of the population – temporary or recurrent poverty and presence of multiple social risks in the same households. This area includes the most Southern macro-regions of South Europe. In this area around 45 per cent of the population is in a situation of vulnerability. A further risk in these areas is the difficult reconciliation of work and childcare as a consequence of the limited availability of public childcare services. Finally, even material compression resulting from dependency has a high impact in these areas. These results highlight not only that there are diverse vulnerability patterns in Europe, but also that such patterns partly differ within national boundaries. In contrast, several countries and macro-regions display very similar patterns. In this way, we have delineated a social geography of Europe that partly replicates traditional national differences, partly reflects the historical configuration of welfare regimes and partly diverges from these latter. Other models to explain these differences can therefore be considered. Three salient findings emerge from the analysis. First, there is a Central European area that is characterized by high standards of well-being and social security and that comprises the Scandinavian countries and the Northern area of continental Europe, as well as the metropolitan regions of Paris and London. This area concentrates not only the greatest well-being but also the highest degree of protection against new social risks. Second, this analysis confirms the existence of a marked dualism between Northern and Southern Europe: 43–49 per cent of the population in the macro-regions of Southern Europe, in fact, suffers significant vulnerability. The analysis shows that social distress in these areas does not exclusively, or even principally, concern the risk of poverty but rather a stronger interweaving among exposure to poverty, housing risk and precarious employment. In contrast, the richest area of Europe has very low indices of vulnerability: fully two thirds of the population are free from any sort of vulnerability. Third, the spread of social vulnerability does not reproduce the standard taxonomy of three or four welfare regimes in Western Europe. There are

242 Social Vulnerability in Europe

Figure 9.3 Map of vulnerability clusters

several incongruities. In Northern Europe, an area of substantial freedom from the new social risks cuts across the traditional distinctions among welfare systems. This includes not only the Scandinavian countries but also a broad Northern area that extends from the great metropolitan areas of Paris and London to the small countries facing the North Sea: Belgium, the Netherlands and Denmark. It also includes Germany and Austria. What groups these areas together is more their level of well-being and socioeconomic development than the kind of welfare regimes they possess. In Southern Europe some macro-regions – notably North-West and CentralEast Italy – are closer to the centre of Europe than to other Southern areas. Moreover, not surprisingly Spain and Italy show huge macro-regional disparities in the exposure of population to social risks. If the North- South divide is still relevant in Western Europe, more specific divisions cut across

Social Vulnerability: A Multidimensional Analysis

243

these two parts of the continent, driving toward a strong regional differentiation in the living conditions of the population.

Social vulnerability and poverty The data on social vulnerability demonstrate the shortcomings of analyses about inequalities and deprivation predicated entirely on the notion of poverty. Economic poverty, measured on the basis of the income of individuals or families, is a decisive but not unique element for determining the living standards of a population. The incidence of poverty in households suffering severe hardship is equal to 57.8 per cent. In the area of moderate vulnerability, the incidence of economic poverty is only 9.1 per cent. Finally, poverty affects about one fifth (19.8 per cent) of households suffering from multiple vulnerability. Therefore poverty is one aspect of severe hardship, but it is neither the unique nor the most frequent one (Berthoud, Bryan, and Bardasi, 2004; Berthoud, 2004b). Social policies relying on poverty indicators are likely to underestimate the magnitude of social hardship and material deprivation (Whelan and Maitre, 2007a). They fail to consider households suffering from sub-standard housing conditions and problems in reconciling work and childcare. As shown by the logistic regression, more frequently than poverty it is these two latter risk factors which cause severe hardship. If poverty is only a partial predictor of severe hardship, it is not always associated with situations of vulnerability. Only 48.7 per cent of poor families experience severe hardship. Around a quarter (26.5 per cent) of poor households do not experience any other form of vulnerability. In these households the shortage of economic resources is counterbalanced by adequate housing conditions, integration in the labour market (no member of the household is unemployed or in temporary employment) and ability to reconcile work and childcare or to look after a dependent family member without suffering a strong compression of living standards. In these households income poverty cohabits with a normal standard of living. They constitute, therefore, a situation that could be considered as ‘integrated poverty’ (Paugam, 1991, 2000). To sum up, there are various factors that cause a significant proportion of the European population to experience serious difficulties in their everyday lives. A strong attention to poverty has traditionally been paid on the assumption that low income was the main factor in social deprivation (Whelan and Maitre, 2004). However, analysis about social exclusion has already shown that income poverty often combines with other factors of social marginalization (Paugam, 2000; Berthoud, 2004b). From our data it emerges that even the most vulnerable population does not always experience difficulties caused by low income. Other risks carry more weight, both because they are more widespread and because they more frequently provoke

244 Social Vulnerability in Europe

multiple vulnerability. Contemporary societies are characterized by a twofold aspect: proliferating availability of resources and increasing disorganization of material life. It is these problems, connected with organizational difficulties and weak social integration and not simply with poverty, that expose a significant proportion of the population to serious difficulties.

Old and new inequalities The vulnerability indicators considered here designate the emergence of new forms of social inequality that partly add to, and partly cumulate with, more traditional forms of inequality. These new forms of inequality are founded on diversified access to resources essential for the well-being of individuals and families (via either the labour market or the welfare system) as well as the capacity to translate such resources into adequate living standards (in terms of housing conditions and/or reconciliation of care responsibilities with work activity). In this section we discuss whether new inequalities arising from this new array of social risks closely reflect the traditional configuration of social inequality, whether vulnerability is patterned along the traditional lines of class membership or whether the new risk factors generate novel forms of social differentiation. The point at issue is whether exposure to all the considered risk factors produces disparities that are relatively independent of the traditional inequality structure of society. Do only the lower social classes experience the new social risks, or do these risks affect also social groups that are part of the middle class? A hypothesis is that ‘risk society’ extends insecurity and vulnerability to social groups hitherto regarded as ‘invulnerable’. On the one hand, one expects to find high exposure to social vulnerability among the most disadvantaged social groups because of weakening in the main protection systems that were typical of industrial society while, on the other hand, high exposure to vulnerability should also be evident among broad sections of the middle class, which are more liable than in the past to suffer insecurity and social fragility. As social vulnerability is not restricted to poverty alone, it also concerns social groups that live above the poverty line. Social vulnerability, in fact, also affects social groups that are in an intermediate position and therefore intersects with the structure of income inequality. Table 9.7 shows this crossrelationship very clearly. In the first decile of equivalent household income, 85 per cent of families lie in the area of multiple vulnerability, severe hardship and transitional poverty. Income vulnerability has a strong incidence only in the first two income deciles, and the same applies to severe hardship. Multiple vulnerability instead has a high degree of incidence even in the third decile and therefore affects some sections of the middle class. The same is the case for difficulties in work/care reconciliation, which has high concentration up to the fourth decile. Moreover, the presence of a dependent

Table 9.7 Distribution of vulnerability by deciles of household income Vulnerability index

Housing deprivation

Dependent persons in material compression

Difficulty in re-conciling work and care

Moderate multiple vulnerabilities

Intense multiple vulnerabilities

Severe hardship

Total

1.9 6.5 6.9 9.0 9.2 9.6 10.5 9.5 9.6 8.8

2.0 4.6 5.7 6.7 6.1 6.2 5.1 5.4 4.6 3.8

3.0 6.0 4.8 4.3 3.6 1.8 1.5 .8 .6 .5

1.2 4.0 3.5 4.1 3.6 3.5 2.8 2.4 2.2 1.8

18.5 15.6 14.4 10.3 8.2 6.5 4.9 3.5 1.9 2.1

12.4 7.2 6.9 3.3 1.9 .9 .6 .5 .3 .1

35.4 18.1 11.4 5.4 3.6 2.4 1.9 1.2 1.3 .8

100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

8.1

5.1

2.7

2.9

8.8

3.5

8.0

100.0

No vulnerability

Income vulnerability

Work vulnerability

1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00

6.5 22.4 33.4 46.9 56.6 62.8 67.8 72.0 76.6 79.6

19.1 15.6 13.0 10.0 7.2 6.3 4.9 4.7 2.9 2.5

Total

52.2

8.7

Deciles

Source: ECHP, authors’ own calculations.

246

Social Vulnerability in Europe

family member has a high incidence up to the fifth decile. The other two vulnerability factors – work precariousness and housing risk – mainly affect households situated between the fourth and the sixth deciles (in the case of housing) and between the fourth and the ninth deciles (in the case of precarious employment or unemployment. These are therefore risks that are very common in population groups occupying an intermediate or even high position in the income scale. But how are vulnerability profiles distributed in relation to class position? To answer this question, we used the socio- occupational positions of persons in employment by re-aggregating these positions at the household level.8 Table 9.8 displays the distribution of social risks by occupational classes. We used a simplified version of the class classifications usually employed in sociological analysis. We merged together the two lowest positions on the socio- occupational scale (semi-skilled and unskilled manual workers) and separated out (using information on the size of firms) small entrepreneurs and skilled self- employed workers from other occupational groups. These two categories were combined into a new occupational group which were called the ‘urban petite bourgeoisie’. Analysis (see Table 9.8) shows that social vulnerability cuts across large part of the social structure, with a greater intensity among the lowest social strata, above all as regards severe hardship and multiple vulnerability. The hardest hit, in fact, are agricultural workers and unskilled manual workers, but also the urban petite bourgeoisie. Vulnerability therefore affects not only the social groups ranking lowest in the social scale, but also traditionally more protected social groups like skilled manual workers and the urban petite bourgeoisie. That social vulnerability is distributed cross-wise to occupational positions becomes even more evident if we separately consider the five social risks making up the area of moderate vulnerability. Income vulnerability is distributed across all the social strata, but it is more marked in the cases of agricultural workers, sales and personal service workers and the urban petite bourgeoisie, but not in the case of the working class. Work precariousness is diffused across two fronts: it affects on the one hand professionals and sales workers, and on the other the low-skilled working class. Obviously, temporary employment has a very different impact on the former than on the latter, and it is more likely to become permanent for the latter. Turning to housing vulnerability, households with the greatest difficulties mainly belong to the unskilled working class, to routine non-manual employees and to the petite bourgeoisie. Here housing conditions reflect the progressive weakening of the purchasing power of medium-to-low wages and they are also indicative of strong compression in living standards for social groups that were traditionally well protected in terms of income continuity and employment stability. Finally, problems connected with caring for dependent family members or children are distributed quite uniformly

Table 9.8 Distribution of the vulnerability index by social class position of households (occupational category of the head of household) (row percentages) Vulnerability index

Legislators, senior officials, corporate managers Professionals Technicians and associate professionals Routine non-manual employees Service and sales workers Urban petite bourgoisie Manual skilled workers Manual semi skilled and not skilled workers Total

Housing deprivation

Dependent persons in material compression

Difficulty in reconciling work and care

Moderate multiple vulner abilities

Intense multiple vulner abilities

Severe hardship

Total

7.6

2.9

.5

8.1

3.8

1.3

1.4

100.0

5.5 5.9

12.9 11.4

4.5 4.0

.4 1.0

3.8 3.5

5.0 5.5

1.7 1.8

1.3 1.9

100.0 100.0

66.8

5.8

9.0

6.3

.9

3.1

5.3

1.2

1.6

100.0

53.9

8.5

11.4

5.1

1.9

3.4

8.7

3.3

3.8

100.0

47.7

9.3

7.4

6.6

1.1

4.8

15.2

4.1

3.8

100.0

49.2

5.8

9.8

6.0

2.4

4.4

12.3

4.7

5.4

100.0

44.7

6.8

10.3

7.3

3.3

3.0

12.4

5.8

6.4

100.0

53.1

7.1

10.1

5.8

1.9

3.7

9.8

3.8

4.7

100.0

No vulnerability

Income vulnerability

Work vulnerability

68.4

6.0

64.9 65.0

Source: ECHP, authors’ own calculations.

248

Social Vulnerability in Europe

across social classes, with a slightly greater intensity of the former problem among low-skilled groups (where dependency combines with greater economic compression because of their lower wages) and of the latter problem among senior officials, corporate managers and intellectual professionals (for whom leaving work and looking after their own children is very often a matter of choice, not necessity). If we finally consider more severe situations, a more evident class divide is apparent. The only exception is the urban petite bourgeoisie, which has a high incidence of multiple vulnerability (but not severe hardship) and this confirms that the living standards of part of the middle class are strongly influenced by social vulnerability. The foregoing analysis has therefore shown that, although severe vulnerability exhibits a divide between the higher positions and medium-to-low positions on the occupational scale, moderate vulnerability cuts across a large part of the social structure. This provides striking evidence of the crisis that has affected broad sectors of the working class and some middle- class groups in the past ten years. The middle- class groups hardest hit by social vulnerability consist largely of skilled workers in sales and services, small traders, small entrepreneurs and artisans. These categories generally do not enjoy the guarantees afforded to wage- earners in all the countries, and are the least protected by high levels of education. These results certainly do not allow exhaustive conclusions about the role played by social class in determining social vulnerability (see Chapter 10 on this aspect). They nevertheless highlight two facts. First, they show that vulnerability affects social groups situated in both the intermediate and lower levels of the social stratification structure. Second, they show that the position of the urban petite bourgeoisie is closer to the working class in terms of social vulnerability than to other groups of the middle-class. The difficulties considered in this chapter are therefore distributed across a broad section of the population, partially cutting across income levels and class boundaries. As a result, social vulnerability has created a further segmentation in the inequality structure of society. In fact, the distribution of social vulnerability is such that large sectors of the European population – including a broad area of the working class and families with average incomes – are substantially exposed to it. Today, therefore, the traditional mechanisms generating social inequalities are flanked by new ones. We will see the factors that are principally responsible for these new inequalities in the next chapter.

Notes Costanzo Ranci wrote the first six sections. Mauro Migliavacca wrote the last section and performed the statistical analysis. 1. The data about economic vulnerability exclude Sweden as the ECHP database for Sweden is not a panel.

Social Vulnerability: A Multidimensional Analysis

249

2. This figure is lower than that estimated by Layte and Fouarge (2004), mainly because the poverty line has been calculated in different ways (as 60 per cent of the median in Layte and Fouarge’s case, but as 50 per cent in our case). Moreover, Layte and Fouarge consider the period 1994–1997, while the period considered in this study is longer (1995–2001). 3. Indices of affordability are not considered here because they are not calculable for all the countries examined. However, affordability is indirectly considered in the analysis because it results from the relation between income and housing conditions. 4. The following analysis does not consider Sweden as the ECHP database for Sweden is not a panel study. 5. Different scores were assigned to the indices in Table 9.1. Moderate levels in vulnerability were assigned one point, more severe levels were assigned two points and situations of severe hardship were assigned three points. The latter include: permanent poverty, the presence only of unemployed workers in the household, lack of basic elements in the dwelling and households with children aged 12 or less who are looked after by parents of whom one is looking for work, as well as severely hampered persons, in very poor health, in a family with compressed consumption. These scores were then summed to obtain an accumulated index. 6. Vulnerability identifies here household situations that are characterized by the presence of at least one social risk and excludes situations of severe hardship. 7. The following variables have been introduced in the analysis: the proportion of households with temporary or recurrent poverty in the 1997–2001 period, the proportion of households with temporary or unemployed workers, the proportion of households experiencing housing deprivation, the proportion of households with difficulties in reconciliing care and work or with a dependent persons and material compression of their living condition, and the proportion of households with multiple vulnerability. 8. Standard measures of social stratification, like the EGP scale (Erickson-GoldthorpePortocarero), were not used because the ECHP database does not provide proper data on this (Jungblut, 2006). Therefore the occupational scale employed here was constructed on the basis of ISCO88 scale. Persons in employment and pensioners (in relation to their last job) were considered in the analysis. The data were reaggregated on a family basis by using a dominance criterion, whereby the highest occupational position in the household prevailed (Erikson, 1984).

10 Explaining Social Vulnerability Costanzo Ranci, Brunella Fiore, Emmanuele Pavolini

Factors of vulnerability: a look at theories The aim of this final chapter is to identify the main social, economic and institutional factors responsible for social vulnerability. Social vulnerability is considered as a multi-dimensional phenomenon affecting contemporary post-industrial societies of Western Europe as a consequence of the emergence of new social risks in the past two decades. The social risks considered and analysed in the previous chapters of this book are the following: flexibilization of the labour market, polarization in the housing market and the consequent increase in housing deprivation, income insecurity and the spread of dependency as a consequence of the ageing of the population. These new social risks have contributed to a weakening of the coordination among the three fundamental mechanisms of social inclusion that in Keynesian societies ensured the stable and organized distribution and redistribution of social opportunities and social protection: the labour market, the family and the welfare state. As a consequence, in the past decade uncertainty and insecurity have spread across the continent more than before. The following analysis focuses on identifying the main factors that explain why some people (and their families) are more exposed to social vulnerability than others. That is, it seeks to determine what social, economic and institutional factors account for the different degrees of social vulnerability present in a population. As we have already seen in Chapter 9, social vulnerability is not equally distributed in Western European countries. Can we identify which factors shape this uneven distribution? If social vulnerability is a relatively new phenomenon, is its unequal distribution coherent with more traditional forms of social inequality? Is social class a good predictor of social vulnerability? To what extent and how does the welfare state successfully reduce the impact of the new social risks on the most vulnerable groups in the population? Which welfare regimes seem to perform better in protecting the population against social vulnerability? These are the main questions addressed in this chapter. 250

Explaining Social Vulnerability 251

Class structure, education and the welfare state Because social vulnerability is closely related to the life chances and material living standards of individuals and families, we argue that the relationship of people with income-generating mechanisms is the main factor likely to affect their social vulnerability. From this perspective, one would expect it to be the position of individuals in the social stratification structure that mainly determines their exposure to vulnerability. This was especially true in industrial societies, where the social chances of individuals essentially depended on their position in the labour market and the occupational structure was organized in accordance with relatively stable rules and hierarchical principles. The professional status of people and their employment relationships were the basic elements affecting the material living standards of most of the population. In post-industrial societies, however, both professional status and employment relations seem to have lost ground as predictors of living standards. First, opacity and flexibility have increasingly characterized the labour market and employment relations, making careers much more dependent on individual capabilities and strategies than before (Sennett, 1998). Secondly, consumption patterns and lifestyles not directly related to the occupational position of individuals have gained importance in the population, because work is no longer the sole source of individual and collective identity (Bourdieu, 1979; Melucci, 1996). This is not to imply that social class is disappearing as a social determinant of people’s material conditions (Goldthorpe, 1987; Crompton, 1993). However, it is increasingly clear that, in a knowledge society, access to social chances (and protection against the most widespread risks) is crucially determined by the cultural capital of individuals as well as by their capacity to mobilize and expand their skills (or to use them adequately in order to cope with risk situations) (Sen, 1985). It is therefore likely that social vulnerability in contemporary Western European societies is influenced not only by the traditional social stratification of the population, but also by the extent to which people belonging to successive generations are able to use their cultural capital in order to achieve social goals or to protect themselves (and their families) against emergent social risks. The position of individuals in the social stratification structure and their level of education (compared to the average level of education in the generation to which they belong) are therefore the two crucial structural factors affecting social vulnerability, if we consider the labour market to be the basic mechanism by which material resources are distributed in contemporary societies. As many authors have shown, however, social opportunities and protection against social risks are also distributed through the intervention of the welfare state. According to comparative political economy theory (Esping-Andersen, 1990), the uneven exposure of the Western European

252

Social Vulnerability in Europe

population to social risks typical of industrial society (loss of work and of the related income because of invalidity, sickness or old age) was adequately accounted for by the existence of different welfare regimes providing the population with variable amounts and kinds of public benefits. According to Esping-Andersen, moreover, welfare regimes have developed in the interplay between market and state intervention to constitute specific institutional systems of state-market regulation. In the past decade the development of new social risks has shown that traditional welfare regimes are only partially able to deal with the problems arising in the transition from industrial to post-industrial societies (EspingAndersen, 1990; Taylor- Gooby, 2004a). As we have shown in this book, new social risks do not affect the population in the same ways as the old social risks did. While the latter were strongly related to the employment position, the former emerge in the interplay between work and other spheres of everyday life, such as housing conditions, care responsibility and family organization. Furthermore, while the old social risks were concentrated mainly in the last stages of the working life, the new social risks are more intense in the early stages of work careers (as a result of the lengthy transition to stability in the labour market) and they largely affect the young population. Nevertheless, although the risk profile of the population has changed profoundly in the past decade, the structure of welfare provision has remained substantially unchanged. Even though some welfare regimes (especially those with already developed universalist schemes) have performed better than others in protecting people against new social risks (Taylor- Gooby, 2004a), institutional inertia and political ostracism have partially reduced the space and the opportunity for structural reforms (Pierson, 2001). The rise of new social risks, in fact, requires not only reorganization of the existing social protection schemes, but also recalibration of the entire welfare system (Ferrera and Rhodes, 2000; Ferrera and Hemerijck, 2003), in an attempt to meet needs and demands for social protection that were previously almost non-existent or marginal, or at least not fully recognized on the public agenda. New programmes and new targets for public welfare policies must be defined and introduced, partially renegotiating the current resource allocation structure of the welfare state. Recent attempts to re-scale public social programmes to the local and regional levels can puzzlingly be considered both a cost- containment strategy and a policy aimed at freeing space for innovation. Overall, the capacity to innovate welfare states, at both the national and regional levels, can be assessed by determining to what extent states and regions develop social programmes specifically intended to meet the needs of the vulnerable population. These programmes – such as work/care reconciliation measures, long-term care programmes for dependent people, employment regulation to protect temporary workers, housing programmes to assist the younger and less privileged population, etc. – are focused on new social risks and are often clearly

Explaining Social Vulnerability 253

distinct from the previous, more traditional, welfare programmes providing protection against the old social risks.

The role of the household Social class (and level of education) and social welfare protection can therefore be considered the main factors responsible for social vulnerability in Western European countries. However, numerous studies in recent years have emphasized the role played by the household structure in redistributing resources and protection (Crouch, 1999; Lewis, 2006; Saraceno, 2008). The structure of European families has changed in the past decade, shifting to a new equilibrium between individualization and family solidarity (see Chapter 3 for more details on this point). For our purposes here, there are three basic components of the household structure: (i) the number of persons cohabiting in the same dwelling and the kinds of relationship among them (ranging from single individuals to couples, to multiple or extended families); (ii) the presence and number of children (and their ages, whether adult or otherwise); and (iii) the income structure of the household (in terms of number and type of incomes). The current changes relate to all these aspects: family size has been shrinking (increasing the number of singles) while domestic relationships have become more turbulent, children have diminished in number while prolonging their residence in the parental home and finally, the traditional male breadwinner model (based on one earner) has been extensively replaced (even in the Mediterranean countries) by a large array of new family models. However, the family (especially if we consider not only people living in the same household but also inter-generational family ties) is still a strong factor in protection against vulnerability, especially for people with no (or instable) incomes and/or no access to social protection benefits. This potential for social protection arises from various sources. The family, through its internal organization of roles, assumes the primary function of pooling basic material resources, such as income and living space (housing), and redistributing them to people not in the labour market. Although individualization has caused a significant increase in the number of singles and dual- earner couples, there has not yet been a widespread replacement of the family’s pooling activity by other private or public mechanisms. In most cases, housing is still obtained and maintained by pooling the incomes of all the members of the family (including inter-generational transfers within the same kinship network). The costs of children and the dependent elderly are still largely borne by families (Saraceno, 2008). A growing proportion of young people, not only in Mediterranean countries, chooses to remain longer in the parental home, taking advantage of the income pooling and income redistribution functions performed by their families. The spread of temporary jobs seems to have reinforced this trend, because the family’s

254 Social Vulnerability in Europe

pooling activity enables young people in transition to adulthood to avoid the social vulnerability likely to occur in the early phase of their autonomy (see Chapter 8 for more details). Besides income and living space, the family also provides such basic services as care and social contacts and it channels to people in need the help and benefits provided by the welfare state, voluntary associations, neighbourhoods, private agencies, etc. Even in the European countries with the highest state provision of care services (the Scandinavian countries), the family is the first and most important supplier of care. In the past decade, the growing emphasis on the principles of ‘ageing in place’, together with a spreading culture opposing the institutionalization of the dependent elderly, has increasingly shifted the actual responsibility for care to informal caregiving, especially onto family members (Pavolini and Ranci, 2008). The public provision of home care and respite services has proved unable substantially to replace informal caregiving. In some countries (Italy, Spain and Austria) greater reliance on the private market (strongly facilitated by the spread of low- cost immigrant care workers) has transferred part of the care burden to professionals, giving rise to the partial marketization of care. However, to date this process has required greater financial responsibility on the part of family members. Furthermore, in the following decades the ageing process, together with an increase in the female activity rate even in Mediterranean countries, will expand the need for care throughout Western Europe. Feminist critics of the sociological theory underlining the family’s social role have stressed the marked and persistent gender inequalities in family relationships. According to these studies (Lewis, 1992; Sainsbury, 1996a), gender inequalities arising in the labour market impact on the gendered division of roles within families, leaving women in a disadvantaged position with respect to men. Other studies have pointed out the disparities distinctive of family organization as shown by the persistent overload of tasks and responsibilities assumed by women regardless of their employment status (Pfau-Effinger, 1998a; Crompton, 2006). According to this perspective, the protective capacity of the family is grounded on its gendered division of labour and on the asymmetry between men and women in both the labour market and the family. This also holds true in new family organizations, although a certain re-balancing can be observed in some Nordic and continental countries. From this perspective, further de-familization of the pooling and caring functions performed by families (via either marketization or state provision of services) is a process whereby more gender equality and symmetry can be achieved in contemporary societies. However, as Chapter 3 has shown, trends toward de-familization are largely counterbalanced by the rediscovery of family solidarity. To what extent the still high protective capacity of families combines with reduced gender inequality will be a matter for research and interpretation in the next years.

Explaining Social Vulnerability 255

The family is therefore to be regarded as a crucial factor in social vulnerability. However, the protective and redistributive capacity of families is greatly influenced by welfare regimes (Esping-Andersen, 1999). First, in many European countries (and especially in the continental or Mediterranean welfare regimes) crucial welfare benefits are distributed on a contributory basis and only subsequently to the family members of the entitled person. Thus, being a family with a member entitled to welfare benefits (in most cases, a permanent and regular employee) does make a difference ceteris paribus (once all the other conditions have been taken in consideration). Second, some public provisions may require, and not substitute for, the family provision of services. For example, child benefits are mainly distributed as inkind services in Sweden, while they are cash benefits in Italy. While in-kind services replace family work (contributing to the partial de-familization of care), cash benefit measures are intended to sustain the direct delivery of services by families (leaving the responsibility for care to the family and, within it, to women). Only recently have many EU countries introduced new programmes giving people the right to choose between care and cash, or delivering cash benefits conditional on obtaining professional care (or on combining informal and professional care).

The importance of regional disparities Vulnerability has been considered in this book as a syndrome affecting intermediate social strata in contemporary societies. Unlike poverty or social exclusion, vulnerability affects social groups that used to be fully integrated in the most significant systems of resources production and redistribution, as well as in the crucial spheres of social integration. It is therefore likely that the level of vulnerability in a population depends not only on individual or familial aspects, but also on the society’s level of economic development. Although it is true that individual poverty is greater in poorer areas (Pringle et al., 1999), it may be that vulnerability, too, is higher in geographical areas characterized by depressed economic development and lower levels of well-being in the population. Western Europe comprises regions that are among the richest in the world, as well as regions that are still very depressed, although the overall living standard is very high if compared with those in other continents. In any case, the economic competitiveness and general level of development of a region are expected to affect not only the life- chances of its population but also its levels of social integration. In addition, social vulnerability represents the fact that access to fundamental resources has become difficult, not only for a small minority of people, but for a substantial part of the working and middle classes. Paradoxically, large part of comparative welfare research has paid little attention to the importance of contextual factors such as the general level of well-being in a region. Today, this kind of information is widely available to

256 Social Vulnerability in Europe

researchers and is extensively used in regional economic analysis. However, with a few exceptions (Berthoud, 2004a), no comparative analysis has consistently used this information to compare poverty or social exclusion across Europe. The main reasons for neglect of this aspect in comparative research are as follow. ●

●

Measures of the economic performance or wealth of regions have been extensively considered as inappropriate for describing the living standards of a local population. Critics of economicism have often supported this idea. And although they are correct to point out the reductionism implicit in most economic analysis of poverty, this does not mean that the general level of development of a local economy has no effect on the material conditions of the population; indeed, disparities among European regions clearly evince the existence of such gaps (Heidenreich and Wunder, 2008). There has long been strong scepticism about whether the living conditions of a population can be described by considering contextual variables that uniformly affect the entire population of an area. Researchers have therefore concentrated on individual inequalities rather than regional disparities. Only recently (see Chapter 2 for a short review of this debate) has the enlargement of the EU revealed the persistence of regional inequalities as a crucial political and analytical issue.

Bringing the level of economic development of macro-regions back into the analysis has two implications. First, it challenges the idea that ‘policy matters the most’. If welfare regimes are still to be considered important predictors of social vulnerability, we need to understand the extent to which territorial differences in welfare provision are important when the economic development of such territories is also considered. According to some research, indeed, it seems that the income gaps among European regions are better predictors of poverty distribution than welfare regimes (Berthoud, 2004a). Poorer countries also exhibit more inequality to the point that, today, being rich is a similar condition throughout Europe, while being poor differs greatly according to the region. And what about social vulnerability? What proportion of the unequal distribution of vulnerability is explained by welfare policies, and to what extent does the regional level of wealth matter? Second, considering the regional level of development as a factor in vulnerability quite obviously entails that ‘territory matters’. We have already shown the marked differences among the 28 European macro-regions identified in this study. The intention now is to determine whether, and to what extent, such differences shape the exposure to social risks of the European population. Comparative analysis has often considered territorial factors (such as welfare policies) as distinct from individual factors (such as social class and education), but the former have been treated only as exogenous variables.

Explaining Social Vulnerability 257

Analysis has more often compared countries or regions by considering a number of aspects, inferring the existence of territorial or national differences from the observed ‘concomitant variations’. Hardly ever has causal analysis jointly considered regional and individual (or familial) effects in order to explain the distribution of social problems. The importance given to territorial diversity in this book means that such joint consideration is crucial, and it is necessarily the final stage of our analysis. The aim of this chapter is to identify the factors associated with vulnerability, and simultaneously to examine the effect on vulnerability of both individual families and macro-regions considered as independent elements. Our hypothesis is that vulnerability is affected by variables located at two different levels of analysis: some relate to family characteristics, while others are linked with the characteristics of the territorial context (the macroregion). The total variability of social vulnerability will therefore be split into two components: variance within the same macro-region, and variance between macro-regions. A multi-level logistic model will be developed to analyse variance in this particular way.

Variables description In light of the discussion in the previous sections, five types of variables are used in the following analysis in order to evaluate the factors connected with vulnerability levels: 1. 2. 3. 4. 5.

household type occupational status education level GDP per capita at the macro-regional level welfare-state provision and coverage at the macro-regional level.

Some of these variables (occupational status and household type) have been already introduced and discussed in previous chapters, so here they will be presented again only very briefly. The three remaining variables will receive closer attention, and especially the last one related to welfare-state provision. In Chapter 3 a complex household classification was drawn up by jointly considering two dimensions of households: the demographic structure (including the presence of children) and the income structure. As already discussed in the literature, and also illustrated in this book, there is an increasing tendency toward differentiation in household types. The result of our analysis is a classification consisting of 10 household profiles: ●

households composed of pensioners constitute 30 per cent of European households: around 60 per cent of these households consist of single persons, and the rest of elderly couples;

258 ●

●

●

Social Vulnerability in Europe

households composed of adults without children constitute a quarter of European households: around half of them are single-person households, and the rest are couples, with a predominance of dual-earner couples; households composed of adults with children aged under 25 years represent almost 40 per cent of households, with a relative majority exhibiting an male breadwinner model, although there are also dual-earner and one and a half models; households composed of parents and adult children (aged over 25 years) and lone parents with children aged less than 25 years account respectively for 8 per cent and 5 per cent of the total.

In regard to class structure, in Chapter 9 we constructed a simplified fiveprofile classification, quite similar to the classification developed by Erikson and Goldthorpe (1992)1: ●

●

●

●

●

highly-skilled non-manual workers (mostly in ISCO categories one and two), representing around 12 per cent of total household profiles; employees (mostly in ISCO categories three, four and five), equal to around 35 per cent of households; high-skilled manual workers (mostly in ISCO categories six and seven), around 20 per cent of total households; medium-low skilled manual workers (mostly in ISCO category eight), around one household out of four; and urban petite bourgeoisie (independent workers with fewer than 10 employees), 8 per cent of the total of profiles.

Household educational level has been calculated on the basis of the ISCED scale as follows: each individual ISCED value has been used to calculate a ratio between the ISCED value and the average ISCED value of the 5-year cohort to which the individual belongs. The average household level is the maximum of such individual values calculated for all the adult household members. Per capita GDP is a typical economic indicator, and it is used in the analysis as a proxy for the level of economic development and well-being of a macro-region. In 2001, per capita GDP in the 28 European macro-regions was 22,700 euros, with a relatively high standard deviation value (7524 euros). While the above variables have already been introduced in previous chapters, or are relatively widely used in the literature, the last set of variables – relative to welfare provision – has been constructed specifically for this analysis. When developing welfare provision indicators, the prime reference is to Esping-Andersen’s decommodification index (Esping-Andersen, 1990) and to related indices created in the debate that followed the book (Scruggs and Allan, 2006), as explained in the second section of this chapter. The decommodification index has many advantages in its structure, of course, but it has two shortcomings given the aims of the present analysis. First, it

Explaining Social Vulnerability 259

has been constructed to compare countries, whereas this chapter also focuses on macro-regional welfare performances. Second, it bears in mind the traditional Fordist welfare state based on monetary transfers covering ‘old’ risks such as income maintenance in situations of old age, illness and unemployment. The index is less able to capture the concept of defamilization2 (Bambra, 2005) and of programmes addressing the dimensions of social vulnerability considered in this book. The spread of the new social risks (Taylor- Gooby, 2004a) in the past decade has required the devising of new welfare programmes, or the reinforcement of previous ones, in order to protect the population against the negative impacts of those risks. Little account is taken of these programmes in the indices most commonly used to compare national welfare regimes. Moreover, the literature on these aspects has concentrated on the national level in conducting international comparisons, so that no indices were created to make inter-regional comparisons. For these reasons, we have developed a new index based on a dataset consisting of our 28 macro-regions and in which the following variables are used: ●

●

● ● ●

● ●

●

● ●

●

●

level of childcare coverage (in terms of the percentage of children aged under three attending childcare facilities); level of elderly care coverage (in terms of the percentage of people aged over 65 receiving care at home or in residential facilities); level of protection for stable workers (OECD index); level of protection for unstable workers (OECD index); level of coverage of potentially poor households (in terms of the percentage of households that would be poor without welfare benefits); effective parental leave; level of coverage for the unemployed (in terms of the percentage of unemployed persons receiving benefits); level of support for household educational activities (in terms of the percentage of households with children aged under 25 receiving educational benefits); percentage of households with children receiving family benefits; percentage of households in rented dwellings that receive housing allowances; percentage of households in social housing among those in rented dwellings; and old age pension net replacement rate weighted for levels of coverage.

The next step was to understand whether there are one or more latent policy dimensions that frame all the policy measures here considered in the same institutional setting. A principal components analysis (PCA) was therefore performed on these 12 variables, and the results are reported in Table 10.1. The PCA shows that none of these policy instruments can be

260 Social Vulnerability in Europe Table 10.1 Principal Component Analysis of the distribution of 12 policy measures addressing social risks in the macro-regions: communalities and extracted components matrix

Communalities

Component 1 New social risks protection index

0.532

0.646

0.123

0.784 0.541 0.670

0.765 0.66 0.349

0.345 0.072 0.740

0.677

0.705

0.424

0.723 0.678

0.798 0.538

0.224 0.625

0.750

0.710

0.495

0.623

0.787

0.056

0.562

0.680

0.316

0.759

0.777

0.394

0.714

0.354

0.638

Level of coverage of potentially poor households Elderly care coverage rate Childcare coverage rate Level of protection for stable workers Level of protection for unstable workers Effective parental leave Level of coverage of the unemployed Level of support for household education activities Percentage of households with children receiving family benefits Percentage of households in rented dwellings that receive housing allowances Percentage of households in social housing among those in rented dwellings Old age pension net replacement rate weighted for levels of coverage

Component 2 Old social risk protection index

Source: ECHP, authors’ own calculations.

framed in just one single index: there are two components with eigenvalues higher than one. The first of them, called the ‘new social risks protection index’ (IPNSR), explains around 44 per cent of the total variance and it chiefly measures social services coverage (child and elderly care), social housing provision, allowances related to poverty, housing, education and the level of protection for unstable workers. The second index, called the ‘traditional risks protection index’ (IPS), explains around 25 per cent of the total variance and is more closely related to the level of protection provided by industrial welfare state institutions such as pensions schemes, unemployment benefit and protection for stable workers.

Explaining Social Vulnerability 261

This result seems to confirm that the level of social programmes aimed at reducing the new social risks is not correlated with the level of the more traditional welfare programmes aimed at reducing risks closely related to the loss of stable employment. Not only is there a clear division between old and new programmes, but also full independence between the levels of such programmes (although the institutional setting of such measures is organized in accordance with the previous, well- established welfare programmes, as many analyses have pointed out). Throughout Western Europe, old and new social risks receive public responses that are not interdependent in their frequency and intensity.

Territorial variables at the macro-regional level Among the five variables just introduced, the last two – concerning per capita GDP and policies – warrant closer inspection at the macro-regional level because they exhibit quite different patterns of distribution. In regard to per capita GDP, corrected by PPP (purchase power parity index), Figure 10.1 shows the marked differentiation between macro-regions. It is interesting that, even more than for other variables, there is a strong intracountry differentiation, which reaches the highest level in countries like Italy (the Central-Northern part of which is among the richest regions of Europe, whereas the Southern part is the second worst), Spain and Germany. France and the United Kingdom are countries where, as already stated, the main differentiation is between the two largest metropolitan areas (London and Paris) and the rest of the country. Apart from the metropolitan areas of Paris and London and the peculiar position of Ireland, the richest regions of Western Europe are located at the two limits of continental Europe: on the North Sea rim (the Netherlands, Denmark and Northern Germany) and in the Southern part, just across the Alps (Austria, Southern Germany, North-Western Italy and CentralEastern Italy). Four clusters of regions are located in the middle range: the Scandinavian countries (Sweden and Finland), the Central regions of continental Europe (Belgium, Western Germany and Central-Eastern France), the Western regions of continental Europe (Western France, Mediterranean France, Eastern Spain and Madrid) and the Central and Western regions of the United Kingdom. Finally, the poorest regions are located, again, at the two extremes of Europe (the Northern United Kingdom and the Mediterranean regions) and in the former East Germany. The geography of per capita GDP therefore does not overlap with the geography of welfare regimes. Neither does it precisely duplicate the North/South divide of the continent. In fact, the regions with the highest levels of per capita GDP are located both in the Northern and Southern parts of the continental area, and they pertain to different welfare regimes (if we take account of the analysis distinguishing between continental and Mediterranean regimes). Note also that the per capita GDPs of the Northern and Eastern parts of Italy and Spain are closer

262

Social Vulnerability in Europe

Legend < 17000 17001–22000 < 22001–27000 < 27001

Figure 10.1

GDP level (PPT) per capita by macro-region (in euros)

Source: ECHP, authors’ own calculations.

to those of continental regions than to those of the other regions of their own country, which once again shows that regional differentiation within countries is important in Western Europe. Finally to be noted is that continental regions (plus Greater London) are the richest ones, thus confirming the finding by ESPON and others of an affluent area of Europe lying among London, Paris, Hamburg and Milan (our analysis shows that North-Eastern Italy should also be included in this area). Finally, Figure 10.2 shows how the macro-regions score in terms of welfare state protection against the ‘old’ and ‘new’ social risks. Inspection of Figure 10.2 raises two issues. The first concerns the relative differences in terms of protection capacity among macro-regions belonging

Explaining Social Vulnerability 263

2,00000 Sw

Dk Fin

New social risks protection index

UK N

1,00000

UK W

UK C UK L Irl

NI Ge E Ge N

Fr M Fr C Be Fr W

0,00000

Fr P

Au

Ge C Ge S

It NEC

−1,00000

It NW It S

Port Sp EM Sp S Gr

Sp CNW

−2,00000 −2,00000

−1,00000

0,00000

1,00000

2,00000

3,00000

Traditional social risks protection index

Figure 10.2 regions

Level of welfare protection against new and old social risks of the macro-

Source: ECHP and other statistical sources, authors’ own calculations.

to the same country. Traditional welfare protection against social risks shows little variation within one country because the programmes adopted are mostly nationwide. Macro-regional differences within the largest European countries in terms of protection against the new social risks are in some cases visible, but they are not usually of such a magnitude as to determine the allocation of a single country’s macro-region to different clusters. In particular, the United Kingdom, Spain and France exhibit little interregional differentiation, whereas Italy and Germany are more heterogeneous. But even in these two last countries the distances in relative terms among macro-regions are limited if compared to what happens in the overall map of European countries. Hence the analysis presented here reduces the importance of regional social policy differences with a strong impact not only in cross-country comparisons but also in intra- country ones. The second feature to be noted is that the adoption of a welfare state ‘regimes’ approach only partly holds when we introduce reference to protection programmes against the new social risks: the above results show that some regimes remain the same but others split into two sub-groups. We may therefore identify six policy areas, in contrast to the typologies of Esping-Andersen (1990) and Ferrera and Rodhes (2000). More closely

264 Social Vulnerability in Europe

matching our typology seems the one constructed by Anttonen and Sipilä (1996), who described five regimes dividing continental Europe into two areas: a German area and a French one. On looking at the data presented in Figure 10.2, it could be argued that there is a further policy area dividing Southern Europe into two zones, one more Western (Spain and Portugal) and the other more Eastern (Italy and Greece). We now consider the characteristics of these ‘regimes’ in more detail. As already shown by Taylor- Gooby (2004a), Scandinavian countries have been usually better able to combine strategies to protect against the old and new social risks. The United Kingdom and Ireland seem better able to redirect their policies to the new social risks: thus, as already stated by the literature on these countries, they adopt a more ‘liberal’ approach to traditional social risks protection. Central and Southern Europe are the areas that exhibit diversification in relation to their traditional ‘regimes’. In continental Europe, France and Belgium seem to occupy, in relative terms, intermediate positions in both rankings, whereas the German area has more difficulties in developing programmes to protect against the new social risks because it still relies on its (strong) traditional corporatist welfare system. In particular, the German area, on the one hand, has more limited coverage of services, especially child care, compared with the French area, while on the other, it seems better able to protect the unemployed and has a slightly higher old age pension net replacement rate weighted for levels of coverage. Southern Europe exhibits a differentiation between a Western-Iberian sub-area and an Eastern one (Italy and Greece). In relative terms, this group as a whole finds it difficult to develop strategies to tackle the new social risks. The differences between the two sub-areas are related to the characteristics of their old social risks protection models: in relative terms, ‘Eastern’ and ‘Western’ Southern European countries, although they usually have high old age pension net replacement rates like the Western ones, lack broad systems of coverage for unemployment risks, especially in Italy and Greece. In conclusion, introducing ‘new’ social risks protection into the analysis changes the way in which welfare states can be classified and makes it difficult to frame the European social protection models in terms of conventional regime typologies.

Multi-level analysis Multi-level analysis can be considered the most suitable technique with which to describe phenomena lying at different levels and influenced by factors having reference to these different levels. It is a particular kind of regression that enables variance in outcome variables to be analysed at multiple hierarchical levels, whereas in simple regressions all effects are modelled to occur at a single level (Raudenbush and Byrk, 1986; Snijders and

Explaining Social Vulnerability 265

Bosker, 1999; Goldstein, 2003). By focusing attention on hierarchical levels, multi-level analysis enables the researcher to determine where and how events are occurring (Jones, 2006). Our sample data are built on a two-level hierarchical structure, with 53,782 families nested in 26 macro-regions (Luxembourg and Sweden are excluded from this analysis because of the lack of substantive data). In our model, social vulnerability may be influenced by both familial and contextual phenomena, which is why family type, profession and family education level are included in the analysis together with contextual factors related to each macro-region; variables regarding the context are per capita GDP and two indicators of welfare policies. These factors do not refer directly to specific family characteristics, which is why GDP and policies should be considered as belonging to a level independent from that of families.3 This method makes it possible to verify the variability between macroregions by distinguishing it by levels. Then, after general estimation of individual parameters as in a simple regression, each model considers the variability within/between the 26 macro regions in which families live. As stated earlier, families are the first-level units (lower level), and the 26 macro-regions are the second-level units (upper level). The mean for each nested structure is about 1991 families, where the minimum is 608 and the maximum 4851. Against this background, the data will be now analyzed using a logistic multi-level model: the dependent binary response variable refers to whether a family was in a vulnerability condition (44.6 per cent) or whether it was not (55.4 per cent). The dependent variable can be denoted with yij which equals 1 if family i in macro-regions j is vulnerable, and 0 if it is not. The multi-level logistic regression model takes the form: Vulnij | Binomial (1, S ij )

(10.1)

where equation (10.1) states that the response variable follows a binomial distribution with probability pij of having vulnerability for family i in macro-regions j. Equation (10.2) has the general form of a logit model: Logit(S ij )

§ S ij log ¨ ¨ 1 S ij ©

· ¸ ¸ ¹

(10.2)

where the quantity S ij /(1 S ij ) is the odds that yij is equal to 1. ‘Odds’ is the ratio between the probability of success (vulnerability) and the probability of unsuccess (no vulnerability) (Snijders and Bosker, 1999).

266

Social Vulnerability in Europe

Equation (10.2) can also be written as: Logit S ij

E 0 j X1ij E1 X2 ij E 2 X3ij E 3 X4 j E 4 ¦ X5 j E 5

(10.3)

where the entire part after the equals sign refers to the modelling of Logit (S ij ) but, specifically, parameter b0j refers to the intercept, parameter b1 to the variable describing family type, b2 concerns information on family occupational status, and b3 indicates the education level of family members. Parameters b4 and b5 refer to variables concerning only the context (which is why the parameter subscripts do not have the letter i, which identifies a particular family). Parameter b4 refers to per capita GDP, while b5 denotes variables about welfare policies (IPNSR and IPS as described in the previous section). Intercept b0j takes the form:

E0 j

E 0 P0 j

(10.4)

where b0 refers to the fixed part of model, while m0j denotes the random part of the intercept which varies only across macro-regions. Because of this random part of the intercept, the model is also known as the Random Intercept Model: put more simply, m0j denotes the residual (or error) term that is the difference between the observed value of vulnerability for macro-regions and the value predicted by regression (Jones, 2006). The regional effects, m0j component, is assumed to follow a Normal Distribution with mean 0 and covariance matrix ⍀m , which for a random intercept model consists in a single term s 2m0: ª¬ Pij º¼ | N 0, : P : : P

ª¬V P2 0 º¼

(10.5)

The value of variance, s 2m0, is of great important for the following analysis, because it allows calculation of how variance between macro-regions decreases as variables are introduced.

Data analysis4 Now described is how variance in social vulnerability decreases when the above-described variables are introduced (see Table 10.2). Shown first, with the support of Table 10.3, is how variability in the probability of families being vulnerable depends on the macro-region in which they live. In the simplest model (model 0 or null model), only the value of the intercept is estimated because predictors at levels one and two are not introduced.5 Therefore the first step in the analysis is to establish whether vulnerability is influenced by

Explaining Social Vulnerability 267 Table 10.2 Probability of vulnerability by family type (logistic regression) Family typology Couple no children two work incomes Single welfare income Single work income Couple welfare income Couple no children one work income Couple with children ⬍ 25 one work income Couple with children ⬍ 25 two work incomes One parent with children ⬍ 25 Couple with children ⬎ 25 two work incomes One parent with children ⬎ 25

b

S.E.

eb

0.000 0.400 0.134 ⫺0.148 0.417 1.635 0.407 1.136 0.436 0.401

– 0.113 0.082 0.101 0.096 0.092 0.085 0.120 0.069 0.096

1.0 1.5 1.1 ⫺0.9 1.5 5.1 1.5 3.1 1.5 1.5

Note: Bold values are statistically significant for confidence intervals of 1.96. Source: ECHP, authors’ own calculations.

family factors specified in models, temporarily leaving contextual variables aside. M1, M2 and M3 introduce variables relative to family characteristics. A parameter with positive sign signifies a higher probability that families are vulnerable. M1 shows that family consisting of two working partners have significant lower odds (or probability) of being vulnerable compared with, for example, couples with children aged under 25 with only one income, or loneparent families. The odds of vulnerability in these two last types of family are respectively 5.1 and 3.1 times the odds for couples with two incomes and no children (Table 10.3). Therefore one-income couples with young children are the most vulnerable, and more so than lone parents with young children. Couples whose income derives from welfare benefits (mostly pensions) seem less vulnerable, because the odds of vulnerability are 0.9 times that of couples with two incomes. Moreover, a single pensioner is slightly more vulnerable than a single worker: the differential odds between the two types of family are only 0.4. M2 adds information about the highest professional profile in families. In the case of couples with double incomes, managers, professionals and employers have the lowest probability of vulnerability, followed by lowergrade professionals and routine non-manual employees (b2 ⫽ 0,118). The weakest professional categories are the petite bourgeoisie (b2 ⫽ 0.422), skilled workers and farmers (b2 ⫽ 0.608) and, for the most part, unskilled workers (b2 ⫽ 0.658). Educational level,6 introduced with M3, evidently protects against vulnerability: for each unit of increase in the educational level of the family, the probabilities of vulnerability decrease by –0.235 (Hanushek and Kimko, 2000). To sum up, a hypothetical family strongly protected against vulnerability is a couple with no children, with two work incomes, high educational level and high professional skills. However, none of the family variables

Table 10.3 Multi-level determinants of vulnerability in 27 macro-regions; random intercept model (N = 40.648) M0

Intercept Household type Ref.: couple two work incomes Single welfare income Single work income Couple welfare incomes Couple one income Couple with children ⬍ 25 one work income Couple with children ⬍ 25 two work incomes One Parent with children ⬍ 25 Couple with children ⬎ 25 two work incomes One parent with children ⬎ 25

M1

M2

M3

M4

M5

Coeff.

S.E

Coeff.

S.E

Coeff.

S.E

Coeff.

S.E

Coeff.

S.E

Coeff.

S.E

⫺0.268

0.092

⫺0.707

0.104

⫺1.062

0.094

⫺0.768

0.079

⫺0.802

0.080

⫺0.809

0,084

0.400 0.134 ⫺0.148 0.417 1.635

0.113 0.082 0.101 0.096 0.092

0.178 0.144 ⫺0.194 0.336 1.637

0.147 0.069 0.104 0.091 0.101

0.137 0.123 ⫺0.224 0.294 1.593

0.145 0.067 0.100 0.082 0.101

0.156 0.132 ⫺0.226 0.301 1.659

0.154 0.071 0.108 0.093 0.116

0.157 0.132 ⫺0.228 0.301 1.665

0.157 0.072 0.110 0.094 0.116

0.407

0.085

0.397

0.086

0.400

0.086

0.424

0.094

0.424

0.095

1.136 0.436

0.120 0.069

1.039 0.387

0.128 0.070

0.995 0.380

0.130 0.068

1.031 0.386

0.136 0.071

1.034 0.385

0.137 0.071

0.401

0.096

0.342

0.094

0.298

0.094

0.313

0.103

0.311

0.101

Profession Ref.: Higher grade professionals, administrators, managers and large proprietors Lower grade professionals, administrators, managers and routine non-manual employees Skilled workers and farmers Low skilled workers Petite bourgeoisie Mean household education level (Normalized) GPD per capita (Normalized) IPNSR (Normalized) IPS (Normalized) Variance 0,217 0.061 0.190 0.052 Random intercept variance (␴2␮0) % of variance reduction between 12.4 regions Note: Bold values are statistically significant for confidence intervals of 1.96. Coeff. = Coefficients. S.E.= Standard Error. Source: ECHP, authors’ own calculations.

0.118

0.089

⫺0.086

0.069

⫺0.094

0.072

⫺0.093

0,073

0.608 0.658 0.422

0.151 0.142 0.129

0.285 0.312 0.157 ⫺0.235

0.105 0.100 0.121 0.035

0.291 0.313 0.155 ⫺0.248

0.108 0.103 0.128 0.039

0.294 0.317 0.157 ⫺0.247

0,109 0,104 0,129 0.039

⫺0.379

0.086

⫺0.290 ⫺0.158 ⫺0.084

0.088 0,055 0,041

0.079 63.6

0.017

0.061 71.9

0.014

0.173 20.3

0.050

0.162 25.3

0.047

270

Social Vulnerability in Europe

estimated in our model drastically reduces the explained variance of social vulnerability. The change from M0 to M3, where all the family variables are included, reduces the random intercept variance by 25.3 per cent (from 0.217 to 0.162). More specifically, family type explains 12.4 per cent of the variance between regions; introducing occupational status reduces the variance by 7.9 per cent; finally, educational level adds 5 per cent of explained variance. Thus, as only approximately a quarter of the probability of being vulnerable results from differences in family characteristics, a large part of variability is not explained at the first level of analysis. This result supports our hypothesis that macro-regional differences contribute substantially to determining vulnerability. Models M4 and M5 introduce variables relative to the regional context, and they enable examination of the variations in vulnerability conditions attributable to intercept differentials observed at the level of macro-regions. The first variable to be considered is per capita GDP. This factor proves to be of great importance: for each unit of increase in per capita GDP, the probability of family vulnerability decreases by 0.379. Moreover, per capita GDP is particularly significant in explaining variability across macro-regions: indeed, about 40 per cent (from 0.162 in M3 to 0.079 in M4) of random intercept variance on vulnerability results from the socio-economic level of macro-regions. Our last model, M5, shows the effect exerted by policy variables IPS and IPNSR. As said, the former summarizes traditional welfare policies and the latter innovative policies against the new social risks. The parameter for policies addressing the new social risks shows that they have a more incisive effect in decreasing vulnerability: for each unit of implementation of IPNSR the probability of falling into vulnerability becomes –0.158. The effect of IPS (policies addressing the old social risks) in reducing vulnerability is much weaker: for each unit of increase, the coefficient parameter shows that the probability of vulnerability decreases by 0.084. However, the input of these two policy variables in M5 explains nearly 10 per cent of the total variability. Finally, nearly 30 per cent of the total variance has not been explained by the variables considered in the model. A large part of such variance could be considered as deriving from model specification errors. Another possible interpretation is that cultural and anthropological factors (such as a shared language, a common geography or a common culture) related to peculiar characteristics of each region (Castles and Obinger, 2008) have caused the unexplained difference between macro-regions.

Discussion The aim of the foregoing analysis was to identify the main social and political factors contributing to the diffusion of social vulnerability in Western

Explaining Social Vulnerability 271

European countries. The model considered two analytical levels: the role of family factors (position in the class structure, level of education in the household and household organization) and the role of macro-regional factors (welfare policy and level of socio- economic development). In this final section we summarize and interpret the main results of this analysis. The first result is that all the variables introduced in the model contribute to explaining social vulnerability. The variables at the household level are all significant in explaining vulnerability. Social vulnerability has been understood in this book as a syndrome widespread in Europe as a consequence of changes in both household organizations and the occupational structure of society. It can be explained in terms of a progressive disentanglement of work and the family that has weakened the social foundations of post-industrial societies (Touraine, 2007). The analysis carried out in this chapter enables us to identify which family types have effects, either positive or negative, on social vulnerability. The results show that the transformations that have taken place in the household organization over the past 20 years (see Chapter 3 for a description) have changed the exposure of families to social vulnerability. Household organizations that used to be adequately protected by traditional welfare states are today among those most likely to be vulnerable. Generally speaking, the occurrence of two incomes in the same household is the strongest protective factor against social vulnerability: households consisting of two adult workers or two pensioners are adequately protected against social risks, while households with only one income (except singles) are the most exposed. The presence of children in one-income households exacerbates these conditions: one-income families with dependent children are therefore likely to become vulnerable. For many decades this family organization has been typical in many continental countries (especially the Mediterranean and German-speaking ones). According to Crouch (1999), in the second half of the last century the stability of Western European societies was built on gender segregation within families and a clear separation between domestic work and employment in the labour market. The growth in female employment throughout Western Europe in recent decades has enormously altered this situation, although female segregation has not been abolished, but rather shifted from family organization to the labour market (where women have lower wages and less stable jobs than men). Our data show that the transition from the classic male breadwinner family model to new dual-income based family models has significantly increased the protection of the population against the new social risks, while the persistence of the traditional male breadwinner model generates greater exposure to social vulnerability. It is therefore the pooling capacity of families that has preserved women from incurring disadvantages and risks caused by the higher incidence of temporary or low-paid jobs in the female labour force. In addition, it is the capacity

272 Social Vulnerability in Europe

of families to reconcile work and childcare that enables them to acquire a stable position in society. In contrast, social vulnerability is greater in families where a work/childcare balance is not achieved and a gendered division persists between domestic work and employment in the labour market. These conclusions are commonly shared in both academic analysis and public discussion. A less common finding is that one-income families with children are largely exposed not so much to poverty as to social vulnerability: an ‘ordinary’ condition that does not lead automatically to severe hardship. In our vulnerability index, male breadwinner families are indeed in first place for moderate vulnerability, but only in fourth place for severe hardship (see Chapter 9). Male breadwinner families still account for one third of all households in Europe, ranging from 20 per cent in the Scandinavian countries to 67 per cent in the Southern Mediterranean ones (see Chapter 3). As they have decreased in the past decade, they have been replaced by a variety of new family models and, not linearly, by the growth of the dual-income family model. In the Mediterranean countries, for example, ‘extended male breadwinner families’ (comprising two earners but with the woman still restricted in domestic work) have taken the place previously occupied by ‘pure’ male breadwinner families (see Chapter 3 for details). In our causal model, the negative impact of male breadwinner families on social vulnerability persists when the effects of social class and welfare policies are controlled for. There are consequently aspects inherent to the male breadwinner model that make these families vulnerable. The most important of them is the low ratio between the number of income earners and the number of family members. Families with a low ratio are characterized by many subsequent disadvantages, such as housing overcrowding or low per capita income. The decreasing purchasing power of wages in the past decade has contributed to the weakening of one-income households. A second aspect is the sharp division of labour characteristic of these families, which causes rigidity in turbulent situations. Third, male breadwinner families are strongly exposed to hardship in the event of employment shortages or severe economic crises. Because the persistence of this traditional model causes major social disadvantages and exposure to social risks, one wonders whether the continuing large number of male breadwinner families in some areas of Europe is due to social policies that do not favor work/childcare reconciliation, and therefore indirectly encourage the traditional gendered division of labour in families, or to cultural resistance against individualization and greater gender equality. While there is no doubt concerning the role of social policy in favouring, directly or indirectly, this persistence (Lewis, 1992; Orloff, 1993), closer attention should also be paid to the negative impact of traditional cultural orientations in social situations demanding high organizational and cultural flexibility (Sennett, 1998).

Explaining Social Vulnerability 273

The second most vulnerable household model is represented, not surprisingly, by lone-parents families with dependent children. Unlike the male breadwinner family, this model is the result of a general trend toward individualization (with a higher incidence in English-speaking and Scandinavian countries). Lone-parent households with children show a higher level in severe hardship (where they occupy the lowest place) than in multiple vulnerability (where they take the fourth place). It is therefore the accumulation of economic, housing and reconciliation problems that puts these families in difficult situations. High exposure to poverty and social exclusion has indeed been extensively shown in the literature on this topic (Chambaz, 2001). Together with household structure, a second important variable in relation to vulnerability is the position of individuals in the social class structure. In industrial societies, the relationship between vulnerability and class was close because the social chances of individuals depended mostly on their position in a labour market organized in accordance with relatively stable rules and hierarchical principles. In post-industrial societies, however, various studies have challenged the idea that professional status and employment relations are the main predictors of living standards (Beck, 1992; Sennett, 1998). It is argued that social class is still an important variable, but other variables, among them cultural capital, perform significant roles as well. The analysis in the previous sections shows a complex scenario, of which some features are well known in the literature while others require closer examination. First, social class matters (of course). Inspection of model two in Table 10.3, when only the household type and the profession are introduced, shows that, if the upper service class is used as a reference group, some sort of differentiation takes place in terms of social vulnerability: on the one hand, the middle class made up by employees seems to be exposed to vulnerability no differently from the upper service class, while on the other, the two different types of working class (skilled and unskilled) and the petite bourgeoisie share a higher exposure to vulnerability. Therefore, whether skilled or unskilled, the working class occupies, in relative terms, a more vulnerable position in comparison with the service classes. Theories put forward by scholars like Kern and Shumann (1987), who argue that the working class has diversified with the rise of a more skilled and professionalized component, seem inapplicable when studying the relationship between class and vulnerability. Another interesting finding is the relatively high exposure of the petite bourgeoisie to social vulnerability compared with the rest of the middle class, which indicates that at least one section of this class experiences social difficulties. If we introduce a third variable – the household education level – into the regression model, the role of social class persists, but with a slight yet significant change. The differentiation between classes becomes a polarization between the working class and the service class, with the petite bourgeoisie

274 Social Vulnerability in Europe

becoming a group not significantly more vulnerable than the service class. The petite bourgeoisie seems more vulnerable, not because of its position in the labour market, but because of the lower level of cultural capital of workers belonging to this group. Thus, as also shown by the variance analysis in the lowest part of Table 10.3, not only the social class but also the cultural capital of households plays an important and independent role in differentiating exposure to social vulnerability. With regard to the exposure of different social classes to vulnerability in relation to other variables, too, it is interesting that, although differences in cultural capital seem to explain why the petite bourgeoisie is more vulnerable than the service class, the welfare state has no significant impact on the relative distribution of social risks among classes: even when welfare-state coverage indices (model five) are introduced, the polarization between the working class (and the low-skilled petite bourgeoisie) and the middle-upper classes persists. At the second level of our analysis, macro-regional differences are considered as factors in social vulnerability. Two factors have been introduced into the casual model: welfare policies and macro-regional per capita GDP. Territorial variables have greater explanatory capacity than household-level variables as the variance analysis has shown. Macro-regional per capita GDP explains 38 per cent of total variance and welfare policies explain 8 per cent. As a result, regional differences in Western Europe seem to be of great importance for the life- chances of the population, and they are possibly more crucial than factors related to work or family conditions. Social vulnerability is therefore better understood by looking at a map of the macro-regions of Europe than by looking at the individual or family-related characteristics of the population. This general result corroborates the conclusions of recent studies carried out in Europe and focused on regional and territorial differentiation (Berthoud, 2004a; Heidenreich and Wunder, 2008) . As described in Chapter 2, these studies have shown that not only is territorial differentiation persistent in Europe, notwithstanding the supposed integrative policy action of the EU, but also that regions are increasingly diverging. The persistence of regional diversity therefore seems to depend on underlying structural characteristics not significantly affected by policy or economic factors (Castles and Obinger, 2008). The main consequence of this situation is that the European population affected by social vulnerability is concentrated in specific macro-regional areas, while other areas are almost free from this syndrome (see Chapter 9 for details). Specifically, social vulnerability is concentrated in the Southern macro-regions of the Mediterranean area, and it is only residual in a contiguous, central area comprising small continental and Northern countries (Belgium, the Netherlands, Denmark and Finland), the large metropolitan areas of Northern Europe (Paris and London), Germany and North-Eastern

Explaining Social Vulnerability 275

France. However, such differences among macro-regions are consistent not only for the vulnerable population but also for the population unaffected by vulnerability. In other words, the gap between being well- off in Portugal and being well- off in Denmark is approximately the same as that between Portuguese and Danish vulnerable families. Only for the most deprived population do macro-regional differences greatly expand as a consequence of the concentration of social deprivation in specific areas where welfare policies are less developed. Two variables have been considered in our model in order to control for the effect of macro-regional differentiation: the impact of welfare policies on social risks (this variable has been split between policies addressing the new social risks and more traditional policies tackling old social risks) and the level of macro-regional per capita GDP. To date, a model of interpretation based on the concept of welfare regimes has been mostly adopted in international comparative research. In opposition to this tradition, our causal model shows the importance of socio-economic structure variables in determining social vulnerability with respect to policy variables. Moreover, the influence of per capita GDP is only slightly reduced when the effect of welfare policy is introduced into the model. This result supports the contention that underlying structural characteristics of macro-regions affect the spread of social risks throughout Western Europe much more than welfare policy does. Per capita GDP measures the degree of socio- economic development in a specific area as shown by the average level of well-being in the population. In Europe, there are huge differences in living standards among macro-regional areas. The impact of welfare policies slightly reduces this regional diversity, but it seems unable to generate a real process of convergence in the EU (Hoffmeister, 2009). Furthermore, the macro-regional levels of per capita GDP are hardly associated with national standards. Italian macro-regions, for example, occupy the eighth, fifteenth and twenty-fifth places, with a range of 12,000 euros from the highest macro-regions to the lowest. While North-West Italy is closer to the richest macro-regions in Europe, South Italy is among the poorest areas. The same holds, with a reduced range, for Spain and Germany. Differences in the British and French macro-regions are much less pronounced, however, with the important exception of the metropolitan areas of Paris and London. These results are consistent with recent research focused on the territorial distribution of income. For example, Berthoud (2004a) studied the impact exerted on geographical inequalities in Europe by various social and political factors. His conclusion was that inequalities were explained more by the general standard of living than by welfare regimes. Similarly, relative poverty within countries and European regions was found to be greater where the median income level was lower, which induced Berthoud to state that ‘the position of the country on the income scale is a more important

276 Social Vulnerability in Europe

influence than the welfare regime’ (Berthoud, 2004a: 35). Furthermore, a recent study on poverty in European regions (Lelkes and Zolyomi, 2008) states that regional disparity in terms of poverty is rather wide across the countries for which data are available. Although territorial inequalities are considered an important obstacle against the integration process in the EU, and a serious threat to the achievement of social cohesion within it (a goal set by the Lisbon Agenda in 2000), disparities in poverty and inequality among European states and at the EU-wide scale have been shown to be lower than in the USA (Brandolini, 2007). However, it emerges from our regression model that macro-regional disparities persist and are not significantly reduced by welfare policies. Furthermore, macro-regional disparities in social vulnerability are not positioned only along the North/South divide but form at least four specific clusters (see Chapter 9 for details). These clusters do not perfectly reproduce the traditional division among welfare regimes, confirming that the explanatory potential of policy factors in understanding social vulnerability should not be overestimated. Arguing that territorial structural characteristics account for differences in social vulnerability distribution more than welfare state provision does not mean, of course, that the latter has no role in exposure to social vulnerability. The data illustrated in model five of Table 10.3 show that the differences in social policy provision across Western Europe help to explain, independently of other variables, why some households are more vulnerable than others. As stated earlier, we have introduced two different indicators of welfare state protection: one more related to traditional risks coverage and another more related to the ‘new’ social risks. The regression analysis presented in Table 10.3 indicates that both types of welfare coverage have significant impacts on the relative diffusion of social vulnerability, but especially the latter. The result is important, but it should be understood correctly. First, what has been studied here is social vulnerability, not social exclusion. It may therefore be argued that the ‘traditional’ welfare state provision helps (often to a large extent in many European countries and regions) to reduce social exclusion. In the case of social vulnerability, it still has an impact, but it is an impact more limited than that of the ‘new’ and different types of welfare provision (Taylor- Gooby, 2004a). Second, in studying differences in exposure to social vulnerability, we have not simply analysed here the role of welfare state provision, but rather the impact of differences across macro-regions in the level of coverage. On considering the data collected by this study and the literature on the issue, it can be argued that there is greater heterogeneity across European macro-regions in terms of new social risks coverage than of old risks coverage. This is also because many old risks are covered by national programmes, whereas the coverage of many new risks is more territorially-based: for instance, the coefficient of variation across European macro-regions is 0.19 in relation to the old age pension net

Explaining Social Vulnerability 277

replacement rate, whereas it is 0.58 and to 0.49 when referring to, respectively, childcare and elderly care provision. The analysis on the role of welfare state provision against social vulnerability has also related to the discussion on welfare regimes: do the data presented in this chapter still support the idea that there exist three or four such ‘regimes’, when welfare protection in the case of new social risks is also introduced? As shown in the fifth section of this chapter, the answer is complicated. On the one hand, some ‘regimes’, in relative terms, still hold together: the Scandinavian regime and the Anglo-Saxon regime seem relatively homogenous. In the social- democratic Scandinavian welfare regimes, the existence of a broad welfare system and strong active employment policies greatly reduce exposure to the new social risks by providing support for work/care conciliation and for labour-market entry. In Anglo-Saxon liberal welfare regimes, it is market supply that predominates, with the consequence that initiatives designed to reconcile work and care are principally made in the private sector, while active employment policies privilege an approach based on incentives that exclude the most vulnerable groups. On the other hand, the continental and Southern European regimes tend to differentiate. Continental Europe corporatist welfare regimes give greater priority to core workers, while the dominant cultural models, which involve assumptions of principle on gender roles, have traditionally delayed the development of social services. In more recent decades, the French-speaking area has adopted an approach more concerned with developing care and work/care conciliation services, whereas the German area has lagged behind. The approach to welfare in Mediterranean countries is familist, and this has heightened the pressure of new social risks following the increase in the female activity rate: family organization in particular is running into difficulties. Within this cluster of countries, labour policies seem to follow different paths and Italy, together with Greece, has fallen behind. In conclusion, it seems difficult to frame differences among European welfare states by adopting the traditional regime concepts when reference to new social risks protection is made.

Notes The analysis was performed jointly by all the authors. Costanzo Ranci wrote the first four sections. Emmanuele Pavolini wrote the next two, Brunella Fiore the two after that and Costanzo Ranci and Emmanuele Pavolini wrote the last section together. 1. The ‘dominance approach’ has been used to allocate each household to one of the five profiles. 2. Defamilization is usually defined as the degree to which individual adults can uphold a socially acceptable standard of living, independently of family relationships, either through paid work or through social security provisions.

278

Social Vulnerability in Europe

3. If a hierarchical structure is not considered, this may have consequences on the correct meaning of the analysis for two main reasons. Firstly, it violates the assumption of independence of observations where the sample is not a simple random sampling but is built on different levels (Hox and Maas, 2002). Secondly, multi-level analysis makes it possible to achieve the greater parsimony of models: if the structure of the sample is not considered, there is a higher probability of obtaining a good estimation where none exists because simple regression generally causes the standard error coefficient to be underestimated (Hox and Maas, 2002; Goldstein, 2003). 4. The models are estimated with the MLwin program. For discrete response multilevel models, maximum likelihood estimation is computationally intensive and this is why quasi-likelihood is utilized. After the application of a linearization method – which converts discrete response models into a continuous response models – the model is estimated using generalized least squares (IGLS) (Goldstain, 2003). 5. In a null model it is assumed that the probability of vulnerability (yij) depends only on the mean of the entire population (mean equal to –0.287) and on a residual error moj. As said, the intercept is not a fixed part but may vary across macroregions, and this is how variance of the intercept can be analysed. The intercept variability is equal to 0.225 with standard error 0.064. This is the basic model and the starting point for determining how random intercept variance decreases as variables are introduced. 6. Educational levels, per capita GDP, IPNSR and IPS have been normalized.

General Conclusions Costanzo Ranci

The material collected in this book highlights the emergence of a new social question arising from the crisis of the European social model that began in the late 1980s. In Western Europe this model has long prevented poverty and social exclusion from reaching so high an intensity as has happened in other countries of the Western world, such as the USA. In the past two decades, with the exhaustion of economic growth and of the expansionary cycle of welfare policies, the picture has progressively changed and significant regional and social inequalities have reappeared. It became clear in the 1990s that the increasing difficulties were not transitory effects, but rather signalled a deeper-lying social change that required profound reorganization in the economic and social systems. The analyses conducted in this book have demonstrated the existence of a broad area of social vulnerability that has largely crossed the confines of the population suffering from poverty and social exclusion and now affects social areas traditionally considered protected against economic insecurity and uncertainty. Although the features of this social area are still to be explored, its substance is here empirically proven by a wide range of indicators showing the magnitude of temporary income poverty, the extent of work insecurity, the spread of housing deprivation, an increase in care problems connected with the ageing of the population and the difficulties inherent in work/childcare reconciliation and the transition to adulthood. The analyses conducted, in fact, have shown the existence, besides sections of the population in severe hardship, of a broader area of households suffering ordinary difficulties that do not necessarily turn into social exclusion or poverty but still expose them to vulnerability and uncertainty. Overall, this study confirms the hypothesis that uncertainty and social risks are widespread in the European population. Only just over half of European households can be considered endowed with an entirely adequate stock of resources, kinship networks and skills. The other half experiences, on one dimension or another, a deficit or a coordination difficulty that exposes them to difficulties and forces them to develop counter-strategies. 279

280 Social Vulnerability in Europe

This is a social area that in large part comprises unskilled manual workers, but also a significant component of the urban petite bourgeoisie and skilled manual workers. Although social vulnerability largely corresponds to traditional class divisions, it extends beyond the lower working class to encompass social groups occupying intermediate positions in the income scale. One feature shared by all these groups is their low level of education. The employment difficulties of women – still marked in many European countries – constitute one of the main risk factors, and they expose mainly oneincome families with dependent children to vulnerability. The presence of only one income earner in the family, in fact, is the most significant factor in vulnerability. Taken together, the studies in this book aid an understanding of the nature of the current transition from the European social model founded on the ‘mid-century compromise’ (Crouch, 1999) to a new social model. The latter seems characterized by the spread of new social risks, in forms that only rarely assume marked intensity but nevertheless affect a broad section of the population. These new social risks spring from changes in the labour market, the demographic structure of the population and the internal organization of families. They make the generous welfare systems constructed in Europe in previous decades altogether less effective. However, the extent of these new social risks clearly prevails over their accumulation and concentration in one section of the population. In the absence of dynamic data, it is obviously impossible to determine whether the spread of such risks is indeed expanding or whether it is concentrating in a significant minority of the population. Nevertheless, the structural analyses developed in this book clearly show that the extension and diversification of risks are both very large, while their accumulation is very modest. Vulnerability is therefore a syndrome that contributes not so much to existing social polarization as to further social differentiation. New inequalities thus add to previous ones, rendering the pattern of social inequalities much broader and more complex than it used to be. Moreover, the new social risks concern social reproduction more than the production of resources. The risks with the greatest cumulative effects on the living conditions of the population concern access to housing and reconciliation between work and childcare. At stake, therefore, is the capacity of people to convert resources acquired from work into adequate standards of living and/or life-projects. It is in the initial phases of the household lifecourse that the stability of economic and kinship resources performs a crucial role. When the difficulty of finding suitable housing, or an inability to satisfy the care demands of more fragile family members, is combined with the lack of an adequate income, a greater risk of poverty or greater risk of being in a marginal position in the labour market rapidly arises. Moreover, this situation is critical in itself as it exposes people to uncertainty, tradeoffs and difficulty in coordinating between divergent social demands.

General Conclusions

281

The sphere of social reproduction and that of the conversion of acquired resources into well-being therefore constitute the two areas of vulnerability most likely to degenerate into severe hardship and accumulated risks. Finally, it should be borne in mind that the majority of the indicators used in our analyses concern not so much the quantity of available resources as their stability, and the capacity of people to coordinate opposing demands. Rather than on poverty in itself, or permanent exclusion from the labour or housing markets, our analysis has focused on households experiencing social uncertainty This is indicated, on the one hand, by a relative scarcity of resources and, on the other, by the instability of positions acquired or difficulties in coordinating different family functionings. The uncertainty of these life circumstances often gives rise to a strong compression of living standards or to difficulties in developing life-projects consistent with expectations. In general, uncertainty and instability fuel anxiety about the future and a preoccupation with risk-avoidance, and they also hinder longterm investments and projects. Sometimes, but not necessarily, they expose the vulnerable population to sudden and rapid deterioration in its living standards as a result of unexpected events like an illness, a loss of employment or the birth of a child. The onset of social vulnerability is moving onto the agendas of social policy-makers, who are developing new social programmes to furnish more targeted protection against these new social risks. The development of these programmes often comes about in continuity with the traditional intervention models, and along already established institutional paths. The level of generosity of the existing intervention models notably influences the coverage and the amount of the new programmes, not only because the allocation of public financial resources is closely constrained by the existing arrangements, but also because traditional programmes are partly addressed to protect from these new social risks. However, the analyses in this book have highlighted the peculiarity of new programmes that are often intended to meet needs that are under-protected by traditional programmes. The EU is called upon to perform a leading role in this area of innovation. Indeed, our research has shown that the strong national and regional characterization of these policies is at odds with enormous territorial differences. The geographical distribution of social vulnerability is in fact markedly imbalanced along lines that do not correspond to national borders, and even less to the boundaries of the different welfare regimes identified by comparative analysis. It is in fact the degree of socio-economic development of a macroregion, even more than the policies adopted at both national and regional level, that influences the extent to which the population is exposed to vulnerability. The place where people live matters more than social class. Public policies attenuate this severe territorial inequality, which originates from unequal patterns of development, only slightly. The question arising from our research therefore concerns the future role of the EU. To what extent

282

Social Vulnerability in Europe

can a strategy of welfare recalibration (Ferrera and Hemerijck, 2003) take account of the severe territorial disparities in Western Europe? The future challenge for the EU is its capacity to give its citizen more equitable life-opportunities independent of their economic, social and even territorial backgrounds. The fight against social vulnerability can be considered either a problem or an opportunity. In the former case, it points up historical shortcomings in the development of national welfare systems; in the latter, it indicates that there is space for innovating and recalibrating these systems on the basis of a European agenda. The close attention paid to regional differences in this book emphasizes that the supranational role of EU policies should be relaunched. The social turbulence analysed in this study is proceeding slowly, and it has not been fully recorded under the current systems of observation. Yet it has profound social and economic implications. It has yet to be thoroughly recognized and addressed by welfare policies. The emergence of the new social question, in fact, requires both citizens and public policy-makers to devise new coping strategies. New opportunities for experimentation and innovation are therefore up for grabs. Hopefully, this book has documented the need for this innovation and demonstrated its urgency.

References

Aasve, A., Iacovou, M. and Mencarini, L. (2006) ‘Youth poverty and transition to adulthood in Europe’, Demographic Research, 15, 2, 21–50. Alber, J. and Köhler, U. (2004) Health and Care in the Enlarged Europe, European Foundation for the Improvement of Living and Working Conditions, Office for Official Publications of the European Communities, Luxembourg. Anderson, R. and Bury, M. (1988) Living with Chronic Illness. The Experience of Patients and their Families, London: Unwin Hyman. Anttonen, A. and Sipilä, J. (1996) ‘European social care services: Is it possible to identify models?’, in Journal of European Social Policy 6, 2, 87–100. Apospori, E. and Millar, J. (eds.) (2003) The Dynamics of Social Exclusion in Europe— Comparing Austria, Germany, Greece, Portugal and the UK, Cheltenham UK and Northampton, MA, USA: Edward Elgar. Arts, W. and Gellissen, J. (2002) ‘Three worlds of welfare capitalism or more?’, Journal of European Social Policy, 12, 137–158. Arum, R. and Muller, W. (eds.) (2004) The Reemergence of Self-Employment: A Comparative Study of Self-Employment Dynamics and Social Inequality, Princeton: Princeton University Press. Atkinson, A. (1996) ‘Income Distribution in Europe and the United States’, Oxford Review of Economic Policy, 12, 15–28. Atkinson, A., Cantillon, B., Marlier, E. and Nolan, B. (2002) Social Indicators: the EU and Social Inclusion, Oxford: Oxford University Press. Attias-Donfut, C. (ed.) (1995) Les Solidarietés Entre Générations. Vieillesse, Famille, Etat, Paris: Nathan. Attias-Donfut, C., Ogg, J. and Wolff, F.C. (2005) Family support, in Börsch-Supan A., Brugiavini A., Jürges H., Mackenbach J., Siegrist J., and Weber G. (eds.), Health, Ageing and Retirement in Europe- First Results from Share, Manheim: MEA, 171–178. Avramov, D. (2002) People, Demography and Social Exclusion, Strasbourg: Council of Europe Publishing. Avramov, D. (2006) ‘The housing dimensions of population change and welfare reforms’, paper presented at ENHR conference Housing in an Expanding Europe: Theory, Policy, Participation and Implementation, Ljubljana: Slovenia, July 2–5, 2006. Bagnasco, A. (1977) Tre Italie. La problematica territoriale dello sviluppo italiano, Bologna: Il Mulino. Balchin, N.P. (ed.) (1996) Housing Policy in Europe, London; New York: Routledge. Baldini, M. and Toso, S. (2004) Diseguaglianza, Povertà e Politiche Pubbliche, Bologna: Il Mulino. Bambra, C. (2005) ‘Cash Versus Services: “Worlds of Welfare” and the Decommodification of Cash Benefits and Health Care Services’, in Journal of Social Policy, 34, 195–213. Barbagli, M., Castiglioni, M. and Dalla Zuanna, G. (2003) Fare Famiglia in Italia. Un Secolo Di Cambiamenti, Bologna: Il Mulino. Barbieri, P. (2001) ‘Self-employment in Italy: does labour market rigidity matter’, International Journal of Sociology, 31, 60–100. 283

284

References

Barbieri, P. (2007) ‘Atypical employment and welfare regimes’, EqualSoc, Policy paper, Volume 1, Antwerp: Herman Deleek Centre for Social Policy. Barnes, C., Mercer, G., and Shakespeare, T. (1999) Exploring Disability. A Sociological Introduction, Cambridge: Polity Press Beck, U. (1992) Risk Society: Towards a New Modernity, London: Sage. Beck, U. (1998) World Risk Society, Cambridge: Polity Press. Bengtson, V.L. (2001) ‘Beyond the nuclear family: The increasing importance of multigenerational bonds’, in Journal of Marriage and Family, 63, 1. Bernhardt, E. (2004), ‘Cohabitation or marriage? Preferred living arrangements in Sweden’, sdf-puzzle, 04/2004, 1–10. Available: http://www.oif.ac.at/sdf/sdf04-04bernhardt.pdf. Berthoud, R. (2004a) Patterns of Poverty Across Europe, Bristol: The Policy Press. Berthoud, R. (2004b) ‘How Can Deprivation Indicators Help Us to Understand Poverty?’, Benefits, 14(2), 103–114. Berthoud, R., Bryan, M. and Bardasi, E. (2004), ‘The Dynamics of Deprivation: the relationship between income and material deprivation over time’, ISER Working Paper. Berthoud, R. and Iacovou, M. (eds.) (2004) Social Europe-Living Standards and Welfare States, Cheltenham: Edward Elgar. Betti, G. and Verma, V.K. (1998) ‘Measuring the degree of poverty in a dynamic and comparative context: A multi-dimensional approach using fuzzy set theory’, Dipartimento di metodi quantitativi Università di Siena, Working Paper, 22. Betti, G. and Verma, V.K. (2004) ‘A methodology for the study of multi-dimensional and longitudinal aspects of poverty and deprivation’, paper presented at IAOS-IASS Joint Conference, Amman-Jordan, November 29–December 1, 2004. Billari, F.C., Castiglioni, M., Castro Martin, T., Michielin, F. and Ongaro, F. (2002) ‘Household and union formation in a Mediterranean fashion: Italy and Spain’, in E. Klijzing and M. Corijn (eds.), Dynamics of Fertility and Partnership in Europe II, Geneva: United Nations, 17–41. Billari, F.C. and Liefbroer, A. (2007) ‘Should I stay or should I go? The Impact of Age Norms on Leaving Home’, Demography, 44, 1, 181–198. Billari, F.C. and Micheli, G.A. (2001) ‘Social norms on agents’ demographic events. Preliminaries for a (multi-) agent-based approach’, in Conte, R. and Dellarocas, C. (eds.), Social Order in Multiagent Systems, Kluwer: Dordrecht, 201–221. Blossfeld, H.P. and Drobnic, S. (eds.) (2001) Careers of Couples in Contemporary Societies: From Male-Breadwinner to Dual-Earner Families, Oxford: Oxford University Press. Blossfeld, H.P., Klijzing, E., Mills, M. and Kurz, K. (eds.) (2005) Globalization, Uncertainty and Youth in Society, London: Routledge. Boarini, R., Johansson, A. and Mira d’Ercole, M. (2006) ‘Alternative measures of well being’, Paris: Directorate for employment, labour and social affairs committee, OECD Social, Employment and Migration, Working Papers, 33. Bourdieu, P. (1979) La distinction: critique sociale du jugement, Paris: Editions de Minuit. Brandolini, A. (2007) ‘Measurement of income distribution in supranational entities: the case of the European Union’, Banca d’Italia, Temi di Discussione Working Paper, 623, 2007 (also published in Jenkins, S.P. and Micklewright, J. (eds.), Inequality and Poverty Re-examined, Oxford: Oxford University Press, 62–83). Brenner, N. (2004) New State Spaces: Urban Governance and the Rescaling of Statehood, Oxford: Oxford University Press.

References

285

Brousse, C. (2004) The Production of Data on Homelessness and Housing Deprivation in the European Union: Survey and Proposals, Luxembourg: European Communities. Burchardt, T. (2004) ‘Capabilities and disability: The capabilities framework and the social model of disability’, in Disability and Society 19, 7, 735–751. Castel, R. (1995) Les métamorphoses de la question sociale. Une chronique du salariat, Paris: Fayard. Castel, R. (2003) L’insecurité sociale. Qu’est-ce qu’etre protégé?, Paris: Editions du Seuil. Castles, F.G. (1998) Comparative Public Policy: Patterns of Post-war Transformation, Cheltenham: Edward Elgar. Castles, F.G. and Mitchell, D. (1993) ‘Worlds of Welfare and Families of Nations’, in Castles, F.G. (ed.), Families of Nations: Patterns of Public Policy in Western Democracies, Dartmouth: Aldershot, 93–128. Castles, F.G. and Obinger, H. (2008) ‘Worlds, families, regimes: Country clusters in European and OECD area public policy’, West European Politics, 31, 1–2, 321–344. Chambaz, C. (2001) ‘Lone-parent Families in Europe: A variety of economic and social circumstances’, in Social Policy and Administration, 35, 6, 658–671. Cherlin, A.J., Scabini, E. and Rossi, G. (1997) ‘Still in the Nest: Delayed Home-Leaving in Europe and the United States’, Journal of Family Issues, November, 572–575. Chung, H. (2005) Different Strategies for Flexibility. Employment Protection and Temporary Employment: A Study of Cross-national Variance on Temporary Employment in 19, OECD Countries, paper presented at TLM.NET conference, Managing Social Risks through Transitional Labour Markets, May 19–21. Cioni, E. (1999) Solidarietà tra generazioni. Anziani e famiglie in Italia, Milano: Franco Angeli. Clanché, F. (2008), ‘Classifying housing situations: Homeless people in general classifications’, in Marpsat, M. (ed.), The INED Research on Homelessness, 1993–2008, Paris: INED, 29–44. Corijn, M. and Klijzing, E. (eds.) (2001) Transitions to Adulthood in Europe, Bruxelles: Kluwer Academic Press. Costa, G. (2004) ‘Severe family care needs, coping strategies and new forms of social vulnerability’, ISA Ageing 2004 Conference ‘Diversity and Change in a Global World’, Roehampton, Surrey, September 7–9. Costa, G. (2007) Quando qualcuno dipende da te. Per una sociologia della cura, Roma: Carocci. Crane, M. and Warnes, A.M. (2000) ‘Evictions and prolonged homelessness’, Housing Studies, 15, 5, 757–773. Crompton, R. (1993) Class and Stratification: An Introduction to Current Debates, London: Polity Press. Crompton, R. (ed.) (1999) Restructuring Gender Relations and Employment: the Decline of the Male Breadwinner, Oxford: Oxford University Press. Crompton, R. (2006) Employment Family. The Reconfiguration of Work and Family Life in Contemporary Societies, New York: Cambridge University Press. Crouch, C. (1999) Social Change in Western Europe, Oxford: Oxford University Press. d’Addio, A.C. and d’Ercole, M.M. (2005) ‘Trends and determinants of fertility rates: The role of policies’, OECD Social Employment and Migration Working Papers, OECD Publishing, 27. Dale, A., Williams, M. and Dodgeon, B. (1996) Housing Deprivation and Social Change: A Report Based on the Analysis of Individual Level Census Data for 1971, 1981, and 1991, Drawn from the Longitudinal Study and the Samples of Anonymised Records, London: HMSO.

286

References

Dalla Zuanna, G. and Micheli, G.A. (eds.) (2004) Strong Family and Low Fertility: A Paradox?, Dordrecht: Kluwer Academic Press. de Beer, J. and Deven, F. (eds.) (2000) Diversity in Family Formation: the 2nd Demographic Transition in Belgium and the Netherlands, Bruxelles: Kluwer Academic Press. Del Boca, D. (2002) ‘The effect of child care and part time opportunities on participation and fertility decisions in Italy’, Journal of Population Economics, 15, 3, 549–573. Del Boca, D. and Wetzel, C. (2007) Social Policies, Labour Markets and Motherhood, Cambridge: Cambridge University Press. Dewilde, C. (2004) ‘The multidimensional measurement of poverty in Belgium and Britain: A categorical approach’, Social Indicators Research, 68, 3, 331–369. Dovey, K. (1985) ‘Home and homelessness’, in Altman, I. and Werner, C.M. (eds.) Home Environments, New York and London: Plenum Press, 33–64. Edgar, B., Doherty, J. and Meert, H. (2002) Access to Housing: Homelessness and Vulnerability in Europe, Bristol: Policy Press. Edgar, B. and Meert, H. (2005), Fourth Review of Statistics on Homelessness in Europe. The ETHOS Definition of Homelessness, Brussels: FEANTSA. EIRO (European Industrial Relations Observatory on-line) (2002). Economically Dependent Workers, Employment Law and Industrial Relations, EU Countries. Available: http://www.eiro.eurofound.eu.int/2002/05/study/tn0205101s.html. Ellingsaeter, A.L. (1998) ‘Dual breadwinner societies: Provider models in the Scandinavian welfare states’, Acta Sociologica, 41, 59–73. Erikson, R. (1984) ‘Social class of men, women and families’, Sociology, 18, 500–514. Erikson, R. and Goldthorpe, J.H. (1992) The Constant Flux. A Study of Class Mobility in Industrial Societies, Oxford: Clarendon. Esping-Andersen, G. (1990) The Three Worlds of Welfare Capitalism, Oxford: Polity Press. Esping-Andersen, G. (1999) Social Foundations of Post-industrial Economies, Oxford: Oxford University Press. Esping-Andersen, G. (2002) A new gender contract, in Esping-Andersen, G., Gallie, D., Hemerijck, A. and Myles, J. (eds.), Why We Need a New Welfare State, Oxford: Oxford University Press, 68–95. Esping-Andersen, G., Gallie, D., Hemerijck, A. and Myles, J. (eds). (2002) Why We Need a New Welfare State, Oxford: Oxford University Press. ESPON, European Spatial Planning Observation Network, Territory matters for competitiveness and cohesion. Facets of regional diversity and potentials in Europe, ESPON Synthesis Report III, results by autumn 2006. Eurobarometer (2007) Health and Long-Term Care in the European Union, Luxembourg: European Commission. EUROFAMCARE Consortium (2006) Carers of Older Dependent People in Europe: Characteristics, Coverage and Usage. Available: http://www.uke.uni-hamburg.de/ extern/eurofamcare. EC—European Commission (1999) Employment in Europe, Luxembourg: European Commission. EC—European Commission (2005) Employment in Europe, Luxembourg: European Commission. EC—European Commission (2006) ‘Flexibility and security in the EU labour markets’ in EC—European Commission, Employment in Europe, Luxembourg: European Commission. EC—European Commission, Directorate-General (2008) ‘Employment, social affairs and equal opportunities’ in Social Inclusion and Income Distribution in the European Union 2007, Luxembourg: European Commission.

References

287

European Foundation for the Improvement of Living and Working Conditions (1995) Who Will Care? Future Prospects for Family Care of Older People in the European Union. Available: www.eurofound.europa.eu. European Foundation for the Improvement of Living and Working Conditions (2003) First European Quality of Life Survey. Available: www.esds.ac.uk/findingData/ snDescription.asp?sn=5260. Eurostat (2000) European Social Statistics: Income Poverty & Social Exclusion (1st Report), Luxembourg: Eurostat. Eurostat (2003a) European Social Statistics: Income Poverty & Social Exclusion (2nd Report), Luxembourg: Eurostat. Eurostat (2003b) Labour Force Survey 2002, Luxembourg: Eurostat. Eurostat (2003c) Feasibility Study. Comparable Statistics in the Area of Care of Dependent Adults in the European Union, Luxembourg: Eurostat. Eurostat (2007) Living Conditions in Europe, 2007, Luxembourg: Eurostat. Ferrera, M. (1996) ‘The “Southern Model” of welfare in social Europe’, Journal of European Social Policy, 6, 1, 17–37. Ferrera, M. (2005) The Boundaries of Welfare. European Integration and the New Spatial Politics of Solidarity, Oxford: Oxford University Press. Ferrera, M. and Hemerijck, A. (2003) Recalibrating European Welfare Regimes, in Zeitlin, J. and Trubeck, D. (eds.), Governing Work and Welfare in a New Economy: European and American Experiments, Oxford: Oxford University Press, 88–128. Ferrera, M. and Rhodes, M. (eds.) (2000) Recasting European Welfare States, London: Frank Cass. Finch, J. (1989) Family Obligations and Social Change, Cambridge: Polity Press. Fitzpatrick, S. and Christian, J. (2006) ‘Comparing homelessness research in the US and Britain’, European Journal of Housing Policy, 6, 3, 313–333. Forrest, R. and Kearns, A. (2001) ‘Social cohesion, social capital and the neighbourhood’, Urban Studies, 38, 12, 2125–2143. Förster, M., Jesuit, D. and Smeeding, T. (2002) ‘Regional poverty and income inequality in Central and Eastern Europe: Evidence from the Luxembourg Income Study’, Luxembourg Income Study Working Paper Series Luxembourg Income Study (LIS) Working Paper, 324. Forster, M. and Mira d’Ercole, M. (2005) ‘Income distribution and poverty in OECD countries in the second half of the nineties’, OECD Social, Employment and Migration Working Papers, 22. Freedman, V. and Martin, L. (2000) ‘Contribution of chronic conditions to aggregate changes in old-age functioning’, American Journal of Public Health, 90, 11, 1755–1760. Friedrichs, J., Galster, G.C. and Musterd, S. (eds.) (2005) Life in Poverty Neighbourhoods: European and American Perspectives, London: Routledge. Fuà, G. (1980) Problemi dello sviluppo tardivo in Europa, Bologna: Il Mulino. Fujiura, G.T. and Rutkowski-Kmitta (2001) Counting disability, in Albrecht, G.L., Seelman, K.D. and Bury, M. (eds.), Handbook of Disability Studies, Sage, Thousand, 69–96. Giorgi, L., Tentschert, U. and Avramov, D. (2001) ‘Housing situation’, in EurohomeImpact, Housing Stress: An Overview of Risk Populations and Policies in the EU, Vienna: Eurohome-Impact Deliverable R1. Available: http://www.iccr-international.org/ impact/docs/deliverable1.pdf. Goldstein, H. (2003) Multilevel Statistical Models, London: Arnold. Goldthorpe, J. (1987) Social Mobility and Class Structure in Modern Britain, Oxford: Clarendon Press.

288

References

Graham, H. (1983) Caring: A Labour of Love, in J. Finch e D. Groves (ed.), Labour of Love: Women, Work and Caring, London: Routledge, 13–30. Grammenos, S. (2003) ‘Illness, disability and social inclusion’, European Foundation for the Improvement of Living and Working Conditions, Luxembourg: European Community. Grasland, C. (2005) ‘Spatial analysis tools and the measurement of territorial cohesion’, paper presented at the 1st ESPON Scientific Conference on ‘European Territorial research in progress’, Luxembourg, October 13–14, 2005. Gray, A. (2001) Definitions of Crowding and the Effects of Crowding on Health: A Literature Review, Wellington: Ministry of Social Policy. Hallett, G. (ed.) (1993) The New Housing Shortage: Housing Affordability in Europe and the USA, London and New York: Routledge. Hancock, K.E. (1993) ‘ “Can Pay? Won’t Pay?” or Economic Principles of “Affordability” ’, Urban Studies 30, 1, 127–145. Hantrais, L. (2004) Family Policy Matters: Responding to Family Change in Europe, Bristol: The Policy Press. Hanushek, E.A. and Kimko, D.D. (2000) ‘Schooling, labor-force quality, and the growth of nations’, American Economic Review 90, 5, December, 1184–1208. Harding, A., Phillips, B. and Kelly, S. (2004) ‘Trends in housing stress’, paper presented at National Summit on Housing Affordability, Canberra June 28, 2004. Harloe, M. (1991) ‘Social housing and the “Social Question”: Early housing reform and its legacy’, in Smith, M.P. (ed.), Breaking Chains. Social Movements and Collective Actions, New Brunswick: Transaction Publishers, 69–107. Heath, S. and Miret, P. (1996) ‘Living in and out of the parental home in Spain and in Great Britain: A comparative approach’, Working Paper Series, Cambridge Group for the History of Population and Social Structure, 2. Heidenreich, M. (2003) ‘Regional inequalities in the enlarged Europe’, Journal of European Social Policy 13, 313–333. Heidenreich, M. and Wunder, C. (2008) ‘Patterns of regional inequality in the enlarged Europe’, European Sociological Review 24, 1, 19–36. Hoffmeister, O. (2009) ‘The spatial structure of income inequality in the enlarged EU’, Review of Income and Wealth 55, 1, 101–127. Holdsworth, C. (2000) ‘Leaving home in Britain and Spain’, European Sociological Review 16, 2, 201–222. Hulchanski, J.D. (1995) ‘The concept of housing affordability: Six contemporary uses of the housing expenditure-to-income ratio’, Housing Studies, 10, 4, 471–491. Iacovou, M. (2002) ‘Regional differences in the transition to adulthood’, The Annals of the American Academy, 580, 1, 40–69. Iacovou, M. (2004) Patterns of family living, in Berthoud, R. and Iacovou, M. (eds.), Social Europe. Living Standards and Welfare States, Cheltenham: Edward Elgar, 21–45. Jacobs, J., and Gerson, K. (2004) The Time Divide: Work, Family, and Social Policy, Cambridge: Harvard University Press. Jacobzone, S. (1999) ‘Ageing and care for frail elderly persons: an overview of international perspectives’, OECD, Labour Market and Social Policy Occasional Papers, 38. Jacobzone, S., Cambois, E., Chaplain, E. and Robie, J.M. (1999) ‘The health of older persons in OECD countries: Is it improving fast enough to compensate for population ageing?’, OECD Labour Market and Social Policy Occasional Papers, 37. Jesuit, D., Rainwater, L. and Smeeding, T. (2002) ‘Regional poverty within the rich countries’, Luxembourg Income Study Working Paper Series, 318.

References

289

Jones, K. (2006) ‘Multilevel analysis: Practical applications’, Essex Summer Schools in SSDA, University of Essex. Jungblut, J.M. (2006) ‘The Erikson-Goldthorpe-Portocarero (EGP) classification applied to surveys with scarce information on occupation’, Mannheim Center for European Social Research Working Paper. Kalleberg, A.L. (2000) ‘Nonstandard employment relations: Part-time, temporary and contract work’, Annual Review of Sociology, 26, 341–365. Kangas, O. and Ritakallio, V.-M. (2004) ‘Relative to what? Cross-national picture of European poverty measured by regional, national and European standards’, Luxembourg Income Study Working Paper Series, 384. Keating, M. (1998), The New Regionalism in Western Europe. Territorial Restructuring and Political Change, Cheltenham: Edward Elgar. Kearns, A. and Parkinson, M. (2001) ‘The significance of neighbourhood’, Urban Studies 38, 12, 2103–2110. Kern, H. and Schumann, M. (1987) ‘Limits of the division of labour. New production and employment concepts in West German industry, in Economic and Industrial Democracy 8, 2, 151–170. Kiernan, K. (2002) ‘The state of European unions: an analysis of partnership formation and dissolution’, in Macura, M. and Beets, G. (eds.), Dynamics of Fertility and Partnership in Europe. Insights and Lessons from Comparative Research, Geneva: United Nations, 57–76. Kittay, E.F. (2001) ‘A feminist public ethic of care meets the new communitarian family policy’, Ethics 11, April, 523–547. Kurz, K. and Blossfeld, H.-P. (eds.) (2004) Home Ownership and Social Inequality in Comparative Perspective, Stanford: Stanford University Press. Kutty, N.K. (1999) ‘Determinants of structural adequacy of dwellings’, Journal of Housing Research 10, 1, 27–44. Kutty, N.K. (2005) ‘A new measure of housing affordability: Estimates and analytical results’, Housing Policy Debate 16, 1, 113–142. Lafortune, G., Balestat, G. and the Disability Study Expert Group Members (2007) ‘Trends in severe disability among elderly people: Assessing the evidence in 12 OECD countries and the future implications‘, OECD Health Working Papers, 26. Lamura, G., Burholt, V., Melchiorre, M.G., Quattrini, S., Balducci, C. and Spazzafumo, L. (2003) ‘The role of material security for quality of life in old age: Preliminary findings from an international survey on ageing well’, Revista Espanola de Geriatrìa e Gerontologìa 38(supl.1), 47. Land, H. and Rose, H. (1985) ‘Compulsory altruism for some or an altruistic society for all?’, in Bean, P., Ferris, J. and Whynes, D. (eds.), In Defense of Welfare, London: Tavistock, 74–96. Lash, S. and Urry, J. (1987) The End of Organized Capitalism, Cambridge: Polity Press. Layte, R. and Fouarge, D. (2004) ‘The dynamics of income poverty’, in Berthoud R. and Iacovou, M. (eds.), Social Europe. Living Standards and Welfare States, Cheltenham: Edward Elgar, 202–224. Layte, R. and Whelan, C.T. (2003) ‘Moving in and out of poverty’, European Societies 5, 2, 167–191. Layte, R. and Whelan, C.T. (2005) ‘Cumulative disadvantage or individualization? A comparative analysis of poverty risk and incidence’, EPAG Working Paper, 21. Lazarus, R.S. and Folkman, S. (1984) Stress, Appraisal and Coping, New York: Springer. Leibfried, S. (1993) ‘Towards a European welfare state?’, in Jones, C. (ed.), New Perspectives on the Welfare State in Europe, London: Routledge, 133–156.

290 References Leira, A. (1999) ‘Reflections on caring, gender and social rights’, paper presented at ESA Will Europe Work?, Amsterdam. Leira, A. and Saraceno, C. (2002) ‘Care: Actors, relationships, contexts’, in Hobson, B., Lewis, J. and Siim, B. (eds.), Contested Concepts in Gender and Social Politics, Cheltenham: Edward Elgar, 55–83. Leisering, L. and Leibfried, S. (1999) Time and Poverty in the Welfare State, Cambridge: Cambridge University Press. Lelkes, O. and Zolyomi, E. (2008) Poverty Across Europe: The Latest Evidence Using the EU-SILC Survey, European Centre for Social welfare Policy and Research, Vienna: Policy brief. Lenoir, R. (1974) Les exclus: un Français sur dix, Paris: Seuil. Lewis, J. (1992) ‘Gender and the development of welfare regimes’, Journal of European Social Policy 2, 159, 159–173. Lewis, J. (1993) ‘Women, work, family and social policies in Europe’, in Lewis, J. (ed.), Women and Social Policies in Europe. Work, Family and the State, Aldershot: Edward Elgar, 1–24. Lewis, J. (2002) ‘Gender and welfare state change’, in European Societies 4, 4, 331–357. Lewis, J. (2005) ‘The changing context for the obligation to care and to earn’, in Maclean, M. (ed.), Family Law and Family Values, Oxford: Hart Publ., 59–80. Lewis, J. (2006) Children, Changing Families and Welfare States, Cheltenham: Edward Elgar. Lewis, J. and Giullari, S. (2005) ‘The adult worker model family, gender equality and care: The search for new policy principles and the possibilities and problems of a capabilities approach’, in Economy and Society 34, 1, 76–104. Lewis, J. and Meredith, B. (1988) Daughters Caring for Mothers, London: Routledge. Livi Bacci, M. (2005) ‘Il paese dei giovani vecchi’, Il Mulino, 3, 409–421. Lucifora, C. (2000) ‘Working poor e redditi familiari: un’equazione sbagliata’, I Box della fondazione Rodolfo Debenedetti. Lyons, R.F. and Sullivan, M.J. (1998) ‘Curbing loss in illness and disability: A relationship perspective’, in Harvey, J. (ed.), Perspectives on Loss: A Sourcebook, Philadelphia: Brunner/Mazel. Lyons, R.F., Sullivan, M.J. and Ritvo, P. (1995) Relationships in Chronic Illness and Disability, London: Sage. Maitre, B., Nolan, B. and Whelan, C.T. (2005) ‘Welfare regimes and household packaging in the European Union’, Journal of European Social Policy, 15, 157–171. Maclennan, D., Stephens, M. and Kemp, P. (1997) Housing Policy in the EU Member States, Luxembourg: European Parliament. Maclennan, D. and Williams, R. (1990) Affordable Housing in Britain and America, York: Joseph Rowntree Foundation. Marpsat, M. (2005) ‘Beyond literal homelessness’, paper presented at 5th CUHP Workshop, Milan, April 21–22, 2005. Marsh, A., Gordon, D., Heslop, P. and Pantazis, C. (2000) ‘Housing deprivation and health: A longitudinal analysis’, Housing Studies 15, 3, 411–428. Melucci, A. (1996) The Playing Self: Person and Meaning in the Planetary Society, Cambridge: Cambridge University Press. Micheli, G.A. (1999) Cadere in povertà. Le situazioni a rischio, i processi, i terreni di coltura dell’impoverimento, Milano: Franco Angeli. Micheli, G.A. (2000) ‘Kinship, family and social network. The anthropological embedment of fertility change in Southern Europe’, Demographic Research 3, 13. Micheli, G.A. (2003) ‘La famiglia forte mediterranea alla prova dei grandi cambiamenti demografici e sociali’, in Micheli, G.A. and Ranci, C. (eds.), Equilibri

References

291

Fragili. Vulnerabilità e vita quotidiana delle famiglie lombarde , Milano: Guerini, 433–505. Micheli, G.A. (2004) ‘Kinship, family and social networks. The anthropological embedment of the demographic change in Southern Europe’, in Dalla Zuanna, G. and Micheli, G.A. (eds.), Strong Family and Low Fertility: A Paradox?, Dordrecht: Kluwer Academic, 77–104. Migliavacca, M. (2008) Famiglie e lavoro. Trasformazioni ed equilibri nell’Europa mediterranea, Milano: Bruno Mondadori. Modell, J. (1997) ‘What do life-course norms mean?’, Human Development 40, 282–286. Moisio, P. (2004) ‘A latent class application to the multidimensional measurement of poverty’, Quality and Quantity 38, 6, 703–717. Moreno Minguez, A. (2005) ‘The persistence of male breadwinner model in southern European countries in a compared perspective: Familism, employment and family policies’, MCFA Annals, IV. Moreno Minguez, A. and Crespo, E. (2005) ‘The persistence of male breadwinner model in the Southern European countries in a compared perspective: Familialism, employment and family policy’, paper presented at ESPANET Conference 2005. Murie, A. (1983) Housing Inequality and Deprivation, London: Heinemann. Myers, D., Baer, W.C. and Choi, S.Y. (1996) ‘The changing problem of overcrowded housing’, Journal of the American Planning Association 62, 1, 66–84. Naldini, M. (2003) The Family in the Mediterranean Welfare States, London: Frank Cass. Naldini, M. (2006) Le politiche sociali in Europa, Roma: Carocci. Navarro, C. and Ayala, L. (2008) ‘Multidimensional housing deprivation indices with application to Spain’, Applied Economics 40, 4–6, 597–612. Nussbaum, M.C. (2000) Women and Human Development, Cambridge: Cambridge University Press. OECD (1999) Employment Outlook, Paris: OECD. OECD (2000) Employment Outlook, Paris: OECD. OECD (2002) Employment Outlook, Paris: OECD. OECD (2004) Employment Outlook, Paris: OECD. OECD (2006a) Employment Outlook, Paris: OECD. OECD (2006b) Society at a Glance: OECD Social Indicators 2006 Edition, Paris: OECD. OECD (2008a) Employment Outlook, Paris: OECD. OECD (2008b) Growing Unequal? Income Distribution and Poverty in OECD Countries, Paris: OECD. Orloff, A. (1993) ‘Gender and the social rights of citizenship: The comparative analysis of gender relations and welfare states’, in American Sociological Review 58, 303–328. Österle, A. (2001) Equity Choices and Long-Term Policies in Europe, Aldershot: Ashgate. Paci, M. (2005), Nuovi lavori, nuovo welfare: sicurezza e libertà nella società attiva, Bologna: Il Mulino. Palier, B. (2000) ‘ “Defrosting” the French welfare state’, in West European Politics 23, 2, 113–136. Paugam, S. (1991) La disqualification sociale: essai sur la nouvelle pauvreté, Paris: PUF. Paugam, S. (ed.) (1996) L’exclusion, l’état des savoirs, Paris: Editions de La Découverte. Paugam, S. (1997) ‘Poverty and social disqualification: A comparative analysis of cumulative social disadvantage in Europe’, in Journal of European Social Policy 6, 4. Paugam, S. (2000) Le salarié de la précarité, Paris: PUF.

292

References

Pavolini, E. and Ranci, C. (2008) ‘Restructuring the welfare state: reforms in longterm care in Western European countries’, Journal of European Social Policy, 18, 3, 246–259. Pfau-Effinger, B. (1998a) ‘Gender cultures and the gender arrangement. A theoretical framework for cross-national gender research’, Innovation 11, 2, 147–166. Pfau-Effinger, B. (1998b) ‘Culture or structure as explanations for differences in parttime work in Germany, Finland and the Netherlands’, in O’Reilly, J. and Fagan, C. (eds.), Part-time Prospects: An International Comparison of Part-time Work in Europe, North America and the Pacific Rim, London: Routledge, 177–198. Pfau-Effinger, B. (1999) ‘The modernisation of family and motherhood in Western Europe’, in Crompton, R. (ed.), Restructuring Gender Relations and Employment: The Decline of the Male Breadwinner, Oxford: Oxford University Press, 60–79. Pfau-Effinger, B. (2005) ‘Culture and welfare state policies: Reflections on a complex interrelation’, in Journal of Social Policy 34, 3–20. Pierson, P. (ed.) (2001) The New Politics of the Welfare State, Oxford: Oxford University Press. Pizzuti, F.R. (ed.) (2006) Rapporto sullo stato sociale 2006, Novara: Utet. Power, A. (1993) Hovels to High Rise: State Housing in Europe since 1850, London: Routledge. Pringle, D.G, Walsh, J. and Hennessy, M. (1999) Poor People, Poor Places: The Geography of Poverty and Deprivation in Ireland, Taylorville: Oak Tree Press. Purcell, K. and Purcell, J. (1999) ‘In-sourcing outsourcing and growth of contingent labour as evidence of flexible employment strategies’, European Journal of Work and Organizational Psychology 7, 1, 39–59. Raudenbush, S.W. and Byrk, A.S. (1986) ‘A hierarchical model for studying school effects’, Sociology of Education 59, 1–17. Reher, D.S. (1998) ‘Family ties in Western Europe: Persistent contrasts’, Population and Development Review 24, 2, 203–234. Rokkan, S. (1970) Citizens, Elections, Parties, New York: McKay. Rosina, A. (2004) ‘Family formation in Italy: A cohort approach’, in Dalla Zuanna, G. and Micheli, G.A. (eds.), Strong Family and Low Fertility: A Paradox?, Dordrecht: Kluwer Academic Press, 23–43. Rosina, A. and Fraboni, R. (2004) ‘Is marriage losing its centrality in Italy?’, Demographic Research 11, 6. Rosina, A., Mencarini, L. and Rettaroli, R. (2005) ‘Alcuni primi risultati dell’indagine I.D.E.A’, paper presented at the Conference Famiglie, Nascite e Politiche Sociali, Accademia dei Lincei, Roma. Rubery, J., Smith, M., Fagan, C. and Grimshaw, D. (1998) Women and European Employment, London: Routledge. Sainsbury, D. (1996a) Gender, Equality and Welfare States, Cambridge: Cambridge University Press. Sainsbury, D. (1996b) ‘The male breadwinner model and women’s entitlements as wives’, in Sainsbury, D. (ed.), Gender Equality and the Welfare State, Cambridge: Cambridge University Press, 47–126. Saraceno, C. (1994) ‘The ambivalent familism of Italian welfare state’, Social Politics 1, 60–82. Saraceno, C. (ed.) (2008) Families, Ageing and Social Policy. Intergenerational Solidarity in European Welfare States, Cheltenham: Edward Elgar. Saraceno, C., Olagnero, M. and Torrioni, P. (2005) First European Quality of Life Survey: Families, Work and Social Networks, Dublin: European Foundation for the Improvement of Living and Working Conditions.

References

293

Sarasa, S. (2008) ‘Do welfare benefits affect women’s choices of adult care giving?’, in European Sociological Review 24, 37–51. Saunders, P. and Williams, P. (1988) ‘The constitution of the home: towards a research agenda’, Housing Studies 3, 2, 81–93. Scherer, S. and Reyneri, E. (2008) ‘Com’è cresciuta l’occupazione femminile in Italia: fattori strutturali e culturali a confronto’, Stato e Mercato 2, 183–216. Schmid, G. and Gazier, B. (eds.) (2002) The Dynamics of Full Employment: Social Integration Through Transitional Labour Markets, Cheltenham: Edward Elgar. Schmid, G. and Shömann, K. (2004) ‘Managing social risks through transitional labour markets: Towards a European social model’, TLM.NET Working Papers, 01. Scruggs, L. and Allan, J. (2006) ‘Welfare-state decommodification in 18 OECD countries: A replication and revision’, Journal of European Social Policy 16, 55–72. Schulz, R. and Beach, S.R. (1999) ‘Caregiving as a risk factor for mortality: The caregiver health effects study’, Journal of the American Medical Association, 282, 23, 2215–2219. Sen, A.K. (1985) Commodities and Capabilities, Amsterdam: North Holland. Sen, A.K. (1987) The Standard of Living, Cambridge: Cambridge University Press. Sen, A.K. (1992) Inequality Re-Examined, New York: Russell Sage Foundation. Sen, A.K. (1999) Development as Freedom, New York: Knopf. Sennett, R. (1998) The Corrosion of Character. The Personal Consequences of Work in the New Capitalism, New York: Norton. Settersten, R.A. and Mayer, K.U. (1997), ‘The measurement of age, age structuring and the life course’, Annual Review of Sociology 23, 233–261. Sleebos, J.E. (2003) Low Fertility Rates in OECD Countries: Facts and Policy Responses, OECD Social, Employment and Migration Working Papers n° 15. Available: http:// www.oecd.org/dataoecd/13/38/16587241.pdf. Snijders, T.A.B. and Bosker, R.J. (1999) Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modelling, Sage: London. Spiess, C.K., Iacovou, M., Robson, K.L. and Uunk, W. (2004) ‘Family effects on employment’, in Berthoud, R. and Iacovou, M. (eds.), Social Europe. Living Standards and Welfare States, Cheltenham: Edward Elgar, 69–98. Springer, S. (2000) ‘Homelessness: A proposal for a global definition and classification’, Habitat International 24, 4, 475–484. Stewart, K. (2002) ‘Measuring well-being and exclusion in Europe’s regions’, CASE paper, Centre for Analysis of Social Exclusion, London School of Economics. Stewart, K. (2003) ‘Monitoring social inclusion in Europe’s regions’, Journal of European Social Policy 13, 4, 335–356. Stewart, K. (2005) ‘Dimensions of well-being in EU Regions: Do GDP and unemployment tell us all we need to know?’, Social Indicators Research 73, 221–246. Stockard, J. and O’Brien, R.M. (2006), ‘Cohort variations in suicide rates among families of nations’, International Journal of Comparative Sociology 47, 1, 5–33. Stone, M.E. (1993) Shelter Poverty: New Ideas on Housing Affordability, Philadelphia: Temple University Press. Stone, M.E. (2006a) ‘A housing affordability standard for the UK’, Housing Studies 21, 4, 453–476. Stone, M.E. (2006b) ‘What is housing affordability? The case for the residual income approach’, Housing Policy Debate 17, 1, 151–184. Taylor-Gooby, P. (ed.) (2004a) New Risks, New Welfare, Oxford: Oxford University Press. Taylor-Gooby, P. (2004b) ‘New risks and social change’, in Taylor-Gooby, P. (ed.), New risks, New Welfare—The Transformation of the European Welfare State, Oxford: Oxford University Press, 1–29.

294

References

Thomas, C. (1993) ‘Deconstructing concepts of care’, Sociology 27, 4, 649–669. Till, M. (2005) ‘Assessing the housing dimension of social inclusion in six European countries’, Innovation: The European Journal of Social Sciences 18, 2, 153–181. Tomassini, C., Wolf, D. and Rosina, A. (2003) ‘Parental housing assistance and parent-child proximity in Italy’, Journal of Marriage and Family 65, 3, 700–715. Tomlinson, M., Walker, R. and Williams, G. (2008) ‘Measuring poverty in Britain as a multi-dimensional concept, 1991 to 2003’, Journal of Social Policy 37, 04, 597–620. Tosi, A. (ed.) (1994) La casa: il rischio e l’esclusione. Rapporto IRS sul disagio abitativo in Italia. Milano: Franco Angeli. Tosi, A. and Torri, R. (2005) ‘Extending the concept of homelessness’, paper presented at 5th CUHP workshop, Milan, April 21–22, 2005. Touraine, A. (2007) New Paradigm for Understanding Today’s World, Cambridge: Polity Press. Townsend, P. (1979) Poverty in the United Kingdom: A Survey of Household Resources and Standards of Living, Harmondsworth: Penguin Books. Twigg, J. (ed.) (1993) Informal Care in Europe, York: The University of York. Ungerson, C. (1983) ‘Why do women care?’, in Finch, J. and Groves, D. (eds.), Labour of Love: Women, Work and Caring, London: Routledge and Kegan Paul, 31–49. Vatsa, K.S. (2004) ‘Risk, vulnerability, and asset-based approach to disaster risk management’, International Journal of Sociology and Social Research 24, 10–11. Villa, P. (2004) ‘La diffusione del modello di famiglia a doppia partecipazione nei paesi europei e in Italia’, Inchiesta 146, 6–20. Vogel, J. (1999) ‘The European welfare mix: Institutional configuration and distributive outcome in Sweden and the European Union. A longitudinal and comparative perspective’, Social Indicators Research 56, 245–297. Vogel, J. (2002) ‘European welfare regimes and the transition to adulthood: A comparative and longitudinal perspective’, in Social Indicators Research 59, 275–299. Whelan, C.T., Layte, R., Maitre, B. and Nolan, B. (2001) ‘Income, deprivation, and economic strain. An analysis of the European Community Household Panel’, in European Sociological Review 17, 4, 357–372. Whelan, C.T. and Maitre, B. (2004) ‘Vulnerability and multiple deprivation. Perspectives on social exclusion in Europe: A latent class analysis’, EPG Working Papers, 52. Whelan, C.T. and Maitre, B. (2005) ‘Economic vulnerability, multidimensional deprivation and social cohesion in an enlarged European community’, in International Journal of Comparative Sociology 46, 3, 215–239. Whelan, C.T. and Maitre, B. (2007a) ‘Measuring material deprivation with EU-SILC: Lessons from the Irish survey’, in European Societies 9, 2, 147–173. Whelan, C.T. and Maitre, B. (2007b), The Dynamics of Economic Vulnerability: A Comparative European Analysis, Dublin: ESRI Working Paper N. 202. Whelan, C.T. and Maitre, B. (2008) “New and Old” Social Risks: Life Cycle and Social Class Perspectives on Social Exclusion in Ireland, Dublin: ESRI Working Paper N. 226. Wilthagen, T. and Tros, F. (2004) ‘The concept of “flexicurity”: A new approach to regulating employment and labour markets’, in Transfer 2, 166–186.

Index activity rate, 164–5 and caregiving, 177–8 female-, 10–11, 45, 49–50, 111, 223 ageing, 12–13 Alber, J., 172 Allan, J., 258 Anderson, R., 159 Anttonen, A., 161, 264 Apospori, E., 58 Arum, R., 91 Atkinson, T., 28, 126 Attias-Donfut, C., 172 Avramov, D., 134, 140, 142 Ayala, L., 133, 144 Baer, W.C., 138 Bagnasco, A., 30 Balchin, N.P., 133, 139 Baldini, M., 86 Balestat, G., 162, 184 Bambra, C., 259 Barbagli, M., 193 Barbieri, P., 89, 91 Bardasi, E., 243 Beach, S.R., 176 Beck, U., 4, 17, 88, 273 Bengtson, V.L., 38 Bernhardt, E., 191 Berthoud, R., 27, 29, 243, 256, 274, 275–6 Betti, G., 140, 142 Billari, F.C., 191, 192, 217 Blossfeld, H.P., 37, 127, 189 Boarini, R., 85, 86 Bologna, S., 91 Bosker, R.J., 265, 266 Bourdieu, P., 158, 251 Brandolini, A., 276 Brenner, N., 26 Brousse, C., 133 Bryan, M., 243 Burchardt, T., 162 Burholt, V., 13 Bury, M., 159 Byrk, A.S., 265

Cambois, E., 160 Cantillon, B., 28, 126 care, 11–14, 19, 21, 159–61, 171–4, 175–84, 225–9 caregivers, 12–13, 174–81 childcare, 9–11, 49–53, 223–4 regimes, 161, 169–71 and vulnerability, 175–81 and work, 177–8 Castel, R., 4, 5, 17, 158, 219 Castiglioni, M., 193 Castles, F.G., 29, 34, 209, 270, 274 Castro Martin, T., 191 Chambaz, C., 273 Cherlin, A.J., 110 Choi, S.Y., 138 Christian, J., 132 Chung, H., 88, 122 Cioni, E., 172 Clanché, F., 133 Corijn, M., 190 Costa, G., 176 Crane, M., 134 Crespo, E., 110 Crompton, R., 223, 234, 251, 254 Crouch, C., 3, 11, 39, 253, 271, 280 Cullingworth, J.B., 138 D’Addio, A.C., 223 Dale, A., 133 Dalla Zuanna, G., 193 de Beer, J., 191 Del Boca, D., 11, 46, 50, 52, 57 demographic patterns, 190–4 dependency, 11–14, 159–71, 181–4, 225–6 definition of, 161–3 and vulnerability, 163–9 d’Ercole, M.M., 7, 10, 11, 62, 85, 86, 223 Deven, F., 191 Dewilde, C., 134 Dodgeon, B., 133 Doherty, J., 220 Dovey, K, 127 Drobnic, S., 37 295

296 Index Edgar, B., 132, 133, 220 EIRO (European Industrial Relations Observatory on-line), 123 elderly, 12–13, 42, 152–3 Ellingsaeter, A.L., 47 employment female, 9–11, 45–6 and fertility, 49–50 temporary, 7–9 unstable, 91–101, 108–9 and work flexibility, 88–90 Erikson, R., 249, 258 Esping-Andersen, G., 4, 5, 11, 15, 21, 24, 26, 29, 63, 89, 96, 193, 208, 209, 223, 251–2, 255, 258–9, 264 ESPON, 238, 262 eurobarometer, 172 EUROFAMCARE Consortium, 160 European Commission (EC), 96, 123 European Foundation for the Improvement of Living and Working Conditions, 160, 190, 200 Eurostat, 9, 10, 12, 91, 99, 111, 126, 131, 159, 162, 184 Fagan, C., 46 family, 11 demographic composition of, 41–2 formation of, 190–2 gender arrangements in the, 49–53 income structure of, 45–9 models, 37–8, 53–5 solidarity, 5, 12–14, 38–40, 171–4, 192–4 types, 42–5, 46–9 vulnerability and family models, 234–6, 253–5, 271–3 Ferrera, M., 25, 26, 63, 192, 209, 252, 264, 282 Finch, J., 172 Fitzpatrick, S., 132 Folkman, S, 176 Forrest, R., 127 Forster, M., 7, 10, 11, 62 Fouarge, D., 6, 59, 70, 73, 249 Fraboni, R, 193 Freedman, V., 160, 184 Frick, J., 87 Friedrichs, J., 127 Fuà, G., 65

Fujiura, G.T., 161, 162 functioning, 18–21 Galster, G.C., 127 Gazier, B., 122 generation inter generational exchange, 13, 38–9, 42, 44–5, 172–4, 175, 208–11, 253 turnover, 13 Gerson, K., 223 Giorgi, L., 140 Giullari, S., 11 Goldstein, H., 265, 278 Goldthorpe, J.H., 249, 251, 258 Gordon, D., 144 Graham, H., 176 Grammenos, S., 159, 225 Grasland, C., 31 Gray, A., 134, 137 Grimshaw, D., 46 Hancock, K.E., 134 Hantrais, L., 172 Hanushek, E.A., 270 Harding, A., 150 Harloe, M., 133 Heath, S., 110 Heidenreich, M., 27, 29, 256, 274 Hemerijck, A., 252, 282 Hennessy, M., 225 Heslop, P., 144 Hoffmeister, O., 275 Holdsworth C, 110 housing affordability of, 134–7 deprivation, 7, 126–8, 132–46, 197–9, 220–1 overcrowding, 137–9 ownership, 128–32 physical quality of, 139–42 regimes, 146–50 and vulnerability, 150–3 Hox, J.J., 278 Hulchanski, J.D., 135 Iacovou, M., 50, 194 income inequality, 27, 62–7, 83, 244–6 instability, 6–7, 75–83 mobility, 70–3 support, 169–71

Index 297 individualization, 38–40 inequalities old and new, 244–8 territorial, 25–9, 62–7, 92–8, 208–16, 236–43, 261–4, 274–7 insecurity, 3–4, 244, 279 job insecurity, 4–6, 7–9, 96 instability, 17–18, 281–2 work instability, 90–1, 105–8, 116–19

Leisering, L., 24, 58 Lelkes, O., 276 Lenoir, R., 158 Lewis, J., 11, 37, 38, 46, 47, 50, 176, 177, 253, 254, 273 Liefbroer, A., 192 Livi Bacci, M, 193 Lucifora, C, 116 Lyons, R.F., 159, 180

Jacobs, J., 223 Jacobs, T., 176 Jacobzone, S., 12, 160 Jesuit, D., 27, 34 Johansson, A., 85, 86 Jones, K., 265, 266 Jungblut, J.M., 249

Maas, M.J.M., 278 Maclennan, D., 135, 139 macro-regions, 29–31, 32–4, 40–1 disparities among, 60–5, 93–5, 146–53, 211–12, 236–43, 256–7, 258–9, 261–4, 270, 274–6 Maître, B., 58, 83, 87, 142, 220, 226, 243 male breadwinner family (MBW), 37–40, 45–9, 50–3, 53–5, 234–6, 267, 271–3 Marlier, E., 28, 126 Marpsat, M., 133 Marsh, A., 144 Martin, L, 160, 184 Mayer, K.U., 217 Meert, H., 133, 220 Melchiorre, M.G., 13 Melucci, A., 251 Mencarini, L., 110 Meredith, B., 176 Micheli, G.A., 38, 39, 110, 193, 201, 210, 217 Michielin, F., 191 Migliavacca, M., 123 Millar, J., 58 Mills, M., 189 Mira d’Ercole, M., 7, 10, 11, 62, 85, 86 Miret, P., 110 Mitchell, D., 209 Modell, J., 217 Moisio, P., 133 Moreno Minguez, A., 110 Muffels, R.J.A., 87 Muller, W., 91 Murie, A., 132 Musterd, S., 127 Myers, D., 138

Kalleberg, A.L., 90 Kangas, O., 27 Kearns, A., 127, 134 Keating, M., 26 Kelly, S., 150 Kemp, P., 139 Kern, H., 273 Kiernan, K., 191 Kimko, D.D., 270 kinship network, 279–80 support, 5, 13 Kittay, E.F., 180 Klijzing, E., 190 Köhler, U., 172 Kurz, K., 127 Kutty, N.K., 134, 135, 154 labour market, 4–5, 7–9, 45–6, 89–90, 92–8, 101–5 and caregiving, 177–8 exclusion from the, 105–8 Lafortune, G., 162, 184 Lamura, G., 13 Land, H., 176 Lash, S., 3 Layte, R., 6, 58, 59, 70, 73, 75, 249 Lazarus, R.S., 176 leaving home, 189–201 regional differences in, 208–16 and vulnerability, 201–16 Leibfried, S., 24, 58, 209 Leira, A., 160, 172, 176

Naldini, M., 37, 50 Navarro, C., 133, 144

298

Index

Nolan, B., 87 Nussbaum, M. C, 184 Obinger, H., 29, 34, 270, 274 O’Brien, R.M., 209 OECD, 60, 62, 78, 85, 86, 87, 88, 91, 92, 96, 121, 122, 123, 124, 155, 165, 184, 259 Ogg, J., 172 Olagnero, M., 39 Ongaro, F., 191 Orloff, A., 273 Österle, A., 160, 172, 225 Paci, M, 88 Palier, B., 22 Pantazis, C., 144 Parkinson, M., 134 Paugam, S., 6, 158, 243 Pavolini, E., 254 Pfau-Effinger, B., 37, 38, 46, 254 Phillips, B., 150 Pierson, P., 26, 252 Pizzuti, F.R., 193 poverty, 6–7, 63, 65–7, 73–5, 83–5, 116–19, 196–9 anti-poverty policy, 170–4 longitudinal, 70–3, 75–7 persistence of, 6–7, 67–70, 77–83 regional, 27, 28–9, 276 risk of, 10–11, 13 studies, 28, 58–9, 256 and transition to adulthood, 201–8 and vulnerability, 220, 221, 222, 223, 224, 225, 226, 243–4 Power, A., 139 Pringle, D.G., 225 Purcell, J., 88 Purcell, K., 88 Putnam, R. Quattrini, S., 13 Rainwater, L., 27 Ranci, C., 254 Raudenbush, S.W., 265 Raymond, M.E. Reher, D.S., 42, 110, 209, 210–11, 214, 216 Rettaroli, R., 110

Reyneri, E., 88, 90, 91, 110 Rhodes, M., 252 Ritakallio, M., 27 Ritvo, P., 180 Robie, J.M., 160 Robson, K.L., 23 Rokkan, S., 26 Rose, H., 176 Rosina, A., 110, 191, 193 Rossi, G., 110 Rubery, J., 46 Rutkowski-Kmitta, 161, 162 Sainsbury, D., 254 Saraceno, C., 172, 193, 253 Sarasa, S., 178 Saunders, P., 127 Scabini, E., 110 Scherer, S., 110 Schmid, G., 88, 122 Schneider, U.A., 23 Schulz, R., 176 Scruggs, L., 258 Sen, A.K., 18, 158, 184, 251 Sennett, R, 251, 273 Settersten, R.A., 217 Shömann, K., 88 Shumann, M., 273 Sipilä, J, 161, 264 Sleebos, J.E., 223 Smeeding, T., 27 Smith, M., 46 Snijders, T.A.B., 265, 266 social exclusion, 158, 232–3 social protection, 5–6, 21–3, 192–4, 252–3, 275–7 social risk, 3–6, 58 analysis, 16 interactions among, 229–33 and negative outcomes, 14–18 new social risks, 6–14, 219–26 social policies contrasting, 21–3, 252–3, 259–64 Spazzafumo, L., 13 Spiess, C.K., 23 Springer, S., 133 Stephens, M., 139 Stewart, K., 28, 29, 34, 237 Stockard, J., 209 Stone, M.E., 135, 154 Sullivan, M.J., 159, 180

Index 299 Taylor-Gooby, P., 5, 14, 21, 26, 58, 88, 189, 193, 252, 259, 264, 276 Tentschert, U., 140 territorial social formation, 30–1, 32–4 Thomas, C., 176 Till, M., 134 Toharia, L., 92 Tomassini, C., 193 Tomlinson, M., 134 Torri, R., 133 Torrioni, P., 39 Tosi, A., 128, 133 Toso, S., 86 Touraine, A., 271 Townsend, P., 158 transition to adulthood, 189–216 changes in the, 190–2 regional variations in the, 192–6 and vulnerability, 196–9, 201–16 Tros, F., 89, 90 Twigg, J., 176 unemployment, 7–9, 101–5, 105–8, 112–16, 221–2 Ungerson, C., 176 Urry, J., 3 Uunk, W., 87 Vatsa, K.S., 16 Verma, V.K., 140, 142 Villa, P., 110 Vogel, J., 192, 194

vulnerability concept of, 14–18 dimensions of, 14–21, 219–26 factors of, 250–7, 264–77 geographical distribution of, 236–43 and household models, 234–6 multidimensional analysis of, 226–33 and social class, 244–8 Walker, R., 134 Walsh, J., 225 Warnes, A.M., 134 welfare system, 5–6, 21–3, 251–3, 258–61, 262–4, 275–7 regimes and vulnerability, 62–3, 192–4, 241–3, 275–7 Wetzel, C., 11 Whelan, C., 6, 58, 70, 75, 83, 87, 126, 133, 142, 220, 226, 243 WHO (1980), 161, 184 Williams, G., 134 Williams, M., 133 Williams, P., 127 Williams, R., 135 Wilthagen, T., 89, 90 Wolf, D., 193 Wolff, F.C., 172 work/family models, 110–16, 122 and vulnerability, 116–19 work instability, 90–1 Wunder, C., 27, 256, 274 the Young, 41–2, 97–8, 101, 189–218 Zolyomi, E., 276